Introduction
Acute pancreatitis (AP) is usually a self-limiting disease resolving within days. Some patients, however, develop overwhelming systemic inflammation, which contributes to the development of vital organ dysfunction, the major cause of mortality in AP [
1,
2]. Systemic inflammation is designated by activation of circulating cells of both the innate immune system, such as monocytes [
3], and the adaptive immune system, such as CD4+ T-helper (Th) -lymphocytes, CD8+ -lymphocytes [
4,
5], and CD19+ B -lymphocytes [
6]. Cellular activation results in the systemic release of pro- and anti-inflammatory mediators, such as tumor necrosis factor (TNF) and interleukin (IL) -10 [
7]. The latter promotes immune suppression, increasing the risk of secondary infections and multiple organ dysfunction syndrome [
8‐
11]. Also, experimental [
12] and clinical [
13] studies suggest that the host's defense against infection is further depressed as the Th1 subpopulation of CD4+ T-cells becomes more strongly suppressed than the Th2 cells in the course of AP, leading to Th1/Th2 cytokine imbalance.
The molecular mechanisms involved in the pathogenesis of systemic inflammation and subsequent immune suppression include multiple signaling pathways and families of transcription factors, such as signal transducers and activators of transcription (STATs), nuclear factor-κB (NFκB), and members of the mitogen-activated protein (MAP)-kinase family [
14]. In lymphocytes, STAT1 is activated by IL-6 and pro-inflammatory interferons [
15,
16], which support Th1 polarization of Th-lymphocytes, STAT3 by anti-inflammatory IL-10 [
17] and by IL-6, which supports the Th17 lineage of lymphocyte differentiation [
18] and STAT6 by IL-4, which supports Th2 polarization of Th-cells [
19]. The lymphocyte NFκB is activated by TNF [
20,
21], while MAP-kinases ERK (extracellular signal-regulated kinase) 1/2 and p38 are phosphorylated by cytokine receptor activation [
22], or by co-stimulation of lymphocytes with calcium ionophore and PMA (phorbol 12-myristate 13-acetate).
We recently described monocyte signaling profiles in 13 AP patients with vital organ dysfunction using phospho-specific whole blood flow cytometry [
23]. In the present study we describe signaling profiles of the patients' circulating lymphocytes.
Material and methods
Patients and controls
The study comprises 16 men with AP admitted to the intensive care unit (ICU) at Helsinki University Central Hospital. The first 13 patients, whose monocyte signaling profiles are described in our previous study [
23], were admitted between September 2007 and January 2009 and the last three patients between January and May 2010 (Table
1). In addition to severe AP, the inclusion criterion for the study was that the proportion of HLA-DR-positive monocytes in circulation was less than 80%. Sixteen healthy volunteers (median age 45 years, range 25 to 66, 13 women) served as reference subjects. The study protocol was approved by the Surgical Ethical Review Board of the Joint Authority for the Hospital District of Helsinki and Uusimaa, and informed consent was obtained from each patient, or their next of kin.
Table 1
Characteristics of the patients
Character | Median (range) |
Age, years | 47 (30 to 74) |
ICU stay, days | 23 (4 to 54) |
Respirator, days | 18 (0 to 42) |
Dialysis, days | 9 (0 to 40) |
Sampling time1, days | 13 (3 to 39) |
CRP at transfer to ICU, mg/l | 230 (3 to 566) |
APACHE at transfer to ICU | 14 (8 to 20) |
SOFA at transfer to ICU | 5 (4 to 13) |
CRP2, mg/ml | 215 (80 to 451) |
APACHE II2 | 15 (6 to 24) |
SOFA2 | 8 (2 to 17) |
Highest CRP (mg/l) | 409 (207 to 566) |
The diagnosis of AP was based on typical clinical findings (acute onset of epigastric pain, nausea, and vomiting), elevated serum amylase levels (more than three times the upper reference limit), and/or typical findings in computed tomography performed on all patients within zero to two days after admission to ICU. The AP severity was determined according to the Atlanta classification [
24].
Blood samples and study design
Blood samples were collected within three to five days after the initial determination of the patients' monocyte HLA-DR expression status. Each patient had a reference subject of his own, with the exception of two patients who shared one reference subject. Of the two data sets from this reference subject, the latter was excluded from the data analyses. Parallel peripheral blood samples were obtained from the patient and from his reference subject on two consecutive days, with the exception of one patient, who had different reference subjects on Days 1 and 2. On Day 1, a 4-ml sample of peripheral venous blood was taken (i) for the study of phosphorylation promoted by soluble leukocyte agonists and (ii) for the re-study of the proportion of HLA-DR-positive monocytes. On Day 2, a 4-ml blood sample was taken for the study of leukocyte phosphorylation in response to bacterial cells. The blood samples were collected into Falcon polypropylene tubes (Becton Dickinson, Lincoln Park, NJ, USA) containing pyrogen-free citrate phosphate dextrose (ACD, Baxter Health Care Ltd, Norfolk, England, UK, 0.14 ml/ml blood), kept at the room temperature, and transported within 15 minutes to the laboratory. All aliquots of the blood samples were stimulated within four hours of sampling.
While the study was in progress the patients' lymphocytes showed constitutive STAT3 activation, that is, STAT3 RFU-values in resting cells were consistently higher in patients than healthy subjects. The original protocol was adapted to measure constitutive STAT3 activation and determinate the involved lymphocyte subsets. The revised protocol was applied to patients 11 to 13 and the STAT3 part of the revised protocol to patients 14 to 16 (Table
1).
Biological agents and leukocyte agonists
We purchased fluorescein isothiocyanate (FITC)-conjugated monoclonal antibodies (mAbs) to CD14 clone MϕP9 (IgG2b), CD19 clone SJ25C1 (IgG1), CD4 clone SK3 (IgG1), and CD8 clone SK1 (IgG1), and peridinine-chlorophyll protein (PerCP)- conjugated mAb to CD3 clone SK7 (IgG1), phycoerythrin (PE)-conjugated anti-HLA-DR mAb clone L243 (IgG2a) and its isotype control (mouse IgG2a), Alexa 647-labeled phospho-specific mAbs to pNF-kB p65 (pS529) clone K10-895.12.50 (IgG2b), STAT1 (pY701) clone 4a (IgG2a), ERK1/2 (pT202/pY204) clone 20A (IgG1), and PE-labeled phospho-specific mAbs to STAT3 (pY705) clone 4/P-STAT3 (IgG2a) and STAT6 (Y641) clone18 (IgG2a) from Becton-Dickinson Biosciences (San Jose, CA, USA). Infliximab (Remicade®) was from Schering-Plough Co. (Kenilworth, NJ, USA) and anakinra (Kineret®) from Amgen Inc. (Thousand Oaks, CA, USA).
The recombinant cytokines TNF, IL-4, and IL-6 were purchased from R&D (Minneapolis, MN, USA) and phorbol-12-myristate-13-acetate (PMA), E. coli O111:B4 lipopolysacharide (LPS), N-acetylmuramyll-alanyl-D-isoglutamine (MDP), and calcium ionophore A23187 from Sigma (St. Louis, MO, USA).
S. aureus (IHT 61972) and E. coli (IH 3080) were kind gifts from Jaana Vuopio, MD, PhD (The National Institute for Health and Welfare, Helsinki, Finland). S. epidermidis was the strain ATCC 53103. Each strain was initially grown in a brain heart infusion (BHI) medium for 18 hours at room temperature (E. coli) or at 37°C (S. aureus and S. epidermidis). After incubation, the bacteria were pelleted by centrifugation, re-suspended in fresh BHI medium, and further cultured at 37°C for five hours. The secondary cultures were pelleted and washed twice with phosphate buffered saline (PBS). From an aliquot of each cell suspension a diluted culture was made to quantify viable bacteria. The rest of the bacteria were pelleted and re-suspended in a glycerol-tryptone soya broth medium and stored in 1.8-ml aliquots at -70°C.
Ex vivo stimulation and immunolabeling of blood samples for three-color flow cytometry
Flow cytometry was first used by Fleisher and co-workers [
25] to demonstrate intracellular phosphorylation of STAT1 in density gradient-separated monocytes activated
ex vivo with IFN-γ. We recently developed a whole blood modification of the method [
26]. In the present setup, lymphocytes were first delineated according to their light scattering properties using electronic gates and monocytes were recognized by CD14-FITC label positivity and excluded. To study lymphocyte subpopulations, the CD19-FITC-label was used to identify B -lymphocytes and the CD3-PerCP-label was used in combination with the CD4-FITC- or CD8-FITC-labels to identify CD3+CD4+T lymphocytes and CD3+CD8+T lymphocytes. The Alexa647-label was used to detect phosphorylated forms of NF-κB, STAT1, and ERK1/2 and the PE-label to detect phosphorylated STAT3 and STAT6.
The 4-ml blood sample was divided into 90 μl aliquots in Falcon polystyrene tubes (Becton Dickinson, Lincoln Park, NJ, USA) and placed at 37°C. Aliquots of FITC-conjugated mAbs to CD14, CD4, and CD8 were added to the tubes. Next the tubes were supplemented with TNF at the final concentration of 10 ng/ml (10 patients) and incubated for 5 minutes at 37°C, IL-6 100 ng/ml (13 patients) for 5 minutes, IL-4 100 ng/ml (10 patients) for 5 minutes, a combination of PMA 1 μM and calcium ionophore A23187 1 μM (10 patients) for 5 minutes, LPS 100 ng/ml (10 patients) for 10 minutes, MDP 100 ng/ml (10 patients) for 20 minutes, E. coli 50 cells/leukocyte (10 patients) for 10 minutes, S. aureus 50 cells/leukocyte (9 patients) for 20 minutes, or S. epidermidis 50 cells/leukocyte (9 patients) for 40 minutes. Reference tubes were left without stimulus. In one experimental series infliximab (final concentration 10 μg/ml), anakinra (100 μg/ml), or both were mixed with the blood obtained from healthy volunteers before addition of bacteria.
After incubation, red cell lysis, leukocyte fixation, and leukocyte permeabilization were performed according to BD Phosflow Protocol III for Human Whole Blood [
27], as described in detail. Briefly, a 1× BD PhosFlow Lyse/Fix Buffer (1.9 ml) pre-warmed to 37°C was added to each tube, they were then incubated for 10 minutes at 37°C, and washed once with Hank's balanced salt solution (Life Technologies, Paisley, UK). For permeabilization, the cell pellet was re-suspended in 1 ml of BD Phosflow Perm Buffer III, pre-cooled at -20°C. The tubes were stored at -20°C until staining with phosphospecific mAbs. After permeabilization, the cell samples were washed twice with ice-cold BD Pharmingen Stain Buffer (BD Sciences, San Jose, CA, USA) and re-suspended in 100 μl of the buffer. Aliquots of Alexa647- and PE-labelled phosphospecific mAbs were then added to the stimulus-treated samples and respective reference samples. Also, aliquots of CD3-PerCP mAb were added at this stage, because PerCP did not withstand the permeabilization procedure. The samples were further incubated in the dark for 30 minutes at 0°C, washed once, and re-suspended in 500 μl of the ice-cold stain buffer. The samples were stored at 0°C and analyzed by flow cytometry within three hours.
Data acquisition and analysis were done with a FACS Calibur flow cytometer and Cell Quest software (BD Sciences, San Jose, CA, USA). A total of 1×104 lymphocytes, 2×103 CD3+CD4+ T lymphocytes, 2×103 CD3+CD8+ T lymphocytes, and 2×103 CD19+B lymphocytes were collected. Finally, appropriate Alexa647- and PE-histograms were developed to determine the intracellular signaling profiles of the cells.
First, data are presented as median fluorescence intensity, in other words, median RFU (relative fluorescence units) of the whole lymphocyte population, and second, as a proportion of positively fluorescing cells and third, as mean fluorescence intensity of the positively fluorescing cells. If a patient's lymphocytes respond normally, but in reduced number compared to healthy subjects, the mean fluorescence intensity of positively fluorescing cells would remain the same. On the other hand, if a patients' lymphocytes were in a reduced phosphorylation state, this would lead to decreased mean fluorescence intensity of positively fluorescing cells. The proportion of positively fluorescing cells was determined using a threshold method, where an electronic gate was manually set to include the brightest 2 to 4% of the cells in non-stimulated sample. Then the same gate served to determine the proportion of positively fluorescing cells in the respective stimulus-treated sample. Thus, the values < 5% indicate cells not responding to the stimulus. During the study, it emerged that in a proportion of a patient's lymphocyte STAT3 was constitutively active. To determine the proportion of lymphocytes showing constitutive activated STAT3 lymphocytes, the electronic gate of non-stimulated lymphocytes of the healthy subject studied concomitantly was used in the analysis of the patients' samples.
The coefficient of variation was 5% within experiments and 10% between experiments.
The activation of signaling proteins have been confirmed using Western blot analyses by us [
26] concerning STAT1 and by Grammer [
28] concerning multiple components of NFκB, STAT and MAP-kinase pathways.
Monocyte surface expression of HLA-DR, expressed as the proportion (%) of monocytes positive for HLA-DR fluorescence, was determined as described previously [
10].
Statistical analysis
Results are shown in mean ± SEM or median (range). Using the nonparametric Wilcoxon-Mann-Whitney test we carried out statistical comparisons between the groups. Statistical analyses were performed with SPSS software (v15, SPSS Inc, Chicago, IL, USA). Probabilities were regarded as statistically significant at the 0.05 level.
Discussion
Our results show multiple aberrations in pro- and anti-inflammatory signaling pathways of lymphocytes, determined by phospho-specific whole blood flow cytometry, in AP patients with vital organ dysfunction and immune suppression. Patients' lymphocytes showed decreased NFκB phosphorylation in whole blood samples supplemented with exogenous TNF. The defect appeared to involve all TNF-responding lymphocytes rather than a subset, because the pNFκB histograms were uniform. Both the proportion of pNFκB-positive cells and their fluorescence intensity were also reduced among CD3+CD4+ and CD3+CD8+ T lymphocytes. NFκB phosphorylation promoted by
E. coli,
S. aureus, and
S. epidermidis was also reduced. Lymphocyte activation was probably secondary to phlogistic mediators generated in whole blood samples supplemented with bacteria. Indeed, the TNF-blocking agent infliximab in co-culture with
S. epidermidis, but not with
E. coli or
S. aureus, reduced NFκB phosphorylation, while the IL-1 receptor antagonist, anakinra, had no effect, suggesting that lymphocyte activation involved TNF among other factor(s). We have recently found that monocytes of the patients also showed reduced NFκB phosphorylation in response to bacterial stimuli [
23], which agrees with our finding that TNF production by anergic monocytes is reduced [
29]. Collectively, the above data shows disturbances in collaboration between patients' lymphocytes and monocytes. The impaired collaboration, together with a significant reduction of circulating T- and B-lymphocytes in severe AP [
13,
30] may contribute to the development of secondary infections in the patients. The risk of infections, however, remains to be determined in prospective follow-up studies of leukocyte NFκB phosphorylation profiles in relation to clinical outcome.
In contrast to NFκB activation, the proportion of pp38-positive lymphocytes induced by
E. coli or
S. aureus were higher in patients than reference subjects, indicating that patients had an increased number of responding cells. Of note, the lymphocytes were double-stained with pNFκB and pp38 mAbs, and, consequently, activity of the two signaling pathways could be evaluated simultaneously. Unlike lymphocytes, the p38 phosphorylation of the patients' monocytes was normal [
23]. Given that MAP-kinases are associated with the development of systemic inflammation [
14], our finding raises the question of whether enhanced p38 activation provides a target for immune suppression in AP patients or if it represents a vital counter-reaction of cells to inhibit NFκB, and should therefore be strengthened rather than depressed.
Phosphorylation of the other MAP-kinase protein ERK1/2 was depressed in this study. ERK1/2 is functionally related to the migration of inflammatory cells, including lymphocytes [
31]. Monocytes of our patients showed low pERK1/2 levels and poor transmigration
in vitro, whereas their pp38 levels were normal [
23]. The possibility that the inhibition of p38 expression and upregulation of ERK1/2 expression is beneficial in human AP, as suggested by studies in experimental AP [
32], warrants further studies.
The results show that STAT3 is constitutively activated in the patients' lymphocytes, particularly in CD3+CD4+ and CD3+CD8+ T-lymphocytes. The constitutive STAT3 activation was confined to lymphocytes and did not occur in the patients' monocytes [
23]. Constitutive STAT3 expression has been described in a variety of disorders [
33‐
35] and may involve a complex crosstalk between different signaling pathways [
36]. Because STAT3 mediates anti-inflammatory signals its constitutive activation may denote an attempt to down-regulate inflammation. The activation was partial and could be completed with IL-6, which promoted pSTAT3 levels in patients' lymphocytes similar to those in reference lymphocytes of healthy subjects. Unlike lymphocytes, the IL-6 induced STAT3 phosphorylation of the patients' monocytes was depressed [
23]. Thus, STAT3 signaling pathway was more impaired in monocytes than in lymphocytes. In addition to IL-6, STAT3 is activated by IL-10, and circulating levels of both are elevated in AP patients [
11], thus possibly contributing to the constitutive pSTAT3 expression. In this context it is of interest that IL-6 together with IL-21 and IL-23 promote sustained STAT3 activation, which favors the Th17 developmental program of CD4+ lymphocytes [
18]. Serum levels of IL-17 were also elevated and served as prognostic markers in patients with severe AP [
37]. The Th17 immune pathway has not, to our knowledge, been thoroughly explored in patients with AP. If involved in the pathogenesis of tissue injury in severe AP, the Th17 pathway may reveal novel prognostic markers and therapeutic possibilities.
STAT1 is activated by multiple cytokines [
15,
16] and associated with pro-inflammatory signaling and development of inflammatory tissue injury. We used IL-6 to phosphorylate STAT1 and found lower levels in patients' lymphocytes than in reference cells. The proportion of pSTAT1-positive cells and their fluorescence intensity were also lower, indicating that the defect involved all IL-6-responding lymphocytes rather than a subset of them. Also the patients' monocytes were depressed [
23]. Unlike pSTAT1 levels, the IL-4-induced levels of p-STAT6, which mediates Th2 signals and is associated with less injurious tissue reactions, were higher in patients' cells. The results suggest that the patients' immune system tries to inhibit inflammatory tissue injury by shifting the tissue-destructive Th1 type of immune response to the less injurious Th2 type. Our results are in accordance with the findings regarding experimental AP [
12] and patients with AP, indicating that although both Th1 and Th2 cytokines are elevated, the magnitude of elevation of the latter is much higher [
13].
Taken together, our data concerning the patients' lymphocytes and monocytes indicate that signaling pathways are more impaired in monocytes. The strong depression of monocytes is meaningful because monocytes are powerful mediators of tissue destruction.
Despite our findings being in accordance with clinical findings and current concepts of immune pathogenesis of severe systemic inflammation, the data should be interpreted with caution. First, the study was confined to a limited number of patients, who had vital organ dysfunction and whose clinical outcome and immune inflammatory status were reasonably comparable. We focused on these patients to screen for the affected signaling profiles in the most severe form of AP. In the future, the aberrant profiles of lymphocytes and monocytes [
23] need to be studied prospectively during follow-up of patients with varying AP severity. Second, phospho-specific whole blood flow cytometry is susceptible to methodological errors. A meticulous sample handling is needed to avoid inappropriate cell activation
ex vivo [
23]. Another critical step is permeabilization of the cells, because it permits the phospho-specific antibody molecules to enter the intracellular compartments. Leukocyte activation may increase cellular resistance to membrane-active agents [
38]. Although we cannot exclude the possibility that patient lymphocytes are more resistant to permeabilization, such a difference is not supported by our findings that pSTAT3 levels were constitutively increased in the patient lymphocytes and double-stained lymphocytes showed simultaneously enhanced p38 activation and depressed NFκB activation in the same cell. Considering caveats, our present results and previous studies [
21,
23,
26] suggest that whole blood phosphor-specific flow cytometry is a suitable method for immune monitoring of patients with systemic inflammation.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
JO collected clinical data and participated in data analyses and the writing of the manuscript. HM participated in design of the study and drafting of the manuscript and performed statistical analysis. MLK and LK participated in design and coordination of the study and drafting of the manuscript. KK participated in flow cytometry and drafting of the manuscript. SS was responsible for the flow cytometry and participated in drafting of the manuscript. EK and PP participated in design and coordination of the study and helped draft the manuscript. HR conceived the study, participated in its design, and helped draft the manuscript.