Proteomic profiling of lung immune cells reveals dysregulation of phagocytotic pathways in female-dominated molecular COPD phenotype

Abstract

Background: Smoking is the main risk factor for chronic obstructive pulmonary disease (COPD). Women with COPD who smoke experienced a higher risk of hospitalization and worse decline of lung function. Yet the mechanisms of these gender-related differences in clinical presentations in COPD remain unknown. The aim of our study is to identify proteins and molecular pathways associated with COPD pathogenesis, with emphasis on elucidating molecular gender difference.

Method: We employed shotgun isobaric tags for relative and absolute quantitation (iTRAQ) proteome analyses of bronchoalveolar lavage (BAL) cells from smokers with normal lung function (n = 25) and early stage COPD patients (n = 18). Multivariate modeling, pathway enrichment analysis, and correlation with clinical characteristics were performed to identify specific proteins and pathways of interest.

Results: More pronounced alterations both at the protein- and pathway- levels were observed in female COPD patients, involving dysregulation of the FcγR-mediated phagocytosis-lysosomal axis and increase in oxidative stress. Alterations in pathways of the phagocytosis-lysosomal axis associated with a female-dominated COPD phenotype correlated well with specific clinical features: FcγR-mediated phagocytosis correlated with FEV1/FVC, the lysosomal pathway correlated with CT < -950 Hounsfield Units (HU), and regulation of actin cytoskeleton correlated with FEV1 and FEV1/FVC in female COPD patients. Alterations observed in the corresponding male cohort were minor.

Conclusion: The identified molecular pathways suggest dysregulation of several phagocytosis-related pathways in BAL cells in female COPD patients, with correlation to both the level of obstruction (FEV1/FVC) and disease severity (FEV1) as well as emphysema (CT < -950 HU) in women.

Trial registration: No.: NCT02627872 , retrospectively registered on December 9, 2015.

Keywords: Bronchoalveolar lavage; Chronic obstructive pulmonary disease; Gender difference; Isobaric tags for relative and absolute quantitation; Orthogonal projection to latent structure-discriminant analysis; Proteomics; Smoking.

Conflict of interest statement

Ethics approval and consent to participate

The study was approved by the Stockholm regional ethical board (Case no. 2006/959–31/1), and written informed consent was obtained from all subjects.

Consent for publication

Not applicable

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Figures

Fig. 1

Flow chart outlining the study design and emphasis groups of the current vs. the companion paper [16]. A total of 69 subjects from the Karolinska COSMIC cohort, well-matched in terms of age, gender, and lung functions, were selected for iTRAQ proteomic investigations, including 18 healthy Never-smokers (9 male, 9 female), 25 Smokers with normal lung function (11 male, 14 female), 18 current smokers with COPD (10 male, 8 female), and 8 ex-smokers with COPD (3 male, 5 female). This manuscript focus on the alterations in proteomes and pathways related to COPD pathology (i.e., comparisons of Smoker vs. smokers with COPD, to some extent related to comparisons of healthy Never-smokers vs. ex-smokers with COPD (exCOPD)). The companion paper focuses on the effects of smoking prior to disease presentation, i.e., comparison of the Never-smoker vs Smoker groups [16]

Fig. 2

a) The protein (Isochorismatase domain-containing protein 2, ISOC2, Uniprot: Q96AB3) exemplifies the profile of a protein that was identified as significantly altered in the joint gender analysis (p = 0.01), in spite of a high heterogeneity (I2 = 0.88). Stratification by gender revealed that this significance was completely driven by that of in female population (p = 0.0002), and was not altered in males (p = 0.66) b) Venn diagram showing the overlap in alterations of proteomic profiles between Smokers and COPD groups in joint gender and gender stratified univariate statistical analyses. Only three proteins were altered in both male and female COPD patients, and the majority of protein alterations in female patients were unique

Fig. 3

Multivariate OPLS-DA models comparing Smokers vs COPD groups before and after stratification by gender. OPLS-DA modeling showed significant separations between Smoker and COPD groups for a) joint gender (R2 = 0.85, Q2 = 0.66, p[CV-ANOVA] = 4.6 × 10−8, 116 proteins), b) females (R2 = 0.85, Q2 = 0.81; p = 1.9 × 10−7, 145 proteins) and d) males (R2 = 0.78, Q2 = 0.73, p = 9.4 × 10−6, 24 proteins). However, the predictive performances (Q2) were better following stratification by gender for both gender models. c) Loadings of the top 24 proteins out of 145 significant variables in the female COPD vs Smoker model; e) All 24 significant proteins from the male COPD vs Smokers model. There was no overlap among 24 top proteins of both gender models. A comprehensive list of loadings along with protein names and statistics are provided in Additional file 2: Table S2. f) Venn diagram displaying overlap between genders in protein alterations due to COPD based on the OPLS-DA models displayed in b) and d). Only four proteins (Q9NSE4, P02751, O95470, and P01876) were altered in both male and female smokers with COPD

Fig. 4

Dysregulation of FcγR-mediated phagocytosis in female COPD patients. The protein levels of ARPC4, ARPC5, ARPC5L, ARPC1B, ARPC2, ARPC3 decreased (blue) and Rac as well as RHOA increased (red) in BAL cells. Rearrangement of the actin cytoskeleton is a necessary driving force for FcγR-mediated phagocytosis [41, 42]. The decreased levels of Arp2/3 and actin cytoskeletal processes may thus imply that FcγR-mediated phagocytosis was hampered in spite of up-regulations of Rac and RhoA in COPD patients. The majority of proteins in the downstream regulation of actin cytoskeleton- and lysosome pathways were down-regulated in female COPD patients (Additional file 2: Table S3) Blue: down-regulated; red: up-regulated. This figure was created with KEGG pathway tool with minor modification

Fig. 5

Correlations of the pathways of interest with clinical parameters and T-cell subpopulations in BAL in female COPD patients. a) Correlation between the proteins in the lysosomal pathway and CT attenuation values <−950 HU in female COPD group (PLS inner relation, R2 = 0.81, p = 0.002). b) The correlation between the proteins in the regulation of actin cytoskeleton pathway and FEV1/FVC (PLS inner relation, R2 = 0.83, p = 0.002) in female COPD patients. c) The proteins in the FcγR-mediated phagocytosis pathway highly correlated with chemokine receptor CXCR4 on CD4+ T cells (R2 = 0.97, p < 0.0001). d) The proteins in the regulation of actin cytoskeleton pathway correlated with CXCR4-expressing on CD8+ T cells (R2 = 0.86, p = 0.001)

Fig. 6

Comparison of results from 2D–DIGE intact proteomics [15] and iTRAQ shotgun proteomics analysis on the same cohort. a) Venn diagram showing that the majority of protein species identified by 2D–DIGE overlapped with proteins identified by iTRAQ. Ninety percent of the proteins identified by iTRAQ were not identified by 2D–DIGE. However, several of the proteins identified by 2D–DIGE served as validation of key proteins in pathways found to be enriched in the current study: Actin-related protein 3 (ARP3; panel b;, 2D–DIGE: p = 0.008; iTRAQ: p = 0.01, is involved in the pathways FcγR-mediated phagocytosis and regulation of actin cytoskeleton; Beta-hexosaminidase subunit beta (HEXB; panel c;)2D–DIGE: p = 0.01; iTRAQ: p = 0.02,); and ATP synthase subunit beta (ATP5B, panel d;) (2D–DIGE: p < 0.0001; p = 0.04,) are both involved in the lysosomal- and oxidative phosphorylation pathways,; e), Leukotriene A-4 hydrolase (LTA4H) (2D–DIGE: p < 0.0001; iTRAQ: p = 0.01,). is one of the most prominent proteins for driving the separation between female COPD and Smokers groups by both proteomics platforms. Data in b, c, d and e are expressed as mean ± SE