Heart failure (HF) patients often develop pulmonary hypertension (PH) that leads to transition into a biventricular HF with poor prognosis. There are two PH components: 1) passive transmission of increased left atrial pressure, 2) heart failure (HF) related pulmonary vascular dysfunction (PVD) with increased vascular resistance. Intriguingly, only some, but not all HF patients develop heart failure-related PVD. The mechanisms and non-invasive detection of HF-PVD are poorly understood and are the focus of the current grant application. Development of PVD is linked to insufficiently characterized metabolic factors that may be mediators of HF-PVD. Untargeted metabolomics is an emerging powerful platform for the discovery of pathways linked to diseases. Its specificity can be further enhanced using transpulmonary gradient sampling. Part A of the project aims to identify novel metabolites associated with the presence of PVD in patients with HF that can serve as biomarkers or targets and will provide biologic insights into PVD. Part C will assess the effects of reverting of metabolic alterations (identified in part A) by a drug/diet on pulmonary vasculature in experimental HF-related PVD. The gold standard for the detection of PVD is right heart catheterization, which is invasive and risky. Heart failure-related PVD is therefore often diagnosed late. There is a need for noninvasive tests that may help to detect PVD in early stages and can be done repeatedly. Recent advances in artificial intelligence (AI)-assisted automated quantitative analysis of lung texture from low-dose contrast-free high-resolution CT images allow to quantify lung water content, interstitial changes or vessel volume, and may provide clues for detection of heart failure-related PVD. Such an approach, not tested yet, will be utilized for the detection of HF-PVD (part B). Clinical and functional characteristics of lung circulation (exercise hemodynamics, diffusion capacity, perfusion) will be analyzed in relation to quantitative CT data.