Iron, arginine, and redox metabolism in peripheral blood mononuclear cells distinguishes sickle cell disease and pulmonary hypertension

Abstract

Pulmonary hypertension (PH) is a severe vascular complication of sickle cell disease (SCD); yet, not all patients with SCD develop PH, and PH also arises independently. This duality underscores the need to understand their intersecting biology. We integrated metabolomic, proteomic, and elemental analyses of human peripheral blood mononuclear cells (PBMCs) from individuals with SCD, PH, combined SCD–PH, and healthy controls to define shared and distinct mechanisms. PBMCs from SCD patients, regardless of PH status, displayed significantly elevated intracellular iron, consistent with chronic hemolysis and erythrophagocytosis. Multi‐omic profiling revealed condition‐specific immune–metabolic signatures: SCD PBMCs showed mitochondrial suppression and reduced oxidative phosphorylation; PH PBMCs showed dysregulated arginine and creatine metabolism, implicating nitric oxide and polyamine pathways; and SCD–PH PBMCs displayed amplified hemoglobin/iron handling, oxidative stress, and immune activation. Unsupervised clustering confirmed discrete phenotypes, with greatest overlap between SCD and SCD–PH, reflecting the additive impact of hemolysis‐driven iron loading and PH‐driven metabolic remodeling. Histological validation of SCD–PH lung tissue demonstrated iron accumulation in perivascular macrophages, supporting a mechanistic link between systemic PBMC remodeling and pulmonary vascular pathology. Together, these findings establish PBMCs as a readily accessible compartment that mirrors disease‐specific metabolic and immune alterations. By capturing iron, arginine, and redox pathways across SCD, PH, and SCD–PH, our study positions PBMC profiling as a novel tool for mechanistic insight, patient stratification, and biomarker discovery and novel interventions.

INTRODUCTION

Sickle cell disease (SCD) is an inherited genetic hemoglobinopathy caused by a single base pair mutation in the β‐globin gene (HBB) that results in a β6Glu→ Val substitution within the β‐globin(s). ¹ , ² The severity of anemia is affected by allelic combinations and haplotypes that most commonly result in HbSS, HbSC, and HbSβ‐thal disease. ³ The prevalence of SCD is approximately 8 million worldwide, of which roughly 100,000 are in the United States. ⁴ The biological consequence of this mutation causes the polymerization of HbS within the red blood cell (RBC) and RBC sickling, membrane dysfunction, factors that favor intravascular hemolysis, as well as splenic sequestration and erythrophagocytosis. ⁵ Clinically, RBC dysfunction and hemolysis lead to several chronic SCD sequelae, at least in part mediated by excess circulating heme and iron, such as thrombosis, vascular occlusion, acute chest syndrome, stroke, leg ulcers, end organ damage, pain, vision loss, stroke, leg ulcers, and pulmonary hypertension. ⁶

Pulmonary hypertension (PH) is a vasculopathy characterized by medial hypertrophy, occlusions, and the plexiform lesions caused from smooth muscle and endothelial cell proliferation. ⁷ The consequence of pulmonary vascular remodeling is an increase in pulmonary arterial pressure and cardiac dysfunction. The prevalence of PH in the general population is relatively rare, with 1–7 out of million people being diagnosed. ⁷ In contrast, PH occurs in approximately 10% of SCD patients, but the underlying molecular mechanisms remain to be fully elucidated. ⁸ The World Health Organization (WHO) groups PH into five categories, and PH associated with SCD is assigned to group 5, which is described as PH secondary to multifactorial causes with undefined mechanistic contributors. ⁹ The development of PH in SCD can either be a result of precapillary or postcapillary (52.6% or 47.4% of patients, respectively) disease. Multiple subtypes of PH pathophysiology occur in the 10% of patients who are diagnosed and include thromboembolic pulmonary hypertension (CTEPH, 2%), pulmonary arterial hypertension (PAH, 5%), and left ventricular failure (high pulmonary arterial occlusion pressure, 3%), and even combined forms of disease. ¹⁰ In all circumstances, the pulmonary vascular pathogenesis can be linked to hemolysis, anemia, nitric oxide (NO) depletion, hypoxia, and inflammation. Within the context of pre‐capillary disease, inflammation is likely caused from the combination of hypoxia and macrophage reprograming in response to erythrophagocytosis as well as exposure to Hb and its degradation products, heme, and iron. ¹¹

Activated macrophages in the pulmonary adventitia are increasingly recognized as key drivers of vascular remodeling across all types of PH (WHO groups I–V). ¹² These macrophages are thought to be recruited from circulating monocyte and tissue macrophage pools; yet, their phenotype is highly diverse and spans across M1 to M2 subsets. ¹³ In SCD‐associated PH, emerging data reveal a particularly distinct macrophage phenotype marked by oxidative stress inflammation that is driven by ongoing clearance of damaged or sickled RBCs causing intracellular iron accumulation and RBC injury causing exposure to DAMPs, cell‐free Hb, and heme. ¹⁴ Importantly, this Hb/iron‐loaded macrophage signature appears to be unique to SCD–PH and is generally absent in other PH etiologies such as hypoxia‐driven PH or idiopathic PAH. ¹⁵ However, while lung tissue studies have demonstrated these striking differences, the circulating immune precursors that give rise to adventitial macrophages in SCD–PH remain poorly defined. Peripheral blood mononuclear cells (PBMCs) provide a window into both systemic immune alterations and the pool of circulating cells capable of trafficking to the pulmonary vasculature. PBMCs are composed predominantly of lymphocytes (~80%–90%), with monocytes (~5%–10%) representing a smaller but functionally important subset that disproportionately contributes to iron handling and inflammatory signaling. ¹⁶ PBMCs also reflect iron handling, redox stress, and metabolic remodeling that may influence macrophage differentiation and function in the lung. ¹⁷ , ¹⁸ Despite their accessibility, the iron content and multi‐omic phenotype of PBMCs in SCD–PH compared to SCD without PH, PH without SCD, or healthy controls have not been systematically investigated.

Here, we hypothesize that use of an integrated metabolomic and proteomic profiling approach can identify specific phenotypes of PBMCs across these four conditions. Consistent with high RBC turnover, PBMCs from SCD patients, regardless of PH status, contain elevated iron, supporting the concept that iron‐handling preceded over vascular disease. Multi‐omic clustering further revealed that each disease group, controls, PH, SCD, and SCD–PH, harbors a distinct molecular phenotype, with the greatest overlap between SCD and SCD–PH. These findings position PBMCs as a circulating cellular compartment that not only reflects systemic immune–metabolic remodeling but may also provide mechanistic insight into the pathogenesis and progression of SCD‐associated PH.

MATERIAL AND METHODS

Patients

Collection of PBMC from control adults and SCD patients was approved by the University of Colorado Anschutz Medical Center Institutional Review Board (Inflammation and cellular function in sickle cell disease protocol number 20‐0505). All blood samples were obtained from consenting adults (n = 31 SCD patients and n = 18 control adults) at the time of either routine visits to the Colorado Sickle Cell Treatment and Research Center over a 2‐year period, or control adults scheduled volunteer appointments at the Thrombosis Center protocol number. Peripheral blood mononuclear cells (PBMCs) were collected from 18 healthy adults without chronic disease, 26 individuals with sickle cell disease (SCD) without a diagnosis of pulmonary hypertension (PH), 7 SCD patients with confirmed PH, and 10 non‐SCD adults with PH (3 group 1 and 7 group 3). The demographic and clinical characteristics are summarized in Supporting Information S1: Table 1. PBMC samples from PH patients were obtained from the Biobank at National Jewish Health.

Human autopsy SCD–PH and PAH lung tissue

Deidentified human lung tissue was obtained from deceased SCD patients both with and without PH from the University of Colorado Denver Anschutz Medical Center, Aurora, CO and Royal London Hospital, United Kingdom. Tissues were deemed not to be human research samples by both institution Review Boards. Similarly, control and idiopathic pulmonary arterial hypertension (IPAH) lung tissues were obtained from National Jewish Health, Denver, CO, and deemed not to be human research samples.

Sample collection and handling

All blood samples (10 mL) were collected from patients into sodium (0.109 M, 3.2%) Vacutainer blood collection tubes (Becton Dickinson). Samples were processed as described previously. ¹⁹ Briefly, blood was diluted with sterile phosphate‐buffered saline at a 2:1 v.v. ratio. Next, density separation of PBMC was performed with a lympholyte solution added in a 4:1 v.v. ratio.

PBMC isolation

PBMCs were isolated using a Ficoll gradient protocol. ²⁰ , ²¹ Briefly, blood was placed on a rocker at room temperature for 1 h. Then, the samples were transferred to a 15 mL conical tube and diluted with sterile PBS at a 1:2 (blood: PBS) ratio. PBMCs were isolated following the manufacturer's protocol (Lympholyte® Mammal Cell Separation media, Cedarlane Labs, product #CL5110). After isolation, samples were washed with sterile PBS and spun down. Excess PBS was removed, and the pellet was resuspended in 1–3 mL of PBS for cell counting and aliquoting. Each aliquot was frozen in liquid nitrogen after removal of PBS and stored at −80°C.

Histology

Hematoxylin and eosin staining. For H&E staining of mouse and human lung, tissue sections were deparaffinized and rehydrated in distilled water. Then, sections were incubated in Mayer′s hematoxylin (Abcam; 3 min, RT) and immediately washed in water. The slides were then counterstained in alcoholic eosin (Abcam) for 1 min. Finally, the slides were dehydrated in alcohol gradients and mounted with coverslips.

Perls' staining with diamino benzidine intensification. Lung sections were dewaxed‐hydrated, incubated (45 min, RT) in 0.3 M HCl (ACROS Organics) and 2.5% (w/v) potassium ferrocyanide trihydrate (Thermo Fisher), and washed with distilled water. Sections were then incubated (30 min, RT) in methanol containing 0.3% H2O2 (LabChem) and 0.01 M NaN3 (Sigma‐Aldrich). Sections were washed in 0.1 M phosphate buffer (pH 7.4) and incubated with 3,3′‐diaminobenzidine (DAB) and H2O2 (SIGMAFAST ™; Sigma‐Aldrich) for 3 min. at RT. Sections were then counterstained in hematoxylin (Gil no. 2; Fisher Scientific). Lung sections were imaged. A quantitative image analysis was performed by obtaining n = 10 images per slide at ×63 objective magnification to account for the entire lung section. The total number of iron‐stained macrophages was counted per field, and the total macrophage count was divided by the number of fields obtained.

Proteomic analysis

The protein pellets obtained after metabolomics extraction were solubilized in 4% SDS in 100 mM triethylammonium bicarbonate (TEAB) pH 7.5 lysis buffer. The samples were digested using a 96‐well S‐Trap plate (Protifi) following the manufacturer's procedure. Samples were reduced with 10 mM DTT at 55 °C for 30 min, cooled to room temperature, and then alkylated with 25 mM iodoacetamide in the dark for 30 min. Next, a final concentration of 1.2% phosphoric acid and then six volumes of binding buffer (90% methanol; 100 mM triethylammonium bicarbonate, TEAB; and pH 7.1) were added to each sample. After gentle mixing, the protein solution was loaded to the 96‐well S‐Trap plate, spun at 1500g for 2 min, and the flow‐through was collected and reloaded onto the 96‐well S‐Trap plate. This step was repeated three times, and then the 96‐well S‐Trap plate was washed with 400 μL of binding buffer 3 times. Finally, 1 μg of sequencing‐grade trypsin (Promega) and 125 μL of digestion buffer (50 mM TEAB) were added onto the filter and digested at 37°C for 6 h. To elute peptides, three stepwise buffers were applied, with 100 μL of each with one more repeat, including 50 mM TEAB, 0.2% formic acid in H₂O, and 50% acetonitrile and 0.2% formic acid in H₂O. The peptide solutions were pooled, lyophilized, and resuspended in 100 μL of 0.1% FA. Raw data file conversion into peak lists in the MGF format, downstream identification, validation, filtering, and quantification were managed using FragPipe version 13.0. MSFragger version 3.0 was used for database searches against a mouse database with decoys and common contaminants added. ²²

Metabolomic analyses

Briefly, metabolites were extracted at 2 × 10⁶ cells per milliliter at 4°C in cold methanol:acetonitrile:water (5:3:2, v/v/v). After vertexing at 4°C for 30 min, extracts were separated from the protein pellet by centrifugation for 10 min at 10,000g at 4°C. Ultra‐high‐pressure liquid chromatography‐mass spectrometry analyses were performed using a Vanquish UHPLC coupled online to a Q Exactive mass spectrometer (Thermo Fisher) using a 5 min C18 gradient in positive and negative ion modes (separate runs) as previously described. Metabolite peaks were integrated and annotated using El‐MAVEN with the KEGG database. ²³ , ²⁴ , ²⁵

Mass spectrometry‐based metal analysis

Prior to ICP‐MS analysis, pelleted cells (2 × 10⁶ cells per milliliter) were aliquoted into a 15 mL conical tube. 200 μL of 65% nitric acid and 20 ng/mL of gold were added to each sample, followed by addition of 100 μL of 30% hydrogen peroxide and brief vortexing. Samples were then incubated in an oven at 70°C for approximately 2 h. Following incubation, 2190 μL of MilliQ water was added to each tube (final nitric acid percentage of ∼5%) and all samples were vortexed briefly. All samples were then diluted 1:15 in a solution consisting of 20 ng/mL and 5% nitric acid. Final dilutions of 1:250 and 1:3750 were then analyzed via ICP‐MS. Different dilutions were used to ensure that all analytes fell within the calibration curves. ²³ , ²⁴ , ²⁵ All samples were analyzed on a Thermo Scientific iCAP RQ ICP‐MS coupled to an ESI SC‐4DX FAST autosampler system utilizing a peristaltic pump. Optimization of the system was performed before the run by first calibrating the system with ICP‐MS iCAP Q/Qnova Calibration Solution, Specpure. The system was subsequently tuned using a tuning solution consisting of Ba, Bi, Ce, Co, In, Li, and U at 1.00 ± 0.05 μg/L. To monitor performance while the system was running, we continually pumped internal standard mix via the peristaltic pump and monitored the signal throughout the run. Additionally, quality controls of a known concentration of each analyte were injected at the beginning, throughout the run between samples, and at the end of the run. The acceptance criterion for all QCs was ±25% of the known concentration. Thermo Scientific Qtegra software was used for all data acquisition and analysis. ²⁶

Statistics

Data are represented as means ± SEM. Statistical tests were performed with GraphPad Prism V10.3. Heat maps, metabolic pathway analysis, partial least squares‐discriminant analysis, and hierarchical clustering were performed using the MetaboAnalyst 6.0 package. ²⁵ A two‐tailed Student's t‐test was used for comparisons of continuous variables between two groups, and P values were calculated using the log‐rank test. Pathway graphs were prepared on BioRender.com.

RESULTS

Lung histopathology and iron deposition

Histological assessment revealed similar pulmonary vascular remodeling in lungs from SCD–PH patients compared to IPAH, with prominent medial hypertrophy and occlusive lesions observed on H&E staining in both lung tissues (Figure 1A). Non‐heme iron staining demonstrated iron accumulation in SCD–PH lungs, localized predominantly within alveolar and perivascular macrophages, whereas iron staining was minimal in IPAH, but still recognizable, lungs (Figure 1B). Quantitative analysis confirmed significantly higher total iron levels in SCD–PH lungs compared to IPAH (P = 0.0023; Figure 1C). Moreover, the number of iron‐positive macrophages per tissue area was markedly increased in SCD–PH (P < 0.0001; Figure 1D). Together, these findings indicate that pulmonary iron deposition and macrophage iron loading are distinct features of SCD‐associated PH lungs that were studied, suggesting a difference in macrophage populations between the two forms of PH.

Figure 1.

Pulmonary vascular remodeling and iron in IPAH and SCD pulmonary hypertension. (A) Hematoxylin and eosin staining of lung tissue sections from control (IPAH−), IPAH (+), SCD no PH (SCD–PH (−)), and SCD with PH (SCD–PH (+)) at ×4 objective (top row) and ×20 objective (bottom row). Boxed areas indicate regions where higher magnification images were obtained. (B) Perls' iron staining with 3,3′‐diaminobenzidine intensification; the sequence is the same as in (A). All scale bars are represented as 100 μm. (C) Iron content in PBMCs from IPAH (+) patients compared to SCD–PH (+) patients. (D) Counts of iron‐positive macrophages per lung tissue section for IPAH (+) and SCD–PH (+) patients (see Methods). All data are presented as mean ± SEM; P < 0.05 was considered to indicate statistical significance. IPAH, idiopathic pulmonary arterial hypertension; PH, pulmonary hypertension; SCD, sickle cell disease.

Molecular alterations associated with SCD or PH

Having established that iron deposition and macrophage iron loading are distinctive features of SCD‐associated PH lungs, we next asked whether circulating immune cells reflect disease‐specific metabolic reprogramming. To address this, we performed integrated proteomic and metabolomic analyses of PBMCs and used linear regression models to identify metabolites and proteins that distinguish SCD and PH from controls.

Focusing first on SCD, heatmap analysis revealed coordinated shifts across multiple proteasome and protein‐quality‐control components (Figure 2A). An increase in gene‐ontology enrichment of proteins in SCD highlighted ubiquitin‐dependent proteolysis, protein localization, and macromolecule catabolism (Figure 2B). Pathway analysis further indicated suppression of central energy and amino acid metabolism—including 2‐oxocarboxylic acid metabolism, the glyoxylate/dicarboxylate pathway, the tricarboxylic acid (TCA) cycle, and oxidative phosphorylation (Figure 2C).

Figure 2.

Integrated proteomic and metabolomic profiling of PBMCs highlights distinct disease‐specific programs in SCD and PH. (A) Top 50 hierarchical clustering analysis of proteins and metabolites significantly altered by SCD status. (B) Pathway enrichment of metabolites and proteins increased in SCD. (C) Pathway enrichment analysis of metabolites and proteins decreased in SCD. (D) Top 50 Hierarchical clustering of proteins and metabolites, and proteins altered by PH status. (E) Pathway enrichment of metabolites and proteins increased in PH. (F) Pathway enrichment analysis of metabolites and proteins decreased in PH. P < 0.05 by one‐way ANOVA. ANOVA, analysis of variance; PBMC, peripheral blood mononuclear cells; PH, pulmonary hypertension; SCD, sickle cell disease.

By contrast, PH PBMCs showed a distinct molecular profile. Heatmap analysis revealed coordinated variation across sugars and amino acids (Figure 2D). Proteins that increased in PH were enriched for chromatin‐ and transcription‐related programs—nucleosome/chromosome organization, protein–DNA complex assembly, and DNA replication‐dependent chromatin assembly—consistent with heightened epigenetic and transcriptional activity (Figure 2E). Proteins that decreased in PH were linked to vesicular trafficking (early endosome‐to‐Golgi and vesicle‐mediated transport), protein localization, and contractile programs (myofibril and striated muscle cell differentiation), indicating diminished structural and transport functions (Figure 2F).

Together, these analyses highlight divergent immune–metabolic programs: SCD is dominated by enhanced proteostasis with suppression of mitochondrial and amino‐acid metabolism, whereas PH shows increased chromatin/transcriptional activity coupled to reduced vesicular transport and structural pathways.

Distinct metabolomic and proteomic signatures define SCD PBMCs

To further delineate molecular alterations specific to SCD, we performed paired metabolomic and proteomic profiling of PBMCs (Figure 3A). Principal component analysis revealed clear separation of SCD patients from controls at both the metabolite and protein levels (Figure 3B), indicating broad immune–metabolic reprogramming. Heatmap visualization confirmed distinct clustering, with widespread shifts across mitochondrial, glycolytic, and signaling proteins (Figure 3C). Pathway enrichment analysis pointed to increased activity of glutathione and arginine metabolism in SCD PBMCs (Supporting Information S2: Figure 2A). Differential expression analysis identified numerous proteins to be significantly altered in SCD (Figure 3D), including immune and inflammatory mediators as well as mitochondrial regulators. Functional annotation of these proteins revealed overrepresentation of opsonin binding, complement binding, phospholipase activity, and chemokine signaling pathways (Figure 3E).

Figure 3.

Multi‐omic profiling of PBMCs from SCD patients highlights distinct molecular signatures. (A) Schematic of the study design. (B) Principal component analysis (PCA) of metabolomics (left) and proteomics (right) data. (C) Top 50 heatmap of significantly altered proteins and metabolites. (D) Volcano plot of proteomics and metabolomics data identifies proteins that were up‐ (red) and downregulated (blue) in SCD relative to controls. (E) Gene ontology enrichment of SCD‐altered proteins. PBMC, peripheral blood mononuclear cells; SCD, sickle cell disease.

Together, these results demonstrate that SCD PBMCs harbor a coordinated proteomic and metabolomic signature characterized by mitochondrial and metabolic remodeling, coupled with heightened immune activation.

SCD–PH PBMCs are enriched for oxygen‐binding and redox‐regulatory pathways

To investigate the molecular alterations associated with PH in the context of SCD (SCD–PH group), we performed integrated metabolomic and proteomic profiling of PBMCs (Figure 4A). Principal component analysis revealed clear separation of SCD–PH from controls at both metabolite and protein levels (Figure 4B). Heatmap analysis highlighted distinct clustering of SCD–PH samples, with upregulation of proteins involved in mitochondrial metabolism, oxidative stress responses, and inflammatory signaling (Figure 4C). Key proteins included electron transport components (MT‐CO2 and ATP6V1D), redox regulators (TXN, NNT, and SCO1), and stress‐adaptation mediators (CASP8, DIABLO, and GNS).

Figure 4.

PBMC multi‐omics highlights iron‐ and hemoglobin‐related remodeling in SCD–PH. (A) Schematic of the study design. (B) Principal component analysis (PCA) of metabolomics (left) and proteomics (right) data. (C) Top 50 heatmap of significantly altered proteins and metabolites. (D) Volcano plot of proteomics and metabolomics data identifies proteins that were up‐ (red) and downregulated (blue) in SCD–PH relative to controls. (E) Gene ontology enrichment of SCD‐altered proteins. PBMC, peripheral blood mononuclear cells; PH, pulmonary hypertension; SCD, sickle cell disease.

Volcano plot analysis confirmed numerous significantly altered proteins spanning mitochondrial regulators and immune effectors (Figure 4D). Enrichment mapping further demonstrated overrepresentation of hemoglobin‐ and haptoglobin‐binding proteins, oxygen carrier activity, and oxidoreductase/peroxidase pathways (Figure 4E). The strong enrichment of antioxidant and heme‐binding programs suggests that SCD–PH PBMCs activate compensatory mechanisms to buffer chronic oxidative stress and dysregulated oxygen handling. Pathway analysis also revealed increased activity in glutathione and arginine metabolism (Supporting Information S2: Figure 2B).

Together, these results show that SCD–PH PBMCs are defined by a molecular profile distinct from both controls and SCD alone, characterized by enrichment of oxygen‐binding, antioxidant, and redox‐regulatory pathways. This signature underscores the heightened oxidative burden and altered oxygen transport dynamics that accompany pulmonary vascular complications in SCD.

PH PBMCs show dysregulated arginine metabolism and altered proteomic signatures

We next examined PH in the absence of SCD using integrated metabolomic and proteomic profiling of PBMCs (Figure 5A). Principal component analysis revealed clear separation of PH from controls at both metabolite and protein levels (Figure 5B), reflecting broad molecular reprogramming. Heatmap analysis highlighted enrichment of amino acid metabolism—including arginine, citrulline, ornithine, lysine, and proline—as well as energy‐related intermediates such as creatine and creatinine (Figure 5C).

Figure 5.

PBMC multi‐omics highlights arginine and creatine metabolism remodeling in PH. (A) Schematic of the study design. (B) Principal component analysis (PCA) of metabolomics (left) and proteomics (right) data. (C) Heatmap of significantly altered proteins and metabolites highlights broad shifts in metabolic and signaling pathways. (D) Volcano plot of proteomics and metabolomics data identifies proteins that were up‐ (red) and downregulated (blue) in PH relative to controls. (E) Arginine pathway and altered metabolites. Significance *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. PBMC, peripheral blood mononuclear cells; PH, pulmonary hypertension.

Volcano plot analysis confirmed widespread proteomic remodeling, with significant alterations in regulators of cytoskeletal organization, redox balance, and signaling (Figure 5D). Pathway enrichment further identified arginine metabolism as a central feature of the PH phenotype (Supporting Information S2: Figure 2B), reinforced by metabolite mapping that placed arginine and creatine pathways at the core of the signature (Figure 5E). Multiple components of the arginine–nitric oxide (NO) axis and downstream polyamine biosynthesis were dysregulated, including accumulation of arginine, ornithine, and citrulline, along with changes in creatine/creatinine and polyamine intermediates (putrescine, spermidine, and spermine). Quantitative analyses showed significant increases in arginine, ornithine, citrulline, creatine, and creatinine, with variable alterations in polyamines, suggesting an imbalance between NO production, arginine recycling, and polyamine metabolism.

Together, these results indicate that PH PBMCs are characterized by coordinated dysregulation of amino acid and creatine metabolism, particularly within the arginine–NO and polyamine pathways, coupled to proteomic changes affecting cytoskeletal and stress‐response programs. This metabolic signature underscores disrupted arginine homeostasis as a defining feature of PH pathophysiology.

SCD–PH shows amplified oxidative stress and immune activation compared to PH

To assess overlapping and distinct molecular features of PH in the context SCD–PH compared with PH alone, we performed integrative omics profiling of the human PBMCs across controls, PH, and SCD–PH groups (Figure 6A). Principal component analysis showed a clear separation among groups, with PH and SCD–PH clustering distinctly from controls and SCD–PH forming a separate subgroup from PH alone (Figure 6B). Heatmap visualization highlighted broad metabolite and protein shifts—especially in the PH group—particularly involving amino acid metabolism (e.g., arginine, lysine, proline, and methionine), energy‐related intermediates (creatinine, lactate, and pyruvate), and oxidative stress‐linked metabolites (mannitol, xanthine, and oxaloacetate) (Figure 6C). Enrichment analysis of differentially abundant proteins indicated significant overrepresentation of pathways associated with hydrogen peroxide catabolism, immune activation, and responses to external stressors (Figure 6D). Targeted analysis of redox‐ and hemoglobin‐related proteins demonstrated significant upregulation of antioxidant enzymes (CAT, SOD1, SOD2, and PRDX4), oxidative stress carriers (HPX, HP, HBD, and HBG2), and immune regulators (IL6ST) in both PH and SCD–PH, with SCD–PH generally showing the greatest significance (Figure 6E). Interestingly, direct comparison of SCD and SCD–PH revealed minimal differences at both the metabolite (Supporting Information S2: Figure 3A) and proteomic (Supporting Information S2: Figure 3B) levels. This suggests that SCD–PH is not defined by the emergence of entirely new pathways but rather by the amplification of stress programs already present in SCD. While PH and SCD–PH share a core signature of amino acid dysregulation and oxidative stress activation, SCD–PH demonstrates heightened induction of hemoglobin‐binding proteins and redox defense pathways. These findings highlight the development of PH in the context of SCD compounds pre‐existing immune and oxidative stress programs, resulting in an amplified molecular burden.

Figure 6.

Integrated PBMC analyses reveal compounded oxidative and immune stress programs in SCD–PH. (A) Schematic of the study design. (B) Principal component analysis (PCA) of metabolomics (left) and proteomics (right) data. (C) Top 50 heatmap of significantly altered proteins and metabolites. (D) Gene ontology enrichment of altered pathways. (E) Levels of selected proteins involved in redox and hemoglobin/iron handling signatures are amplified in SCD–PH. Significance *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001. PBMC, peripheral blood mononuclear cells; PH, pulmonary hypertension; SCD, sickle cell disease.

Comparative analyses reveal condition‐specific and overlapping metabolic programs in SCD, PH, and SCD–PH

To directly compare molecular differences between SCD and PH, we performed a four‐way volcano analysis integrating both metabolomic and proteomic data sets (Figure 7A). This approach highlighted effect size differences in each condition relative to controls (x‐axis: SCD vs. control; y‐axis: PH vs. control), allowing visualization of shared versus divergent molecular alterations. The majority of features clustered near the diagonal, reflecting overlapping biology; yet, each disease displayed unique directional changes, with select metabolites and proteins enriched specifically in SCD or PH.

Figure 7.

Comparative multi‐omic analysis highlights distinct and compounded immune–metabolic remodeling in SCD and PH. (A) Four‐way volcano plot of effect sizes (Hedges' g) comparing SCD, PH, and controls integrates metabolomic and proteomic data sets, highlighting broad shifts across disease states. (B) Schematic model summarizing key findings: SCD PBMCs show suppression of mitochondrial/TCA cycle activity (orange), PH PBMCs show increased urea cycle metabolism (green), and SCD–PH PBMCs demonstrate compounded erythrophagocytosis, iron/redox stress, and PPP activation (red).

To contextualize these findings, we constructed an integrated schematic model summarizing the major metabolic pathways distinguishing SCD, PH, and SCD–PH presented herein (Figure 7B). In PH, alterations centered on arginine metabolism and the urea cycle, with elevated arginine, ornithine, and creatine/creatinine pointing to dysregulated nitrogen handling. In contrast, SCD was marked by broad suppression of mitochondrial metabolism, particularly reduced TCA cycle activity and oxidative phosphorylation, underscoring impaired energy metabolism. SCD–PH showed additive features of both conditions, with further amplification of erythrophagocytosis, phagocytic‐immune activation, and oxidative stress pathways, consistent with the more severe clinical phenotype observed in this group. Together, these results demonstrate that while SCD and PH share certain metabolic shifts, they are distinguished by condition‐specific molecular programs: suppressed mitochondrial energy metabolism in SCD, altered arginine metabolism in PH, and compounded phagocytosis and redox stress in SCD–PH.

DISCUSSION

Our study identifies circulating PBMCs as a key cellular compartment that mirrors disease‐specific immune and metabolic remodeling in SCD associated with PH. By integrating proteomic, metabolomic, and iron‐handling profiles, we uncovered distinct molecular programs: mitochondrial suppression in SCD, arginine activation in PH, and compounded erythrophagocytosis and oxidative stress in SCD–PH. These results provide new insight into how circulating immune cells may contribute to pulmonary vascular remodeling, while also highlighting PBMCs as an accessible platform for mechanistic discovery and biomarker development in SCD–PH and PH. Previous work has largely focused on macrophages in the lung vascular adventitia, where activated and iron‐loaded macrophages promote remodeling in SCD–PH, but are absent in hypoxia‐induced PH and idiopathic PAH models.

Our findings extend these observations by showing that PBMCs, precursors to vascular macrophages, already carry disease‐specific iron signatures and metabolic remodeling. Although monocytes comprise a minority of PBMCs relative to lymphocytes, their metabolic and iron‐related activity can disproportionately influence bulk omic pathway signatures. This suggests that systemic immune alterations may “prime” circulating cells for pulmonary recruitment and differentiation. Our observations in human and murine SCD–PH lungs suggest that pulmonary vascular adventitial regions are populated with iron‐rich macrophages that are also pro‐oxidative and pro‐inflammatory, in line with prior work showing that heme/iron loading imprints a distinct macrophage activation state. ⁸ , ²⁷ Within these lung adventitial “niches,” it is clear that chemokines, specifically IL‐6, are upregulated in CD163 (+) macrophages, providing a conduit for recruitment of peripheral monocytes and macrophages. ¹⁴ Moreover, while metabolic dysregulation has been described in both SCD and PH, few studies have directly compared these conditions side by side using applied integrated multi‐omics at the PBMC level. Also, to our knowledge, no study has made this comparison with a focus on SCD–PH.

Our multi‐omic results provide several mechanistic insights. First, the suppression of mitochondrial metabolism and TCA cycle activity in SCD PBMCs is consistent with the chronic hemolysis, oxidative stress, and metabolic inflexibility observed in RBCs and endothelial cells in this disease. ²⁸ , ²⁹ Mitochondrial dysfunction is central not only to hemolysis‐driven oxidative stress in SCD but also to endothelial and immune dysregulation in PH. Second, PH‐specific enrichment of the arginine cycle and creatine metabolism pathways aligns with known alterations in nitric oxide bioavailability and vascular tone in pulmonary vascular disease along with markers of aging. ³⁰ , ³¹ Thus, arginine dysregulation represents a shared metabolic axis linking hemolysis‐driven NO depletion in SCD with arginase‐mediated vascular remodeling in PH. Third, the compounded phenotype in SCD–PH, characterized by heightened erythrophagocytosis, hemoglobin/iron handling, and redox stress, suggests that the combination of hemolysis‐driven iron loading and PH‐driven metabolic shifts creates a unique immune–metabolic environment that may accelerate vascular remodeling. ²¹ , ³² Iron‐driven redox signaling is well established in SCD and is increasingly recognized as a contributor to macrophage activation and vascular remodeling in PH. The demonstration that PBMCs capture disease‐specific phenotypes has several important implications. First, along with other markers such as plasma NT‐proBNP and Troponins, PBMCs could serve as a surrogate for the harder to access vascular tissue environment, enabling early‐stage discovery of cellular signatures for early identification of patients at risk for SCD–PH. ³³ Multi‐omic signatures identified here may also guide patient stratification, distinguishing PH secondary to SCD from idiopathic or hypoxia‐driven forms. Beyond diagnostics, understanding immune–metabolic remodeling in PBMCs may uncover novel therapeutic targets such as strategies to restore mitochondrial function, or modulate arginine metabolism in circulating immune cells before they contribute to pulmonary vascular remodeling.

Our PBMC findings highlight arginine metabolic reprogramming as a central axis in PH and particularly SCD–PH biology. For example, we showed that arginine utilization increases with aging across healthy and SCD cohorts, correlating with impaired RBC function and RBC circulatory capacity. ³⁴ In SCD, low plasma L‐arginine is linked to hemolysis, cardiorenal dysfunction, and increased risk of PH. ³⁵ Additionally, the role of arginine metabolism in PH is dichotomous: while arginine depletion and impaired NO bioavailability contribute to the risk of PH in SCD, elevated arginase activity and arginine diversion in polyamine/creatine pathways are also implicated in nonsickle PH pathobiology. ³⁶ Our PH PBMC signature—marked by elevated arginine, ornithine, citrulline, creatine, and creatinine—echoes these systemic patterns and underscores immune cells as active participants, not passive observers.

From an immunometabolic standpoint, macrophage function is tightly regulated by arginine routing: iNOS‐derived NO supports classically activated (M1) inflammatory programs, while arginase‐driven conversion toward ornithine, polyamines, and creatine aligns with repair or remodeling (M2‐like) states. ³⁷ , ³⁸ Metabolic reprogramming governs macrophage activation and epigenetic readiness. ³⁹ , ⁴⁰ Our PH PBMCs' arginine‐through‐polyamine/creatine signature, especially combined with chromatin/transcriptional activation, suggests an arginase‐skewed, epigenetically primed immune phenotype. Elevation in circulating levels of polyamines has also been linked to macrophage immunomodulation in the context of efferocytosis, including phagocytosis of bacterial cells. ⁴¹ , ⁴² Alterations of NADPH‐dependent pathways—such as NO synthesis from arginine ‐ have also been linked to heme catabolism in the context of erythrophagocytosis of healthy or sickle RBCs. ⁴³ , ⁴⁴ In SCD–PH, the overlay of hemolysis‐driven iron and oxidative stress likely amplifies this, favoring accumulation of highly activated, redox‐adapted monocytes. ⁴⁵ Consistent with prior work, chronic exposure to heme and labile iron has been shown to shift monocytes/macrophages toward a highly oxidative, iron‐recycling phenotype characterized by HO‐1/ferritin induction, altered cytokine production, and enhanced efferocytic activity features that can sustain vascular inflammation and promote maladaptive remodeling in SCD and PH. ⁴⁶ , ⁴⁷ Such heme/iron‐driven macrophage reprogramming is highly relevant to SCD–PH pathogenesis, as these iron‐loaded, oxidative macrophage phenotypes are known to amplify vascular inflammation, promote adventitial remodeling, and contribute to the progression of pulmonary hypertension. ⁴⁸ Moreover, renal dysfunction further perturbs arginine homeostasis: decreased filtration and altered creatine metabolism can exacerbate systemic arginine dysregulation and the uremic milieu—amplifying PBMC metabolic shifts and compounding vascular risk. ⁴⁹ , ⁵⁰ Together, these insights position arginine metabolism not only as a biomarker of systemic aging and SCD severity but also as a mechanistic fulcrum in PBMC and macrophage reprogramming. They suggest actionable pathways—from arginase modulation to creatine cycle interference—to disrupt the immune–metabolic priming that may drive pulmonary vascular remodeling in SCD–PH.

Several limitations warrant consideration: our study was limited by human sample size, which may affect the resolution of subgroup analyses. Cross‐sectional sampling does not allow assessment of longitudinal dynamics, and it remains unclear whether PBMC signatures precede PH development or reflect established disease. Iron quantification was performed in bulk PBMCs, and future studies should resolve cell type‐specific contributions (e.g., monocytes, macrophages, and neutrophils). In addition, because lymphocytes constitute the majority of PBMCs, and group‐related differences in lymphocyte and monocyte proportions may influence bulk molecular profiles, macrophage‐associated signatures identified here reflect pathway enrichment rather than assumptions about underlying cell abundance. Thus, the macrophage‐ and monocyte‐associated pathways that we report should be viewed as functional signatures arising within bulk PBMCs rather than direct quantification of specific cell subsets. Future work should validate these findings in larger, longitudinal cohorts to determine whether PBMC signatures predict progression from SCD to SCD–PH. Single‐cell multi‐omics approaches could further resolve immune subpopulations and identify transcriptional programs driving iron handling and metabolic reprogramming. Together, such studies could establish PBMCs not only as a biomarker source but also as a therapeutic target to intercept SCD–PH progression.

In summary, our study demonstrates that PBMCs reflect disease‐specific metabolic and immune alterations in SCD, PH, and SCD–PH. By linking systemic iron handling and multi‐omic remodeling to pulmonary vascular disease, we identify PBMCs as a novel and accessible platform to investigate mechanisms of SCD‐associated PH and to inform future biomarker and therapeutic development.

AUTHOR CONTRIBUTIONS

Francesca I. Cendali: Conceptualization; data curation; formal analysis; writing—original draft; writing—review and editing; methodology; software; validation; visualization; investigation; project administration. Christina Lisk: Conceptualization; writing—review and editing; investigation; formal analysis; project administration; visualization. Amy Argabright: Methodology; data curation; writing—review and editing; formal analysis. Monika Dzieciatkowska: Writing—review and editing; data curation; formal analysis. Nishant K. Rana: Investigation; writing—review and editing. Delaney Swindle: Writing—review and editing; investigation; methodology. Daniel Stephenson: Data curation; formal analysis; writing—review and editing. Julie McAfee: Project administration; writing—review and editing. Natalie Westover: Writing—review and editing; investigation. Melissa Lucero: Investigation; writing—review and editing. Aneta Gandjeva: Investigation; writing—review and editing. Kurt Stenmark: Resources; writing—review and editing; supervision. Rubin Tuder: Writing—review and editing; resources; supervision. Brian B. Graham: Writing—review and editing; supervision; project administration. Tim Lahm: Writing—review and editing; investigation. Vijaya Karoor: Writing—review and editing; investigation; data curation. Gemlyn George: Writing—review and editing; data curation; investigation; supervision. Kathryn Hassel: Writing—review and editing; project administration; supervision. Rachelle Nuss: Writing—review and editing; investigation. Pavel Davizon‐Castillio: Writing—review and editing; investigation. Paul W. Buehler: Writing—review and editing; writing—original draft; resources; project administration; visualization; funding acquisition; supervision; formal analysis; investigation; data curation; conceptualization. Angelo D'Alessandro: Conceptualization; methodology; investigation; supervision; funding acquisition; visualization; project administration; resources; writing—original draft; writing—review and editing. David C. Irwin: Writing—review and editing; writing—original draft; funding acquisition; project administration; visualization; resources; supervision; investigation; methodology; conceptualization.

CONFLICT OF INTEREST STATEMENT

The authors declare that AD are founders of Omix Technologies Inc and Altis Biosciences LLC. AD is a Scientific Advisory Board (SAB) member for Hemanext Inc, Macopharma Inc and Synth‐Med Biotechnologies. GG is a scientific consultant for Agios and Sanofi. All the other authors have no conflicts to disclose in relation to this study.

FUNDING

S. Department of Health and Human Services, National Institutes of Health, and National Heart, Lung, and Blood Institute, HL1598629, HL161004, P01 HL152961, U.R01 HL158076.

Supporting information

Human PBMC Supplemental_Table 1_09_232_025.

Human PBMC Supplemental_Table 1_09_232_025.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available in the supplementary material of this article.