Differential down-regulation of HLA-DR on monocyte subpopulations during systemic inflammation

Patients scheduled for abdominal aortic surgery (AAS) and carotid artery surgery (CAS) were recruited at the Pitié-Salpêtrière Hospital after approval of the study protocol by the Ethics Committee for Human Research of this hospital (Session of April 4^th, 2007). The following patients were excluded: those undergoing coeloscopic surgery or surgery on the thoracic aorta, those with signs of pre-operative infection, undergoing chronic dialysis, under anti-inflammatory medication or antibiotic treatment before surgery, presenting an on-going or neoplastic hematologic pathology, or in an immunodepressed state. Finally, 20 AAS patients (17 males and 3 females; age 67.0 ± 2.9 years) and 20 CAS patients (13 males and 7 females; age 73.9 ± 2.8 years) were included in this study. There were no significant differences in age or proportion of gender between the two surgery groups. The two groups showed similar medical history (that is, hypertension, diabetes mellitus, angina pectoris, myocardial infarction, heart failure, coronary bypass, chronic obstructive pulmonary disease, renal failure). The protocol followed for preoperative medication and anesthesia was similar in both groups of patients. The only difference was that treatment with anti-platelet aggregation agents was discontinued five days before surgery for AAS patients, whereas it was continued until the day of surgery for CAS patients. The usual premedications were maintained except for converting enzyme inhibitors and angiotensin II antagonists, which were discontinued the day before surgery. All patients were premedicated with 5 mg of midazolam given orally one hour before surgery. During the operative period, all patients were anesthetized by target-controlled infusion of propofol, sufentanil, and cisatracurium. Antibioprophylaxis was performed using cefamandole. Depending on patient hemodynamics and hematocrit, fluid loading was performed using crystalloid infusion (lactated Ringer's solution or isotonic saline) and colloid infusion (hydroxyethylstach 130/0.4), associated with blood transfusion if necessary to maintain hemoglobin levels above 10 g/dl. Approximately 30 minutes before the end of surgery, all patients received paracetamol for postoperative analgesia, which was completed in the recovery room with intravenous morphine until pain relief was achieved. Healthy volunteers were recruited (ICaReB) in order to determine the main mediators responsible for the down-regulation of HLA-DR expression on CD14^HIGH and CD14^LOW monocyte subpopulations (n = 9, three males and six females; age 37 ± 5 years). Informed consent was obtained from each patient and volunteer.

Blood samples from patients were collected into sodium citrate vacuum tubes as follows: immediately before anesthesia induction (T₁); before incision (T₂), before vascular clamping (aortic clamping (AAS patients) or carotid artery clamping (CAS patients)) (T₃), after blood reperfusion (T₄) during the surgery, and on postoperative Days 1 (POD1) and 2 (POD2) after the surgery. Blood samples from some patients were collected on POD4 (CAS patients, n = 7) or POD7 (AAS patients, n = 10). Blood from healthy controls (12 ml/each) was collected into sodium citrate vacuum tubes.

Whole blood (100 μl) was immediately processed for double staining with 20 μl of fluorescein isothiocyanate (FITC)-anti-HLA-DR antibody (Beckman Coulter, Marseille, France) or 20 μl FITC-anti-CD16 antibody (Beckman Coulter) and 4 μl of phycoerythrin (PE)-anti-CD14 antibody (MY4-RD1, Beckman Coulter, Fullerton, CA, USA). For IL-10 receptor expression, 100 μl of whole blood was incubated with 10 μl of FITC-anti-CD14 antibody (MY4, Beckman Coulter), 10 μl of allophycocyanin (APC)-anti-CD16 antibody (Miltenyi Biotec, Bergisch Gladbach, Germany) and 20 μl of PE-anti-IL-10R (CD210) antibody (Biolegend, San Diego, CA, USA). As isotype controls, 2 μl of FITC-mouse IgG₁ or IgG2b (Sigma-Aldrich, St Louis, MO, USA), 10 μl of PE-mouse IgG₂a or IgG2b (Miltenyi Biotec) and/or 10 μl of APC-mouse IgM (Miltenyi Biotec) were used. After 20 minutes of incubation in the dark, 1 ml of lysis buffer (BD FACS™ lysing solution, BD Bioscience, Franklin Lakes, NJ, USA) was added to stained samples to lyse erythrocytes. After a further 10-minute incubation and centrifugation (300 g for five minutes, 4°C), the supernatant was removed and 300 μl of MACS buffer (DPBS with 2 mM EDTA and 0.5% fetal calf serum) was added to cells. The expression of surface markers was immediately measured by flow cytometry (FACScan, BD Bioscience). The settings of the flow cytometer were maintained constant during the whole study, which was performed with the same batch of antibodies for all patients, allowing a similar signal for the monocyte subsets throughout the investigation. The values were expressed as mean fluorescence intensity (MFI). Data analysis was performed using CellQuest software (BD Bioscience, Franklin Lakes, NJ, USA).

Whole blood samples from healthy volunteers were incubated with each or a combination of the following molecules for 24 hours (37°C, 5% CO₂): norepinephrine (MERCK, Lyon, France), acetylcholine (Sigma-Aldrich), vasoactive intestinal peptide (VIP) (Sigma-Aldrich), pituitary adenylate cyclase-activating polypeptide (PACAP) (Sigma-Aldrich), substance P and enkephalin (kind gifts of Dr Catherine Rougeot, Institut Pasteur), transforming growth factor-β (TGF-β) (R&D Systems, Abingdon, Oxfordshire, UK), tumor necrosis factor-α (TNF-α) (R&D systems), interleukin-10 (IL-10) (Genzyme, Saint Paul, MN, USA), prostaglandin E₂ (PGE₂) (Sigma-Aldrich), adrenocorticotropic hormone (ACHT) (Novartis, Rueil-Malmaison, France), glucocorticoid (hydrocortisone, HC) (Sigma-Aldrich), blocker of corticoid receptor (RU486; mifepristone, Sigma-Aldrich), and pathogen-associated molecular patterns (PAMPs) (Pam3CysSK4 (EMC microcollection, Tübingen, Germany), muramyl dipeptide (MDP; Sigma-Aldrich), E. coli lipopolysaccharide (LPS; Alexis, Enzo Life Sciences Inc., Farmingdale, NY, USA)). The cells were then stained and flow cytometry was performed following the procedure described above.

Plasma levels of cortisol before anesthesia (T₁) and at POD1 were measured using enzyme immunoassays (AbCys S.A., Paris, France). Plasma levels of IL-10 were measured by an enzyme-linked immunosorbent assay (ELISA) (DuoSet, R&D Systems). The assays were carried out according to the manufacturer's instructions. The resultant color reaction was read using a MRX ELISA microplate reader (Revelation, DYNEX, Magellan Science, Gaithersburg, PA, USA) at 450 nm.

Whole blood from healthy volunteers was centrifuged and the plasma replaced with that from AAS or CAS patients collected after blood reperfusion (T₄). The samples were incubated for 24 hours (37°C, 5% CO₂). In some samples, RU486 (20 μM), a glucocorticoid receptor antagonist, was added simultaneously. The cells were stained and flow cytometry was then performed following the procedure described above to determine HLA-DR expression. The results, expressed as the mean of % change of HLA-DR expression, were compared to HLA-DR expression after 24 hours incubation at 37°C in the presence of autologous plasma.

Whole blood was subjected to Ficoll separation (MSL, Les Ullis, France) in order to isolate peripheral mononuclear cells (PBMCs). Monocytes were isolated from PBMCs using MACS CD14 magnetic beads (Miltenyi Biotec), and total RNA was extracted using the RNeasy miniprep kit (Qiagen, Valencia, CA, USA) following the manufacturer's protocol. cDNA was generated by reverse transcription as previously described [31]. Quantitative real-time polymerase chain reaction (qPCR) was performed on a Stratagene MX3005P^® using Brilliant^®II SYBR^®Green qPCR Master mix (Agilent Technologies, Massy, France), and 10 μM of each primer (custom synthesis by Sigma Oligo, St Louis, MO, USA). Primer sequences for MARCH1 are the following: hMARCH1 E1-258 F1 TCCCAGGAGCCAGTCAAGGTT, hMARCH1 E2-385 R1 CAAAGCGCAGTGTCCCAGTG[23]. The PCR consisted of 40 cycles at 94°C for 40 sec, 58°C for 30 sec and 72°C for 40 sec. The specificity of the SYBR green-amplified product was confirmed by dissociation curve analysis. Transcript levels for the MARCH1 gene were normalized against those of the housekeeping gene GAPDH [32].

As shown in Table 1, the group of patients who underwent AAS was not statistically different from those who underwent CAS. In contrast, all parameters linked with this type of surgery indicate that AAS was more severe than CAS in terms of duration, length of clamping, blood loss, transfusion, and translocation of microbial products from the gut [33, 34]. This difference in severity was illustrated by significantly higher levels of markers of inflammation (IL-6, C-Reactive Protein (CRP)) one day after surgery in AAS patients. In addition, post-operative complications were more frequent in AAS: nine AAS patients and three CAS had cardiac and/or pulmonary complications (P = 0.038). Infections occurred in some of the patients with post-operative infection without reaching statistical significance between AAS and CAS patients.

Monocytes were analyzed after exclusion of the other cells using side scatter (SSC) and forward scatter (FSC) parameters (Figure 1A). We also checked that CD14^HIGH monocytes were CD16^- and that CD14^LOW monocytes were CD16⁺, as previously reported [35] (Figure 1B). Similar patterns were obtained for patients before surgery and for healthy controls (data not shown). HLA-DR expression before surgery in AAS and CAS patients was not significantly different from that measured in healthy volunteers. HLA-DR MFI absolute values for AAS and CAS before anesthesia and healthy controls on CD14^HIGH monocytes were 282 ± 31, 285 ± 23 and 211 ± 37, respectively. HLA-DR MFI absolute values for AAS and CAS before anesthesia and healthy controls on CD14^LOW monocytes were 426 ± 34, 418 ± 31 and 452 ± 31, respectively. Figure 1C shows a representative flow cytometric analysis of HLA-DR expression on CD14^HIGH and CD14^LOW monocyte subpopulations monitored at T₁ (before anesthesia) and POD1 for two AAS patients. It can be seen that HLA-DR expression on CD14^HIGH and CD14^LOW monocytes at POD1 was decreased as shown by a leftward shift as compared to levels at T₁.

At T₁, the percentage of CD14^HIGH monocytes was of 6.02 ± 0.45 and 5.99 ± 0.69, and that of CD14^LOW was of 0.80 ± 0.06 and 0.78 ± 0.08 for AAS and CAS patients respectively. These values were similar to those obtained with healthy donors, and didn't vary significantly during the survey (Figure 2).

Figure 3A shows that in AAS patients, HLA-DR expression on CD14^HIGHcells was reduced after blood reperfusion (T₄) (-49%) and continued to decrease by POD1 (-60%) and POD2 (-73%). In contrast, the expression of HLA-DR on CD14^LOW cells did not change at T₄, and was only reduced at POD1 (-40%) and POD2 (-51%) (Figure 3B). The kinetics for the two subpopulations were significantly different (P < 0.01). On POD4 or POD7, HLA-DR expression on both subpopulations returned to close to the normal range as measured before surgery. AAS patients were compared with a second group of patients (CAS) for whom the inflammatory insult was less severe. For CAS patients, the kinetics of the reduction of HLA-DR expression on CD14^HIGH and CD14^LOW cells were not different of that of AAS patients, but the reduction was significantly less severe.

We measured plasma levels of cortisol and IL-10 in AAS and CAS patients in order to investigate their possible relationship with the modulation of HLA-DR expression on both CD14 monocyte subpopulations. Plasma cortisol levels before anesthesia (T₁) were not significantly different among AAS patients (110.8 ± 8.7 ng/ml), CAS patients (104.8 ± 10.8 ng/ml) or healthy controls (95.5 ± 10.7 ng/ml). Both groups of patients showed a significant increase in plasma levels of cortisol at POD1 (Table 1), in agreement with previous studies, which reported peak level of cortisol one day after surgery [36, 37]. A negative correlation between the percent change in the levels of plasma cortisol and the percent change in HLA-DR expression on CD14^HIGH and CD14^LOW monocytes was observed when comparing values obtained at T₁ and POD1 (Figure 4).

Plasma levels of IL-10 were measured during the observational period in both groups of patients. The levels were almost undetectable until T₃ in both groups. A peak of IL-10 in the AAS group occurred at T₄ (mean ± SEM (standard error of the mean) = 30 ± 9 pg/ml; median = 15.0 pg/ml) and IL-10 could be still detected at POD1 (mean ± SEM = 23 ± 6 pg/ml, median = 17.5 pg/ml), whereas IL-10 was below the detection limit in most CAS patients. Thus, the relationship between changes in HLA-DR expression and IL-10 levels could only be analyzed in AAS patients. However, the absence of detectable IL-10 in CAS patients does not mean that there was no IL-10 produced, since detectable circulating cytokines only represent the tip of the iceberg [38]. A significantly negative correlation between IL-10 levels and the alteration of HLA-DR expression could be obtained for CD14^HIGH monocytes (r = -0.465, P = 0.039, when comparing T₁ and T₄, and r = -0.516, P = 0.030, when comparing T₁ and POD1), but not for CD14^LOW monocytes.

In order to identify the mechanism of HLA-DR down-regulation on CD14^HIGH and CD-14^LOW monocytes, we tested various mediators produced during stress that are known to interfere with the immune response. Blood from healthy volunteers was incubated with each molecule or a combination of the molecules for 24 hours, and HLA-DR expression on both CD14^HIGH and CD14^LOW monocytes was analyzed by flow cytometry. IL-10 down-regulated HLA-DR expression on CD14^HIGH monocytes, but not on CD14^LOW cells (Figure 5). Hydrocortisone (HC) down-regulated HLA-DR expression on both monocyte subpopulations, and the effect of HC was inhibited by the glucocorticoid receptor antagonist RU486 (Figure 5). As the effect of IL-10 was not the same on the two monocyte subsets, we investigated whether this could be linked to a different expression of the receptor for IL-10 (IL-10R). The expression of the IL-10R was analyzed by flow cytometry on monocytes from healthy volunteers. As shown in Figure 6, its expression was significantly higher on the CD14^HIGH CD16^-, the population that was sensitive to IL-10 effects in vitro.

Neurotransmitters are among the other mediators associated with stressful situations. However, none of the tested neuromediators (norepinephrine, acetylcholine, vasoactive intestinal peptide, pituitary adenylate cyclase-activating polypeptide, substance P, and enkephalin) alone or together had any effect on HLA-DR expression (Figure 6). Similarly, prostaglandin E2, adrenocorticotropin hormone or TGFβ had no effect. In contrast, TNFα increased the expression of HLA-DR, particularly on CD14^HIGH cells (CD14^HIGH: + 263%; CD14^LOW: + 87%) (Figure 7).

We previously showed that translocation of microbial products occurs in AAS patients [33, 34]. Thus, we studied the capacity of several microbial products, including agonists of TLR2, TLR4 or NOD2 (Pam3CysSK4, LPS and MDP, respectively), to modulate the expression of HLA-DR on monocytes. As shown in Figure 8, HLA-DR expression on both CD14^HIGH and CD14^LOW monocytes was up-regulated by these pathogen-associated molecular patterns (PAMPs), particularly on CD14^HIGH cells. We then investigated the capacity of IL-10 or hydrocortisone to counteract the effects of these PAMPs. As compared to IL-10, hydrocortisone had a greater capacity to reduce the up-regulation of HLA-DR induced by PAMPs, especially on CD14^HIGH monocytes. When exposed to both IL-10 and HC, the levels were further reduced.

We then investigated whether the plasma of AAS patients could suppress the expression of HLA-DR on CD14^HIGH and CD14^LOW monocyte subpopulations of healthy donors. Whole blood from healthy volunteers was incubated for 24 hours at 37°C, after replacing their plasma with that from AAS patients (n = 13) sampled at POD1. Decreased expression of HLA-DR was obtained for 69% of the plasma samples (9 out of 13 patients), with a 42 ± 9% decrease for CD14^HIGH and a 11 ± 8% decrease for CD14^LOW monocytes. In addition, for four AAS plasma, which led to a 37.3 ± 10.4% decrease of HLA-DR expression onto CD14^HIGHmonocytes from healthy controls, the co-incubation with RU486, a glucocorticoid receptor antagonist, fully abolished the inhibitory effect of the active AAS plasma (data not shown).

Because of the known role of MARCH1 on HLA-DR trafficking [24, 25], we investigated whether MARCH1 was modulated by glucocorticoids. As shown in Figure 9A, incubation of whole blood from healthy donors with hydrocortisone led to a three-fold increase in MARCH1 gene expression in their monocytes after 24 hours of incubation. Similarly, in AAS patients MARCH1 gene expression was increased at POD1 as compared to T₁ (Figure 9B).