Over 500 million individuals reside at altitudes of 1,500 m or higher above sea level, where the unique environment significantly influences human physiology, health, and disease. High-altitude pulmonary hypertension (HAPH) is a progressive condition characterized by pulmonary vascular remodeling and hypertension, primarily caused by chronic hypoxia and other factors in those living above 2,500 m. Classified as the third type of hypoxia- or lung disease-related pulmonary hypertension1, HAPH affects 6–10% of high-altitude populations2. This condition severely impacts their quality of life and can lead to right heart failure, ultimately resulting in premature death3. Inflammation, oxidative stress, endothelial dysfunction, and pulmonary vascular remodeling are all implicated in HAPH4, although its exact pathogenesis remains unclear.

Advancements in multi-omics technology have significantly improved our comprehension of the roles played by genes, proteins, and metabolites in the development of pulmonary arterial hypertension (PAH). Mutations in key genes linked to PAH, such as BMPR2, SMAD1, SMAD4, and SMAD9, have been identified5. Additionally, proteomic analyses have connected proteins like ApoE, prekallikrein, IGFBP-1, and TIMP-2 to the prognosis of primary pulmonary hypertension6. These findings have paved the way for the creation of targeted medications for pulmonary hypertension, such as prostacyclin analogs, endothelin receptor antagonists, and phosphodiesterase-5 inhibitors, which have demonstrated potential in enhancing exercise capacity and pulmonary hemodynamics in patients with pulmonary hypertension1. Nevertheless, the efficacy of these medications in high-altitude populations affected by pulmonary hypertension, as well as the specific proteins and pathways involved in its development, remains uncertain, impeding the progress of personalized targeted therapies. Hence, the application of multi-omics strategies is essential to pinpoint crucial proteins and pathways contributing to high-altitude pulmonary hypertension (HAPH).

This study specifically targeted the Tibetan population living at altitudes above 4,500 m. Through gathering extensive data on demographic features, physiological and biochemical parameters, as well as clinical symptoms, we employed integrated metabolomic and proteomic analyses to pinpoint crucial pathways and potential proteins associated with the development of high-altitude pulmonary hypertension (HAPH). The results of our research aim to offer important insights into the global understanding of HAPH pathogenesis.

Materials and methods

Study design and population

The study population was recruited by the Very High Altitude Program (ChiCTR2100047945) and approved by the Ethics Committee of People’s hospital of Xizang Autonomous Region (Approval No. ME-TBHP-21-028). All methods were carried out following the relevant guidelines and regulations. Written informed consent was obtained from all participants prior to their inclusion in the study.

Between July and August 2021, serum samples were collected and cryopreserved from pulmonary hypertension patients. Of the 2,912 participants aged > 18 years in the Very High Altitude Program, 102 (3.50%) were diagnosed with pulmonary hypertension, and 75 patients were eligible for enrollment. After applying eligibility criteria, 60 individuals were enrolled: 30 in the HAPH group and 30 healthy controls matched for demographic information (sex, age, BMI) and clinical characteristics (blood pressure, oxygen saturation, echocardiographic indices). To further validate the proteomic analysis results, an additional cohort of 30 HAPH patients and 30 healthy controls were recruited.

HAPH was defined based on the 2004 Criteria for Diagnosis of High Plateau Pulmonary Hypertension7, with a transthoracic pulmonary artery systolic pressure > 50 mmHg (pulmonary artery systolic pressure was estimated using tricuspid regurgitation spectral peak velocity detected by echocardiography, and pulmonary artery pressure was estimated using Doppler flow imaging and spectral Doppler techniques). Two ultrasound experts reviewed all echocardiographic images. Patients with chronic diseases such as congenital heart disease, left heart disease, heart valve disease, abnormal pulmonary function, hepatic or renal insufficiency, hematologic disorders, and current medication use were excluded.

Sample collection and Preparation

All participants provided 2 mL of venous blood after overnight fasting. The blood was immediately centrifuged at 3,500 rpm for 15 min, aliquoted into standard serum collection tubes, and stored at − 80 °C until analysis.

Non-targeted metabolomics analysis

Sample Preparation and metabolite extraction

Metabolomic analysis was conducted on 60 serum samples divided into HAPH (n = 30) and control (n = 30) groups. For metabolite extraction, 100 µL of each serum sample was transferred to an EP tube with 400 µL of extract solution (acetonitrile: methanol = 1:1, containing isotopically-labeled internal standard mixture). Samples were vortexed for 30 s, sonicated for 10 min in an ice-water bath, and incubated for 1 h at − 40 °C to precipitate proteins. After centrifugation at 12,000 rpm (RCF = 13800(×g), R = 8.6 cm) for 15 min at 4 °C, the resulting supernatant was transferred to fresh glass vials for analysis. Quality control (QC) samples were prepared by mixing equal aliquots of the supernatants from all samples.

UHPLC-orbitrapMS analysis

Non-targeted metabolomic analyses were performed using an Ultra-High Performance Liquid Chromatography (UHPLC System Software (Thermo Fisher Scientific)) with a Waters ACQUITY UPLC BEH Amide column (2.1 mm × 100 mm, 1.7 μm) coupled to an Orbitrap Exploris 120 mass spectrometer (Thermo Fisher Scientific). The mobile phase consisted of 25 mmol/L ammonium acetate and 25 mmol/L ammonium hydroxide in water (pH = 9.75) (A) and acetonitrile (B). The detailed HPLC gradient parameters were as follows(see supplementary table).

The auto-sampler temperature was maintained at 4 °C, and the injection volume was 2 µL

The Orbitrap Exploris 120 mass spectrometer was operated using Xcalibur Software (version 4.4; Thermo Fisher Scientific, Waltham, MA, USA) to acquire MS/MS spectra in information-dependent acquisition (IDA) mode. The ESI source conditions were: sheath gas flow rate 50 Arb, auxiliary gas flow rate 15 Arb, capillary temperature 320 °C, full MS resolution 60,000, MS/MS resolution 15,000, collision energy 10/30/60 in NCE mode, and spray voltage 3.8 kV (positive) or −3.4 kV (negative).

Data processing and analysis

Raw data were converted to mzXML format using ProteoWizard Software (http://proteowizard.sourceforge.net/) and processed with an in-house program developed using R and based on XCMS Online (https://xcmsonline.scripps.edu/) for peak detection, extraction, alignment, and integration8,9. The XCMS method used was centWave with parameters: ppm 10, peak width 5–20, and signal-to-noise ratio 3. Pre-filtering was performed to retain only peaks with at least 3 intensities ≥ 1000. The function used to calculate the center of the chromatographic peak m/z was wMean, representing the intensity-weighted average of the peak’s m/z values. The minimum m/z dimension difference required for peaks with overlapping retention times was−0.001.

An in-house MS2 database Biotree Database (version 2.1; Biotree, Shanghai, China) was applied for metabolite annotation with a cutoff value of 0.3 for algorithmic scoring. The qualitative scoring was based on Euclidean distance and dot product algorithms, which improved the accuracy of MS annotation9.

To identify differentially expressed metabolites (DEMs), we employed a dual statistical approach. First, univariate analysis using Student’s t-test identified metabolites with significant differences between groups (P < 0.05), with Benjamini-Hochberg false discovery rate correction for multiple testing (P.adjust < 0.2).

Second, multivariate analysis using Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) evaluated the variable importance in projection (VIP), with VIP >1.5 considered significant. Metabolites meeting both criteria (P< 0.05, P.adjust < 0.2, and VIP > 1.0) were classified as differentially expressed. This dual approach was chosen to combine the strengths of both univariate testing (identification of independently changing significant metabolites) and multivariate analysis (selection of metabolites with high contribution to overall intergroup differences), allowing observation from different statistical perspectives while minimizing false positive errors and model overfitting.

Principal component analysis (PCA) was performed using R Statistical Computing Software (version 3.6.3; R Foundation for Statistical Computing, Vienna, Austria) or SIMCA Multivariate Data Analysis Software (version 16.0.2; Sartorius, Göttingen, Germany), with data logarithmically transformed and centrally processed before modeling.

Proteomics analysis

Sample Preparation

For proteomic analysis, 15 HAPH cases and 15 control cases were analyzed using data-independent acquisition (DIA). Samples were reconstituted and centrifuged at 12,000 rpm at 4 °C for 10 min. From each sample, 30 µL was withdrawn into new EP tubes as assay samples. From the supernatant, 10 µL samples were taken into new EP tubes, mixed with 190 µL of 50 mM NH₄HCO₃, vortexed, and protein concentration was measured after centrifugation.

For library building, 5 µL supernatants from each sample were pooled. After vortexing, 100 µL was processed using the Thermo De-High Abundance Protein Kit. The processed samples were transferred to 10KD ultrafiltration tubes and centrifuged at 12,000 rpm at 4 °C for 20 min. The ultrafiltration tubes were washed twice with 200 µL of 50 mM NH₄HCO₃, and the protein concentration was measured using the BCA method.

Protein processing

Protein concentrations were determined using the BCA assay. For reduction and alkylation, 100 µL protein complex solution was sonicated in a water bath for 3 min, followed by addition of 1 µL DTT (5 mM final concentration) and incubation at 55 °C for 20 min with vibration. After cooling to room temperature, 3 µL IAA (15 mM final concentration) was added and samples were incubated for 30 min in the dark.

For protein digestion, trypsin was dissolved to 0.5 µg/µL in resuspension buffer and incubated for 5 min at room temperature. Trypsin was added at a 1:50 ratio (trypsin: protein), mixed well, and incubated overnight at 37 °C with 1000 rpm shaking.

Peptide desalting and fractionation

For peptide desalting, 1 µL of 10% TFA was added to acidify peptides (final concentration 0.1%). After centrifugation at 12,000 rpm for 10 min, 50 µg of supernatant peptide was diluted to 1 mL with Buffer A for desalting using C18 columns. The desalting columns were activated with Buffer C, equilibrated with Buffer A, loaded with sample, washed with Buffer A, and peptides were eluted with 400 µL Buffer B. The eluate was dried by vacuum centrifugation.

For high pH pre-fractionation of library-building samples, lyophilized peptide samples were reconstituted to 50 µL with mobile phase A and separated using RPUPLC under alkaline conditions. Fractions were collected every 1 min for a total of 12 components, vacuum dried, and re-solubilized with a solution containing iRT reagent.

LC-MS/MS analysis

LC-MS/MS analyses were performed using an UltiMate 3000 system (Thermo Fisher Scientific) coupled to a timsTOF Pro2 mass spectrometer (Bruker Daltonics). For library samples, 200 ng of sample was injected onto an analytical column (25 cm × 75 μm i.d., IonOpticks) maintained at 50 °C. The column flow rate was controlled at 300 nL/min with a gradient starting at 4% B phase, increasing to 28% in 45 min, 44% in 10 min, and 90% in 10 min, maintained at 90% for 7 min, then equilibrated at 4% for 8 min.

The mass spectrometer was operated in DDA PASEF mode for library building, with 10 PASEF scans per cycle and 100 ms accumulation time. The scanning range was 100–1700 m/z, with ion mobility 1/K0 set to 0.6–1.6. The target value was 10,000 with a dynamic exclusion time of 0.4 min. The isolation window width was set to 2 Th for m/z < 700 and 3 Th for m/z > 700.

For test samples, the mass spectrometer was operated in diaPASEF mode for DIA data acquisition, with a scanning range of 349–1229 m/z and isolation window width of 40 Da. During PASEF MS/MS scanning, the collision energy increased linearly with ion mobility from 59 eV (1/K0 = 1.6 Vs/cm²) to 20 eV (1/K0 = 0.6 Vs/cm²).

Data analysis

DDA data were processed using Spectronaut Software (version 16; Biognosys AG, Schlieren, Switzerland).

Databases with default parameters. The sequence database was UniProt Homo sapiens (version 2022, https://www.uniprot.org/), with trypsin digest set to a maximum of 2 missed cleavages. Fixed modifications included carbamidomethylation (C) + 57.02 Da, and variable modifications included oxidation (M) + 15.99 Da. The false discovery rate (FDR) was set to 1% at both peptide and protein levels.

DIA data were analyzed using Spectronaut 16 with default parameters, using iRT peptides for automatic correction of retention times and determination of extraction windows. Decoy databases were generated using a mutated strategy, with a minimum of 2 and maximum of half the total length of amino acids scrambled. Data were normalized using a local normalization strategy, and protein quantification was based on the average peak areas of the top 3 peptides with < 1.0% FDR.

Differentially expressed proteins (DEPs) were identified using Student’s t-test with criteria of P < 0.05, P.adjust < 0.2(Benjamini-Hochberg correction), and fold chang ≤ 0.83 or ≥ 1.2.

Proteomics parallel reaction monitoring (PRM) validation

To validate candidate proteins, a validation set of 30 HAPH cases and 30 control cases was established. Twenty proteins were initially selected for PRM validation based on their differential expression, biological relevance to HAPH pathophysiology, and technical feasibility for detection. These included MYH10, EGFR, HNRNPA3, GXYLT2, ACTG1, IGHV3-35, RPS7, HNRNPM, VNN1, YWHAZ, IGLV1-40, PCBP2, GGH, IL1RAP, H6PD, H2AC8, VIM, APOF, CRISP3, and CST3.

PRM was performed on a Q-Exactive HFX high-resolution mass spectrometer (Thermo Fisher Scientific) coupled with a nanoUPLC system (EASYnLC 1200 System Software (Thermo Fisher Scientific)) for separation. Data acquisition was conducted in positive ion detection mode using a nano-liter ion source. Following validation experiments, five proteins (ACTG1, VNN1, CKB, APOF, and CST3) showed consistent differential expression patterns in both discovery and validation cohorts and were selected for further functional analysis.

Integrated analysis

Kyoto Encyclopedia of Genes and Genomes (KEGG; https://www.kegg.jp/) was used for cohomology analysis and identification of potential pathways involving DEPs and DEMs. The five validated proteins were investigated for potential functional and biological associations with differential metabolites to understand their interactions in the context of HAPH pathophysiology. This integrated analysis aimed to identify pathway modifications involving both proteins and metabolites that contribute to HAPH development.

Hierarchical clustering was employed to assess correlations between DEPs and DEMs using Euclidean distance as the distance criterion. Pathway enrichment analysis was performed to identify significantly altered metabolic and signaling pathways in HAPH.

Statistical analysis

Data are expressed as means for normally distributed continuous variables and medians (interquartile range) for non-parametrically distributed variables. Student’s t-test and Mann-Whitney U-test were used to compare continuous variables, while chi-square or Fisher’s exact test was used for categorical variables. For omics data analysis, multiple testing correction was performed using the Benjamini-Hochberg procedure with the p.adjust function in R to control the false discovery rate. Both raw P-values and adjusted Q-values are reported. The OPLS-DA model evaluated the discriminative ability of DEPs and DEMs, with R² indicating model performance and Q² indicating repeatability. Multivariate logistic regression models were used to determine correlations between patient characteristics and HAPH. Variables with P < 0.05 on univariate analysis were included in multivariate logistic regression models to screen for DEPs and DEMs.

Results

Clinical baseline characteristics

The HAPH and control groups were matched for sex, age, and weight, with no significant differences in baseline hemoglobin, blood pressure, or oxygen saturation. However, the right ventricular internal diameter was significantly larger in the HAPH group than in the control group, consistent with the pathology of pulmonary hypertension (Table 1).

Table 1 Baseline characteristics and echocardiographic parameters of the study population.
Full size table

Metabolomic profiling reveals distinct metabolic signatures in HAPH

OPLS-DA analysis revealed distinct clustering of metabolomic data between HAPH patients and healthy controls. The OPLS-DA scatter plots demonstrated clear separation between the two groups, suggesting significant differences in their metabolic profiles. Multi-group differential expression analysis identified a total of 26 metabolites that were differentially expressed in HAPH, with 12 upregulated and 14 downregulated compared to controls. Hierarchical clustering analysis further illustrated the distinct metabolic patterns between HAPH patients and controls. Our metabolomic analysis identified significant alterations in multiple metabolite classes between HAPH patients and healthy controls. The most significantly upregulated metabolites included: Pantothenol (FC = 2.253, P.adjust = 0.182), 2,2-Methylpropanoic acid (FC = 1.784, P.adjust = 0.199), Plastoquinone 3 (FC = 2.699, P.adjust = 0.182), 2’-O-Methylisoguanosine (FC = 1.601, P.adjust = 0.012). Among the downregulated metabolites, we observed: L-Hexahydro-3-hydroxypyridine (FC = 0.387, P.adjust = 0.091), Ascochinline (FC = 0.365, P.adjust = 0.182), Octadecylamine (FC = 0.822, P.adjust = 0.047), Blenin B (FC = 0.601, P.adjust = 0.182). Lipid Metabolism Alterations: Several glycerophospholipids showed significant changes, including: PC(20:4/8Z,11Z) (FC = 2.804, P.adjust = 0.047), PC(P-18:1/9Z) (FC = 2.288, P.adjust = 0.047), PC(22:1/13Z) (FC = 2.458, P.adjust = 0.146). Amino Acid and Peptide Metabolism: We identified significant changes in amino acid-related metabolites: Aspartyl-alanine (FC = 0.864, P.adjust = 0.182), N-Feruloylaspartic acid (FC = 0.362, P.adjust = 0.091). These findings suggest that HAPH is characterized by widespread metabolic reprogramming, particularly affecting lipid metabolism and amino acid pathways. The significant alterations in glycerophospholipids suggest membrane remodeling may play a crucial role in high-altitude adaptation mechanisms. (Fig. 1).

Fig. 1
figure 1

Multi-omics analysis reveals distinct molecular signatures in HAPH patients. (A) OPLS-DA score plot showing the separation between HAPH patients and controls based on metabolomic profiles (R2Y(cum) = 0.69, Q2(cum) = −0.85). (B) PCA score plot demonstrating metabolic differences between HAPH patients (P) and controls (C).The ellipse represents the 95% confidence interval. (C) Heatmap visualization of differentially expressed metabolites between HAPH patients and controls. Red indicates upregulation and blue indicates downregulation. (D) Volcano plot showing differentially expressed metabolites. Red dots represent upregulated metabolites (n = 12), blue dots represent downregulated metabolites (n = 14), and grey dots represent non-significant changes (|log2FC| > 1, p.adjust < 0.05). (E) Fold change analysis of significantly altered metabolites. The x-axis represents log2 fold change, and the y-axis shows metabolite names. (F) KEGG pathway enrichment analysis of differential metabolites. The size of dots represents the count of enriched metabolites, and the color indicates the p-value.

Full size image

Top 15 differentially expressed metabolites in HAPH patients compared to healthy controls were identified based on VIP > 1, P < 0.05 and p.adjust < 0.2 (Table 2).

Table 2 Top 15 differentially expressed metabolites in HAPH patients compared to healthy controls.
Full size table

Proteomic profiling reveals distinct metabolic signatures in haph

Quantitative proteomic analysis using LC-DIA-MS identified a total of 688 proteins. Of these, 35 proteins were differentially expressed in HAPH compared to controls, with 5 proteins upregulated and 30 downregulated. Hierarchical clustering analysis demonstrated distinct protein expression patterns between HAPH patients and healthy controls. Our transcriptomic analysis revealed significant alterations in gene expression profiles between HAPH patients and healthy controls. The most significantly differentially expressed genes were identified using stringent statistical criteria (P.adjust < 0.2). The most significantly upregulated genes included: MYH10 (FC = 7.574, P.adjust = 0.189), HNRNPA3 (FC = 2.103, P.adjust = 0.126), GXYLT2 (FC = 2.027, P.adjust = 0.033), ACTG1 (FC = 1.815, P.adjust = 0.117). Several genes showed significant downregulation: IGLL1 (FC=−1.557, P.adjust = 0.003), IGHV3-35 (FC=−1.375, P.adjust = 0.154), APOF (FC=−1.295, P.adjust < 0.001), IGHV3OR16-12 (FC=−1.162, P.adjust = 0.115), IGHV1-69 (FC=−1.026, P.adjust = 0.110). Notable changes were observed in immunoglobulin-related genes: Multiple IGHV family members showed consistent downregulation, IGLV3-9 (FC=−0.938, P.adjust = 0.189), IGLV5-39 (FC=−0.878, P.adjust = 0.189). Several regulatory proteins showed significant changes: VIM (FC = 0.854, P.adjust = 0.169), ORM1 (FC=−0.753, P.adjust = 0.185), CRISP3 (FC=−0.687, P.adjust = 0.189), GGH (FC=−0.679, P.adjust = 0.131). These findings suggest that HAPH is associated with substantial changes in protein expression, particularly affecting cytoskeletal organization (MYH10, ACTG1), immune response (immunoglobulin genes), and regulatory pathways. The downregulation of multiple immunoglobulin-related genes suggests potential alterations in immune function in HAPH patients. (Fig. 2). (Table 3).

Fig. 2
figure 2

Integrated proteomics analysis and pathway enrichment in HAPH. (a) Volcano plot showing the distribution of differentially expressed genes. Blue dots represent down-regulated genes, red dots represent up-regulated genes, and grey dots indicate genes with no significant changes (|log2 fold change| > 1, p.adjusted < 0.2). (b) Hierarchical clustering heatmap of differentially expressed genes across different samples. The color scale represents the normalized expression values, with red indicating up-regulation and blue indicating down-regulation. Sample groups are shown at the bottom of the heatmap. (c) Distribution of differentially expressed genes across major COG/KOG functional categories. The y-axis shows the frequency of annotated genes, and different colors represent distinct functional categories: RNA processing and modification (A), Active ion transport and metabolism (E), Signal transduction mechanisms (T), Defense mechanisms (V), and others. (d) Gene Ontology (GO) enrichment analysis of differentially expressed genes. The bar plot shows the number of genes (Count) associated with each GO term. The x-axis displays the GO terms, and the y-axis represents the gene count. (e) Enrichment analysis of key biological pathways. The dot plot shows the enrichment factors of significant pathways (p < 0.05). The size of dots represents the gene count, and the color intensity indicates the p-value significance level. Notable pathways include complement and coagulation cascades, cytoskeleton in muscle cells, regulation of actin cytoskeleton, and cholesterol metabolism. Data are presented as mean ± SD. Statistical significance was determined using Student’s t-test with Benjamini-Hochberg correction for multiple testing. OPLS-DA: Orthogonal Partial Least Squares Discriminant Analysis; PCA: Principal Component Analysis; FC: Fold Change; KEGG: Kyoto Encyclopedia of Genes and Genomes; GO: Gene Ontology.

Full size image
Table 3 Top 15 differentially expressed proteins in HAPH patients compared to healthy controls.
Full size table

Validation of HAPH candidate proteins by PRM

PRM was used to validate the expression of 20 candidate proteins, including MYH10, EGFR, HNRNPA3, GXYLT2, ACTG1, IGHV3-35, RPS7, HNRNPM, VNN1, YWHAZ, IGLV1-40, PCBP2, GGH, IL1RAP, H6PD, H2AC8, VIM, APOF, CRISP3, and CST3. Five proteins were successfully validated in the independent cohort: ACTG1, VNN1, and CKB were confirmed to be upregulated, while APOF and CST3 were downregulated in HAPH patients.

Integrated pathway analysis of metabolomic and proteomic data

To comprehensively understand the molecular mechanisms underlying HAPH, we performed an integrated analysis of metabolomic and proteomic data using KEGG pathway mapping. The integration revealed several significantly enriched pathways that demonstrate the complex interplay between metabolic alterations and protein expression changes.(a) Lipid Metabolism Network: Glycerophospholipid metabolism was significantly enriched (P.adjust < 0.05): Upregulated phospholipids: PC(20:4/8Z,11Z) (FC = 2.804, P.adjust = 0.047), Associated proteins: APOF (FC=−1.295, P.adjust < 0.001), This suggests membrane remodeling and lipid homeostasis disruption. (b) Immune System Regulation: Immunoglobulin-related proteins: IGLL1 (FC=−1.557, P.adjust = 0.003), Complement cascade components: C8A (FC=−0.354, P.adjust = 0.155), Associated metabolites: Aspartyl-alanine (FC = 0.864, P.adjust = 0.182). (c) Amino Acid Metabolism: Protein components: SERPINA1 (FC=−0.515, P.adjust = 0.033), Related metabolites: L-Hexahydro-3-hydroxypyridine (FC = 0.387, P.adjust = 0.091).(d) Cellular Structure and Organization: Cytoskeletal regulation pathway showed significant enrichment: Structural proteins: MYH10 (FC = 7.574, P.adjust = 0.189), ACTG1 (FC = 1.815, P.adjust = 0.117), Associated metabolites: Plastoquinone 3 (FC = 2.699, P.adjust = 0.182). (e) Oxidative Stress Response: Integrated analysis revealed activation of: Antioxidant proteins: GGH (FC=−0.679, P.adjust = 0.131), Related metabolites: Pantothenol (FC = 2.253, P.adjust = 0.182)10,11.

Discussion

This study represents the first comprehensive integration of proteomics and metabolomics in the analysis of serum from HAPH patients. Our findings revealed significant alterations in both protein expression and metabolite levels, providing new insights into the molecular mechanisms underlying HAPH. The research identified 35 differentially expressed proteins (DEPs) and 26 differentially expressed metabolites (DEMs), and further validated the upregulation of ACTG1, VNN1, and CKB, and the downregulation of APOF and CST3 as candidate proteins associated with HAPH.

Candidate protein biomarkers in HAPH

The five proteins validated by PRM represent potential biomarkers for HAPH. ACTG1, which encodes γ-actin, is involved in muscle contraction and cytoskeletal assembly. In adult rhabdomyoblasts, γ-actin participates in force transduction and transmission in muscle cells12, which is vital for maintaining cell morphology. ACTG1 has been identified as a candidate protein in studies of atrial defects combined with pulmonary hypertension13, but its specific role in HAPH remains unclear and warrants further investigation.VNN1, a pantetheine hydrolase, is involved in glucose metabolism, oxidative stress, inflammation, and cell migration. It has been implicated in the pathogenesis of several chronic diseases14, although its association with pulmonary hypertension has not been previously established. The significant increase in VNN1 expression in HAPH patients suggests its potential involvement in disease development, possibly through its role in oxidative stress regulation under hypoxic conditions. CKB, the brain-type creatine kinase, plays a crucial role in energy metabolism through the phosphocreatine pathway. Its upregulation in HAPH patients may reflect an adaptive response to the increased energy demands under chronic hypoxic conditions. Creatine kinase is essential for maintaining ATP levels in tissues with high energy requirements, and its dysregulation has been implicated in various cardiovascular and metabolic disorders. Conversely, APOF, which is involved in lipid metabolism and cholesterol transport, was significantly downregulated in HAPH patients. APOF helps maintain lipid homeostasis by inhibiting cholesterol ester transfer protein15. and its downregulation may contribute to alterations in lipid metabolism observed in HAPH. CST3, another downregulated protein, functions as a cysteine protease inhibitor and has been associated with kidney function and cardiovascular disease. Its reduced expression in HAPH patients may reflect dysregulation of protease activity and extracellular matrix remodeling, processes that are crucial in pulmonary vascular pathology.

Disrupted lipid metabolism and membrane remodeling

One of the most striking findings was the significant dysregulation of lipid metabolism, particularly in glycerophospholipid pathways. The upregulation of several phosphatidylcholines, notably PC(20:4/8Z,11Z) (FC = 2.804, P.adjust = 0.047), coupled with downregulation of APOF (FC=−1.295, P.adjust < 0.001), suggests substantial membrane remodeling in HAPH. This finding aligns with previous studies showing that hypoxia-induced membrane restructuring is crucial for cellular adaptation to high altitude16. The altered lipid profile may contribute to pulmonary vascular remodeling, HAPH pathogenesis, consistent with findings lipid metabolism in high altitude pulmonary arterial hypertension17,18.

Immune system modulationd

Our integrated analysis revealed significant changes in immune-related proteins and metabolites. The consistent downregulation of immunoglobulin-related proteins, including IGLL1 (FC=−1.557, P.adjust = 0.003) and multiple IGHV family members, suggests altered immune function in HAPH patients. This immune modulation may represent an adaptive response to chronic hypoxia, potentially influencing inflammatory processes in pulmonary vasculature19,20. The concurrent changes in complement cascade components, such as C8A (FC=−0.354, P.adjust = 0.155), further support the involvement of immune system regulation in HAPH development20.

Cytoskeletal reorganization and cellular structure

The significant upregulation of cytoskeletal proteins, particularly MYH10 (FC = 7.574, P.adjust = 0.189) and ACTG1 (FC = 1.815, P.adjust = 0.117), indicates substantial cellular structural reorganization. This finding aligns with previous work demonstrating that cytoskeletal remodeling plays a crucial role in pulmonary vascular smooth muscle cell proliferation21,22,23. The associated changes in metabolites like Plastoquinone 3 (FC = 2.699, P.adjust = 0.182) further support the involvement of structural reorganization in HAPH pathogenesis24.

Oxidative stress response

Our data revealed a coordinated response to oxidative stress, evidenced by changes in both antioxidant proteins and related metabolites. The downregulation of GGH (FC=−0.679, P.adjust = 0.131) coupled with upregulation of Pantothenol (FC = 2.253, P.adjust = 0.182) suggests activation of protective mechanisms against oxidative damage. This finding is consistent with Stenmark’s research on the role of oxidative stress in pulmonary hypertension25.

Amino acid metabolism and protein homeostasis

The alterations in amino acid metabolism, including changes in SERPINA1 (FC=−0.515, P.adjust = 0.033) and related metabolites like L-Hexahydro-3-hydroxypyridine (FC = 0.387, P.adjust = 0.091), indicate significant changes in protein homeostasis. These modifications may reflect cellular adaptation to chronic hypoxic conditions and could influence vascular remodeling processes26.

Conclusions

This comprehensive study provides novel insights into the molecular mechanisms of HAPH through integrated metabolomic and proteomic analyses. Our findings demonstrate that HAPH pathogenesis involves complex interactions among multiple biological systems, particularly: Disrupted lipid metabolism and membrane remodeling, evidenced by alterations in glycerophospholipid pathways and associated proteins. Modified immune system function, characterized by downregulation of immunoglobulin-related proteins and complement cascade components. Extensive cytoskeletal reorganization, indicated by upregulation of structural proteins and associated metabolites. Activated oxidative stress response mechanisms, reflected in coordinated changes of antioxidant proteins and metabolites.

Limitations

While this study offers valuable insights, it is essential to acknowledge its limitations. Although we adhered to established guidelines for diagnosing HAPH through ultrasound-estimated pulmonary artery systolic pressure, it is crucial to recognize the inherent limitations of this method. Future research should consider integrating multiple diagnostic modalities, including right heart catheterization, to delve deeper into the pathogenesis of HAPH. The sample size, while appropriate for a discovery study, limits the statistical power for detecting subtle alterations in protein and metabolite levels. Additionally, the cross-sectional design prevents assessment of temporal changes in biomarker levels and their relationship to disease progression. Longitudinal studies with larger cohorts are needed to address these limitations and validate our findings.