Reduced airway levels of fatty-acid binding protein 4 in COPD: relationship with airway infection and disease severity

This was a prospective, multicenter, cross-sectional study that included clinically stable COPD patients and healthy volunteers with normal lung function, who served as controls. Participants were recruited from five university tertiary hospitals [Hospital de la Santa Creu i Sant Pau (Barcelona, Spain), Hospital del Mar (Barcelona, Spain), Hospital Universitari Parc Taulí (Sabadell, Spain), Hospital Germans Trias i Pujol (Barcelona, Spain) and Hospital Clínic (Barcelona Spain)]. The study protocol was approved by the local institutional review board (IIBSP-MIC-2015-57) and all subjects gave signed informed consent.

The diagnosis of COPD was established according to the Global Initiative for Chronic Obstructive Lung Disease (GOLD) guidelines [12]. The inclusion criteria were: 8 weeks of clinical stability (defined by the absence of an exacerbation that required oral corticosteroids and/or antibiotic treatment), age between 40 and 75 years, and FEV₁ between 20 and 70% predicted. All patients underwent a computerized tomography scan, and those with bronchiectasis, lung cancer, pneumonia, and/or interstitial lung diseases were excluded. Other exclusion criteria were patients with active malignant disease and/or any type of immunosuppression, drug addiction or alcohol abuse. As controls, we included adult volunteers without respiratory diseases and with normal spirometry recruited in these same centers [13].

Demographic data, the number of exacerbations in the previous year, the time from last exacerbation, relevant comorbid conditions and current treatments were recorded at inclusion using standardized questionnaires. All patients underwent spirometry (Datospir-600; Sibelmed SA, Barcelona, Spain) following international recommendations. Reference values were those of Mediterranean populations [14].

Bronchoalveolar lavage fluid (BALF), induced sputum and plasma were obtained from all participants and were processed immediately. BALF samples were recovered using 150 ml saline lavage with the bronchoscope wedged in the right middle lobe. BALF samples were centrifuged at 800 xg for 10 min to obtain the cellular pellet and the supernatant. Induced sputum was collected just before the bronchoscopy as previously described [15]. Sputum samples were disaggregated using dithiothreitol (Oxoid Ltd., Hampshire, United Kingdom) for 15 min and were centrifuged at 600 xg for 6 min to obtain the supernatant. Proteases inhibitors (Calbiochem, San Diego, CA) were added to the supernatants during thawing. Plasma samples were obtained from blood collected in ethylenediamine tetra-acetic acid (EDTA) tube and centrifuged at 850 xg for 10 min. BALF and sputum supernatants and plasma were stored immediately at − 80 °C until analysis.

Samples were processed for qualitative and quantitative bacteriology, as previously described [16]. Airway infection was defined as the presence of potentially pathogenic bacteria (PPB) at ≥10³ colony-forming units/ml in BALF [17] in clinically stable patients.

FABP4 levels were measured by validated, commercially available ELISA kit (RayBiotech, Peachtree Corners, GA) according to the manufacturer’s instructions. The limit of kit detection was 38 pg/ml. The dilutions used were 1/7 for BALF supernatants, 1/5 for sputum supernatants and 1/75 for plasma. BALF FABP4 levels were adjusted to the total protein content quantified using Qubit fluorometer (Invitrogen, Carlsbad, CA).

In COPD patients, cellular pellet obtained from BALF samples was lysed to avoid red blood cells contamination (RBC lysing solution; BioLegend, San Diego, CA). Cells were resuspended in one ml of PBS supplemented with 2% bovine serum albumin (BSA; Roche Diagnostics GmbH, Mannheim, Germany) to quantify the number of total cells in a MACSQuant cytometer (Miltenyi Biotec, Bergisch Gladbach, Germany). Cells were adjusted to 1x10⁶cells/ml and stained for 15 min at room temperature in dark with viability dye (Zombie NIR; BioLegend), CD45-FITC, CD14-APC (Immunotools GmbH, Friesoythe, Germany) and CD15-PE (Biolegend, San Diego, CA). Aggregated and non-viable cells were excluded from the analysis. Mɸ were gated according to CD45 positive population, CD15 negative, CD14 positive and high side scatter (SSC) parameter.

Fifty-two COPD patients and 29 controls were included. Table 1 shows their demographic and clinical characteristics. COPD patients were older (65.2 ± 7.9 vs. 55.0 ± 12.3, p = 0.0002) than controls but gender (75.9 vs. 65.4% males, p = 0.6) and BMI (26.3 ± 4.7 vs. 28.7 ± 9.7, p = 0.7) were similar in both groups. Airflow limitation in patients with COPD ranged from mild to severe (FEV₁ of 59 ± 15% of predicted) whereas spirometry was normal in controls by design. Twenty-four patients (46%) were classified as GOLD grade C and D, and 21 patients (40%) were frequent exacerbators, defined as those patients who suffered from 2 or more exacerbations during the previous year to the inclusion.

Twelve COPD patients (23%) had airway infection. Patients with airway infection were predominantly males (92%) and older than non-infected ones (Table 2). Haemophilus influenzae was the most common PPB isolated (n = 9, 75%), followed by Streptococcus pneumoniae (n = 2, 17%) and Moraxella catharralis (n = 1, 8%).

Airway FABP4 levels were lower in COPD patients than in controls, reaching statistical significant differences in BALF [219.2 (96.0–319.6) vs. 273.4 (203.1–426.7) (pg/ml)/protein, p = 0.03], but not in sputum [12.2 (0.7–27.5) vs. 14.5 (0.5–45.6) ng/ml, p = 0.4] (Fig. 1A, C). Yet, we observed a positive significant correlation between BALF and sputum FABP4 levels in the entire population (rho = 0.31, p = 0.01) and in COPD patients (rho = 0.37, p = 0.01). On the other hand, plasma FABP4 levels were similar in patients and controls [26.6 (19.3–39.5) vs. 25.2 (18.3–36.5) ng/ml, p = 0.4)] and no significant correlation was observed between systemic and airway FABP4 levels.

Airway FABP4 levels were lower in COPD patients with airway infection vs. those without it, reaching statistical significant differences in sputum [0.73 (0.35–15.3) vs. 15.6 (2.0–29.4) ng/ml, p = 0.02] but not in BALF [181.5 (28.7–395.5) vs. 228.5 (106.3–319.1) (pg/ml)/protein, p = 0.7) (Fig. 1B, D). No differences in plasma FABP4 levels were observed between COPD patients with and without airway infection [29.8 (20.4–45.0) vs. 26.1 (19.3–37.9) ng/ml, p = 0.6].

Airway (but not plasma) FABP4 levels were related with several measures of disease severity, including GOLD stage, airway limitation and quality of life. Patient with GOLD D had the lowest values both in BALF (Fig. 2A) and sputum (Fig. 2B). In addition, a positive correlation among airway FABP4 levels and FEV₁ (% predicted) (Fig. 3A, B) and a negative correlation with levels of symptoms measured in CAT questionnaire (Fig. 3C, D) were observed.

Alveolar Mɸ represented 49.9 (26.4–76.8) % of cells in BALF of COPD patients, corresponding to an absolute number of 140 (48.9–311) Mɸ per μl. Both proportion and absolute number of alveolar Mɸ were significantly related with BALF FABP4 levels [proportion (rho = 0.54, p = 0.0003) and absolute number (rho = 0.52, p = 0.0006)] (Fig. 4). No significant differences among infected and non-infected patients and related to GOLD stage were found.

FABP4 is an adipokine widely studied in metabolic and cardiovascular diseases [18]. Its role in chronic respiratory diseases is poorly understood, although it has been recognized that FABP4 is associated with airway inflammation [19]. Likewise, a dysregulation of airway FABP4 has been previously described in asthma [20] where it has been shown in experimental models that FABP4 participates in the recruitment and activation of eosinophils [21]. In patients with COPD, a dysregulation of systemic FABP4 compared with controls has been suggested [22], which is at variance with our findings here. Differences may be related to distinct inclusion criteria. Whereas Zhang et al excluded patients with metabolic and vascular comorbidities [22], known to be associated with elevated systemic FABP4 levels [23, 24], we did not because cardiovascular comorbidities are highly prevalent in patients with COPD (74% in our cohort) and we did not want to bias our study population. In any case, to our knowledge, our study is the first to investigate airway FABP4 levels, and their relationship with airway infection and disease severity in COPD.

Bacterial airway infections are relevant in the natural history of COPD [3] because it worsens clinical outcomes, including mortality and costs [4]. The molecular and cellular mechanisms that favor bacterial infection in some COPD patients are, however, not yet fully elucidated. We observed that airway FABP4 levels were reduced in COPD patients, particularly in those with evidence of airway infection, and we found a relationship between BALF Mɸ (absolute number and proportion) and FABP4 levels in COPD. Given that alveolar Mɸ produce FABP4 [9], these observations support a pathogenic role for a locally defective Mɸ production of FABP4. In support of this interpretation is the fact that FABP4 facilitates the interaction between Mɸ and neutrophils through the regulation of CXCL1, a chemokine secreted by Mɸ to recruit neutrophils to the site of infection [11, 25]. Further works studying Mɸ subpopulations would help to better understand the immunological mechanism related to FABP4 production by Mɸ.

On the other hand, we found that patients with severe COPD showed the lowest airway FABP4 levels. We also found that there was a direct relationship between airway FABP4 concentration and FEV₁ levels and an inverse relationship between the former and CAT score (i.e. worse health status). In keeping with these observations, some previous studies had demonstrated that patients with severe COPD have altered airway innate immunity [16, 26, 27] that leads to chronic airway inflammation and dysregulation of normal alveolar Mɸ function [28]. Yet, because our study is cross-sectional, we cannot infer what is the cause and what is the consequence, this is, what comes first.

Our study has some limitations that deserve comment. First, controls were younger than COPD patients, but we did not observe any relationship between airway FABP4 levels and the age of the participants. Second, induced sputum was collected just before the bronchoscopy so we cannot discard any interaction with BALF samples, although this procedure had been applied to all the participants. Third, we did not perform flow cytometry in BALF from controls, but it is well reported in the literature that COPD alveolar Mɸ differ from healthy controls in phenotype [29] and in the ability to phagocytose bacteria and apoptotic cells [30, 31]. Forth, we have not determined FABP4 in infected patients without COPD and it would be of great interest to better understand its role in the pathogenesis or airway infection. Finally, we did not obtain follow-up samples, so we cannot infer if any therapeutic intervention (e.g., treatment with low dose azithromycin) can restore airway FABP4 levels and what clinical consequences that may have.