Skip to main content
NCBI home page
Blood logoLink to Blood
. 2026 Jul 1;148(16):2072–2082. doi: 10.1182/blood.2025032905

Arterial iron regulates vasodilation during anemia via endothelial holo-α-globin

Luke S Dunaway 1,2, Shruthi Nyshadham 1, Skylar A Loeb 1,2, Brooke L O’Donnell 1, Nasim Abib 1, Wyatt J Schug 1,2, Zuzanna J Juśkiewicz 1,2, Melissa A Luse 1,2, Macy E Stahl 3, Timothy M Sveeggen 4, Pooneh Bagher 4, Jason D Allen 1,3, Stefano Toldo 1,5, Adam N Goldfarb 6, Brant E Isakson 1,2,∗
PMCID: PMC13675883  PMID: 42371808

Key Points

  • •

    Iron regulation varies across the vascular endothelium, with resistance arteries having the lowest iron stores.

  • •

    Arterial iron deficiency and the subsequent loss of α-globin is required for increased NO signaling in IDA.

Visual Abstract

graphic file with name BLOOD_BLD-2025-032905-ga1.webp

Open in a new tab

Abstract

Iron deficiency is a highly prevalent nutrient deficiency and the most common cause of anemia. Although iron deficiency exacerbates cardiovascular disease, the direct impact of iron deficiency on the vasculature remains unstudied. We assessed iron levels across the vascular endothelium in mouse and human endothelial cells and found that resistance artery endothelial cells have the lowest iron stores, suggesting they may be especially affected by iron deficiency. Anemia has previously been shown to increase arterial nitric oxide (NO) signaling in patients, and we have previously shown endothelial α-globin scavenges NO in the resistance artery endothelium. We hypothesized that iron regulates vascular function through regulation of endothelial α-globin. To test this, we used a mouse model of iron deficiency anemia (IDA). In female mice, IDA increased NO signaling, which was rescued to control levels by repletion of vascular iron with iron dextran. Despite being similarly anemic and having a similar reduction in α-globin protein, there were no changes in NO signaling across groups in male mice. We further measured whether α-globin was in its heme-bound (holo-α-globin) or heme-free (apo-α-globin) state and found males did not fully lose its functional, heme-bound α-globin. Using endothelial-specific α-globin knockout mice, we show loss of endothelial α-globin is necessary for increased NO signaling in IDA and for the rescue of NO signaling by iron dextran in female mice. Altogether, the data presented here demonstrate iron is a determinant of endothelial identity and modulates endothelial NO signaling through the regulation of α-globin.

Are the hemodynamic changes observed in iron deficiency anemia solely due to reduced red cell hemoglobin and mass, or does iron insufficiency directly affect the vasculature as well? Dunaway et al reveal that it is the latter. They provide compelling genetic evidence from murine models that endothelial iron deficiency directly alters vascular tone through reduction in α-hemoglobin with consequent increases in nitric oxide signaling and vasodilation of resistance arteries. Intriguingly, this phenomenon is most pronounced in females.

Introduction

Iron deficiency is a common nutrient deficiency and is associated with chronic diseases such as heart failure,1 chronic kidney disease,2 and other diseases with chronic inflammation.3 Each of these diseases is also associated with endothelial dysfunction. The endothelium regulates diverse processes (eg, vascular resistance, tissue perfusion, and immune cell trafficking), thereby requiring specialization of endothelial cells (ECs) along the vascular tree. Because iron is an important cofactor for proteins involved in transcriptional regulation, metabolism, and redox signaling, iron likely plays an important role in endothelial biology.4 However, the role of iron on endothelial phenotypes and function remains understudied.

Perhaps the most central endothelial function is the regulation of tissue perfusion. ECs modulate hemodynamics to ensure adequate nutrient and oxygen delivery to tissues. Previous studies investigating hemodynamic consequences in iron deficiency has focused on the impact of anemia and left the direct consequences of vascular iron deficiency unexplored. Anemia has been previously shown to increase nitric oxide (NO) signaling.5 This has been attributed to changes in blood hemoglobin because hemoglobin is a potent NO scavenger, but animal models demonstrate that erythrocytes minimally affect arterial NO signaling in perfused preparations, suggesting that the decreased hematocrit in anemia may be insufficient to increase NO signaling.6,7

We and others have previously demonstrated the alpha chain of hemoglobin (α-globin) is expressed in the endothelium of resistance arteries, the arteries that control tissue perfusion.8, 9, 10, 11 In the endothelium, α-globin binds directly to endothelial NO synthase (eNOS) and scavenges NO. Because hemoglobin is tightly regulated by systemic iron homeostasis in erythrocytes, we hypothesize that vascular iron deficiency downregulates α-globin, thereby contributing to the increased NO signaling in iron deficiency anemia (IDA).

Methods

Cell culture models

To deplete iron, human coronary artery ECs (CC-2585; Lonza, Basel, Switzerland) and human dermal lymphatic ECs (C-12217; PromoCell, Heidelberg, Germany) were plated at a seeding density of 167 × 103 cells per well in a 6-well plate. After 48 hours, cells had reached confluence and were treated with 100μM deferoxamine (DFO; MedChemExpress, Monmouth Junction, NJ) or vehicle control (diH2O) for 24 hours. To model resistance arteries, ECs were cultured following the vascular cell coculture model as previously reported.12,13 RNA isolation, sequencing, and transcriptomic analysis were performed as described in the supplemental Methods, available on the Blood website.

Model of IDA

All experiments were approved by the University of Virginia Animal Care and Use Committee and followed the National Institutes of Health guidelines for the care and use of laboratory animals. C57BL/6 mice were bred in-house from mice purchased from Taconic Biosciences. Endothelial Hba1 knockout (Cdh5-CreERT2+/− Hba1fl/fl) mice were generated as previously reported.10 Cre+ and Cre− littermates received intraperitoneal (IP) tamoxifen injections (100 μL at 10 μg/mL in peanut oil), which were administered for 10 days starting at 6 weeks of age. To model IDA, male and female mice were fed an iron-deficient diet (2-6 ppm; TD.80396; Inotiv, West Layfayette, IN) beginning at 3 weeks of age. Phlebotomies of 10% blood volume were performed at 9 and 11 weeks to drive the progression of anemia. A subset of mice (IDA + iron dextran [FeDex]) received an IP injection of FeDex (20 mg/kg; Sigma, St. Louis, MO). Control mice were fed a nutrient-matched control diet (48 ppm; TD.220330; Inotiv) and received an IP injection of saline (0.9% NaCl; 5 mL/kg) at 12 weeks. All experiments were performed 7 to 8 days after injection. Body composition was measured by EchoMRI body composition analyzer (EchoMRI, Houston, TX). Terminal blood collections were taken from the heart. Complete blood counts were analyzed on a VETSCAN HM5 hematology analyzer (Abaxis, Union City, CA). Liver iron was assessed as described in the supplemental Methods.

Laser speckle contrast imaging (LSCI)

To assess NO signaling, blood flow was assessed with PeriCam PSI (PeriMed, Järfälla, Sweden), similar to previously published methods.14,15 The mice were anesthetized with 4% isoflurane delivered with room air, then reduced to 1.5%. Core body temperature was maintained between 37.0°C to 37.4°C. After at least 15 minutes of equilibration, blood flow was measured in both hind paws of the mice, then saline (0.9% NaCl; 5 mL/kg) was injected IP as vehicle control. After 15 minutes, mice were injected IP with nitro-l-arginine methyl ester hydrochloride (L-NAME; 100 mg/kg). The percent change was calculated as follows: (postinjection blood flow – preinjection blood flow)/preinjection blood flow. The magnitude of response to L-NAME has been used as a measure of NO signaling in vivo in mice and humans.5,16

Results

Iron regulates endothelial identity

To better understand how iron deficiency affects endothelial biology, we investigated how iron is distributed across the microvascular endothelium. We performed pseudobulk analysis of a previously published single-cell RNA sequencing data set of human ECs17 (Figure 1A; supplemental Tables 1-8). We found both ferritin light chain (FTL) and ferritin heavy chain (FTH; FTH1) expression was highest in lymphatic ECs and lowest in arterial and capillary ECs (Figure 1B). Using a single-cell RNA sequencing data set of mesenteric and adipose ECs previously published by our group,18 we found similar results, with Fth1 being most highly expressed in lymphatic ECs and least in arterial ECs regardless of the tissue (Figure 1C; supplemental Tables 9-20). To investigate labile iron in this data set, we analyzed expression of transferrin receptor 1 (Tfrc), which is negatively regulated by labile iron.19 We found Tfrc expression to be highest in arterial ECs and lowest in lymphatic ECs regardless of tissue (Figure 1D). Labile iron was directly assessed by flow cytometry of freshly isolated, FerroOrange-stained mesenteric ECs, which demonstrated labile iron to be significantly higher in lymphatic ECs than blood ECs from the same mice (Figure 1E).

Figure 1.

Figure 1.

Open in a new tab

In the microvasculature, the arterial endothelium is iron depleted. (A) UMAP and principal component analysis plots of previously published single-cell RNA sequencing (scRNA-seq) of human ECs reanalyzed with pseudobulk analysis. Statistical analysis was performed with DESeq2. (B) Expression of FTH1 and FTL in previously published scRNA-seq data of human ECs from males (blue) and females (pink). (C-D) Ferritin heavy chain (Fth1) and transferrin receptor (Tfrc) expression in previously published scRNA-seq data of ECs isolated from adipose and mesentery. ∗P < .0001 vs artery; #P < .0001 vs capillary; &P < .0001 vs vein. Statistical analysis was performed with Wilcoxon rank-sum test. (E) Mesenteric endothelial labile iron assessed with FerroOrange by flow cytometry. Statistical analysis was performed with paired t test. (F-J) Human coronary artery ECs and human dermal lymphatic ECs were treated with DFO or vehicle control for 24 hours; labile iron was assessed with FerroOrange by flow cytometry (F) and microscopy (G; scale bar, 50 μm); and FTH was assessed by western blot (H). Statistical analysis was performed with 2-way analysis of variance (ANOVA) and Holms-Šídák post hoc analysis. (I-J) The expression of SOX17, PROX1, and markers of arterial and lymphatic EC phenotypes were assessed by bulk RNA sequencing. Statistical analysis was performed with DESeq2. ∗P < .05. AEC, arterial ECs; Avg Exp, average expression; BEC, blood ECs; Con, control; LEC, lymphatic ECs; MFI, mean fluorescence intensity; PC, principal component; Pct Exp, percent expressed; UMAP, uniform manifold and projection.

Similar results were observed in human arterial and lymphatic ECs in vitro. TFRC expression was significantly lower in lymphatic ECs, and this difference was eliminated by iron chelation with DFO, further suggesting the differences in Tfrc expression in vivo are likely due to differences in labile iron (supplemental Figure 1A). Direct assessment of labile iron by FerroOrange further demonstrated arterial ECs have significantly less labile iron than lymphatic ECs. DFO treatment effectively chelated labile iron in both cell types (Figure 1F-G). Ferritin expression was assessed by western blot and demonstrated FTH, but not FTL, to be significantly higher in lymphatic ECs than arterial ECs. As expected, the expression of both was decreased by DFO (Figure 1H; supplemental Figure 1B). These data demonstrate that iron is differentially regulated among endothelial subtypes and that these differences are maintained in vitro.

To better understand the impact and importance of differential iron regulation among EC subtypes, we performed bulk RNA sequencing on human artery ECs and human lymphatic ECs in the presence or absence of DFO (Figure 1I-J; supplemental Tables 21-25). We used this data set and the single-cell transcriptomic data sets to assess the expression of iron-dependent genes across endothelial subtypes. We found the differential regulation of iron across endothelial subtypes does not correlate with the expression of iron-dependent genes (supplemental Figure 2).

Therefore, we investigated whether the low iron in the arterial endothelium is related to arterial identity. To test this, we assessed transcriptomic changes in human artery ECs and human lymphatic ECs with and without iron chelation (Figure 1I-J). SOX17 and PROX1 are important transcription factors that promote arterial and lymphatic EC identity, respectively. We found iron chelation upregulates SOX17 expression in both EC subtypes, whereas PROX1 is minimally expressed in arterial ECs and is downregulated in lymphatic ECs (Figure 1I). Examining the expression of genes that have been previously used to assess endothelial identity,20 we found most arterial markers to be upregulated by iron chelation in both cell types and lymphatic EC markers to be downregulated in lymphatic ECs (Figure 1J). These data demonstrate the importance of low endothelial iron for arterial EC identity in the microvasculature.

Resistance artery ECs have low intracellular iron

Microvascular artery ECs have distinct biology compared to large conduit arteries.21 We therefore compared FTL expression in conduit and resistance arteries prepared en face to examine whether iron stores are uniformly low across the arterial endothelium. Using aortas as a representative conduit artery and third-order mesenteric arteries as a representative resistance artery, we found ferritin expression to be higher in the conduit artery endothelium than the resistance artery endothelium (Figure 2A). Biological differences between the conduit and resistance arterial endothelium can be dictated by the presence of myoendothelial junctions (MEJs) in resistance arteries and/or the high flow rates in conduit arteries (Figure 2B).22,23 To investigate the contribution of MEJs to arterial iron stores, we used a vascular cell coculture model previously established by our laboratory that promotes the formation of MEJs and recapitulates the cell biology of resistance artery ECs.8,12,24 We cultured human coronary artery ECs under 3 conditions: cultured without smooth muscle cells (SMCs; monolayer); cocultured without direct contact with SMCs (no contact); and cocultured with SMCs such that MEJs could form (contact; Figure 2C). Bulk RNA sequencing of ECs grown in these 3 conditions resulted in >2700 genes being differentially regulated, which demonstrates that communication with SMCs has broad implications for EC phenotypes (Figure 2D; supplemental Figure 3; supplemental Tables 26-29). Labile iron was directly assessed by FerroOrange, and we found contact with SMCs does not affect endothelial iron (Figure 2E). We also found no evidence of differential iron regulation as assessed by western blot of transferrin receptor, FTH, and FTL (Figure 2F). To check whether shear stress was responsible for the difference in iron regulation across the arterial endothelium, we reanalyzed published bulk RNA-sequencing data of ECs exposed to continuous shear stress (11.5 dynes/cm2) for up to 48 hours, which revealed >4300 differentially expressed genes (Figure 2G-H; supplemental Tables 30-34).23 We found shear stress increased FTH1 and FTL and decreased TFRC (Figure 2I) and thus is likely responsible for differences across the arterial endothelium. Altogether, our findings demonstrate iron levels in resistance artery ECs are among the lowest observed across EC subtypes, suggesting resistance arteries may be especially susceptible to changes in iron homeostasis (Figure 2J).

Figure 2.

Figure 2.

Open in a new tab

Shear stress dictates intracellular iron in conduit and resistance arteries. (A) En face imaging FTL in a conduit artery (aorta) and resistance artery (third-order mesenteric artery). (B) Schematic highlighting shear stress and MEJs as key determinants of conduit and resistance artery endothelial biology. (C-F) Human coronary artery ECs were grown in transwells either in a monolayer, cocultured with SMCs without direct contact (no contact), or cocultured in direct contact with SMCs (contact). (D) Heat map of differentially expressed genes (DEGs) across all 3 conditions as assessed by bulk RNA sequencing. (E) Labile iron was assessed by flow cytometry with FerroOrange. Statistical analysis was performed with Student t test. (F) Transferrin receptor (TFRC; TfR1), ferritin light chain (FTL; FTL), and ferritin heavy chain (FTH1; FTH) were assessed by western blot. Statistical significance was determined by 2-way ANOVA and Holms-Šídák post hoc analysis. (G) Reanalysis of human umbilical vein ECs grown in static or flow conditions for up to 48 hours. (H) Heat map of genes differentially regulated by shear stress. (I) TFRC, FTL, and FTH1 expression in static and flow conditions. Statistical analysis was performed with DESeq2. ∗∗∗P < .001 vs TFRC 0 minute; ∗∗∗∗P < .0001 vs TFRC 0 minute; &&P < .01 vs FTH1 0 minute; ##P < .01 vs FTL 0 minute; ###P < .001 vs FTL 0 minute. (J) Schematic summarizing iron across the vascular endothelium. Art., artery; DAPI, 4′,6-diamidino-2-phenylindole; IEL, internal elastic lamina; MFI, mean fluorescence intensity.

IDA increases NO signaling

The resistance artery endothelium plays a central role in the regulation of blood flow through the production of vasoactive compounds, such as NO, and it has been previously shown that patients with anemia have altered hemodynamics due to elevated NO signaling.5 To investigate whether changes in endothelial iron contribute to the increased NO, we created a mouse model of IDA with replete vascular iron. Mice were placed on an iron-deficient diet at weaning, and phlebotomies were performed when the mice were aged 9 and 11 weeks. Mice then received an IP injection of saline or FeDex (20 mg/kg) to replete vascular iron (Figure 3A). Both male and female IDA mice had decreased liver iron, demonstrating iron deficiency, as well as blood hemoglobin, mean corpuscular hemoglobin concentration, and hematocrit, demonstrating anemia (Figure 3B-C; supplemental Table 35; supplemental Figures 4 and 5). FeDex did not fully rescue anemia in these mice. Vascular iron stores were assessed by FTL expression. FTL was not different between sexes at baseline (supplemental Figure 6). In both sexes, FTL was decreased in IDA mice but similar to controls in IDA+FeDex mice, demonstrating that vascular iron was depleted during IDA and repleted with FeDex (Figure 3D). This resulted in a mouse model with anemia but replete vascular iron.

Figure 3.

Figure 3.

Open in a new tab

Model of IDA with replete vascular iron. (A) Experimental timeline. Female (pink) and male (blue) mice were fed control (48 ppm Fe) or iron-deficient diet (2-6 ppm Fe) beginning at weaning. Phlebotomies were performed in iron-deficient mice at 9 weeks and 11 weeks to aid in the progression of anemia. Mice received a single injection of FeDex (20 mg/kg) or vehicle (saline) at 11 weeks. All experiments were performed at 13 weeks. (B) Liver iron assessed by colorimetric assay. (C) Hemoglobin, mean corpuscular hemoglobin concentration (MCHC), and hematocrit from complete blood counts. Dashed line denotes cutoff for anemia. (D) FTL assessed by western blot in isolated mesenteric arteries. Statistical significance was determined by 1-way ANOVA and Holms-Šídák post hoc analysis.

To examine how changes in vascular iron regulate NO signaling during IDA, we used LSCI to measure changes in blood flow in response to L-NAME. The response to a NO synthase inhibitor, such as L-NAME, is commonly used to assess NO signaling in mice and humans, in whom a larger change in blood flow indicates higher basal NO signaling.5,16 No differences were observed in baseline blood flow or in response to the saline control (supplemental Figure 7). However, we found female IDA mice had a greater decrease in blood flow in response to L-NAME, whereas mice with replete vascular iron had a similar response to L-NAME as control mice (Figure 4A). In male mice, we observed no differences in NO signaling between groups (Figure 4B). We also assessed hemodynamic changes and found increased cardiac output and decreased peripheral resistance in female but not male mice, in agreement with the increased NO signaling (supplemental Figure 8).

Figure 4.

Figure 4.

Open in a new tab

Vascular IDA increases endothelial NO. (A-B) LSCI was used to assess percent change in blood flow in hind paws of female (A) and male mice (B) in response to IP injection of the NO synthase inhibitor L-NAME. Statistical significance was determined by 1-way ANOVA and Holms-Šídák post hoc analysis. Con, control.

Loss of endothelial holo-α-globin increases NO in IDA

NO signaling may become elevated due to either increased NO production or decreased NO scavenging. We assessed the NO metabolites nitrate and nitrite in plasma and observed no change in either across groups (Figure 5A-B). Similarly, we found IDA did not increase eNOS expression or phosphorylation at the activating site S1177 (Figure 5C-D). We have previously shown α-globin is expressed in the endothelium and inhibits NO signaling by scavenging NO.8, 9, 10, 11 We measured α-globin expression in isolated vessels and found no differences in baseline α-globin between sexes (supplemental Figure 6). However, IDA decreased α-globin expression in both sexes, and repletion of vascular iron with FeDex rescued α-globin expression only in female mice (Figure 5E). Heme proteins can either exist in an apo (heme-free) or a functional holo (heme-bound) state. To assess holo-α-globin levels in our model, α-globin was immunoprecipitated from isolated vessels. The amount of heme pulled down was measured and normalized to the amount of α-globin pulled down. We pulled down similar amounts of α-globin in males and females. In females, we found holo-α-globin was at the detection limit for the assay in control and IDA conditions and increased when mice were treated with FeDex. In males, we found a much larger proportion of holo-α-globin in controls, which was decreased in IDA and not rescued with FeDex (Figure 5F). Together, this suggests that although both males and females lose α-globin expression in IDA, females start with a very small pool of holo-α-globin, which is easily lost in IDA and increases with FeDex. Males start with a large pool of holo-α-globin in the endothelium, which, although decreased in IDA, is still detectable and sufficient for maintaining homeostatic levels of NO signaling (Figure 5G).

Figure 5.

Figure 5.

Open in a new tab

Vascular iron regulates endothelial α-globin but not eNOS. (A-B) Plasma nitrate and nitrite concentration. (C-D) S1177 phosphorylation (p-eNOS) and total eNOS assessed by western blot. (E) Endothelial α-globin expression assessed by western blot. (F) α-globin was immunoprecipitated, and heme was measured in the immunoprecipitate to assess holo-α-globin. Similar amounts of α-globin were pulled down across groups. Heme was normalized to the amount of α-globin pulled down for each sample, then scaled relative to the average of male and female controls. (G) Summary of α-globin and holo-α-globin levels in each condition. (H-I) Cdh5-CreERT2+/−Hba1fl/fl mice were fed control (48 ppm Fe) or iron-deficient diet (2-6 ppm Fe) beginning at weaning. Tamoxifen was administered IP at 6 weeks for 10 consecutive days. Phlebotomies of 10% blood volume were performed in iron-deficient mice at 9 weeks and 11 weeks to aid in the progression of anemia. Mice received a single injection of FeDex (20 mg/kg) or vehicle (saline) at 11 weeks. All experiments were performed at 13 weeks. Changes in blood flow in hind paws of female (H) and male mice (I) were assessed by LSCI in response to IP injection of the NO synthase inhibitor L-NAME. All statistical analysis were determined by 1- or 2-way ANOVA and Holms-Šídák post hoc analysis. Con, control.

Based on these results, we hypothesized that the increased NO signaling occurring in IDA females is due to the loss of functional α-globin and that males do not experience this same effect because, although decreased, a functional pool of α-globin remains. To test this hypothesis, we used endothelial-specific α-globin knockout mice (Cdh5-CreERT2+/− Hba1fl/fl), which we have previously shown selectively deplete α-globin from the endothelium.10 If our hypothesis is correct, we expect genetic deletion of endothelial α-globin to prevent iron regulation of NO signaling and eliminate the sex differences observed in response to IDA.

Deletion of Hba1 in the endothelium does not affect arterial iron stores or blood hemoglobin levels (supplemental Figure 9; supplemental Table 36). We again measured NO signaling using LSCI. In female Cre− mice, we found IDA increased NO signaling, which was rescued by repleting vascular iron. However, in Cre+ mice, NO signaling was not rescued by repletion of vascular iron, demonstrating that iron regulates arterial NO signaling through endothelial α-globin. Similar to what was observed in male wild-type mice, male Cre− mice had no change in NO signaling in response to IDA, whereas male Cre+ mice had increased NO signaling, which was not rescued by FeDex (phenocopying female Cre+ mice; Figure 5H-I). These data demonstrate the loss of α-globin is necessary for increased NO signaling in IDA.

Discussion

It is well recognized that IDA is associated with perturbed hemodynamics, but the literature investigating these changes has largely neglected direct actions of iron deficiency on the vasculature. Here, we demonstrate iron levels vary across the vasculature and directly affect endothelial phenotypes. Furthermore, we identify a role for endothelial α-globin in mediating the increase in NO signaling observed in anemia.

The endothelium is specialized into distinct phenotypes that vary along the vascular tree. We used single-cell transcriptomic data sets of ECs from mice and humans to investigate whether iron varies between endothelial subtypes and validated these findings both in vivo and in vitro. We found iron levels to be lowest in arterial ECs and highest in lymphatic ECs. Notably, this pattern did not match the expression of iron-dependent proteins, suggesting differential iron regulation is not due to increased iron utilization. Therefore, we investigated the impact of iron chelation on EC phenotypes. We found iron chelation upregulated the expression of arterial markers and decreased lymphatic markers. Not all markers of arterial and lymphatic ECs changed in the hypothesized direction, suggesting iron is not sufficient to fully shift EC identity. Many signals regulate EC identity, and it remains to be studied how iron interacts with or mediates the known effects of vascular endothelial growth factor, shear stress, Notch signaling, and cell cycle state on endothelial differentiation.25

Arteries themselves are heterogeneous. Large conduit arteries, such as the aorta, are exposed to high flow rates and have thick internal elastic lamina, whereas resistance arteries are exposed to lower flow rates and have a thin, discontinuous internal elastic lamina, allowing for the formation of MEJs. We modeled both flow and MEJs in vitro to examine whether they affect endothelial iron. Although the presence of MEJs does not change iron stores or the labile iron pool, exposing ECs to flow increases intracellular iron, consistent with the phenotype observed in conduit artery ECs. These results explain the difference in iron between arterial EC subtypes but do not explain the differences we observed among other vascular segments because iron is highest in lymphatic ECs even though flow is lowest in lymphatics. These differences seem to be intrinsic to the cell’s identity because artery and lymphatic differences were maintained in culture despite removal from their native microenvironment. Ultimately, our data suggest the resistance artery endothelium is iron deplete relative to other EC subtypes. Having limited iron stores may make these ECs especially susceptible to iron deficiency. We therefore investigated how the resistance artery endothelium is affected in IDA.

Increased arterial NO signaling has been previously observed in patients with anemia and attributed to decreased scavenging of NO by erythrocytes.5 Free hemoglobin is a well-recognized and potent NO scavenger, and in silico modeling suggests changes in hematocrit can affect NO signaling in the vascular wall.26,27 However, ex vivo artery preparations demonstrate limited impact of erythrocytes on arterial NO signaling. For example, studies in the rat renal microvasculature demonstrated minimal scavenging of arterial NO by erythrocytes, with the maximum effect at 1% hematocrit,6 well below what is observed in anemia. Other studies in porcine coronary arteries found no effect of erythrocytes (60% hematocrit) on flow-mediated dilation, an NO-dependent process.7 In addition, studies in a phlebotomy-driven mouse model of anemia have demonstrated impaired endothelial function attributed to erythrocyte dysfunction, which is contrary to what is expected if reduced hematocrit is sufficient to increase NO.28 Taken together, these studies suggest changes in hematocrit may be insufficient to increase NO signaling in anemia.

For this reason, we developed a mouse model of IDA, in which we can replete vascular iron without fully rescuing the anemia. Blood flow was assessed by LSCI, a semiquantitative measurement of erythrocyte movement through tissue. We observed similar baseline blood flow between groups. Others have demonstrated that volumetric blood flow is increased in anemia.5 This is in agreement with our findings of increased cardiac output and decreased peripheral resistance. To assess NO signaling, we measured blood flow in the hind paws before and after administration of L-NAME. This method is analogous to the studies performed in the forearm of patients with anemia.5 L-NAME causes vasoconstriction by inhibiting NO synthesis. Therefore, the magnitude of the response to L-NAME can be used as a measure of NO signaling. We observed elevated NO signaling in female, but not male, IDA mice, which was rescued by repletion of vascular iron with FeDex. We found IDA did not affect nitrite and nitrate levels, eNOS expression, or eNOS activation by phosphorylation at S1177, suggesting that NO synthesis was not responsible for the increase in NO signaling.

We and others have previously shown α-globin is expressed in the resistance artery endothelium and inhibits NO signaling by directly binding to eNOS and scavenging NO.8,10,11,22,29 We found IDA decreases α-globin in both male and female mice, and FeDex rescues α-globin expression in female mice. We further measured the heme-bound portion of α-globin in our model and found holo-α-globin to be at the detection limit of the assay in females but easily detectable in males. IDA decreased holo-α-globin in males, which was not rescued by repleting vascular iron, whereas holo-α-globin was increased in iron-replete females. Based on these data, we hypothesized changes in NO signaling may be due to the regulation of holo-α-globin. To test this hypothesis, we used endothelial-specific Hba1 knockout mice and found that deletion of endothelial α-globin prevented the regulation of NO signaling by FeDex in female mice, demonstrating iron regulates NO signaling through endothelial α-globin. In male Cre+ mice, IDA increased NO, which was not rescued by FeDex phenocopying the female Cre+ mice. Based on these findings, we conclude loss of α-globin is necessary for increased NO signaling in IDA in both males and females. However, the increased NO signaling does not occur in males because they have a larger pool of holo-α-globin, which is not fully lost in IDA.

Notably, loss of endothelial α-globin is necessary for increased NO signaling in IDA and for the rescue of NO signaling by iron repletion, but it is not sufficient to increase NO in this model, suggesting IDA-induced increases in NO are multifactorial. It may be that neither loss of endothelial α-globin nor loss of circulating hemoglobin is sufficient to increase NO in IDA, but rather both are required. Indeed, blood transfusion has been shown to rescue NO signaling in patients with anemia.5 However, interpretation is potentially complicated by free hemoglobin and erythrocyte-derived microparticles, which are potent NO scavengers often found in stored blood.30, 31, 32, 33 The complexity of NO regulation in anemia is further demonstrated by genetic anemias such as β-thalassemia and sickle cell anemia, which present with decreased NO signaling, along with hemolysis, iron overload, and inflammation.34, 35, 36 Together, these observations demonstrate NO bioavailability in anemia is determined by the full disease milieu rather than by hematocrit alone. Although our model enables assessment of NO signaling in vivo, it does completely isolate individual components that contribute to NO signaling. Future studies using more reductionist models, including isolated arteries, cell culture, and in silico models, may provide further clarification regarding how hematocrit and endothelial α-globin interact to regulate NO signaling in anemia. In addition, development of genetic models that manipulate endothelial iron independent of systemic iron will provide novel insights.

The mechanisms by which endothelial α-globin expression and stability are regulated in IDA warrant further study. α-hemoglobin–stabilizing protein has been demonstrated in the endothelium and likely accounts for the stability of the apo-protein.29,37,38 Other pathways that regulate α-globin in erythrocytes during iron deficiency may also contribute to the downregulation of α-globin in ECs during iron deficiency. In erythrocytes, iron restriction decreases α-globin translation through heme-regulated inhibitor and promotes α-globin autophagy through mechanistic target of rapamycin kinase complex 1/ULK1-mediated signaling.39,40 The genes necessary for both heme-regulated inhibitor and mechanistic target of rapamycin kinase pathways are detected in the endothelial transcriptomic data sets included in supplemental Tables 3-34.

The specificity of these regulatory pathways may also account for why α-globin, but not other hemoproteins such as eNOS, are downregulated in IDA. Notably, the NO receptor soluble guanylyl cyclase is a hemoprotein. Nevertheless, NO signaling is maintained. Early studies demonstrated NO-heme complexes can transfer from holo-hemoproteins to apo–soluble guanylyl cyclase, and more recent studies have demonstrated NO mobilizes intracellular heme.41,42 These or similar mechanisms may redistribute heme to maintain NO signaling in IDA.

Anemia has long been recognized to regulate hemodynamics by decreasing systemic vascular resistance,5 but the role of endothelial iron deficiency in these changes has been overlooked. Our data demonstrate that iron regulates endothelial specification and that resistance arterial iron and its regulation of α-globin are critical for modulating hemodynamics in IDA. Furthermore, our study demonstrates a novel role for α-globin in mediating sex differences in vascular biology. Although there were similar levels of iron deficiency in males and females, female mice were more sensitive to vascular iron deficiency due to their nearly absent holo-α-globin. In humans, the impact of this sensitivity may be potentiated by the high prevalence of IDA in females relative to males.43 Understanding the role of iron in vascular biology will help elucidate the hemodynamic consequences of one of the most common nutrient deficiencies.44

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Acknowledgments

This study was supported by the National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH) grants HL007284 (L.S.D., S.A.L., W.J.S., M.A.L.), HL172605 (L.S.D.), HL176103 (S.A.L.), HL137112 and HL171997 (B.E.I.), HL155618 (B.E.I. and P.B.), and HL172874 (A.N.G.); American Heart Association (AHA) grants 25CDA1444441 (L.S.D.), 25PRE1375860[(W.J.S.), 25PRE1376958 (Z.J.J.), and AHA915176 (M.A.L.); University of Virginia (UVA) Comprehensive Cancer Center Farrow Fellowship (B.L.O.); and UVA Comprehensive Cancer Center Pediatric Pilot Award FY2025 (B.E.I.).

Authorship

Contribution: L.S.D. and B.E.I. conceived and designed experiments; T.M.S., P.B., J.D.A., S.T., A.N.G., and B.E.I. contributed critical expertise and resources; L.S.D., S.N., W.J.S., S.A.L., B.L.O., Z.J.J., N.A., M.A.L., M.E.S., and S.T. performed experiments; L.S.D. and B.E.I. wrote the manuscript; and all authors reviewed the manuscript.

Footnotes

Raw data and normalized counts from bulk RNA sequencing of cocultured endothelial cells and deferoxamine-treated endothelial cells have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (accession numbers GSE314128 and GSE314015). Full results of differential expression analysis are available in the supplemental Materials.

The online version of this article contains a data supplement.

There is a Blood Commentary on this article in this issue.

The publication costs of this article were defrayed in part by page charge payment. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Supplementary Material

Supplemental Methods, Figures, Table Legends, and References
BLOOD_BLD-2025-032905-mmc1.pdf (2.8MB, pdf)
Supplemental Tables 1-8
BLOOD_BLD-2025-032905-mmc2.xlsx (15.7MB, xlsx)
Supplemental Tables 9-20
BLOOD_BLD-2025-032905-mmc7.xlsx (3.5MB, xlsx)
Supplemental Tables 21-25
BLOOD_BLD-2025-032905-mmc3.xlsx (10.4MB, xlsx)
Supplemental Tables 26-29
BLOOD_BLD-2025-032905-mmc4.xlsx (5.9MB, xlsx)
Supplemental Tables 30-34
BLOOD_BLD-2025-032905-mmc5.xlsx (12.6MB, xlsx)
Supplemental Tables 35 and 36
BLOOD_BLD-2025-032905-mmc6.xlsx (16.5KB, xlsx)

References

  • 1.Anand IS, Gupta P. Anemia and iron deficiency in heart failure: current concepts and emerging therapies. Circulation. 2018;138(1):80–98. doi: 10.1161/CIRCULATIONAHA.118.030099. [DOI] [PubMed] [Google Scholar]
  • 2.Awan AA, Walther CP, Richardson PA, Shah M, Winkelmayer WC, Navaneethan SD. Prevalence, correlates and outcomes of absolute and functional iron deficiency anemia in nondialysis-dependent chronic kidney disease. Nephrol Dial Transpl. 2021;36(1):129–136. doi: 10.1093/ndt/gfz192. [DOI] [PubMed] [Google Scholar]
  • 3.Ganz T. Anemia of inflammation. N Engl J Med. 2019;381(12):1148–1157. doi: 10.1056/NEJMra1804281. [DOI] [PubMed] [Google Scholar]
  • 4.Andreini C, Putignano V, Rosato A, Banci L. The human iron-proteome. Metallomics. 2018;10(9):1223–1231. doi: 10.1039/c8mt00146d. [DOI] [PubMed] [Google Scholar]
  • 5.Anand IS, Chandrashekhar Y, Wander GS, Chawla LS. Endothelium-derived relaxing factor is important in mediating the high output state in chronic severe anemia. J Am Coll Cardiol. 1995;25(6):1402–1407. doi: 10.1016/0735-1097(95)00007-Q. [DOI] [PubMed] [Google Scholar]
  • 6.Imig JD, Gebremedhin D, Harder DR, Roman RJ. Modulation of vascular tone in renal microcirculation by erythrocytes: role of EDRF. Am J Physiol. 1993;264(1 pt 2):H190–H195. doi: 10.1152/ajpheart.1993.264.1.H190. [DOI] [PubMed] [Google Scholar]
  • 7.Liao JC, Hein TW, Vaughn MW, Huang KT, Kuo L. Intravascular flow decreases erythrocyte consumption of nitric oxide. Proc Natl Acad Sci U S A. 1999;96(15):8757–8761. doi: 10.1073/pnas.96.15.8757. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Straub AC, Lohman AW, Billaud M, et al. Endothelial cell expression of haemoglobin alpha regulates nitric oxide signalling. Nature. 2012;491(7424):473–477. doi: 10.1038/nature11626. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Straub AC, Butcher JT, Billaud M, et al. Hemoglobin alpha/eNOS coupling at myoendothelial junctions is required for nitric oxide scavenging during vasoconstriction. Arterioscler Thromb Vasc Biol. 2014;34(12):2594–2600. doi: 10.1161/ATVBAHA.114.303974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Keller TCS, Lechauve C, Keller AS, et al. Endothelial alpha globin is a nitrite reductase. Nat Commun. 2022;13(1):6405. doi: 10.1038/s41467-022-34154-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Keller TCS, IV, Butcher JT, Broseghini-Filho GB, et al. Modulating vascular hemodynamics with an alpha globin mimetic peptide (HbalphaX) Hypertension. 2016;68(6):1494–1503. doi: 10.1161/HYPERTENSIONAHA.116.08171. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Biwer LA, Lechauve C, Vanhoose S, Weiss MJ, Isakson BE. A cell culture model of resistance arteries. J Vis Exp. 2017;127:55992. doi: 10.3791/55992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Heberlein KR, Straub AC, Best AK, et al. Plasminogen activator inhibitor-1 regulates myoendothelial junction formation. Circ Res. 2010;106(6):1092–1102. doi: 10.1161/CIRCRESAHA.109.215723. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Dunaway LS, Saii K, LoBue A, et al. The hemodynamic response to nitrite is acute and dependent upon tissue perfusion. Nitric Oxide. 2024;150:47–52. doi: 10.1016/j.niox.2024.07.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Sveeggen TM, Isakson BE, Straub AC, Bagher P. Bedding as a variable affecting fasting blood glucose and vascular physiology in mice. Am J Physiol Heart Circ Physiol. 2023;325(2):H338–H345. doi: 10.1152/ajpheart.00168.2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Dunaway LS, Cook AK, Kellum CE, et al. Endothelial histone deacetylase 1 activity impairs kidney microvascular NO signaling in rats fed a high-salt diet. Acta Physiol (Oxf) 2024;240(9) doi: 10.1111/apha.14201. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Barnett SN, Cujba AM, Yang L, et al. An organotypic atlas of human vascular cells. Nat Med. 2024;30(12):3468–3481. doi: 10.1038/s41591-024-03376-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Dunaway LS, Luse MA, Nyshadham S, et al. Obesogenic diet disrupts tissue-specific mitochondrial gene signatures in the artery and capillary endothelium. Physiol Genomics. 2024;56(2):113–127. doi: 10.1152/physiolgenomics.00109.2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Wilkinson N, Pantopoulos K. The IRP/IRE system in vivo: insights from mouse models. Front Pharmacol. 2014;5:176. doi: 10.3389/fphar.2014.00176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Chavkin NW, Genet G, Poulet M, et al. Endothelial cell cycle state determines propensity for arterial-venous fate. Nat Commun. 2022;13(1):5891. doi: 10.1038/s41467-022-33324-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.King DR, Sedovy MW, Eaton X, et al. Cell-To-cell communication in the resistance vasculature. Compr Physiol. 2022;12(4):3833–3867. doi: 10.1002/cphy.c210040. [DOI] [PubMed] [Google Scholar]
  • 22.Shu X, Ruddiman CA, Keller TCS, et al. Heterocellular contact can dictate arterial function. Circ Res. 2019;124(10):1473–1481. doi: 10.1161/CIRCRESAHA.118.313926. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Afshar Y, Ma F, Quach A, et al. Transcriptional drifts associated with environmental changes in endothelial cells. Elife. 2023;12 doi: 10.7554/eLife.81370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Biwer LA, Taddeo EP, Kenwood BM, Hoehn KL, Straub AC, Isakson BE. Two functionally distinct pools of eNOS in endothelium are facilitated by myoendothelial junction lipid composition. Biochim Biophys Acta. 2016;1861(7):671–679. doi: 10.1016/j.bbalip.2016.04.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Trimm E, Red-Horse K. Vascular endothelial cell development and diversity. Nat Rev Cardiol. 2023;20(3):197–210. doi: 10.1038/s41569-022-00770-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Kuhn V, Diederich L, Keller TCS, et al. Red blood cell function and dysfunction: redox regulation, nitric oxide metabolism, anemia. Antioxid Redox Signal. 2017;26(13):718–742. doi: 10.1089/ars.2016.6954. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Jeffers A, Gladwin MT, Kim-Shapiro DB. Computation of plasma hemoglobin nitric oxide scavenging in hemolytic anemias. Free Radic Biol Med. 2006;41(10):1557–1565. doi: 10.1016/j.freeradbiomed.2006.08.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Chennupati R, Solga I, Wischmann P, et al. Chronic anemia is associated with systemic endothelial dysfunction. Front Cardiovasc Med. 2023;10 doi: 10.3389/fcvm.2023.1099069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Lechauve C, Butcher JT, Freiwan A, et al. Endothelial cell alpha-globin and its molecular chaperone alpha-hemoglobin-stabilizing protein regulate arteriolar contractility. J Clin Invest. 2018;128(11):5073–5082. doi: 10.1172/JCI99933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Nishiyama T, Hanaoka K. Free hemoglobin concentrations in patients receiving massive blood transfusion during emergency surgery for trauma. Can J Anaesth. 2000;47(9):881–885. doi: 10.1007/BF03019668. [DOI] [PubMed] [Google Scholar]
  • 31.Donadee C, Raat NJ, Kanias T, et al. Nitric oxide scavenging by red blood cell microparticles and cell-free hemoglobin as a mechanism for the red cell storage lesion. Circulation. 2011;124(4):465–476. doi: 10.1161/CIRCULATIONAHA.110.008698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Liu C, Liu X, Janes J, et al. Mechanism of faster NO scavenging by older stored red blood cells. Redox Biol. 2014;2:211–219. doi: 10.1016/j.redox.2013.12.014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Schaid TR, Jr., Cohen MJ, D'Alessandro A, et al. Trauma induces intravascular hemolysis, exacerbated by red blood cell transfusion and associated with disrupted arginine-nitric oxide metabolism. Shock. 2023;59(1):12–19. doi: 10.1097/SHK.0000000000002036. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Aggeli C, Antoniades C, Cosma C, et al. Endothelial dysfunction and inflammatory process in transfusion-dependent patients with beta-thalassemia major. Int J Cardiol. 2005;105(1):80–84. doi: 10.1016/j.ijcard.2004.12.025. [DOI] [PubMed] [Google Scholar]
  • 35.Detterich JA, Kato RM, Rabai M, Meiselman HJ, Coates TD, Wood JC. Chronic transfusion therapy improves but does not normalize systemic and pulmonary vasculopathy in sickle cell disease. Blood. 2015;126(6):703–710. doi: 10.1182/blood-2014-12-614370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Kato GJ, Steinberg MH, Gladwin MT. Intravascular hemolysis and the pathophysiology of sickle cell disease. J Clin Invest. 2017;127(3):750–760. doi: 10.1172/JCI89741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Kiger L, Keith J, Freiwan A, et al. Redox-regulation of alpha-globin in vascular physiology. Antioxidants (Basel) 2022;11(1):159. doi: 10.3390/antiox11010159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Ghosh A, Garee G, Sweeny EA, Nakamura Y, Stuehr DJ. Hsp90 chaperones hemoglobin maturation in erythroid and nonerythroid cells. Proc Natl Acad Sci U S A. 2018;115(6):E1117–E1126. doi: 10.1073/pnas.1717993115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Voon HP, Vadolas J. Controlling alpha-globin: a review of alpha-globin expression and its impact on beta-thalassemia. Haematologica. 2008;93(12):1868–1876. doi: 10.3324/haematol.13490. [DOI] [PubMed] [Google Scholar]
  • 40.Crosby JS, Chefalo PJ, Yeh I, et al. Regulation of hemoglobin synthesis and proliferation of differentiating erythroid cells by heme-regulated eIF-2alpha kinase. Blood. 2000;96(9):3241–3248. [PubMed] [Google Scholar]
  • 41.Hanna DA, Harvey RM, Martinez-Guzman O, et al. Heme dynamics and trafficking factors revealed by genetically encoded fluorescent heme sensors. Proc Natl Acad Sci U S A. 2016;113(27):7539–7544. doi: 10.1073/pnas.1523802113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Ignarro LJ, Adams JB, Horwitz PM, Wood KS. Activation of soluble guanylate cyclase by NO-hemoproteins involves NO-heme exchange. Comparison of heme-containing and heme-deficient enzyme forms. J Biol Chem. 1986;261(11):4997–5002. [PubMed] [Google Scholar]
  • 43.Lopez A, Cacoub P, Macdougall IC, Peyrin-Biroulet L. Iron deficiency anaemia. Lancet. 2016;387(10021):907–916. doi: 10.1016/S0140-6736(15)60865-0. [DOI] [PubMed] [Google Scholar]
  • 44.Pasricha SR, Tye-Din J, Muckenthaler MU, Swinkels DW. Iron deficiency. Lancet. 2021;397(10270):233–248. doi: 10.1016/S0140-6736(20)32594-0. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Methods, Figures, Table Legends, and References
BLOOD_BLD-2025-032905-mmc1.pdf (2.8MB, pdf)
Supplemental Tables 1-8
BLOOD_BLD-2025-032905-mmc2.xlsx (15.7MB, xlsx)
Supplemental Tables 9-20
BLOOD_BLD-2025-032905-mmc7.xlsx (3.5MB, xlsx)
Supplemental Tables 21-25
BLOOD_BLD-2025-032905-mmc3.xlsx (10.4MB, xlsx)
Supplemental Tables 26-29
BLOOD_BLD-2025-032905-mmc4.xlsx (5.9MB, xlsx)
Supplemental Tables 30-34
BLOOD_BLD-2025-032905-mmc5.xlsx (12.6MB, xlsx)
Supplemental Tables 35 and 36
BLOOD_BLD-2025-032905-mmc6.xlsx (16.5KB, xlsx)

Articles from Blood are provided here courtesy of The American Society of Hematology

RESOURCES