Hypoxia and low temperature upregulate transferrin to induce hypercoagulability at high altitude

Abstract

Studies have shown significantly increased thromboembolic events at high altitude. We recently reported that transferrin could potentiate blood coagulation, but the underlying mechanism for high altitude-related thromboembolism is still poorly understood. Here, we examined the activity and concentration of plasma coagulation factors and transferrin in plasma collected from long-term human residents and short-stay mice exposed to varying altitudes. We found that the activities of thrombin and factor XIIa (FXIIa) along with the concentrations of transferrin were significantly increased in the plasma of humans and mice at high altitudes. Furthermore, both hypoxia (6% O2) and low temperature (0°C), 2 critical high-altitude factors, enhanced hypoxia-inducible factor 1α (HIF-1α) levels to promote the expression of the transferrin gene, whose enhancer region contains HIF-1α binding site, and consequently, to induce hypercoagulability by potentiating thrombin and FXIIa. Importantly, thromboembolic disorders and pathological insults in mouse models induced by both hypoxia and low temperature were ameliorated by transferrin interferences, including transferrin antibody treatment, transferrin downregulation, and the administration of our designed peptides that inhibit the potentiation of transferrin on thrombin and FXIIa. Thus, low temperature and hypoxia upregulated transferrin expression-promoted hypercoagulability. Our data suggest that targeting the transferrin-coagulation pathway is a novel and potentially powerful strategy against thromboembolic events caused by harmful environmental factors under high-altitude conditions.

Graphical abstract

Figure 1.

Enhanced thrombin and FXIIa activities, decreased iron levels, and elevated transferrin levels in plasma of long-term human residents and short-stay mice. (A) Graphical representation of plasma collection from long-term human residents and short-stay mice at different altitudes. (B-C) Relative enzymatic activity of thrombin in human (n = 20) (B) and mouse (n = 30) (C) plasma at different altitudes. (D-E) Relative enzymatic activity of FXIIa in human (n = 30) (D) and mouse (n = 30) (E) plasma at different altitudes. (F-G) Iron level in human (n = 111-128) (F) and mouse (n = 30) (G) plasma at different altitudes. (H-I) Transferrin concentration in human (n = 111-128) (H) and mouse (n = 30) (I) plasma at different altitudes. (J-K) Fibrinogen concentration in human (n = 20) (J) and mouse (n = 30) (K) plasma at different altitudes. Each experiment was independently repeated in triplicate. Data represent mean ± SD. Panels B, G, J, and K, ∗P < .05, ∗∗P < .01 by unpaired t test; panels C-F, H, and I, ∗P < .05, ∗∗P < .01 by Mann-Whitney U test.

Figure 2.

Hypoxia- and low temperature–induced HIF-1α activation to promote transferrin expression both in vitro and in vivo. (A-C) After hypoxia treatment (1% O₂, 5% CO₂, and 94% N₂), HIF-1α and transferrin proteins in BNL CL.2 cells were analyzed by western blot analysis (A), and corresponding quantifications are shown in panels B and C. β-actin was used as loading control in panel A (n = 6). (D-F) Western blot detection of HIF-1α and transferrin levels in livers of normal and hypoxia-induced mice and corresponding quantifications (n = 6). (G-I) Western blot analysis of HIF-1α and transferrin levels in liver of normal and low temperature–induced mice and corresponding quantifications (n = 6). β-actin was used as loading control in panels D and G. (J-L) Transferrin levels in supernatant of HepG2 cells transfected by transferrin expression plasmid of wild-type hypoxia response elements (HRE; WT-HRE) or mutated HRE (MT-HRE) were analyzed by western blot analysis (J,K) and ELISA (L) after hypoxia treatment (n = 6). β-actin was used as loading control in panel J. Each experiment was independently repeated in triplicate. Data represent mean ± SD. Panels B, C, E, F, H, I, K, and L, ∗∗P < .01 by unpaired t test. Western blots were from different membranes, and representative blots are shown in panels A, D, G, and J. NC, normal control.

Figure 3.

Hypoxia- and low temperature–induced transferrin upregulation to potentiate enzymatic activities of thrombin and FXIIa. (A-D,G-J) Effects of anti–Tf-AB, IgG control, RNR-Tf virus, blank (RNR) virus, and HIF inhibitor LW6, on HIF-1α and transferrin expression in liver or transferrin expression in plasma of mice following hypoxia and low-temperature treatment was determined by western blotting (A-C,G-I) or ELISA (D,J) (n = 6-7). (E-F,K-L) Relative activities of thrombin (E and K) and FXIIa (F,L) in plasma were examined (n = 6-7). β-actin was used as loading control in panels A and G. Animal experiments were repeated 3 times, independently. Data represent mean ± SD. Each point represents 1 mouse. Panels B-F and H-L, ∗P < .05, ∗∗P < .01 by unpaired t test. Western blots were from different membranes, and representative blots are shown in panels A and G.

Figure 4.

Hypoxia-induced hypercoagulability and thrombosis aggravation, which were reversed by transferrin knockdown and functional interference. (A-C) Effects of anti–Tf-AB, IgG control, RNR-Tf virus, blank (RNR) virus, HIF inhibitor LW6, or peptides TH16 and FX18 on APTT (A), PT (B), and bleeding time (C) in hypoxia-treated mice (n = 6-8). (D-E) Representative images of carotid artery blood flow (left) in FeCl₃-treated mice by laser speckle perfusion imaging (D), with region of interest (rectangle in white) placed in carotid artery to quantify blood flow change (E) (n = 6). Color bar at bottom indicates perfusion unit scale (0-989); scale bar represents 1 mm. Mice were subject to inferior vena cava (IVC) stenosis for 24 hours to evaluate venous thrombogenesis. (F-G) The pathological changes were observed through hematoxylin and eosin staining (F) and calculating thrombus weight (G) (n = 6-7); scale bar represents 200 μm. (H-I) Representative images of TTC-stained coronal brain sections (H) and quantitative analysis of stained area (I) on day 1 after tMCAO. Ischemic infarctions appear white, and brain infarct volumes were measured by planimetry (percentage of whole volume). (J-K) Bederson (J) and grip test (K) scores were also measured (n = 6-7); scale bar represents 0.5 cm. Animal experiments were repeated 3 times, independently. Data represent mean ± SD. Each point represents 1 mouse. Panels A-C, E, G, and I-K, ∗P < .05, ∗∗P < .01 by unpaired t test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups.

Figure 4.

Hypoxia-induced hypercoagulability and thrombosis aggravation, which were reversed by transferrin knockdown and functional interference. (A-C) Effects of anti–Tf-AB, IgG control, RNR-Tf virus, blank (RNR) virus, HIF inhibitor LW6, or peptides TH16 and FX18 on APTT (A), PT (B), and bleeding time (C) in hypoxia-treated mice (n = 6-8). (D-E) Representative images of carotid artery blood flow (left) in FeCl₃-treated mice by laser speckle perfusion imaging (D), with region of interest (rectangle in white) placed in carotid artery to quantify blood flow change (E) (n = 6). Color bar at bottom indicates perfusion unit scale (0-989); scale bar represents 1 mm. Mice were subject to inferior vena cava (IVC) stenosis for 24 hours to evaluate venous thrombogenesis. (F-G) The pathological changes were observed through hematoxylin and eosin staining (F) and calculating thrombus weight (G) (n = 6-7); scale bar represents 200 μm. (H-I) Representative images of TTC-stained coronal brain sections (H) and quantitative analysis of stained area (I) on day 1 after tMCAO. Ischemic infarctions appear white, and brain infarct volumes were measured by planimetry (percentage of whole volume). (J-K) Bederson (J) and grip test (K) scores were also measured (n = 6-7); scale bar represents 0.5 cm. Animal experiments were repeated 3 times, independently. Data represent mean ± SD. Each point represents 1 mouse. Panels A-C, E, G, and I-K, ∗P < .05, ∗∗P < .01 by unpaired t test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups.

Figure 5.

Low temperature–induced hypercoagulability and thrombosis aggravation, which were reversed by transferrin knockdown and functional interference. (A-C) APTT (A), PT (B), and bleeding time (C) in mice are shown (n = 6-7). (D-E) Representative images of carotid artery blood flow in FeCl₃-treated mice by laser speckle perfusion imaging (D), with region of interest (rectangle in white) placed in carotid artery to quantify blood flow change (E) (n = 6). Color bar at bottom indicates perfusion unit scale (0-989); scale bar represents 1 mm. Mice were subject to IVC stenosis for 24 hours to evaluate venous thrombogenesis. (F-G) The pathological changes were observed through hematoxylin and eosin staining (F) and calculating thrombus weight (G) (n = 6-7); scale bar represents 200 μm. (H-I) Representative images of TTC-stained coronal brain sections (H) and quantitative analysis of stained area (I) on day 1 after tMCAO. Ischemic infarctions appear white, and brain infarct volumes were measured by planimetry (percentage of whole volume). (J-K) Bederson (J) and grip-test (K) scores were also measured (n = 6). (L) Survival rate of mice in anti-Tf-AB–, RNR-Tf–, RNR virus–, and TH16- and FX18-peptide–treated groups under low temperature (n = 20). (M) Graphical representation of iron and O₂ replenishment at high altitude, contributing to thromboembolic disorders. Detrimental environmental factors (hypoxia and low temperature) and high altitude–induced iron deficiency increased HIF-1α levels, which upregulated transferrin to increase iron transport to compensate for erythropoiesis and O₂ supply. Simultaneously, abnormally upregulated transferrin caused hypercoagulability by interacting with multiple plasma proteins to potentiate thrombin and FXIIa and inhibit antithrombin. Animal experiments were repeated 3 times, independently. Data represent mean ± SD. Each point represents 1 mouse. Panels A-C, E, G, and I-K, ∗P < .05, ∗∗P < .01 by unpaired t test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups. Panel L, ∗P < .05 by log-rank test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups.

Figure 5.

Low temperature–induced hypercoagulability and thrombosis aggravation, which were reversed by transferrin knockdown and functional interference. (A-C) APTT (A), PT (B), and bleeding time (C) in mice are shown (n = 6-7). (D-E) Representative images of carotid artery blood flow in FeCl₃-treated mice by laser speckle perfusion imaging (D), with region of interest (rectangle in white) placed in carotid artery to quantify blood flow change (E) (n = 6). Color bar at bottom indicates perfusion unit scale (0-989); scale bar represents 1 mm. Mice were subject to IVC stenosis for 24 hours to evaluate venous thrombogenesis. (F-G) The pathological changes were observed through hematoxylin and eosin staining (F) and calculating thrombus weight (G) (n = 6-7); scale bar represents 200 μm. (H-I) Representative images of TTC-stained coronal brain sections (H) and quantitative analysis of stained area (I) on day 1 after tMCAO. Ischemic infarctions appear white, and brain infarct volumes were measured by planimetry (percentage of whole volume). (J-K) Bederson (J) and grip-test (K) scores were also measured (n = 6). (L) Survival rate of mice in anti-Tf-AB–, RNR-Tf–, RNR virus–, and TH16- and FX18-peptide–treated groups under low temperature (n = 20). (M) Graphical representation of iron and O₂ replenishment at high altitude, contributing to thromboembolic disorders. Detrimental environmental factors (hypoxia and low temperature) and high altitude–induced iron deficiency increased HIF-1α levels, which upregulated transferrin to increase iron transport to compensate for erythropoiesis and O₂ supply. Simultaneously, abnormally upregulated transferrin caused hypercoagulability by interacting with multiple plasma proteins to potentiate thrombin and FXIIa and inhibit antithrombin. Animal experiments were repeated 3 times, independently. Data represent mean ± SD. Each point represents 1 mouse. Panels A-C, E, G, and I-K, ∗P < .05, ∗∗P < .01 by unpaired t test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups. Panel L, ∗P < .05 by log-rank test to compare; for example, Tf-AB and IgG groups or RNR-Tf and RNR groups.