Angiopoietin-2 is critical for cytokine-induced vascular leakage

Abstract

Genetic experiments (loss-of-function and gain-of-function) have established the role of Angiopoietin/Tie ligand/receptor tyrosine kinase system as a regulator of vessel maturation and quiescence. Angiopoietin-2 (Ang-2) acts on Tie2-expressing resting endothelial cells as an antagonistic ligand to negatively interfere with the vessel stabilizing effects of constitutive Ang-1/Tie-2 signaling. Ang-2 thereby controls the vascular response to inflammation-inducing as well as angiogenesis-inducing cytokines. This study was aimed at assessing the role of Ang-2 as an autocrine (i.e. endothelial-derived) regulator of rapid vascular responses (within minutes) caused by permeability-inducing agents. Employing two independent in vivo assays to quantitatively assess vascular leakage (tracheal microsphere assay, 1-5 min and Miles assay, 20 min), the immediate vascular response to histamine, bradykinin and VEGF was analyzed in Ang-2-deficient (Ang-2(-/-)) mice. In comparison to the wild type control mice, the Ang2(-/-) mice demonstrated a significantly attenuated response. The Ang-2(-/-) phenotype was rescued by systemic administration (paracrine) of an adenovirus encoding Ang-2. Furthermore, cytokine-induced intracellular calcium influx was impaired in Ang-2(-/-) endothelioma cells, consistent with reduced phospholipase activation in vivo. Additionally, recombinant human Ang-2 (rhAng-2) alone was unable to induce vascular leakage. In summary, we report here in a definite genetic setting that Ang-2 is critical for multiple vascular permeability-inducing cytokines.

Figure 1. Time and dose dependency of histamine-induced vascular leakage in the tracheal microsphere extravasation assay.

Mice were intravenously injected with increasing concentrations of histamine (0.1, 0.5, 2.5, 12.55, 62.5 and 312.5 µg/g) or saline as negative control together with 100 nm green fluorescent microspheres. Tissue samples were whole-mount stained for CD31 (A) and microspheres area density was quantified as indicated in Materials and Methods (B, C). To establish the dose response of histamine-induced vascular leakage, fluorescent microspheres were injected together with histamine and mice were sacrificed after 1 min (B). To establish the time course of histamine-induced vascular leakage, microspheres were injected after the indicated times of exposure to submaximal concentrations of histamine (62.5 µg/g) and mice were sacrificed within 1 min (C). Values are expressed as mean ± SD (n = 5, saline n = 3). **P<0.01 compared to saline. Scale bar 100 µm.

Figure 2. Histamine-induced vascular leakage is reduced in Ang-2^−/− mice.

WT and Ang-2^−/− mice were intravenously injected with 12.5, 62.5 and 312.5 µg/g histamine or saline as negative control together with 100 nm green fluorescent microspheres. Tracheas were whole-mount stained for CD31 (red, A), and microspheres area density was measured (B). The response to locally administered histamine on vascular leakage was measured using the Miles assay. Increasing concentrations of histamine (25 ng, 125 ng, 625 ng) or saline as negative control, were injected intradermally in WT and Ang-2^−/− mice (C) and dye extravasation was quantified (D). Values are expressed as mean ± SD (n = 5, trachea assay; n = 4, Miles assay). *p<0.05 compared to saline; **p<0.01 compared to saline; #p<0.05 compared to corresponding WT histamine treatment group; ##p<0.01 compared to corresponding WT histamine treatment group. Scale bar 100 µm.

Figure 3. VEGF-induced vascular leakage is reduced in Ang-2^−/− mice.

WT and Ang-2^−/− mice were intravenously injected with 0.25 µg/g VEGF or saline as negative control together with 100 nm green fluorescent microspheres. Tracheas were whole-mount stained for CD31 (red, A), and microspheres area density was measured (B). For Miles assays, WT and Ang-2-deficient mice were intradermally injected with 25 ng VEGF or saline as negative control (C). Skin patches were removed and extravasated dye was measured (D). Results are expressed as mean ± SD (n = 5, trachea assay; n = 3, Miles assay). **p<0.01 compared to saline; ##p<0.01 compared to WT and VEGF, respectively. Scale bar 100 µm.

Figure 4. Bradykinin-induced vascular leakage is reduced in Ang-2^−/− mice.

Bradykinin (100 µg/g) was injected intravenously together with 100 nm fluorescent microspheres in WT and Ang-2^−/− mice. CD31 whole-mount stained tracheas (A) demonstrated higher microsphere accumulation in WT mice compared to Ang-2^−/− mice (B). Miles assays was performed in WT and Ang-2^−/− mice. Bradykinin (100 ng) or saline as negative control were injected intradermally (C). Skin patches were taken and extravasated dye was measured (D). Results are expressed as mean ± SD (n = 5, trachea assay; n = 3, Miles assay). **p<0.01 compared to saline; ##p<0.01 compared to WT and Bradykinin, respectively. Scale bar 100 µm.

Figure 5. Exogenous Ang-2 does not induce vascular leakage.

Increasing concentrations of recombinant Ang-2 (0.1, 0.5 and 1 µg/g) or CHAPS buffer as negative control were intravenously injected together with 100 nm green fluorescent microspheres (A). CD31 whole-mount stained tracheas from WT mice were quantitated for microsphere extravasation. Miles assay was performed using increasing concentrations of Ang-2 (5, 25, 125 ng), Evans blue dye was extracted from skin samples and quantified (B). Results are expressed as mean ± SD (n = 3, trachea assay; n = 5, Miles assay).

Figure 6. Adenoviral delivery of Ad-Ang-2 rescues the loss of histamine response in Ang-2^−/− mice.

WT and Ang-2^−/− mice received an intravenous injection of 10⁹ PFU Ad-LacZ or Ad-Ang-2, respectively. Ang-2 in the serum was semi-quantitatively determined by Western blot after 3 days showing a myc-tagged band of approx. 60 kDa in Ad-Ang-2 treated mice alone (A). Histamine (62.5 µg/g) or saline as negative control were injected together with 100 nm green fluorescent microspheres in either Ad-LacZ or Ad-Ang-2 treated WT and Ang-2^−/− mice. CD31 whole-mount stained tracheas were quantitated for microsphere extravasation (B). Adenoviral delivery of Ang-2 in Ang-2^−/− mice rescued the phenotype of Ang-2^−/− mice and facilitated the histamine response. Moreover, delivery of Ad-Ang-2 in WT mice enhanced the response to histamine. Values are expressed as mean ± SD (n = 6, saline n = 3). **p<0.01 compared to Ad-LacZ and histamine.

Figure 7. Loss of Ang-2 results in attenuated intracellular calcium increase in response to VEGF and histamine in vitro.

Immortalized murine lung endothelial cells from WT and Ang-2^−/− mice (WT EOMA, KO EOMA) were loaded with Fluo-4 calcium sensitive dye and the relative fluorescence (stimulated/unstimulated fluorescence) was quantified in response to VEGF (25 ng/ml) and histamine (10⁻⁵M) after 10 min stimulation. HUVEC served as positive control (n = 3, **p<0.005 WT vs. KO).

Figure 8. Loss of Ang-2 reduces PLCγ2 activation in vivo.

(A) WT and Ang-2^−/− mice were intravenously injected with 62.5 µg/g histamine or saline as negative control. Mice were sacrificed after 90 s. The lungs were excised (1,2,3 represent different mice), and tissue lysates were used for immunoprecipitation and Western blotting. The blots were probed for PLCγ2, phospho-PLCγ2 (Tyr 1217), Tie-2, phospho-tyrosine and actin as indicated. Quantification of the relative intensity of phospho-PLCγ2 to total PLCγ2 (B) and phospho-tyrosine to total Tie-2 (C) revealed a significant reduction in PLCγ2 phosphorylation following histamine stimulation in Ang-2^−/− mice compared to WT mice and no reduction in pTie-2 following histamine in KO mice. (n = 3 animals for each group. *p<0.05). Additionally, HUVEC stimulated with recombinant Ang-2 demonstrated an increase in activation of PLCγ2 compared to unstimulated cells indicating the role of Ang-2 in PLCγ2 activation thereby affecting intracellular calcium influx (D).