Acute effects of exercise on the inflammatory state in patients with idiopathic pulmonary arterial hypertension

Patients were recruited from the Center of Pulmonary Arterial Hypertension at the University Medical Center Hamburg-Eppendorf with the following inclusion criteria: (1) prevalent diagnosis of IPAH in accordance with the current diagnostic recommendations and classification [9], (2) >18 years of age, (3) World Health Organization functional class (WHO FC) II-III, (4) no evidence for current or past malignant or autoimmune diseases, (5) without exercise limitation due to an orthopedic reason, and (6) without current infectious disease. Healthy subjects underwent echocardiography to rule out signs of right or left heart failure and were matched with regard to age, gender and body mass index (BMI). Cardiopulmonary exercise testing (CPET) was performed as symptom-limited incremental cycling exercise according to the ERS task force (2007) [17]. The rate of workload increase was 10 Watt/min. The anaerobic threshold (AT) was indirectly measured as the point when the VCO₂ − VO₂ slope abruptly increases. The peak oxygen consumption (Peak VO₂/kg [ml/min/kg]) was used for analysis because it represents the most widely used CPET parameter for therapeutic decision making in PAH according to current recommendations [9]. No interruption of drug intake was required and the time between exercise testing and drug intake (PAH targeted therapy) was 2–4 h. The following assessments were obtained from all study subjects at the day of CPET: 6MWD was obtained using a standardized protocol in accordance with the American Thoracic Society statement (Measuring the distance patients can quickly walk on a flat, hard surface 30 m hallway in a period of 6 min. Patients are allowed to stop and rest during the test), functional class according to the modified WHO classification was assessed by the attending physician, N-terminal pro-hormone of brain natriuretic peptide (NT-proBNP) concentrations were measured using a commercially available assay (Dimension Vista, Siemens Healthcare Diagnostics, IL, USA; Catalogue No. K6423A), echocardiography and pulmonary function test (PFT) including spirometry, body plethysmography and gas dilution.

Peripheral venous whole blood samples were taken at rest before, directly after and 1 h after CPET. Differential blood count was obtained (Coulter Ac-T diff 2, Beckmann Coulter, Krefeld, Germany) for absolute and relative numbers of leukocytes, neutrophils, lymphocytes and monocytes. Immunophenotyping was performed on whole blood samples with multicolor (4–6 colors) flow cytometry (LSR Fortessa, BD Bioscience, NJ, USA). In brief, 100 μl blood samples were incubated with staining buffer containing a pre-titrated concentration of fluorescent monoclonal antibodies or with immunoglobulin (Ig) isotype-matched controls, respectively (all BioLegend, CA, USA). Flurochromes were Alexa fluor 700 (anti-CD3), allophycocyanin (anti-CD8, -CD45, -CD127, -CD183), brilliant Violet (BV) 421 (anti-TCRγδ), BV605 (anti-CD56), BV650 (anti-CD196), fluorescein isothiocyanate (anit-CD3, -CD14), pacific blue (anti-CD19), R-phycoerythrin (anti-CD4, -CD16, -CD194) and peridinin-chlorophyll-protein complex (anti-CD4). Compensation has been performed computationally before analysis. Erythrocytes were lysed (RBC Lysis/Fixation solution, BioLegend, CA, USA), samples were washed twice with buffer solution (phosphate-buffered saline, 0.1 % bovine serum and 0.02 % NaN₃), pelleted by centrifugation (350 × g for 5 min) and stored in buffer solution in the dark at 4 °C until analysis. Flow cytometric analysis was performed immediatly after blood sampling. T and B lymphocytes were gated by expression of CD3 or CD19 out of CD45⁺/CD16^low/CD14⁻ cells (Table 1). T lymphocyte phenotypes were gated by size exclusion, CD3⁺ and subsequent expression of CD4⁺/CD8⁻ (T helper), CD4⁻/CD8⁻ (cytotoxic T) or CD4⁻/CD8⁻/TCRγδ⁺ (TCRγδ) (Table 1). T helper lymphocytes were gated by size exclusion, CD3+/CD4+ and subsequent expression of surface chemokine receptors, which allowed for the distinction of four different memory T helper lymphocytes, CXCR3⁺/CCR4⁻/CCR6⁻ (Th1), CXCR3⁻/CCR4⁺/CCR6⁻ (Th2), CXCR3⁻/CCR4⁺/CCR6⁺ (Th17) and CXCR3⁺/CCR4⁻/CCR6⁺ (Th1/Th17), as described previously (Fig. 1 and Table 1) [1, 4]. Regulatory T lymphocytes (Treg) were identified as T helper cells expressing high levels of CD25 and low levels of CD127 (Table 1). Natural killer cells were gated by expression of CD45/CD56 and subsequent expression of CD3 (Table 1). The flow cytometry data were collected using FACS-Diva software (BD Bioscience).

Serum samples were stored at −80 °C and thawed only for analyses. Serum concentrations of IL-6 and IL-17A were measured using human IL-6 ELISA kit (Pierce Biotechnology, IL, USA) and Biosource IL-17 Cytoscreen (Life Technologies, CA, USA) following the manufacturer’s instructions. ELISA standards and samples were run in duplicate. The sensitivity of the assays was 1 pg/l for IL-6 and 2 pg/l for IL-17A. The optical density (450 nm) of each sample was determined using a microplate reader (Sunrise, Männedorf, Switzerland). Serum concentrations of IL-1-beta, IL-1ra, IL-2, IL-4, IL-10, IL-12p70, IL-13 and TNF-alpha were analyzed using a magnetic bead-based multiplex assay solution (Bio-Plex Pro TM Human Cytokine Standard 27-Plex Group I, Bio-Rad Laboratories, CA, USA) according to manufacturer’s protocol. In short, standards and samples were diluted 1:4 in buffer and added to microplates containing assay beads in duplicate. The data were collected using the Bio-Plex 200 suspension array system (Bio-Rad Laboratories).

The differences between patients and healthy controls were compared using the unpaired Student’s t test in parametric data and the Chi-square-test in nominal data. Correlations were assessed by Pearson’s coefficient. One-way repeated measures analysis of variance (rANOVA) was conducted to analyze the difference within patients or healthy subjects, respectively, over time (i.e. before, direct after and 1 h after CPET). Two-way repeated measures rANOVA was conducted to analyze the differences between patients and healthy subjects over time. P-values as result of rANOVA are presented following Wilks’ Lambda. All reported p-values were two-sided and considered significant at levels <0.05. All statistical calculations were performed using SPSS statistics version 20 software (IBM, NY, USA).

Hemodynamic parameters of 16 IPAH patients were mean pulmonary arterial pressure (mPAP) of 51 ± 19 mm Hg, pulmonary vascular resistance (PVR) of 1065 ± 809 dyne*s*cm⁵, pulmonary artery wedge pressure (PAWP) of 10 ± 3 mm Hg, right arterial pressure (RAP) of 7 ± 5 mm Hg and cardiac index (CI) of 2.18 ± 0.6 L/min*m². Mean time between right heart catheterization and study inclusion was 18 ± 10 months. At time of study, six IPAH patients were in WHO FC II and 10 in WHO FC III. Fifteen patients were treated with an endothelin-receptor antagonist (ERA), 14 patients with a phosphodiesterase type 5 inhibitor (PDE5-I), 1 patient with a soluble guanylate cyclase stimulator (sGC-S) and 3 patients with a prostacyclin analog. Only one patient was treated with monotherapy and none of the patients received long-term oxygen treatment.

Matching of healthy control subjects to patients with IPAH was achieved for age, gender and BMI as intended. All tested CPET parameters were significantly reduced in IPAH patients compared to healthy control subjects. FEV₁ and FEV₁/FVC were significantly decreased in IPAH patients yet still within normal ranges. Levels of c-reactive protein (CRP) were < 5 mg/dl in all study subjects. Table 2 summarizes the baseline characteristics of the study subjects.

There was no significant difference in absolute levels of leukocytes, lymphocytes, neutrophils and monocytes before exercise between IPAH patients and control subjects (Fig. 2). Within the CD4⁺ T helper lymphocyte compartment patients with IPAH showed elevated subsets of Th2 and regulatory T cells, while the subset of Th1/Th17 lymphocytes was reduced as compared to healthy subjects and as indicated by relative and absolute cell counts (all p < 0.05; two-way rANOVA; Figs. 4 and 5). Prior to exercise, at rest, patients with IPAH demonstrated elevated levels of IL-6 and TNF-alpha, whilst level of IL-4 was reduced (all p < 0.05; two-way ANOVA; Fig. 6).

Overall, exercise provoked a significant, transient elevation of absolute numbers of leukocytes, lymphocytes in patients with IPAH as well as in healthy subjects (all p < 0.001; one-way rANOVA; Fig. 2a-b). Given that no difference occurred between patients and healthy subjects, two-way rANOVA revealed the number of leukocytes and lymphocytes significantly depended on relation to exercise (p < 0.001 and p = 0.005), but not on IPAH (p = 0.63 and p = 0.28). The general increase of lymphocytes in patients with IPAH and healthy was also found for T lymphocytes (both p < 0.01; one-way rANOVA), and by a non-significant trend for B lymphocytes, T helper lymphocytes and cytotoxic T lymphocytes (Fig. 3a-b and e-f). Interestingly, exercise led to a relative reduction of B lymphocytes in healthy subjects but not in patients with IPAH (Fig. 3b-d). However, exercise did not influence the level of TCR-γδ lymphocytes, nor natural killer cells or natural killer T lymphocytes (data not shown).

Within the T helper lymphocyte compartment throughout exercise no significant differences occurred regarding the subsets of Th1, Th2, Th1/Th17 and regulatory T lymphocytes comparing patients with IPAH and healthy subjects (Figs. 4 and 5). Exercise, however, caused a relative reduction of Th17 lymphocytes in patients with IPAH, but not in healthy subjects (p = 0.049; one-way rANOVA; Fig. 4). However, the differences appeared to be very small with significant overlap in standard deviation and no differences were observed in regard of absolute numbers of Th17 cells (Fig. 4).

Exercise caused a sustained reduction of the IL-6 level in patients with IPAH but not in healthy subjects (p = 0.022; one-way rANOVA; Fig. 6a). Since healthy subjects showed only a marginal increase in IL-6 following exercise, two-way rANOVA revealed that the level of IL-6 significantly depended on IPAH (p = 0.009), but not on relation to exercise (p = 0.094). In addition, the immediate and 1 h lasting reduction of the IL-6 level correlated significantly with the peak VO₂/kg in patients with IPAH (p = 0.022 and p = 0.025; Fig. 7). Furthermore, exercise caused a reduction of the IL-1-beta level in patients with IPAH but not healthy subjects (p = 0.022; one-way rANOVA; Fig. 6e). Herein, two-way rANOVA revealed no significant dependence. The change of IL-6 and IL-1-beta was not significantly related to the time to exhaustion in the CPET and findings were not influenced by change in hemoconcentration (data not shown). Exercise did not influence differently the level of TNF-alpha, IL-1ra, IL-12p70, IL-13 or IL-10 in patients with IPAH as compared to healthy subjects (Fig. 6). The cytokines IL-17 and IL-2 were detectable in only 1 out of 16 patients with IPAH, respectively.