Plasma Proteomics Linking Primary and Secondary diseases: Insights into Molecular Mediation from UK Biobank Data

hanyu qian, Chao Wu, Bo Li, Anthony Rosenzweig, Meng WangmedRxiv 2025 · DOI

Abstract

Diabetes, hypertension, and dyslipidemia are well-established risk factors for cardiovascular (CVD), cerebral, and renal diseases (RD). However, the molecular mechanisms linking these primary conditions to downstream secondary diseases remain incompletely understood. Using data from approximately 50,000 UK Biobank participants, this study investigates the role of plasma proteins as mediators of disease progression. Across three primary diseases and 18 subsequent conditions, we identified 1,461 significant mediation pathways involving 395 unique plasma proteins. Notable mediators include GDF15, which consistently links diabetes to CVD, and ADM, which mediates the pathway between hypertension and pulmonary disease. These proteins are highly enriched in immune, metabolic, and cytokine-related pathways. Mendelian randomization (MR) supported causal roles for 84 proteins, highlighting their potential as therapeutic targets. Furthermore, machine learning models incorporating these mediating proteins significantly improved predictive accuracy for secondary disease risk. For instance, in hypertensive individuals, top mediating proteins improved the C-index for glomerular disease risk by approximately 14%. Stratified analyses also revealed severity-dependent pathways, such as APOE mediating the link between severe diabetes and Alzheimer’s disease. These findings position plasma proteins as central molecular bridges and nominate promising targets for biomarker development and therapeutic intervention.

Introduction

While the clinical association between metabolic disorders—such as diabetes, hypertension, and dyslipidemia—and secondary complications like heart failure or kidney disease is well-documented, the biological "black box" connecting them has remained elusive. Traditional clinical risk factors like age, BMI, and blood pressure provide a baseline for risk assessment but often fail to capture the specific molecular shifts that drive an individual from a primary diagnosis to a secondary complication.

This study leverages large-scale proteomic data to map the molecular architecture of disease progression. By identifying specific plasma proteins that act as mediators, we can gain mechanistic insight into biological paths (e.g., systemic inflammation and oxidative stress) and identify actionable biomarkers for early risk stratification and drug development.

Methodology

The researchers utilized a multi-layered computational framework to analyze data from the UK Biobank (UKB), focusing on approximately 53,030 participants.

Data Sources

  • Proteomics: Plasma protein levels were quantified using the Olink Explore 3072 platform, covering 2,923 unique proteins.
  • Clinical Data: Longitudinal health records were used to track 18 secondary diseases across cardiovascular, renal, pulmonary, and neurological domains.
  • Genetics: GWAS summary statistics from the IEU OpenGWAS database supported Mendelian Randomization (MR) analyses.

Analytical Approach

  1. Mediation Analysis: Using Cox proportional hazards models, the team tested if the effect of a primary disease on a secondary outcome was mediated through protein expression. A Proportional Indirect Effect (PIE) of at least 10% was required for significance.
  2. Causality Assessment: MR was employed to determine if the identified mediators had a causal impact on secondary diseases, using genetic variants (cis-pQTLs) as instrumental variables.
  3. Predictive Modeling: LASSO and XGBoost survival models compared the predictive power of mediating proteins against traditional clinical factors.
  4. Severity Stratification: Participants were stratified by continuous biomarkers (e.g., HbA1c for diabetes) to identify non-linear "acceleration points" where risk increases sharply.

Key Insights and Results

The Landscape of Molecular Mediation

The study identified 1,461 significant mediation pathways involving 395 unique proteins. These mediators were predominantly enriched in pathways related to immune response, metabolic regulation, and cytokine activity, suggesting that systemic inflammation is a central driver of multi-organ decay.

  • Universal vs. Specific Mediators: Certain proteins emerged as "super-mediators." GDF15 was found to consistently bridge diabetes with various cardiovascular and renal complications. In contrast, ADM (Adrenomedullin) was specifically linked to the hypertension-pulmonary disease pathway.
  • Causal Evidence: MR analysis supported causal roles for 84 proteins, providing a high-confidence list of potential therapeutic targets to halt disease progression.

Enhancing Clinical Prediction

One of the study's most significant findings is the clinical utility of these proteins in risk forecasting. Machine learning models demonstrated that "biologically informed" feature selection—using only the top 10 mediating proteins—often outperformed models using the entire 3,000-protein panel.

In participants with hypertension, adding the top mediating proteins to traditional risk factors improved the prediction accuracy (C-index) for glomerular disease by 14%. This suggests that these proteins capture critical physiological information that standard metrics miss.

Severity-Dependent Pathways

Disease progression is not always linear. Stratified analyses revealed that some mediation paths only activate at specific levels of severity. For example, APOE was identified as a mediator specifically linking severe diabetes (indicated by high HbA1c) to Alzheimer’s disease. The study further identified specific "acceleration points" for Alzheimer’s risk at HbA1c levels of 5.7% and 7.9%, offering concrete targets for clinical glucose management.

Discussion and Limitations

This research provides a comprehensive map of how the proteome mediates the transition from primary metabolic stress to secondary organ damage. By identifying 84 proteins with causal evidence, the study moves beyond correlation toward actionable biology.

However, several limitations remain. The UK Biobank cohort is predominantly of European descent, necessitating validation in more diverse populations to ensure global applicability. Additionally, while the study focused on individual proteins, future research should explore protein-protein interactions and joint mediation effects to better reflect the complexity of human biology. Finally, the reliance on linear mediation models may oversimplify certain non-linear biological processes, and low prevalence rates for some secondary diseases (as low as 0.2%) suggest that those specific findings require further validation in larger datasets.

Conclusion

By mapping the proteomic bridges between primary and secondary diseases, this study identifies high-priority biomarkers and therapeutic targets. These insights pave the way for more precise risk stratification and the development of interventions designed to interrupt the molecular pathways of disease progression.

Summary updated

Abstract

Diabetes, hypertension, and dyslipidemia are major risk factors for cardiovascular (CVD), cerebral, and renal diseases (RD). However, the underlying molecular mechanisms, particularly the biological paths linking these primary conditions to downstream diseases remain incompletely understood. In this study, we investigated the role of plasma proteins as mediators of secondary disease development using data from ~50,000 UK Biobank participants. Across three primary diseases and 18 subsequent conditions, we identified 1,461 significant mediation pathways involving 395 unique plasma proteins. Notable examples included GDF15 consistently mediating the diabetes-CVD link and ADM mediating the hypertension-pulmonary disease pathway. Protein mediators of secondary disease development were highly enriched in immune, metabolic, and cytokine-related pathways. Mendelian randomization supported causal roles for 84 proteins, highlighting their potential as therapeutic targets. Moreover, the identified mediating proteins improved predictive accuracy for secondary disease risk compared to other proteins and traditional clinical risk factors from machine learning methods. For example, in individuals with hypertension, the inclusion of top mediating proteins improved prediction accuracy for glomerular disease risk by approximately 14% measured by C-index. Stratified analyses based on disease severity revealed additional disease progression pathways and mediating proteins, such as APOE mediating the association between severe diabetes and Alzheimer's disease. Together, these findings implicate plasma proteins as central molecular mediators linking these primary diseases to subsequent development of a secondary condition and nominate promising targets for biomarker development and therapeutic intervention.
BibTeX
@misc{qian2025plasma,
  author = {hanyu qian and Chao Wu and Bo Li and Anthony Rosenzweig and Meng Wang},
  title = {{Plasma Proteomics Linking Primary and Secondary diseases: Insights into Molecular Mediation from UK Biobank Data}},
  year = {2025},
  doi = {10.1101/2025.08.29.25334726},
  howpublished = {medRxiv},
}
Figure