Protein expression profile changes of lung tissue in patients with pulmonary hypertension

Clinical specimens and patient enrollment

Our research plan and an informed patient consent form for the present study were submitted to the Medical Research Ethics Committee of Guangdong Provincial People’s Hospital. The Medical Ethics Committee reviewed and approved our submitted proposal, ethical approval number Guangdong Medical Ethics 2016220H (R1)/GDREC2016220H (R1). All patients signed the written informed consent form prior to being included in the study.

Patients were enrolled in strict accordance with the following inclusion and exclusion criteria, as specified in the experimental protocol. For the experimental group, the included patients had been diagnosed as having pulmonary hypertension with an atrial septal defect that could not be corrected by internal closure. In the resting state, with the patient lying so that the left atrium was in a horizontal position and the measurement determined at a distance midway between the middle of the sternum and the bedside, the average pressure of the pulmonary artery measured by a catheter located in the right ventricle of the heart was ≥25 mmHg. For the control group, the included patients did not have pulmonary hypertension and had received a diagnosis of central lung cancer requiring lobectomy, but the cancer tissue had not invaded the peripheral lung tissue used in this study.

iTRAQ labeling and fractionation by cation exchange chromatography

The 8-plex iTRAQ labeling method was performed as previously described (Ding et al., 2019; Liu et al., 2019). The reagents were used to label protein peptides from each group multiplex kit (ABI, Foster City, CA, USA) (isobaric tags 113, 114, and 116 for the control group, and isobaric tags 118, 119, and 121 for the pulmonary hypertension group). For each specimen, protein (200 µg) was precipitated with acetone at −20 °C overnight. After centrifugation for 10 min, the protein pellet was dissolved in 60 µL of iTRAQ dissolution buffer (Applied Biosystems). The iTRAQ labeling reagents were added to the corresponding peptide specimen to react at room temperature for 1 h. After the reaction for the labeling was stopped and an extraction efficiency test was performed, the specimen was ZipTip desalted and subjected to a matrix-assisted laser desorption ionization procedure. Each group was pooled and then vacuum-dried. Each pooled specimen containing mixed peptides was lyophilized and dissolved in solution A (2% acetonitrile and 20 mM ammonium formate, pH 10). This solution was loaded on a reverse-phase column (Luna C18, 4.6 × 150 mm; Phenomenex; Torrance, CA, USA) and eluted using a step linear elution program at a flow rate of 0.8 mL/min: 0%–10% buffer B (500 mM KCl, 10 mM KH₂PO₄ in 25% acetonitrile, pH 2.7) for 10 min, 10%–20% buffer B for 25 min, 20%–45% buffer B for 5 min, and 50%–100% buffer B for 5 min. The eluent fractions were collected and centrifuged for 5–45 min. The fractions were finally combined into six pools and desalted using C18 Cartridges (Empore standard density Solid Phase Extraction C18 Cartridges, bed I.D. seven mm, three mL volume; Sigma; St. Louis, MO, USA).

LC–MS/MS analysis

We performed MS using a TripleTOF 5600 system (AB SCIEX) combined with a nanoliter spray III ion source (AB SCIEX, USA). The spray voltage was set at 2.5 kV; the air curtain pressure, at 30 psi; the atomization pressure at 5 psi; and the heater temperature, at 150 °C. The scanning mode used was information-dependent acquisition. The scan time of the first time-of-flight (TOF)–MS single image was 250 ms. Each information-dependent acquisition cycle collected 35 secondary TOF–MS images. Each cycle time was fixed at 2.5 s. The dynamic exclusion was set to 18 s, which was approximately equal to half the width of the chromatogram.

Sequence database search and data analysis

The data were processed according to the flowchart shown in Fig. 1. After the data were collected, we used Protein Pilot Software v. 5.0 (AB SCIEX, USA) for the analysis. Human proteome databases containing reviewed Uniprot sequences were used to perform peptide identification. Decoys for the database search were generated with the revert function. We chose the following options to identify the proteins: cysteine alkylation = iodoacetamide; enzyme = trypsin; search effort = thorough. The proteins that showed a relative upregulated or downregulated fold change of at least 1.2 (in both replicates) and that were significantly different between the groups at P <0.05 were determined to be differentially expressed between the groups.

Gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses

We performed GO analyses using Metascape, a web-based resource for gene and protein annotation, visualization, and integration discovery (http://metascape.org) (Fang et al., 2019; Soonthornvacharin et al., 2017). KEGG pathway analyses of the DEPs were performed using the KOBAS online analysis database (http://kobas.cbi.pku.edu.cn/) (Kanehisa & Goto, 2000). A two-sided P <0.05 was considered statistically significant.

Protein–protein interaction (PPI) network construction and module analyses

PPI analysis was used to assess the functions associated with the DEPs and to determine the general organizational principles of the functional cellular networks. The functional relationships between proteins were identified using the Search Tool for the Retrieval of Interacting Genes (STRING; http://string.embl.de/) (Fang et al., 2019; Von Mering et al., 2003). The PPI networks associated with the respective DEPs were constructed to predict the interaction of those proteins.

Specimen collection and patient characteristics

From April to September 2016, we collected lung tissue specimens from six patients with atrial septal defect and pulmonary hypertension. The patients were all women, with a mean (SD) age of 46.5 ± 4.7 years, and their condition could not be corrected with an internal medical occlusion device. For the control group, nine patients with lung cancer who underwent lobectomy were enrolled: six men and three women, with a mean (SD) age of 56.7 ± 1.7 years. Among the patients with pulmonary hypertension, the mean pulmonary systolic pressure was 66.7 ± 5.5 mmHg; mean pulmonary artery diastolic pressure, 23.2 ± 1.6 mmHg; mean pulmonary artery pressure, 39.7 ± 3.4 mmHg; mean pulmonary artery resistance, 3.27 ± 0.28 Wood units; pulmonary oxygen saturation, 87.3% ± 0.84%; and mean patient body surface, 1.52 ± 0.06 m².

Identification of DEPs in pulmonary tissue

After performing the quantitative analyses with Protein Pilot Software, we identified 2,953 proteins using a detection protein threshold (Unused ProtScore [Conf]) cutoff of >1.3; 95% confidential interval) (Table 1), and we obtained the protein expression fold changes between the control specimens and the pulmonary hypertensive lung tissue specimens. The data indicated marked protein expression profile changes between the control lung tissue and the pulmonary hypertensive lung tissue. Using our selection criteria, we identified 74 significantly upregulated DEPs, including integrin subunit α1 (ITGA1) and the voltage-gated calcium (Ca²⁺) channel α2/ δ1 subunit, and 88 significantly downregulated DEPs, including Rho-GDP dissociation inhibitor 2 (Table 2, Fig. 2).

GO functional enrichment analysis

GO analysis is used to hierarchically classify genes or gene products into categories organized in an ontology. The analysis is based on three categories: (1) molecular function, to describe the molecular activity of a gene; (2) biological process, to characterize the larger cellular or physiological role of the gene; and (3) cellular component, to indicate the location of the gene or gene product in the cell. We used Metascape to enrich all identified proteins. Among the top 20 terms in the biological process category, the highest percentages of the proteins were associated with the term single-organism process (Figs. 3 and 4) (n = 119 proteins; with the top 3 upregulated proteins in this process being LDLRAP1, FTL, and CHUK). The maximum levels of the expressed proteins this category were detected for the term protein complex subunit organization (n = 27 proteins; with the top 3 upregulated proteins being SLC6A4, FRYL, and HIST2H3A). Among the top 20 terms in the GO category cellular component, the highest percentages of proteins were associated with the term cytoplasm (n = 104 proteins, with the top 3 upregulated proteins for this term being LDLRAP1, FTL and CHUK. The maximum levels of the expressed proteins in this category were in the term major histocompatibility complex (MHC), which is a group of proteins located on the cell surface (n = 9 proteins, with the top 2 upregulated proteins being HLA-A and HLA-DPA1). In the top 20 terms of the Go category molecular function, the highest percentages of proteins were associated with the term binding (113 proteins, with the top 3 upregulated proteins being LDLRAP1, FTL and CHUK), and the maximum levels of the expressed proteins in this category were for the term ferric iron binding (with the top 2 upregulated proteins being FTL and FTH1). The GO functional enrichment analysis assessing the distributions of the proteins with the greatest fold changes in the biological process, molecular function, and cellular component categories indicated that the proteins associated with the terms organism process, protein binding, and extracellular vesicle, respectively, were markedly changed.

KEGG signaling pathway analysis and PPI network construction

To identify the functions associated with the DEPs, we used the KOBAS online analysis tool to determine the KEGG signaling pathways, and we used the STRING online analysis tool to construct the PPI networks. Through KEGG signaling pathway analysis, the gene content in the provided genome is compared with the information in the KEGG pathway database to clarify which pathways and associated functions link to the genes in the genome. The PPI network indicates physical contacts between two or more proteins resulting from biochemical events controlled by electrostatic forces, including hydrophobic effects. The top 10 KEGG signaling pathways were enriched, and the PPI networks were constructed (Figs. 5–9). We found that the top 10 signaling pathways included hypertrophic cardiomyopathy, systemic lupus erythematosus, arrhythmogenic right ventricular cardiomyopathy, dilated cardiomyopathy, pathogenic Escherichia coli infection, viral myocarditis, phagosome, alcoholism, cardiac muscle contraction, and ribosomes, which are predominantly associated with human diseases (Fig. 5). Some of the DEPs crossed several different signaling pathways. For example, the proteins highlighted in red boxes in Figs. 6, 7 and 8 including DHPR, ACTB, desmin, ACTG1, PRKAA1 and ITGA1, were all involved in hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, and dilated cardiomyopathy (Table 2). The results of the KEGG pathway enrichment and PPI network construction analyses indicated that many of the upregulated proteins (shown as brown dots in Fig. 9) were involved in hypertrophic cardiomyopathy, By contrast, many of the downregulated proteins (shown as green dots in Fig. 9) were related to ribosome function. Taken together, our results suggest that the heart is a major organ impaired by pulmonary hypertension.