Copper Transport Protein Antioxidant-1 Promotes Inflammatory Neovascularization via Chaperone and Transcription Factor Function

Abstract

Copper (Cu), an essential micronutrient, plays a fundamental role in inflammation and angiogenesis; however, its precise mechanism remains undefined. Here we uncover a novel role of Cu transport protein Antioxidant-1 (Atox1), which is originally appreciated as a Cu chaperone and recently discovered as a Cu-dependent transcription factor, in inflammatory neovascularization. Atox1 expression is upregulated in patients and mice with critical limb ischemia. Atox1-deficient mice show impaired limb perfusion recovery with reduced arteriogenesis, angiogenesis, and recruitment of inflammatory cells. In vivo intravital microscopy, bone marrow reconstitution, and Atox1 gene transfer in Atox1(-/-) mice show that Atox1 in endothelial cells (ECs) is essential for neovascularization and recruitment of inflammatory cells which release VEGF and TNFα. Mechanistically, Atox1-depleted ECs demonstrate that Cu chaperone function of Atox1 mediated through Cu transporter ATP7A is required for VEGF-induced angiogenesis via activation of Cu enzyme lysyl oxidase. Moreover, Atox1 functions as a Cu-dependent transcription factor for NADPH oxidase organizer p47phox, thereby increasing ROS-NFκB-VCAM-1/ICAM-1 expression and monocyte adhesion in ECs inflamed with TNFα in an ATP7A-independent manner. These findings demonstrate a novel linkage between Atox1 and NADPH oxidase involved in inflammatory neovascularization and suggest Atox1 as a potential therapeutic target for treatment of ischemic disease.

Figure 1. Atox1 in tissue resident cells, but not BM cells, is required for ischemia-induced neovascularization.

(A) Atox1 protein expression in ischemic and non-ischemic muscles from wild type (WT) and Atox1 knockout (KO) mice at day 7 post-surgery. Actin is loading control (n = 5). (B) Blood flow recovery after hindlimb ischemia in WT and Atox1 KO mice, as determined by the ratio of foot perfusion between ischemic (left) and non-ischemic (right) legs. Bottom panels show representative laser Doppler images (white arrows shows ischemic foot) and toe necrosis characterized by edematous fingers and degenerative nail beds (black arrows) in Atox1 KO mice (n = 10). (C) Bone marrow transplantation (BMT) showing a role of Atox1 in tissue resident cells for ischemia-induced neovascularization. After 6 weeks of BMT, mice were subjected to hindlimb ischemia and limb blood flow was measured (n = 6). (D) Atox1 gene transfer in Atox1 KO mice rescues blood flow recovery after hindlimb ischemia. Purified adenoviruses (Ad.Atox1-WT, Ad. LacZ (control), 1 × 10⁹ pfu) were injected into the adductor and gastrocnemius muscles in Atox1 KO mice at one day prior to ischemic injury, and limb blood flow was measured. Representative laser Doppler images (white arrows shows ischemic foot) and quantitative analysis are shown (C,D). **p < 0.01, ***p < 0.001.

Figure 2. Atox1 is required for ischemia-induced arteriogenesis and angiogenesis.

(A) Representative micro-computed tomography (micro-CT) angiograms (16 μm resolution; arteriogenesis) of ischemic legs in WT and Atox1 KO mice at day 7 post-surgery. Bottom panel shows quantitative analysis of micro-CT, as the total number of vascular structures in 1355 z-axis slices over the ranges of sizes from 4 different mice. (B,C) Representative images of CD31 (EC marker)(B) and αSMA (arteriole)(C) staining in ischemic and non-ischemic gastrocnemius muscles of WT and Atox1 KO mice at day 7. Scale bars = 50 μm. Bottom panels show their quantitative analysis (n = 4). *p < 0.05.

Figure 3. Atox1 promotes angiogenesis via activating Cu enzyme lysyl oxidase in ECs in an ATP7A-dependent manner.

(A) Sponge implant assay was performed by implanting polyvinyl alcohol sponge containing VEGF subcutaneously into WT and Atox1 KO mice. Representative images for H&E staining and isolectin immunostaining for blood vessel formation in sponges harvested on day 21. Right panels show quantitative analysis of the number of red blood cells (RBC) and isolection+ ECs. Scale bars = 50μm. (B) HUVECs were transfected with control, Atox1 or ATP7A siRNAs or treated with Cu chelator TTM (20 nM, 24 hrs) and seeded on Matrigel-coated plates in culture media containing VEGF for 6 h. Four random fields per well were imaged, and representative pictures are shown (left). Averaged numbers of capillary tube branches, branching points, and tube length per field are shown (Right). (C) HUVECs were treated with LOX inhibitor β-aminopropionitrile (BAPN, 100 μM) for 24 hrs and capillary tube formation was measured (n = 3). (D) Activity of LOX was measured in ischemic gastrocnemius muscle of WT and Atox1 KO mice (left)(n = 4) or in culture medium from VEGF (20ng/ml)-stimulated HUVECs transfected with siControl or siAtox1 (right)(n = 4). *p < 0.05.

Figure 4. Atox1 is required for inflammatory cell recruitment to the ischemic tissues or inflamed ECs.

(A) Representative images (left) and quantification (right) for Mac3 (for macrophage) and Gr1 (for neutrophils) staining in ischemic muscles in WT and Atox1 KO mice at day 3 (n = 4). (B,C) Expression of TNFα and IL-6 mRNAs (B), (n = 4) and VEGF protein (C), (n = 3) in non-ischemic and ischemic muscles of WT and Atox1 KO mice at day 3. (D) Intravital microscopy analysis for neutrophil rolling and adhesion on TNFα-inflamed endothelium in vivo. At 3 hours after TNFα injection to WT and Atox1 KO mice, the cremaster muscle was exposed. Mouse neutrophils were monitored in the inflamed cremaster muscle venules by infusion of Alexa 647-labeled anti-Gr-1. Time lapse are shown in WT (adherent cells shown in white arrows) vs. Atox1 KO mice (the rolling cells shown in black and white arrows). In bottom, quantification of the number of adherent cells over 5 min and the number of detaching cells after firm adhesion, rolling influx (cells/min) and velocity (mm/sec) of WT and Atox1 KO neutrophils. (n = 15–16 venules in WT and Atox1 KO mice). (E) Confluent HUVEC monolayers transfected with Atox1, CTR1, ATP7A, or control siRNAs or treated with the Cu chelator BCS (200 μM) for 48 hrs were stimulated with TNFα (10 ng/ml) for 18 hours. The numbers of bound THP1 monocytes (fluorescently labeled) to ECs were measured with a fluorescence microscope. Bottom panel showed quantification (n = 15). *p < 0.05, **p < 0.01, ***p < 0.001

Figure 5. Atox1 is required for adhesion molecules expression in a Cu-dependent manner via activating NFκB in inflamed ECs.

(A) Representative images for VCAM-1 staining in ischemic adductor muscles in WT and Atox1 KO mice at day 1 (n = 4). Scale bars = 100μm. (B) Expression of ICAM-1 and VCAM-1 mRNAs in ischemic and non-ischemic muscles at day 3 (n = 4). (C) HUVECs transfected with Atox1 or control siRNAs, or treated with BCS were stimulated with TNFα (10ng/ml), and ICAM-1 and VCAM-1 mRNAs were measured (n = 12). (D) HUVECs transfected with Atox1 or control siRNAs were incubated with TNFα for 18 hours, and VCAM-1 and ICAM-1 protein expression was measured. Actin was loading control (n = 8). (E) ChIP assay showing a role of Atox1 in TNFα-induced p65 NFkB binding to the ICAM-1/VCAM-1 promoter in vivo. HUVECs transfected with Atox1 or control siRNAs stimulated with TNFα were precipitated with the anti-p65NFkB antibody. The ICAM-1/VCAM-1 promoter region was amplified by PCR. Input of nuclear DNA was used as PCR control. (n = 4). (F) In vivo bioluminescence imaging of NFkB reporter mice (HLL mice) and HLL mice crossed with Atox1 KO mice before and after hindlimb ischemia at day 3. (n = 3). Representative images (left) and quantification of bioluminescence intensity (right). (G) Luciferase activity in homogenates of non-ischemic and ischemic gastrocunemious tissue from HLL and HLL/Atox1 KO mice at day 3 (n = 5). *p < 0.05, **p < 0.01, ***p < 0.001

Figure 6. Atox1 is involved in endothelial ROS production in ischemic tissues and ECs stimulated by TNFα in a Cu-dependent and ATP7A-independent manner.

(A) O₂^•- detection probe, dihydroethidium (DHE) was injected into mice at 30 min before sacrifice at day 7, and tissues were stained with CD31. Representative pictures for CD31 staining, DHE fluorescence, and their merged images in ischemic muscles in WT and Atox 1 KO mice (10μm thickness). DHE+/CD31+ double positive cells (yellow) are shown in white arrows. Bars = 20μm. Lower panel shows quantification of number of CD31+ cells and ratio of CD31+/DHE+ cells (n = 3). (B) HUVECs transfected with siRNAs for Atox1, CTR1, ATP7A, control, or treated with either BCS (200uM for 48hrs) or PEG-catalase (500U/ml for 1 hr) were stimulated with TNFα (10ng/ml) for 18hrs. Representative images for DCF fluorescence and DAPI staining (blue, nucleus marker) (left) and quantification of fluorescence intensity (right) (n = 3). *p < 0.05 and ***p < 0.001.

Figure 7. Atox1-mediated p47phox transcription is required for adhesion molecule expression in a Cu-dependent manner in ECs.

(A) HUVECs transfected with Atox1 or control siRNAs, or treated with BCS (200uM for 48 hours) were stimulated with TNFα (10ng/ml), and p47phox and p22phox mRNAs were measured (n = 4). (B,C) p47phox protein expression in HUVECs transfected with Atox1 or control siRNAs (B) or ECs isolated from WT and Atox1 KO mice stimulated with TNFα (C)(n = 3). (D) ECs isolated from WT and Atox1 KO mice were transfected p47 phox-EGFP plasmid, and stimulated with TNFα for 18 hrs. Lysates were used to measure VCAM-1, p47phox-EGFP, and actin proteins by western analysis (n = 3). *p < 0.05, ***p < 0.001.

Figure 8. Atox1 promotes p47phox transcription by binding to its promoter in ECs activated by TNFα in a Cu-dependent manner.

(A) Immunofluorescence staining of Atox1. HUVECs stimulated with TNFα (10 ng/ml) pretreated with or without Cu chelator BCS were immunostained for Atox1 or nuclear marker DAPI. Percentage of Atox1 + cells in nucleus was shown in right (n = 3). (B) Identification of Atox1 responsive elements (Atox1-RE) in a p47phox promoter. HEK293 cells were transiently transfected with indicated promoter luciferase constructs along with pcDNA/Atox1-WT or mutants Atox1 mutated at nuclear translocation signal (K56,60E) or Cu-binding domains (C12,15S). Two days after transfection, the luciferase activity was measured (n = 3). (C) HUVECs were transfected with a p47phox promoter luciferase reporter construct (pGL3-p47phox promoter (−1217/+52)) with or without mutation or deletion of the Atox1-RE (−110 to −105 region) (n = 3). (D) ChIP assay showing Atox1 binding to the p47phox promoter in vivo. HUVECs transfected with Atox1 or control siRNAs were treated with TNFα (10ng/ml). After precipitation with anti-Atox1 antibody, the p47phox promoter region (the Atox1-RE (−110 to −105 region)) was amplified by PCR (n = 4). Input nuclear DNA was used as PCR control. *p < 0.05. ^#p < 0.05, **p < 0.01, ***p < 0.001. (E) Models for role of Atox1 in inflammatory neovascularization, which is dependent on arteriogenesis/angiogenesis and inflammation. Atox1 functions as a Cu chaperone mediated through ATP7A to increase LOX activity involved in VEGF-induced angiogenesis as well as a Cu-dependent transcription factor for NADPH oxidase organizer p47phox to increase the ROS-NFkB-VCAM-1/ICAM-1 axis in ECs inflamed with TNFα. This in turn promotes recruitment of inflammatory cells which secrete TNFα and VEGF. This represents a novel positive feedback loops whereby Cu transport protein Atox1 promotes inflammatory neovascularization.