COMBINED HAPTOGLOBIN AND HEMOPEXIN THERAPY FOR THE TREATMENT OF CARDIOPULMONARY DYSFUNCTION IN SICKLE CELL DISEASE

Abstract

Hemolysis and the downstream consequences of cell-free hemoglobin (Hb) and heme contribute to the development of sickle cell disease pulmonary hypertension (SCD-PH). The plasma concentrations of Hb and heme scavenger proteins haptoglobin (Hp) and hemopexin (Hpx) in sickle cell patients are observed to be significantly lower than healthy donors. The unchecked exposure to Hb and heme contribute to vasculopathy and aberrant cardiac function. This is consistent with vascular remodeling co-localized within iron rich macrophages. Based on these observations in patients, we hypothesize that a joint Hb and heme scavenger approach, combining Hp + Hpx as a therapeutic will attenuate hemolysis driven SCD-PH progression in a SCD mouse model. To test the hypothesis, we utilized our validated Berk-SS mouse model of SCD-PH driven by a 10-week moderate hypoxia exposure and weekly subcutaneous administration of Hp + Hpx. At study termination, we analyzed changes in cardiopulmonary iron deposition, right ventricular and pulmonary functional parameters, and multi-omic indices associated with SCD-PH. Our data demonstrates that Hp+Hpx improves pulmonary vascular resistance and right ventricular function including stiffness, afterload, cardiac output, ventricular to vascular coupling ratio, pulmonary vascular resistance and medial hypertrophy. Histological evaluation of lung and right ventricular tissue demonstrates attenuation of cardiopulmonary pathology. Finally, a multi-omic analysis of whole lung and heart tissue demonstrates a rebalancing of proteins related to PH, iron, inflammation, and oxidative stress. This data provides strong pre-clinical evidence for the clinical study of combined Hb and heme scavenger proteins in the treatment of PH-associated SCD.

INTRODUCTION

In sickle cell disease (SCD), the genes responsible for the β-globin component in the tetrameric hemoglobin (Hb) protein contain the nucleotide mutation, GAG to GTG (β6^Glu→Val) [1]. The presence of only one gene containing the altered β-globin subunit, results in sickle cell trait, the carriers usually experience minimal distress, but individuals in which both genes carry the mutation and are homozygous with the mutated hemoglobin (HbS) suffer from sickle cell disease. When HbS becomes deoxygenated (including in tissue after delivering the O₂ load), HbS polymerizes, forming long, rigid fibers within red blood cells (RBCs), causing membrane distortion, reduced deformability, and the crescent-shaped RBC morphology characteristic of the disease [2]. The mechanical stiffness and poor flexibility of sickled RBCs serve as the basis for the vascular obstruction and hemolysis triggering the acute and chronic SCD sequelae, including acute chest and pain syndromes, stroke, cardiac dysfunction, irreversible end-organ damage, and pulmonary hypertension (PH).[3]

The development of PH as a severe complication of SCD occurs in approximately 6–10% of the SCD patient population when verified with right heart catheterization[4] and serves as an independent risk factor for mortality.[5] While multifactorial in origin, a causal link exists between development of PH in SCD and cell free Hb extravasation into pulmonary adventitial regions. [6] The biochemical processes producing vascular injury, promoting macrophage recruitment, and provoking of pulmonary vascular remodeling include NO consumptive reactions and met-Hb accumulation followed by heme and iron exposure that contribute toward oxidation. Studies in our group focus on Hb and heme scavenger administration and dose response in our Berkeley SCD (Berk-SS) mouse model of hemolysis driven SCD-PH[7]. In this model, Berk-SS mice exposed to moderate hypoxia (8,000 ft; P_B ~ 564 mmHg, 15.4 % O₂) for 10-weeks, resulting in progressive cardiopulmonary disease[8, 9]. Our previous work demonstrates that hemopexin (Hpx) administered either subcutaneously at 300 mg/kg three times weekly, or by aerosol for intrapulmonary delivery at 100 mg/kg twice weekly partially rebalances lung oxidative stress and inflammation[7, 10]. Herein, we describe data to support our hypothesis that co-administration of haptoglobin (Hp) with Hpx provides a dual function strategy for detoxifying Hb and heme to attenuate hemolysis-driven SCD-PH.

MATERIALS AND METHODS

Ethical approval and animal care

Young-adult male and female C57Bl/6J (WT) and Berk-SS mice (8 weeks old; ~ 20–25 G) were obtained from Jackson Laboratories (Bar Harbor, ME, USA). Mice were housed and bred in an AAALAC accredited animal facility at the University of Colorado, Denver, Anschutz Medical campus and maintained on a 12:12 light-dark cycle with food and water available ad libitum. Female heterozygous Berk-SS mice were bred with male homozygous Berk-SS mice to generate homozygous offspring. Specifically, Berk-SS mice with genotype Tg (Hu-miniLCR α1 Gγ Aγ δ βs) Hba0/0 Hbb0/0 and the hemizygous with genotype Tg(Hu-miniLCR α1 Gγ Aγ δ βs) Hba0/0 Hbb0 Hbb+ were littermates. Genotyping of mice used for breeding and experiments was performed by Transnetyx (Cordova, TN, USA)[7, 11, 12]. A total of 37 mice WT: n=12, Berk-SS mice: n=25) were used in the present investigation and levels of discomfort and distress were monitored daily by the in-house animal care staff, with a veterinarian available, as needed. Mice presented no pain or discomfort associated with hypoxia and were alert as well as eating, drinking, and grooming normally while housed. All experimental procedures were conducted under the guidelines recommended by The Journal of Physiology[13] and the National Institutes of Health. The use of animals in the experiments were approved by the Institutional Animal Care and Use Committee at the University of Colorado, Denver, Anschutz Medical Campus.

Ethical approval and human subjects

Lung tissue: Deidentified human lung tissue was obtained from deceased SCD patients with identified PH from the University of Colorado Denver Anschutz Medical Center and Royal London Hospital. Use of deceased tissue was considered by the Colorado Multiple Institution Review Board and deemed as nonhuman subject research and given an exempt status.

Plasma samples: Stable adult SCD patients were consented at the time of regularly scheduled visits at the University of Maryland Medical Centers. Stability was defined as a period of at least 3 weeks since a hospitalization, emergency department or clinic visit for acute chest syndrome, pain crisis, infection, or other sickle cell disease-related complication prior to sample collection. Patients were not consented if they were within 2 weeks of a simple or exchange transfusion. Healthy donors were age, sex and race matched, in all cases healthy donor and patient specimens were deidentified (Approved under UMB-IRB-HP-00112426, HP#−00086767).

Experimental design

The overarching experimental design is shown in Figure 1. Berk-SS mice (n=10; male, and 12 female) and wildtype mice (n=6; male, and n=6 female) were continuously exposed to oxygen levels simulating a moderate altitude of 8,000 ft elevation/564 mmHg. Wildtype (WT) mice served as a healthy group for comparisons to Berk-SS, while the Berk-SS mice were divided into to two treatments groups to receive either: (1) saline (VC) n=12(5 male and 7 female) or (2) haptoglobin (Hp; 150 mg/kg) + hemopexin (Hpx; 150 mg/kg) n=10 (5 male and 5 female), twice weekly for 11±1 weeks. As controls, the WT and Berk-SS VC groups were obtained from concurrent therapeutic intervention studies evaluating aerosol and gavage administration. All dosing is based on pilot studies to define an effective dosing strategy for the combination of Hp with Hpx. A dose dependent approach was not feasible for inclusion in this primary study. For the Berk-SS VC group we pooled mice from concurrent VC groups to avoid excessive use of mice, while maintaining comparable ages, weights. Examining the different routes of VC administration between study groups we observed no irregularities between delivery methods suggesting there was no “route of delivery effect” (Figure S1). In some mice we were unable to obtain occlusions to complete right ventricular mechanical function and exact numbers are represented in figure legends. Investigators performing the right ventricular function, morphology, and multi-omic group analyses were blinded to the treatment groups.

Figure 1. Experimental Design and Vaildation.

Schematic illustration of the experimental workflow (from left to right).

Plasma haptoglobin and hemopexin levels

Plasma Hp and Hpx levels were measured in Healthy and SCD patients using a human specific Hp ELISA Kit (Abcam, Waltham MA, Catalog #, ab108856) and human specific Hpx ELISA Kit (Abcam, Waltham MA, Catalog #, ab108859) according to the manufacturer’s protocol.

Open chest solid state catheterization for pulmonary arterial pressures and right ventricular function analysis

At the end of the experiment, mice underwent terminal open chest right ventricular function measurements with a 1.2F, FTE-1212B-4018 pressure/volume catheter (Transonic Systems Inc., Ithaca, NY) inserted by direct cardiac puncture. Briefly, mice were anesthetized with isoflurane (4–5%), and a tracheal incision (~ 1 cm) was performed. A tracheal tube was inserted and connected to an Anesthesia Workstation or Hallowell EMC Microvent with an anesthetic plane maintained at 1.0–2.5% isoflurane in 100% oxygen. A thoracotomy was then performed exposing the heart, the pericardium resected, and a small hole made at the base of the right ventricle with a 30g needle for insertion of the pressure-volume catheter. Steady state hemodynamics were collected with short pauses in ventilation (up to 10 seconds) or high-frequency oscillatory ventilation to eliminate ventilator artifact from the pressure-volume recordings. Occlusions of the inferior vena cava were performed by applying pressure to the inferior vena cava (up to 10 seconds) through the abdominal opening. After pressure, volume and hemodynamic measurements concluded, mice were humanely euthanized by exsanguination and cervical dislocation. Data was recorded continuously using LabScribe2 and analyzed offline. To prevent bias, investigators performing solid state cardiac function measurements were blinded to the dosing groups.

Blood and organ collection

At the conclusion of the experimental protocol, 0.8 mL of blood was collected and placed in an EDTA-K+ vacutainer and a hematocrit (Hct) tube for analysis of plasma and hematocrit, respectively. The plasma was collected, snap frozen in liquid nitrogen, and immediately stored at −80 °C until analysis. Tissues were collected after PBS perfusion as reported previously[11]. Upon heart removal, the right ventricle (RV) and left ventricle with septum (LV+S) were weighed for the assessment of the Fulton Index (RV/LV+S).

Lung morphology

In a subset of mice from each group (WT n=6; VC n=6 and Hp+Hpx n=5), the vasculature in the left lung lobe was quantified for medial thickening. The H&E-stained scanned lung images were equally divided into 100 equal segments. Images that did not have sufficient tissue in the segment were excluded and from the remaining images, 10 randomly selected segments were used for analysis. Using the program STEPanizer (Tschanz & Weibel, 2011), a grid was overlaid on each image and was used to quantify the vasculature within each image. The percentage of vascular tissue to parenchymal tissue in the lung was calculated.

Lung Perls iron staining with DAB intensification

Lung sections were dewaxed-hydrated, incubated (45 min., RT) in 0.3M HCl (ACROS Organics, Geel, Belgium) and 2.5% (w/v) potassium ferrocyanide trihydrate (ThermoFisher, Waltham, MA, USA), and washed with distilled water. Sections were then incubated (30 min., RT) in methanol containing 0.3% H₂O₂ (LabChem, Zelienople, PA, USA) and 0.01 M NaN3 (Sigma-Aldrich, St. Louis, MO, USA). Sections were washed in 0.1 M phosphate buffer (pH 7.4) and incubated with 3,3’-diaminobenzidine (DAB) and H₂O₂ (SIGMAFAST ^™, Sigma-Aldrich, St. Louis, MO, USA) for 3 min. at RT. Sections were then counterstained in hematoxylin (Gil no. 2, Fisher Scientific, Waltham, MA, USA). Berk-SS (VC, n=7) and Berk-SS (Hp + Hpx, n=7) lung sections were imaged. A quantitative image analysis was performed by obtaining n=5–10 images per slide at 63x objective magnification to account for the entire lung section. The total number of iron-stained macrophages were counted per field, and the total macrophage count was divided by the number of fields obtained.

Lung 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, Waltham, Massachusetts, USA; 3 min, RT) and immediately washed in water. The slides were then counterstained in alcoholic eosin (Abcam, Waltham, Massachusetts, USA) for 1 min. Finally, the slides were dehydrated in alcohol gradients and mounted with coverslips.

Proteomics analyses

The protein pellets remaining 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, Huntington, NY) 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 minutes. Next, a final concentration of 1.2% phosphoric acid and then six volumes of binding buffer (90% methanol; 100 mM triethylammonium bicarbonate, TEAB; pH 7.1) were added to each sample. After gentle mixing, the protein solution was loaded onto the 96 well S-Trap plate, spun at 1500x g for 2 min, flow-through collected and reloaded onto the 96 well S-Trap plate. This step was repeated three times, before the 96 well S-Trap plate was washed with 400 μL of binding buffer, three times. Finally, 1 μg of sequencing-grade trypsin (Promega) and 125 μL of digestion buffer (50 mM TEAB) were added onto the filter and digestion for 6 h was carried out at 37 °C. 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 with 0.2% formic acid in H₂O. The peptide solutions were pooled, lyophilized and resuspended in 100 μL of 0.1 % FA. Raw data files were converted to peak lists in the MGF format, while 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 added decoys and common contaminants.

Metabolomic Analyses

Metabolomic analyses for the whole lung (right lobe) and right ventricle tissue in cohorts of normoxic wildtype (WT-Nx), wildtype hypoxic (WT-Hx), normoxic Berk-SS mice (Berk-SS-Nx), Berk-SS hypoxic (Berk-SS-Hx or VC), and Berk-SS hypoxic mice treated with Hp-Hpx (Berk-SS-Hp+Hpx). Metabolomics analyses were performed as previously described[14]. Briefly, metabolites were extracted 10 mg frozen tissue at 4 °C in cold methanol:acetonitrile:water (5:3:2, v/v/v). After vortexing at 4 °C for 30 min, extracts were separated from the protein pellet by centrifugation for 10 min at 10,000 g at 4 °C. Analysis by Ultra-High-Pressure Liquid Chromatography-Mass Spectrometry was performed using a Vanquish UHPLC coupled online to a Q Exactive mass spectrometer (ThermoFisher, Bremen, Germany) using a five 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.

Multi-omic enrichment and pathway analysis

We performed proteomic and metabolomic enrichment and pathway analyses for whole tissue of the lung and right ventricle. Analysis of the MS intensity data in MetaboAnalyst 6.0 resulted in 2- and 3-D Principal Component Analysis (PCA) score plots and hierarchical clustering analysis to build heatmaps. The heatmaps for each -omic profile represent the top 25 significantly differentiated proteins/metabolites unless otherwise stated. We also employ the MetaboAnalyst 6.0 site for metabolomic pathway and enrichment analyses utilizing their MSEA database. Pathway and enrichment analysis for proteomics were conducted on the Omics.net.ca site, accessing the KEGG (gene+protein), Reactome, GO:BP, and KEGG (protein+metabolite) databases.

Statistical analysis

Data are presented as a mean ± standard error of the mean (SEM). Statistical comparisons for y data measurements were completed with one-way analysis of variance (ANOVA) with the same standard deviations between vehicle control and treated Berk-SS mice. Post-hoc analyses are performed with the Tukey-Kramer multiple comparison tests, unless otherwise noted. Treatment effects between the Berk-SS VC and Berk-SS Hp+Hpx treated mice are completed using a student’s t-test. Statistical analysis was completed using the statistical software package GraphPad (Version 10.3.1) for MacOS, GraphPad software, Boston, MA, USA, Statistical significance is defined as p ≤ 0.05.

RESULTS

Haptoglobin and hemopexin plasma concentrations

Hp and Hpx levels follow a distinct pattern of depletion in SCD that is characterized by the inability of Hp synthesis to keep up with Hb-Hp clearance. Here, steady state SCD patients showed a 5-fold lower Hp concentration compared to healthy donors (Figure 2A). As a complementary detoxification system, Hpx levels decline after plasma Hp depletion. Also, SCD patients showed a 3-fold lower Hpx concentration compared to healthy donors (Figure 2B). Consistent with previous reports[15, 16], these data suggest that heme exposure and Hpx depletion are dependent on Hp levels in human SCD. Therefore, we posit that hemolysis driven cardiopulmonary complications of SCD may be attenuated by a co-formulated Hp and Hpx therapeutic cocktail.

Figure 2. Haptoglobin and hemopexin are decreased in sickle cell disease patient plasma.

(A) Data shows haptoglobin concentrations in healthy donors (n=19) and sickle cell patients (n=19) plasmas. (B) Data shows hemopexin concentrations in healthy donors (n=19) and sickle cell patients (n=19) plasmas. In each analysis the data is expressed as mean ± SEM and plotted as individual values. Significance: **** p < 0.0001. The * next to the symbols indicates statistical outliers determined by a Grubb’s test in GraphPad.

Hp+Hpx cocktail improves right ventricular function in SCD mice

In our Berk-SS model of hemolysis driven cardiopulmonary disease, symptoms manifested as increased right ventricular systolic pressure, a less compliant right ventricle, and high pulmonary vasculature resistance. This aberrant physiology led to a chronically increased right ventricular stiffness, increased afterload, and right ventricular hypertrophy that caused a reduction in ventricular to vascular coupling ratio, predisposing to right heart failure. Figure 3 and Figure 4 illustrate the functional differences between each cohort. Right ventricular pressure volume loop analysis is shown in Figure 3A and B and individual bar graphs for specific functional parameters are presented in Figures 3C–I. These data indicate that subcutaneous administration of Hp+Hpx, improved right ventricular stiffness (~20%) towards that of a WT mouse and significantly reduced afterload (~30%) relative to vehicle control (1.57±0.09 VC vs. 1.12±0.05 Hp+Hpx, p=0.0144 t-test). Additionally, stroke volume was preserved and accompanied the reflexive increase in heart rate, which improved cardiac output by 67% (7.6±0.56 VC vs. 12.69 ± 0.51 Hp+Hpx; p<0.001 t-test). The effectiveness of scavenging Hb and heme is further demonstrated as a modest improvement in the ventricular to vascular coupling ratio toward that of WT mice [17]. Examining sex as a biological variable did not reveal any obvious differences between male and female mice (data not shown).

Figure 3. Right ventricular mechanical functional analysis in Berk-SS after Hp+Hpx cocktail teatment.

(A) Representative tracings of Pressure Volume (PV) Loops for WT, vehicle control (VC) and Hp+Hpx treated mice. In (B), the corresponding schematics of PV loops illustrate the differences in RV function between WT vs VC, and VC vs Hp+Hpx. Figures C-H contain bar graphs of Right Ventricular Stiffness: WT (N=12 0.03±.006 mm Hg·μl⁻¹); VC (N=15 0.08±0.01 mm Hg·μl⁻¹); Hp+Hpx (N=7 0.06±0.0.01 mm Hg·μl⁻¹); Afterload: WT (N=12 1.12±0.08 mm Hg·μl⁻¹); VC (N=15 1.56±0.09 mm Hg·μl⁻¹); Hp+Hpx (N=7 1.12±0.05 mm Hg·μl⁻¹); Stroke Volume: WT (N=12 26.89±2.16 μl); VC (N=12 20.39±0.99 ml); Hpx (N=8 25.95±1.63 ml); Cardiac Output: WT (N=12 13.19±1.62 ml·min⁻¹); VC (N=15 7.61±y0.56 ml·min⁻¹); Hp+Hpx (N=8 12.69±0.50 ml·min⁻¹); Ventricular-to-Arterial Coupling Ratios: WT (N=12 0.84±0.06); VC (N=15 0.45±0.04); Hp+Hpx N=7 0.67±0.0.18); Contractility: WT (N=12 1.25±0.22 mm Hg·μl⁻¹); VC (N=15 0.76±0.11 mm Hg·μl⁻¹); Hp+Hpx (N=7 0.88±0.20 mm Hg·μl⁻¹); and Heart Rate: WT (N=12 440±9 beats·min⁻¹; VC (N=15 415±10 beats·min⁻¹), Hp+Hpx (N=8 500±23 beats·min⁻¹), respectively. Data is expressed as mean ± SEM and plotted as individual animals. Significance: * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001.

Figure 4. Pulmonary vascular analysis in Berk-SS mice after Hp+Hpx cocktail treatment.

(A) Pulmonary Vascular Resistance: WT (N=12 2.70±0.20 Woods unit); VC (N=15, 4.747±0.58 woods unit); Hp+Hpx (N=8, 2.91±0.35 wood unit); (B) Right Ventricular (RV) Weight: WT (N=12, 0.024±0.0.002 g); VC (N=15, 0.05±0.0.003 g); Hp+Hpx (N=9 0.04±0.0.002 g). Data is expressed as mean ± SEM and plotted as individual animals. Significance: * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.00. (C) Right ventricular systolic pressure (RVSP): WT (N=12 30.88±0.0.90); VC (N=15 31.09±0.0.92); HP+Hpx (N=9, 28.6±1.6) (D)Fulton Index (RV/LV = S): WT (N=12 0.24±0.0.02); VC (N=15 0.39±0.0.02); HP+Hpx (N=9, 0.34±0.0.02); and (E) Pulmonary arterial medial thickness: WT N=6, 8±0.85); VC (N=6, 7.9±0.59); Hp+Hpx (N=5, 6.42±0.23). Representative H&E tissue staining of mouse lungs are shown in the images below the bar graph. Arrows indicate initial remodeling of the vasculature, scale bars = 100 microns. (F) Human SCD-PH lung lesion: Top (1) (20x objective) and (2) (40x objective) indicate endothelial proliferation extending through the vascular lumen, arrow. Bottom (3) (20x objective) and (4) (40x objective) indicate both increased muscularization and neo-intimal formation, filled arrow. In all figures, scale bars = 100 microns.

Hp+Hpx cocktail improves pulmonary vascular resistance in SCD mice

During the pathogenesis of pulmonary hypertension, the medial layer of the pulmonary artery undergoes gradual thickening. Decreasing the arterial diameter increases pulmonary vascular resistance. In this study, Hp+Hpx cocktail treated Berk-SS mice exhibited a diminution in pulmonary vascular resistance (Figure 4A). However, this correction did not significantly improve either the Fulton Index or right ventricular weights (Figure 4B and C). Consistent with a reduction in pulmonary vascular resistance, we observed a reduction in lung medial thickening (Figure 4D). It is accepted that pre- and post-capillary pulmonary vessels do not dramatically remodel in mice as compared to vascular remodeling in human PH and this is also true of SCD[18]. For comparison, images of remodeled vasculature from long standing PH in human SCD lungs are shown (Figure 4E). Nonetheless, cardiopulmonary dysfunction in our Berk-SS model demonstrate improved right ventricular and pulmonary function that is consistent with the response observed with our earlier Hpx monotherapy studies[7, 10, 19], suggesting that heme scavenging results in an improvement in reactive right ventricular afterload. This observation reflects increased stroke volume – which, in turn, accompanies an attenuation of pulmonary vascular resistance and right ventricular stiffness. In humans with PH, studies demonstrate a strong correlation between stroke volume and the six-minute walk distance[20, 21].

Hp+Hpx cocktail attenuates pulmonary vascular macrophage accumulation in SCD mice

Based on our observations of human lung tissue obtained at autopsy from individuals with confirmed SCD-PH we note a distinct pattern iron macrophages surrounding the pulmonary vasculature (Figure 5 A–D)[6, 7]. Our Berk-SS mouse model of SCD-PH suggests pulmonary vascular iron accumulates in similar pulmonary vascular regions as humans, albeit not to the same degree of severity (Figure 5 E–F). Here, we observe treatment with the Hp+Hpx cocktail significantly attenuates pulmonary vascular iron rich macrophage accumulation surround the lung vasculature (Figure 5I).

Figure 5. Iron rich macrophage vascular distribution in human and Berk-SS pulmonary hypertension:

We have previously identified that cells surrounding the pulmonary vasculature in SCD pulmonary hypertension are CD163(+) macrophages [6]. (A) pulmonary vasculature iron rich macrophage (arrows) deposition shown in the box region at low magnification (20x obgective magnification). (B-D) Show each vessel independently at high magnification (63x objective magnification); (E-F) Show representative lung staining of VC treated Berk-SS mice that indicate pulmonary vascular macrophages localized in perivascular regions; (G-H) Show representative lung staining of Hp + Hpx treated Berk-SS mice. Representative images in both E and G were taken at 20x objective magnification, while images in both F and G were obtained at 63x objective magnification; (I) Shows the quantitation of iron rich macrophages from VC and Hp+Hpx treated mice (n=6 per group). The mean number of macrophages per section was 11.3 ± 1.96 in the VC treated group and 1.54 ± 0.478 in the Hp+Hpx group. Data are represented as mean ± SEM. **p=0.0012 and all scale bars are = 100 microns.

We next performed multi-omics on RV and lung tissue analyses were performed as both untargeted and then targeted. Groupings and categories that are representative of the expected treatment effects were selected targeted analysis. For proteomics these include the following categories: Oxidative stress; Pulmonary hypertension; Inflammation; Fibrosis. In each of the targeted groupings, the top 25 gene products were extracted to illustrate the up- and down-regulated proteins. The differences between vehicle control (VC) and Hp + Hpx dosed mice were compared using a student’s t-test.

Multi-omic analysis of right ventricle tissue

Proteomic analyses:

We conducted untargeted hierarchical clustering analysis, creating heatmaps for all proteins with p ≤ 0.6 and the top 25 differentially expressed RV proteins for five combinations of the various mouse cohorts, including: (1) All Five: WT-NX, WT-Hx, Berk-SS Nx, Berk-SS-Hx and Berk-SS-Treated(Hp +Hpx Cocktail); (2) Three Berk-SS: Berk-SS Nx, Berk-SS-Hx and Berk-SS-Treated(Hp +Hpx Cocktail); (3) Hypoxia-Exposed: WT-Hx, Berk-SS-Hx and Berk-SS-Treated(Hp +Hpx Cocktail) and (4) Berk-SS VC vs Berk-SS-Treated(Hp +Hpx Cocktail). Consideration of the Principal Component Analysis (PCA) Scores plot for the Hypoxia-Exposed group demonstrates a clear separation between healthy WT mice and both hypoxia-exposed cohorts of Berk-SS mice, which exhibit a distinct proteomic profile. These heatmaps and PCA Scores plots may be found in the Supplemental Material as Figures S2 and S3.

The top 25 untargeted analyses provide insight into the effects of treatment with Hp+Hpx, including upregulation of proteins important to the metabolism of lipids (Plpp1) and the mitochondrial electron transport chain and oxidative phosphorylation, OXPHOS, (Coxba1, AIP, and Cox6c). We also see downregulation of cellular stress response and growth proteins (RO60, Raf1, and Map2k3).

Targeted Proteomic analyses:

To understand how Hp-Hpx therapy improves cardiopulmonary function as it relates to hemolysis-driven PH [7], we screened the 3996 proteins in the raw MS data file by cross-referencing them to four curated lists of proteins with established relevance to the pathophysiology and pathways of SCD-PH. These curated lists focused on PH, oxidative stress, inflammation, and fibrosis. Again, utilizing hierarchical cluster analysis, we created heatmaps for the top 25 significantly differentially expressed proteins for each list, as shown in Figure 6A–D. (The screening statistics are presented in Table S1, while Table S2 summarizes the differentiated proteins for each of the four subsets.) The bar plots to the left of each heatmap indicate proteins of interest. Overall, the four heatmaps demonstrate that Hp+Hpx treatment results in a shift toward in downregulated proteins. We observe noteworthy changes in the expected areas of heme scavenging proteins including: oxidative stress, Figure 6A, where proteins involved in redox homeostasis, (Nqo1, Txnrd2), mitophagy at hypoxia[22] (Fundc1), lipid metabolism and ferroptosis[23] (Lpcat3), and the TGF-β pathway (Me1, Rbx1). Proteins of significance to PH (Figure 6B) include regulators of asymmetric dimethylarginine levels[24] (Ddah2, Ddah1), endothelial cell dysfunction (Stat1, Gnaq, RhoB), and regulation of the coagulation cascade (Dcn, Myo1c[22], Serpinc1.) Substantial differentiation among inflammatory proteins, Figure 6C, includes those important to mitochondrial function (Hccs, Adrm1, C1qbp, Ndufs4, Ndufs6), the Ras/MapK pathway relating to cardiomyopathies (Lmna, Raf1), and complement reception (C1qbp). The influence on fibrotic processes (Figure 6D) appears as reduced expression of proteins involved in platelet activation (Cdc42, Map2k2 and Pecam1), saturated fatty acid metabolism (Map2k2, Map2k6, Raf1, Rock1), cell migration and/or proliferation (Pecam1/CD31, Cdc42, Lgals1, Msn), cell adhesion (Cdh2, Jam2, Mypn, Map2k3), and EGFR signaling (Dnm1, Dnm3, Gpc6) Note that also Hp appears as downregulated. Tabulated data for each subset and the top 25 resides in Supplementary Material (Tables S3–S7.)

Figure 6. Targeted Proteomics and Metabolomics with Pathway and Enrichment Analysis for the RV.

Heatmaps arising from screening for the top 25 proteins related to SCD-PH using the (A) oxidative stress, (B) pulmonary hypertension, (C) inflammation, and (D) fibrosis subsets. The bar plots to the right of each heatmap indicate proteins interest, also highlighted in bold color on the heatmaps. Other highlighted proteins are mentioned in the text. The PCA score plots in (E) confirm the validity of the comparison of the metabolites in the heatmap (F), demonstrating clear separation between WT and Berk-SS mice upon hypoxic exposure. The large pie chart in (G) represents the most relevant KEGG pathways involved given the conditions under study, while the smaller pie charts in (I) indicate the altered abundance of each metabolite in terms of increase or decrease and the table in (I) summarizes the enrichment of significant KEGG pathways in terms of p-value derived from pathway topology analysis. The cut-off for significance is p ≤ 0.05.

Metabolomic Analysis of the Right Ventricle:

The PCA plots in Figure 6E demonstrate the statistical validity of our comparison between Berk-SS (VC) and the Berk-SS treated cohort for RV metabolomics. As with the proteomics, hierarchical clustering analyses was completed for the five cohorts as described above for our metabolomic data. Heatmaps for the specific comparisons among the five groups are presented in the Supplementary Material, Figure S4 The two PCA score plots in Figure 6E demonstrate the separation between WT and Berk-SS mice in two and three dimensions, validating the comparisons of only Berk-SS mice in the heatmap of the top 25 most differentiated metabolites comparing Berk-SS VC vs. Berk-SS Hp+Hpx cohorts, shown in Figure 6F. Arranging the metabolites in order of assigned KEGG pathways and then by relative abundance (Figures 6G and 6H), indicate an increased presence of unsaturated fatty acids, while the concentrations of saturated fatty acids decreased. In health, cardiomyocytes rely upon fatty acid oxidation (FAO) to fulfill up to 70% [25] of their ATP needs, with the rest provided via glucose, the TCA cycle, and the amino acids cysteine, glutamate, histidine, and lysine[26]. The dominant pathways of relevance are arranged in Figure 6I, with the polyunsaturated linoleic acids increasing, while ammonia recycling remains up. While upregulation of ammonia recycling in pulmonary hypertension can impact on arginine availability for nitric oxide synthesis, compete with NOS by arginase, and affect endothelial function and oxidative stress[23–25], we see downregulation of arginine metabolism and decreases in antioxidant production through the glutathione, methionine, and glutamate metabolism pathways. In the Supplementary Material, Tables S9 and S10 summarize metabolomic pathway analysis.

Proteomic and metabolomic associations with right ventricular function:

To examine the relationship between the -omics data as they relate to right ventricular function, we performed Pearson correlation analysis on the Berk-SS (VC vs Hp+Hpx) cohorts. The associations between right ventricular mechanical functional parameters and protein and metabolite expression are presented in Figures 7 and Figure 8, respectively. Each set contains plots for the top 10 positive and negative correlates with p-values ≤ 0.05 highlighted in red and blue, respectively. If a specific RV functional metric lacked 10 significant protein or metabolite associations, data with 0.05 ≤ p-values ≤ 0.1 are represented in gray. Data for the top 10 RV proteins from this analysis is contained in the Supplementary Material (Table S8.)

Figure 7. Proteomic Hemodynamic Associations:

Pearson correlation plots for hemodynamic parameters and the proteomic analysis. The top 10 positive and negative correlations with a p-value ≤ 0.05 are highlighted in red or blue, respectively. Proteins correlate with 0.05 ≤ p-value ≤ 0.01 are gray.

Figure 8. Metabolomic Hemodynamic Associations:

Pearson correlation plots for RV hemodynamic parameters and RV metabolites. The top 10 positive and negative associations with a p-value ≤ 0.05 are highlighted in red or blue, respectively. Metabolic correlates with 0.05 ≤ p-value ≤ 0.01 are gray.

Multi-omic analysis of lung tissue

Following the same approach used for the RV, we performed untargeted proteomic analysis of whole lung tissue, analyzing the data using PCA and hierarchical clustering. The heatmaps for the five cohort groups and PCA scores plots for comparison among the five cohorts (q.v.) are presented in in the Supplementary Material (Figure S5, with corresponding data in Table S21.)

Targeted Lung Proteomics:

Investigating the influence of Hp-Hpx cocktail on the lung proteome within the context of iron scavenging Hb and heme iron attenuates the progression of SCD-PH, [27] we screened the 6714 detected proteins of raw whole tissue lung proteomic data with the same four subsets utilized in our targeted proteomic analysis of the RV tissue. In the Supplementary Material, Table S15 provides descriptions of each subset, the raw data, and the resultant whole lung tissue proteins matching to each subset, while Table S16 summarizes the proteins ranked in order of p-value, based on a Student’s t-test between the Berk-SS VC and Berk-SS-Hp+Hpx cohorts, and Tables S17–S21 contain names for all proteins in the targeted heatmaps. Heatmaps for the top 25 significantly differentially expressed proteins for each list are shown in Figure 9. Interestingly, heatmaps for the 25 most differentiated proteins related to oxidative stress, PH, inflammation, and fibroses in lung tissue after 10 weeks of Hp+Hpx therapy exhibit protein expression shifted toward upregulation, in contrast to that of the RV.

Figure 9. Targeted Proteomics and Metabolomics with Pathways and Enrichment Analysis for the Lung.

Heatmaps arising from screening for the top 25 proteins related to SCD-PH using the (A) Oxidative stress, (B) pulmonary hypertension, (C) inflammation, and (D) fibrosis subsets. The bar plots to the right of each heatmap indicate proteins interest, also highlighted in bold color on the heatmaps. Other highlighted proteins are mentioned in the text. In (E), the PCA score plots in confirm the validity of the comparison, of the for the heatmap in (F) demonstrating the separation, in different planes, of WT and Berk-SS mice after hypoxic exposure. The large pie chart in (G) represents the most relevant KEGG pathways involved given the conditions under study, while the smaller pie charts indicate the altered abundance of each metabolite in terms of increase or decrease. The table in (H) summarizes the enrichment of significant pathways in terms of p-value, calculated using pathway topology analysis. The cut-off for significance is p ≤ 0.05.

Lung metabolomic analysis:

The relatively low number of metabolites render filtering by p-value or screening via list for targets unnecessary. The PCA score plots in Figure 9E validate the direct comparison, between WT and both cohorts of Berk-SS, indicating A 3-D representation best elucidates the separation. Considering the heatmap in Figure 9F, which contains the top 25 differentially abundant metabolites between Berk-SS VC vs. Berk-SS Hp+Hpx, the data suggests that Hp+Hpx therapy increases indole and tryptophan metabolism (5-hydroxyindoleacetate, 6-hydroxykynurenic acid, γ-oxalocrotonate); glutathione homeostasis metabolites, (oxidized glutathione, S-glutathionyl-L-cysteine); and a participant in the pentose phosphate pathway (Ribose 1-phosphate.) Further, hypoxanthine and ribose 1-phosphate, resulting from purine degradation through phosphorylation of inosine, also increase. The metabolites decreasing include succinate and phosphoenolpyruvate (PEP) – the final step of glycolysis to generate ATP signify a defect at key point between aerobic and anaerobic mitochondrial function. The pie charts in Figure 9G summarize simple KEGG analysis of significantly changed metabolic pathways, revealing that fatty acid metabolism possesses less relevance in the lung than in the RV. The table in Figure 9H summarizes that relatively few enriched pathways were significantly impacted upon treatment with the Hp+Hpx cocktail. The Supplementary Material (Tables S22 and S23) contains expanded data for lung metabolomic pathway analysis.

DISCUSSION

Chronic hemolysis and the down-stream consequences of cell free Hb and heme play a critical role in the development and progression of cardiopulmonary complications including PH in SCD patients [26]. We previously reported that subcutaneous administration of Hpx (300 mg/kg) given three times weekly[7], as well as intrapulmonary administration of Hpx (100 mg/kg) delivered directly to the lungs twice weekly[10], attenuates pulmonary vascular pathology, right ventricular dysfunction and improves exercise tolerance in our Berk-SS model. In the plasma of patients with SCD, Hp synthesis cannot keep up with Hb-Hp clearance, depleting Hp levels, leaving Hpx to function as a final defense against heme loss once free Hb cannot be cleared by Hp [15, 16]. Consequently, this detoxification system serves in sequence to limit chronic Hb exposure-induced pathophysiology[27, 28]. Depletion of both proteins due to ongoing hemolysis, the pathophysiologic contributions of Hb and heme become important features that contribute to SCD sequela.

Based on this knowledge, we hypothesized that systemic delivery of co-formulated Hp and Hpx as a therapy would effectively attenuate the hemolysis-driven, progressive cardiopulmonary disease in our sub-chronic (approximately 3-month) hypoxia exposed Berk-SS mice. The current investigation integrated hemodynamic, immunohistochemical, proteomic, and metabolomic analysis to evaluate the effectiveness of Hp+Hpx administered twice weekly by subcutaneous administration. Our data provides compelling evidence that the combination of Hp and Hpx dosed simultaneously protects the pulmonary vasculature from hemolysis-mediated oxidant stress and accompanying right ventricular dysfunction.

In SCD-PH patients, pulmonary vascular remodeling drives increased pulmonary arterial pressures, pulmonary vascular resistance, increased right and left ventricular afterload and ventricular hypertrophy [7]. Berk-SS mice exposed to moderate hypoxia, in a sub-chronic time frame, share some similarities in cardiopulmonary function tests with humans suffering from SCD-PH. Comparable parameters include: increased pulmonary vascular resistance; increased right ventricular afterload, a stiffer non-compliant right ventricle, and a reduced ventricular to vascular coupling ratio, which are all hallmarks of right ventricular dysfunction [7, 29]. The dosing schedule of Hp+Hpx improved right ventricular stiffness, afterload, and pulmonary vascular resistance, which resulted in increased cardiac output. [7] Our dosing strategy for combined Hp+Hpx therapy also alleviated iron rich macrophage accumulation in proximity to lung vasculature, as well as pulmonary vascular remodeling. These data are consistent with our studies that evaluated higher (300 mg/kg) and more frequent (three times weekly) Hpx dosing, suggesting that a lower, and less frequently administered combination of Hp+Hpx (150 mg/kg of each protein, twice weekly) may offer a sequential protection against the hemolysis driven effects of Hb and heme on cardiopulmonary dysfunction. Pre-capillary PH is thought to be driven by hemolysis in SCD [30]. Therefore, we suggest that sequestering both Hb and heme using a combination treatment could be more advantageous than monotherapy (i.e., Hp or Hpx alone). Compared to our previous study that tested Hpx alone [7], here we observed ventricular to vascular coupling (VVCR), RV cardiac output as well as pulmonary vascular resistance that were all normalized to that of WT mice. Indeed, this observation suggests a role for Hp combined with Hpx as a therapeutic strategy to attenuate SCD PH.

Proteomic data from the right ventricular tissue suggests that global protein changes support the positive effects of Hp+Hpx therapy in a manner that extends beyond right ventricular function. In the Berk-SS mice that received the Hp+Hpx therapy, we observed a rebalancing of cardiopulmonary proteins that are recognized to have critical roles in the development of oxidative stress, inflammation as well as PH and fibrosis.[31] Many of the changes in protein expression described in the categorical subsets (PH, inflammation and fibrosis) can be traced to oxidative stress. Thus, changes associated with Hp+Hpx treatment are congruent with the expected mechanism of action of scavenger proteins. In addition, our data demonstrates that the pattern of right ventricular protein expression corrects various dysfunctional complexes essential to the cardiac mitochondrial respiratory chain (e.g., Ndufs3, Ndufs6 and Bcs1l, Uqcr11) that contribute toward an improvement in ventricular function. At the metabolic level, changes observed in right ventricular mitochondrial respiratory proteins after Hp+Hpx therapy relate to a shift toward upregulation of unsaturated fatty acid biosynthesis. Unsaturated fatty acids are a more efficient source of energy compared to their saturated counterparts and appear to be more desirable than glycolysis in cardiomyocytes [31, 32].

In contrast to the shift toward protein downregulation associated with improved ventricular mechanical function, we observe a shift in upregulated proteins accompanying improvement in peripheral vascular resistance. We also note increased expression of the oxidative stress proteins Adam17, Sod1, Thioredoxin 2 (Txn2), and down regulation in Icam, Pdk1 and Fech. The protein subsets that impact PH suggest a link between lung oxidative stress, inflammation and fibrosis. All subsets show a shift toward upregulation, a likely divergence arising from cardiomyocyte vs. pulmonary vascular cell expression.

Summary and Conclusion

The findings presented above highlight the potential for Hp+Hpx therapy to attenuate SCD-PH progression. Nonetheless, some limitations to this study must be acknowledged. Clearly, mouse models of SCD-PH cannot fully reflect the severity of pulmonary vascular remodeling seen in human SCD-PH. However, pulmonary hemodynamics and right ventricular disease in our Berk-SS mouse model of sub-chronic hypoxia exposure do approximate human SCD cardiopulmonary dysfunction. Further, our model reveals therapeutic treatment effects attenuating Hb and heme exposure. Recall, this study utilizes whole lung tissue to evaluate multi-omic parameters, therefore data reflect effects on the entire tissue proteome and metabolome rather than that of lung vasculature, specifically. Regardless, multi-omic data does indicate metabolite and protein correlates to cardiopulmonary dysfunction and PH in our model.

Our data indicates further studies will be necessary to fully understand the relationships between cardiopulmonary function and the multi-omic associations observed in Berk-SS mice treated with Hp+Hpx. The precise mechanisms that drive Hb and heme driven right ventricular and pulmonary vascular pathophysiology in SCD-PH will need definition at the cellular level. Nevertheless, this applied preclinical evidence suggests further exploration of scavenger protein combinations provide a promising treatment strategy for SCD-associated cardiopulmonary complications.

Supplementary Material

Supplemental material for this article is available online.