From Blood to Vessel: Lipid Ratios in Pulmonary Hypertension

Circulating metabolites represent the integrated effects of gene expression, enzymatic activity, and environmental influences across different tissues. Compared with gene transcripts and proteins, they are considered more closely related to the extant pathology, making them particularly valuable as biomarkers for assessing disease risk and gaining insights into pathobiological mechanisms (1, 2). A growing body of evidence has demonstrated that pulmonary arterial hypertension (PAH) is characterized by aberrant metabolism, particularly altered bioenergetics. In addition to a glycolytic shift, abnormalities in fatty acid (FA) metabolism have been documented in the affected pulmonary vasculature and failing right ventricle (RV).

In small pulmonary arteries (PAs) from patients with PAH, gene transcripts related to FA uptake and processing have been reported to be upregulated (3). FA oxidation has been implicated in the development of preclinical models of pulmonary hypertension (PH) and pulmonary vascular remodeling (3–5). The potential relevance of FAs to PAH is further supported by findings from blood-based studies, which have shown increased circulating levels of free FAs (FFAs) in patients with PAH compared with healthy controls (HCs), although not all studies have reported consistent results (6–8). Notably, the directionality of associations between FA, FA derivates, and other complex lipids with PAH was divergent, with some lipid species increased and others decreased (6, 8). These opposing trends support the use of composite biomarkers, potentially enhancing diagnostic sensitivity and specificity. For instance, whereas circulating levels of FFAs and long-chain acylcarnitines (FA derivatives conjugated to carnitine for mitochondrial transport) were found to be increased in PAH, levels of phosphatidylcholine species (complex lipids that are major source of endothelial arachidonic acid) were reduced compared with HCs (6, 7, 9).

In addition to the pulmonary vasculature, changes in FA metabolism have also been implicated in the adaptation of the RV under conditions of chronic pressure overload (4, 10). Increased FA levels have been observed in RV tissue from patients with PAH compared with HCs and diseased controls (DCs) with dilated cardiomyopathy (6). Interestingly, transcriptomic analysis suggests that mitochondrial FA transport enzymes may not be the primary drivers of FA accumulation in the RV, pointing instead to alterations in downstream effects such as FA oxidation (6, 10). The contribution of the RV to circulating FAs was supported by metabolite gradients measured between the vena cava and PA, with the most pronounced differences observed for lipids (11).

In this issue of the Journal, Bordag and colleagues (pp. 1264–1276) leveraged the metabolic dysregulation found in PH to identify blood-based metabolic markers for the detection of the condition (12). The authors quantified 164 known metabolites in two independent cohorts of patients with various forms of PH alongside HCs and DCs in whom significant PH had been excluded by noninvasive clinical tests but who had metabolic syndrome, chronic obstructive pulmonary disease, or interstitial lung disease. They found differences between the metabolomes of patients with PH compared with the HC and DC groups, which were driven by specific FAs. Although machine learning tools were initially employed in their data analysis, simple ratios of FFAs to complex lipids outperformed machine learning models. These ratios yielded high area under the curve (AUC) values greater than 0.8 in the validation cohort when comparing patients with PH versus HCs, offering a more interpretable alternative to complex statistical models. Diagnostic performance depended on FA structure, with ratios containing monounsaturated FAs outperforming those with other degrees of saturation, regardless of chain length. Three top-performing ratios, each comprising between six and eight different metabolites, demonstrated AUCs greater than 0.89 in the validation set. These ratios were associated with patient survival in the overall cohort, suggesting potential prognostic value in addition to their diagnostic utility.

To further assess the biological role of increased circulating FFAs, the authors examined PA sections from patients with idiopathic PAH and HCs, demonstrating lipid accumulation within the pulmonary vasculature of patients. Finally, to mimic increased circulating FFA levels, the authors exposed primary human pulmonary artery endothelial cells and smooth muscle cells from healthy donors to a defined FFA mix. This led to impaired endothelial function with reduced nitric oxide production and loss of barrier integrity, as well as increased proliferation in smooth muscle cells.

Bordag and colleagues are to be commended for investigating the biological effects of lipids on pulmonary vascular cells as well as highlighting the utility of lipid profiling in blood samples to detect PH (Figure 1). However, there are also limitations to the study that should be considered, including the relatively small sample size, the unbalanced case–control ratio, and the absence of DCs in the independent validation cohort. The impact of the control group on marker performance has been shown previously. Discriminant models based on combination of seven metabolites (some increased and others reduced in PAH) demonstrated strong diagnostic performance when comparing patients with PAH versus HCs (AUCs of 0.95 and 0.94 in two validation cohorts) but more modest accuracy when DCs were used as the comparator (AUCs of 0.72 and 0.75) (8). In addition, although the inclusion of multiple PH subtypes may strengthen the identification of metabolic markers with diagnostic potential independent of the underlying disease etiology, between-group differences and the specific biological roles of these markers also warrant further investigation.

Figure 1.

Bordag and colleagues investigated the diagnostic potential of lipid ratios in patients with different forms of pulmonary hypertension compared with healthy and diseased controls. They identified high diagnostic accuracy for ratios of free fatty acids to complex lipids and demonstrated direct effects of free fatty acids on pulmonary vascular cells. Figure includes elements from Servier Medical Art, licensed under CC BY 4.0. FFA = free fatty acid; IPAH = idiopathic pulmonary arterial hypertension; PH = pulmonary hypertension; SMC = smooth muscle cell.

Conceptually, their findings raise two interdependent scenarios: 1) impaired FA metabolism in the RV under chronic overload leads to increased circulating FA levels, which in turn exert harmful effects on the pulmonary vasculature; and 2) disturbances in lipid metabolism within the small PAs may drive vascular dysfunction and increased RV afterload, creating a vicious circle. It should also be noted that some high-performing ratios (including the three top performing ones) contained nonendogenous odd-chained FAs, suggesting a possible microbiome origin, consistent with studies linking gut dysbiosis to PAH (13). Finally, there is increased skeletal muscle uptake and oxidation of FFAs during physical activity, and circulating FFA levels increase only when lipolysis exceeds uptake, which is typically during prolonged exercise (14). Differences in physical activity could contribute to variability in circulating FFAs.

In summary, the identification by Bordag and colleagues of interpretable lipid ratios as a biomarker in the context of PH is novel and may accelerate future biomarker discovery, but further validation in a clinical setting is needed. Although informative, lipid ratios may not be as practical a measurement as other accessible risk factors, including other potential circulating biomarkers such as proteins. The demonstrated direct effects of FFAs on pulmonary vascular cells open exciting new avenues for research and underscore the relevance of FA metabolism as a potential therapeutic target. These findings also advance our understanding of the cross-talk between the pulmonary vasculature and the RV under chronic pressure overload. Incorporating measures of RV function including load-independent metrics will be important for future studies.