SEARCH

SEARCH BY CITATION

Keywords:

  • dendritic cells;
  • differentiation;
  • innate immunity;
  • monocytes;
  • sepsis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Sepsis-induced immune depression is characterized by infection susceptibility and monocyte early deactivation. Because monocytes are precursors for dendritic cells (DC), alterations in their differentiation into DC may contribute to defective immune responses in septic patients. We therefore investigated the ability of monocytes to differentiate into functional DC in vitro in patients undergoing surgery for peritonitis. Monocytes from 20 patients collected immediately after surgery (D0), at week 1 and at weeks 3–4 and from 11 control donors were differentiated into immature DC. We determined the phenotype of monocytes and derived DC, and analysed the ability of DC to respond to microbial products and to elicit T cell responses in a mixed leucocyte reaction (MLR). We show that, although monocytes from septic patients were deactivated with decreased responses to lipopolysaccharide (LPS) and peptidoglycan and low human leucocyte antigen D-related (HLA-DR) expression, they expressed the co-stimulatory molecule CD80, CD40 and CCR7. Monocytes collected from patients at D0 and week 1 differentiated faster into DC with early loss of CD14 expression. Expression of HLA-DR increased dramatically in culture to reach control levels, as did responses of DC to LPS and peptidoglycan. However, although patient and control immature DC had similar abilities to induce T cell proliferation in MLR, maturation of DC derived from patients did not increase T cell responses. These results show that circulating monocytes from septic patients express markers of activation and/or differentiation despite functional deactivation, and differentiate rapidly into phenotypically normal DC. These DC fail, however, to increase their T cell activation abilities upon maturation.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Sepsis represents a major cause of mortality in developed countries intensive care units (ICU) [1–3]. For between 10% and more than 30% of patients, abdominal infection is the origin of the disease [1,4]. Initially, patients show signs of systemic inflammation (systemic inflammatory response syndrome, SIRS) including fever, tachycardia, leucocytosis and hyperventilation, which can lead further to sepsis and eventually septic shock [5–7]. The latter, defined as an infection with arterial hypotension refractory to vasopressors, is in most cases associated with multiple organ dysfunctions [8].

Septic patients are subject to infection with commensal microorganisms and this may be due to a state of immunosuppression [9,10] with defective adaptive as well as innate immune responses. A decreased expression of human leucocyte antigen D-related (HLA-DR) on the surface of monocytes was recognized as a marker for this immunosuppression with predictive value on the clinical evolution [11–14]. Low expression of HLA-DR is associated with decreased responses of monocytes to ex vivo activation, such as decreased production of tumour necrosis factor (TNF)-α, interleukin (IL)-1α, IL-1β, IL-6, IL-10 and IL-12p40 [15,16] (Payen et al., unpublished observations). Monocytes are precursors for dendritic cells (DC) in vitro[17,18] as well as in vivo[19–28], and the effects of immunosuppression on monocyte differentiation and DC generation are, to date, not well understood. DC are professional antigen-presenting cells and are the only cells able to activate naive T cells [29]. After encounter with a pathogen, DC capture, process and present the antigens complexed to major histocompatibility complex (MHC) class II molecules, such as HLA-DR, to the T cells. In addition, DC can deliver a wide range of signals through expression of co-stimulatory molecules, such as CD80 and CD86, and the secretion of various cytokines, including IL-12 and IL-10. These signals contribute to orientate the adaptive immune response towards Th1, Th2, regulatory or tolerogenic responses [29]. Because DC are essential for immune responses development and regulation, alterations in their generation and functions could contribute to immunosuppression.

Based on the evidence that systemic inflammation and surgical hit might down-regulate monocyte functions, the aim of this study was to assess the ability of circulating monocytes from post-operative peritonitis patients to differentiate into DC in vitro, in comparison with control donors, and to characterize the functions of the DC. Blood samples were obtained immediately, 7 days and 4 weeks after surgery to study the effect of clinical evolution on monocyte functions. Data were analysed according to patients' disease severity. Phenotype and activation status of both monocytes and DC produced by differentiation in culture with IL-4 and granulocyte–macrophage colony-stimulating factor (GM-CSF) were studied, especially the expression of HLA-DR, Toll-like receptor (TLR)2, TLR4, CD40, CD80 and CCR7. The ability of DC to induce T cell proliferation in mixed leucocyte reaction (MLR) was also investigated.

Materials and methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Patients and control donors

The design of the study was approved by the institutional ethical committee. Twenty patients undergoing surgery for peritonitis were enrolled in the study. Written informed consent was obtained as soon as possible from the patient or from their next of kin. Patients' clinical characteristics are summarized in Table 1. At inclusion time, organ failures were quoted using logistic organ dysfunction (LODS) criteria [30]. In addition, patients' severity was characterized using the standard clinical scoring systems simplified acute physiology score (SAPS II) quoted at day 0, which includes comorbidity, biological and functional parameters [31,32].

Table 1.  Clinical characteristics of peritonitis patients.
PatientsAgeGenderDiagnosisSAPS IIMicro organisms in peritoneal fluidNo. of organ failure(s)Outcome
194MSigmoid perforation70Bacteroides fragilis, Strepto D, Escherichia coli4Death 1st week
253MSecondary peritonitis after sigmoidectomy48E. coli multiR, Streptococcus (no group), Proteus4Death 2nd week
369MSecondary peritonitis after small intestine fistula45E. coli3Alive
453FDiverticular peritonitis50E. coli, Klebsiella pneumoniae3Death 1st week
576FStercoral peritonitis after sigmoiditis3803Alive
677FSecondary peritonitis after anastomotic small intestine fistula44E. coli (abcess)2Alive
748FAcute necrotizing pancreatitis with infection42B. fragilis, S. anginosus2Death at D30
889FAcute angiocholitis with cholecystis44Enterococcus faecalis, E. coli2Alive
994FGastric ulcer perforation49E. faecalis, group C streptococcus, Proteus vulgaris2Alive
1089FSecondary peritonitis after total sigmoido-proctectomy43E. coli, Bacteroides1Alive
1121MDudonenum ulcer perforation30Not determined1Alive
1253MSecondary peritonitis after trauma3101Alive
1370FNecrotizing ischemic colitis and sigmoid perforation3801Alive
1439MSecondary peritonitis after anastomosis fistula12Citrobacter koseri, E. coli, S. intermedius1Alive
1564MBiliary secondary peritonitis3901Alive
1655MGastric ulcer perforation2400Alive
1755FGastric ulcer perforation2400Alive
1831FGastric ulcer perforation1600Alive
1986FAppendicular peritonitis24S. constellatus0Alive
2060MSecondary peritonitis after anastomosis fistula20E. coli0Alive

Peritoneal fluid samples were withdrawn in 11 patients during surgery and kept in sterile polystyrene containers at 4°C until analysis. In 16 patients, the first post-operative blood sample [day 0 (D0) sample] was drawn 9·5 ± 10·2 h (mean ± s.d.) after the end of surgery, and 17·5 ± 10·2 h after the onset of the first organ failure when present. Second (week 1) and last (weeks 3–4) blood samples were drawn 7 ± 1 and 25 ± 5 days after surgery.

In four patients (5, 14, 15 and 20), blood samples were drawn only 8 ± 2 days after surgery and used for monocyte differentiation and subsequent MLR.

Bacteriological routine analysis was performed in peritoneal fluid from patients indicated in Table 1.

Blood samples from six healthy donors were collected to compare results obtained in the patients' group for lymphocyte subsets, cell differentiation in culture and stimulation with microbial products. Eight other healthy controls were used to compare plasma cytokine concentrations with patients, and five additional donors were used to compare MLR results with patients 5, 14, 15 and 20 (Table 1).

Peritoneal fluid analysis

Peritoneal fluid samples were centrifuged for 10 min at 300 g. Supernatants were frozen immediately at −20°C until cytokine measurement by enzyme-linked immunosorbent assay (ELISA). Cell pellets were resuspended in complete RPMI medium (RPMI-1640) supplemented with 10% heat-inactivated fetal calf serum (FCS), 2 mM l-glutamine, 100 U/ml penicillin, 100 µg/ml streptomycin and 10 mM HEPES (all from Gibco, Invitrogen Life Technologies, Cergy Pontoise, France). Cell counts and viability were assessed by Trypan blue exclusion (Eurobio, Les Ulis, France) on a Mallassez chamber. Smears were stained with Wright–Giemsa colouration (Accustain™, Sigma Diagnostics, St Louis, MO, USA) and stored for later microscopy analysis.

Blood sample and cell separation

Arterial or venous blood (4·5 ml) was drawn in sterile tubes (Vacutainer, Becton Dickinson, Plymouth, UK) containing lithium heparinate. Plasma was separated by centrifugation (700 g) and stored at −20°C until IL-10, IL-12p40 and macrophage migration inhibitory factor (MIF) measurements were performed by ELISA. In addition, 20–25 ml blood was drawn in sterile tubes (Vacutainer) containing sodium citrate for peripheral blood mononuclear cells (PBMC) separation by density-gradient centrifugation (Eurobio). Cell viability was assessed by Trypan blue.

Cell culture

PBMC (3·6 × 106 ml−1) were resuspended in complete RPMI medium. Cell culture was performed in polystyrene six- or 24-well plates (Falcon®, Becton Dickinson, Le Pont De Claix, France), depending on cell numbers, with 3 ml or 0·75 ml of cell suspension per well, respectively, as described previously [19]. Briefly, monocytes were allowed to adhere for 2 h; supernatants and non-adherent cells were then removed to eliminate lymphocytes, and fresh complete RPMI medium supplemented with 1000 U/ml rHu IL-4 (Promocell, Heidelberg, Germany) and 800 U/ml rHu GM-CSF (Leucomax®, Novartis, Huningue, France), preheated at 37°C, was added. At days 3 and 6, half the culture medium was replaced by fresh preheated complete RPMI medium with 500 U/ml rHu IL-4 and 800 U/ml rHu GM-CSF. Non-adherent cells were harvested at D1, D2, D3 and D7 for fluorescence activated cell sorter (FACSs) analysis. Cells harvested at D7 were defined as immature DC (iDC). Mature DC (mDC) were obtained after an additional 48 h stimulation with 10 ng/ml sonicated lipopolysaccharide (LPS) from Escherichia coli O111:B4 (Sigma).

Cell stimulation

PBMC (1 × 106 ml−1) and non-adherent cultured cells (2·7 × 105 ± 1·9 × 105 ml−1) harvested at D7 were resuspended in complete RPMI medium and stimulated for 24 h with either 10 ng/ml LPS from E. coli O111:B4 or 1 µg/ml peptidoglycan (PGN) (Invivogen, San Diego, CA, USA) or sterile non-pyrogenic H20 (Laboratoires Aguettant, Lyon, France) as control. After 24 h, cells were centrifuged for 10 min at 300 g and supernatants were stored at −20°C until IL-10 and IL-12p40 measurement by ELISA.

Analysis of cell populations and cell surface molecules during differentiation and after stimulation

Cell surface markers were analysed by flow cytometry with double or triple surface immunostaining, with the following antibodies: anti-HLA-DR (L243)–phycoerythrin (PE) (BD Biosciences, San Jose, CA, USA), anti-CD1a (HI149)–phycoerythrin-cyanine 5 (PC5) (BD PharMingen, San Diego, CA, USA), anti-CCR7 (150503)–PE (R&D Systems Europe, Lille, France), anti-CD3 (UCHT1)–fluorescein isothiocyanate (FITC), anti-CD3 (UCHT1)–PC5, anti-CD14 (RMO52)–FITC, anti-CD19 (J4·119)–FITC, anti-CD4 (13B8·2)–PE, anti-CD8 (B9·11)–PE, anti-CD56 (N901)–PE, anti-CD66b (80H3)–FITC, anti-CD1a (BL6)–PC5, anti-CD25 (B1·49·9)–PC5 (all from Beckman Coulter Immunotech, Marseille, France), anti-TLR2 (TL2·1)–PE, anti-TLR4 (HTA125)–PE, anti-CD40 (5C3)–PE and anti-CD80 (2D10·4)–PE (all from eBioscience, San Diego, CA, USA).The expression of HLA-DR, TLR2, TLR4, CCR7, CD40 and CD80 was quantified in number of sites per cell (nb sites) after calibration with Quantibrite™ (BD Biosciences).

T cell separation

Residual blood samples after cytapheresis from healthy donors (Blood Bank, Saint-Louis Hospital, Paris, France) were used for isolation of PBMC by density-gradient centrifugation (Eurobio). Cell viability was assessed by Trypan blue. For MLR assays, T cells were separated by magnetic depletion (StemCell Technologies, Meylan, France and Miltenyi Biotec, Paris, France) according to the manufacturers' recommendations. Cells were resuspended at 10 × 106/ml in RPMI-1640 containing 50% heat-inactivated FCS (Gibco) and 10% dimethylsulphoxide (DMSO) (Sigma), frozen at −80°C overnight, and then stored in liquid nitrogen.

MLR

Frozen T cells were removed from liquid nitrogen, thawed immediately at 37°C in a water bath, resuspended in the same volume of preheated complete RPMI medium, and centrifuged at 300 g for 10 min. Cells were washed with complete RPMI medium. Viability was assessed by Trypan blue. T cells were incubated with 5 μM carboxyfluorescein diacetate (CFDA-SE, Invitrogen Molecular Probes, Cergy Pontoise, France) for 10 min at 37°C, washed with complete RPMI medium and cultured with monocyte-derived iDC or mDC. DC and T cell were co-cultured at 1/10 ratio, with 5 × 104 DC for 5 × 105 T cells per ml. After 5 days, cells were harvested and stained with anti-CD3 for flow cytometry analysis. FL1 signals obtained from CFDA-SE intracellular esterification were analysed for CD3-positive cells. T cell division induced the decrease of intracellular content of CFDA-SE, resulting in a multiple-peak signal. The peak with the highest mean fluorescence intensity (MFI) characterized cells that did not decrease CFDA-SE content and hence did not divide. Peaks with lower MFI represented cells engaged in proliferation.

The percentage of T cells engaged in proliferation was calculated as:

  • (all events − number of events for peak of non-dividing cells) × 100/all events.

In addition, the percentage of initial progenitors T cells undergoing division was calculated as described previously [33].

Cytokines

Plasma and supernatants IL-10, IL-12p40 and MIF concentrations were determined by an immunoenzymatic method (ELISA). IL-10 was measured using a kit (optEIA™ set; PharMingen) according to the manufacturer's recommendations. IL-12p40 and MIF were measured using unlabelled capture and biotinylated detection monoclonal anti-human (R&D Systems, Abingdon, Oxon, UK) and recombinant human standards (PharMingen). Plates were read with an automatic plate reader at 450 nm (EL800, Bio-tek Instruments, Winooski, VT, USA). Standard samples ranged from 7·8 to 500 pg/ml for IL-10 and from 31·25 to 2000 pg/ml for IL-12p40 and MIF. Detection thresholds were 2·7 ± 3·1 pg/ml for IL-10, 25·8 ± 33·3 pg/ml for IL-12p40 and 25·7 ± 34·6 pg/ml for MIF.

Statistical analysis

Changes in the percentages of CD14+, CD1a+ and in the number of HLA-DR sites between patients and controls during the 7 days of culture were analysed by two-way analysis of variance (anova) for repeated measures and Sheffé's post-hoc test.

Mann–Whitney and Wilcoxon non-parametric tests were used to compare plasma and peritoneal cytokines, TLR2, TLR4, CCR7, CD40 and CD80 baseline expression, as well as CD40, CD80, cytokine response to stimulation and percentage of T cell proliferation in MLR between patients and controls. P < 0·05 was considered significant.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Patients

Sixteen patients (all except patients 5, 14, 15 and 20 in Table 1; six men/10 women) undergoing surgery for peritonitis were enrolled as early as possible after the end of surgery (D0). The four remaining patients (three men/one woman) were enrolled 8 ± 2 days after surgery and the samples were used for MLR assays. Mean age was 64 ± 21 (mean ± s.d.) and the calculated SAPS II was 37 ± 14. Because transitory single cardiovascular failure often occurs after surgery, patients were classified according to the presence of two or more organ failures versus less than two organ failures at D0. Microbiological analysis revealed either polymicrobial contamination of peritoneal fluid, mainly including E. coli, or fluid sterility, especially when peritonitis originated from gastric ulcer perforation (Table 1).

Plasma cytokines

To define further our patient population, we measured the plasma concentration of three cytokines related to monocyte/DC function: IL-10 that is involved in down-regulation of monocytes HLA-DR [34,35] and in generating tolerogenic DC [36,37], IL-12p40 involved in DC-induced lymphocyte Th1 orientation and MIF, which acts as a counter-regulatory molecule for corticoids [38]. In patients, results were expressed according to the number of organ failures present at D0 (Table 2). Plasma IL-10 was elevated in both patient groups compared to controls at D0, and then decreased in week 1 and the last samples. In contrast, IL-12p40 was almost undetectable in both patient groups at D0, and increased only in patients with less than two organ failures in the last sample. This may reflect an early predominance of anti-inflammatory cytokines in plasma and a shift to immunodepression in blood and peripheral tissues. MIF was higher in patients compared to controls until week 4, suggesting a different regulation for this mediator.

Table 2.  Concentrations of interleukin (IL)-10, IL-12p40 and macrophage migration inhibitory factor (MIF) in patients' peritoneal fluid (n = 11), in plasma in D0 (n = 16), week 1 (n = 11) and late (n = 6) samples and in control donors (n = 8). Patients were separated into two groups according to the number of organ failures present at D0.
 Peritoneal fluidPlasma D0 samplePlasma week 1 samplePlasma weeks 3–4 samplePlasma controls
< 2 organ failures≥ 2 organ failures< 2 organ failures≥ 2 organ failures< 2 organ failures≥ 2 organ failures< 2 organ failures≥ 2 organ failures
  • *

    P < 0·05, Mann–Whitney between patients' plasma and control donors' plasma.

  • P < 0·05, Wilcoxon between patients' plasma D0 sample and patients' peritoneal fluid. Data in median ± interquartile range.

IL-10 (pg/ml)516 ± 687288 ± 75734 ± 19*64 ± 381*7 ± 174 ± 3710 ± 114 ± 280·3 ± 7
IL-12p40 (pg/ml)0·0 ± 500·0 ± 410·0 ± 0·00·0 ± 0·00·0 ± 750·0 ± 0·059 ± 42*0·0 ± 0·00·0 ± 0·0
MIF (ng/ml)20·4 ± 30·324·0 ± 25·45·2 ± 9·0*2·9 ± 2·5*10·9 ± 18·3*7·8 ± 5·1*34·0 ± 10·9*4·6 ± 20·11·0 ± 1·2

Peritoneal fluid analysis and comparison with plasma cytokines

To assess the abdominal inflammatory response, peritoneal fluid samples were collected from 11 patients (numbers 2, 8, 9, 10, 11, 12, 13, 16, 17, 18 and 19). Cellularity was analysed and cytokine concentrations were determined and compared with those in plasma in patients to assess the existence of an inflammation gradient between abdomen and systemic compartment. Mean (± s.d.) cell count was 102 ± 135·106 ml−1. As expected, smear analysis revealed that most of these cells were polymorphonuclear neutrophils and macrophages, with the presence of many intracytoplasmic vacuoles. In patients 10, 13, 16, 17 and 18 many cells showed an apoptotic nucleus (data not shown). In peritoneal fluid from 11 patients at D0 (Table 2), IL-10 and MIF were higher than in plasma only in patients with less than two organ failures (P < 0·05).

White blood cell populations in septic patients' peripheral blood

Variations in counts of various leucocyte populations in sepsis have been described [39–41]. We therefore quantified the following cell populations among PBMC in our patients: CD3+, CD3+/CD4+, CD3+/CD4+/CD25+, CD3+/CD8+ and CD3+/CD56+ T cells, CD19+ B cells and CD14+ monocytes. Contamination of PBMC with granulocytes was assessed by the presence of CD66b+ cells.

As shown in Table 3, T cell counts were lower in PBMC from patients compared to controls at D0. Population analysis showed that reduced numbers were due primarily to decreased numbers of both CD4+ and CD8+ T cells, confirming previous reports [42]. In contrast, patients CD4+/CD25+ T cell counts were not different from healthy controls. CD14+ monocytes were increased in patients immediately after surgery and were still elevated in week 1 samples. CD19+ B cell counts were also high at this time. No difference between patients and control donors was detected for CD3+/CD56+ cells.

Table 3.  Cell count in healthy donors and peritonitis patients in median ± interquartile range.
106 cells/ml bloodHealthy (n)Patients D0 (n)PPatients D7 (n)PPatients D28 (n)P
  1. Absolute numbers of monocytes (CD14+), T cells (CD3+), CD4+ T cells (CD3+/CD4+), T regulatory (Treg) cells (CD3+/CD4+/CD25+), CD8+ T cells (CD3+/CD8+), natural killer (NK) T cells (CD3+/CD56+) and B cells (CD19+) were determined in patients' samples drawn at D0, D7 and D28, and compared to healthy donors' cell counts (Mann–Whitney non-parametric test).

CD14+0·07 ± 0·08 (6)0·20 ± 0·33(16)0·0220·20 ± 0·27 (11)0·0440·14 ± 0·16 (6)n.s.
CD3+0·70 ± 0·35 (6)0·29 ± 0·52 (16)0·0150·65 ± 0·51 (11)n.s.0·83 ± 0·72 (6)n.s.
CD3+/CD4+0·50 ± 0·22 (6)0·17 ± 0·32 (15)0·0130·43 ± 0·21 (11)n.s.0·50 ± 0·22 (6)n.s.
CD3+/CD4+/CD25+0·04 ± 0·05 (4)0·04 ± 0·05 (10)n.s.0·08 ± 0·11 (10)n.s.0·08 ± 0·14 (5)n.s.
CD3+/CD8+0·23 ± 0·17 (6)0·08 ± 0·16 (16)0·0180·21 ± 0·23 (11)n.s.0·22 ± 0·62 (6)n.s.
CD3+/CD56+0·04 ± 0·06 (6)0·02 ± 0·04 (16)n.s.0·04 ± 0·14 (11)n.s.0·09 ± 0·15 (6)n.s.
CD19+0·05 ± 0·03 (6)0·05 ± 0·11 (16)n.s.0·10 ± 0·07 (11)0·0180·07 ± 0·05 (6)n.s.

We observed a significant contamination of patient PBMC with CD66b+ cells in all samples (17·5 ± 38·8%; 27·2 ± 25·0%; and 2·3 ± 5·8% in postoperative week 1 and the last samples, respectively, P < 0·05 versus 0·4 ± 1·4% in controls).

Evolution of CD14, CD1a and HLA-DR expression during monocyte differentiation

To monitor the differentiation into DC of monocytes isolated from patients and cultured in the presence of IL-4 and GM-CSF, we analysed the expression of CD14 (Fig. 1a,b,c) and CD1a (Fig. 1d–f), which are monocyte and DC markers, respectively. We also analysed HLA-DR (Fig. 2a,b,c), as its low level of expression correlates with monocyte deactivation in sepsis. Expression was analysed by flow cytometry at days 1, 2, 3 and 7 of the culture (D1, D2, D3 and D7) and we determined the percentage of cells within the cell population that was positive for CD14, CD1a or both markers. In patients' initial PBMC samples collected immediately after surgery and in healthy donor samples this cell population included 99·3 ± 1·0 and 99·8 ± 0·3% of CD14+ cells, respectively, and no statistically significant difference between the percentage of CD1a+ cells was detected (5·0 ± 11·9 and 1·8 ± 3·6).

image

Figure 1. Evolution of cell phenotype during 7 days of culture with interleukin (IL)-4 and granulocyte–macrophage colony-stimulating factor (GM-CSF). Monocytes (CD14+/CD1a cells at D0 of culture) were separated from peripheral blood mononuclear cells (PBMC) by adhesion. Cells that became non-adherent were analysed by flow cytometry at D1, D2, D3 and D7. (a,b,c) Percentage of CD14+ cells and (d,e,f) percentage of CD1a+ cells within a population gated according to positive expression of CD14 and/or CD1a in samples collected at D0, week 1 and weeks 3–4 from peritonitis patients (black circles) and control donors (white circles). D0 sample was obtained from patients 1, 3, 6, 10, 15, 16 and 18, week 1 sample was obtained from patients 3, 6, 10, 15 and 18, and weeks 3–4 sample was obtained from patients 3, 6, 7, 9, 11 and 12. Mean ± s. d. Patient groups were compared to one control donors group (n = 6) by two-way analysis of variance for repeated measures. *P < 0·05 Sheffépost-hoc test between patients and control donors.

Download figure to PowerPoint

image

Figure 2. Evolution of human leucocyte antigen D-related (HLA-DR) expression during 7 days of culture with interleukin (IL)-4 and granulocyte–macrophage colony-stimulating factor (GM-CSF). Monocytes (CD14+/CD1a cells at D0 of culture) were separated from peripheral blood mononuclear cells (PBMC) by adhesion. Cells that became non-adherent were analysed by flow cytometry at D1, D2, D3 and D7. (a,b,c) Number of HLA-DR sites per cell within a population gated according to positive expression of CD14 and/or CD1a in samples collected at D0, week 1 and weeks 3–4 from peritonitis patients (black circles) and control donors (white circles). D0 sample was obtained from patients 1, 3, 6, 10, 15, 16, and 18, week 1 sample was obtained from patients 3, 6, 10, 15 and 18, and weeks 3–4 sample was obtained from patients 3, 6, 7, 9, 11 and 12. Mean ± s.d. Patient groups were compared to one control donors group (n = 6) by two-way analysis of variance for repeated measures. *P < 0·05 unpaired t-test for each culture day, Sheffe post-hoc test between patients and control donors.

Download figure to PowerPoint

Surprisingly, during cell culture in IL-4 and GM-CSF more patients' cells than control cells became CD14 negative on first and second culture days of week 1 samples (Fig. 1b, P = 0·003 two-way anova; P = 0·005 and P = 0·0005 versus control on the second and third days of culture, respectively, Sheffé's test). A tendency for the same accelerated evolution was observed with D0 and weeks 3–4 samples, but without statistical significance. Thus, on the third culture day of D0 sample (Fig. 1a), there were 60·6 ± 16·8% CD14+ cells in the patients' group but still 82·7 ± 11·7% CD14+ in controls.

In contrast, cells from patients and healthy controls acquired CD1a expression along a similar kinetic (Fig. 1d,e,f). After 7 days of culture, cells from patients and healthy controls showed a similar immature DC phenotype with 2·8 ± 5·7% and 3·3 ± 4·7% of CD14+ cells, and 97·8 ± 4·9% and 99·7 ± 0·4% of CD1a+ cells, respectively. As shown in Fig. 2a, the analysis of HLA-DR expression in CD14+ cells in PBMC confirmed previous reports [11–14] with very low levels of HLA-DR compared to healthy controls (1947 ± 1714 sites per cell and 17722 ± 12219, respectively, P < 0·0001, Sheffé's test). However, HLA-DR expression increased dramatically in patient cells at D2 and D3 (P = 0·01, two-way anova) to reach normal levels at D7 of culture in the presence of IL-4 and GM-CSF (after 7 days' culture: 88847 ± 50981 sites per cell in patients and 111206 ± 20812 in controls, n.s.).

In patients' week 1 samples (Fig. 2b), CD14+ cells in PBMC also tended to express lower HLA-DR than healthy control cells (3233 ± 2920 sites) and restoration to normal levels during culture.

Three to 4 weeks after surgery (Fig. 2c), CD14+ cells in patients' PBMC did not show decreased HLA-DR expression anymore.

Expression of CD40, CD80, TLR2, TLR4 and CCR7 on monocytes and after 7 days of culture

To define further the differentiation status of both monocytes and DC, phenotypic assessment of pathogen recognition receptors, chemokine receptor and antigen presentation co-stimulatory molecules was performed by flow cytometry. CD40 (Fig. 3a,b,c), CD80 (Fig. 3d,e,f), CCR7 (Fig. 3g,h,i), TLR2 (data not shown) and TLR4 expression (data not shown) were quantified on CD14+ cells in PBMC and after 7 days of culture in the presence of IL-4 and GM-CSF in samples from patients on D0, at week 1 and at weeks 3–4, and expression was compared to that of control donors. Patients were grouped according to the number of organ failure.

image

Figure 3. CD40, CD80 and CCR7 expression on monocytes and immature dendritic cells (iDC) by flow cytometry analysis. (a,b,c) Number of CD40 sites per cell; (d,e,f) number of CD80 sites per cell; (g,h,i) number of CCR7 sites per cell within the gated population for D0, week 1 and weeks 3–4 samples, respectively, in peritonitis patients having less (grey circles) and ≥ 2 (black circles) organ failure(s) and control donors (white circles). For CD40 and CD80, D0 sample was obtained from patients 2, 4, 7, 8, 9, 10, 11, 12, 13, 17 and 18, week 1 sample was obtained from patients 2, 7, 9, 10, 11, 12, 13 and 18, and weeks 3–4 sample was obtained from patients 7, 9, 11 and 12. Individual values. For CCR7, D0 sample was obtained from patients 2, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 15, 16, 17 and 18, week 1 sample was obtained from patients 2, 3, 6, 7, 9, 10, 11, 12, 13, 15 and 18, and weeks 3–4 sample was obtained from patients 3, 6, 7, 9, 11 and 12. Individual values. *P < 0·05, non-parametric Mann–Whitney between patient groups and control donors group (n = 6). No difference was observed between patient groups. §P < 0·05 non-parametric Wilcoxon between monocytes and dendritic cells within each group.

Download figure to PowerPoint

As described previously [19,22,43,44] CD40 and CD80 expressions on CD14+ monocytes were not detected in four of six and in five of six controls, respectively. In contrast, in monocytes from PBMC isolated from 10 patients immediately after surgery, CD40 and CD80 were expressed in seven and four patients, respectively, although this did not amount to statistically significant differences. For CD40, this profile persisted 1 week and 3–4 weeks after surgery, especially in patients with less than two organ failures. Fewer patients showed detectable expression of CD80 at these time-points. Both CD40 and CD80 expression increased during culture in the presence of IL-4 and GM-CSF (P < 0·05, Wilcoxon for control and for all post-operative patients and week 1 samples, P = 0·068 for all late samples from patients). The significant difference in CD80 expression on DC between patients with two or more organ failures at D0 and controls is due probably to one healthy volunteer with very high CD80 expression.

TLR2 expression on monocytes and on DC was highly heterogeneous in both controls and patients in all samples (data not shown). Consequently, no significant differences were observed between groups and between monocytes and DC within each group.

TLR4 expression in monocytes was less heterogeneous than TLR2 (data not shown). In healthy controls, one individual showed high TLR4 expression while for the five others it was either undetectable of very weak. In patients, TLR4 expression was usually low, except in the week 1 sample where three patients (two with less and one with more than two organ failures) of 11 patients exhibited TLR4 expression higher than 300 sites per monocyte. These results are in agreement with expression profiles described in the literature, with septic patients exhibiting higher expression than healthy donors [45–47]. TLR4 expression tended to increase during culture in the presence of IL-4 and GM-CSF, but the amplitude of variation was heterogeneous between individuals, leading to heterogeneous expression on DC in both patients and healthy donors. However, this increase was significant for all patient post-operative samples (P < 0·01) (data not shown).

CCR7 expression on monocytes was not detected in three of six healthy donors. When detectable, it did not exceed 637 sites per cell. In contrast, we observed up to 1900 CCR7 sites per monocyte after surgery in 11 of 15 patients. One week after surgery, nine patients of 11 expressed up to 4975 CCR7 sites per monocyte, and at the time of late sample collections we found up to 4059 sites CCR7 per monocyte in five of six patients. As for TLR4, the rate of increase of CCR7 expression during the 7-day culture was heterogeneous, and led to heterogeneous expression on DC in both patients and healthy donors. However, as for TLR4, this increase was significant for all patients in post-operative samples (P < 0·05).

Stimulation test

To assess phenotypic modifications and cytokine production capabilities of monocytes and DC after exposure to microbial products, PBMC and DC were stimulated during 24 h with LPS or PGN (Fig. 4).

image

Figure 4. Human leucocyte antigen D-related (HLA-DR), CD40 and CD80 expression on CD14+ monocytes (a,b,c) and on immature dendritic cells (iDC) (d,e,f) after 24 h stimulation with lipopolysacharide (LPS) or peptidoglycan (PGN). Peripheral blood mononuclear cells (PBMC) and monocyte-derived iDC from control donors and from peritonitis patients D0, week 1 and weeks 3–4 samples were incubated in complete RPMI with LPS (10 ng/ml, black bars) or PGN (1 µg/ml, grey bars) for 24 h and analysed by flow cytometry. Expression was also measured before incubation (white bars) and after 24 h with sterile non-pyrogenic H20 (hatched bars). For PBMC stimulation, D0 sample was obtained from patients 2, 4, 7, 8, 9, 10, 11, 12 and 13, week 1 sample was obtained from patients 2, 7, 9, 10, 11, 12, 13 and 18, and weeks 3–4 sample was obtained from patients 7, 9, 11 and 12. Mean ± s. d. For iDC stimulation, D0 sample was obtained from patients 2, 4, 7, 8, 9, 10, 11, 12, 13 and 18, week 1 sample was obtained from patients 7, 9, 11, 12, 13 and 18, and weeks 3–4 sample was obtained from patients 7, 9, 11 and 12. Mean ± s.d. *P < 0·05, non-parametric Mann–Whitney between patients and control donors (n = 6). §P < 0·05 non-parametric Wilcoxon between LPS or PGN stimulation and control condition within each group.

Download figure to PowerPoint

Within PBMCs, monocytes were analysed for the expression of HLA-DR (Fig. 4a), CD40 (Fig. 4b) and CD80 (Fig. 4c) before and after stimulation. The same markers were analysed in DC (Fig. 4d,e,f) stimulated after 7 days of culture with IL-4 and GM-CSF.

Expression of CD40 and CD80 tended to increase with both agonists in monocytes from healthy donors. HLA-DR modification of expression seemed, however, more related to incubation procedure than to specific agonist effect, as suggested by Monneret et al. who showed a spontaneous increase of monocytes HLA-DR in whole blood [48].

In patients, we observed that at D0 monocyte responses to LPS and PGN were almost totally inhibited with respect to expression of HLA-DR, CD40 and CD80, reflecting the deactivated state of monocytes in sepsis patients. Induction of CD40 and CD80 expression was recovered progressively in monocytes from week 1 and late samples, while HLA-DR response recovery seemed more delayed.

In DC, HLA-DR (Fig. 4d), CD40 (Fig. 4e) and CD80 (Fig. 4f) responses to LPS were increased significantly in healthy and patients, suggesting that the depression observed in monocytes was reversed. PGN effects were less pronounced in both groups and for all markers.

IL-10 production from patients PBMCs (Fig. 5a) was significantly down-regulated, especially after LPS stimulation, in all D0, week 1 and the last samples. In contrast, IL-12p40 production was unexpectedly not depressed (Fig. 5c). In contrast, DC responses to LPS and PGN (Fig. 5b,d) were not different between patients and healthy controls.

image

Figure 5. Interleukin (IL)-10 and IL-12p40 concentration in peripheral blood mononuclear cells (PBMC) (a,c) and immature dendritic cells (iDC) (b,d) supernatant after 24 h stimulation with lipopolysaccharide (LPS) or peptidoglycan (PGN). PBMCs and monocyte-derived iDC from control donors and from peritonitis patients. D0, week 1 and weeks 3–4 samples were incubated in complete RPMI with LPS (10 ng/ml, black bars) or PGN (1 µg/ml, grey bars), and supernatants were analysed by enzyme-linked immunosorbent assay. Cytokine concentration was also measured after 24 h with sterile non-pyrogenic H20 (hatched bars). For PBMC stimulation, D0 sample was obtained from patients 2, 4, 7, 8, 9, 10, 11, 12 and 13, week 1 sample was obtained from patients 2, 7, 9, 10, 11, 12, 13 and 18, and weeks 3–4 sample was obtained from patients 7, 9, 11 and 12. Mean ± s. d. For iDC stimulation, D0 sample was obtained from patients 2, 4, 7, 8, 9, 10, 11, 12, 13 and 18, week 1 sample was obtained from patients 7, 9, 11, 12, 13 and 18, and weeks 3–4 sample was obtained from patients 7, 9, 11 and 12. Mean ± s.d. *P < 0·05, non-parametric Mann–Whitney between patients and control donors (n = 6); §P < 0·05 non-parametric Wilcoxon between LPS or PGN stimulation and control condition within each group.

Download figure to PowerPoint

MLR

As we observed that the phenotype and the responses to microbial products of DC derived from monocytes of patients were similar to those of healthy controls, we assessed their capacity to induce lymphocyte proliferation in allogeneic conditions (Fig. 6). We focused on samples that showed the most striking difference between patients and controls, i.e. the loss of CD14 expression in dendritic cells at 8 ± 1 days after surgery, and performed MLR with monocyte-derived DC from four patients. Figure 6 shows that T cell activation abilities of iDC derived from patients or from healthy controls were not significantly different in terms of percentage of progenitors that had undergone one or more division within the initial T cell population (Fig. 6a) or of the percentage of dividing T cells in the culture (Fig. 6b). After maturation by exposure to LPS, in contrast to control DC which significantly increased recruitment of T cell progenitors (Fig. 6a) and the percentage of proliferating T cells (Fig. 6b), mDC from patients showed roughly equivalent activity as iDC in T cell progenitor activation (Fig. 6a) and only a trend to increased percentage of proliferating T cells (Fig. 6b). These results indicate that, although dendritic cells derived from patients expressed normal levels of HLA-DR, of co-stimulatory molecules and of cytokines after stimulation, they were still not able to raise the capacity to initiate T cell activation and proliferation after maturation. Alternatively, patients' iDC have already acquired the capacity to activate T cell progenitors (Fig. 6a). In that case, however, signals delivered by patients' iDC would not allow a full development of the T cell response (Fig. 6b).

image

Figure 6. Allogeneic T cell proliferation in mixed leucocyte reaction with immature dendritic cells (iDC) or mature DC (mDC) from peritonitis patients 8 days after surgery (n = 4) and from control donors (n = 5). Percentage of progenitors engaged in proliferation within the initial T cell population (a) and percentage of proliferating T cells (b) were calculated as described in the Materials and methods section. Median ± interquartile range. No difference was observed between patients' iDC and control donors iDC, and between patients mDC and control donors mDC. §P < 0·05 non-parametric Wilcoxon between control iDC and mDC.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study provides new insights into monocyte function and differentiation capabilities during human sepsis. Monocytes isolated from septic patients have been described as deactivated. Thus, our results using samples obtained during the immediate post-operative period showed low expression of HLA-DR and decreased responses to LPS and PGN, confirming previous studies [11–16]. However, expression of CD40, CD80 and CCR7 were higher than in control monocytes. These markers are usually described as markers for activated monocytes and macrophages [49,50] or DC [19,20,22,23,51,52]. Monocytes from patients could therefore display mixed characteristics of deactivation and activation states. Alternatively, these cells have acquired markers of differentiation, suggesting that in the context of surgery-induced systemic inflammation or sepsis, monocytes could be orientated towards DC differentiation. Supporting this latter view, patient monocytes, drawn immediately after surgery or several days or weeks later and cultured in IL-4 and GM-CSF, were able to differentiate into DC. Surprisingly, differentiation was faster compared with controls, as suggested by the more rapid decrease of CD14 expression, especially for samples drawn 7 days after surgery. Because in other pathological states the yield of DC differentiation was decreased, as described in renal metastatic carcinoma patients [53], we also assessed that the cell yield after DC differentiation was not different between patients at any sample time and healthy controls (data not shown).

Determination of plasma cytokine concentrations confirmed the concept of systemic anti-inflammation predominance during sepsis [9], especially with elevated IL-10 concentration. As IL-10 or other anti-inflammation mediators could modulate DC differentiation in vitro[36,37,54–56], patient monocyte differentiation could have been modified by their in vivo‘priming’ in an anti-inflammatory environment. Our findings of unexpected CD40, CD80 and CCR7 monocyte expression, and more importantly the ability of these cells to differentiate rapidly into DC, raise the question of a pre-orientated status of monocytes, which is repressed actively in vivo by systemic anti-inflammation but allowed to proceed in in vitro conditions. Such a control may limit the extension and amplification of immune response and could contribute to susceptibility to infection. In addition, IL-12p40 production by patient PBMC was not depressed, reinforcing further the concept of selective deactivation of monocyte functions and possibly of pre-orientation toward differentiation. Moreover, extensive DC apoptosis has been described during sepsis [57–60], suggesting the need for a rapid turnover of this cell population. Circulating monocytes, which act as a reserve of DC progenitors, would be less likely to differentiate in septic patients in that particular cytokine environment.

The expression of TLR2 and TLR4 on monocytes was heterogeneous. TLR4 was, however, increased in the patient group compared to healthy controls, as suggested by previous studies performed at both mRNA and protein levels [45–47].

Although DC derived in vitro from patient monocytes acquired a phenotype and responses to TLR agonists comparable to control monocyte-derived DC, they showed no increased capacity to activate T cells after maturation induced by LPS. Because the cells expressed similar levels of HLA-DR, co-stimulatory molecules and IL-12 as control cells, the mechanisms for this lack of enhanced ability associated with maturation remains unclear. Alternatively, as patients' iDC were able to recruit as many T cell progenitors as their mature counterparts (Fig. 6a), they may already acquire some capacity to activate T cells during differentiation. However, this progenitors recruitment is not associated with mature-like induction of proliferation of T cells, suggesting that signals delivered by patients' iDC to T cells were partly defective. Analysis of the efficiency of DC signal on T cell responses has shown that signal strength was determinant for the proliferative response of T cell and may depend on the dose of antigen presented, the co-stimulatory activity or the duration of DC–T cell interaction [61].

Persistence of unusual T cell activation abilities after 9 days in tissue culture media suggests an incomplete recovery of the differentiation potential in patient monocytes. Exposure to the anti-inflammatory environment in patients may thus hinder the functional properties of the cells long-term, with implications for strategies of DC therapies in septic patients [60,62].

In conclusion, this study demonstrated that monocytes from SIRS or septic surgical patients are able to differentiate in vitro into functional DC. The largely admitted concept of immune cells depression after surgery and in the context of systemic inflammation and sepsis might, in large part, be more related to active control of cells by their microenvironment in vivo, as opposed to an intrinsic status. Functional defects in mature DC to activate T cells may, however, point to some long-lasting effects determined probably by exposure to the unique inflammatory/anti-inflammatory balance found in septic patients.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was supported by the Association pour la Recherche sur le Cancer (grants 7639 and 4286) to A. Haziot, by Albsten (FP6 503319) to D. Charron, and plan quadriennal EA322 Université Paris 7, Ministère de l'Education et de La Recherche to V. Faivre, A. C. Lukaszewicz and D. Payen. This work was part of the PhD thesis of V. Faivre.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References