COPD target protein based on plasma proteomics, application and medicine for treating COPD

Abstract

The invention provides a COPD (chronic obstructive pulmonary disease) target protein based on plasma proteomics. The target protein comprises one or more of the following four proteins: HLA-DRA (human leukocyte antigen-DRA), CXCL9, TNFRSF8 and RNASET2. The invention not only highlights the important value of longitudinal proteomics analysis in revealing the early biological change of COPD, but also opens up a new path for the development of personalized treatment schemes. In view of the complex function of protein in the biological regulation and control process, the research result of the invention provides an important framework for understanding the pathological mechanism of COPD, and lays a foundation for developing a treatment strategy based on proteomics in the future. The invention also provides an application based on the COPD target protein and a medicine for treating COPD.

Description

Technical Field

[0001] This invention relates to a COPD target protein based on plasma proteomics, and its application and treatment of COPD. Background Technology

[0002] Chronic obstructive pulmonary disease (COPD), characterized by persistent airflow obstruction and chronic airway inflammation, is the third leading cause of death worldwide. Despite its high prevalence, COPD is often diagnosed at a late stage, limiting the window for early intervention. Early diagnosis and timely, effective treatment are crucial for slowing disease progression and improving prognosis. While many treatments for COPD are currently available, they primarily serve to relieve symptoms and have limited effectiveness in preventing the progressive decline of lung function.

[0003] COPD is a multifactorial disease driven by genetic susceptibility and environmental factors such as tobacco smoke, air pollution, and poor diet, but its pathogenesis remains incompletely understood. Elucidating its complex mechanisms and the relationship between risk factors and disease progression is crucial for developing effective prevention and treatment strategies for COPD. Advances in genomics and proteomics have provided key insights into translational regulation and its role in various cellular processes and disease pathophysiology. Analyzing plasma protein alterations before the onset of COPD can reveal early biomarkers and molecular mechanisms, facilitating early diagnosis, timely intervention, and the development of targeted therapeutic strategies.

[0004] Multiple case-control studies have shown significant differences in plasma protein expression profiles between COPD patients and healthy individuals. While these findings are of significant research value, these studies are all cross-sectional, limiting the determination of whether observed protein changes are a cause or consequence of disease development, highlighting the need for longitudinal studies. To date, only two studies have explored the association between baseline plasma protein levels and the subsequent risk of developing COPD. The first was a prospective cohort study of 5247 men, identifying several inflammation-related proteins associated with an increased long-term COPD risk. However, limitations such as limited proteomics data, a small number of COPD-related events, and a focus only on hospitalized cases of severe COPD restricted a broader understanding of COPD-related proteomic changes. The second study used a multi-omics approach to identify SCARF2 as a potential biomarker for COPD risk. Despite extensive validation, its limited protein coverage restricts the identification of a wider range of biomarker candidates. Therefore, future work should integrate large-scale proteomics, prospective cohort studies, and genetic analyses to comprehensively understand proteomic markers in COPD risk. The UK Biobank (UKB) cohort provides comprehensive phenotypic and proteomics data, laying a solid foundation for studying the association between plasma proteomic profiles and COPD pathogenesis.

[0005] To investigate the relationship between large-scale proteomics and the risk of COPD, this invention integrates evidence from observational and genetic analyses. First, the association between 2726 baseline plasma proteins and COPD incidence was analyzed in 52,513 UKB participants. Further investigation was conducted into the association between COPD-related proteins and genetic susceptibility, lung function, lifestyle behaviors, and environmental exposures (including air pollution). In addition, a series of biological functional analyses were performed, including enrichment analysis, transcription factor (TF) analysis, protein-protein interaction (PPI) analysis, and single-cell expression analysis, to systematically elucidate the pathogenesis of COPD. The therapeutic potential of COPD-related proteins was systematically analyzed through drug accessibility assessment. Mendelian randomization (MR) and colocalization analysis were used to assess the causal relationship between plasma proteins and COPD and identify potential therapeutic targets. Finally, a COPD diagnostic model was constructed based on proteins identified by MR. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the main objective of this invention is to provide an application of COPD target proteins based on plasma proteomics and a drug for treating COPD.

[0007] To achieve the above-mentioned main objectives, on the one hand, the present invention provides a COPD target protein based on plasma proteomics, which includes one or more of the following four proteins: HLA-DRA, CXCL9, TNFRSF8, and RNASET2.

[0008] On the other hand, the present invention provides an application of the above-mentioned COPD target protein based on plasma proteomics, wherein the target protein is used in the preparation of a drug for treating COPD.

[0009] In another aspect, the present invention provides a drug for treating COPD, which targets one or more of the following four proteins: HLA-DRA, CXCL9, TNFRSF8, and RNASET2.

[0010] The present invention has the following beneficial effects:

[0011] This invention not only highlights the significant value of longitudinal proteomics analysis in revealing early biological changes in COPD, but also opens new avenues for the development of personalized treatment regimens. Given the complex functions of proteins in regulating biological processes, the findings of this invention provide an important framework for understanding the pathological mechanisms of COPD and lay the foundation for future development of proteomics-based therapeutic strategies.

[0012] The present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0013] Figure 1 This is a research overview, which includes:

[0014] Top panel: The data used in this study include chronic obstructive pulmonary disease (COPD) diagnostic information, genotypic data, proteomics data, lung function data, as well as modifiable lifestyle factors and COPD risk factors such as air pollution;

[0015] Upper left panel: Identification of proteins significantly associated with new-onset COPD using a Cox proportional hazards regression model;

[0016] Top right panel: This study explored the association between plasma proteins and COPD-related factors, including lung function, genetic factors, and modifiable lifestyle and air pollution risk factors.

[0017] The lower middle panel: Through a variety of methods such as pathway enrichment analysis, phenotypic enrichment analysis, transcription factor analysis, protein-protein interaction network analysis and single-cell expression analysis, we explore the biological functions of identified proteins and reveal their potential mechanisms of action in COPD.

[0018] Bottom panel: Mendelian randomization analysis was used to explore the potential causal relationship between plasma proteins and COPD, aiming to identify potential drug targets for COPD.

[0019] In addition, colocalization analysis was conducted, integrating data from multiple databases such as DrugBank, Open Targets, and Therapeutic Target Database to screen for drugs with potential for treating COPD. Finally, a diagnostic model for COPD was constructed based on plasma proteins identified by MR.

[0020] Figure 2 shows the proteomic identification of proteins associated with the occurrence of COPD, where:

[0021] Figure 2a It is a volcano plot showing the association between protein levels and the occurrence of COPD using a Cox proportional hazards regression model;

[0022] Figure 2b This is a volcano plot, showing the results of a replication analysis performed using a linear regression model;

[0023] Figure 2c This indicates a comparison of the roles of COPD-related proteins (blue dots) in this study with those in previous studies (red dots).

[0024] Figure 2dThis indicates the association between COPD-related proteins and baseline lung function.

[0025] Figure 3 shows the association between COPD-related proteins and genetic and environmental factors, among which...

[0026] Figure 3a We used a linear regression model to study the association between COPD-related proteins and modifiable lifestyle factors.

[0027] Figure 3b A linear regression model was used to study the correlation between COPD-related proteins and air pollution.

[0028] Figure 3c This indicates the association between COPD-related proteins and the genetic risk of COPD at five p-thresholds;

[0029] Figure 3d The forest plot shows the estimated genetic correlation (rg) between plasma proteins and COPD.

[0030] Figure 4 shows the biological function analysis of proteins associated with chronic obstructive pulmonary disease, in which:

[0031] Figure 4a The results of the biological pathway enrichment analysis are shown on the horizontal axis; the horizontal axis represents the negative logarithm (-lg) of the p-value for each term. Only terms that remain statistically significant after FDR correction are displayed. The vertical axis lists different terms, with the source of each term distinguished by different colors. Abbreviations: GO: BP (Gene Ontology Biological Process Term); GO: MF (Gene Ontology Molecular Function Term); KEGG (Kyoto Encyclopedia of Genes and Genomes Pathway);

[0032] Figure 4b The results of the phenotypic enrichment analysis are shown on the horizontal axis, which represents the logarithmic value of each term, and the vertical axis, which corresponds to different phenotypes in the MGI database. P-values ​​were obtained using the Fischer exact test. Phenotypes that were significantly enriched after FDR correction are clearly labeled and displayed.

[0033] Figure 4c The results of transcription factor enrichment analysis are shown in the figure; the figure displays the transcription factors that were significantly enriched after FDR correction. Each enriched transcription factor target (colored) is linked to its associated gene (gray). The color bars indicate the strength of the association between the protein and the occurrence of COPD; the thicker the line, the stronger the association.

[0034] Figure 4d This represents a protein-protein interaction network of COPD-related proteins. Nodes represent proteins, and edges represent the associations between them. Larger and more central nodes correspond to proteins with higher height numbers, indicating stronger evidence supporting their interactions.

[0035] Figure 5 shows the single-cell expression of genes encoding 468 candidate proteins in COPD patients and healthy controls, among which:

[0036] Figure 5a This indicates that six cell types were identified in COPD patients through atlas-based tSNE clustering analysis.

[0037] Figure 5b The t-SNE plot shows the expression of each candidate protein-coding gene in different cell types in COPD patients and healthy controls.

[0038] Figure 5c The bubble chart shows the expression proportion and level of each candidate protein-coding gene in different cell types in COPD patients and healthy controls.

[0039] Figure 5d This indicates cell-specific enrichment of each candidate protein-coding gene.

[0040] Figure 6 This represents a COPD diagnostic model based on four MR-supported plasma proteins (CXCL9, RNASET2, HLA-DRA, and TNFRSF8). Detailed Implementation

[0041] Example 1

[0042] The prospective relationship between plasma proteomics characteristics and chronic obstructive pulmonary disease (COPD) remains largely unexplored. Studying pre-COPD proteomics alterations will provide important insights into disease pathogenesis and potential therapeutic interventions.

[0043] Using data from 52,513 participants with baseline COPD-free status in the UK Biobank, a Cox model was employed to examine the association between 2,726 plasma proteins and the risk of COPD. Linear regression analysis was used to systematically investigate the associations between COPD-related proteins and lung function, modifiable risk factors, air pollution, and polygenic risk scores. A series of research methods, including phenotypic enrichment analysis, pathway enrichment analysis, protein-protein interaction network (PPI) analysis, and single-cell expression analysis, were employed to reveal the biological significance of candidate proteins. Mendelian randomization (MR) analysis was used to identify potential causal proteins. Finally, based on the proteins identified by MR, a COPD diagnostic model was constructed using machine learning.

[0044] After Bonferroni correction, a total of 468 proteins were significantly associated with the occurrence of COPD, of which 83.5% of COPD-related proteins were positively correlated with COPD pathogenesis. Furthermore, genetic susceptibility to COPD and modifiable risk factors were significantly associated with COPD-related proteins. Bioenrichment analysis revealed the crucial role of immune and inflammatory responses in COPD development, and the PPI network identified two pivotal proteins, EGFR and TNF. Single-cell analysis showed that COPD-related proteins were enriched in epithelial cells, endothelial cells, macrophages, and T cells. MR analysis identified four proteins causally related to COPD: HLA-DRA (P = 2.09 × 10⁻¹²), CXCL9 (P = 1.48 × 10⁻⁵), TNFRSF8 (P = 8.32 × 10⁻⁵), and RNASET2 (P = 1.79 × 10⁻⁴). In a COPD diagnostic model, the area under the curve (AUC) for the four protein models was 0.77.

[0045] This study identified specific proteomic features causally related to the pathogenesis of COPD, advancing our molecular understanding of the etiology of COPD. These findings provide potential targets for future COPD prevention strategies and personalized treatment plans.

[0046] I. Methods

[0047] 1. Study population

[0048] This study used data from the UKB cohort study, which included approximately 500,000 participants from 22 assessment centers. The data encompassed multiple follow-up records, updated health records, and information on sociodemographic characteristics, lifestyle, and blood samples. All participants signed written informed consent forms for the UKB project, which had received ethical approval from the National Research Ethics Service Northwest (NRWS) (approval numbers: 11 / NW / 0382, 16 / NW / 0274, and 21 / NW / 0157). Data access was obtained through project code 83339, in accordance with UKB ethical guidelines and access protocols.

[0049] External validation was performed using data from the European Prospective Study of Cancer and Nutrition (EPIC) cohort, consisting of 25,639 participants recruited from Norfolk. The ethical approval was granted by the Norfolk Research Ethics Committee (reference number: 05 / Q0101 / 191), and all participants provided written informed consent.

[0050] 2. Blood proteomics

[0051] Blood samples collected by UKB were processed by Olink Analytical Services in Sweden using antibody-based Olink Explore. TMThe proximity extension assay ultimately identified 2726 proteins. Detailed information regarding sample selection, protein data processing, quality control procedures, and protein concentration standardization has been documented in previous studies. This study used the standardized protein expression (NPX) values ​​for each protein from each participant.

[0052] Proteomic analysis of serum samples collected from the EPIC cohort was performed using Olink Explore 1536 and the Olink Explore Expansion panel, targeting a total of 2923 unique proteins. Subjects with failed proteomics quality control or missing data on age, sex, body mass index (BMI), or smoking status were excluded.

[0053] 3. Chronic obstructive pulmonary disease status assessment

[0054] COPD cases were identified based on hospital inpatient records (field 41270) and the diagnosis date (field 41280) extracted from associated medical records. COPD diagnoses were coded as J40-J44 according to the International Classification of Diseases, Tenth Revision (ICD-10). Subjects without COPD at baseline were followed up until the earliest of the following events: COPD diagnosis, death, or review date (October 31, 2022). Patients diagnosed with COPD before baseline were excluded from the analysis.

[0055] 4. Selection of covariates

[0056] The study adjusted for demographic variables, including sex (male / female), age, education level (college degree or not), race (white / other ethnicities), physical activity (frequent / infrequent exercise), BMI, Thomson Deprivation Index (TDI), alcohol consumption frequency (divided into six levels from abstaining to daily drinking), smoking status (never smoked / used to smoke / current smoker), dietary habits (healthy diet / poor diet), and income level (low income). Middle-income high income Diabetes, high blood pressure, and asthma.

[0057] 5. COPD-related factors

[0058] The lung function parameters analyzed included forced vital capacity (FVC), forced expiratory volume in one second (FEV1), and the FEV1 / FVC ratio. Modifiable lifestyle factors considered in the analysis included smoking status, alcohol consumption, BMI, physical activity, and diet quality. Alcohol consumption was categorized as ≥3 times per week or <3 times per week; smoking status was categorized as never-smoker or former / current smoker; and BMI was categorized as <18.5 or ≥18.5 kg / m². 2Physical activity is categorized as regular or irregular; dietary status is categorized as healthy or unhealthy; estimates of long-term exposure to air pollution are based on annual average particulate matter concentrations (PM2.5, diameter ≤ 2.5 μm). 25 ) and ≤10μm (PM 10 ) and nitrogen oxides (NO2) and nitrogen oxides (NO) x ).

[0059] 6. The association between plasma proteins and COPD

[0060] This study used a Cox proportional hazards regression model to assess the association between plasma protein levels and the risk of COPD in the UKB cohort, and calculated the hazard ratio (HR) and its 95% confidence interval (CI). The model adjusted for the following covariates: age, sex, education level, race, alcohol consumption, smoking status, TDI, BMI, income level, physical activity, dietary structure, diabetes, hypertension, and asthma. All covariates were obtained from baseline data, with less than 20% of the observed missing values. In this study, chain equation multiple imputation was used to imputate missing covariate data. Multiple tests were performed with false discovery rate (FDR) correction, and a p-value <0.05 after correction was defined as statistically significant. In sensitivity analysis, individuals who developed COPD within two years after baseline and individuals participating in the UKB-Pharma Proteomics Project (UKB-PPP) consortium were excluded to reduce adverse causality, and the time association between baseline plasma protein and COPD occurrence was examined before Cox regression analysis.

[0061] To conduct external validation, Cox regression analysis was used to assess the relationship between plasma proteins and the risk of COPD in the EPIC cohort. The model adjusted for the following covariates: age, sex, race, TDI, BMI, alcohol consumption, diabetes, and hypertension.

[0062] In addition, linear regression analysis was performed in the UKB cohort to examine the relationship between plasma proteins and lung function (FVC, FEV1, FEV1 / FVC), using the same covariates as in the Cox analysis described above.

[0063] 7. Correlation between plasma proteins and COPD risk factors

[0064] Linear regression models were used to assess the relationship between plasma protein levels and modifiable lifestyle factors and air pollution. Modifiable lifestyle models were adjusted for age, sex, race, TDI, education level, income, diabetes, hypertension, and asthma. Air pollution models were adjusted for age, sex, race, education level, physical activity, TDI, BMI, income, smoking status, alcohol consumption, diet, diabetes, hypertension, and asthma. Multiple-test corrections were performed on all models using FDR correction.

[0065] 8. Calculation of Polygenic Risk Score (PRS)

[0066] The PRS (Prescription Score) is based on genome-wide association studies (GWAS) of COPD, reflecting an individual's genetic susceptibility to COPD. Higher PRS scores correlate with an increased risk of developing the disease. This score is calculated by weighting the number of risk alleles carried by each participant, specifically by extrapolating the effect alleles and odds ratios from the summary statistics of clustered COPD GWAS.

[0067] P-value-guided clustering analysis was performed using PLINK (r² = 0.1, within a 250 kbps window), with major histocompatibility complex (MHC) regions represented by single SNPs. PRS was calculated at multiple representative p-value thresholds (pT), including P < 0.001 (PT_0.001), P < 0.005 (PT_0.05), P < 0.01 (PT_0.01), P < 0.02 (PT_0.02), P < 0.03 (PT_0.03), P < 0.04 (PT_0.04), and P < 0.05 (PT_0.05). Multiple imputation of missing genotypes or phenotypes was performed using the chain equation method.

[0068] 9. Relationship between COPD plasma proteins and PRS

[0069] To investigate the impact of genetic susceptibility on plasma proteins and COPD risk, a linear model was used to assess the association between COPD-related proteins and COPD PRS. The same covariates as in the Cox analysis were adjusted, and multiple tests were performed using FDR correction.

[0070] 10. Genetic correlation between plasma proteins and COPD

[0071] Genetic associations between plasma proteins and COPD were calculated using linkage disequilibrium score regression (LDSC) based on protein quantitative trait loci (pQTL) data and GWAS pooled data. The pQTL data for plasma proteins were derived from 54,219 individuals measured on the Olink platform in the UKB-PPP, while the GWAS data for COPD, including 58,559 cases and 937,358 controls, came from the Global Biobank Meta-Analysis Project (GBMI), covering data from 23 biobanks. All individuals in both datasets were of European descent. Covariance relationships between datasets were assessed, taking into account linkage disequilibrium structures, based on pre-calculated European linkage disequilibrium (LD) scores from the 1000 Genomes Project reference data. To ensure the reliability of linkage disequilibrium estimates, the linkage disequilibrium correlation coefficient (LDSC) analysis was limited to single nucleotide polymorphisms (SNPs) within the HapMap3 reference panel. To investigate whether specific genetic variations affect protein expression associated with chronic obstructive pulmonary disease (COPD), 468 proteins identified in previous studies were analyzed. Multiple tests are performed using FDR correction.

[0072] 11. Functional enrichment analysis

[0073] To systematically investigate the biological functions of COPD-related proteins, this study utilized the Metascape platform, which integrates information from multiple biological pathways and annotation resources. Pathway enrichment analysis was performed based on pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO), covering three main categories: biological processes, molecular functions, and cellular components. Screening criteria included: p-value <0.01, minimum number of genes ≥3, and enrichment factor greater than 1.5 (derived by calculating the ratio of observed to expected values). Based on previous research, these terms were clustered according to member similarity.

[0074] To explore phenotypic enrichment, the Mouse Genome Informatics (MGI) database was used. This database provides new insights into biological processes and is widely used in biomedical research to study human diseases. The MGI database provides 19,323 phenotypic-annotated genes, of which 2,286 genes encoding 2,726 plasma proteins serve as a reference set. Of the 468 genes encoding COPD-related proteins, 400 were mapped to the MGI-annotated gene list. Fisher's exact test was applied to assess the statistical significance of each phenotype by comparing the proportion of genes in the target gene set that were associated with the phenotype to the proportion of genes in the background gene set.

[0075] To investigate the upstream regulatory mechanisms of COPD-related proteins, transcription factor analysis was performed using the TRRUST database. TRRUST is a systematically organized database containing 9396 transcription factor-target regulatory interactions involving 798 human transcription factors. By integrating data from Metascape and TRRUST, transcription factors potentially involved in the regulation of COPD-related proteins were systematically screened. The 2726 protein-coding genes of interest in this study were used as a reference gene set.

[0076] 12. PPI Network Mapping

[0077] To investigate potential interactions among COPD-related proteins, a PPI network was constructed and visualized using Metascape and Cytoscape. The Molecular Complex Detection (MCODE) algorithm was applied to identify highly interconnected protein clusters. Pathway and process enrichment analyses were performed on each cluster detected by MCODE, and the three most significant terms were retained for functional annotation.

[0078] 13. Single-cell expression enrichment analysis

[0079] To assess cell type-specific expression of COPD-related genes, this study used the single-cell RNA sequencing dataset (GSE279570), which included 11 COPD patients and 7 healthy controls (HCs). Using the Seurat R package (version 3.2.0), cells expressing 500-6000 genes, with a mitochondrial gene percentage below 10%, and expressed in at least 50 cells were selected. Data were then normalized to exclude the influence of mitochondrial gene percentage. The top 10 principal components obtained through principal component analysis (PCA) were used in downstream procedures. Cluster analysis was performed at a resolution of 0.5, and t-distributed random neighborhood embedding (t-SNE) was applied to the integrated Seurat unified data objects. This analysis included expression profiles of 12,193 genes from 42,764 cells.

[0080] Differential expression analysis was performed using the Wilcoxon rank-sum test to compare the expression levels of candidate protein-coding genes in specific enriched cell types between COPD and HCs. Specific enriched cell types for a gene were identified based on a mean |log2FC| greater than 0.25 and an FDR-corrected p-value less than 0.05. This criterion was also used to identify differentially expressed genes in the COPD and HC groups within specifically enriched cell types. Cell types were annotated using the SingleR software package, with the human primary cell atlas as a reference.

[0081] 14. Evaluation of druggability

[0082] Genome Relocation Programming (GREP) was used to identify candidate therapeutics by associating gene sets with known drug targets. Based on this, this study evaluated whether genes encoding proteins related to chronic obstructive pulmonary disease (COPD) showed enrichment in various diseases according to the International Classification of Diseases, 10th Revision (ICD-10). The analysis integrated information from the DrugBank, Open Targets, and Therapeutic Target databases, using a background gene set containing 2726 test proteins as the basis for the analysis.

[0083] 15. Mendelian randomization analysis

[0084] Using pQTL data of 468 plasma proteins that were identical to those used in the LDSC analysis and GWAS data of COPD, a two-way MR analysis was employed to investigate the potential causal relationship between COPD-related proteins and COPD risk. Proteins with significant differences at the genome-wide level (P < 5 × 10⁻⁶) were screened. -8 And meets the chain imbalance criterion (r) 2 Single nucleotide polymorphisms (SNPs) with a value <0.01 and a distance >1000 kb were used as instrumental variables. To ensure the strength of the instrumental variables, only SNPs with an F-statistic greater than 10 were included in the analysis. In addition, the Steiger filter was used to exclude SNPs with intermediate frequencies or palindromic alleles.

[0085] In the primary analysis, the Wald ratio method was used when only a single SNP was present, while the inverse variance weighted (IVW) method was used when two or more SNPs were present. To assess the robustness of the results, sensitivity analyses were performed using complementary methods, including MR-Egger, weighted median, simple mode, and weighted mode. The MR-Egger intercept test allowed for non-zero intercepts, which helped detect and correct for level pleiotropy. The MR-PRESSO (pleiotropy residuals and outliers) method was used to further evaluate pleiotropy. Furthermore, Cochran's Q test was used to assess the heterogeneity of the genetic instruments. Results were considered statistically significant using an FDR < 0.05. All analyses were performed using the TwoSampleMR and MR-PRESSO packages in R software (version 4.4.1).

[0086] 16. Bayesian Colocality Analysis

[0087] To further elucidate the causal relationship between plasma proteins and COPD discovered in MR analysis, colocalization analysis was employed to assess the probability that the two features share a common causal variant, rather than an association solely due to LD. Posterior probabilities were calculated for five hypotheses using the default parameter settings of the R package "coloc" (version 5.2.1). A posterior probability (PPH4) ≥ 0.7 for hypothesis 4 was considered strong evidence of colocalization, indicating that the association between the plasma protein and COPD is driven by the same causal variant.

[0088] 17. Diagnostic Model Construction

[0089] Based on plasma proteins identified by MR, this study developed a COPD diagnostic model using the Lightweight Gradient Boosting Machine (LightGBM) algorithm. Subjects diagnosed with COPD at baseline in the UKB cohort were included in this analysis and randomly assigned to the training set (80%) and the test set (20%). To minimize overfitting, internal 5-fold cross-validation was applied to the training set, the maximum number of iterations was limited to 100, and early stopping was implemented. Balanced class weights were used to adjust for class imbalance. The diagnostic performance of the model was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). The 95% confidence interval of the AUC was estimated using a bootstrap method (1000 iterations).

[0090] II. Results

[0091] 1. Participant characteristics

[0092] A total of 52,513 participants who were not diagnosed with COPD at baseline were included in the analysis. The mean age of the participants was 56.76 years (SD = 8.21), of whom 54.0% were female and 93.7% were white. The median follow-up time was 13.57 years (interquartile range [IQR] = 12.76–14.35; range = 0.01–16.63; mean = 12.99). During the follow-up period, 2,678 participants were diagnosed with COPD. All participants used Olink Explore. TM The platform performs blood proteomic analysis. A total of 2,726 plasma proteins passed quality control and were included in the analysis. Figure 1 This demonstrates the framework of this study.

[0093] 2. The association between plasma proteins and COPD

[0094] Using a Cox proportional hazards model, this study assessed the relationship between plasma proteins and COPD incidence in individuals without COPD at baseline. After Bonferroni correction (P = 0.05 / 2726 = 1.83 × 10⁻⁶), the results were statistically significant. -5), 468 proteins were significantly associated with new-onset COPD ( Figure 2a Among the plasma proteins associated with COPD, 391 proteins were positively correlated with COPD risk, while 77 proteins were negatively correlated with COPD risk.

[0095] Notably, whey acidic protein tetradisulfide core domain protein 2 (WFDC2) (HR = 2.18 [2.05–2.33], P = 2.22 × 10⁻⁶) -125 ), CXC motif chemokine ligand 17 (CXCL17) (HR=1.89[1.78-2.01], P=1.21×10 ), -90 ), urokinase plasminogen activator surface receptor (PLAUR) (HR=2.96 [2.66–3.30], P=4.11×10 ), -87 ), Growth / differentiation factor 15 (GDF15) (HR=1.77[1.66–1.87], P=9.54×10), -82 ), and lysosome-associated membrane protein 3 (LAMP3) (HR = 1.81 [1.70–1.92], P = 1.24 × 10⁻⁶). -78 It is most strongly associated with COPD.

[0096] In addition, tendonogen-R (TNR) (HR = 0.53 [0.49–0.58], P = 3.33 × 10⁻⁶) -53 Carbonic anhydrase 6 (CA6) (HR=0.75 [0.71–0.79], P=7.19×10⁻⁶) -24 ), contactin-1(CNTN1)(HR=0.57[0.51–0.64], P=8.62×10 -24 ), Advanced Glycosylation Advanced Product Specific Receptor (AGER) (HR = 0.66 [0.61–0.72], P = 5.75 × 10⁻⁶) -22 ), and the Brevican core protein (BCAN) (HR = 0.62 [0.56–0.68], P = 6.37 × 10⁻⁶). -22 Elevated expression levels of this substance are significantly negatively correlated with the incidence of COPD.

[0097] In the sensitivity analysis, after excluding participants diagnosed with COPD within two years of baseline and those enrolled in the UKB-PPP consortium, 96.8% of COPD-related plasma proteins remained statistically significant after FDR correction. An additional sensitivity analysis using extrapolated protein data yielded consistent results. In the validation analysis, 505 patients diagnosed with COPD at baseline (excluded from the main analysis) had their plasma protein levels compared to 49,835 COPD-free controls. The results showed that all 468 COPD-related proteins exhibited consistent trends, with 441 proteins (94%) remaining statistically significant after FDR correction. Figure 2b In addition, a previous study examined 377 out of 468 plasma proteins and found that 124 of them were significantly associated with COPD (P < 2.7 × 10⁻⁶). -8 ; Figure 2c ).

[0098] In external validation, 308 out of 468 COPD-related proteins in the EPIC cohort were also associated with COPD (P<0.05). These results demonstrate that the COPD-related proteins identified in this study are highly reproducible, supporting the robustness and generalizability of existing findings.

[0099] 3. The association between COPD-related proteins and lung function

[0100] To explore the biological significance of COPD-related proteins, this study assessed their association with lung function parameters, including FEV1, FVC, and the FEV1 / FVC ratio. Of the 468 plasma proteins associated with COPD risk, 424, after FDR correction, showed a significant correlation with at least one lung function indicator. Figure 2d It is noteworthy that 386 proteins (such as WFDC2, PLAUR, CXCL17, CXCL9, GDF15, and TNF superfamily members) showed inverse effects on COPD risk and lung function parameters, suggesting that these proteins may be involved in the development of lung dysfunction.

[0101] 4. The association between COPD-related proteins and COPD risk factors

[0102] To further investigate the epidemiological relevance of COPD-related proteins, this study assessed their association with known modifiable risk factors, including unhealthy lifestyle factors and air pollution exposure. After FDR correction, 426 of the 468 COPD-related proteins were significantly associated with alcohol consumption (P<0.05), while 404 proteins were associated with unhealthy diets, 396 with lack of exercise, 373 with smoking behavior, and 297 with underweight. Notably, 448 proteins were significantly associated with three or more risk factors (P<0.05). Figure 3a Furthermore, after FDR correction, among the 468 proteins detected, 46 were significantly associated with PM2.5 exposure (P<0.05), while 25 were associated with NO2 exposure, 20 with NOx exposure, and 8 with PM10 exposure. Figure 3b ).

[0103] 5. The relationship between COPD-related proteins and COPD genetic factors

[0104] To assess the impact of genetic factors on COPD risk, this study evaluated the association between COPD-related proteins and COPD genetic susceptibility. Of the 468 COPD-related proteins, 300 were nominally significantly associated with PRS, and 181 of these remained statistically significant after FDR correction. Figure 3c However, LDSC analysis showed that among the 468 COPD-related proteins, 294 proteins had a significant genetic association with COPD. Figure 3d Overall, most COPD-related proteins are influenced by genetic factors, and a significant proportion of these proteins may be driven by protein-specific genetic determinants.

[0105] 6. Biological functions of COPD-related proteins

[0106] Pathway enrichment analysis revealed that 468 COPD-related proteins are primarily involved in immune pathways, such as signal receptor regulation, cytokine-cytokine receptor interaction, bacterial response, humoral immune response, and cell activation regulation. Figure 4a ).

[0107] Phenotypic enrichment analysis revealed that 468 COPD-related proteins showed the strongest enrichment in the immune system phenotype (Pexact = 7.26 × 10⁻⁶). -7 Furthermore, it was significantly enriched in the muscle phenotype (Pexact = 7.11 × 10⁻⁶). -3 Embryo phenotype (Pexact = 1.19 × 10⁻⁶) -2 ) and respiratory phenotype (Pexact = 1.48 × 10) -2 ) phenotype ( Figure 4b ).

[0108] In transcription factor network analysis, 16 transcription factors regulating COPD-related protein-coding genes were identified. Figure 4c The most significantly associated transcription factors include: nucleoside κB subunit 1 (NFKB1), NF-κB p65 subunit (RELA), transcription factor Sp1 (SP1), Jun proto-oncogene (JUN), transcription factor Sp3 (SP3), and signal transduction and transcription activator 3 (STAT3), which play key roles in inflammation and immunity.

[0109] PPI analysis of 468 COPD-related proteins revealed a protein expression network consisting of 379 nodes and 12 edges. Figure 4d Among them, epidermal growth factor receptor (EGFR) and tumor necrosis factor (TNF) are central nodes. Furthermore, MCODE analysis detected three highly interconnected modules enriched in cytokine-cytokine receptor interactions, humoral immune responses, and inflammatory responses.

[0110] 7. Cell-specific enrichment of genes encoding COPD-related proteins

[0111] In single-cell expression analysis, the cell type enrichment of 468 genes encoding COPD-related proteins and their expression differences between COPD patients and HCs were investigated. Of the 468 genes, 100 genes were significantly enriched in COPD-specific cell types, which were categorized into six types: endothelial cells, epithelial cells, macrophages, monocytes, natural killer (NK) cells, and T cells. Figure 5a ). Figure 5b This study shows the single-cell expression patterns of the top 12 cell type-specific genes in COPD patients and HCs. Expression data for all 12 genes were present in both COPD patients and HCs. Figure 5c ).

[0112] 8. Drug target discovery and diagnostic model development

[0113] A two-way, two-sample MR analysis was used to assess the potential causal relationship between COPD-related proteins and the disease. After FDR correction, the human leukocyte antigen D-associated α subunit (HLA-DRA) was found to be positive; OR = 1.07, 95% CI: 1.05–1.09, P = 2.09 × 10⁻⁶. -12 ), Cyclic CXC motif chemokine ligand 9 (CXCL9; OR = 1.12, 95% CI: 1.06–1.18, P = 1.48 × 10⁻⁶). -5), tumor necrosis factor receptor superfamily member 8 (TNFRSF8; OR = 1.08, 95% CI: 1.04–1.12, P = 8.32 × 10⁻⁶). -5 ), and ribonuclease T2 (RNASET2; OR = 1.05, 95% CI: 1.02–1.08, P = 1.79 × 10⁻⁶). -4 The expression levels of these proteins were significantly associated with increased COPD risk. Sensitivity analysis suggested no evidence of level pleiotropic or heterogeneous effects, supporting the robustness of the results. In contrast, reverse MR analysis showed no significant causal effect of COPD on these protein levels, with all FDR-corrected p-values ​​greater than 0.05. Of the four potential drug targets identified by MR analysis, only the co-localization of the TNFRSF8 gene with COPD showed strong evidence (PPH4 = 75.8%). This finding indicates a shared causal variation between TNFRSF8 expression and COPD risk (rs62065216).

[0114] To assess therapeutic potential, existing drugs targeting COPD-related protein-coding genes were searched in DrugBank, Open Targets, and Therapeutic Target databases. Currently, 125 COPD-related genes have been identified as therapeutic targets for 146 diseases, covering malignancies, tumors of unknown or undetermined nature, hemolytic anemia, ophthalmic diseases, non-tumor digestive system diseases, and inflammatory polyarthritis. Notably, TNF, EGFR, IL1B, RET, and FLT3 genes are each associated with more than 30 diseases. Of the four targets identified by MR analysis, TNFRSF8 is a target of Brentuximab vedotin, an approved drug for the treatment of lymphoma. Furthermore, Iratumumab and XmAb 2513 both target TNFRSF8 and are currently undergoing clinical trials for lymphoma. A COPD diagnostic model was developed using a machine learning approach (LightGBM) based on four plasma proteins (CXCL9, RNASET2, HLA-DRA, and TNFRSF8) supported by MR analysis. In the test set, the median AUC of the diagnostic models constructed from these four proteins reached 0.77 (95% CI: 0.73–0.82), indicating good diagnostic performance. Figure 6 ).

[0115] III. Discussion

[0116] In this large-scale prospective cohort study based on UKB, we identified 468 plasma proteins significantly associated with the pathogenesis of COPD. Among these, 386 proteins were also strongly associated with key lung function parameters. Furthermore, biological enrichment analysis highlighted the crucial role of immune and inflammatory responses in the pathogenesis of COPD, with EGFR and TNF identified as central hub proteins in the PPI network. Single-cell enrichment analysis showed that genes encoding COPD-related proteins were significantly enriched in epithelial cells, endothelial cells, macrophages, and T cells. MR analysis revealed CXCL9, HLA-DRA, TNFRSF8, and RNASET2 as target proteins of COPD, with an AUC of 0.77 for a COPD diagnostic model constructed using these four proteins. In conclusion, this study provides new insights into the molecular landscape of COPD, highlighting the potential of proteomic biomarkers in risk prediction, diagnosis, and personalized treatment strategies, while also suggesting that dysregulation of immune and inflammatory signaling is associated with COPD progression.

[0117] Using the UKB cohort, 468 plasma proteins were found to be significantly associated with COPD incidence, including 391 positively associated with COPD risk and 77 negatively associated with COPD risk. In this study, 83.5% of COPD-related proteins were positively associated with COPD incidence, a finding consistent with a previous cross-sectional proteomics analysis in which approximately 73.7% of COPD-related proteins were positively associated with COPD status. The higher proportion of positively associated proteins observed in prospective analyses may be related to differences in study design, statistical analysis methods, or protein detection methods. Cross-sectional studies typically collect proteomics data after clinical diagnosis, and their measurements may be affected by treatment interventions or disease progression. In contrast, the prospective cohort study of this invention used a Cox regression model to identify proteins associated with new disease risk, thereby detecting early systemic changes that may reflect causality or precursor biological processes. Given that COPD patients typically have high levels of inflammatory markers, the predominance of positively associated proteins may suggest their important role in the inflammatory response during COPD development. These results reinforce the crucial role of systemic inflammation in the initiation and progression of COPD. Furthermore, 90.6% of COPD-related proteins were significantly associated with at least one lung function parameter, indicating their close relationship with respiratory physiology. Multiple sensitivity and external validation analyses confirmed most of the COPD-related proteins identified in the main analysis, further enhancing the robustness and reproducibility of our findings.

[0118] Understanding the multifactorial nature of COPD, including genetic susceptibility, modifiable risk factors, and environmental exposures such as air pollution, is crucial for elucidating the underlying pathogenesis of COPD. The development of COPD is considered a result of disruption of a complex biological network involving genes and proteins. In our analysis, COPD-related proteins were significantly associated with the plasma reactivity disorder (PRS) of COPD, suggesting that overall genetic susceptibility to COPD may influence the expression of certain COPD-related proteins. Furthermore, COPD-related proteins showed a significant genetic association with COPD, suggesting a shared genetic structure. This dual evidence suggests that some plasma proteins may act as important molecular intermediates linking genetic risk to COPD development, highlighting the critical role of genetic factors in COPD pathogenesis. In addition, this study found that modifiable COPD risk factors, including alcohol consumption, unhealthy diet, lack of exercise, smoking, and underweight, were broadly associated with COPD-related proteins. These factors have been shown to promote systemic inflammation and oxidative stress, both of which are crucial in COPD pathogenesis. Unhealthy diets have been reported to promote COPD through pro-oxidative and pro-inflammatory mechanisms involving NF-κB activation. The findings of this invention suggest that these risk factors may influence COPD-related proteomic alterations and potentially affect inflammatory pathways associated with COPD progression. While long-term exposure to air pollution is associated with increased COPD risk, particularly in individuals with high genetic susceptibility and unhealthy lifestyles, proteomic analysis revealed that common air pollutants, including PM2.5, PM10, NOx, and NO2, are associated with only a limited number of COPD-related proteins. This relatively limited number of significant associations may indicate that chronic air pollution exerts subtle or localized biological effects that may not be adequately captured in systemic circulation. Alternatively, responses induced by key pollutants may occur in the airways or lung tissue, rather than in the plasma. These results underscore the need for future research that integrates tissue-specific and time-resolved omics data to fully elucidate the local or global biological effects of air pollution on COPD.

[0119] To gain a broader understanding of the potential biological functions of COPD-related proteins, this invention employed a comprehensive analysis, including a series of enrichment analyses, PPI network mapping, and TF network analysis. Pathway and phenotypic enrichment analyses revealed that these COPD-related proteins were primarily enriched in immune-related signaling pathways, suggesting an immune-mediated mechanism in COPD pathogenesis. Previous studies have confirmed that both innate and adaptive immune cells play crucial roles in COPD progression. Consistent with these findings, single-cell analysis showed significant enrichment of macrophages, monocytes, T cells, and NK cells. Macrophages and monocytes are the main immune cells driving lung inflammation in COPD. Monocytes are recruited to the lungs by chemokines and subsequently differentiate into dysfunctional macrophages, thereby reducing the clearance of pathogenic bacteria and promoting COPD progression. T lymphocytes, especially cytotoxic CD8 cells, are abundant in the airways of COPD patients. + A significant increase in T cells, with their increased abundance correlated with decreased lung function, suggests a crucial role in lung inflammation and alveolar destruction. Increased NK cells in the lungs of COPD patients have been reported, exhibiting enhanced cytotoxicity against autologous lung epithelial cells, which is associated with increased disease severity. Furthermore, significant enrichment of structural cells, including epithelial and endothelial cells, has been observed. Airway epithelial cells, as the first line of defense in the respiratory tract, regulate host immunity, inflammation, and remodeling. However, in COPD, they exhibit impaired repair capacity, excessive mucus secretion, and increased oxidative stress, leading to persistent inflammation, airway wall thickening, and airflow limitation. Endothelial dysfunction and aging are key factors in the pathogenesis of COPD. As regulators of immune cell migration, vascular permeability, and remodeling, lung endothelial cells play an active role in maintaining lung homeostasis. Their dysfunction leads to persistent inflammation, barrier disruption, and vascular remodeling, which are associated with airflow limitation and pulmonary hypertension. Recent evidence further suggests that endothelial-specific loss or inhibition of PRMT1 exacerbates alveolar destruction and aging in COPD models through excessive NF-κB activation. The findings of this invention highlight the crucial role of immune and inflammatory processes in COPD and support the biological relevance of the protein enrichment observed in structural and immune cells, further emphasizing that COPD is caused by complex and dysregulated interactions among multiple cell types.

[0120] Proteins are fundamental building blocks of cellular systems, and their functional interactions establish an integrated network structure crucial for revealing the complexity of biological phenomena. In the PPI network constructed from COPD-related proteins, EGFR and TNF appear as core hub nodes, indicating their potential key regulatory roles in COPD-related biological functions and signaling pathways. Previous evidence shows that aberrant activation of EGFR signaling contributes to epithelial barrier dysfunction and airway remodeling, characteristic manifestations of COPD. Mechanistically, EGFR activation has been shown to promote inflammatory responses through downstream pathways, particularly the NF-κB signaling cascade. However, the role of EGFR in the pathogenesis of COPD remains controversial. It has been reported that weakened EGFR signaling induces alveolar cell apoptosis, impairs alveolar regeneration, and exacerbates lung function decline. Future research is needed to elucidate the role of EGFR signaling in COPD. Furthermore, TNF has been identified as the second most important node in the PPI network of this invention. Multiple studies have shown that the pro-inflammatory cytokine TNF-α is involved in airway inflammation, tissue remodeling, and the development of emphysema. It also contributes to systemic inflammation and disease progression by activating the NF-κB signaling pathway. Although the pathogenic role of TNF-α has been well-established, clinical trials targeting this receptor have shown limited efficacy in COPD patients. This reflects the complexity of COPD and underscores the necessity of developing comprehensive or multi-target therapeutic strategies. Transcriptional factors (TFs) are considered major regulators of gene transcription. Studies have found that two transcription factors constituting the classical NF-κB signaling complex, NFKB1 and RELA, are associated with COPD-related proteins. Our findings are consistent with previous evidence confirming the crucial role of NF-κB signaling in COPD pathogenesis and further support the involvement of inflammation-related transcriptional regulation in disease progression.

[0121] MR analysis revealed a causal relationship between four candidate proteins (TNFRSF8, CXCL9, HLA-DRA, and RNASET2) and COPD, with TNFRSF8 showing strong evidence of colocalization. TNFRSF8, also known as CD30, is a member of the tumor necrosis factor receptor superfamily and is involved in the activation of the NF-κB and mitogen-activated protein kinase pathways, which are central to the regulation of inflammation and immune responses. This study, using cohort studies, MR analysis, and colocalization, provides strong evidence for a causal relationship between elevated plasma TNFRSF8 levels and increased COPD risk. Furthermore, in COPD model rats, treatment with anti-CD30 antibodies reduced TNFRSF8 levels and alleviated pulmonary vascular remodeling, highlighting the potential of TNFRSF8 as a therapeutic target for COPD. In terms of clinical relevance, TNFRSF8 targets Brentuximab vedotin, an antibody-drug conjugate approved for the treatment of lymphoma, as well as Iratumumab and XmAb 2513, two investigational drugs currently under clinical evaluation. Given the elevated expression of TNFRSF8 in COPD, the potential for repurposing these drugs in COPD treatment warrants further investigation. CXCL9, an interferon (IFN)-γ-induced inflammatory chemokine, plays a crucial role in immune activation and inflammatory responses. Its effects have been extensively documented in chronic inflammatory diseases and various cancers, with multiple case-control studies consistently demonstrating significantly elevated serum CXCL9 levels in COPD patients compared to healthy controls. However, these studies have limitations in cross-sectional design and small sample sizes. To further elucidate this phenomenon, this study, through a large-scale cohort study and multivariate regression analysis, found a significant association between elevated plasma CXCL9 levels and increased COPD risk. CXCL9 is primarily produced by alveolar macrophages and bronchial epithelial cells, and it promotes inflammatory cell recruitment by binding to the CXCR3 receptor, highlighting its key role in COPD-related immune responses. These results suggest that CXCL9 is a promising therapeutic target for COPD. HLA-DRA, a subunit of HLA-DR, is crucial for immune function through its role in antigen presentation. A comparative observational study found that HLA-DR was present in COPD patients. + Myeloid-derived regulatory cells accumulate in the airways, promoting autologous CD4. +T cell proliferation and pro-inflammatory activity were observed. In mouse models mimicking asthma and COPD, transcriptomic analysis identified HLA-DRA as a key upstream regulator of inflammatory signaling. RNASET2, a highly conserved endonuclease, is involved in RNA metabolism, autophagy, and innate immune activation, playing a crucial role in maintaining immune homeostasis. Although evidence directly linking plasma HLA-DRA and RNASET2 to COPD is lacking, their involvement in immune and inflammatory processes suggests a potential role in the pathogenesis of COPD.

[0122] This study has several significant advantages. First, it provides a comprehensive analysis of a large number of proteomic biomarkers from a large prospective cohort to examine protein alterations prior to COPD onset, thereby deepening our understanding of potential molecular changes in COPD pathogenesis. Multiple sensitivity analyses were also performed to enhance the reliability of the results. Second, given the complexity of COPD, this study integrates genetic, environmental, and risk-related factors with biological processes to elucidate its underlying mechanisms and capture phenotypic variations across different populations. Third, the robustness of the results was validated through cross-platform comparisons of the UKB Olink dataset and the Icelandic SomaScan dataset. The comparisons showed a moderate correlation between the two datasets in terms of cross-sectional association. Furthermore, the high consistency of the findings across the UKB and EPIC cohorts, two large prospective studies, further strengthens the external generalizability and reliability of proteomics research results.

[0123] However, this study still has some limitations. First, the study primarily involved European populations, and considering that different genetic backgrounds may lead to differences in protein expression, this may limit the generalizability of the findings to other populations. Second, due to the lack of prospective large-scale cohort data including proteomics features of lung-related tissues or body fluids, it is difficult to directly generalize plasma-based research conclusions to target organs. Nevertheless, plasma proteomics remains valuable for early disease feature identification due to its ease of access. Finally, although the biological mechanisms of the candidate proteins were investigated, their therapeutic efficacy and safety need to be validated in future clinical trials.

[0124] In summary, this invention, by integrating observational analysis and genetic methods from large-scale proteomics studies, has created a comprehensive atlas of plasma protein changes prior to the onset of COPD. The study identified 468 proteins significantly associated with COPD, including key proteins such as TNFRSF8, CXCL9, HLA-DRA, RNASET2, EGFR, and TNF, which may play a central role in the pathophysiological process of the disease. This research not only highlights the crucial value of longitudinal proteomics analysis in revealing early biological changes in COPD but also opens new avenues for the development of personalized treatment strategies. Given the complex functions of proteins in regulating biological processes, the findings of this invention provide an important framework for understanding the pathological mechanisms of COPD and lay the foundation for future development of proteomics-based therapeutic strategies.

[0125] While the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the scope of the invention. Any person skilled in the art can make modifications without departing from the scope of the invention; all equivalent modifications made in accordance with the invention should be covered by the scope of the invention.

Claims

1. COPD target proteins based on plasma proteomics, characterized in that, The target proteins include one or more of the following four proteins: HLA-DRA, CXCL9, TNFRSF8, and RNASET2.

2. An application of COPD target proteins based on plasma proteomics as described in claim 1, characterized in that, The target protein is used in the preparation of drugs for treating COPD.

3. A drug for treating COPD, characterized in that, The drug targets one or more of the following four proteins: HLA-DRA, CXCL9, TNFRSF8, and RNASET2.

Images