The relationship between CD4⁺CD25⁺CD127^- regulatory T cells and inflammatory response and outcome during shock states

All consecutive patients newly admitted to the medical ICUs of two French teaching hospital between January 2007 and June 2007 were prospectively enrolled in the study if they were suffering from shock whatever its etiology. Septic shock was defined by the standard criteria of the American College of Chest Physicians/Society of Critical Care Medicine consensus conference [14], and non-septic shock states were defined according to definitions used by Antonelli and colleagues [15]. Patients who were immunocompromised were not enrolled (treatment with corticosteroids > 0.5 mg/kg equivalent prednisolone, bone marrow or organ transplant recipients, leukopenia (white blood cell count < 1000/mm³) or neutropenia (polymorphonuclear granulocyte count < 500/mm³), hematological malignancy or AIDS). Patients who presented with early death or discharge (within 12 hours after admission) were also excluded. Seven healthy volunteers from the laboratory served as controls: four men, three women, mean age 37.5 ± 3.6 years, without chronic or acute diseases. Approvals from the Comité de Protection des Personnes-EST III and the ethical committee of the Société de Réanimation de Langue Française were obtained, and the informed consents of the patients and volunteers were sought prior to inclusion.

Upon admission to the ICU, the following data were recorded for each patient: age; sex; severity of underlying medical condition according to the criteria of McCabe and Jackson [16]; Simplified Acute Physiology Score (SAPS) II [17]; Sepsis-related Organ Failure Assessment (SOFA) score [18]; reason for admission into the ICU; principal diagnosis; vital signs; respiratory parameters; routine blood tests and microbiological culture results. Survival or death in the ICU was assessed during a follow-up period of up to 28 days.

Within 12 hours upon admission, peripheral whole blood was collected on citrated tubes. Sampling was repeated at day three, five and seven provided that the patient was still in the ICU.

Flow cytometry analyses were conducted on a Coulter Cytomics FC500 cytometer (Beckman-Coulter, Hialeah, FL, USA). All the monoclonal antibodies and the isotypic controls (except anti-Foxp3) used were from Immunotech (Marseille, France): PE-Texas Red (ECD)-labelled anti-CD3, fluorescein isothiocyanate (FITC)-labelled anti-CD14, phycoerythrin (PE)-labelled anti-HLA-DR (clone IM 1639), FITC-labelled anti-CD3, PC5-labeled anti-CD56, PC5-labelled anti-natural killer (NK) G2D, PC5-labeled anti-CD4, FITC-labelled anti-CD25 and PE-labelled anti-CD127; the Foxp3 monoclonal antibody was from eBiosciences (San Diego, CA, USA). After red blood cells lysis (Q-prep system, Beckman Coulter, Hialeah, FL, USA), naturally occurring Tregs were characterized by the expression of CD4 and CD25 and the lack of expression of CD127. We also characterized the number of monocytes and NK cells based upon the expression of the usual markers (CD45, CD14, HLA-DR, CD16, and NKG2D, respectively). Results are expressed as percentages of the CD4⁺ lymphocyte population and as number of cells per microliter of whole blood.

Plasma concentrations of IL-2, IL-4, IL-5, IL-10, interferon (INF) γ and TNFα were quantified through fluorescence activated cell sorting (FACS) by the use of human Th1/Th2 CBA kit (BD Biosciences, San Diego, CA, USA) according to the recommendations of the manufacturer. Intra- and inter-assay coefficients of variation were less than 8%. Concentration of TGFβ was determined by ELISA (R&D Systems, Lille, France).

All experiments were approved by the Institutional Animal Care and Use Committee. Adult (5 to 6 weeks, 20 to 23 g) male Balb/c mice were purchased from the Centre d'élevage Dépré (Saint-Doulchard, France).

For induction of polymicrobial sepsis, mice were anesthetized by intraperitoneal administration of ketamine and xylazine in 0.2 ml sterile pyrogen-free saline. The cecum was exposed through a 1.0 cm abdominal midline incision and subjected to ligation of the distal half followed by two punctures with a 21G needle. A small amount of stool was expelled from the punctures to ensure patency. The cecum was replaced into the peritoneal cavity and the abdominal incision closed in two layers. After surgery, all mice were injected subcutaneously with 0.5 ml of physiologic saline solution for fluid resuscitation and subcutaneously every 12 hours with 1.25 mg (i.e., 50 μg/g) of imipenem. A sham operation was performed by isolating the cecum without ligation or puncture.

When specified, animals were given either an intraperitoneal injection of 500 μg of affinity purified CD25 (PC61 hybridoma)-depleting antibody (BioExpress, West Lebanon, NH, USA) or an identical volume (250 μl) of sterile normal saline 72 hours before the caecal ligation and puncture (CLP). Administration of the antibody (250 μg) or normal saline was repeated two hours before the procedure.

Animals were observed for up to seven days to determine survival. For certain experiments, 4 to 6 mice were euthanized at 24, 48 or 96 hours after surgery to harvest blood or splenocytes. Among splenocytes, the Tregs were considered to be the CD4⁺CD25⁺Foxp3⁺ cells on FACS.

At indicated times, plasma concentrations of TNF-α, IL-1β and IL-6 were measured by ELISA.

Changes in time or between groups were analyzed by one-way analysis of variance. Comparisons between patients with septic shock, patients with non-septic shock and/or healthy volunteers were performed by using Student's t test as normally distributed values (Kolmogoroff-Smirnov test), Mann-Whitney U test, or Kruskal Wallis test when indicated. Data are expressed as mean ± standard deviation or median ± interquartile range when appropriate. Correlations were established using Spearman's test. Survival analysis (mice) was performed using the log rank test after drawing the Kaplan Meier survival curves. Analysis was completed with the Statview software (Abacus Concepts, Berkeley, CA, USA), with a two-tailed P value less than 0.05 considered as statistically significant.

Naturally occurring Tregs were defined as CD4⁺CD25⁺CD127^- cells (Figure 1a).

At day one, the absolute number of Tregs was lower in patients than in healthy volunteers (Figure 1b and Table 2) without any difference between the two groups of patients. When expressed as a percentage of CD4⁺ lymphocytes, no differences were observed between patients and volunteers on admission (Figure 1c and Table 2).

By day three, both Treg numbers and percentages progressively increased in both groups of patients, especially the septic ones, although there was no statistical difference between the two groups of patients (Figures 1b and 1c). At day seven, although the percentage of Tregs was higher in the septic patients than in the non-septic patients and healthy volunteers, there were no differences in terms of absolute count, reflecting the persistence of the CD4⁺ lymphopenia in the septic shock group.

We next sought to evaluate the relation between Treg kinetics and survival. Considering the whole group of patients, or those suffering from a non-septic shock, there were no differences between survivors and non-survivors in terms of percentage or number of Tregs. By contrast, within the septic shock patients, those surviving had constantly higher Tregs counts and percentages than non-survivors (Figure 2).

Considering the correlation between Tregs and outcome, at least in septic patients, we evaluated the relation between these cells and severity. Indeed, there was an inverse correlation between the percentage of Tregs and SAPS II (not shown), SOFA score or arterial lactate level (Figure 3).

By contrast, we were unable to find any correlation between Tregs and plasma cytokine (IL-2, IL-4, IL-5, IL-10, INFγ, TNFα and TGF-β) concentrations, either taken individually or expressed as Th1/Th2 ratios (data not shown). This absence of correlation was found in septic as well as non-septic patients.

NK cells have been advocated as an important component influencing outcome during septic shock [18].

We observed that, although the percentage of NK cells was not different between patients and healthy volunteers throughout the study period, patients with shock presented with constantly decreased NK absolute numbers (Figure 4). As we observed for the Treg counts, there were no differences between septic and non-septic shock patients, or between survivors and non-survivors. There was no correlation between the percentages or absolute counts of Tregs and NK cells.

To get further insight in to the role of Tregs during septic shock, we used a CLP-induced model of peritonitis in mice. We first observed that as early as 24 hours after surgery, there was an increase in the percentage of Tregs among splenocytes (CD4⁺CD25⁺Foxp3⁺ cells) of septic mice as compared with the sham-operated ones (Figure 5).

We next sought to investigate the influence of Tregs on survival. We depleted animals of CD25⁺ cells by using repeated injection of anti-CD25 antibody. Depletion was highly effective with less than 0.2% of CD25⁺ cells left 24 hours after the last antibody injection (Figure 6). This Treg depletion persisted at day four but was not checked thereafter.

Animal survival was not affected by the CD25⁺ cells status with 40% of CD25-depleted mice surviving as compared with 31% for the non-depleted group (P = 0.62; Figure 7).

Finally, plasma concentrations of TNF-α (Figure 8), IL-1β or IL-6 (not shown) were not different between the two groups of mice.