Decreased Expression of the Fractalkine Receptor CX3CR1 on Circulating Monocytes as New Feature of Sepsis-Induced Immunosuppression

Septic syndromes (i.e., sepsis, severe sepsis, and septic shock) represent a dramatic problem in critically ill patients. Their incidence has increased continuously over the last two decades (1). Meanwhile, despite major advances in supportive therapy and the use of potent antibiotics, mortality has only modestly decreased. Nowadays, the overall mortality rate for septic syndromes is estimated at 28% in the U.S. (2). Among them, septic shock represents the most severe stage with a mortality ranging from 40 to 60%, accounting for >135,000 deaths per year in Europe and in the U.S. (2, 3).

Septic shock is characterized by major perturbations of the immune homeostasis. Initially, patients develop a tremendous systemic inflammatory response (leading to refractory arterial hypotension, shock, and multiple organ dysfunctions), immediately followed by an antiinflammatory process, acting as a negative feedback (1, 4, 5). This compensatory inhibitory response may become deleterious as nearly all immune functions are compromised. Indeed, it is speculated that a high proportion of septic shock-related deaths occur in a delayed immunosuppressive state (4, 6), probably in relation with patients’ inability to kill invading microorganisms and their increased risk of contracting nosocomial infections (7, 8, 9). In line, clinical studies have shown that the spontaneous recovery of a functional proinflammatory response after septic shock is associated with survival (6, 10, 11). However, much remains to be learned about the pathophysiology of this state of immunosuppression, and the characterization of its mechanistic bases remains a major challenge to improve care for critically ill patients.

Based on a microarray study, we recently identified a panel of 28 genes whose peripheral blood mRNA expression could efficiently discriminate survivor from nonsurvivor septic shock patients (12). Numerous genes in this list suggested a better recovery of immune functions in patients who were going to survive. Among the genes up-regulated in survivors compared with nonsurvivors, the chemokine receptor CX3CR1 showed the highest factor of change (≈8-fold increase in comparison with nonsurvivors). CX3CR1 is a seven transmembrane-spanning G protein-coupled receptor expressed on monocytes, NK cells, and some lymphocyte subpopulations (13). CX3CR1 ligand (CX3CL1), also called fractalkine, is the sole member of the CX3C-chemokine subfamily. The molecule exists in two forms, membrane anchored or shed chemotactic soluble glycoprotein (14, 15). The soluble CX3CL1 (sCX3CL1)² exerts a potent chemotactic activity (16). The membrane-bound CX3CL1 is expressed on endothelial cells upon stimulation by proinflammatory cytokines (17, 18, 19) and induces leukocyte adhesion and transmigration into injured tissue (16, 20, 21). Interestingly, CX3CR1 is preferentially expressed on proinflammatory, Th1, and/or cytotoxic cells, and its interaction with CX3CL1 has been proposed as an amplification process for polarized Th1/proinflammatory response (17, 22, 23, 24).

The goal of the current study was to further investigate the expression of CX3CR1 during septic shock. We examined its expression in circulating leukocytes focusing on monocytes, which are known not only to express high levels of CX3CR1 but also to play a central role in septic shock pathophysiology (5). We observed that CX3CR1 expression (both mRNA and protein) was severely down-regulated in monocytes from septic shock patients. Importantly, nonsurvivors presented with significantly sustained lower expression in comparison with surviving patients, confirming our microarray data. Most importantly, this down-regulation was reproduced by the incubation of PBMCs from healthy individuals with LPS, whole bacteria (Escherichia coli and Staphylococcus aureus), and, to a lower extent, with corticosteroids. rIL-10 and sCX3CL1 had no effect. We thus propose that the decreased CX3CR1 expression observed in patients represents a new feature of sepsis-induced immunodepression.

Patients older than 18 years of age admitted to two intensive care units (ICUs) of a university hospital for septic shock were included. The protocol was approved by the local Ethical Committee for Clinical Research, and all patients or relatives gave their informed consent. The diagnosis of septic shock was based on ACCP/SCCM criteria (25). The onset of the septic shock was defined as the beginning of vasopressor therapy. All patients were treated similarly according to the standardized recommendations of our ICU. Severity was assessed using the simplified acute physiologic score II (SAPS II) (26). Mortality was defined as death occurring within 28 days after the onset of shock. For flow cytometry analysis, serial blood samples were collected every 2 or 3 days until day 14. For all the other analyses, a first blood sample was collected during the first 72 h following the onset of shock and a second one was collected between day 4 and 10. Eighty-nine healthy individuals (age: 51 ± 5 years; 42 females and 47 males) with no known comorbidities were also included after informed consent was given, to provide panels of control values for mRNA quantification, ELISA measurements, and flow cytometry analysis.

PBMCs were isolated by means of Ficoll-Plaque density gradient centrifugation (Amersham Biosciences) and washed with PBS while the remaining RBC were lysed using the BD Pharm Lyse buffer (BD Biosciences). For some experiments, monocytes were isolated by positive selection (based on CD14 expression) using MACS isolation kits (Miltenyi Biotech). Cells (PBMCs or monocytes) were cultured in serum-free medium BioWhittaker X-Vivo 20 (Cambrex) in 24-well ultra-low attachment plates (Corning-Costar). The number of cells per well was adjusted to 2 × 10⁶ cells/ml. Cells were treated or not with LPS from E. coli O55:B5 (10 ng/ml; Sigma-Aldrich), heat-killed E. coli (American Type Culture Collection number: 11775; 10⁷ CFU/ml), heat-killed S. aureus (ATCC number: 12600; 10⁷ CFU/ml), dexamethasone (10⁻⁶ M; Merck), recombinant human sCX3CL1 (10 ng/ml; R&D Systems), or IL-10 (10 or 1 ng/ml; R&D Systems) for 3, 6, and 24 h at 37°C in a humidified 5% CO₂ incubator. Cells were used to extract RNA as described below.

Peripheral blood mRNA expression was measured using samples collected directly in PAXGene Blood RNA tubes (PreAnalytix). Total RNA was extracted using the PAXGene Blood RNA kit (PreAnalytix). Regarding cell culture (PBMCs and purified monocytes), total RNA was extracted using RNAlater reagent (Ambion) and RNeasy kits (Qiagen). For each RNA extraction, the residual genomic DNA was digested using the RNase-Free DNase set (Qiagen). The integrity and quality of the total RNA was assessed on an Agilent 2100 Bioanalyzer (Agilent) by means of the RNA 6000 Nano assay (Agilent).

Total RNA was reverse transcribed into cDNA using ThermoScript RT-PCR system (Invitrogen) according to the manufacturer’s instructions. mRNA expression was quantified using quantitative real-time PCR as previously described (27). Briefly, PCR reactions were performed using a LightCycler instrument with the Fast-Start DNA Master SYBR Green I real-time PCR kit according to the manufacturer’s instructions (Roche). Thermocycling was performed in a final volume of 20 μl containing 3 mM MgCl₂ and 0.5 μM each of the required primers. PCR was performed with an initial denaturation step of 10 min at 95°C, followed by 40 cycles of a touch-down PCR protocol (10 s at 95°C, 10 s annealing at 68–58°C, and 16 s extension at 72°C). mRNA expression of CX3CL1 and the housekeeping gene Peptidylpropyl isomerase B (PPIB) encoding for cyclophilin B were investigated using specific cDNA standards and ready-to-use primer mixes obtained from Search-LC. The cDNA standard for CX3CR1 (accession number: NM_001337) was prepared from purified PCR amplicons obtained with the following primer combinations: CX3CR1: forward primer: 5′-TGACTGGCAGATCCAGAGGTT-3′, reverse primer: 5′-GTAGAATATGGACAGGAACAC-3′ (amplicon size: 163 bp) as previously described (27). The LightCycler 2.0 was used to determine the crossing point for individual samples. Serial dilutions of the cDNA standard were prepared in quadruplicate to generate standard curves. Relative standard curves describing the PCR efficiency of target genes and PPIB were created and used to perform efficiency-corrected quantification with the LightCycler Software version 4. The results were expressed as a concentration ratio between the target gene mRNA and PPIB mRNA levels. The efficiency of PPIB mRNA as a reference for target mRNA quantification in humans has been previously described (28).

Microarray experiments were performed as previously described (29). Briefly, 100 ng of total RNA were used to prepare double-stranded cDNA containing the T7 promoter sequence using the Two-Cycle Target Labeling assays (Affymetrix). cRNA was synthesized and labeled with biotinylated ribonucleotide by in vitro transcription using the T7 promoter coupled double stranded cDNA as template and the GeneChip IVT labeling kit (Affymetrix). A second round of cDNA synthesis and in vitro transcription reactions was performed as described by the GeneChip Eukaryotic Small Sample Target labeling assay version II (Affymetrix). The fragmented cRNA was hybridized onto HG-U133 PLUS 2.0 oligonucleotide arrays (Affymetrix). The arrays were washed and stained according to the Affymetrix protocol EukGE-WS2v4 using an Affymetrix fluidic station FS450. The array was scanned with the Agilent G2500A GeneArray scanner. The perfect match only model robust multichip average quantil normalization (30, 31) was conducted using BioConductor package Affy_1.2.30 (Bioconductor project; http://www.bioconductor.org). Expression levels were expressed as 2e^{robust multichip average expression level}.

Flow cytometric (FC500; Beckman Coulter) expression of cell surface CX3CR1 was assessed on peripheral whole blood collected on EDTA anticoagulant tubes. mAbs were used according to the manufacturer’s recommendations: PE-labeled anti-CX3CR1 (clone 2A9–1; Medical & Biological Laboratories), FITC-labeled anti-CD14, and PC5-labeled anti-CD16 (Immunotech). The red cells were lysed with Versalyse reagent (Beckman Coulter). Monocytes were characterized based on their CD14 expression, proinflammatory monocytes on their CD14^dim/CD16^bright coexpression. Monocyte HLA-DR (mHLA-DR) measurements were obtained as previously described (6).

CD11b expression was measured as previously described (32). This protocol permitted functional testing with septic cells (available in small amount) without cell purification required for cell migration experiments. Leukocytes (100 μl of whole blood) diluted in PBS were incubated with chemokines (0.2, 2, and 20 ng/ml; sCX3CL1 and MCP-1 from R&D Systems) for 10 min at 37°C. The incubations were terminated by the addition of ice-cold PBS and centrifugation. The cells were then incubated for 30 min at 4°C with a mixture of anti-CD11b and anti-CD14. Finally, red cells were lysed with the FACS lysing solution (30-min incubation, 4°C). Finally, after washing, cells were resuspended in PBS containing 4% formaldehyde for immediate flow cytometry acquisition. After positive-CD14 monocyte gating, the measurement of CD11b expression was expressed as mean fluorescence intensity (MFI).

sCX3CL1 was measured by ELISA using capture and detection of goat IgG anti-CX3CL1 Abs and recombinant human CX3CL1 as a standard (all reagents from R&D Systems). The standard curve was prepared using recombinant human CX3CL1 (range: 8–1000 pg/ml). The absorbance was measured at a 450 nm wavelength (Dynatech Laboratories).

Data are presented either as the mean ± SEM or as box and whisker plots, with representation of the median, 75^th and 90^th percentiles, and outliers. Comparisons between groups were made with the nonparametric Mann-Whitney U test for continuous variables without correction on the number of tests performed whereas categorical variables were compared with the Pearson χ² test or Fisher exact test when appropriate. The Wilcoxon nonparametric paired test was used to compare data from the same group of patients measured at different time points. CX3CR1 mRNA level was assessed as a prognostic marker using receiver operating characteristics curve and its area under the curve. The results were considered significant at a values of p < 0.05.

Fifty-eight patients with septic shock were enrolled in the study. The demographic and clinical characteristics of the cohort are presented in Table I. The mean value of the SAPS II at diagnosis (mean = 51) and the Sequential Organ Failure Assessment Score at days 1–2 (mean = 10) illustrate the high severity of the septic shock patients studied in this cohort. Mortality in the overall population was 38%. As indicated in Table I, every patient exhibited decreased mHLA-DR expression both at days 1–3 and 4–10 (median values < 50%; control values > 90% (33)) illustrating the development of a state of immunosuppression.

To validate our previous microarray results (12), PAXgene blood samples were obtained from 47 additional septic shock patients. By using quantitative real-time PCR, we observed a major and significant down-regulation of CX3CR1 mRNA in circulating cells from septic shock patients in comparison with healthy individuals (p < 0.0001). Most importantly, and in accordance with our preliminary microarray study (12), we observed that CX3CR1 mRNA expression was significantly higher in survivors in comparison with nonsurvivors at both time points (p < 0.005; Fig. 1,A). No significant difference was observed between patients dying early after the onset of shock (<48 h) and those with a late death (data not shown). Regarding the time-course analysis of paired samples, a significant decrease in the circulating cell CX3CR1 mRNA level was observed in nonsurvivors (Wilcoxon paired test p = 0.03), while it remained stable in survivors. CX3CR1 mRNA level measured at days 1–3 was assessed as a prognostic marker for mortality using a receiver operating characteristics analysis (Fig. 1 B). The area under the curve was found at 0.88. To predict unfavorable outcome, the best threshold for CX3CR1 mRNA level (i.e., with maximized sensitivity and specificity) was calculated at 0.00135. Using this value, nonsurviving patients could be identified with a sensitivity and specificity of 88 and 78%, respectively.

To determine whether the transcriptional inhibition of CX3CR1 expression was associated with alterations in its protein expression, we assessed CX3CR1 on circulating cells from 22 consecutive patients using flow cytometry (17 survivors and 5 nonsurvivors). We focused on monocytes, as they are both key players in septic shock pathophysiology and high-expressing CX3CR1 cells (16). In agreement with the mRNA results, the patients displayed a decrease in CX3CR1 expression in comparison with healthy individuals (p < 0.0001). When the patients were stratified according to mortality, monocyte CX3CR1 expression remained significantly decreased in survivors and each group (i.e., survivors and nonsurvivors) compared with healthy individuals (p < 0.0001). Although no significant difference was observed between groups (Fig. 2,A), nonsurviving patients showed, as observed at the mRNA level, a significant decrease in CX3CR1 expression on monocytes between days 1–3 and 4–10 (Wilcoxon paired test p = 0.018), whereas survivors did not. Because proinflammatory monocytes overexpressing CX3CR1 can be identified on the basis of increased CD16 expression (associated with a diminished CD14 expression) (22, 34), we then monitored CX3CR1 down-regulation on this specific monocyte subpopulation. As previously reported (35), we observed an increased percentage of CD14^dimCD16^high monocytes in septic shock patients in comparison with healthy individuals (mean ± SEM: 29 ± 5.4% vs 4.4 ± 1.2%). Both monocyte subsets (i.e., CD16^high and CD16^negative) were significantly affected by the loss of CX3CR1 expression (Fig. 2,B). However, this expression remained higher on the proinflammatory subset of monocytes than on their CD14^highCD16^negative counterparts (Fig. 2 B).

To assess whether this loss of CX3CR1 might impact monocyte functionality, we investigated CD11b up-regulation, which constitutes a key step in monocytes chemotaxis and in their infiltration into tissues (36). As shown in Fig. 3,A, fractalkine clearly increased the expression of CD11b on monocytes from healthy controls. In contrast, it had no effect on cells from septic patients. In these experiments we used MCP-1 as a positive control because this molecule is known for having very potent chemotactic effect on monocytes. Interestingly, as observed with fractalkine, cells from septic patients did not respond to MCP-1 challenge, whereas we observed a marked elevation of CD11b expression (∼70% above baseline) in control monocytes (Fig. 3 B). Collectively, these data indicated that the decrease in monocyte CX3CR1 expression was associated with a loss of functionality upon fractalkine challenge.

In an inflammatory context, the membrane-bound CX3CL1 is highly expressed by endothelial cells and can be shed as a soluble chemotactic form by proteolytic cleavage. Therefore, we measured sCX3CL1 concentration in the serum of patients with septic shock and healthy individuals. We observed a significant increase in septic shock patients (p < 0.0001). When patients were stratified according to mortality, sCX3CL1 levels remained significantly higher in both groups (i.e., survivors and nonsurvivors) in comparison with healthy individuals (p < 0.0001; Fig. 4), and higher sCX3CL1 concentrations were observed in nonsurvivors in comparison with survivors. As CX3CL1 can be produced by certain type of cells circulating in small numbers in the bloodstream (i.e., dendritic cells, endothelial cells, and macrophages) (37), we therefore measured CX3CL1 mRNA levels in peripheral blood using quantitative real-time PCR. We observed very low and heterogeneous concentrations in patients, close to the quantitative real-time PCR detection threshold. However, our results tended to indicate a slight decrease in CX3CL1 mRNA expression during septic shock (Fig. 5). No difference was observed between survivors and nonsurvivors at any time point (data not shown).

Because monocyte deactivation observed after a septic insult can be reproduced in vitro using low doses of endotoxin (38, 39), we therefore examined the effect of LPS on CX3CR1 mRNA expression. We studied the effect of LPS on purified CD14⁺ monocytes and PBMCs from healthy individuals. A challenge with 10 ng/ml LPS led to a strong and significant down-regulation of monocyte CX3CR1 mRNA expression at 6 and 24 h (% relative to control samples at 6 h: 32.2 ± 13.2%; and at 24 h: 12.2 ± 5%) (Fig. 6,A). This down-modulation was also observed, although to a lower extent, in the LPS primed-PBMCs after 24 h (21.2 ± 4.9% of control value) (Fig. 6,A). Similar experiments were performed with heat-killed Gram⁺ (S. aureus) or Gram⁻ (E. coli) bacteria. As observed with LPS, both types of bacteria induced a significant decrease in CX3CR1 mRNA expression in PBMCs from healthy individuals (Fig. 6 B).

To understand which mechanisms may sustain this down-regulation, we performed microarray analysis on LPS-stimulated monocytes at 24 h. We observed a transcriptional induction of several endogenous negative regulators of TLR signaling concurrently with the decreased expression of CX3CR1 mRNA (Table II). Additional investigations are now warranted to delineate the precise signaling pathway regulating CX3CR1 expression in this model.

As IL-10 and cortisol are both known to be increased in septic shock patients and to be involved in sepsis-induced monocyte deactivation, we studied their effect (with dexamethasone reproducing the effect of endogenous cortisol) on CX3CR1 expression in vitro (10, 40). IL-10 used at 10 ng/ml induced a slight, although nonsignificant, decrease in CX3CR1 mRNA expression, whereas the lower concentration tested (1 ng/ml) had no effect. In contrast, dexamethasone had a significant negative effect on CX3CR1 mRNA expression after 24 h of culture (Fig. 7). We also assessed the effect of sCX3CL1 on its own receptor. However, as shown in Fig. 7, we did not observe any effect of sCX3CL1 on CX3CR1 mRNA expression.

It is now agreed that septic shock patients rapidly present with severely depressed immune functions. The condition includes enhanced leukocyte apoptosis (41, 42), defective lymphocyte proliferation in response to recall Ags or mitogens (43, 44), marked elevation of the percentage of circulating regulatory CD4⁺CD25⁺ T cells (45), and deactivated monocyte functions. Monocytes are one of the main effectors of innate immunity against infection. They have the capacity to phagocyte microorganisms, to sense microbial products, and, in response, to release a large number of inflammatory mediators that contribute to defense against infection. As APCs they represent a link with adaptive immunity by their capacity to induce specific T cell activation (46). However, after the onset of septic shock, monocytes have been shown to rapidly exhibit an impaired production of proinflammatory cytokines in response to additional bacterial challenge (also known as endotoxin tolerance) (47), and a reduced Ag presentation capacity likely due to their decreased expression of HLA-DR (43). Most importantly, it has been observed that these alterations develop to a larger extent in nonsurviving patients in comparison with survivors. Thus, the comprehension of the underlying molecular mechanisms responsible for monocyte deactivation may be an important way to define new strategies for restoring immune functions in this deadly clinical situation.

In a recent microarray study performed in septic shock patients, we observed that circulating cell CX3CR1 mRNA expression was strongly decreased in nonsurvivors while their immune functions were severely depressed (>48 h after the onset of shock and as established by the decreased mHLA-DR expression observed in patients) (12). As CX3CR1 is known to be highly expressed on monocytes and as it was the gene presenting with the highest factor of change between survivors and nonsurvivors, we speculated that CX3CR1 down-regulation in circulating cells could represent a new feature of sepsis-induced immunodeficiency. Therefore, we further investigated the regulation of its expression as well as that of its specific ligand CX3CL1 after septic shock.

We report for the first time the persistent decrease of the systemic CX3CR1 expression in septic shock patients, both at the mRNA and protein level. Importantly, CX3CR1 down-modulation was significantly more pronounced in nonsurvivors. The observed transcriptional repression represents an early event in the course of the disease because CX3CR1 mRNA expression was already severely decreased in the first days of shock. These data are in line with those of Feezor et al. (48), who recently described a strong decrease in CX3CR1 mRNA expression in circulating leukocytes from patients after major surgery, a clinical condition also known to induce immunosuppression. Interestingly, CX3CR1 is preferentially expressed by cells presenting a proinflammatory, Th1, and/or cytotoxic phenotype, and the interaction of CX3CR1 and CX3CL1 has been proposed as an amplification process for polarized Th1/proinflammatory responses (22, 17). In addition, CX3CR1 is also highly expressed by the proinflammatory CD14^dimCD16^high monocytes (22) as confirmed by our results. Therefore, the decreased CX3CR1 expression that we observed in patients might be related to the decrease of proinflammatory/Th1 immune responses described in septic patients (40, 44, 49, 50). Moreover, the persistent decrease observed in nonsurvivors is in accordance with the hypothesis that a recovery/restoration of the proinflammatory response after septic shock is observed in survivors and is necessary to survive after septic shock. Finally, considering that CX3CR1/CX3CL1 interaction preferentially mediates arrest and migration of inflammatory cells (22), the down-regulation of CX3CR1 expression and the loss of functional response (i.e., lack of CD11b up-regulation) to fractalkine in monocytes from septic patients observed in our study might directly impair their tissue recruitment and contribute to patients’ inability to kill invading microorganisms. Similar decreases in chemotaxis receptor leading to failures in chemotaxis have been already described in sepsis (51, 52). Moreover, in accordance with this hypothesis, Auffray et al. (53) recently observed, by direct examination of blood monocyte functions in vivo, that the subset of monocytes expressing CX3CR1 patrols healthy tissues through long-range crawling on the resting endothelium. They showed that these monocytes are required for rapid tissue invasion at the site of an infection where they can initiate an early immune response.

This diminished expression of CX3CR1 is likely amplified by the cleavage of CX3CL1 from endothelial cells. Indeed, we also report a significant increase in serum sCX3CL1 concentrations from septic shock patients. CX3CL1 is constitutively expressed by nonhematopoietic tissues (54) but is barely expressed by PBMCs. Accordingly, we found very low (even decreased) systemic levels of CX3CL1 mRNA expression, which could be attributed to circulating dendritic cells, macrophages, and/or endothelial cells (1, 37). Upon stimulation, endothelial cells express several adhesion molecules, which can be cleaved off to yield soluble adhesion molecules. For example, increased levels of ICAM-1 and E-selectin have been observed in patients with sepsis (55). The soluble forms of these molecules are determined as a variable for endothelial activation. Similarly, the release of sCX3CL1 from the endothelial membrane-bound molecule may occur by the action of different proteases, such as TNF-α-converting enzyme (14, 15). Therefore, we may speculate that the increased concentrations of sCX3CL1 in septic patients may be caused by an induction of proteolytic cleavage of CX3CL1 from endothelial cells. Together with the decreased expression of CX3CR1 on circulating monocytes, this phenomenon might be involved in the altered recruitment of proinflammatory cells into infected tissue. Among other alterations, this new feature of monocyte deactivation may therefore contribute to the inability of many patients to clear the primary infection correctly, and to their increased risk of acquiring secondary, nosocomial infections (56). Additional studies are needed to confirm the hypothesis of an induction of CX3CL1 proteolytic cleavage during septic shock and to investigate whether CX3CL1 mRNA is also depressed in endothelial cells.

We also observed that the CX3CR1 loss depicted in septic patients could be mimicked by incubation with LPS or heat-killed bacteria. The challenge of monocytes induced, after a slight and transient up-regulation of CX3CR1 mRNA expression, a major transcriptional down-regulation of this gene. These results are very consistent with the concept of LPS-induced monocyte deactivation, characterized by changes in intracellular pathways and cell surface markers similar to that observed in cells from septic patients (1, 39). Accordingly, our microarray study revealed that a significant transcriptional induction of several endogenous negative regulators of TLR signaling (i.e. SOCS1, ABIN-3, IRAK-M, Monoarch-1, and NOD2) occurred concurrently with the down-regulation of CX3CR1 expression. These molecules, which can directly or indirectly impair NF-κB activation, might thus be involved in the control CX3CR1 mRNA expression (57). Similarly, the decrease of monocyte HLA-DR expression observed in septic shock patients can be reproduced upon LPS challenge of cells from healthy individuals (6, 38). As several mediators released in high concentrations during septic shock have immunosuppressive effects (1) (mainly IL-10 and cortisol), we tested their effect as well (with dexamethasone reproducing the effect of endogenous cortisol). As previously observed for other cell surface markers, such as HLA-DR (10), we observed a slight but significant down-modulating effect of dexamethasone on CX3CR1 transcription level. However, and in accordance with other results obtained on microglial cell cultures (58), IL-10 had a only modest impact on the CX3CR1 mRNA expression. Lastly, we also observed that sCX3CL1 was unable to down-regulate in vitro the expression of its own receptor. Finally, regarding mechanisms sustaining CX3CR1 down-regulation, we cannot rule out an involvement of the massive apoptotic processes occurring in septic shock (59, 60). This point deserves additional specific experiments.

Overall, we demonstrated here for the first time that the expression of CX3CR1 is strongly down-regulated on monocytes from septic shock patients both at the mRNA and the protein level. Importantly, nonsurvivor patients presented with significantly lower expression during the course of the disease. In line with the loss of other crucial molecules involved in LPS recognition (i.e., CD14 (29)), Ag presentation (i.e., HLA-DR (6)), or T cell activation (i.e., costimulatory molecules CD86 (61)), our clinical results as well as the effects depicted for LPS and corticosteroid on CX3CR1 expression are very consistent with the concept of monocyte deactivation after septic shock. Moreover, the present study provides another argument to support the concept that patients who spontaneously restore their immune functions (or those with less pronounced alterations) are those who will survive. Although the results should be confirmed in larger clinical studies, they illustrate, together with our previous study (12), the potential of a global transcriptional analysis to identify new molecules that may play a role in sepsis pathogenesis. The full understanding of the mechanisms responsible for sepsis-induced immunosuppression remains a worthwhile challenge to identify new monitoring tools and therapeutic strategies to save patients from this hitherto deadly disease.

We thank H. Thizy, S. Conrozier, and F. Gueyffier from the Centre d’Investigation Clinique (Clinical Research Center) of Institut National de la Santé et de la Recherche Médicale and Hospices Civils de Lyon for logistic support.

A. Pachot, M. A. Cazalis, C. Faudot, J. Diasparra, N. Bourgoin, F. Turrel, and B. Mougin are employees of bioMérieux. G. Monneret and A. Lepape are working in collaboration with bioMérieux based on a global contract between the Hospices Civils de Lyon (University Hospital of Lyon) and bioMérieux.