Inflammatory Role of Milk Fat Globule–Epidermal Growth Factor VIII in Age‐Associated Arterial Remodeling

Abstract

Background

Age‐associated aortic remodeling includes a marked increase in intimal medial thickness (IMT), associated with signs of inflammation. Although aortic wall milk fat globule–epidermal growth factor VIII (MFG‐E8) increases with age, and is associated with aortic inflammation, it is not known whether MFG‐E8 is required for the age‐associated increase in aortic IMT. Here, we tested whether MFG‐E8 is required for the age‐associated increase in aortic IMT.

Methods and Results

To determine the role of MFG‐E8 in the age‐associated increase of IMT, we compared aortic remodeling in adult (20‐week) and aged (96‐week) MFG‐E8 (−/−) knockout and age matched wild‐type (WT) littermate mice. The average aortic IMT increased with age in the WT from 50±10 to 70±20 μm (P<0.0001) but did not significantly increase with age in MFG‐E8 knockout mice. Because angiotensin II signaling is implicated as a driver of age‐associated increase in IMT, we infused 30‐week‐old MFG‐E8 knockout and age‐matched littermate WT mice with angiotensin II or saline via osmotic mini‐pumps to determine whether MFG‐E8 is required for angiotensin II–induced aortic remodeling. (1) In WT mice, angiotensin II infusion substantially increased IMT, elastic lamina degradation, collagen deposition, and the proliferation of vascular smooth muscle cells; in contrast, these effects were significantly reduced in MFG‐E8 KO mice; (2) On a molecular level, angiotensin II treatment significantly increased the activation and expression of matrix metalloproteinase type 2, transforming growth factor beta 1, and its downstream signaling molecule phosphorylated mother against decapentaplegic homolog 2, and collagen type I production in WT mice; however, in the MFG‐E8 knockout mice, these molecular effects were significantly reduced; and (3) in WT mice, angiotensin II increased levels of aortic inflammatory markers phosphorylated nuclear factor‐kappa beta p65, monocyte chemoattractant protein 1, tumor necrosis factor alpha, intercellular adhesion molecule 1, and vascular cell adhesion molecule 1 molecular expression, while in contrast, these inflammatory markers did not change in knockout mice.

Conclusions

Thus, MFG‐E8 is required for both age‐associated proinflammatory aortic remodeling and also for the angiotensin II–dependent induction in younger mice of an aortic inflammatory phenotype observed in advanced age. Targeting MFG‐E8 would be a novel molecular approach to curb adverse arterial remodeling.

Nonstandard Abbreviations and Acronyms

AT1

angiotensin II receptor type 1

AT2

angiotensin II receptor type 2

ICAM1

intercellular adhesion molecule 1

IMT

intimal medial thickness

MCP1

monocyte chemoattractant protein 1

MFG‐E8

milk fat globule–epidermal growth factor VIII

MMP2

matrix metalloproteinase type 2

NF‐κB p65

nuclear factor‐kappa beta p65

p‐NF‐κB p65

phosphorylated nuclear factor‐kappa beta p65

SMAD2

mother against decapentaplegic homolog 2

TGF‐β1

transforming growth factor beta 1

TNF‐α

tumor necrosis factor alpha

VCAM1

vascular cell adhesion molecule 1

VSMC

vascular smooth muscle cell

WT

wild‐type

Clinical Perspective.

What Is New?

• Milk fat globule–epidermal growth factor 8 increases with age and is required for age‐associated proinflammatory arterial remodeling that results in aortic intimal medial thickening.

• In younger mice, milk fat globule–epidermal growth factor 8 is also essential for angiotensin II–induced adverse arterial remodeling that mimics that which occurs in advanced age via the increase of proinflammation, intimal medial thickening, elastin fragmentation, collagen deposition, and vascular smooth muscle cell proliferation.

What Are the Clinical Implications?

• Since milk fat globule–epidermal growth factor 8 is required for age‐associated proinflammation in arterial wall remodeling, targeting milk fat globule–epidermal growth factor 8 is a potential molecular approach to curb inflammatory arterial remodeling, thus maintaining the health of the arterial system during aging and in arterial diseases such as hypertension.

Milk fat globule‐epidermal growth factor 8 (MFG‐E8) is a secreted extracellular glycoprotein that was initially discovered as a bridging molecule between apoptotic cells and macrophages for the clearance of cellular debris, also known as efferocytosis. ¹ , ² Recent studies have also shown that MFG‐E8 exerts an inflammatory role in vascular remodeling. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ In vivo studies have demonstrated that arterial MFG‐E8 protein levels increase during aging, arterial injury, and also in arterial walls of hypertensive, diabetic, and atherosclerotic animal models. ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ¹⁰ , ¹¹ These in vivo and in vitro studies suggest that MFG‐E8 signaling plays an important role in the initiation and progression of adverse arterial remodeling such as intimal medial thickening.

Angiotensin II, a major metabolic component of the renin‐angiotensin‐aldosterone signaling system, produces multiple cardiovascular inflammatory effects. ¹² , ¹³ Physiologically, angiotensin II signaling is critical in regulating blood pressure. In pathological conditions, angiotensin II stress signaling induces inflammation, contributing to the development of arterial aging and hypertensive vasculopathy. ¹² , ¹⁴ Our previous study showed that younger adult animals infused with angiotensin II exhibited an older arterial phenotype, including increased intimal vascular smooth muscle cell (VSMC) infiltration, intimal medial thickening (IMT), matrix metalloproteinase type 2 (MMP2) activation, transforming growth factor beta 1 (TGF‐β1) activation, and collagen deposition, eventually leading to elevated systolic blood pressure. ¹⁵ In vitro studies have demonstrated that MFG‐E8 is a pivotal relay element within the angiotensin II/monocyte chemoattractant protein 1 (MCP1) signaling cascade that mediates VSMC inflammation, invasion, and proliferation. ³ , ⁴

Although both MFG‐E8 and angiotensin II increase in the inflamed aging arterial wall or in cultured aging VSMCs, ³ , ⁴ whether the underlying proinflammatory role of MFG‐E8 is required for age‐associated remodeling and aortic remodeling caused by angiotensin II remains unknown. We hypothesized that a chronic infusion of angiotensin II to younger mice, like younger rats, ¹⁵ would generate an “older” arterial inflammatory phenotype, which is impacted by MFG‐E8 signaling. To this end, younger (30‐week‐old) wild‐type (WT) and MFG‐E8 knockout (−/−) (KO) mice infused with angiotensin II or saline were used and compared with untreated control mice with a wide range of ages (from 4 to 96 weeks) to investigate if MFG‐E8 is necessary for angiotensin II–associated arterial inflammatory remodeling at the molecular, cellular, and tissue levels.

METHODS

All data, analytic methods, and study materials are available to other researchers for purposes of reproducing the results or replicating a procedure. The material that supports the findings of this study is available from the corresponding author upon reasonable request.

Experimental Animals

All experiments were conducted according to the protocols (445‐LCS‐2022) approved by the National Institute on Aging in accordance with the National Institutes of Health Animal Care and Use Committee. MFG‐E8 (−/−) knockout mice that were generated, characterized, and genotyped as described previously, were obtained from Dr Mark Udey at the National Cancer Institute. ¹⁶ , ¹⁷ Transgenic Rip1‐Tag2 mice were obtained from the National Cancer Institute. MFG‐E8 knockout mice were generated by replacing exons 2 to 6 of the gene encoding MFG‐E8 in 129SvJ embryonic cells with a neomycin‐resistant cassette. Mice were genotyped by polymerase chain reaction using the following primers: Rip1‐Tag2, GGACAAACCACAACTAGAATGCAGTG (forward) and CAGAGCAGAATTGTGGAGTGG (reverse); Neo, GCCAGAGGCCACTTGTGTAG; and MFG‐E8, CTCTCAGATTCACCTGCTCGTG and CACCGTTCAGGCACAGGCTG. Thirty‐week‐old male MFGE8 knockout mice (homozygous for lack of MFG‐E8) and age‐matched WT mice were used in this angiotensin II infusion study. In addition, 4‐, 8‐, 20, 50‐, and 96‐week‐old male WT and age‐matched knockout archival aortic sections or frozen tissue were used for the study of age‐related arterial remodeling.

Angiotensin II Infusion

Osmotic mini‐pumps (Alzet Model 2004) were implanted dorsally and subcutaneously in anesthetized (2% isoflurane) mice (n=14–18/group) to deliver angiotensin II (500 ng/kg per min) or 0.9% saline (placebo) for 28 days.

Sample Collection and Preparation

Blood Collection

Ocular blood was collected at 14 days after mini‐pump implantation. At the end of the experiment, mice were euthanized under sodium pentobarbital anesthesia, and blood was collected from the right atrium into a 4‐mL EDTA‐coated Eppendorf tube immediately. The blood samples were centrifuged for 15 minutes at 3000 rpm (1500g) at 4 °C. The supernatant was carefully transferred into a 0.5‐mL Eppendorf tube and stored at −80 °C until use.

Tissue Collection

Tissue samples were isolated from the ascending aorta to the bifurcation of the common iliac artery of mice. For morphological analysis, animals were perfused with 4% paraformaldehyde‐ PBS at physiological pressure for 5 minutes; and aortic tissues were paraffin‐embedded and sectioned for histological analysis. For western blotting, animals were perfused with 0.9% saline and fresh aortic tissues were snap‐frozen in liquid nitrogen and stored at −80 °C until use.

Histology, Immunostaining, and Morphometric Analyses

To quantitate aortic remodeling, IMT was measured using hematoxylin and eosin staining; elastin breaks were counted via Elastin Verhoeff's–Van Gieson staining; and collagen deposition was evaluated with Masson's trichrome staining. Hematoxylin and eosin, Elastin Verhoeff's–Van Gieson, and Masson trichrome staining were performed using MasterTech stain kits (StatLab, McKinney, TX). Staining of aortic walls was performed as described in previous studies. ¹⁸ In brief, aortic paraffin sections (5 μm in thickness) were used for immunostaining with antiproliferating cellular nuclear antigen, tissue necrosis factor‐alpha (TNF‐α), intercellular adhesion molecule 1 (ICAM1), vascular cell adhesion molecule 1 (VCAM1), and collagen type I antibodies. Details of primary antibodies used are listed in Table 1. The ratios of target immunohistochemical staining positive area to the total tissue or cell area were determined via a computer‐imaging program according to the instruction provided by manufacture (MetaMorph Imaging System; Molecular Devices, San Jose, CA).

Table 1.

Primary Antibodies

Antibody	Company	Catalog no.	Titer for WB	Titer for IHC
Angiotensin II	Sigma	A9525		1:350
AT1	Santa Cruz	sc‐515 884	1:500	1:100
AT2	Abcam	Ab92445	1;1000	1:100
Collagen I	Santa Cruz	sc‐59 772		1:50
GAPDH	Cell Signaling technology	5014	1:1000
ICAM1	Santa Cruz Biotechnology	Sc‐8439		1:50
MCP1	R&D Systems	AF479	1:200
MFG‐E8	R&D Systems	AF2805	1:1000
MMP2	R&D Systems	AF1488	1:500
MT1‐MMP	Thermo Fisher Scientific	PA5‐16514	1:500
NF‐κB p65	Santa Cruz Biotechnology	sc‐8008	1:500
p‐NF‐κB p65	Santa Cruz	Sc‐33 020	1:500	1:50
PCNA	Santa Cruz	sc‐9857		1:50
p‐Smad2	Santa Cruz Biotechnology	3108	1:500
Smad2	Santa Cruz Biotechnology	5339	1:1000
TGF‐β1	Santa Cruz Biotechnology	sc‐146	1:200
TNF‐α	Abcam	ab9739	1:1000	1:100
VCAM1	Santa Cruz Biotechnology	Sc‐13 160		1:50

AT1 indicates angiotensin II receptor type 1; AT2, angiotensin II receptor type 2; ICAM1, intercellular adhesion molecule 1; IHC immunohistochemistry; MCP1, monocyte chemoattractant protein 1; MMP2, matrix metalloproteinase type 2; MT1MMP, member type I of MMP activator; NF‐κB p65, nuclear factor‐kappa beta p65; p‐NF‐κB p65, phosphorylated nuclear factor‐kappa beta p65; PCNA, proliferating cell nuclear antigen; SMAD2, mother against decapentaplegic homolog 2; p‐SMAD2 phosphorylated mother against decapentaplegic homolog 2; TGF‐β1, transforming growth factor beta 1; TNF‐α tumor necrosis factor alpha; VCAM1, vascular cell adhesion molecule 1; and WB, western blotting.

Western Blotting

Western blot analysis was performed as described previously. ¹⁸ In brief, total protein was quantified using the Pierce Coomassie (Bradford) protein assay kit (Hercules, CA), following the manufacturer's instructions. Ten micrograms of total protein was run on 4% to 12% NuPAGE gels (Thermo Fisher Scientific, Waltham, MA), then transferred to a polyvinylidene fluoride membrane and immunoblotted with the antibodies listed in Table 1. Western blotting bands were quantified using National Institutes of Health image J (http://rsb.info.nih.gov/nih‐image/) and the intensity values were normalized to loading control GAPDH or β‐actin.

PAGE Zymography

MMP‐2 activity was determined via PAGE gelatin zymography (Thermo Fisher Scientific) as described previously. ¹⁹

Angiotensin II Quantification

RayBio Mouse Angiotensin II ELISA kit, a commercially available ELISA (RayBiotech Life, Peachtree Corners, GA) was used to measure the amount of angiotensin II in mouse plasma.

Statistical Analysis

All data were presented as mean±SEM. Statistical analyses used 1‐way, 2‐way, or 2‐way repeated‐measures ANOVA followed by Bonferroni post hoc tests used for multiple comparisons. These statistical analyses were performed using Prism version 8.4.1 (GraphPad Software, San Diego, CA). For systolic blood pressure, the repeated‐measures data were also analyzed using a linear mixed‐effects model. ²⁰ A value of P≤0.05 was considered statistically significant.

RESULTS

Age‐Associated Increase in IMT Does Not Occur in MFG‐E8 Knock‐Out Mice

To determine the role of MFG‐E8 in the age‐associated increase of IMT, we compared aortic remodeling in adult (20‐week) and aged (96‐week) MFG‐E8 knockout and age‐matched WT littermate mice. We confirmed that aortic MFG‐E8 protein was markedly elevated in untreated 50‐week‐old versus 8‐week‐old WT mice but was not detected in knockout mice (Figure 1A). The average aortic IMT increased with age in the WT from 50±10 to 70±20 (P<0.0001) but did not significantly increase with age in MFG‐E8 knockout mice. In 96‐week‐old untreated WT, aortic IMT was markedly increased versus 20‐week‐old, but no aging increase was observed in the KO mouse (Figure 1B).

Figure 1. Age‐associated characteristics of the MFG‐E8 knockout mouse.

A, Representative western blots of aortic MFG‐E8 (left panel). Quantitative data of western blots show aortic MFG‐E8 protein abundance (P<0.01 for main age effect, P<0.0001 for main genotype effect, P<0.01 for age×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM combined with individual data points for knockout and WT mice. ***P<0.001 by Bonferroni post‐hoc tests following 2‐way ANOVA. B, Morphometric analysis of the aortic IMT (P<0.0001 for main age effect, P<0.01 for main genotype effect, P<0.0001 for interaction by 2‐way ANOVA). Graph showing mean±SEM combined with individual data points for knockout and WT mice. **P<0.01; and ****P<0.0001 by Bonferroni post‐hoc tests following 2‐way ANOVA. IMT indicates intimal medial thickness; KO, knockout; MFG‐E8, milk fat globule–epidermal growth factor VIII; MW, molecular weight; and WT, wild‐type.

In addition, aortic angiotensin II protein was significantly increased in untreated aging WT mice, but not in the MFG‐E8 KO (Figure S1A). Immunostaining demonstrated that the angiotensin II receptor type 1 (AT1) was markedly increased in the aortic walls of older untreated WT and reduced in the older MFG‐E8 KO (Figure S1B); and the angiotensin II receptor type 2 (AT2) was decreased in both aging WT and KO aortic walls (Figure S1C). We confirmed the immunostaining AT1 and AT2 aging effect with Western blot analysis and observed no AT1 increase or AT2 decrease in aging MFG‐E8 KO mice (Figure S1D). Notably, aortic MFG‐E8 protein levels were significantly decreased in the AT1 knockout mouse, but these levels were not altered in the AT2 knockout mouse (Figure S2) suggesting that MFG‐E8 modulates age‐associated adverse arterial remodeling and angiotensin II/AT1 signaling.

MFG‐E8 Is Required for the Angiotensin II–Induced Aortic Remodeling That Accompanies Advanced Age

Because angiotensin II signaling is implicated as a driver of age‐associated increase in IMT, we infused 30‐week‐old MFG‐E8 knockout mice and age‐matched littermate WT mice with angiotensin II or saline via osmotic mini‐pumps to determine whether MFG‐E8 is required for angiotensin II induced aortic remodeling. Compared with saline, angiotensin II infusion markedly increased the levels of circulating angiotensin II in both WT and KO mice accessed on days 14 and 28, to an even greater extent in MFG‐E8 KO than in WT (Figure 2A). Angiotensin II infusion increased the AT1 receptor expression and reduced the AT2 receptor in both genotypes (Figure 2B and 2C). Angiotensin II dramatically increased the expression of aortic MFG‐E8 in WT mice, but aortic MFG‐E8 protein was not detectable in the knockout mice infused with either angiotensin II or saline (Figure 2D).

Figure 2. MFG‐E8 is required for the angiotensin II–induced signaling.

A, Plasma angiotensin II concentrations (P<0.001 for overall treatment effect by repeated 2‐way ANOVA with the factors of treatment and genotype). Graph showing mean±SEM with individual animal data from knockout and WT mice with angiotensin II or saline infusion over time. **P<0.01 and ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. B, Representative western blots of AT1 receptors (left panel). Western blotting analysis of the AT1 abundance (P<0.0001 for main angiotensin II treatment effect, P<0.0001 for main genotype effect, P<0.0001 for treatment×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM combined with individual data points for knockout and WT mice. ***P<0.001 and ****P<0.0001 by Bonferroni post hoc tests following 2‐way ANOVA. C, Representative western blots of AT2 receptor (left panel). Western blotting analysis of the AT2 abundance (P<0.001 for main angiotensin II infusion effect, P<0.05 for main genotype effect by 2‐way ANOVA). Graph (right panel) showing mean±SEM combined with individual data points for knockout and WT mice. **P<0.01 by Bonferroni post hoc tests following 2‐way ANOVA. D, Representative western blots of aortic MFG‐E8 (left panel). Quantitative data of Western blots of aortic MFG‐E8 protein abundance as normalized by GAPDH (P<0.001 for main angiotensin II infusion effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. Ang II indicates angiotensin II; AT1, angiotensin II receptor type 1; AT2, angiotensin II receptor type 2; KO, knockout; MFG‐E8, milk fat globule–epidermal growth factor VIII; MW, molecular weight; and WT, wild‐type.

Immunohistochemistry of mouse aortic walls was performed after 4 weeks of angiotensin II or saline infusion. The morphometrical analysis indicated that adverse arterial remodeling was dependent upon the presence of MFG‐E8. In WT mice, a marked increase in IMT was observed in the aortic walls of mice infused with angiotensin II compared with saline; however, in knockout mice, no significant difference in IMT was observed (Figure 3A). Elastin Verhoeff's–Van Gieson staining indicated a significant increase in the number of the elastin fiber breaks in WT mice treated with angiotensin II versus saline, while this effect was substantially reduced in knockout mice (Figure 3B). In WT mice infused with angiotensin II, Masson's trichrome staining showed that the fraction of intimal medial extracellular matrix, mainly collagen, was increased versus saline, while this effect was not observed in knockout mice (Figure 3C). VSMC proliferation is a key cellular event of arterial thickening in hypertension induced by angiotensin II. ²¹ In WT mice infused with angiotensin II, the percentage of proliferating cell nuclear antigen–positive VSMCs, an index of cellular division, was markedly elevated when compared to saline; however, no difference in proliferating cell nuclear antigen was seen in knockout animals (Figure 3D). Thus, MFG‐E8 plays a necessary role in angiotensin II‐induced IMT, elastic lamina degradation, collagen deposition, and proliferation of VSMCs; and in general, MFG‐E8 deficiency alleviates these effects in aortic walls.

Figure 3. MFG‐E8 is required for the angiotensin II–induced aortic remodeling.

A, Photomicrographs of hematoxylin and eosin staining of aortic walls. Morphometric analysis of the aortic IMT (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA) (far right panel) B, Photomicrographs of EVG staining of aortic walls. Quantification of elastin breaks (P<0.05 for main treatment effect, P<0.001 for main genotype effect, P<0.01 for treatment×genotype, by 2‐way ANOVA) (far right panel). C, Photomicrograph of Masson's trichrome staining of aortic walls. Morphometric analysis of intimal medial ECM (blue color) (P>0.05 for main treatment effect, P<0.01 for main genotype effect, P<0.01 for treatment×genotype, by 2‐way ANOVA) (far right panel). D, Photomicrographs of PCNA immunostaining of aortic walls. Quantification of relative PCNA stained nuclei area (P>0.05 for main treatment effect, P<0.01 for main genotype effect, P<0.01 for treatment×genotype, by 2‐way ANOVA) (far right panel). Graph showing mean±SEM with individual animal data points for knockout and WT mice with angiotensin II or saline infusion. *P<0.05, **P<0.01 and ***P<0.001 by Bonferroni post‐hoc tests following 2‐way ANOVA. Scale bar=100 μm. Ang II indicates angiotensin II; ECM, extracellular matrix; EVG, Elastin Verhoeff's–Van Gieson; HE, hematoxylin and eosin; IMT, intimal medial thickness; KO, knockout; L, lumen; M, media; PCNA, proliferating cell nuclear antigen; and WT, wild‐type.

Angiotensin II Activation of Aortic MMP2 Is Dependent on MFG‐E8

Prior studies indicate that increased levels of MMP2 and MMP2 activation occur in response to angiotensin II signaling. PAGE zymography demonstrated that the levels of activated MMP2 within the aortic walls of younger mice after angiotensin II infusion were dependent upon the presence of MFG‐E8 (Figure 4A): in WT mice, angiotensin II infusion markedly increased the levels of activated aortic MMP2 compared with saline; in KO mice, activated MMP2 in angiotensin II infused animals was significantly higher than saline, but was still significantly less than that of the angiotensin II–treated WT mice (Figure 4A). Western blot analysis showed that the abundance of aortic MMP2 protein induced by angiotensin II was dependent upon the presence of MFG‐E8 (Figure 4B): in WT mice, angiotensin II markedly increased the levels of MMP2 protein in aortic walls when compared with saline (Figure 4B middle panel); in knockout mice, the MMP2 protein levels were significantly increased in angiotensin II– versus saline‐treated mice but was significantly less than that of angiotensin II–treated WT mice (Figure 4B middle panel). Similarly, angiotensin II increased the levels of member type I of MMP activator, an activator of MMP2, and was dependent on the presence of MFG‐E8 (Figure 4B, right panel); in WT mice, angiotensin II significantly increased the levels of member type I of MMP activator protein in the aortic wall when compared with saline; however, in knockout animals, there was no significant increase in MT1MMP protein levels (Figure 4B, right panel). MFG‐E8 in necessary for the increased activation and expression of MMP2 induced by angiotensin II in the aortic walls of younger animals.

Figure 4. Angiotensin II activation of aortic MMP2 is dependent on MFG‐E8.

A, Representative zymograms of aortic gelatinases (left panel). Quantitative graph shows activated MMP2 protein abundance (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. *P<0.05 and ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. B, Representative western blots of aortic MMP2, MT1MMP, and GAPDH (left panel). Quantitative data of aortic MMP2 protein abundance (P<0.001 for main treatment effect, P<0.01 for main genotype effect, P<0.01 for treatment×genotype, by 2‐way ANOVA) (middle panel). Quantitative data of aortic MT1MMP2 protein abundance (P<0.001 for main treatment effect, P<0.01 for main genotype effect, P<0.01 for treatment×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. *P<0.05, **P<0.01, and ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. Ang II indicates angiotensin II; IMT, intimal medial thickness; KO, knockout; MFG‐E8, milk fat globule–epidermal growth factor VIII; MMP2, matrix metalloproteinase type 2; MT1MMP, type I of MMP activator; MW, molecular weight; and WT, wild‐type.

Angiotensin II Increases Aortic TGF‐β1 Fibrotic Signaling Through MFG‐E8

Angiotensin II signaling and increased MMP2 activation, all promote the activation of TGF‐β1 and its fibrotic effect within arterial walls in vivo and in cultured VSMCs. ¹⁵ , ²² , ²³ Western blotting analysis demonstrated that the angiotensin II‐induction of activated arterial TGF‐β1 was dependent upon the presence of MFG‐E8 (Figure 5A and 5B): in WT mice, angiotensin II infusion markedly increased the levels of activated TGF‐β1 protein in the aortic walls; while in knockout mice there was a smaller increase when compared to saline (Figure 5A and 5B). Notably, in angiotensin II–infused mice, the levels of TGF‐β1 protein in the WT mice was significantly higher than those found in knockout mice (Figure 5A and 5B). Similarly, western blot analysis demonstrated that angiotensin II infusion increased SMAD2 phosphorylation, a TGF‐β1 downstream signaling molecule, and was also dependent upon the presence of MFG‐E8 (Figure 5A and 5B). In WT mice, angiotensin II infusion markedly increased the levels of activated phosphorylated SMAD2 protein in the aortic walls compared with saline; in contrast, in knockout mice, there was no significant increase of these proteins with angiotensin II infusion (Figure 5A and 5B). Expectedly, immunohistostaining and western blotting analyses indicated that the angiotensin II induction of the arterial collagen I levels, a TGF‐β1/ phosphorylated SMAD2 downstream signaling product, were dependent upon MFG‐E8 (Figure 5C and 5D). In WT mice, angiotensin II infusion markedly increased the levels of collagen type I protein in the aortic walls when compared with saline; in contrast, in knockout mice, there was no significant increase of collagen type I when treated with angiotensin II (Figure 5C and 5D). Thus, MFG‐E8 plays a necessary role in the angiotensin II–induced activation of TGF‐β1, its downstream signaling molecule, phosphorylated SMAD2, and collagen I in the aortic wall of mice.

Figure 5. Angiotensin II increases aortic TGF‐β1 fibrotic signaling through MFG‐E8.

A, Representative western blots of aortic TGF‐β1, p‐SMAD2, and SMAD2. B, Quantitative data of aortic activated TGF‐β1 protein abundance (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA) (left panel). Quantitative data of aortic p‐SMAD2/SMAD2 ratio (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.01 for treatment×genotype, by two‐way ANOVA). Graph (right panel) showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. *P<0.05 and ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. C, Photomicrographs of immunostaining collagen type I. Morphometric analysis shows relative collagen I immunostaining area (%) (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA). Graph (far right panel) showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. D, Representative western blots of aortic collagen I. Quantitative data of aortic collagen I (P<0.01 for main treatment effect, P>0.05 for main genotype effect, P<0.0001 for treatment×genotype, by 2‐way ANOVA). Graph (right panel) showing mean±SEM with individual data points for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. Scale bar=100 μm. Ang II indicates angiotensin II; KO, knockout; L, lumen; M, media; MW, molecular weight; p‐SMAD2, phosphorylated mother against decapentaplegic homolog 2; SMAD2, mother against decapentaplegic homolog 2; TGF‐β1, transforming growth factor beta 1; and WT, wild‐type.

Angiotensin II Activation of Nuclear Factor‐Kappa Beta Signaling and Its Downstream Proinflammatory Effects Require MFG‐E8

Nuclear factor‐kappa beta (NF‐κB) is a prominent proinflammatory transcription factor in inflamed arterial walls. ²⁴ , ²⁵ NF‐κB p65, an element of NF‐κB activation, facilitates the upregulation of its downstream inflammatory factors, MCP1, tissue necrosis factor‐alpha 1 (TNF‐α), ICAM1, and VCAM1, which are directly involved in the process of arterial inflammation. ²⁵ , ²⁶ , ²⁷ Aging markedly increased aortic phosphorylated nuclear factor‐kappa beta p65 (p‐NF‐κB p65) in the aging WT mice but did not increase in the knockout mice (Figure S3A and S3B).

Western blot analyses showed that the abundance of arterial p‐NF‐κB p65 in mice infused with angiotensin II was dependent on the presence of MFG‐E8: p‐NF‐κB p65 protein abundance was markedly upregulated in WT mice infused with angiotensin II versus saline, whereas the protein levels remained at similar levels in both groups of knockout animals (Figure 6A and 6B). Western blotting analysis indicated that the abundance of aortic MCP1 protein induced by angiotensin II in mice was also dependent upon the presence of MFG‐E8: WT mice infused with angiotensin II showed a significant increase in the levels of activated MCP1 protein in the aortic wall when compared with saline; conversely, knockout mice infused with angiotensin II exhibited similar levels of MCP1 protein expression when compared to saline (Figure 6A and 6C). Notably, when mice were treated with angiotensin II, MCP1 protein levels were significantly lower in knockout versus WT mice (Figure 6A and 6C). Western blotting and immunostaining analyses also revealed that angiotensin II infusion altered TNF‐α protein levels in mice and was dependent upon MFG‐E8: The level of TNF‐α protein in WT mice was significantly increased in the aortic wall when infused with angiotensin II and compared with saline; however, the protein levels in knockout mice were not significantly altered in either treatment group (Figure 6A, 6D and 6E). In addition, immunohistostaining and morphometric analysis demonstrated that the increased levels of arterial adhesive molecules ICAM1 and VCAM1 protein in mice treated with angiotensin II were also dependent upon MFG‐E8 (Figure 6G, 6H, 6I and 6J). In WT mice, angiotensin II infusion markedly increased the protein levels of both ICAM1 and VCAM1 in the aortic wall when compared with saline; however, in knockout mice, angiotensin II infusion did not significantly alter either ICAM1 or VCAM1 protein abundance versus saline (Figure 6G, 6H, 6I and 6J). MFG‐E8 is required for angiotensin II–induced increases in p‐NF‐κB p65, both MCP‐1, TNF‐α, ICAM1, and VCAM1 protein expression in aortic walls.

Figure 6. Angiotensin II activation of NF‐κB signaling and its downstream proinflammatory effects require MFG‐E8.

A, Representative western blots of aortic p‐NF‐κB p65, NF‐κB p65, MCP‐1, TNF‐α, and GAPDH. B, Quantitative data of p‐NF‐κB p65 protein abundance (P<0.001 for main treatment effect, P<0.01 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA). Graph showing mean±SEM and individual data points for knockout and WT mice with angiotensin II or saline infusion. **P<0.01 and ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. C, Quantitative data of MCP1 protein abundance (P<0.01 for main treatment effect, P<0.05 for main genotype effect, P<0.001 for treatment×genotype, by two‐way ANOVA). Graph showing mean±SEM with individual data points for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post‐hoc tests following two‐way ANOVA. D, Quantitative data of TNF‐α protein abundance (P<0.001 for main treatment effect, P>0.05 for main genotype effect, P<0.05 for treatment×genotype, by two‐way ANOVA). Bar graph showing mean±SEM with individual data points for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. E, Photomicrographs of immunostaining of TNF‐α. F, Quantitative data of aortic TNF‐α immunostaining area (%) (P<0.01 for main treatment effect, P<0.05 for main genotype effect, P<0.01 for treatment×genotype, by two‐way ANOVA). Graph showing mean±SEM with individual data for KO and WT mice with angiotensin II or saline infusion. *P<0.05, **P<0.01, and ***P<0.001 by Bonferroni post hoc tests following two‐way ANOVA. G, Photomicrographs of immunostaining ICAM1. H, Morphometric analysis of relative ICAM1 immunostaining area (%) (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by 2‐way ANOVA). Graph showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post hoc tests following 2‐way ANOVA. I, Photomicrographs of immunostaining VCAM1. J. Morphometric analysis of relative VCAM1 immunostaining area (%) (P<0.001 for main treatment effect, P<0.001 for main genotype effect, P<0.001 for treatment×genotype, by two‐way ANOVA). Graph showing mean±SEM with individual data for knockout and WT mice with angiotensin II or saline infusion. ***P<0.001 by Bonferroni post‐hoc tests following 2‐way ANOVA. Scale bar=100 μm. Ang II indicates angiotensin II; ICAM1, intercellular adhesion molecule 1; KO, knockout; L, lumen; M, media; MCP1, monocyte chemoattractant protein 1; MFG‐E8, milk fat globule–epidermal growth factor VIII; MW, molecular weight; NF‐κB p65, nuclear factor‐kappa beta p65; p‐NF‐κB p65, phosphorylated nuclear factor‐kappa beta p65; VCAM1, vascular cell adhesion molecule 1; TNF‐α, tumor necrosis factor alpha; and WT, wild‐type.

DISCUSSION

Role of MFG‐E8 in Age‐Associated Aortic Remodeling

The primary objective of this study was to determine the requirement of MFG‐E8 in age‐associated aortic inflammation and structural remodeling. The general strategy we employed was to age MFG‐E8 knockout mice, and their WT littermates, to determine whether aortic inflammation and arterial wall thickening in old KO differed from that in old WT. Our first major finding was that MFG‐E8 was indeed required for adverse aortic remodeling that accompanies advanced age, because the hallmarks or the extent of this remodeling, that is, inflammation and intimal medial thickening that were present in WT mice at 96 weeks of age were not observed in age‐matched MFG‐E8 KO mice, in which MFG‐E8 was not expressed. However, it is noteworthy that IMT increased between 20 to 30 weeks of age in the KO mice but not in the WT mice (Figure 1B versus Figure 3A), and the elucidation of this unexpected phenotypic characteristic of the MFG‐E8 knockout mice that occurred between 20 and 30 weeks merits further study. Nevertheless, in advanced age (96 weeks) the IMT of the MFG‐E8 knockout mice was substantially reduced compared with age‐matched WT mice (Figure 1B).

Role of MFG‐E8 in Angiotensin II–Induced Aortic Remodeling

It is well documented that aortic angiotensin II increases with advancing age, and that angiotensin II is the quintessential perpetrator of age‐associated adverse aortic remodeling. The marked increase in arterial wall MMP2 activation in WT mice in response to angiotensin II infusion is attributable, in part at least, to an angiotensin II–induced increase of member type I of MMP activator, an activator of MMP2. We previously observed that treating VSMCs with recombinant human MFG‐E8 activates MMP2, ²⁸ , ²⁹ suggesting that MFG‐E8 is necessary for angiotensin II associated MMP2 activation. Notably, active MMP2 has a high capacity to degrade aortic elastin laminae, releasing activated TGF‐ß1. ¹⁵ , ²² , ²³ , ³⁰

Angiotensin II infusion induces the activation of aortic MMP2 and TGF‐ß1, which promotes arterial elastic fiber degeneration and fibrosis and is dependent on MFG‐E8. In addition, MMP2, a potent activator of latent TGF binding protein, facilitates the conversion from the latent to the active form of TGF‐ß1. ²² We had previously demonstrated that treating VSMCs with recombinant human MFG‐E8 increases the fibrogenic TGF‐ß1/SMAD2/collagen type I signaling cascade in VSMCs. ²⁸ Thus, this also suggested that MFG‐E8 may be vital for angiotensin II–induced MMP2 activation/TGF‐β1 fibrogenic signaling during adverse arterial remodeling. ²² , ³¹

To determine whether MFG‐E8 is required for angiotensin II–induced aortic remodeling, we implanted osmotic mini‐pumps to infuse either angiotensin II or saline into both MFG‐E8 knockout and WT mice at 30 weeks of age. The second major finding of our study was that although circulating levels of angiotensin II increased to even a greater extent in MFG‐E8 knockout versus WT mice, aortic IMT in WT mice increased in response to elevations of angiotensin II but did not increase in the MFG‐E8 KO. Further, except for a small increase in TGF‐β, there were no significant changes in molecular and cellular mechanisms that underlie adverse age‐associated aortic remodeling in response to angiotensin II infusion, including inflammation, MMP2 activation, fibrosis, elastin fragmentation, and VSMCs proliferation ³ , ⁴ , ⁵ , ⁶ , ⁸ , ⁹ , ¹³ , ¹⁹ , ²³ , ³² , ³³ in response to angiotensin II infusion in MFG‐E8 knockout mice compared with those in WTmice . Thus, angiotensin II–induced aortic remodeling requires MFG‐E8 signaling.

Our results also indicate that increased NF‐κB induced inflammation in response to angiotensin II is dependent upon MFG‐E8. Angiotensin II treatment markedly increased p‐NF‐κB p65, a core element of inflammatory arterial wall remodeling, ³⁴ in MFG‐E8 knockout mice but not in WT mice, indicating that MFG‐E8 is an essential element in the signaling by which angiotensin II markedly increases NF‐κB. Several studies indicate that NF‐κB activation facilitates the production of MCP1, TNF‐α, ICAM1, and VCAM1 during the processes of arterial inflammation and remodeling ¹⁰ , ²⁵ , ²⁶ , ²⁷ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ and that MFG‐E8 increases MCP1 activation in VSMCs. ⁴ In response to angiotensin II infusion, MCP1, TNF‐α, ICAM1, and VCAM1 are increased in WT mice but not in MFG‐E8−/− KO mice, further indicating that MFG‐E8 is a crucial player required for angiotensin II–induced arterial inflammation.

In addition, to the aforementioned proinflammatory factors, the degradation fragment of MFG‐E8, medin, per se, induces proinflammatory endothelial activation. ⁴⁰ , ⁴¹ , ⁴² It is well known that medin is the most common amylogenic protein found in aged arterial walls. ⁷ , ¹³ When angiotensin II induces aortic remodeling, excessive medin fragments may contribute to fibrosis and elastin fragmentation due to a marked increase in MFG‐E8.

In summary, our current findings demonstrate that MFG‐E8 plays a necessary inflammatory role in age‐associated vascular remodeling at the molecular, cellular and tissue levels by promoting increased inflammation, VSMC proliferation, collagen deposition, and elastic fiber fragmentation. Thus, targeting MFG‐E8 is a potential molecular approach for the prevention or treatment of adverse arterial remodeling during aging.

Acknowledgments

L.N. and M.W. conceived and designed experiments and wrote the manuscript; L.N., L.L., W.Z., R.T., J.Z., and P.G.G. performed experiments; L.N., L.L., C.H.M., E.G.L., and M.W. analyzed the data; L.N., L.L., C.L., J.L.L, and E.G.L. interpreted results of experiments; L.N. and M.W. wrote the manuscript; R.E.M. and K.R.M. edited the manuscript; and M.W. approved the final version of manuscript.

Sources of Funding

This research was supported by the Intramural Research Program of the National Institute on Aging, National Institutes of Health, the National Natural Science Foundation of China (No. 81470585) and the Fundamental Research Funds for the Central Universities (No. 3332019028).

Disclosures

None.

Supporting information

Data S1–S3

Figures S1–S5

References ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵²