Endothelial insulin-like growth factor-1 signalling regulates vascular barrier function and atherogenesis

Abstract

Aims: Progressive deposition of cholesterol in the arterial wall characterizes atherosclerosis, which underpins most cases of myocardial infarction and stroke. Insulin-like growth factor-1 (IGF-1) is a hormone that regulates systemic growth and metabolism and possesses anti-atherosclerotic properties. We asked whether endothelial-restricted augmentation of IGF-1 signalling is sufficient to suppress atherogenesis.

Methods and results: We generated mice with endothelial-restricted over-expression of human wild-type (WT) IGF-1R (hIGFREO/ApoE-/-) or a signalling-defective K1003R mutant human IGF-1R (mIGFREO/ApoE-/-) and compared them with their respective ApoE-/- littermates. hIGFREO/ApoE-/- had less atherosclerosis, circulating leucocytes, arterial cholesterol uptake, and vascular leakage in multiple organs, whereas mIGFREO/ApoE-/- did not exhibit these phenomena. Over-expressing WT IGF-1R in human umbilical vein endothelial cells (HUVECs) altered the localization of tight junction proteins and reduced paracellular leakage across their monolayers, whilst over-expression of K1003R IGF-1R did not have these effects. Moreover, only over-expression of WT IGF-1R reduced HUVEC internalization of cholesterol-rich low-density lipoprotein particles and increased their association of these particles with clathrin, but not caveolin-1, implicating it in vesicular uptake of lipoproteins. Endothelial over-expression of WT vs. K1003R IGF-1R also reduced expression of YAP/TAZ target genes and nuclear localization of TAZ, which may be relevant to its impact on vascular barrier and atherogenesis.

Conclusion: Endothelial IGF-1 signalling modulates both para- and transcellular vascular barrier function. Beyond reducing atherosclerosis, this could have relevance to many diseases associated with abnormal vascular permeability.

Keywords: Atherosclerosis; Endothelial; Insulin-like growth factor-1 receptor; Vascular permeability.

Graphical Abstract

Figure 1

hIGFREO/ApoE^−/− mice develop less atherosclerotic plaque. (A) Schema of hIGFREO/ApoE^−/− generation. In comparison with ApoE^−/− littermates, hIGFREO/ApoE^−/− develop: (B) reduced atherosclerotic plaque area defined by Oil-Red O staining in thoraco-abdominal aorta (n = 11, 13); scale bar denotes 5 mm; (C) reduced atherosclerotic plaque area defined by Oil-Red O staining in aortic arch (n = 11, 13); scale bar denotes 2000 μm; (D) reduced atherosclerotic plaque area in the aortic sinus (n = 6, 6); scale bar denotes 500 μm; (E) F4/80⁺ macrophage abundance (n = 5, 5); scale bar denotes 250 μm; (F) circulating CD45⁺ leucocytes (n = 15, 15); (G) bone marrow total CD45⁺ leucocytes (n = 8, 13). Data expressed as mean (SEM). *P < 0.05. n denotes the number of mice per group. All statistical comparisons are made with unpaired Student’s t-tests.

Figure 2

hIGFREO/ApoE^−/− exhibit morphological features of VE-Cadherin junction stabilization, reduced vascular paracellular permeability, and reduced endothelial LDL-cholesterol uptake. (A) Representative images of femoral artery VE-Cadherin junctions in hIGFREO/ApoE^−/− and ApoE^−/− controls (scale bar denotes 50 μm). (B) VE-Cadherin junctions are thinner in hIGFREO/ApoE^−/−. (C) VE-Cadherin junction area is lower in hIGFREO/ApoE^−/− (n = 8, 8). (D) Vascular permeability of Evans Blue dye is reduced in multiple vascular beds of hIGFREO/ApoE^−/− (n = 11, 8). (E) Vascularity of bone marrow is similar in hIGFREO/ApoE^−/− and ApoE^−/− controls; in representative images, red denotes VE-Cadherin (VECAD) and blue denotes DAPI (n = 9, 10); scale bar denotes 100 μm. (F) Uptake of BODIPY-labelled LDL-cholesterol is reduced in the aortic endothelium of hIGFREO/ApoE^−/− at 8 weeks of age without western diet feeding (representative images on left show BODIPY in green and DAPI in blue; n = 5, 5); scale bar denotes 25 μm. Data expressed as mean (SEM). *P < 0.05. n denotes the number of mice per group; all statistical comparisons are made with unpaired Student’s t-tests.

Figure 3

IGF-1 receptor over-expression reduces paracellular leakage and LDL-cholesterol uptake in human endothelial cells via its kinase domain. (A) Schema of lentiviral constructs used to achieve doxycycline-inducible WT or kinase-dead (K1003R) IGF-1R over-expression. (B) WT and K1003R IGF-1R protein expression are increased approximately three-fold by doxycycline. (C) Representative images of HUVEC VE-Cadherin with either WT or K1003R IGF-1R over-expression (VE-Cadherin—red, DAPI—blue). D) VE-Cadherin junction thickness is reduced in confluent HUVEC exposed to 100 μM hydrogen peroxide over-expressing WT IGF-1R, but remains unchanged with cells over-expressing K1003R IGF-1R. (E) VE-Cadherin junction area is similar in confluent HUVEC over-expressing WT or K1003R IGF-1R. (F) Paracellular leakage of 40 kDa FITC-Dextran is reduced in HUVEC over-expressing WT IGF-1R, but not K1003R, IGF-1R. G) Uptake of BODIPY-labelled LDL-cholesterol is reduced in HUVEC over-expressing WT IGF-1R, but not K1003R, IGF-1R (representative images on left; BODIPY—green, DAPI—blue). (H) Representative immunoblots of YAP/TAZ and loading controls in nuclear and cytosolic lysates, accompanied by quantification of normalized nuclear TAZ expression and nuclear:cytosolic TAZ ratio, showing reduced nuclear TAZ in HUVEC over-expressing WT vs. K1003R IGF-1R. ns denotes non-specific band in α-tubulin blot. Data expressed as mean (SEM). *P < 0.05; n = 3, 3 in a–g and n = 4, 4 in h. All statistical comparisons are made with unpaired Student’s t-tests.

Figure 4

ApoE^−/− mice with endothelial-specific over-expression of the kinase-dead K1003R mutant IGF-1R (mIGFREO/ApoE) exhibit comparable atherosclerosis, circulating leucocytes and vascular permeability to ApoE^−/− mice. After 12 weeks of western diet feeding, there was no difference between mIGFREO/ApoE^−/− and ApoE^−/− in: (A) lipid deposition in aorta (representative images on left; n = 10, 21); scale bar denotes 5 mm; (B) circulating CD45⁺ leucocytes (n = 10, 8); (C) circulating CD45⁺CD11b⁺ myeloid cells (n = 10, 8); (D) circulating CD45⁺CD11b⁺Ly6G^hiLy6C^hi neutrophils (n = 10, 8); (E) circulating CD45⁺CD11b⁺Ly6G⁻Ly6C⁺ monocytes (n = 10, 8); (F) circulating CD45 ⁺ CD11b ⁺ Ly6G⁻Ly6C^hi ‘inflammatory’ monocytes (n = 10, 8); (G) circulating CD45⁺CD11b⁺Ly6G⁻Ly6C^lo ‘patrolling’ monocytes (n = 10, 8); (H) ratio of ‘inflammatory’ to ‘patrolling’ circulating monocytes (n = 10, 8); or (I) vascular permeability of Evans Blue dye in multiple vascular beds (n = 9, 6). Data expressed as mean (SEM). n denotes the number of mice per group. All statistical comparisons are made with unpaired Student’s t-tests.