Abstract
Acute mountain sickness (AMS) is frequently experienced by non-high-altitude natives at high altitudes, affecting health and productivity. Diagnosis currently relies on subjective self-questionnaires, highlighting a need for reliable, objective biomarkers. This study profiled plasma proteomic changes in a longitudinal cohort of 53 individuals using proximity extension assays (PEA) and multiple reaction monitoring (MRM). By integrating 65 clinical indexes and 22 symptom phenotypes, researchers identified biomarkers for AMS diagnosis, prognosis, protection, and pathogenesis. Key findings include the identification of RET and ADAM15 as potential therapeutic targets, and PHGDH and TRAF2 as candidate predictive and diagnostic biomarkers. Notably, higher baseline gluconeogenesis activity at sea level was found to predict AMS susceptibility. These results provide biological insights into AMS and promote the development of precision medicine for high-altitude conditions.
Introduction
Acute Mountain Sickness (AMS) is a common condition affecting individuals who ascend rapidly to altitudes above 2,500 meters. Symptoms range from debilitating headaches and nausea to life-threatening edema. Currently, the "gold standard" for diagnosis is the Lake Louise Score (LLS), a subjective self-questionnaire. This subjectivity complicates precise medical care and prevents the identification of high-risk individuals before ascent. This research aims to establish objective molecular markers to improve diagnosis and predict susceptibility.
Methods and Study Design
The researchers conducted a two-stage longitudinal study following 53 Chinese Han male subjects as they traveled from 1,200m (plain) to 4,300m (high altitude).
- Discovery Stage: Proximity Extension Assay (PEA) technology was used to screen 1,069 proteins in a subset of participants to identify differentially expressed proteins (DEPs).
- Validation Stage: Multiple Reaction Monitoring (MRM) validated 102 key proteins across the entire cohort (30 AMS patients, 23 non-AMS controls).
- Machine Learning: Four XGBoost models were developed to classify subjects based on pathogenesis, protection (resilience), prediction (pre-ascent), and diagnosis (at altitude).
Key Insights and Results
Biomarkers for Clinical Scenarios
The XGBoost models demonstrated high accuracy in identifying individuals across four critical clinical contexts:
- Pathogenesis (AUC 0.90): Identified a 7-protein panel, with RET as a top-weighted variable. RET is linked to pain-sensing neurons, potentially explaining AMS headaches.
- Protection (AUC 0.95): Found that individuals who remained healthy (nAMS) significantly downregulated ADAM15 at high altitude, suggesting a protective role for suppressing certain inflammatory pathways.
- Prediction (AUC 0.91): Identified PHGDH as a critical biomarker at sea level that can predict who will fall ill before they ascend.
- Diagnosis (AUC 0.89): Utilized a 13-protein panel featuring TRAF2 and AGT to objectively diagnose AMS at altitude.
Metabolic Pre-conditioning and Networks
A major finding was that individuals susceptible to AMS already exhibited higher gluconeogenesis activity while still at sea level. This suggests that a person's baseline metabolic state is a primary determinant of altitude tolerance. Furthermore, protein-symptom-clinical index functional network analysis revealed how these molecular markers correlate with specific clinical phenotypes.
Discussion and Future Directions
While these findings offer a wealth of biological insights, the study has limitations. The cohort was relatively small and limited to young Han Chinese males; future research must validate these biomarkers in larger, more diverse populations. Additionally, functional in vivo experiments are needed to confirm the causal roles of proteins like RET and ADAM15. Ultimately, integrating these molecular "fingerprints" with traditional clinical scores could revolutionize the management of AMS and move the field toward true precision medicine for high-altitude health.
Summary updated