Iron Sucrose Accelerates Early Atherogenesis by Increasing Superoxide Production and Upregulating Adhesion Molecules in CKD

Abstract

High-dose intravenous iron supplementation is associated with adverse cardiovascular outcomes in patients with CKD, but the underlying mechanism is unknown. Our study investigated the causative role of iron sucrose in leukocyte–endothelium interactions, an index of early atherogenesis, and subsequent atherosclerosis in the mouse remnant kidney model. We found that expression levels of intracellular cell adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) and adhesion of U937 cells increased in iron-treated human aortic endothelial cells through upregulated NADPH oxidase (NOx) and NF-κB signaling. We then measured mononuclear–endothelial adhesion and atherosclerotic lesions of the proximal aorta in male C57BL/6 mice with subtotal nephrectomy, male apolipoprotein E–deficient (ApoE^−/−) mice with uninephrectomy, and sham-operated mice subjected to saline or parenteral iron loading. Iron sucrose significantly increased tissue superoxide production, expression of tissue cell adhesion molecules, and endothelial adhesiveness in mice with subtotal nephrectomy. Moreover, iron sucrose exacerbated atherosclerosis in the aorta of ApoE^−/− mice with uninephrectomy. In patients with CKD, intravenous iron sucrose increased circulating mononuclear superoxide production, expression of soluble adhesion molecules, and mononuclear–endothelial adhesion compared with healthy subjects or untreated patients. In summary, iron sucrose aggravated endothelial dysfunction through NOx/NF-κB/CAM signaling, increased mononuclear–endothelial adhesion, and exacerbated atherosclerosis in mice with remnant kidneys. These results suggest a novel causative role for therapeutic iron in cardiovascular complications in patients with CKD.

Anemia is frequently encountered and associated with cardiovascular (CV) outcomes in patients with CKD.¹ Correcting anemia usually requires erythropoiesis-stimulating agents (ESAs). The use of intravenous iron with ESA is required for optimal management of anemia in CKD patients.² The reasons for intravenous iron therapy are that it helps reduce ESA requirements and increases hemoglobin levels.³ As a major transition metal, excess iron is a potent pro-oxidant capable of redox cycling. Rooyakkers et al.⁴ have disclosed that intravenous iron administration generates bioactive iron, increases reactive oxygen species (ROS) in plasma, and reduces forearm flow-mediated dilatation in healthy individuals. A cross-sectional study has shown an interrelation among administered annual intravenous iron dose, carotid intima media thickness, and generation of advanced oxidation products of proteins in patients under maintenance hemodialysis (HD).⁵ Recently, our study⁶ and a study by Kalantar-Zadeh et al.⁷ have suggested that high-dose intravenous iron supplementation is associated with adverse CV outcomes and increased mortality in HD patients. However, the underlying mechanisms of this association have not been fully elucidated.

The molecular link between iron and endothelial damage and subsequent formation of atherosclerosis in vivo remains to be clarified. Pang et al.⁸ have shown colocalization of ferritin and iron in human atherosclerotic lesions, implicating the pathologic evidence of iron in the development of atherosclerosis. Our previous study⁹ has disclosed that iron deposition is closely associated with the progression of atherosclerosis in apolipoprotein E–deficient (ApoE^−/−) mice. Restrictions in dietary iron intake to ApoE^−/− mice led to a significant reduction in atherosclerotic lesion formation through decrease of LDL oxidation.⁹ In contrast, Kirk et al.¹⁰ have found that elevated serum tissue levels of iron induced by a 2% carbonyl iron diet were not atherogenic in ApoE^−/− mice. Dietary iron overload caused a 30% rise in plasma triglyceride and cholesterol but reduced the severity of atherosclerosis by 50%. Using a rat CKD model, Lim and Vaziri¹¹ showed that intravenous iron dextran administration induced oxidative stress in the aorta and heart but failed to show an impact on atherosclerotic lesions. Therefore, the contention that iron participates in endothelial dysfunction and subsequent atherosclerosis in CKD animal models remains to be confirmed.

In initiating atherosclerosis, cell adhesion molecules, such as intracellular cell adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1), are upregulated on the endothelial cell surface to mediate the adhesion of circulating mononuclear cells (MNCs) to endothelial cells and the migration of MNCs to the subendothelial space. During disease progression, ROSs serve as common intracellular messengers for redox-sensitive pathways and play a role in the development of vascular disease.¹² Because renal dysfunction is associated with higher oxidative stress and because iron provokes ROS production in vivo,^5,13 we postulated that iron accelerates atherogenesis by increasing the adhesion of MNCs to endothelial cells and the development of atherosclerotic lesions in CKD.

Therefore, first, we conducted an in vitro assay of MNC–endothelial cell adhesion to investigate the signaling cascade in iron sucrose-treated human aortic endothelial cells (HAECs). Second, a novel MNC–endothelial cell adhesion assay, an index of early atherogenesis, was carried out in subtotal nephrectomized (SNx) C57BL/6 mice to assess the in vivo effects of parenteral iron sucrose on endothelial damage. Third, uninephrectomized (UNx) ApoE^−/− mice were examined to assess the chronic cumulative effects of parenteral iron sucrose on the progression of atherosclerotic lesions. Fourth, a human study was performed to validate whether intravenous iron sucrose could promote circulating MNCs in ROS production and endothelial adhesiveness of circulating MNCs in CKD patients undergoing HD.

Results

Iron Sucrose Enhanced Adhesion of Monocytic U937 to HAECs and Increased Intracellular ROS Production in HAECs

Noncytotoxic iron sucrose concentrations≤160 μg/ml were used based on the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide cell viability assay (Supplemental Figure 1). Iron sucrose concentration of 160 μg/ml is within a clinically achievable range of plasma iron concentration.¹⁴ We found that HAECs treated with iron sucrose had distinct intracellular uptake of Prussian blue (Figure 1A). Then, we discovered that iron sucrose significantly induced a time-dependent increase in intracellular ROS production in HAECs between 1 and 3 hours at a concentration of 160 μg/ml, but the effect diminished at 4 hours (Figure 1B). The effects of iron sucrose on endothelial adhesion were further examined, and 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester-labeled U937 cells and iron-treated HAECs were submitted to adhesion assay (Figure 1, C and D). Iron sucrose at a concentration of 160 μg/ml significantly increased endothelial adhesion in HAECs in a time-dependent manner, but the effect diminished at 5 hours (Figure 1C). Iron sucrose also increased endothelial adhesion in a dose-dependent manner at concentrations of 40–160 μg/ml (Figure 1D). In addition, iron sucrose could significantly increase NADPH oxidase (NOx) activity of HAECs between 0.5 and 1 hours at a concentration of 160 μg/ml (Figure 1E). Cotreatment of HAECs with N-acetylcysteine, apocynin, or p22^phox small interfering RNA (siRNA) significantly attenuated the iron-induced intracellular ROS production (Figure 1, F, H, and J) and endothelial adhesion (Figure 1, G, I, and K).

Figure 1.

Iron sucrose increased intracellular ROS production in HAECs and enhanced adhesion of monocytic U937 to HAECs. (A) HAECs were incubated with 160 μg/ml iron sucrose or vehicle, and cellular labeling was visualized with Prussian blue staining for 4 hours. (B) Time-dependent effect of iron sucrose on intracellular ROS production in HAECs. (C) Time-dependent and (D) dose-dependent effects of iron sucrose on adhesion of U937 to HAECs. (E) Time-dependent effect of iron sucrose on NOx activity. Effects of N-acetylcysteine (NAC; 100 mmol/L) on (F) ROS production and (G) U937 adhesion in iron-treated HAECs, apocynin (100 nM) on (H) ROS production and (I) U937 adhesion in iron-stimulated HAECs, and p22^phox siRNA on (J) ROS production and (K) U937 adhesion in iron-stimulated HAECs. Cell adherence was expressed as the fold difference from the untreated control. Values are means±SEMs of five separate experiments. Scale bar, 100 μm in B–D. *P<0.05 compared with baseline; ^†P < 0.05 compared with 160 µg/ml iron sucrose for 3 hours for ROS production and 4 hours for U937 adhesion; ^#P<0.05 compared with iron-treated groups.

Iron Sucrose Enhanced NF-κB Binding Activity and Increased Expression of Adhesion Molecules in HAECs

The effects of iron sucrose on NF-κB binding activity and downstream expression of adhesion molecules were further examined. Our data disclosed that iron sucrose significantly induced an increase of NF-κB binding activity in HAECs in dose- (Figure 2A) and time-dependent (Figure 2B) manners. Furthermore, iron sucrose at 80 and 160 μg/ml significantly upregulated VCAM-1 and ICAM-1 expressions in HAECs (Figure 2C). Cotreatment of HAECs with NF-κB p65 siRNA significantly attenuated iron-induced VCAM-1 and ICAM-1 expressions (Figure 2D) and endothelial adhesion (Figure 2E). Similarly, cotreatment of HAECs with anti–VCAM-1 antibody (10 μg/ml) or anti–ICAM-1 antibody (10 μg/ml) attenuated endothelial adhesion (Figure 2F).

Figure 2.

Iron sucrose enhanced NF-κB binding activity and increased expression of VCAM-1 and ICAM-1 in HAECs. NF-κB binding activity was determined by assaying nuclear lysates using DNA binding ELISA. Protein levels were determined with Western blotting. (A) Dose-dependent and (B) time-dependent effects of iron sucrose on NF-κB binding activity in HAECs. (C) Dose-dependent effect of iron sucrose on expression of VCAM-1 and ICAM-1 in HAECs. Effect of NF-κB p65 siRNA on (D) VCAM-1/ICAM-1 expression and (E) U937 adhesion in iron-stimulated HAECs and (F) antagonist effect of anti–VCAM-1 (10 μg/ml) antibody or anti–ICAM-1 (10 μg/ml) antibody on U937 adhesion in iron-treated HAECs. Values are expressed as means±SEMs of five separate experiments. *P<0.05 compared with baseline; ^#P<0.05 compared with iron-treated groups.

Intraperitoneal Iron Increased Leukocyte–Endothelium Adhesion in SNx Wild-Type Mice

We investigated whether iron sucrose enhanced leukocyte–endothelium adhesion in SNx or sham-operated C57BL/6 mice. After intraperitoneal (intraperitoneal) administration of iron for 5 days, our data showed that the serum iron parameters were highest in the iron-treated SNx mice (Supplemental Table 1). We found that the number of leukocytes adhering to the aortic endothelium was significantly increased in intraperitoneal iron-treated SNx mice compared with that in the SNx groups without iron (Figure 3, A and B). Likewise, the number of adhered leukocytes in aortic endothelium was significantly increased in SNx mice compared with sham groups (Figure 3B). To avoid subjective bias, a novel in vivo fluorescence-labeled exogenous U937–endothelium adhesion assay was performed to further quantify the degree of endothelial adhesion. The procedure is illustrated in Figure 3C. Similarly, fluorescence-labeled adhesion analysis disclosed that SNx mice exhibited significantly greater fluorescence intensity in aortic endothelium than sham mice without iron. Intraperitoneal iron exacerbated fluorescence intensity in the aortic endothelium of SNx mice (Figure 3D).

Figure 3.

Intraperitoneal iron increased leukocyte–endothelium adhesion in SNx wild-type mice. (A) Representative images of leukocyte–endothelium adhesion in sham or SNx mice; the arrows indicate the adhered leukocytes in aortic endothelium. (B) Quantitative data of the effect of iron on leukocyte–endothelium adhesion in sham or SNx mice. Data are expressed as means±SEMs. Number of animals examined in each group was 10. (C) Protocol illustration of the in vivo fluorescence-labeled exogenous U937–endothelium adhesion assay. (D) Quantitative data of the effect of iron on 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM) -labeled endothelium adhesion in sham or SNx mice. Data are expressed as means±SEMs. Number of animals examined in each group was 10. Adherence was expressed as the fold difference from the untreated control.

Intraperitoneal Iron Enhanced Superoxide Production and Upregulated Adhesion Molecule Expression in SNx Wild-Type Mice

We measured in vivo superoxide counts in aortic tissue using a chemiluminescence (CL) analyzing system¹⁵ and found that the CL count in aortic tissue was significantly increased in SNx mice compared with sham groups without iron. Likewise, intraperitoneal iron exacerbated the increased CL count in aortic tissue of SNx mice (Figure 4, A and B). Downstream expressions of aortic VCAM-1 and ICAM-1 were also examined. Similarly, tissue ELISA analysis disclosed that SNx mice exhibited significantly higher VCAM-1 and ICAM-1 expressions in aortic tissues than sham groups without iron, and intraperitoneal iron aggravated VCAM-1 and ICAM-1 expressions in aortic tissues of SNx mice (Figure 4, C and D). Eight weeks after SNx, SNx mice significantly had increased ROS production by circulating MNCs compared with sham-operated mice (Supplemental Figure 2). Furthermore, the source of ROS in aorta was determined using immunohistochemical staining of the p22^phox subunit of NOx in aortic tissues (Figure 4E). We found that the positive immunoreactivity was observed mainly in the endothelium but scantly in the media in sham-operated mice. SNx mice had higher p22^phox expression in aortic tissues than sham groups without iron. Intriguingly, intraperitoneal iron enhanced the expression of p22^phox in aortic tissues of SNx mice with increased p22^phox staining in not only the endothelium but also, the media.

Figure 4.

Intraperitoneal iron enhanced superoxide production and upregulated the expression of adhesion molecules in SNx wild-type mice. (A) Representative images of superoxide production in iron-loaded sham or SNx mice. (B) Quantitative data of CL counts in iron-loaded sham or SNx mice; the data were presented as CL counts per 10 s/μg protein from aortic tissue lysates. Expression of (C) VCAM-1 and (D) ICAM-1 in aortic tissue lysates in iron-loaded sham or SNx mice. (E) Representative images of immunohistochemical staining of p22^phox in aorta of iron-loaded sham or SNx mice. The arrows indicate positive signals of p22^phox in endothelium, and the arrowheads indicate positive signals of p22^phox in the media layer. Data are expressed as means±SEMs. Number of animals examined in each group was 10. Scale bar, 20 μm.

Intraperitoneal Iron Sucrose Exacerbated Lipid Peroxidation and Atherosclerosis in UNx ApoE^−/− Mice

Early atherosclerotic lesions, including foam cell formation and intermediate lesion consisting of a mixture of smooth muscle cells, were observed in ApoE^−/− mice at age 21 weeks. Likewise, the UNx ApoE^−/− mice developed more advanced atherosclerotic lesions than the sham-operated ApoE^−/− mice. This progression of atherosclerosis was further accentuated by intraperitoneal iron administration in the UNx ApoE^−/− mice (Figure 5A). As expected, Perls’ staining of the sectioned aortas revealed that iron deposition was more evident in the endothelium of intermediate lesions from iron-loaded UNx ApoE^−/− mice (Figure 5A). We also measured in vivo superoxide counts of aortic tissue using a CL analyzing system and tissue lipid peroxide level in the aorta using thiobarbituric acid (TBA) reactive substance. We found that CL counts from aortic tissue were significantly increased in UNx mice compared with sham groups. Likewise, iron sucrose further increased CL count in aortic tissue of UNx ApoE^−/− mice (Figure 5B). Similarly, UNx mice exhibited significantly higher malondialdehyde (MDA) level (Figure 5C) and ICAM-1 expression (Figure 5E) in aortic tissues compared with the sham group without iron, and iron sucrose accelerated MDA level and ICAM-1 expression in the aortic tissues of UNx ApoE^−/− mice. Iron sucrose increased the trend of VCAM-1 expression in UNx ApoE^−/− mice, but this effect did not reach statistical significance (Figure 5D).

Figure 5.

Intraperitoneal iron sucrose exacerbated lipid peroxidation and atherosclerosis in UNx ApoE^−/− mice. (A) Left panel indicates representative images of artherosclerotic lesions and iron deposition in sham or UNx mice; right panel indicates quantitative data of the effect of iron on artherosclerotic lesions in sham or UNx mice. Quantitative data of (B) CL counts and (C) lipid peroxidation from aortic tissue lysates in sham or UNx mice. Expression of (D) VCAM-1 and (E) ICAM-1 in aortic tissue lysates in sham or UNx mice. Data are expressed as means±SEMs. Numbers of animals examined in each group were 6 in the sham group, 9 in the UNx group, and 10 in the UNx+iron group. H&E, hematoxylin and eosin.

Intravenous Iron Sucrose Increased ROS Production, Soluble Adhesion Molecules, and Enhanced Mononuclear–Endothelial Adhesion in CKD Patients

Healthy subjects (n=20) and stage 5 CKD patients receiving (n=20) and not receiving (n=20) intravenous iron sucrose administration were recruited in this experiment. CKD patients did not differ significantly in age or sex. Two weeks after a total of 1000 mg iron sucrose supplementation, CKD patients significantly had the highest basal and activated counts of intracellular superoxide production by circulating MNCs compared with CKD patients without iron supplementation and healthy subjects (Figure 6A). Basal counts of superoxide production in CKD patients receiving iron, CKD patients without iron, and healthy controls were 22.5 (interquartile range [IQR]=20.1–24.3)×10³, 21.3 (IQR=17.9–22.6)×10³, and 17.7 (IQR=16.6–22.9)×10³ counts/s (ANOVA: P<0.05), respectively. Similarly, soluble VCAM-1 and ICAM-1 levels in serum and MNC–endothelial cell adhesion using an ex vivo assay were the highest in CKD patients receiving intravenous iron followed by CKD patients without iron supplementation and healthy subjects (Figure 6, B–D). In addition, serum iron parameters of these three groups are shown in Supplemental Table 2. As expected, serum levels of iron and ferritin were highest in the iron-treated CKD patients.

Figure 6.

Intravenous iron sucrose increased intracellular superoxide production and endothelial damage in CKD patients. Activated intracellular production of superoxide in (A) circulating MNCs, plasma levels (s) of (B) VCAM-1 and (C) ICAM-1, and (D) endothelial adhesiveness of circulating MNCs in healthy subjects (n=20) and stage 5 CKD patients undergoing maintenance hemodialysis without (n=20) or with (n=20) intravenous iron sucrose therapy. Whisker plots show the 10th, 25th, 50th, 75th, and 90th percentile distributions in each panel.

Discussion

Iron is essential for many physiologic processes, including oxygen transport and enzymatic reactions. Nevertheless, excess iron can damage tissue by promoting the generation of ROS through the Fenton reaction and the iron-catalyzed Haber–Weiss reaction.¹⁶ The link between body iron status and risk of CV disease was first proposed by Sullivan¹⁷ in the early 1980s. The iron hypothesis suggested that iron depletion may protect against ischemic CV disease. Subsequent cohort studies showed that elevated body iron stores are associated with an increased risk of myocardial infarction.^18,19 These observations were supported by several studies finding that increased tissue iron levels occurred in human atherosclerotic plaques.^8,20,21 In CKD patients, the adverse effects of iron supplementation on CV outcomes remain a matter of debate owing to a lack of long-term randomized controlled studies. Although we⁶ and Kalantar-Zadeh et al.⁷ have suggested that high-dose intravenous iron administration is associated with adverse CV outcomes in HD patients, the contention that iron participates in atherosclerosis and subsequent CV disease in CKD remains unconfirmed.^22,23

Until now, evidence from experimental studies to support the iron hypothesis was fragmentary and contradictory. intravenous iron preparations inhibited proliferation and promoted apoptosis of cultured endothelial cells^24,25 as well as increased MNC–endothelial adhesion in vitro^6,26; however, a reliable cell model for atherogenesis^27–29 and the exact signaling pathway have not been fully established. Li and Frei³⁰ have found that iron increased endothelial NOx activity and ROS production in cultured endothelial cells. Additionally, ROS is thought to induce endothelial damage through activation of NF-κB, a major redox-sensitive transcription factor that is a key regulator of cytokines, chemokines, and cellular adhesion molecules.^29,31 Moreover, the signal mechanisms of ROS with vascular inflammation have been documented.³² Previous studies proposed the hypothesis that iron is a pro-oxidant provoking intracellular ROS production, which may induce redox-sensitive transcription pathway activation and adhesion molecule expression, subsequently promoting MNC–endothelial adhesion. This hypothesis is validated by our in vitro experiments (Figures 1 and 2), and the ROS–NF-κB–VCAM/ICAM signaling is further established based on our findings that the iron effects on MNC–endothelial adhesion were abrogated by N-acetylcysteine (ROS scavenger), apocynin (NOx inhibitor), p22^phox, or NF-κB p65 siRNA and anti–ICAM-1/VCAM-1 antibodies in vitro. Taken collectively, iron-induced ROS generation through activation of NAPDH oxidase and its downstream redox-sensitive transcription factors may initiate early vascular inflammation and endothelial damage.

To our knowledge, this study is the first to show the involvement of iron in early endothelial damage and subsequent atherosclerosis in mice with renal dysfunction. During the initiation of atherogenesis, an in vivo leukocyte–endothelium adhesion model was established in SNx wild-type mice with intraperitoneal administration of iron. As expected, iron sucrose exacerbated early vascular damage in mice with remnant kidney (Figures 3 and 4). However, wild-type rodents seldom developed atherosclerotic plaques, making the long-term impact of iron on the acceleration of atherosclerosis difficult to validate. Genetically engineered ApoE^−/− mice, which have delayed lipoprotein clearance, developed hyperlipidemia and atherosclerosis, generalizing similar features of human atherosclerosis.³³ Therefore, ApoE^−/− mice were used in our study. Atherosclerotic process has been known to be accelerated by reduction of renal parenchyma in the UNx ApoE^−/− mice,^34–37 but the mortality rate for UNx is lower compared with that for SNx. Moreover, the UNx ApoE^−/− mice exhibited significant atherosclerotic lesions compared with the sham-operated mice.^34,35 Accordingly, we chose the UNx ApoE^−/− mice model to validate where iron could exacerbate the atherosclerosis in ApoE^−/− mice at age of 21 weeks. Likewise, more advanced atherosclerotic lesions developed in UNx ApoE^−/− mice, and parenteral iron supplementation further accelerated the atherosclerosis progression in UNx ApoE^−/− mice. Our findings clearly support our hypothesis that iron initiates early atherogenesis and then, accelerates the progression of atheroscerosis.

From clinical viewpoints, several issues merit discussion in this study. NOx-derived ROSs, such as superoxide radicals, have been implicated in the progression of human atherosclerotic lesions.³⁸ Literature has revealed a close link between iron status and NOx activity in clinical situations. Investigators reported that NOx activity was significantly lower in patients with iron-deficient anemia and increased after iron supplementation.³⁹ Because CKD is a proinflammatory status and associated with aggravated oxidative stress, the superoxide production of MNCs in CKD patients is easily primed by proinflammatory cytokines in vivo or PMA in vitro. In our study, for augmenting superoxide production of the isolated MNCs, PMA was used in activation of MNCs ex vivo to mimic the CKD inflammatory status in vivo. We showed that intravenous iron significantly enhanced activated superoxide production of circulating MNCs in CKD patients. Our finding corroborates the previous study. We further found that leukocyte–endothelial cell adhesiveness was modulated by VCAM-1 and ICAM-1 (Figures 2, 4, and 5), indicating that iron supplementation plays a pivotal role in vascular inflammation. Some studies disclosed that endothelial activation and inflammation occurred in early atherosclerosis, and high levels of circulating adhesion molecules were associated with future CV events in non-CKD patients.^40,41 In CKD patients, prospective studies showed that increased serum-soluble adhesion molecules were associated with carotid intima media thickness, a surrogate marker of preclinical atherosclerosis.^42,43 Moreover, elevated serum levels of soluble adhesion molecules predicted mortality in predialysis CKD patients⁴⁴ and patients undergoing HD⁴⁵ and peritoneal dialysis.⁴⁶ Our clinical study showed that intravenous iron significantly increased serum levels of adhesion molecules and endothelial adhesion of circulating MNCs in CKD patients (Figure 6). The iron-related detrimental outcome was not followed in this short-term study; however, Yin et al.⁴⁷ indicated that endothelial adhesiveness of circulating MNCs was related to major adverse cardiac events in patients with chronic heart failure during a median follow-up period of 182 days.

In conclusion, our in vitro experiments showed that iron sucrose increased adhesion molecule expression and aggravated MNC–endothelial cell adhesion by NOx/NF-κB/cell adhesion molecule signaling in cultured HAECs. In the animal models of remnant kidney, parenteral iron sucrose increased mononuclear–endothelial adhesion in SNx wild-type mice and exacerbated atherosclerosis in UNx ApoE^−/− mice. intravenous iron sucrose is also an in vivo pro-oxidant to enhance superoxide production by circulating MNCs, soluble adhesion molecule levels in plasma, and endothelial adhesion of circulating MNCs in CKD patients. Our study highlights a novel causative role for therapeutic iron in CV complications of CKD patients. Likewise, our results, together with our previous study findings, propose the compelling evidence of potential adverse CV outcomes with excessive intravenous iron administered to highly vulnerable CKD patients.

Concise Methods

In Vitro Studies

Reagents

Information on reagents and assay kits is provided in Supplemental Material.

Cell Culture

HAECs and U937 cells were cultured as described in the work by Chen et al.²⁷ (Supplemental Material).

Iron Staining

Iron deposits in HAECs were examined using Perls’ Prussian blue reaction. Cells were incubated with 4% potassium ferrocyanide in 4% hydrochloric acid for 30 minutes and counterstained with neutral red.

In Vitro Mononuclear–Endothelial Cell Adhesion Assay

To explore the effect of iron sucrose on mononuclear–endothelial cell adhesion, we examined the adherence of U937 cells to HAECs under static conditions. Adhesion assays were performed as previously described (Supplemental Material).⁶

Detection of Intracellular ROS Production

The effect of iron sucrose on ROS production in HAECs was determined with fluorometric assay using 2′,7′-dichlorofluorescein diacetate as the probe (Supplemental Material).²⁷

Measurement of NOx Activity

NOx activity was assessed using a CL-based assay. Briefly, the cell lysates were subjected to 1000×g (4°C; 10 minutes) to remove cell debris. The chemiluminescent probe lucigenin (Sigma-Aldrich) was added, resulting in a final concentration of 5 µM. After an incubation period of 20 minutes, the reaction was started by the addition of 100 μM NADPH (Sigma-Aldrich). NOx activity was measured by a luminescence plate reader (Victor II).

siRNA Transfection

HAECs were transfected with control siRNA, p22^phox siRNA, or NF-κB p65 siRNA (Santa Cruz Biotechnology) using siRNA Transfection Reagent in siRNA Transfection Medium (Santa Cruz Biotechnology) for 6 hours and serum-containing medium for 16 hours. Cells were refed with fresh medium, and experiments were performed 24 hours later.

Assays of NF-κB Binding Activity

We measured the capacity of NF-κB to bind to consensus oligonucleotides in vitro by assaying nuclear lysates using DNA binding ELISA following the manufacturer’s instructions (Active Motif).

Western Blot

Protein extracts were prepared as previously described (Supplemental Material).⁶

In Vivo Studies

Mice

All animal experiments were approved by the Animal Care and Utilization Committee of National Yang-Ming University. To evaluate the acute effects of parenteral iron on mononuclear–endothelium adhesion in mice with renal dysfunction, we established four groups of mice (n=10 in each group): sham with saline, sham with iron sucrose, SNx mice with saline, and SNx mice with iron sucrose. C57BL/6 mice were from the National Laboratory Animal Center. SNx was induced in 8-week-old male C57BL/6 mice using a two-step surgical nephrectomy as previously described.²⁸ Briefly, under intraperitoneal anesthesia with sodium pentobarbital at 65 mg/kg, two of three branches of the left renal artery were ligated through a lateral incision. One week after the first operation, the right kidney was removed after ligation of the renal blood vessels and ureter under anesthesia. Eight weeks after the procedure, BUN was measured. Mice with BUN values of 48 and 109 mg/dl were allocated to experimental groups. Previous study implies that an iron sucrose-loaded rodent model by intraperitoneal administration may reflect well the results from intravenous administration in human.⁴⁸ Moreover, the tail veins of mice are too small to endure repeated intravenous injections. Therefore, iron sucrose (2 mg/25 g) or 0.9% saline (2 ml/25 g) was administered intraperitoneal one time per day for 5 days as previously described.¹⁴ Later, the experimental mice underwent in vivo fluorescence-labeled exogenous U937–endothelium assay. Blood samples were also taken, and retrograde perfusion by ice-cold NaCl was performed. A 2-mm segment of the aorta approximately 3 mm distal of the aortic valve at the side of the aortic arch was prepared and embedded in paraffin. Sections (6-µm thick) were cut and stained with hematoxylin and eosin or immunohistochemical staining. The number of in vivo leukocyte–endothelium adhesions at the aortic sinus was counted directly. Meanwhile, mouse aorta was further homogenized for measurement of tissue superoxide production using a CL analyzing system and adhesion molecules using cell lysate ELISA.

To evaluate the cumulative dose effect of iron on aortic atherosclerosis in mice with renal dysfunction, we devised three ApoE^−/− mice groups: sham with saline (n=6), UNx mice with saline (n=9), and UNx mice with iron sucrose (n=10). For UNx, mice underwent left sham operations or left nephrectomy at 8 weeks of age. One week after the procedure, intraperitoneal iron sucrose (2 mg/25 g) or 0.9% saline (2 ml/25 g) was administered weekly for 12 weeks in UNx mice. The ApoE^−/− mice were euthanized at age 21 weeks. The atherosclerotic plaque areas at the aortic sinus were measured. Meanwhile, the remaining aorta, including the distal part, was further homogenized for measurement of tissue superoxide production using a CL analyzing system, lipid peroxidation, and adhesion molecules using cell lysate ELISA. Homozygous ApoE^−/− mice of C57BL/6 background were purchased from The Jackson Laboratory, bred, and maintained under conventional housing conditions in our animal facility.

In Vivo Mononuclear–Endothelial Cell Adhesion Assay

Adherence of U937 cells to vehicle or iron-treated sham or SNx mice was examined under static conditions. For fluorescein staining, U937 cells were incubated with 2′,7′-bis(2-carboxyethyl)-5(6)-carboxyfluorescein acetoxymethyl ester at 37°C for 1 hour, washed, and suspended in RPMI 1640 to a concentration of 10⁷ cells/ml. One hundred microliters labeled U937 (10⁶ cells) or RPMI 1640 only was injected through the tail vein of sham or SNx mice. Sixty minutes after injection, the mice were killed by anesthesia overdose. After perfusion of ice-cold PBS through the left ventricle, the aorta was dissected and homogenized in lysis buffer with proteinase inhibitor. After centrifugation at 24,000×g for 30 minutes at 4°C, the supernatant was used to measure fluorescence intensity (relative fluorescence units) at 485 nm excitation and 530 nm emission with a fluorescence microplate reader (Victor II; PerkinElmer).

Determination of Tissue Superoxide Production

Superoxide production was measured as previously described¹⁵ with minor modification. Aortic tissue was homogenized in 0.1 ml PBS (pH 7.4). CL was measured in a completely dark chamber of a CLD-110 CL analyzing system (Tohoku Electronic Industry Co.). After a 50-second background-level determination, 1.0 ml 0.1 mM lucigenin in PBS (pH 7.4) solution was injected into the sample. CL was monitored continuously for 180 seconds. For quantification of the CL counts, total CL was calculated by integrating the area under the curve and subtracting it from the background level. The results were further normalized by total sample protein and expressed by CL counts per 10 s/μg protein.

Measurement of Lipid Peroxidation

Tissue lipid peroxide levels in mouse aorta were determined as TBA-reactive substances using a reagent kit (Abnova). MDA-TBA adducts formed by the reaction of MDA and TBA under high temperature, and acidic conditions were measured colorimetrically at 530 nm.

Tissue Lysate ELISA of ICAM-1 and VCAM-1

Tissue was homogenized in lysis buffer per the manufacturer’s instructions and stored at −70°C until assay. Mouse VCAM-1 and ICAM-1 levels were quantified with an ELISA kit (Biorbyt or RayBiotech). Assay sensitivity was 5 pg/ml for ICAM-1 and 20 pg/ml for VCAM-1, and the intra-assay coefficient of variation was <10%.

Histologic Examination and Immunohistochemical Assessment

Mouse hearts were harvested and fixed with 4% paraformaldehyde, embedded in paraffin, and serially sectioned at 6 μm. For C57BL/6 mice, section slides of aorta were undergone by hematoxylin and eosin or immunohistochemical staining. To detect the expression of p22^phox in aorta, sections were deparaffinized, rehydrated, and covered with 3% H₂O₂ for 10 minutes. After blocking with BSA, slides were incubated with anti-p22^phox antibody (Santa Cruz Biotechnology) overnight at 4°C and then, the corresponding secondary antibody for 1 hour. Antigenic sites were visualized by the addition of 3,3′-diaminobenzidine. and nuclear sites were counterstained with methyl green. For a negative control, the primary antibody was replaced with rabbit IgG (data not shown). For ApoE^−/− mice, the atherosclerotic plaque areas at the aortic sinus were quantified using Motic Images Plus 2.0.

Iron Histochemistry

Iron deposits on atherosclerotic lesions were examined with Perls’ Prussian blue reaction using 3,3′-diaminobenzidine intensification (Supplemental Material).

Human Studies

Patients and Study Protocol

A cohort of 60 subjects, including 20 healthy subjects and 40 CKD stage 5 patients receiving maintenance HD for more than 3 months, was recruited for measurement of superoxide production of MNCs and circulating levels of soluble adhesion molecules and ex vivo mononuclear–endothelial cell adhesion assay, which served as an index of early atherogenesis.^6,47 Patients with malignancy, infectious disease, sepsis, or hepatobiliary disease were excluded. All CKD patients were randomly allocated into two groups receiving intravenous iron sucrose or normal saline. intravenous iron sucrose (100 mg elemental iron diluted in 250 ml 0.9% saline) or 250 ml 0.9% saline (control) was administered for 60 minutes postdialysis weekly for 10 weeks. For CKD patients, blood samples were taken 2 weeks after administration of the last iron dose. The study was approved by the Committee on Human Research at Taipei Tzu Chi General Hospital. Written informed consent was obtained from each subject before enrollment.

Isolation of PBMCs

Circulating MNCs from saline or iron-treated CKD patients and healthy subjects were isolated and extracted using density ultracentrifugation⁶ with minor modifications (Supplemental Material).

Assessment of Superoxide Generated by MNCs

Using ultraweak and luminol-enhanced CL, superoxide generation by circulating MNCs was measured as previously described.⁶ Briefly, after blood sampling, MNCs were immediately isolated by centrifugation (1500×g) of whole blood in CPT tubes. The MNC suspension was then adjusted to 10⁶ cells/ml (average lymphocytes:monocytes was 93:7), which was determined using a Coulter counter (STKS; Coulter Electronics). After adding 1 ml solution of 0.25 mm lucigenin to 100 μl cell suspension, the photon emission at 200–750 nm was measured as a basal photon count using a BJL-Ultra-Weak CL Analyzer (American Biologics; sensitivity=1.85×10⁻¹⁷ W/cm² count) and recorded as the basal superoxide generated by MNCs. MNCs (suspended at 1×10⁶/ml) were then activated by adding PMA (1 μg/ml), and the activated photon counts were recorded for 30 minutes. The total count of superoxide was the sum of the basal and activated counts.

Measurement of Plasma Levels of Soluble Adhesion Molecules

Plasma levels of human soluble VCAM-1 and ICAM-1 were determined with ELISA using commercial kits (BioSource) according to the manufacturer’s instructions.

Statistical Analyses

Results are expressed as the mean±SEM. The Mann–Whitney U test was used to compare two independent groups. The Kruskal–Wallis test followed by Bonferroni post hoc analysis was used for multiple testing. Potential differences among the three patient groups were assessed with multivariate ANOVA.

Disclosures

None.