Impaired Lung BCAA Metabolism Promotes Ferroptosis and Resultant Pulmonary Arterial Hypertension-Associated Hepatopathy Authors: Madelyn J Blake, Jason Hong, Adam Brownstein, Christopher J Rhodes, Emilia M Swietlik, Jeffrey C Blake, Ryan A Moon, Lynn M Hartweck, Jenna B Mendelson, Sasha Z Prisco, Todd Markowski, LeeAnn Higgins, Kevin Murray, Candance Guerrero, Sandra Breuils-Bonnet, Steeve Provencher, Joanna Pepke-Zaba, Mark Toshner, Martin Wilkins, Sebastien Bonnet, Kurt W Prins Published: 2026 ABSTRACT Visual AbstractKey Words: branched-chain amino acids, ferroptosis, hepatopathy, pulmonary arterial hypertension, right ventricular failure Valine, leucine, and isoleucine are the 3 branched-chain amino acids (BCAAs), and dysregulation of BCAA homeostasis is implicated in multiple cardiovascular diseases.1 Emerging data suggest that impaired BCAA metabolism contributes to pulmonary arterial hypertension (PAH) via effects on pulmonary artery smooth muscle cell (PASMC) biology.2 In particular, branched-chain ketoacids, downstream metabolites of BCAAs, aerobically activate hypoxia-inducible factor 1 and induce PAH-like phenotypes in PASMC in vitro.2 However, the therapeutic relevance of modulating BCAA metabolism in PAH is undefined. Alterations in BCAA catabolism are linked to ferroptosis, a type of cell death triggered by lipid peroxidation caused by mitochondrial dysfunction.3 Inhibition of branched-chain amino acid aminotransferase 2, an enzyme required for BCAA catabolism, triggers ferroptosis in cancer cells.4 In cerebral ischemia-reperfusion, activation of BCAA oxidation with the small molecule BT2 alleviates neuronal ferroptosis and improves neurological recovery after injury.5 In the cisplatin mouse model of chronic kidney disease, up-regulation of microRNA 429-3P suppresses BCAA catabolism and promotes renal ferroptosis, while BT2-induced BCAA oxidation rescues this phenotype.6 We7 and others8, 9, 10 have demonstrated that ferroptosis promotes PAH through metabolic and inflammatory mechanisms, but whether modulating BCAA homeostasis alters PAH severity and associated pulmonary vasculature ferroptosis phenotypes is unknown. In addition to BCAAs, other amino acids modulate ferroptosis induction. Perhaps the best studied amino acid pathway in ferroptosis is the transmembrane system xc− pathway,11 an amino acid antiporter that exports glutamate and imports cystine, which is then converted to cysteine. Intracellular cysteine is used to synthesize glutathione, an antioxidant that maintains the antiferroptotic protein glutathione peroxidase 4.11 The trans-sulfuration pathway also modulates ferroptosis,12 as it facilitates the conversion of methionine to homocysteine, and then cysteine, which feeds back to regulate glutathione peroxidase 4.11 Moreover, recent data show arginine has a proferroptotic function through its downstream metabolites, ornithine and polyamines.13 At present, the interplay among BCAAs, system xc−, trans-sulfuration, and arginine metabolites in PAH lungs is not well defined. Disrupted hepatic function is a risk factor for mortality in PAH14 because slight alterations in hepatic function tests are highly predictive of survival.15 Furthermore, heightened liver stiffness is observed in PAH, and liver stiffness correlates with PAH severity.16 Currently, the mechanisms underlying hepatic dysfunction in PAH are poorly understood. Disrupted and turbulent venous flow caused by right heart failure is a likely contributor14; however, a detailed evaluation of altered cellular mechanical stress phenotypes caused by disturbed venous flow in PAH is lacking. Two classic and repeatedly identified cellular phenotypes of cells exposed to excess mechanical stress/shear stress are expansion in nuclear size17,18 and disrupted mitochondrial regulation.17 Thus, these phenotypes may serve as readouts for mechanical stress, but the impact of PAH on liver nuclear morphology and mitochondrial function is not well described. Methods A detailed methods section is available in the Supplemental Appendix. To evaluate the effects of small-molecule-induced BCAA catabolism in the monocrotaline (MCT) PAH model, male Sprague-Dawley rats (7-8 weeks old, Charles River) were randomly assigned to 3 groups: 1) controls receiving a single subcutaneous injection of phosphate buffered saline (n = 10); 2) MCT-injected rats (60 mg/kg) treated daily with intraperitoneal vehicle (2% dimethyl sulfoxide, 50% polyethylene glycol, 5% Tween 80, and 43% double distilled water) starting 14 days post-MCT injection (n = 20); and 3) MCT-injected rats treated daily with intraperitoneal BT2 (20 mg/kg, MedChemExpress) starting 14 days post-MCT injection (n = 15). A low-BCAA dietary intervention was also performed in MCT-injected male Sprague-Dawley rats (7-8 weeks old, Charles River). Rodents were randomized to the following: 1) standard chow diet throughout (Teklad, 2918) (n = 10); or 2) standard chow diet (Teklad, 2918) for 14 days post-MCT, then a low BCAA chow (Envigo, TD.140714) for 10 days (n = 10). Endpoint analysis occurred 25 days post-MCT, with right ventricular systolic pressure (RVSP) as the primary physiological readout. Female rats were not evaluated because of the more robust PAH phenotype in male rats.7 Immunoblots analyzed mTORC1 signaling in rodent lung samples as well as KLF2 and PAkt/Akt in rodent liver samples. Confocal microscopy evaluated perivascular C3 abundance and transferrin receptor activation in rodent lung samples. Antibodies used and their respective dilutions are provided in Supplemental Table 1. Proteomics and metabolomics analyses of rodent lungs, serum, and liver and human liver specimens were carried out as described.7 Statistical analyses were performed using GraphPad Prism version 10.4.0 (GraphPad Software). The association between 2 continuous variables was determined using Pearson's correlation coefficient (r). For comparisons between 2 groups, normality was first assessed using the Shapiro-Wilk test. Normally distributed data were compared using an unpaired Student's t-test, while non-normally distributed data were compared using the Wilcoxon rank-sum test. For comparisons among 3 groups, if data were normally distributed and there was equal variance as determined using the Brown-Forsythe test, 1-way analysis of variance with Tukey's post hoc test for multiple pairwise comparisons was performed. If there was unequal variance, Welch's analysis of variance with Dunnett's post hoc was completed. If the data were not normally distributed, the Kruskal-Wallis test and Dunn's post hoc test were used when comparing more than 2 groups and the Wilcoxon rank-sum test for comparing 2 groups. Sex-stratified analyses were performed in the human bulk RNAseq and smooth muscle cell (SMC) snRNAseq data sets using pairwise 2-sided Wilcoxon rank-sum tests. Graphs display the mean or median value and all individual values. A P value < 0.05 was considered statistically significant. Hierarchical cluster analyses, sparse least discriminant analysis, random forest classification, and correlational heatmapping were performed using MetaboAnalyst (https://www.metaboanalyst.ca/).19 The relative abundance of each protein was determined using Proteome Discover. Transcripts and proteins that were significantly correlated (P < 0.05) with central venous pressures and right ventricular systolic pressure were processed for Kyoto Encyclopedia of Genes and Genomes and WikiPathways (Wiki) using ShinyGO 0.76.3 (http://bioinformatics.sdstate.edu/go/).20 Because human PAH tissue is rare and difficult to obtain, there was a small sample size for the liver proteomics analysis; thus, the Benjamini-Hochberg false discovery rate for pathway enrichment was increased to 0.1. Human liver tissue use adhered to the principles of the Declaration of Helsinki, and all experimental procedures were approved by the Institutional Review Boards at the Centre de recherche de l’Institut de cardiologie et de pneumologie de Québec (CRIUCPQ) ethics committee (CER #20773). Written informed consent was obtained from all participants. For the UK plasma metabolite gradient cohort, all participants provided written informed consent, and the study was approved by local research ethics committees (reference numbers 17/LO/0563 and 15/EE/0201). Lung transcriptomic data were obtained from the Pulmonary Hypertension Breakthrough Initiative, for which all participants provided informed consent and studies were approved by Institutional Review Boards at participating centers, as previously described. All animal studies were approved by the University of Minnesota Institutional Animal Care and Use Committee. Results Metabolomics analysis identified altered BCAA metabolites in human pulmonary vascular disease To determine if BCAA homeostasis was disrupted in patients with pulmonary vascular disease, we probed BCAAs and BCAA catabolite levels (Figure 1A) in arterial (ART), superior vena cava (SVC), and pulmonary artery (PA) blood samples and gradients across those sites from patients with PAH (n = 18) or chronic thromboembolic pulmonary hypertension (n = 68) (Figure 1B). We observed alterations in BCAAs as the isoleucine SVC-PA gradient and leucine PA-ART gradient were significantly increased in pulmonary hypertension patients (Figure 1C). In addition, further downstream metabolites in the BCAA pathway were modified as the SVC-PA gradients of 2-hydroxy-3-methylvalerate, 2-methylbutyrylcarnitine, and 3-hydroxy-2-ethylpropionate were altered in patients with pulmonary vascular disease (Figure 1C). Finally, the PA-ART gradients of 3-hydroxy-2-ethylpropionate and ethylmalonate were deranged in the diseased patients (Figure 1C). Human Metabolomics, Bulk RNAseq, and snRNAseq Data Linked Altered BCAA Catabolism and Ferroptosis to PVD (A) Schematic representation of branched-chain amino acid (BCAA) metabolism with associated metabolites. (B) Graphical depiction of gradient blood sampling for human metabolomics analysis. (C) Summary of BCAA-associated metabolite alterations at each individual site and gradients with associated P values in patients with pulmonary vascular disease (PVD). (D) Lung bulk RNAseq demonstrating alterations in BCAA catabolism and ferroptosis pathways in pulmonary arterial hypertension (PAH) lungs, with smooth muscle cell (SMC) deconvolution displaying a positive, but statistically insignificant, trend between SMC abundance and BCAA catabolism, and a significant negative correlation between SMC abundance and ferroptosis activity. P values determined by 2-sided Wilcoxon rank-sum test for group comparisons and Pearson correlation for associations. (E) Lung snRNAseq mirrors bulk RNAseq with PAH SMCs showing alterations in both BCAA catabolism and ferroptosis activity scores relative to controls. P values determined by 2-sided Wilcoxon rank-sum test. ∗∗Data in D and E are presented as individual data points with box-and-whisker plots representing the median and IQR. Correlation plots display individual data points with the line of best fit and 95% CI. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Bulk and snRNAseq data from human PAH samples demonstrated dysregulation of the BCAA catabolic pathway and ferroptosis in human SMC Then, we evaluated the BCAA and ferroptosis pathways in a lung bulk RNAseq analysis comparing 52 control samples to 96 PAH patients. There were significant reductions in both the BCAA catabolism and ferroptosis pathway scores in PAH lungs as compared with controls (Figure 1D). Then, we probed the relationship between BCAA score and deconvoluted SMC relative abundance and found a nonsignificant (P = 0.071) association between SMC abundance and BCAA score (Figure 1D). However, there was a highly significant and negative association between ferroptosis score and SMC abundance (r = −0.27; P < 0.001) (Figure 1D). A targeted analysis of BCAA and ferroptosis pathways in SMCs was performed in human lung snRNAseq data from 25 control and 42 PAH patients (Brownstein et al, manuscript in preparation). In agreement with the bulk RNAseq data, both the BCAA and ferroptosis scores were significantly lower in PAH SMC than controls (Figure 1E). When evaluating ferroptosis. Excess BCAAs promoted a proferroptotic phenotype in human PASMC in vitro To directly examine the link between BCAAs and ferroptosis in human PASMC, we evaluated how excess BCAAs modulated PASMC ferroptotic phenotypes in vitro. Heightened BCAAs induced mitochondrial fragmentation (Figure 2A), altered mitochondrial membrane potential (Figure 2B), and triggered mitochondrial reactive oxygen species generation (Figure 2C). Moreover, elevated BCAAs increased lipid peroxidation levels, the trigger for ferroptosis21 (Figure 2D). Finally, PASMCs treated with BCAAs exhibited heightened sensitivity to cell death, which was rescued by ferrostatin-1, a small molecule inhibitor of ferroptosis22 (Figure 2E), suggesting that excess BCAAs promoted a proferroptotic phenotype in human PASMC. Excess BCAAs Promoted a Proferroptotic Phenotype in Human Pulmonary Artery Smooth Muscle Cell (A) Representative confocal micrographs of pulmonary artery smooth muscle cells stained with MitoTracker Orange and quantification of mitochondrial organization in control and branched-chain amino acid (BCAA)–treated pulmonary artery smooth muscle cells (control pulmonary artery smooth muscle cells [Con]: 1.1 ± 0.6, BCAA-treated pulmonary artery smooth muscle cells [BCAA]: 0.7 ± 0.3); P values determined by Wilcoxon rank-sum test. (B) Confocal micrographs of TMRE-stained pulmonary artery smooth muscle cells in control and BCAA-treated media, with corresponding quantification of mitochondrial membrane hyperpolarization, indicated by fluorescence intensity (con: 8.3 ± 17.2, BCAA: 42.7 ± 9.7). P values determined by Wilcoxon rank-sum test. (C) Excess BCAAs increase mitochondrial reactive oxygen species, demonstrated by confocal micrographs of control and BCAA-treated pulmonary artery smooth muscle cells stained with MitoSox Red (con: 6.4 ± 2.1, BCAA: 14.5 ± 6.9). P values determined by Wilcoxon rank-sum test. (D) Representative confocal micrographs of control, BCAA-treated, and BCAA and ferrostatin-1 (FER-1)–treated pulmonary artery smooth muscle cells incubated with 50 mmol/L oleate and 50 mmol/L palmitate and stained for lipid peroxidation using BODIPY (Con: 1.8 ± 0.3, BCAA: 1.4 ± 0.1, BCAA-treated with 5 μmol/L ferrostatin-1 [BCAA-FER-1]: 1.7 ± 0.3). P values determined by Kruskal-Wallis test and Dunn's multiple comparisons test. (E) Incubation with BCAAs induces ferroptotic cell death, demonstrated by viability staining of control, BCAA-treated, and BCAA-treated with FER-1 pulmonary artery smooth muscle cells via Trypan Blue (Con: 9.4 ± 7.5, BCAA: 31.7 ± 11.7, BCAA-FER-1: 13.0 ± 11.1). White arrows indicate trypan blue-positive cells. P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test. Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Activation of BCAA catabolism with BT2 mitigated PAH severity, improved right ventricular function, and enhanced exercise capacity in monocrotaline rats To understand how modulating BCAA metabolism affected PAH pathobiology, we treated MCT rats with the small molecule BT2, a well-validated activator of BCAA catabolism.23 Hemodynamic evaluation demonstrated BT2 counteracted PAH as RVSP, mean pulmonary arterial pressure, and total pulmonary resistance index all decreased compared with MCT-vehicle rodents (Figures 3A to 3C). Histological analysis confirmed BT2 treatment suppressed small pulmonary arterial remodeling (Figure 3D). There was blunted right ventricular (RV) hypertrophy, both at the organ level demonstrated by a decreased Fulton index (Figure 3E) and at the cardiomyocyte level with smaller cardiomyocyte cross-sectional area (Figure 3F). In addition, BT2 treatment reduced RV fibrosis (Figure 3G) and augmented RV function as tricuspid annular plane systolic excursion, RV free wall thickening, and cardiac output normalized to body weight were all significantly improved (Figures 3H to 3J). Maximal right atrial pressures were also minimized in MCT-BT2 animals (control: 6.0 ± 2.0 mm Hg, MCT-vehicle: 14.0 ± 3.0 mm Hg, MCT-BT2: 8.0 ± 3.0 mm Hg), suggesting an enhancement of RV diastolic function. Finally, we showed that BT2 affected systemic biology because it augmented maximal oxygen consumption rates and total distance run on the treadmill in rodents subjected to maximal exercise testing (Figures 3K and 3L). BT2-Mediated Activation of BCAA Catabolism Mitigated Cardiopulmonary Alterations in Monocrotaline Rats (A) BT2 (20 mg/kg) treatment reduced right ventricular systolic pressure (RVSP) (control [con]: 32 ± 3 mm Hg [n = 10], MCT-vehicle [MCT-V]: 84 ± 19 mm Hg [n = 13], MCT-BT2: 53 ± 5 mm Hg [n = 10]), (B) mean pulmonary artery pressure (mPAP) (con: 21 ± 2 mm Hg, MCT-V: 53 ± 12 mm Hg, MCT-BT2: 34 ± 3 mm Hg), and (C) total pulmonary resistance index (TPRi) (con: 66 ± 18 mm Hg/mL/g, MCT-V: 380 ± 126 mm Hg/mL/g, MCT-BT2: 143 ± 60 mm Hg/mL/g); P values determined by 1-way analysis of variance with Tukey's multiple comparison test. (D) Small vessel pulmonary arterial remodeling was mitigated by BT2 (median values averaged from 5 animals per group: con: 29.1 ± 2.3%, MCT-V: 53.4 ± 2.5%, MCT-BT2: 44.7 ± 6.9%; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test). Lung tissue was fixed in 10% formalin, paraffin embedded, sectioned at 4 μm, and stained with hematoxylin and eosin. (E) BT2 diminished right ventricular (RV) hypertrophy at the organ level as assessed by the Fulton index (con: 0.23 ± 0.02, MCT-V: 0.45 ± 0.06, MCT-BT2: 0.36 ± 0.07; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test) and (F) at the cardiomyocyte level (median cross-sectional area values averaged from 5 animals per group: con: 252.2 ± 39.3 μm2, MCT-V: 602.8 ± 70.3 μm2, MCT-BT2: 375.2 ± 41.8 μm2). RV sections were stained with hematoxylin and eosin. (G) RV fibrosis was also mitigated by BT2 treatment (con: 2.8 ± 0.3%, MCT-V: 4.5 ± 0.4%, MCT-BT2: 3.0 ± 0.2%; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test). RV sections were stained with Picrosirius Red. Arrows indicate areas of fibrosis. Inducing BCAA catabolism augmented (H) tricuspid annular plane systolic excursion (TAPSE) (con: 2.6 ± 0.3 mm, MCT-V: 1.5 ± 0.2 mm, MCT-BT2: 2.2 ± 0.4 mm) and (I) RV free wall thickening (con: 119.5 ± 19.5%, MCT-V: 36.4 ± 12.5%, MCT-BT2: 76.0 ± 10.4%); P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test. (J) BT2 treatment also improved cardiac output normalized to body weight (con: 0.31 ± 0.10 mL/min/g, MCT-V: 0.14 ± 0.04 mL/min/g, MCT-BT2: 0.26 ± 0.1 mL/min/g; P values determined by Kruskal-Wallis test with Dunn's multiple comparison test). (K) Endpoint treadmill assessments demonstrated BT2 improved maximal exercise capacity (VO2 max, con: 63 ± 12 mL/kg/m, MCT-V: 19 ± 12 mL/kg/m, MCT-BT2: 52 ± 16 mL/kg/m), and total distance traveled on the treadmill (con: 128.1 ± 29.5 m, MCT-V: 24.6 ± 19.1 m, MCT-BT2: 235.3 ± 85.8 m); P values determined Kruskal-Wallis test with Dunn's multiple comparison test. Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. BT2 treatment reduced lung BCAA levels and restructured BCAA metabolic proteins Next, we evaluated how BT2 modulated lung BCAA levels and BCAA catabolizing proteins using a combination of metabolomics and proteomics analyses. First, analysis of proteins that were significantly altered in the lungs of the 3 groups identified BCAA catabolism, further suggesting a link between BCAA, metabolic pathways, and PAH pathobiology (Supplemental Figure 1A). When we specifically profiled BCAA metabolism, we found that BT2 treatment induced an intermediate phenotype between control and MCT-vehicle lungs as it mitigated down-regulation of several BCAA catabolic proteins (Supplemental Figure 1B). Then, metabolomics data demonstrated BT2 significantly reduced lung levels of all 3 BCAAs (Supplemental Figure 1C). Notably, MCT-BT2 animals also had a parallel decrease in serum BCAAs (Supplemental Figure 1D). Finally, we evaluated another readout of BCAA accumulation, the mTORC1 pathway in lung samples from the 3 experimental groups. BT2 blunted mTORC1 induction as lung levels of phosphorylated mTOR and phosphorylated S6 kinase were reduced by BT2 (Supplemental Figure 1E). BT2 treatment suppressed multiple proferroptotic phenotypes in the lungs of monocrotaline rats To better examine the link between altered BCAA catabolism and ferroptosis, we evaluated both lung metabolomics and proteomics data. Random forest classification identified arginine, ornithine, sulfur-containing amino acids, calculated rate of homocysteine synthesis (homocysteine/methionine), and homocysteine as important variables that differentiated the 3 groups (Figure 4A). Arginine and ornithine are members of the polyamine catabolic pathway,24 while cysteine, methionine, and homocysteine are related to the transmembrane system xc− and trans-sulfuration pathways,25,26 all of which are implicated in ferroptosis13,27,28 (Figure 4B). MCT lungs had elevated levels of arginine, ornithine, methionine, cysteine, and homocysteine, which BT2 diminished (Figure 4C). Importantly, the observed increases in arginine, cysteine, and homocysteine were lung-enriched because these metabolites were not significantly altered in the serum (Supplemental Figure 2A). BT2 Combatted a Proferroptotic Molecular Signature in the Lungs of Monocrotaline Rats (A) Random forest classification of top lung metabolites and lipids distinguishing control, MCT-V, and MCT-BT2. (B) Schematic representation of amino acid-induced ferroptosis, consisting of the polyamine catabolic pathway, transmembrane system xc−, and the trans-sulfuration pathway. (C) BT2 treatment significantly reduced lung concentrations of arginine (Con: 12.3 ± 7.1 pmol/mg, MCT-V: 77.7 ± 25.8 pmol/mg, MCT-BT2: 17.6 ± 10.0 pmol/mg), ornithine (Con: 12.6 ± 7.3 pmol/mg, MCT-V: 67.6 ± 29.0 pmol/mg, MCT-BT2: 18.1 ± 8.6 pmol/mg), methionine (Con: 172.5 ± 67.2 pmol/mg, MCT-V: 493.6 ± 118.2 pmol/mg, MCT-BT2: 307.0 ± 87.0 pmol/mg), cysteine (Con: 52.0 ± 21.3 pmol/mg, MCT-V: 212.1 ± 75.3 pmol/mg, MCT-BT2: 84.7 ± 43.3 pmol/mg), homocysteine (Con: 0 ± 0 pmol/mg, MCT-V: 1.76 ± 0.64 pmol/mg, MCT-BT2: 0 ± 0 pmol/mg), and the calculated rate of homocysteine synthesis (Con: 0 ± 0, MCT-V: 0.004 ± 0.0009, MCT-BT2: 0 ± 0). P values for arginine, ornithine, methionine, and cysteine determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test, Kruskal-Wallis test, and Dunn's multiple comparison test for homocysteine and homocysteine synthesis. Hierarchical cluster analyses of proteins in the (D) ferroptosis and (E) classical complement pathways, red text denotes statistically significant proteins with P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test. Representative confocal micrographs show increased (F) complement deposition (Con: 14.4 ± 5.2, MCT-V: 22.8 ± 9.0, MCT-BT2: 16.3 ± 9.9 C3 fluorescence intensity) and (G) soluble transferrin receptor concentration (Con: 9.8 ± 5.6, MCT-V: 36.6 ± 13.5, MCT-BT2: 10.9 ± 9.7 STFR fluorescence intensity) around the pulmonary vasculature in MCT-V animals, which was mitigated by BT2 treatment. P values determined by the Kruskal-Wallis test and Dunn's multiple comparison test. Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. White arrows indicate pulmonary artery smooth muscle cell layer. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Abbreviations as in Figure 3. When probing lung proteomics data, we found that multiple ferroptosis proteins were up-regulated in MCT-vehicle rats, which BT2 combatted (Figure 4D). Next, we further leveraged our proteomics data to evaluate complement regulation as complement deposition is a marker of lung ferroptosis.7 Proteomic analysis revealed BT2 partially negated the accumulation of several complement proteins in the lungs (Figure 4E). Likewise, immunohistochemistry analysis revealed perivascular complement deposition was present in MCT lungs but reduced with BT2 (Figure 4F). Finally, PASMC-specific expression of transferrin receptor, a validated histological marker of ferroptosis,29 was increased in MCT rats but reduced by BT2 treatment (Figure 4G). A low BCAA diet mitigated PAH and restructured systemic metabolism in monocrotaline rats We then assessed how a nonpharmacological intervention using a low BCAA diet modulated PAH severity. Although not as potent as BT2, a low BCAA diet imparted notable therapeutic effects in MCT rats. In particular, the dietary intervention reduced RVSP, mean pulmonary arterial pressure, and total pulmonary resistance index (Figures 5A to 5C). Histological analysis also revealed diminished RV fibrosis in the MCT-low BCAA diet animals (Figure 5D). Next, the low BCAA diet improved RV function as tricuspid annular plane systolic excursion, RV free wall thickness change, and cardiac output normalized to body weight were all increased (Figures 5E to 5G). Because dietary BCAAs alter metabolism in several organs, we profiled serum metabolomics to gain additional insights into how this dietary intervention impacted systemic metabolism. The low BCAA diet did not significantly decrease any BCAAs in the serum, which may have contributed to its less robust effects compared with BT2 (Figure 5H). However, it did alter the systemic metabolome (Figure 5I) and unbiased random forest classification identified the ratio of homoarginine to symmetric dimethylarginine and predicted rate of homoarginine synthesis as important metabolites in distinguishing the MCT-standard and MCT-low BCAA diet rodents (Figure 5J). Notably, the low BCAA diet increased both homoarginine levels and the ratio of homoarginine to symmetric dimethylarginine (Figures 5K and 5L). In total, these data supported the hypothesis that modulating dietary BCAAs can affect pulmonary vascular disease and indicated that a low BCAA diet may be a nonpharmacological approach to combat pulmonary vascular remodeling.w A Low BCAA Diet Imparted Mild Therapeutic Effects in Monocrotaline Rats Low BCAA diet significantly reduced (A) RVSP (MCT-standard diet [MCT-standard]: 85.4 ± 9.4 mm Hg [n = 10], MCT-low BCAA diet [MCT-low BCAA]: 59.3 ± 18.0 mm Hg [n = 10]), (B) mPAP (MCT-standard: 54.1 ± 5.7 mm Hg, MCT-low 38.1 ± 11.0 mm Hg), and (C) TPRi (MCT-standard: 387.9 ± 236.6 mm Hg/mL/g, MCT-low BCAA: 182.3 ± 122.3 mm Hg/mL/g); P values determined by unpaired Student's t-test. These pulmonary vasculature alterations were matched by significant improvements in (D) RV fibrosis (MCT-standard: 3.23 ± 0.08%, MCT-low BCAA: 1.99 ± 0.36%, RV sections stained with Picrosirius Red) and (E) tricuspid annular plane systolic excursion (MCT-standard: 1.75 ± 0.14 mm, MCT-low BCAA: 2.00 ± 0.29 mm), with noticeable, but not statistically significant, alterations in (F) RV free wall thickening (MCT-standard: 57.4 ± 9.1%, MCT-low BCAA: 71.7 ± 20.8%) and cardiac output normalized to body weight (MCT-standard: 0.168 ± 0.060 mL/min/g, MCT-low BCAA: 0.251 ± 0.109 mL/min/g) (G); P values determined by unpaired Student's t-test. (H) Serum concentrations of isoleucine (MCT-standard: 125.0 ± 60.8 pmol/mg, MCT-low BCAA: 85.8 ± 44.3 pmol/mg) and valine (MCT-standard: 206.0 ± 101.0 pmol/mg, MCT-low BCAA: 166.0 ± 89.6 pmol/mg) were nonsignificantly reduced by the Low BCAA diet, with leucine concentrations unchanged; P values determined by Wilcoxon rank-sum test. (I) Hierarchical cluster analysis of the top 200 serum metabolites/lipids from MCT-standard Diet (n = 9) and Low BCAA diet rodents (n = 9). (J) Random forest classification of top serum metabolites and lipids important for distinguishing MCT-standard Diet and Low BCAA diet rodents. (K) Low BCAA diet significantly increased serum homoarginine concentrations (MCT-standard: 0.3 ± 0.5 pmol/mg, MCT-low BCAA: 6.3 ± 2.8 pmol/mg) and (L) the calculated ratio of homoarginine to symmetric dimethylarginine (MCT-standard: 0 ± 0, MCT-low BCAA: 8.2 ± 4.8); P values determined by unpaired Student's t-test and Wilcoxon rank-sum test, respectively. Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. Black arrows indicate areas of fibrosis. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Abbreviations as in Figure 3. BT2 combatted liver histological, metabolomic, and proteomic dysregulation in monocrotaline rats Next, we investigated how BT2-mediated correction of PAH severity affected hepatic hemodynamic stress and hepatic histological/molecular phenotypes in our 3 experimental groups. BT2 significantly reduced maximal right atrial pressures (Figure 6A). Histological analysis of liver sections showed MCT rats had increased hepatocyte nuclear size, a readout of shear stress,18 which BT2 partially corrected (Figure 6B). To further probe molecular shear stress markers, we evaluated hepatic expression of KLF230 and pAkt31 by immunoblot. KLF2 was elevated and pAkt/Akt reduced in MCT-vehicle livers, but both were normalized by BT2 (Figure 6C). Likewise, BT2 partially mitigated the accumulation of shear stress response proteins in proteomics analysis (Figure 6D). BT2 Treatment Mitigated Activation of Hepatic Shear-Stress Phenotypes and Associated Metabolic Derangements (A) BT2 reduced maximal right atrial pressures (Con: 6.0 ± 2.0 mm Hg, MCT-V: 14.0 ± 3.0 mm Hg, MCT-BT2: 8.0 ± 3.0 mm Hg; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test). (B) Hepatocyte nuclear cross-sectional area was decreased in MCT-BT2 livers (Con: 78.5 ± 0.8 μm2, MCT-V: 108.7 ± 8.3 μm2, MCT-BT2: 87.2 ± 2.6 μm2; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test). Nuclei are stained with WGA and DAPI. Arrows indicate hepatocyte nuclei. (C) Normalized expression of the ratio of phosphorylated Akt to Akt (Con: 1.00 ± 0.24, MCT-V: 0.214 ± 0.03, MCT-BT2: 1.55 ± 0.65) and KLF2 (Con: 1.17 ± 0.25, MCT-V: 1.75 ± 0.27, MCT-BT2: 0.86 ± 0.16) were reduced by BT2 treatment; P values determined by ordinary 1-way analysis of variance with Tukey's multiple comparison test. (D) Hierarchical cluster analysis of proteins involved in the shear stress pathway. (E) Hierarchical cluster analysis of top 250 metabolites and lipids from control (n = 8), MCT-V (n = 8), and MCT-BT2 (n = 8) livers. (F) Random forest classification of top metabolites and lipids important for distinguishing between control, MCT-V, and MCT-BT2 livers. Hierarchical cluster analysis of (G) liver ceramides and (H) top 500 proteins from control (n = 5), MCT-V (n = 5), and MCT-BT2 (n = 6) livers. Kyoto Encyclopedia of Genes and Genomes and Wiki pathway analysis of proteins negatively correlated with (I) maximal right atrial pressure and (J) right ventricular systolic pressure. Hierarchical cluster analysis of (K) proteins in the oxidative phosphorylation pathway and (L) top 500 metabolites and lipids from control (n = 10), MCT-V (n = 17), and MCT-BT2 (n = 10) serum samples. (M) Random forest classification of top metabolites and lipids important for differentiating between control, MCT-V, and MCT-BT2 serum. BT2 treatment significantly reduced (N) sum of asymmetric and symmetric arginine methylation (Con: 0.003 ± 0.0003, MCT-V: 0.005 ± 0.0081, MCT-BT2: 0.00 ± 0.0012) and (O) asymmetric dimethylarginine (Con: 0.8 ± 0.2, MCT-V: 1.1 ± 0.2 pmol/mg, MCT-BT2: 0.8 ± 0.1 pmol/mg); P values determined by Kruskal-Wallis test with Dunn's multiple comparison test and ordinary 1-way analysis of variance with Tukey's multiple comparison test, respectively. Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Abbreviations as in Figure 3. To determine whether these structural and protein-level changes coincided with broader liver metabolic reprogramming, we conducted multiomic profiling of rodent livers. Hierarchical clustering of the top 250 liver metabolites and lipids confirmed MCT-BT2 displayed a distinct metabolome relative to control and MCT-vehicle livers (Figure 6E). Given the liver's key role in BCAA degradation,32 we assessed whether hepatic BCAA metabolism was altered by BT2 treatment. Metabolomic and proteomic analyses showed similar liver BCAA concentrations across groups (Supplemental Figure 3A), and BCAA degradation proteins were reduced in both MCT-vehicle and MCT-BT2 compared with control livers (Supplemental Figure 3B). Notably, hexosylceramide d18:1/16:0 emerged as the top metabolite differentiating the 3 groups (Figure 6F), and most ceramides were elevated in MCT livers, which BT2 treatment negated (Figure 6G). Because ceramides regulate mitochondrial homeostasis,33 we then focused on mitochondria-specific alterations linked to PAH severity. Proteomic analysis of liver mitochondrial enrichments confirmed that MCT-BT2 livers had a unique proteomic signature compared with both control and MCT-vehicle livers (Figure 6H). Next, to link liver proteomic restructuring with PAH severity and RV dysfunction, we performed pathway enrichment analysis on proteins negatively correlated with RVSP and maximal right atrial pressures, which identified predominantly oxidative phosphorylation/electron transport chain pathways (Figures 6I and 6J). Importantly, BT2 preserved oxidative phosphorylation protein abundances, in contrast to the down-regulation observed in MCT-vehicle livers (Figure 6K). Finally, to evaluate the potential systemic metabolic effects of altered hepatic function, we profiled the serum metabolome of all 3 groups. Hierarchical clustering of serum metabolites revealed that BT2 treatment modestly altered the serum metabolome (Figure 6L). Random forest classification identified serum asymmetrical and symmetrical arginine methylation and the sum of dimethylated arginines as important components in differentiating the 3 groups (Figure 6M). Further analysis revealed that BT2 treatment blocked the predicted sum of asymmetrical and symmetrical arginine methylation (Figure 6N) and normalized asymmetrical dimethylarginine levels (Figure 6O). PAH livers displayed hepatocyte nuclear expansion and metabolic derangements To examine the human relevance of our rodent hepatic phenotypes, we performed histological and proteomic analysis of control (n = 4) and PAH (n = 3) liver samples. Consistent with our rodent data, PAH hepatocytes exhibited increased nuclear size (Figure 7A). Next, we analyzed liver metabolic regulation by evaluating the mitochondrial proteome of control and PAH livers. Hierarchical clustering of the top 500 liver proteins showed a distinct PAH proteomic signature (Figure 7B). Pathway analysis of enriched proteins in PAH (false discovery rate: 0.10 caused by small sample size) revealed up-regulation of proinflammatory pathways, glycolysis, and vascular endothelial growth factor signaling (Figure 7C). In contrast, pathway analysis of suppressed proteins indicated significant dysregulation of the liver's metabolic capabilities, with reductions in cytochrome P450, bile secretion, and lipid metabolism pathways (Figure 7D). Livers From PAH Patients Exhibited Hepatocyte Nuclear Expansion and Metabolic Dysregulation (A) Confocal micrographs of liver sections stained with WGA and DAPI. PAH patients had a significant increase in mean nuclear cross-sectional area. Nuclei are stained with WGA and DAPI. Arrows indicate hepatocyte nuclei. (Control: 114.3 ± 33.3 μm2, PAH: 336.1 ± 24.3 μm2). P value determined by unpaired Student's t-test. (B) Hierarchical cluster analysis of top 500 proteins from mitochondrial enrichments of control (n = 4) and PAH (n = 3) livers. (C) Kyoto Encyclopedia of Genes and Genomes (KEGG) and Wiki pathway analysis of proteins up-regulated in PAH samples as determined by a false discovery rate of 0.10. (D) KEGG and Wiki pathway analysis of proteins down-regulation in PAH samples determined by a false discovery rate of 0.10. (E) Key genes involved in ketone metabolism were altered in PAH livers. (F) FLT1 expression was nonsignificantly increased in PAH livers (Control: 31.6 ± 12.9, PAH: 120.1 ± 72.2; P value determined by Wilcoxon rank-sum test). Data are presented as individual data points with the horizontal line representing the mean if normally distributed or median if not normally distributed. ∗P < 0.05, ∗∗P < 0.01, ∗∗∗P < 0.001; NS = not significant. Abbreviations as in Figure 3. In accordance with the liver's crucial role in ketogenesis, we then evaluated the abundances of key proteins involved in ketone metabolism.34,35 Of the 5 ketogenic proteins identified, 3 were down-regulated (ACAT1, BDH1, and HMGCS2) and 2 were up-regulated (HMGCL, OXCT1) in PAH, though none reached statistical significance (Figure 7E). Besides ketogenesis, the liver can also affect systemic metabolism through the production of vasoactive mediators. One notable protein associated with pulmonary vascular remodeling,36 Fms-related receptor tyrosine kinase 1 (FLT1, a receptor for the proangiogenic vascular endothelial growth factor pathway37) was elevated in PAH livers (Figure 7F). Collectively, these data demonstrated PAH livers displayed altered nuclear morphology, dysregulation of numerous metabolic pathways, and heightened levels of FLT1. Discussion Here, we provide multiple lines of evidence that disrupted BCAA metabolism contributes to PASMC dysfunction and pulmonary vascular disease. In human studies, we identified altered BCAA metabolite gradients across vascular compartments in PAH and chronic thromboembolic pulmonary hypertension. PASMC-specific transcriptomic deficits in BCAA catabolism provide orthogonal, cell-resolved evidence supporting a pulmonary vascular origin. Lung transcriptomic analyses further demonstrated dysregulation of BCAA catabolic and ferroptosis pathways in PAH PASMCs, and in vitro studies showed that excess BCAAs induce proferroptotic phenotypes in nondiseased PASMCs. In preclinical models, augmenting BCAA catabolism reduced PAH severity, improved RV function, and enhanced exercise capacity. BT2 treatment mitigated lung BCAA accumulation, restored metabolic pathway activity, and suppressed mTORC1 signaling. mTORC1 hyperactivation may mechanistically link impaired BCAA catabolism to broader amino acid accumulation by suppressing autophagy and lysosomal protein degradation, thereby limiting intracellular amino acid recycling.38 Multi-omic and histological analyses suggest that ferroptosis mediates these effects, because BT2 normalized ferroptosis-associated metabolites and reduced perivascular complement deposition and transferrin receptor accumulation in PASMCs. Consistent with these findings, a low-BCAA diet reduced disease severity and altered systemic metabolism in MCT rats. In parallel, BT2 attenuated shear stress signaling, hepatocyte nuclear remodeling, mitochondrial dysfunction, and systemic metabolic perturbations in MCT livers. Human PAH livers recapitulated these phenotypes, including hepatocyte nuclear expansion and suppression of cytochrome P450 and ketone metabolism pathways. Collectively, these data implicate impaired PASMC BCAA metabolism as a driver of pulmonary vascular remodeling via ferroptosis and link PAH severity to liver metabolic dysfunction through mechanical stress–associated pathways. These findings suggest that targeting disruptions in BCAA catabolism and ferroptosis induction may have therapeutic relevance, and this hypothesis could be addressed with a currently available pharmaceutical. 4-phenylbutyrate (4PBA), an ammonia scavenger commonly prescribed for urea cycle disorders,39 is an effective inducer of BCAA catabolism. In acute lung injury, 4PBA treatment reduces pulmonary inflammation and ferroptosis.40 These results align with our data, in which inducing BCAA catabolism via BT2 blunts classical complement and ferroptosis activation in the lungs. Interestingly, 4PBA treatment mitigates PAH severity in both the chronic hypoxia41 and monocrotaline42 models of PAH. Although these data highlight the potential utility of activating BCAA catabolism in PAH, longitudinal clinical studies are still needed. Currently, there is an ongoing phase II clinical trial evaluating 4PBA in idiopathic or hereditary PAH,43 and this trial may provide further insights regarding the role of BCAA metabolism in PAH pathobiology. Our data further positions ferroptosis as a key regulator of pulmonary vascular disease. We showed that ferrostatin-1 reduces PAH severity, mitigates ectopic complement deposition, and exerts potent anti-inflammatory effects in the pulmonary vasculature.7 Here, we propose that the reduced ferroptosis pathway scores in PAH SMCs in our snRNA data likely describes a proferroptotic state honing to the down-regulation of antiferroptotic regulators GPX4 and SLC7A11.44 Our findings are congruent with several independent studies that link PASMC ferroptosis to PAH across diverse experimental models. In vitro, Li et al45 demonstrates ca-circSCN8A, a proferroptotic circular RNA, is up-regulated in pulmonary hypertension, and its overexpression induces ferroptosis in human PASMCs. Similarly, another study finds that a different circular RNA, circMyst4, decreases under hypoxia and that restoring circMyst4 alleviates PASMC ferroptosis.46 Finally, large-scale bioinformatic analyses of human pulmonary vascular tissues reveal altered expression of multiple ferroptosis-associated genes in PAH.47,48 Thus, these data highlight the role that ferroptosis has in pathological remodeling in the pulmonary vasculature. Beyond pulmonary vascular effects, our data also define an emerging lung–right heart–liver axis in PAH. The liver is an increasingly relevant component of PAH pathobiology, as even minor changes in liver functional assessments are highly prognostic15 and PAH patients display elevated liver stiffness,16 which strongly correlates with RV failure. One potential trigger for PAH hepatology is hemodynamic shear stress, because tricuspid regurgitation propels retrograde blood flow into the right atrium and inferior vena cava,49 which could expose downstream organs to turbulent venous flow, disrupt the phospholipid bilayer, and induce sphingolipid production.50 Sphingomyelinases cleave sphingolipids into ceramides when the membrane is altered, which then accumulate in the mitochondria and destabilize the electron transport chain.51,52 This mechanism could potentially account for the reduced abundance of oxidative phosphorylation enzymes in MCT-vehicle livers. We also nominate a potential biomarker asymmetric dimethylarginine (ADMA) for this process because serum ADMA levels are elevated in MCT-vehicle animals. The liver primarily metabolizes ADMA, and high circulating levels are found in liver cirrhosis, alcoholic hepatitis, and acute liver failure.53 However, future comparisons of liver molecular phenotypes across different disease states may further clarify how PAH specifically impacts the liver. Study limitations First, the human analyses consisted of a small sample size (n = 4 control and n = 3 PAH liver samples), which limits statistical interpretations. However, given the rarity of PAH and limited availability of human tissue, these data still provide important biological insights. To partially address this issue, we performed parallel validation in both rodent and human livers whenever feasible. Second, we evaluated only male rats because they develop more severe RV failure than female rats,7 and assessment of sex-specific responses was beyond the scope of this study, a notable limitation given the female predominance in PAH. The monocrotaline model also incompletely recapitulates sex-specific features of human PAH, which may limit the generalizability of these findings. To partially address this, we performed exploratory sex-stratified analyses of BCAA catabolism and ferroptosis scores in our human transcriptomic data set (Supplemental Figure 4). Both male and female PAH patients demonstrated significantly reduced SMC ferroptosis and BCAA catabolism scores relative to sex-matched control subjects. These findings suggest that the observed disruptions in BCAA metabolism and ferroptosis are not sex-restricted, although dedicated studies in female preclinical models will be required to confirm these effects. Finally, the SVC-to-PA BCAA metabolite gradient likely reflects contributions from both the right ventricle and pulmonary vasculature and future studies incorporating RV-specific tissue or metabolomic sampling will be required to formally distinguish these relative effects. Conclusions Here, we demonstrate alterations in BCAA metabolism promote PASMC ferroptosis and resultant pulmonary vascular remodeling. This in turn induces hepatic derangements likely through shear stress pathways. Thus, enhancement of BCAA may be a novel approach to counteract PAH and its associated hepatopathy. CLINICAL COMPETENCY IN MEDICAL KNOWLEDGE: This study identifies impaired BCAA catabolism as a potential mechanistic contributor to PAH through the promotion of PASMC ferroptosis and links pulmonary vascular disease and subsequent RV dysfunction to hepatic metabolic dysfunction. Restoration of BCAA metabolism reduces pulmonary vascular remodeling, improves RV function, and mitigates hepatic shear stress phenotypes and metabolic abnormalities in preclinical models. Thus, clinicians should recognize PAH as a multiorgan disease in which metabolic dysregulation may couple pulmonary vascular remodeling with RV failure and secondary liver injury, presenting a potentially modifiable therapeutic axis. TRANSLATIONAL OUTLOOK: Future studies should evaluate whether augmenting BCAA metabolism could be a novel therapeutic strategy PAH. Repurposing compounds such as 4-phenylbutyrate could be evaluated in early phase clinical trials. Finally, determining how PAH therapies modulate the lung-right heart-liver axis could provide greater insight into the complex integrative biology of PAH and the specific role of the liver in PAH. Funding support and author Disclosures The Orbitrap Eclipse instrumentation platform used in this work was purchased through High-end Instrumentation Grant S10OD028717 from the National Institutes of Health. This work was supported by the NIHR BioResource, which supports the UK National Cohort of Idiopathic and Heritable PAH, the British Heart Foundation (BHF SP/12/12/29,836), and the UK Medical Research Council (MR/K020919/1). Dr Hong is funded by National Institutes of Health K08 HL169982. Dr Rhodes is supported by BHF Basic Science Research fellowship (FS/SBSRF/21/31,025). Dr Mendelson is funded by National Institutes of Health F31 HL170585. Dr Prisco is funded by an American Heart Association Career Development Award (23CDA1049093, https://doi.org/10.58275/AHA.23CDA1049093.pc.gr.167948) and by National Institutes of Health K08 HL168166. Dr Prins is funded by National Institutes of Health R01s HL158795 and HL162927; has served as a consultant to Edwards and Merck; and has received funding from Bayer. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Footnotes The authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the Author Center. Appendix For an expanded Methods section, as well as supplemental figures and a table, please see the online version of this paper. Appendix Supplemental Figures 1 to 4 and Supplemental Table 1 References Associated Data Supplementary Materials Supplemental Figures 1 to 4 and Supplemental Table 1