Endothelial protein kinase MAP4K4 promotes vascular inflammation and atherosclerosis

Abstract

Signalling pathways that control endothelial cell (EC) permeability, leukocyte adhesion and inflammation are pivotal for atherosclerosis initiation and progression. Here we demonstrate that the Sterile-20-like mitogen-activated protein kinase kinase kinase kinase 4 (MAP4K4), which has been implicated in inflammation, is abundantly expressed in ECs and in atherosclerotic plaques from mice and humans. On the basis of endothelial-specific MAP4K4 gene silencing and gene ablation experiments in Apoe(-/-) mice, we show that MAP4K4 in ECs markedly promotes Western diet-induced aortic macrophage accumulation and atherosclerotic plaque development. Treatment of Apoe(-/-) and Ldlr(-/-) mice with a selective small-molecule MAP4K4 inhibitor also markedly reduces atherosclerotic lesion area. MAP4K4 silencing in cultured ECs attenuates cell surface adhesion molecule expression while reducing nuclear localization and activity of NFκB, which is critical for promoting EC activation and atherosclerosis. Taken together, these results reveal that MAP4K4 is a key signalling node that promotes immune cell recruitment in atherosclerosis.

Figure 1. Increased MAP4K4 expression in atherosclerosis.

(a) Eight-to-ten-week-old mice were fed chow or 60% HFD for 16 weeks, messenger RNA (mRNA) was extracted from the indicated tissues, and quantitative RT–PCR was performed for Map4k4 and normalized to 36b4. The data represent the mean±s.e.m. (*P<0.05, **P<0.005, N=3–7). (b–d) Aortas were extracted from age-matched chow-fed wild-type or Apoe^−/− mice or WD-fed Apoe^−/− mice. (b) Immune-complex kinase assays were performed in Map4k4 immunoprecipitates using MBP as an exogenous substrate. Lysates were immunoblotted for tubulin as a loading control. (c) Densitometric quantification of ³²P MBP as normalized to tubulin. (d) Densitometric quantification of immunoprecipitated Map4k4 as normalized to tubulin (analysis of variance ⁺P=0.05, *P<0.05, N=4–5). (e) mRNA was isolated from normal human arteries or atherosclerotic plaques, and quantitative RT–PCR was performed for MAP4K4 or GAPDH (*P<0.05, N=3–5).

Figure 2. Reduced atherosclerosis in M4K4 ECKO mice.

Map4k4 flox/flox animals were crossed with Cdh5(PAC)-ERT2-Cre animals and injected with tamoxifen for 5 consecutive days at 6–8 weeks of age. (a) Cre-mediated Map4k4 exon-7 deletion. (b) Schematic of injection and feeding scheme. (c) Messenger RNA was extracted and qRT–PCR was performed for Map4k4 in primary MLECs, the unselected, non-EC fraction of mouse lung cells and peripheral blood leukocytes. The data represent the mean±s.e.m. as normalized to 36b4 expression (*P<0.05, N=6–8). (d) Immunoblots were performed for Map4k4 or tubulin in immune-selected primary MLECs and the unselected, non-EC fraction of mouse lung cells. Western blots are representative of 6–8 animals per group. (e–j) Flox/flox and MAP4K4 ECKO mice were crossed with Apoe^−/− mice and fed a WD for 16 weeks as in b. (e) Left, Oil Red-O-stained en face aortic preparations from flox/flox and MAP4K4 ECKO animals. Right, quantification of Oil Red-O-stained area. Data represent the mean±s.e.m. (***P<0.0005, N=8–10). (f–j) Aortic root sections of flox/flox and MAP4K4 ECKO Apoe^−/− animals stained with (f) haematoxylin and eosin (H&E) (scale bar, 250 μm), (g) Oil Red-O (scale bar, 250 μm), (h) trichrome (scale bar, 250 μm), (i) smooth muscle actin (scale bar, 250 μm (top image); scale bar, 100 μm (bottom image)) and (j) Cd68 (scale bar, 250 μm (top image); scale bar, 100 μm (bottom image)). Left panel, representative images. Right panels, quantification of stained area or as a percentage of lesion area. Data represent the mean±s.e.m. (*P<0.05, **P<0.005, ***P<0.0005, N=8–11). DAPI, 4,6-diamidino-2-phenylindole.

Figure 3. Reduced atherosclerosis in M4K4 KD mice.

(a) Schematic of the transgenic construct used to generate MAP4K4 KD animals. The U6 promoter becomes reconstituted after cre-mediated recombination to drive tissue-specific shRNA expression. (b,c) Primary lung endothelial cells (MLECs) were derived from control and MAP4K4 KD animals. (b) Messenger RNA (mRNA) was extracted and quantitative RT–PCR was performed for Map4k4 in immune-selected or unselected cells. The data represent the mean±s.e.m. as normalized to 36b4 expression (*P<0.05, N=7). (c) Primary endothelial or unselected cell lysates were immunoblotted for Map4k4 and Vegfr2. Data represent the mean±s.e.m. as normalized to tubulin expression (*P<0.05, N=3–6). (d–g) MAP4K4 KD mice were crossed with Apoe^−/− mice and fed a WD for 16 weeks. (d) Aortic root sections of control and MAP4K4 KD animals stained with haematoxylin and eosin (H&E) (top) and Oil Red-O (bottom). Scale bar, 250 μm. (e) Quantification of aortic root lesion area. Data represent the mean±s.e.m. (**P<0.005, N=9,11). (f) Oil Red-O-stained en face aortic preparations from control and MAP4K4 KD animals. (g) Quantification of Oil red-O-stained area. Data represent the mean±s.e.m. (*P<0.05, N=5–7). (h–i) mRNA was prepared from whole aortas, and qPCR was performed. (h) Macrophage markers F4/80 and Cd68. (i) Chemokines Ccl2, Cxcl1, Ccl3, Ccl4, Ccl5, Ccl7, Cxcl9 and Cxcl10. Data represent the mean±s.e.m. as normalized to 36b4 (*P<0.05, **P<0.005, N=9–10).

Figure 4. Less macrophages in plaques from M4K4 ECKO and KD mice.

(a,b) Messenger RNA was prepared from whole aortas, and qPCR was performed. (a) Macrophage markers F4/80 and Cd68. (b) Chemokines Ccl2, Cxcl1, Ccl3, Ccl4, Ccl5, Ccl7, Cxcl9 and Cxcl10. Data represent the mean±s.e.m. as normalized to 36b4 (*P<0.05, **P<0.005, N=9–10). (c,d) Homing of GFP leukocytes into atherosclerotic lesions 48 h after intravenous injection into control or MAP4K4 KD mice that were fed WD for 16 weeks. (c) Fluorescence micrograph of atherosclerotic plaque demonstrating four GFP leukocytes within the aortic arch. The dashed line indicates the plaque border. Inset, magnification of three GFP leukocytes. Left, 4,6-diamidino-2-phenylindole; middle, GFP; right, merge. Scale bars, 100 μm. (d) Quantification of GFP leukocytes per square millimetre of plaque. Data represent the mean±s.e.m. (**P<0.005, N=4,7).

Figure 5. Reduced EC Icam-1 and Vcam-1 in M4K4 KD plaques.

(a,b) Control and M4K4 KD mice were fed WD for 16 weeks. Aortic arch sections were stained for Cd31 as an endothelial marker (green), Icam-1 (red, (a)), Vcam-1 (red, (b)) or DAPI. From left, CD31, Icam-1/Vcam-1, DAPI and merge. Scale bars, 100 μm. Arrowheads indicate Cd31/Icam-1 or Cd31/Vcam-1 colocalization. Images are representative of at least three control and three M4K4 KD animals. DAPI, 4,6-diamidino-2-phenylindole.

Figure 6. Endothelial MAP4K4 promotes EC activation via NF-κB.

HUVECs were treated with scrambled or MAP4K4 siRNA, and cells were stimulated with 1 or 10 ng ml⁻¹ TNF-α for the indicated times. (a) Cells were seeded onto transwell chambers. Confluent cells were treated overnight with 10 ng ml⁻¹ TNF-α or left untreated, and FITC-labelled dextran that migrated through the HUVEC monolayer was measured. The data represent the mean fluorescence intensity±s.e.m. (analysis of variance (ANOVA) *P<0.05, N=4). (b) THP-1 monocytes were stained with calcein green and adhered to activated endothelium for 30 min. Fluorescence microscopy was performed to determine the number of adherent THP-1 monocytes per microscopic field (× 100). The data represent the mean±s.e.m. as normalized to the unstimulated control time point (ANOVA **P<0.01, ****P<0.0001, N=5). (c–f) Messenger RNA was extracted and quantitative RT–PCR was performed for (c) MAP4K4, (d) ICAM-1, (e) VCAM-1 and (f) SELE. The data represent the mean±s.e.m. as normalized to RPLP0 (*P<0.05, **P<0.005, ^#P<0.0001, N=5–7). (g) Immunoblots were performed for ICAM-1, VCAM-1 and E-selectin. (h) Densitometric analyses from g. The data represent the means±s.e.m. as normalized to VE-cadherin (*P<0.05, N=4–6). (i) Biochemical fractionations were performed, and nuclear fractions were immunoblotted for MAP4K4, p-p65, total p65, p50 and lamin-β1. (j) Densitometric analyses represent the mean±s.e.m. for the 60-min time point as normalized to lamin-β1 (*P<0.05, **P<0.005, N=3). (k) NFκB-luciferase and SV40-Renilla were transfected into HUVECs after treatment with scrambled or MAP4K4 siRNA. Cells were left unstimulated or stimulated with 10 ng ml⁻¹ TNF-α overnight before luciferase and Renilla measurement. The data represent the mean±s.e.m. from four independent experiments. (l–n) HUVECs were transfected with MAP4K4 siRNA and stimulated or not with 1 ng ml⁻¹ TNF-α for 1 h. IgG, p65 and histone antibodies were used to immunoprecipitate chromatin and RT–PCR was used to amplify (l) E-selectin, (m) VCAM-1 and (n) IκBα promoters. The data represent the mean±s.e.m. as normalized to input (*P<0.05, N=3–4).

Figure 7. Pharmacological MAP4K4 inhibition ameliorates atherosclerosis.

(a–c) ApoE^−/− mice were administered PF-06260933 or vehicle twice daily for 6 weeks. (a) Dosing regimen (b) Oil red-O-stained en face aortic preparations from vehicle and PF-06260933-treated animals. (c) Quantification of the Oil Red-O-stained area. Data represent the mean±s.e.m. (*P<0.05, N=14–15). (d–f) Ldlr^−/− mice were fed HFD for 10 weeks followed by PF-06260933 or vehicle administration daily for 10 additional weeks. (d) Dosing regimen. (e) Oil Red-O-stained en face aortic preparations from vehicle and PF-06260933-treated animals. (f) Quantification of the Oil Red-O-stained area. Data represent the mean±s.e.m. (**P<0.005, N=6).