Accelerated atherosclerosis in beta-thalassemia

Abstract

Children with beta-thalassemia (BT) present with an increase in carotid intima-medial thickness, an early sign suggestive of premature atherosclerosis. However, it is unknown if there is a direct relationship between BT and atherosclerotic disease. To evaluate this, wild-type (WT, littermates) and BT (Hbb^th3/+) mice, both male and female, were placed on a 3-mo high-fat diet with low-density lipoprotein receptor suppression via overexpression of proprotein convertase subtilisin/kexin type 9 (PCSK9) gain-of-function mutation (D377Y). Mechanistically, we hypothesize that heme-mediated oxidative stress creates a proatherogenic environment in BT because BT is a hemolytic anemia that has increased free heme and exhausted hemopexin, heme’s endogenous scavenger, in the vasculature. We evaluated the effect of hemopexin (HPX) therapy, mediated via an adeno-associated virus, to the progression of atherosclerosis in BT and a phenylhydrazine-induced model of intravascular hemolysis. In addition, we evaluated the effect of deferiprone (DFP)-mediated iron chelation in the progression of atherosclerosis in BT mice. Aortic en face and aortic root lesion area analysis revealed elevated plaque accumulation in both male and female BT mice compared with WT mice. Hemopexin therapy was able to decrease plaque accumulation in both BT mice and mice on our phenylhydrazine (PHZ)-induced model of hemolysis. DFP decreased atherosclerosis in BT mice but did not provide an additive benefit to HPX therapy. Our data demonstrate for the first time that the underlying pathophysiology of BT leads to accelerated atherosclerosis and shows that heme contributes to atherosclerotic plaque development in BT.

NEW & NOTEWORTHY This work definitively shows for the first time that beta-thalassemia leads to accelerated atherosclerosis. We demonstrated that intravascular hemolysis is a prominent feature in beta-thalassemia and the resulting increases in free heme are mechanistically relevant. Adeno-associated virus (AAV)-hemopexin therapy led to decreased free heme and atherosclerotic plaque area in both beta-thalassemia and phenylhydrazine-treated mice. Deferiprone-mediated iron chelation led to deceased plaque accumulation in beta-thalassemia mice but provided no additive benefit to hemopexin therapy.

INTRODUCTION

Beta-thalassemia (BT) is an autosomal recessive hemoglobinopathy that affects approximately 80 million people worldwide. Advances in iron chelation therapy and regular blood transfusion have extended the life of patients with BT by preventing iron toxicity and severe anemia (1–3), but the extent of the underlying vasculopathy in BT remains largely unknown (4). Recently, several clinical association studies have shown children with BT have increased carotid intimal-medial thickness, a sign of premature atherosclerosis (5–12). Patients with BT also have dyslipidemia characterized by elevated triglycerides and low- or high-density lipoprotein levels resulting in a high-atherogenic index (5, 6). However, it is not known if there is a direct relationship between BT and atherosclerotic disease nor is the mechanism responsible for this phenomenon known.

The potential pathological cause of accelerated atherosclerosis in BT is multifactorial and could be due to anemia, compensatory responses to the anemia, dyslipidemia, and/or a result of hemolysis of red blood cells. BT is characterized by impaired (B⁰) or reduced (B⁺) synthesis of the beta globin chain of hemoglobin which causes excess, unpaired alpha hemichromes that lead to erythrocyte instability (13, 14). Traditionally, the clinical categorization of BT is major, intermedia, minor, or more recently nontransfusion-dependent thalassemia or transfusion-dependent thalassemia. BT is a molecularly heterogenous disease whose severity is affected by the inherited HBB mutation (B⁰ or B⁺), coinheritance of genetic factors, and disease management (13). The hallmark of BT is ineffective erythropoiesis that is characterized by increased but ineffectual erythropoiesis that leads to erythroid progenitor precursors failing to mature, premature destruction of red blood cells, and subsequent anemia. As a result, this leads to increased compensatory mechanisms to improve anemia such as extramedullary hematopoiesis, increased gastrointestinal iron absorption, upregulation of erythropoietin, and decreased hepcidin expression (1, 13).

In BT, mature erythrocytes are microcytic, hypochromic, and abnormally shaped, and have decreased survival times (14, 15). A prominent feature of BT is intravascular hemolysis which exposes the vasculature to elevated levels of hemoglobin and free heme (15). Free heme can promote the generation of reactive oxygen species (ROS) directly through Fenton chemistry and indirectly downstream through Toll-like receptor 4 (TLR4) activation (16). Under normal physiological conditions, hemopexin (HPX) binds to heme and prevents heme’s prooxidant effects (17). HPX is a glycolytic, 57-kDa liver protein that binds free heme with high affinity on an equimolar basis and is cleared by the liver via CD91 (17). Heme oxygenases (HO-1, HO-2) subsequently degrade heme into iron, a porphyrin ring, carbon monoxide, and biliverdin (18). However, in humans with BT, HPX is constantly consumed by the elevated circulating levels of heme resulting in a twofold increase in free heme and a 10-fold decrease in HPX compared with controls (19). Elevated free heme and depleted HPX levels in BT are even more pronounced than those observed in sickle cell disease, a classic hemolytic anemia whose effects of free heme on cardiovascular disease are well known (20).

To our knowledge, our data reveal for the first time that BT definitively causes accelerated atherosclerosis in mice. Mechanistically, we show that heme-mediated oxidative stress that occurs because of intravascular hemolysis is critical to the progression of atherosclerosis. We accomplished this by decreasing free heme using an adeno-associated virus (AAV) vector to overexpress HPX in BT and confirming this effect in a complementary drug-induced (phenylhydrazine) model of hemolysis. To gain a better understanding of the effects of iron on the progression of atherosclerosis in BT, we administered deferiprone (DFP), an orally available iron chelator used in BT that binds to ferric iron (Fe³⁺), in our atherosclerotic model. Globally, we show that BT leads to accelerated atherosclerosis and demonstrate that free heme is critical to this phenotype.

MATERIAL AND METHODS

Animals

Hbb^th3/+ mice (Jax No. 000996) on a C57BL/6 background were used as a model of BT intermedia (21). Aged-matched C57BL/6 mice wild-type (WT) littermates were used as normal controls. Mouse colonies were established, maintained, and bred in-house at the Emory University Department of Animal Resources. Both male and female WT and BT mice (Hbb^th3/+) between 8 and 12 wk old were used in this study to determine the effect of BT on atherosclerosis. For our drug-induced model of intravascular hemolysis, mice received a weekly intraperitoneal injection of fresh phenylhydrazine (25 mg/kg body weight in 0.1 mL of phosphate-buffered saline) (PHZ, Sigma-Aldrich, Cat. No. P26252) (22, 23). PHZ is a widely used chemical model to induce intravascular hemolysis. For our iron chelation experiments, deferiprone (DFP, Sigma-Aldrich, Cat. No. Y0001976) was administered at 1.25 mg/mL in the drinking water. Deferiprone is an orally absorbable, water-soluble molecule that has a neutral charge, binds iron at a 3:1 molar ratio, and has been shown to effectively remove excess iron in BT mouse models (24). Mice were housed and cared for in agreement with the guidelines approved by the Emory University’s Institutional Animal Care and Use Committee.

Serum Assays

Following euthanasia, blood was obtained via cardiac puncture and centrifuged at 6,000 rpm for 15 min to isolate serum. HPX was measured in serum via a Mouse Hemopexin ELISA Kit (Abcam, Cat. No. ab157716). Free heme was obtained by centrifuging fresh serum in <3 kDa Amicon Ultra-0.5 Centrifugal Filter columns (Sigma-Aldrich, Cat. No. UFC500396) for 1.5 h at 10,000 rpm and measured using the QuantiChrom Heme Assay Kit (BioAssay Systems, Cat. No. DIHB-250). Both total iron (directly from serum) and nontransferrin-bound iron (after centrifuging serum in <3-kDa Amicon Ultra-0.5 Centrifugal Filter columns) were measured using an Iron Assay Kit (Sigma-Aldrich, Cat. No. MAK025). Hematocrit was measured via the microhematocrit method.

Models of Atherosclerosis

Susceptibility to atherosclerosis was induced by feeding mice a high-fat diet (HFD, Research Diets, Cat. No. D12336) and inducing low-density lipoprotein receptor suppression by administering gain-of-function proprotein convertase subtilisin-kexin type 9 (Vector Biolabs, AAV8-D377Y-mPCSK9) (25). Mice were retro-orbitally injected with 1E11 genomic copies of AAV8-D377Y-mPCSK9 and then placed on HFD for 12 wk. An AAV8-mouse HPX vector expression was generated from Vector Biosystems (AAV8-ApoE/hAAT1-mHPX/eGFP-WPRE). Hemopexin mice received 4E11 GC AAV-HPX that was coinjected with AAV8-D377Y-mPCSK9. A secondary confirmatory model of atherosclerosis was done by retroorbitally injecting mice with AAV8-D377Y-mPCSK9, subcutaneously implanting osmotic minipump (Azlet, Cat. No. 2004) delivering angiotensin II (Sigma-Aldrich, Cat. No. A9525) at a rate of 0.75 mg/kg per day, and feeding mice an HFD (Research Diets, Cat. No. D12336 containing 15% fat, 1.25% cholesterol) for 4 wk, as previously described (26).

Atherosclerotic End Points

Cholesterol.

Total plasma cholesterol, triglycerides, and high-density lipoprotein cholesterol (HDLc) were determined by enzymatic methods on the Beckman AU480 (Beckman Diagnostics, Brea, CA) automatic chemistry analyzer. Reagents, calibrators, and controls are obtained from Sekisui Diagnostics (Burlington, MA). LDL was calculated per the standard Friedewald formula.

Atherosclerotic lesion area.

Mice were euthanized by CO₂ inhalation at prescribed time points (1 or 3 mo). Heart and aorta were pressure perfused with 0.9% sodium chloride solution, followed by pressure fixation at ≈100 mmHg with a 4% formaldehyde solution. The hearts were embedded in paraffin and 5-μm-thick serial sections were prepared. Atherosclerosis was evaluated by measuring cross-sectional atherosclerotic lesion area in the aortic root and total atherosclerotic lesion area (%) relative to luminal area en face (analyzed via Image J) in the descending thoracic and abdominal aorta, as previously described (26). Histological analysis on the aortic root was used to evaluate lesion complexity through H&E and Mason’s trichrome (Sigma-Aldrich, Cat. No. HT15-1KT) staining. Trichrome sections were used to measure the necrotic core area, which was defined as the region within the aortic root plaque characterized by acellularity and subsequently quantified using ImageJ software (27).

Hydrogen Peroxide Detection

Extracellular H₂O₂ from the aortas was measured using the Amplex Red assay kit (Invitrogen, Cat. No. A22188) as previously described (28) per recommendations from the American Heart Association (29). When exposed to hydrogen peroxide in the presence of horseradish peroxidase, Amplex Red yields the fluorescent product resorufin. Fluorescence was detected on a fluorescence plate reader (Ex/Em = 530/580 nm) with background fluorescence subtracted and normalized to dry tissue weight.

Polymerase Chain Reaction

Total RNA was extracted from homogenized aortas using the RNeasy Mini Kit (Qiagen, Cat. No. 74004) according to the manufacturer’s instructions. RNA was reverse transcribed into cDNA and purified with QiaQuick (Qiagen, Cat. No. 28104). Gene expression was quantified on an Applied Biosystems StepOnePlus Real-Time PCR System. IL-6, TNFα, and 18S Quantitect Primer Assays (Qiagen) were used. Gene expression was normalized to housekeeping gene 18S and expressed relative to the average untreated wildtype (AA) value using the comparative cycle threshold (Ct) method with the formula: Fold change = 2^−ΔΔCt.

Statistical Analysis

Atherosclerotic lesion area in the aortic root and aortic enface preparations was expressed as means ± SE, tested for normality, and parametric test was selected for analysis. All statistical analysis was carried out using Prism (GraphPad Software, La Jolla, CA). Differences between the two experimental groups were evaluated using unpaired Student two-tailed t test with no assumption of equal variance. Two-way ANOVA test was used for comparing more than two groups of data with Tukey’s post hoc test. Corrections were made for multiple comparisons. Normality was tested using Kolmogorov–Smirnov test.

RESULTS

Beta-Thalassemia’s Underlying Pathophysiology Leads to Accelerated Atherosclerosis

Accelerated atherosclerosis.

We observed an increase in plaque area (%) relative to luminal area via aortic en face analysis in BT in both sexes compared with their respective WT counterparts (WT male, 28.1 ± 2.4% vs. BT male, 45.5 ± 2.4%, P < 0.001 and WT female, 10.6 ± 2.9% vs. BT female, 23.3 ± 1.5%, P < 0.001, Fig. 1A). Atherosclerosis was higher in both BT and WT male mice compared with their respective female counterparts. Atherosclerosis in BT male mice was found throughout the entirety of the aorta, while in WT male mice, aortic plaque accumulation was largely regionalized to the aortic arch. There was no significant difference in the distribution (thoracic or abdominal) of atherosclerosis between WT and BT female mice. Likewise, we saw an increase in atherosclerotic lesion area via aortic root analysis in both BT sexes compared with their respective WT counterparts (WT male, 304.6 ± 30.9 µm² × 10³ vs. BT male, 444.7 ± 38.1 µm² × 10³, P < 0.001, and WT female, 136.4 ± 13.6 µm² × 10³ vs. BT female, 241.7 ± 22.1 µm² × 10³, P < 0.05, Fig. 1B). As shown in Fig. 1C, WT mice had more total cholesterol and low-density lipoprotein (LDL) than BT mice (total cholesterol WT, 690.5 ± 32.2 mg/dL vs. BT, 526.8 ± 51.8 mg/dL, P < 0.001 and LDL: WT, 690.5 ± 32.2 mg/dL vs. BT, 526.8 ± 51.8 mg/dL, P < 0.001). BT mice had more high-density lipoprotein than WT mice, but no significant differences were seen in triglycerides.

Figure 1.

Beta thalassemia (BT) is a hemolytic anemia that causes accelerated atherosclerosis. A: representative aortic en face analysis of plaque accumulation (%) of total luminal area (n = 5–11). B: representative hematoxylin and eosin (H&E) staining of the aortic root and analysis of lesion area (µm² × 10³; n = 5–11). C: serum total cholesterol, triglycerides, LDL, and HDL (n = 10–12). Values represent means ± SE. *P < 0.05; **P < 0.01; ***P < 0.001.

To confirm these findings, we also performed studies in an aggressive short-term model of atherosclerosis (AAV-PCSK9 gain of function + angiotensin II infusion + HFD for 1 mo). Similar, to our main atherosclerotic model (AAV-PCSK9 gain of function + HFD for 3 mo), both BT male and female mice developed more severe atherosclerosis via aortic en face analysis compared with their respective WT counterparts (Fig. 2A). In fact, atherosclerosis and subsequent aneurysmal formation was so severe that it led to a 56.6% mortality rate in our BT mice compared with 12.5% in our WT mice (Fig. 2, B and C). Because high mortality and aneurysm formation not being a common phenotype in BT, we focused our studies only on the first model of atherosclerosis shown in Fig. 1

Figure 2.

Secondary model confirms accelerated atherosclerosis is present in beta thalassemia (BT). Short-term model of accelerated atherosclerosis with angiotensin II infusion, 1 mo high-fat diet (HFD), and proprotein convertase subtilisin/kexin type 9 (PCSK9) gain of function confirms accelerated plaque accumulation in BT mice. A: representative aortic en face analysis (n = 4). B: representative of severe aneurysm formation. C: percent survival in wild-type (WT) and BT mice. Values represent means ± SE. *P < 0.05; **P < 0.01.

BT is a hemolytic anemia with intravascular hemolysis.

At baseline, serum hemopexin levels were severely depleted in BT mice (WT, 6.2 ± 0.2 mg/mL vs. BT, 0.5 ± 0.1 mg/mL, P < 0.001, Fig. 3). In addition, BT mice had an increase in free heme compared with their wild-type littermates (WT, 0.06 ± 0.02 µM vs. BT, 0.2 ± 0.04 µM, P < 0.01, Fig. 3).

Figure 3.

Baseline hemopexin and free heme levels in beta thalassemia (BT) mice. A: baseline (normal chow diet) serum hemopexin (HPX) levels (n = 6). B: baseline (normal chow diet) serum free heme (n = 6). Values represent means ± SE. **P < 0.01; ***P < 0.001.

Hemopexin Therapy Decreased Atherosclerosis in BT Mice

To determine whether free heme was contributing to increased atherosclerosis in BT, we studied the effect of hemopexin (HPX), free heme’s endogenous scavenger, on the development of atherosclerosis. HPX therapy led to a significant decrease in plaque accumulation in BT mice as assessed by aortic en face analysis (BT, 45.5 ± 2.4% vs. BT + HPX, 31.7 ± 2.5%, P < 0.01, Fig. 4A). In addition, we saw decreased plaque lesion area via aortic root analysis in our BT mice with HPX therapy (BT, 444.7 ± 38.1 µm² × 10³ vs. BT + HPX, 318.8 ± 28.6 µm² × 10³, P < 0.05, Fig. 4B). Furthermore, aortic root (Mason’s trichrome) histology indicated hemopexin therapy decreased necrotic area (BT, 193.4 ± 8.5 µm² × 10³ vs. BT + HPX, 107.3 ± 5.9 µm² × 10³, P < 0.05, Fig. 4B) and collagen area (BT, 186.6 ± 11.1 µm² × 10³ vs. BT + HPX, 85.4 ± 6.4 µm² × 10³, P < 0.01, Fig. 4B) in BT mice compared with WT mice. HPX therapy functionally decreased serum-free heme levels in the BT mice (BT, 0.2 ± 0.02 μM vs. 0.1 ± 0.02 μM, P < < 0.05, Fig. 4C), but did not significantly increase hemopexin levels in BT mice (BT, 0.8 ± 0.1 mg/mL vs. BT HPX 1.9 ± 0.3 mg/mL, P < = 0.07, Fig. 4D), Furthermore, hemopexin therapy decreased aortic hydrogen peroxide levels (Fig. 4E) as well as aortic IL-6, TNFα and HO-1 expression measured via RT-qPCR. In addition, we saw no change in atherosclerotic plaque in WT mice treated with the control vector (Fig. 5). These data demonstrate that hemopexin therapy decreases atherosclerotic plaque accumulation, lesion complexity, ROS levels, and inflammatory gene expression indicating that free heme plays a central, causal role in the development of atherosclerosis in BT.

Figure 4.

Adeno-associated virus (AAV) overexpression of hemopexin leads to decreased plaque accumulation in beta thalassemia (BT) mice. A: representative aortic en face images and analysis in wild-type (WT), BT, and BT + serum hemopexin (HPX) mice (n = 6–10). B: representative hematoxylin and eosin (H&E) and trichrome staining of the aortic root (n = 6–10). Lesion, necrotic, and collagen area (µm² × 10³) were analyzed. C: serum free heme levels at 1 mo (n = 5–7). D: serum hemopexin (HPX) levels (n = 6) at 3 mo. E: aortic hydrogen peroxide levels at 1 mo (n = 5–7). F: aortic gene expression of IL-6, HOX-1, TNF-α via RT-qPCR at 1 mo (n = 5–7). Values represent means ± SE. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 5.

Control serum hemopexin (HPX) vector does not affect atherosclerotic plaque accumulation. A: representative aortic en face images analysis in wild-type (WT) and WT + HPX mice. B: means data (n = 4). Values represent means ± SE.

Hemopexin Therapy Decreases Atherosclerosis in a Phenylhydrazine-Induced Model of Intravascular Hemolysis

BT is a complex disease that has systemic effects including hemolysis (4). To understand the specific effects of intravascular hemolysis, we employed a complimentary drug-induced model of hemolysis using phenylhydrazine (PHZ) (22). Hematocrit was significantly reduced in our PHZ groups compared with our WT littermate control (WT, 44.8 ± 0.6% vs. PHZ: 31.2 ± 1.8%, P < 0.001, or PHZ + HPX, 33.2 ± 0.6%, P < 0.001, Fig. 6A) at a level that was similar to what was observed in the BT mice (32.7 ± 0.4%). PHZ-treated mice also developed significant splenomegaly similar to what is typically observed in BT mice because of increased extramedullary hematopoiesis and destruction of red blood cells (Fig. 6B) (21). Aortic en face analysis showed increased plaque accumulation in the PHZ group compared with WT controls (WT, 20.3 ± 2.2% vs. PHZ: 41.9 ± 4.9%, P < 0.001, Fig. 6C). Aortic root total lesion area was increased in the PHZ group compared with WT controls (WT, 274.6 ± 20.9 µm² × 10³ vs. PHZ: 470.0 ± 58.7 µm² × 10³, P < 0.01, Fig. 6, D and E). Hemopexin therapy in PHZ-treated mice led to a decrease in plaque accumulation as assayed via aortic enface analysis (PHZ: 41.9 ± 4.9% vs. PHZ + HPX, 23.8 ± 3.1%, P < 0.05, Fig. 6C) and a decrease in plaque area as analyzed via total lesion area of the aortic root (PHZ: 470.0 ± 58.7 µm² × 10³ vs. PHZ+ HPX, 292.5 ± 28.0 µm² × 10³, P < 0.05, Fig. 6, D and E). Free heme levels were increased in PHZ-treated mice compared with WT (WT, 0.11 ± 0.01 μM vs. WT + PHZ: 0.22 ± 0.02 μM, P < 0.001, Fig. 6F), whereas PHZ mice treated with HPX had decreased free heme levels when compared with PHZ mice (WT + PHZ + HPX, 0.11 ± 0.03 μM vs. WT + PHZ: 0.22 ± 0.02 μM, P < <0.01, Fig. 6F). Similar to BT findings, HPX levels were decreased in PHZ mice compared with WT (WT, 6.3 ± 0.6 mg/dL vs. WT ± PHZ: 0.3 ± 0.1 mg/dL, P < 0.001, Fig. 6G) and hemopexin level were not significantly increased after hemopexin treatment. These data demonstrate that drug-induced intravascular hemolysis resulted in an increase in atherosclerosis that was rescued by hemopexin therapy. These finding support the concept that increased free heme occurring as a result of hemolysis is responsible for the increase in atherosclerosis that we observed in BT mice.

Figure 6.

Phenylhydrazine (PHZ)-induced hemolysis increases atherosclerosis and is reversed by serum hemopexin (HPX) expression. A: serum hematocrit at 3 mo (n = 5–7). B: spleen weight at 3 mo (n = 5–7). C: representative aortic en face plaque accumulation in wild-type (WT), PHZ, and PHZ + HPX mice (n = 6–12). D: representative hematoxylin and eosin and trichrome staining of the aortic root (n = 6–10). E: quantitative analysis of aortic root lesion area (n = 6–11). F: serum free heme levels at 3 mo (n = 5–7). G: HPX levels (n = 5–6). Values represent means ± SE. *P < 0.05; **P < 0.01; ***P < 0.001.

Deferiprone Reduces Atherosclerosis in BT

Iron chelation therapy is a mainstay in BT (30). BT led to an increase in iron overload because of constant blood transfusion and/or increased iron absorption (31, 32). Deferiprone (DFP) is an iron chelator that removes iron from ferritin and hemosiderin in intracellular stores, the primary stores of scavenged “free” iron in the body (24). To determine if traditional iron chelation therapy is effective in ameliorating the proatherogenic effects of BT, we studied four treatment groups (WT, BT, BT + DFP, BT + DFP + HPX). BT+ DFP and BT + DFP + HPX mice had decreased plaque area via aortic en face analysis compared with untreated BT mice (BT, 50.12 ± 1.78% vs. BT + DFP, 35.2 ± 2.0%, P < 0.001, or BT + DFP + HPX, 34.4 ± 1.9%, P < 0.001, Fig. 7A). BT+ DFP and BT + DFP + HPX mice had decreased plaque area via aortic root lesion area analysis compared with BT (BT, 349.8 ± 33.5 µm² × 10³ vs. BT + DFP, 216.5 ± 8.9 µm² × 10³, P < 0.01, or BT + DFP + HPX, 224.1 ± 14.52 µm² × 10³, P < 0.01, Fig. 7B). Relative to baseline BT mice, BT + DFP + HPX did not provide an additive benefit compared with only BT + DFP in either aortic enface or aortic root analyses (Fig. 7, A and B). Serum iron levels were decreased in BT + DFP compared with BT but not further decreased in BT + DFP + HPX (Fig. 7C). Non-transferrin and nonhemopexin-bound iron levels indicated successful chelation of iron (Fig. 7D).

Figure 7.

Deferiprone (DFP) decreases atherosclerotic plaque accumulation in beta thalassemia (BT) mice but does not provide an additive benefit over hemopexin therapy. A: representative aortic en face plaque accumulation in WT, BT, and BT + DFP, BT + DFP + HPX mice (n = 5–10). B: representative hematoxylin and eosin and trichrome staining of the aortic root (n = 5–10). C: total serum iron at 3 mo. D: non-transferrin-bound iron at 3 mo. Values represent means ± SE. *P < 0.05; **P < 0.01; ***P < 0.001.

DISCUSSION

Patients with BT have a variety of poorly understood and complex cardiovascular complications such as ineffective erythropoiesis, hypercoagulability, leg ulcers, altered lipid profiles, and impaired flow-mediated dilation of brachial arteries (1, 13). To our knowledge, 12 clinical association studies from 2009 to 2022 have shown an increase in carotid-intimal media thickness in both BT major and BT intermedia with and without chelation history (5–10, 12, 33–37). However, it is not known if there is a direct relationship between BT and atherosclerotic disease. From a cardiovascular risk perspective, patients with BT traditionally have hypocholesteremia with low LDL levels (38, 39), lower blood viscosity due to decreased hematocrit (40), and a lower incidence of hypertension (41). On the other hand, patients with BT have increased iron levels (from both transfusion and increased gastrointestinal iron absorption) (32, 42), and increased triglycerides with low HDL (5, 10). Traditionally, myocardial infarction is not common in BT, but in 2004, the first incidence of myocardial infarction was reported in BT with subsequent sporadic case reports in 2009 and 2023 (43–45). A recent study by Lee et al. (46) demonstrated an increased rate of MI and stroke in transfusion-dependent patients with beta-thalassemia with OR of 3.02 for MI and 3.20 for stroke. In 2016, Hahalis et al. (7) showed that coronary artery calcium score was not significantly changed, but carotid intimal thickness was elevated suggesting a disparate rate of atherosclerosis progression between carotid and coronary arteries among patients with thalassemia. One explanation for the evolving data suggesting increased cardiovascular events is that the life expectancy is increasing in individuals with BT resulting in more cardiovascular events. In 1964, the average life expectancy of patients with BT was 16 yr, which increased to 40–49 yr by 2011, with 37%, and 89% respective survival at average life expectancy (47). As a result, BT disease presentation is a changing landscape that has more comorbidities because of an aging population (47). Ultimately, the true incidence of atherosclerosis in BT is difficult to determine because of numerous factors such as survivor bias, iron chelation history, and small clinical studies. It is clear that BT has potentially very complex cardiovascular effects, but it is unclear if the underlying pathophysiology of BT has a direct, causal role in the pathogenesis of atherosclerosis. The present study supports the hypothesis that BT causes an increase in the development of atherosclerosis.

Our results are striking in that in a mouse model of BT intermedia, we saw markedly accelerated atherosclerosis. We found increased atherosclerosis in both male and female BT mice compared with WT mice in two separate models of atherosclerosis. BT mice had more than a 1.5-fold increase (compared with WT) in plaque area measured via aortic en face analysis. In addition, the plaque accumulation was robust and present throughout the entire aorta. Unsurprisingly, atherosclerosis was more extensive in male mice, a common finding others have reported which may be due to innate differences between males and females or due to decreased PCSK9 expression in female mice (48). BT mice developed more atherosclerosis despite lower levels of total cholesterol and LDL. We did observe higher HDL levels in BT mice. The etiology of this is unclear and higher HDL levels would generally be considered to be atheroprotective, opposite of our observation of increased atherosclerosis in BT mice. Importantly, we found that at baseline, the BT mice exhibited an increase in free heme and a decrease in HPX which is similar to observations in humans with BT (19).

In our experiments, HPX therapy led to decreased plaque accumulation in both BT and PHZ induced model of intravascular hemolysis. Typically, hemopexin is administered with intraperitoneal injections (49–51), but this is a costly and time-consuming process. Therefore, we employed an AAV-mediated approach to overexpress HPX and were able to show that hemopexin therapy reduced atherosclerosis in BT. BT is a multifaceted disease that could have several causes for atherosclerosis such as anemia or hypercoagulability. Therefore, we used a PHZ drug induced model of hemolysis to mechanistically compare the results of HPX therapy in BT. Both PHZ and BT had accelerated atherosclerosis that could be reversed by HPX therapy. Both, PHZ and BT mice demonstrated decreased free heme levels but had no statistically significant change in HPX levels. This may be because PHZ and BT mice have very high levels of free heme that consumed all of the overexpressed hemopexin. Collectively, these studies demonstrate that elevated levels of free heme are the cause of accelerated atherosclerosis in BT.

Previous work supports a proatherogenic role for free heme but the in vivo compensatory mechanisms are nuanced and incompletely understood (22, 52, 53). Blood viscosity levels, free heme-mediated ROS, detoxifying effects of heme-oxygenase, iron metabolism, compensatory erythropoiesis, and lipid metabolism play a complex role in hemolytic anemias (13, 23, 53). Heme has been shown to increase endothelial cell permeability, SMC proliferation, and LDL oxidization (52). Furthermore, heme and other erythrocyte products are found in advanced human atherosclerotic plaques (53). Free heme is catabolized by heme oxygenase in the liver to both pro- (iron) and anti-inflammatory (CO, bilirubin) byproducts. Heme oxygenase knockout mice show increased atherosclerosis (54), but paradoxically, hemin administration (25 mg/kg, 4 times/wk for 6 wk) was previously shown to decrease plaque accumulation by increasing heme oxygenase (55). Likewise, PHZ models of intravascular hemolysis have shown both decreased and increased atherosclerosis (22, 23, 25, 26). The precise reason for these discrepant results is unclear but may reflect systemic toxicity from higher doses used in some studies. Blood viscosity levels, free heme-mediated ROS, detoxifying effects of heme-oxygenase, iron metabolism, compensatory erythropoiesis, and lipid metabolism play a complex compensatory mechanism in hemolytic anemias. SCD mice, another hemolytic hemoglobinopathy, show a paradoxical decrease in plaque accumulation (56) and autoimmune hemolytic models of anemias have been shown to decrease plaque accumulation, but increase plaque vulnerability (57). Further investigation in these models is needed to understand potential in vivo compensatory changes.

Currently, there are no available FDA approved therapies that target hemolysis-mediated effects, but over the past decade, HPX therapy in mice has been used to improve cardiovascular function in hemolytic anemia such as BT and SCD (19, 51, 58–60). Vinchi et al. (51) showed that hemopexin therapy decreases vascular adhesion molecules in the aorta and improves cardiac function in BT and SCD mice. In the context of atherosclerosis, Mehta et al. (61) showed that double knockout HPX^−/− and ApoE^−/− mice developed more atherosclerosis, oxidative stress, and impaired HDL function. Other novel approaches to scavenge free heme in the vasculature such as apohemoglobin-haptoglobin are currently under investigation in BT (62). CSL889, a plasma-derived hemopexin, is in a phase-1 clinical trial (www.clinicaltrials.gov identifier NCT04285827) based on preclinical evidence showing improved endothelial activation (externalization of P-selectin and von Willebrand factor, and expression of IL-8, VCAM-1, and heme oxygenase-1) and decreased heme-mediated vasoocclusion in SCD (49).

Iron chelation therapy is a mainstay therapy in BT due to high-iron levels caused by regular blood transfusions and increased iron absorption (42). The role that iron plays in atherosclerosis is controversial as both iron deficiency and overload have been associated with atherosclerosis (53, 63, 64). Recently, nontransferrin-bound iron (NTBI) and iron tissue accumulation have been highlighted as the most important contributors to a proinflammatory phenotype that generates ROS via Fenton chemistry (63). Here, we have shown that DFP is able to decrease atherosclerosis in BT mice. HPX overexpression did not provide an additive benefit suggesting that HPX and iron chelation affect the same pathways and that the effects of free heme are mediated through local iron deposition. Furthermore, this highlights the fact that iron chelation might provide additional anti-atherogenic benefits in BT. Iron chelation with HPX therapy showed a significant decrease in NTBI, but not in total iron. One possible explanation is that HPX therapy leads to “artificially” increased hemopexin-bound iron in the circulation which blocks heme-mediated ROS formation. Iron chelation therapy decreased serum iron levels and plaque accumulation but did not provide an additive benefit to hemopexin therapy.

Potential limitations of this study include the use of a mouse model which does not completely recapitulate atherosclerotic disease that occurs over decades in humans. In addition, DFP was delivered in the drinking water as opposed to being administered via daily gavage. However, it is important to note that efficacy was confirmed by a uniform decrease in serum iron in the treated animals (Fig. 4C). Similarly, although we studied the efficacy of AAV in preliminary studies, we did not measure AAV transduction efficiency in these animals. However, all animals treated with the PCSK9 vector had similar responses in terms of serum cholesterol levels and all animals treated with the HPX vector demonstrated similar reductions in free heme suggesting uniform efficacy.

In conclusion, our results demonstrate, for the first time, that BT mice are more susceptible to atherosclerosis and suggest that oxidative stress resulting from increased free heme in the circulation is a primary mechanism for increased atherosclerosis.

DATA AVAILABILITY

Data are available upon reasonable request.

GRANTS

This work was funded by National Institutes of Health Grants R01HL131414 (to W.R.T and D.R.A), T32GM00860 (to J.H.), and T32HL007745 (to J.R.W.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.H., D.R.A., and W.R.T. conceived and designed research; J.H., H.S., G.J., and S.G. performed experiments; J.H., C.V.L., and J.R.W. analyzed data; J.H., H.S., G.J., C.R.N., J.R.W., D.R.A., and W.R.T. interpreted results of experiments; J.H., G.J., and C.R.N. prepared figures; J.H. drafted manuscript; J.H., H.S., C.V.L., C.R.N., J.R.W., D.R.A., and W.R.T. edited and revised manuscript; J.H., H.S., G.J., C.V.L., C.R.N., D.R.A., and W.R.T. approved final version of manuscript.