Insulin-Like Growth Factor 1 Receptor Deficiency in Macrophages Accelerates Atherosclerosis and Induces an Unstable Plaque Phenotype in Apolipoprotein E Deficient Mice

Abstract

Background

We have previously shown that systemic infusion of insulin-like growth factor-1 (IGF-1) exerts anti-inflammatory and anti-oxidant effects and reduces atherosclerotic burden in apolipoprotein E (Apoe) deficient mice. Monocytes/macrophages express high levels of IGF-1 receptor (IGF1R) and play a pivotal role in atherogenesis but the potential effects of IGF-1 on their function are unknown.

Methods and Results

To determine mechanisms whereby IGF-1 reduces atherosclerosis and to explore the potential involvement of monocytes/macrophages, we created monocyte/ macrophage specific IGF1R knockout (MΦ-IGF1R-KO) mice on Apoe^−/− background. We assessed atherosclerotic burden, plaque features of stability, and monocyte recruitment to atherosclerotic lesions. Phenotypic changes of IGF1R-deficient macrophages were investigated in culture. MΦ-IGF1R-KO significantly increased atherosclerotic lesion formation, as assessed by Oil-red-O staining of en face aortae and aortic root cross-sections, and changed plaque composition to a less stable phenotype, characterized by increased macrophage and decreased α-smooth muscle actin-positive cell population, fibrous cap thinning, and decreased collagen content. Brachiocephalic artery lesions of MΦ-IGF1R-KO mice had histological features implying plaque vulnerability. Macrophages isolated from MΦ-IGF1R-KO mice showed enhanced proinflammatory responses upon stimulation by IFNγ and oxidized LDL and elevated antioxidant gene expression levels. Moreover, IGF1R deficient macrophages had decreased expression of ABCA1 and ABCG1 and reduced lipid efflux.

Conclusions

Our data indicate that macrophage IGF1R signaling suppresses macrophage and foam cell accumulation in lesions and reduces plaque vulnerability, providing a novel mechanism whereby IGF-1 exerts anti-atherogenic effects.

Introduction

Insulin-like growth factor -1 (IGF-1) is a pleiotropic factor, which is produced and acts locally (i.e. via autocrine or paracrine effects) or that circulates in blood and exerts endocrine effects. Circulating IGF-1 levels reach a peak during the pubertal growth phase, eventually declining with the progression of age. While its role in developmental and pubertal growth as the major mediator of growth hormone’s effects is well-documented, the physiological roles of IGF-1 in aged subjects is not understood. Aging is a major independent risk factor for coronary heart disease; in fact, there is increased cardiovascular and coronary heart disease prevalence with age in both genders. In light of the aging-dependent decline of circulating IGF-1 levels, a potential link between IGF-1 levels and the elevated prevalence of cardiovascular diseases has been suggested. Indeed, epidemiologic data have suggested that low IGF-1 levels are an important predictor of coronary events in aged subjects^1–3. In an animal model of atherosclerosis, we have shown that low levels of circulating IGF-1 are associated with more atherosclerosis⁴ and vice versa an increase in circulating IGF-1 decreases atherosclerotic burden⁵.

The pathogenesis of atherosclerosis is complicated, involving multiple cell types including vascular endothelial cells, smooth muscle cells, and pro-inflammatory cells such as macrophages. To determine the potential target(s) whereby IGF-1 reduces atherosclerosis, we previously investigated whether smooth muscle specific overexpression of IGF-1 alters atherosclerosis⁶. Intriguingly, overexpression of IGF-1 in smooth muscle cells did not alter atherosclerotic burden⁶, but increased features of plaque stability^{6, 7}, suggesting that other cellular components were a potential target for anti-atherogenic effects of IGF-1. Macrophages play a pivotal role in the pathogenesis of atherosclerosis by modulating inflammatory status and by scavenging and accumulating excess lipid to become foam cells. Thus regulation of macrophage functions in terms of inflammation and phagocytosis is key to comprehending the disease process. Since macrophages have a predominant role in the inflammatory status of atherosclerotic lesions, it is critical to determine whether IGF-1 regulates macrophage function, particularly inflammatory activation and phagocytic activity. Thus far, there is very limited information on the potential link between IGF-1 effects and macrophage function, particularly in relation to vascular disease. IGF-1 has been reported to enhance chemotactic macrophage migration⁸, stimulate TNFα expression⁸, and enhance LDL uptake and cholesterol esterification⁹. There are also reports from clinical investigations providing indirect evidence of anti-inflammatory effects of IGF-1, e.g., there is an inverse relation between serum IL-6 and IGF-1 levels¹⁰, IGF-1/IGFBP-3 administration to patients with severe burn injury induced an anti-inflammatory effect and reduced IL -6 and TNFα ^{11, 12}, and low IGF-1 and high IL-6 and TNFα levels are associated with higher mortality in elderly patients ^{13, 14}. In this study we used a well-established animal model of atherosclerosis to examine the role of macrophage IGF-1 signaling in atherosclerosis development and progression.

Materials and Methods

Detailed Materials and Methods are in supplemental materials.

Animals

All animal experiments were performed according to protocols approved by the institutional animal care and use committee. Monocyte/macrophage targeted IGF-1 receptor null mice on Apoe^−/− background (MΦ-IGF1R-KO) and control mice (IGF1R-flox) were generated as described in the online supplemental Materials and Methods. Eight-week old mice were fed a high-fat diet for 8 weeks before assessing atherosclerosis.

Atherosclerotic burden and plaque composition

Atherosclerotic burden was quantified using en face preparations of whole aorta stained with Oil Red O and in cross sections of aortic root. Plaque composition was assessed in cross sections of aortic root by immunostaining for Mac-3 (macrophage) and α-smooth muscle actin (smooth muscle cell), and Masson’s Trichrome staining for collagen. Brachiocephalic artery lesions were analyzed by Carstairs’ method as per Gough et al¹⁵.

Tracing recruitment of circulating monocytes

Circulating Ly6C^hi or Ly6C^lo monocytes were labeled in vivo with polychromatic red microspheres as described¹⁶, and numbers of red microsphere-positive cells were counted in a plaque and normalized to the efficiency of labeling in circulating monocytes (% labeled/total number), assessed by flow cytometry.

Intravital Fluorescence Microscopy

Intravital fluorescence microscopy was performed as described elsewhere¹⁷.

Macrophage culture, Western Blot Analysis, Quantitative Real-Time RT-PCR

Thioglycolate elicited macrophages were obtained from MΦ-IGF1R-KO and IGF1R-flox mice¹⁸. Western blot analysis¹⁹, and total RNA extraction and real-time PCR²⁰ were performed as previously described.

Statistical Analysis

All numeric data are expressed as mean ± SEM. Statistical analyses were performed using GraphPad PRISM (version 6.07) software. Data sets were first assessed for residuals distribution using D’Agostino-Pearson omnibus normality test and for equal variances using Levene’s Test for Equality of Variances. Differences in outcomes were determined by ANOVA and Bonferroni’s multiple comparisons test, Kruskal-Wallis test, unpaired Student’s t-test with or without Welch’s correction, or Mann-Whitney U test, accordingly with the normality of residuals distribution. Differences were considered significant at P<0.05. Fisher’s exact test was used to compare frequency of observed indices of plaque vulnerability (Table 1).

Table 1.

MΦ-IGF1R-KO induced features of ruptured plaques in the brachiocephalic artery

	IGF1R-flox	MF-IGF1R-KO	P value
Fibrous cap disruption	0/22	1/17	0.44
Intraplaque hemorrhage	3/22	8/17	0.03
Fibrin deposition	0/22	0/17	ND
Medial elastin breaks	0/22	4/17	0.03

Data indicates positive observation in a total number of samples. Statistical significance was tested by Fisher’s exact test. ND: not determined.

Results

Generation of MΦ-IGF1R-KO/Apoe^−/− mice

Monocyte/macrophage targeted IGF-1 receptor deficient mice were generated on Apoe^−/− background (LyzCre^/+/Igf1r^flox/flox/Apoe^−/−: MΦ-IGF1R-KO) by crossing LyzCre^/+ mice into Igf1r^flox/flox/Apoe^−/− (IGF1R-flox, served as a control). Genotype segregation in the offspring followed the expected Mendelian frequency, and we did not recognize any developmental/morphological abnormalities. IGF-1 receptor deficiency was confirmed by lack of IGF1R protein detection by western blot, and IGF-1-dependent phosphorylation on Akt in peritoneal macrophages (Suppl. Fig 1A). Consistent with myeloid-selective Cre expression^21–23, we found that the exon 3 of Igf1r was also excised in neutrophils (Suppl. Fig 1B): However we were unable to detect IGF1R protein in IGF1R-flox neutrophils, indicating extremely low levels of IGF1R protein expression in neutrophils (Suppl. Fig 1C).

IGF1R and insulin receptor (InsR) can form a hybrid receptor (heterotetramer consisting of α+β subunits of IGF1R and α+β subunits of InsR)^24–26, which binds IGF-1 with high affinity but not insulin^{27, 28}. Depletion of IGF1R in endothelial cells is reported to allow InsR to form a holotetramer, thereby enhancing insulin sensitivity²⁹. Since insulin signaling in macrophages has significant effects on atherosclerosis (there are contradictory reports showing anti-³⁰ or pro-atherogenic³¹ effects), we assessed insulin signaling activity in MΦ-IGF1R-KO macrophages. IGF1R-deficiency did not alter InsR expression levels (Suppl. Fig 1D). In IGF1R-flox macrophages immunoprecipitation of InsR pulled down 100 % of IGF1R (Suppl. Fig 1E, left panel) whereas anti-IGF1R immunoprecipitation pulled down 50% of InsR (Suppl. Fig 1E right panel), suggesting that the ratio of IGF1R/InsR-hybrid receptors to InsR-holoreceptors was 2:1 without presence of IGF1R-holoreceptors. IGF1R-deficency did not alter insulin induced dose-dependent phosphorylation of Akt (Suppl. Fig 1F).

Animals were assessed for circulating leukocyte counts (Suppl. Table 1), cholesterol levels, and cytokine levels (Suppl. Table 2). As has been reported³², western diet feeding for 8 weeks was associated with elevated monocyte count; however IGF1R deficiency did not result in a significant difference in white blood cell count or in circulating monocyte count (Suppl. Table 1). A subpopulation of circulating monocytes as defined by CD11b⁺/ CD90⁻/B220⁻/CD49b⁻/NK1.1⁻/Ly6G⁻/Ly6C^high cells (Ly6C^hi-monocytes) has been reported to be pro-inflammatory and to be increased under atherogenic conditions such as hyperlipidemia³³. We did not observe a significant difference in Ly6C^hi-monocyte levels between MΦ-IGF1R-KO and IGF1R-flox mice (Suppl. Table 1). MΦ-IGF1R-KO did not significantly alter circulating IGF-1, proinflammatory cytokine (IL-6, TNFα, MCP1), or cholesterol levels (Suppl. Table 2).

Atherosclerosis was enhanced by MΦ-IGF1R-KO

Atherosclerotic lesion formation was assessed after 8-week of high-fat diet feeding. En face oil-red-o staining of aortae revealed a significant ~64 % increase in oil-red-o positive lesion area, and there was a consistent ~34% increase in plaque size at the aortic root in MΦ-IGF1R-KO mice (Fig 1). The effect was confirmed in each gender, indicating there was no gender specific effect of IGF1R deficiency (Suppl. Fig 2). Plaque composition with regard to Mac3 (macrophage)-positive cells, α-smooth muscle actin (αSMA)-positive cells, and collagen content (Masson’s Trichrome stain) was significantly altered by MΦ-IGF1R-KO (Fig 2); there was a 49% increase in Mac3 detection (Fig 2A), and a 31% decrease in αSMA detection (Fig 2B). Plaques in MΦ-IGF1R-KO mice had a thinner smooth muscle cap than IGF1R-flox mice (Fig 2B) and decreased collagen content (Fig 2C). Since we observed a decrease of collagen content in plaques from MΦ-IGF1R-KO mice, we hypothesized that this could be due to enhanced collagen degradation by matrix metalloproteinases (MMPs). Cultured peritoneal macrophages from MΦ-IGF1R-KO animals expressed higher levels of MMP-1, -2, -8, -9, -12, -13 and -14 than those from IGF1R-flox animals, suggesting enhanced MMPs’ activity (Fig 2D). MMP’s protein levels in tissue lysates of ascending aortas after 2-month of Western diet feeding (Suppl. Fig 3A and 3B) showed a significant increase in expression of MMP-1, -2, -8, and -9 in MΦ-IGF1R-KO animals. These observations are consistent with enhanced lesion formation in MΦ-IGF1R-KO mice with a phenotypic shift toward increased features of plaque vulnerability.

Figure 1.

Monocyte/macrophage IGF1R deficiency aggravated atherosclerosis. Atherosclerotic lesion formation was assessed in MΦ-IGF1R-KO and IGF1R-flox mice (control) fed on a high-fat diet for 2 months. (A) Oil red o staining of en face aortae; plaque covered area was determined and expressed as % of total area (C; **P<0.0001 by Mann-Whitney U test, n=25 for IGF1R-flox and n=23 for MΦ-IGF1R-KO). (B) Plaque size was determined in H&E stained cross sections of aortic root; *P<0.05 by Mann-Whitney U test, n=18 for IGF1R-flox and n=19 for MΦ-IGF1R-KO (D).

Figure 2.

Monocyte/macrophage IGF1R deficiency alters atherosclerotic plaque composition. Atherosclerotic plaque composition was assessed in cross sections of aortic root. (A) Macrophage content was assessed by immunostaining against Mac3 antigen. *P<0.05 by Student’s t-test, n=18 for IGF1R-flox and n=17 for MΦ-IGF1R-KO. (B) Smooth muscle cell content assessed as α-smooth muscle actin positive cells by immunostaining (representative images shown in upper-left panels). *P<0.05 by Student’s t-test, n=16 for IGF1R-flox and n=18 for MΦ-IGF1R-KO. Thickness of smooth muscle cell positive fibrous cap was also determined by immunostaining against α-smooth muscle actin (representative images shown in lower-left panels). **P<0.001, by Student’s t-test, n=12 for IGF1R-flox and n=14 for MΦ-IGF1R-KO. (C) Collagen content determined by Masson’s Trichrome stain. **P<0.01 by Mann-Whitney U test, n=17 for IGF1R-flox and n=16 for MΦ-IGF1R-KO. (D) Western blot analysis of matrix metalloproteinases (MMPs) in cultured peritoneal macrophages isolated from IGF1R-flox (Flox) and MΦ-IGF1R-KO (Cre+) mice. Representative results of three independent experiments are shown.

Plaque destabilization in MΦ-IGF1R-KO mice

Since our findings suggested plaque destabilization in MΦ-IGF1R-KO mice, we assessed indicators of vulnerability in brachiocephalic artery plaques (Suppl. Fig 4) by staining cross-sections of brachiocephalic artery with the Carstairs’³⁴ and Verhoeff–Van Gieson methods³⁵. Five μm-thick cross-sections were made every 50 μm along the artery, for a total of 10 sections per artery. If 3 consecutive sections indicated a fibrous cap disruption, intraplaque hemorrhage, fibrin deposition, or medial elastin break, the artery was considered positive for signs of plaque vulnerability. After 8 weeks on a high fat diet, MΦ-IGF1R-KO animals had increased features of vulnerable plaques compared with IGF1R-flox animals, as determined by the presence of intraplaque hemorrhage (IGF1R-flox: 3-positive in 22 animals vs. MΦ-IGF1R-KO: 8-positive in 17 animals, P=0.03, Table 1) and medial elastin breaks (IGF1R-flox: no positive in 22 animals vs. MΦ-IGF1R-KO: 4-positive in 17 animals, P=0.03, Table 1), suggesting that plaques in MΦ-IGF1R-KO mice are more unstable than in IGF1R-flox mice.

Proinflammatory responses are enhanced in IGF1R deficient macrophages

Thioglycolate elicited peritoneal macrophages from the MΦ-IGF1R-KO mice were assessed for proinflammatory cytokine and chemokine secretion (Fig 3). Among tested cytokines and chemokines, IL-1β production was below detectable levels, whereas IL-1α, IL-6, TNF α, MCP-1, and Fractalkine were detected (Fig 3 A–F). IGF1R deficient macrophages secreted significantly higher levels of all the tested cytokines and chemokines, except Fractalkine, which showed a strong trend to an increase (P=0.0508). Interferon γ (IFNγ) significantly enhanced TNFα (Fig 3C) and MCP-1 (Fig 3D) production in IGF1R-flox cells, but not in MΦ-IGF1R-KO cells. Intriguingly, IFNγ did not enhance Fractalkine production in IGF1R-flox macrophages whereas in MΦ-IGF1R-KO cells it decreased Fractalkine production (Fig 3E). In a setting of exposure to IFNγ, MΦ-IGF1R-KO cells secreted higher levels of IL-1α, IL-6 and MCP-1 compared to IGF1R-flox. We tested NFkB DNA binding activity in these cells (Fig 3F). There was significantly higher NFkB-DNA binding activity in MΦ-IGF1R-KO, which was further enhanced by IFNγ, consistent with higher cytokine/chemokine production. To further assess NFkB involvement, we exposed cells to NFkB inhibitors (BMS-345541 and Parthenolide, Fig 3G–J). BMS-345541 completely abolished enhanced TNFα and IL-6 production in MΦ-IGF1R-KO cells (Fig 3G and 3H), and Parthenolide recapitulated the BMS-345541 effect (Fig 3I and 3J), supporting the importance of NFkB-dependent cytokine production in MΦ-IGF1R-KO macrophages.

Figure 3.

Proinflammatory cytokine and chemokine production by cultured macrophages isolated from IGF1R-flox and MΦ-IGF1R-KO mice. IL-1α (A), IL-6 (B), TNFα (C), MCP-1 (D), and Fractalkine (E) secretion by peritoneal macrophages were assessed by respective ELISA. Macrophages were isolated by peritoneal lavage from IGF1R-flox (open column) and MΦ-IGF1R-KO (closed column) mice and allowed to adhere on culture plates overnight. After removal of non-adhering cells, cells were primed by interferon-γ (IFN γ), and conditioned media were collected after 24 hours. *P<0.05, **P<0.01 vs IGF1R-flox; ##P<0.01 vs IGF1R-flox+IFNγ; $P<0.01 vs MΦ-IGF1R-KO. Statistical significance was assessed by 2-way ANOVA and subsequent posthoc analysis using Bonferroni’s multiple comparisons test, n=6. (F) NFκB-DNA binding activity was assessed (expressed as relative light units; RLU) in cell lysates of IGF1R-flox and MΦ-IGF1R-KO macrophages with or without priming by IFNγ for 3 hours. *P<0.05 vs IGF1R-flox by 2-way ANOVA, n=3. (G) – (J) effects of NFkB inhibitors on TNFα (G and H) and IL-6 (I and J) secretion by peritoneal macrophages. IGF1R-flox (open column) and MΦ-IGF1R-KO (closed column) macrophages were exposed to the indicated dose of BMS-345541 or Parthenolide for 1 hour prior to activation by IFNγ or lipopolysaccharide (LPS). Conditioned media were collected after 24 hours and assessed for cytokine concentration using respective ELISA kit (R&D Systems). ##P<0.05 vs. IGF1R-flox+IFNγ, #P<0.05 vs IGF1R-flox+LPS, by Mann-Whitney U test, n=4.

Macrophage polarization

Since IGF1R deficient macrophages manifested enhanced pro-inflammatory responses, we examined whether IGF1R deficiency influenced macrophage polarization. Classical activation (M1) was induced by exposure to IFNγ and subsequently to lipopolysaccharide (LPS; Fig 4A–E). M1-marker gene (Tnf, Nos2, Il6, Ccl2, and Ccl5) expression levels were highly induced by exposure to IFNγ-LPS, and IGF1R deficiency further enhanced the expression, implying enhanced M1 activation in MΦ-IGF1R-KO macrophages. On the other hand, IL-4-induced gene expression levels (i.e. M2 activation markers; Arg1, Mrc1, and Pparg) were not influenced by IGF1R deficiency (Fig 4F–H). It has been reported that oxidatively modified LDL (oxLDL) and oxidized lipid components of oxLDL alter macrophage activation status or induce a distinctive activation status^36–39. Hemeoxygenase-1 (Hmox1) and thioredoxin reductase-1 (Txnrd1) are signature genes, which have been shown to be upregulated in macrophages exposed to oxidized phospholipid, leading to a polarization status of Mox³⁹. In fact, oxLDL exposure enhanced Hmox1 and Txnrd1 expression significantly (Fig 5AB). Intriguingly, IGF1R deficiency did not influence Txnrd1 mRNA levels (Fig 5B) but significantly upregulated Hmox1 mRNA levels (Fig 5A). We further assessed if IGF1R deficiency influences oxLDL’s effect on expression levels of macrophage activation marker genes (Fig 5). With regard to M1-activation markers, oxLDL by itself did not alter mRNA levels in IGF1R-flox cells (Tnf, Nos2, Il6, Ccl2 and Ccl5: Fig 5C–G); however in MΦ-IGF1R-KO cells, in which these mRNA levels were significantly elevated compared to IGF1R-flox cells, oxLDL significantly further upregulated Tnf, Nos2, Il6, and Ccl2 mRNA levels (Fig 5C–F), whereas Ccl5 mRNA was suppressed (Fig 5G). On the other hand, of the M2-activation markers, Mrc1 mRNA levels were moderately upregulated by oxLDL (Fig 5I), while Arg1 and Pparg were not affected by oxLDL (Fig 5H and J). Although there were modestly lower Arg1 mRNA levels in IL-4/oxLDL treated MΦ-IGF1R-KO cells than in IGF1R-flox cells, overall oxLDL or IGF1R-deficiency did not robustly alter M2 marker levels. These data suggested that IGF1R deficiency influenced macrophage polarization, namely enhancing the pro-inflammatory M1 phenotype as evoked by IFNγ and by oxLDL stimulation; however, the effect is not entirely classical activation, as implicated by Ccl5 downregulation (Fig 5G). These results prompted us to examine if Mox-marker gene expression was altered by IGF1R deficiency in IFNγ-primed macrophages. IFNγ suppressed Hmox1 and Txnrd1 mRNA levels in IGF1R-flox macrophages (Fig 5AB); however, in MΦ-IGF1R-KO cells IFNγ exerted the opposite effect leading to an upregulation of Hmox1 mRNA (Fig 5A). IFNγ suppressed Txnrd1 mRNA levels in IGF1R deficient macrophages, however to a lesser extent than in IGF1R-flox cells, resulting in higher expression levels compared to IGF1R-flox cells (Fig 5B).

Figure 4.

Messenger RNA expression levels representing classical activation (M1) are elevated in MΦ-IGF1R-KO macrophages. Classical activation marker (Tnf, A; Nos2, B; Il6, C; Ccl2, D; Ccl5, E) and alternative activation marker (Arg1, F; Mrc1, G; Pparg, H) mRNA levels were assessed by quantitative RT-PCR. After isolation and adhesion to culture plates, macrophages remained untreated (NA), or were exposed to IFNγ and LPS to induce classical (i.e. M1) activation or exposed to IL-4 to induce alternative (i.e. M2) activation. IGF1R-flox, open column; MΦ-IGF1R-KO, closed column. Representative data from 4 independent experiments is shown. **P<0.01 or *P<0.05 vs IGF1R-flox by 2-way ANOVA and Bonferroni’s multiple comparisons test, n=3.

Figure 5.

IGF1R-deficiency skews macrophage activation induced by oxLDL. Macrophages isolated from IGF1R-flox (open column) and MΦ-IGF1R-KO (closed column) mice were primed with IFNγ for 6 hours, and exposed to oxLDL for 18 hours. Total RNA was extracted and subjected to quantitative RT-PCR, testing oxidized lipid-induced activation (Mox) marker (Hmox1; A and Txnrd1; B), M1 activation marker (Tnf, C; Nos2, D; Il6, E; Ccl2, F; Ccl5, G), and M2 activation marker (Arg1, H; Mrc1, I; Pparg, J) mRNA levels. Representative data from 3 independent experiments is shown. ##P<0.01 vs. respective condition without oxLDL, $P<0.01 vs. IGF1R-flox of the respective condition, ¶P<0.01 vs. respective condition without IFNγ, **P<0.01 vs. IFNγ-treated MΦ-IGF1R-KO, #P<0.05 vs. IL-4 and oxLDL-treated IGF1R-flox, *P<0.05 vs. respective condition without oxLDL, by 2-way ANOVA and Bonferroni’s multiple comparisons test, n=3.

Efferocytosis is a process by which apoptotic cells are removed by phagocytosis and is considered a significant mechanism involved in the resolution of inflammation. We evaluated expression levels of genes involved in efferocytosis (Suppl. Fig 5). M1-activation suppressed expression of efferocytosis related genes (Anxa1, Gas6, C1qa, and Mfge8), and M2-activation had no effect (Anxa1, C1qa, Mertk; Suppl. Fig 5) or downregulated (Gas6, Mfge8) gene expression levels. IGF1R-deficiency did not alter these mRNA levels in either activated or non-activated cells. OxLDL by itself did not induce obvious effects, except for an upregulation of Mertk mRNA levels (Suppl. Fig 5D). IGF1R-deficiency upregulated Anxa1 and Mfge8 in the presence of oxLDL (Suppl. Fig 5A and E), however it caused a trend to downregulation of C1qa (Suppl. Fig 5C). Thus IGF1R deficiency did not induce changes suggesting enhanced/reduced efferocytosis.

Lipid internalization and efflux

Macrophage internalization and accumulation of modified LDL leads to foam cell formation, which is a hallmark of atheroma formation. Macrophage lipid internalization was assessed by exposing cells to oxLDL and acLDL (Fig 6AB). Macrophages were exposed to IFNγ/LPS (M1-activation) or IL4 (M2-activation) then tested for lipid incorporation after exposure to modified LDL for 48 hr. M1 cells internalized far lesser amounts of lipid compared to the not-activated cells or to M2 cells (Fig 6AB), whereas, M2-activated cells incorporated more lipid than untreated cells when they were exposed to acLDL (Fig 6A). IGF1R deficiency did not alter lipid incorporation (Fig 6AB), irrespective of activation status and exposure to acLDL or oxLDL. Cholesterol efflux from cells to extracellular lipid acceptors also contributes to lipid accumulation and thus foam cell formation. Thus, activated and lipid laden macrophages were tested for cholesterol efflux activity (Fig 6CD). IGF1R deficiency markedly reduced ApoAI-dependent cholesterol efflux in M1 activated cells (Fig 6C). We measured HDL-dependent cholesterol efflux in non-activated, M1-activated, and M2-activated cells (Fig 6D), and found that IGF1R deficiency caused a small but significant reduction in cholesterol efflux to HDL (Fig 6D). To gain insights into potential mechanisms, we assessed expression levels of ABCA1, ABCG1, and SRB1, which are major lipid-transporters responsible for cholesterol efflux. They were differentially regulated by IGF1R deficiency (Fig 7); ABCG1 expression levels were downregulated by ~50 % in IGF1R deficient macrophages, regardless of activation status (Fig 7B and D), whereas ABCA1 and SRB1 were not regulated. To investigate ABCG1 expression in MΦ-IGF1R-KO cells, we exposed the cells to ligands of LXR, which is a major regulator of ABCG1 gene expression⁴⁰. The LXR agonists, GW3965 or T0901317, induced ABCG1 and ABCA1 expression in IGF1R-flox macrophages, but not in MΦ-IGF1R-KO cells (Fig. 7E). Thus LXR-dependent regulation of ABCG1 expression is compromised by IGF1R deficiency, potentially accounting for downregulation of ABCG1 in MΦ-IGF1R-KO cells.

Figure 6.

Cholesterol efflux from macrophages was adversely altered by IGF1R deficiency. Panel A and B: Peritoneal macrophages were incubated with IFNγ+LPS to induce M1 and IL-4 to induce M2 activation status, subsequently exposed to acetylated LDL (A) or oxidized LDL (B) for 48 hours and assessed for lipid uptake by oil-red-o stain. Fluorescent images of stained cells were assessed using ImagePro software. Oil-red-o staining intensity was normalized to cell number and expressed as relative fluorescent unit (RFU)/ cell number. Panel C and D: Peritoneal macrophages were not activated (NA) or activated to M1 and M2, and then loaded with lipid by exposure to acLDL and ³H-cholesterol for 24 hours. Subsequently, ApoAI-dependent (C) or HDL-dependent (D) efflux of ³H-cholesterol for the subsequent 24 hours were assessed. IGF1R-flox, open column; MΦ-IGF1R-KO, closed column. Statistical significance was assessed by Mann-Whitney U test, n=8.

Figure 7.

IGF1R-deficiency downregulated ABCG1 expression levels in acLDL-loaded macrophages. Peritoneal macrophages were not activated (NA) or activated to M1 and M2, and then exposed to acLDL for 24 hours. ABCA1 (A), ABCG1 (B), and SRB1 (C) expression levels were assessed by Western blot analysis (representative measurements and blots (D) from 3 independent experiments are shown). IGF1R-flox, open column; MΦ-IGF1R-KO, closed column. *P<0.05 vs IGF1R-flox cells by 2-way ANOVA. (E) Western blot analysis for ABCG1 expression, assessing effects of LXR agonists. AcLDL-loaded macrophages were exposed to LXR agonists (1 μM GW3965, and 1 μM T0901317) for 48 hours and assessed for ABCG1 expression levels. Representative results from 3 independent experiments are shown. *P<0.05 vs IGF1R-flox, **P<0.01 vs IGF1R-flox, Student’s t-test, n=3.

Monocyte recruitment, macrophage proliferation, and apoptosis in MΦ-IGF1R-KO mice

Since MΦ-IGF1R-KO caused an increase in Mac3-positivity within plaques, we evaluated monocyte recruitment to lesions, macrophage proliferation⁴¹ and macrophage apoptosis in plaques, three major determinants of macrophage number in lesions⁴². To evaluate monocyte recruitment, we labeled circulating monocytes in vivo by i.v. administration of fluorescent microspheres⁴³. The microspheres are not capable of penetrating into tissue interstitial space, thus red fluorescence-positive cells (identified by DAPI positivity) within plaque are considered recruited and infiltrated cells. We evaluated red microsphere labeling by flow cytometry: as described¹⁶, i.v. injections of red microspheres labeled the Ly6C^lo population among circulating monocytes (Suppl. Fig 6A), whereas clodronate administration 1-day prior to the injection of red microspheres introduced the label to Ly6C^hi monocytes (Suppl. Fig 6B). As shown in Fig 8, we detected a significantly larger number of red-fluorescence positive cells in plaques from MΦ-IGF1R-KO mice both without (Fig 8C) or with clodronate administration (Fig 8D). It is noteworthy that a higher number of red microsphere positive cells are detected following clodronate administration, indicating that Ly6C^hi monocytes are the dominant subpopulation to be recruited to lesions³³. In order to assess leukocyte adhesion and rolling on the luminal side of the endothelium in vivo, we performed intravital microscopy to detect CD11b-positive cells in the mesenteric circulation. MΦ-IGF1R-KO significantly increased CD11b+ leukocyte adhesion (Fig 8E) and rolling (Fig 8F) on the luminal surface of the endothelium, consistent with increased recruitment of CD11b+ cells, i.e. monocytes and neutrophils. Since IGF1R expression was undetectable in neutrophils (Suppl. Fig 1), it is unlikely that IGF1R gene deletion in neutrophils contributed significantly to these findings.

Figure 8.

Monocyte/Macrophage IGF1R deficiency increased monocyte recruitment and infiltration into atherosclerotic plaques. MΦ-IGF1R-KO and IGF1R-flox mice were fed on a high-fat diet for 12 weeks. Red microspheres were administered via i.v. injection at 3 and 7 days before sacrifice (C), or clodronate liposomes were administered i.v. 1 day prior to the red microsphere administration in order to introduce label to the Ly6C^hi-monocyte population (D). Serial frozen sections were obtained at aortic root and stained with Oil Red O (A) or with DAPI (B). Images (B) are magnified to view the corresponding serial sections of (A) at specified areas (rectangle). L – lumen; blue: DAPI-stained nuclei. Arrows indicate red microsphere-positive cells within the lesion. The number of red microsphere-positive cells was counted per plaque (C; without clodronate, D; with clodronate), and results are shown by box-and-whiskers plot, in which a box extends from 25^th to 75^th percentiles with the median indicated by a horizontal line and whiskers indicating the minimum and maximum. *P<0.05, **P<0.01 by Mann-Whitney U test, n=6 for IGF1R-flox and n=5 for MΦ-IGF1R-KO. (E and F) intravital microscopy assessment of CD11b+ cell adhesion (E) and rolling (F) on the luminal surface of the mesenteric circulation endothelium, and the results are shown by box-and-whiskers plot as described above. *P<0.05 and **P<0.01 by Mann-Whitney U test, n=4 for MΦ-IGF1R-KO and n=6 for IGF1R-flox.

To assess proliferation activity of macrophages in plaques, we assessed Ki67 and proliferating cell nuclear antigen (PCNA) gene expression levels^44–46 (Suppl. Fig 6C). Macrophage Marker (a monoclonal antibody raised against isolated macrophages of mouse origin, Santa Cruz, SC-101447)^{47, 48} -positive plaque area was laser dissected from aortic root for RNA isolation, followed by quantitative RT-PCR. Equal amplification of CD68 was confirmed between IGF1R-flox and MΦ-IGF1R-KO lesions. Macrophage rich region from MΦ-IGF1R-KO plaques expressed the same levels of Ki67 or PCNA compared to IGF1R-flox (Suppl. Fig 6C), suggesting no difference in macrophage proliferation in plaques. TUNEL-/Mac3- double positive cell-numbers were not different (Suppl. Fig 6D). Taken together, our data suggests that IGF1R-deficiency enhances monocyte recruitment to lesions, thereby increasing the macrophage cell population.

Oxidized LDL downregulates IGF1R in macrophages

Oxidized LDL plays a critical role in atherogenesis and we have previously shown that oxLDL downregulates IGF1R in vascular smooth muscle cells⁴⁹. To determine potential regulation of macrophage IGF1R expression by oxLDL we used the human monocytic cell line, THP-1, which was differentiated into macrophages and exposed to 100 μg/mL oxLDL (Suppl. Fig 7). OxLDL exposure for 24 hours downregulated IGF1R levels by 80%.

Discussion

IGF-1 production has been documented in macrophages^{8, 9, 50, 51}, however precise effects of IGF-1 in macrophages in relation to the pathogenesis of atherosclerosis have not been elucidated. In this study, we generated macrophage/monocyte specific IGF-1 receptor deficient mice and discovered pivotal roles of IGF-1 in the regulation of inflammatory responses and lipid handling in macrophages, which are relevant to anti-atherogenic effects of IGF-1. Our data indicate that IGF1R deficiency in macrophages enhances pro-inflammatory activation (i.e. M1 polarization), thereby promoting pro-inflammatory cytokine production (Fig 3 and Fig 4), and enhances lipid accumulation due to reduced efflux (Fig 6 and Fig 7). These phenotypic changes resulted in increased recruitment of macrophages to atherosclerotic plaques and increased atherosclerotic burden in MΦ-IGF1R-KO mice (Fig 1 and Fig 2). Moreover, histological evaluation of brachiocephalic arteries indicated that the MΦ-IGF1R-KO induced features of unstable plaques (Table 1).

There is growing interest in the role of IGF-1 in cardiovascular disease. Low circulating IGF-1 levels have been associated with cardiovascular disease risk factors^52–56, and in particular there is growing evidence for a role for IGF-1 deficiency in the pathogenesis of metabolic syndrome⁵⁷. Acromegaly (i.e. excessive growth hormone and IGF-1) or otherwise, growth hormone and IGF-1 deficiency have been linked to cardiovascular complications^{58, 59}. However, epidemiological studies linking IGF-1 with cardiovascular disease report mixed results. Some cross-sectional and prospective studies^60–65 suggest a positive association between IGF-1 (and in some cases IGFBP-3) and atherosclerosis, but others have found that low IGF-1 is a predictor of ischemic heart disease and mortality, consistent with the potential anti-apoptotic, antioxidant, and plaque stabilization effects of IGF-1^{1–3, 13, 66–72}. Methodological constraints could explain these contradictions, because measurement of total IGF-1 levels represents only a crude estimate of the biologically active IGF-1. Thus, an IGF-1-specific kinase receptor activation assay may better reflect IGF-1 bioactivity⁷³ and in fact it has been reported that higher IGF-1 bioactivity is associated with significantly longer survival in subjects with a high inflammatory risk profile or history of cardiovascular disease⁷⁴. Additionally, polymorphisms in the IGF-1 gene promoter region that influence circulating IGF-1 levels have been reported^75–77: The alleles that indicate lower circulating IGF-1 levels have been associated with increased risk for type 2 diabetes, myocardial infarction, left ventricular hypertrophy, higher carotid intima-media thickness, higher aortic pulse wave velocity, and lower endothelium-dependent vasodilation^75–77. Intriguingly, IGF-1-resistance in the endothelium was reported in an animal model of obesity, thereby blunting the vasodilatory response to IGF-1 via attenuated eNOS-phosphorylation and nitric oxide production⁷⁸. Our present findings that reduced macrophage IGF-1 signaling is highly pro-inflammatory and increases atherosclerotic burden is consistent with the growing evidence that decreased IGF-1 action may be a significant contributor to the pathogenesis of atherosclerosis. Of note, we found that oxidized LDL downregulates IGF1R levels in human derived THP1 macrophage (Suppl. Fig 7), which is consistent with our previous reports that both IGF-1 and IGF1R expression were significantly lower in human atherosclerotic plaque intimal regions with macrophage infiltration, where oxidized LDL was highly detected^{79, 80}.

We used MΦ-IGF1R-KO mice, which have one allele of Lys2 ablated, whereas the IGF1R-flox mice (control) have both alleles intact. Previous studies in LysCre^/+ mice with regard to potential alterations in monocyte/macrophage biology due to the hemizygous deficiency of Lys2 showed no evidence of a heterozygous phenotype²¹. In addition, a complete ablation of both alleles of Lys2 does not influence atherosclerosis in Apoe−/− mice⁸¹. Lys2-cre mediated gene excision occurs in monocytes and in neutrophils^21–23. We confirmed deletion of exon 3 of the Igf1r gene in both macrophages and neutrophils isolated from MΦ-IGF1R-KO mouse, however, we were unable to detect IGF1R protein even in Igf1r-normal neutrophils (i.e. IGF1R-flox neutrophils) indicating extremely low expression levels. Thus, although contribution of neutrophils to the phenotype of MΦ-IGF1R-KO mice cannot be excluded, it appears much more likely that macrophages, which express significant levels of IGF1R, play the predominant role. IGF1R and insulin receptor (InsR) are structurally similar and form a heteromeric hybrid receptor, consisted of α+β subunits of IGF1R and α+β subunits of InsR. The hybrid receptor binds IGF-1 with high affinity but does not bind insulin at physiological ranges. To our knowledge, this is the first study evaluating the expression ratio between IGF1R and InsR in macrophages, showing predominant expression of InsR. Thus, IGF1R-flox macrophages (expressing both IGF1R and InsR) responded to physiological dose of insulin (Suppl. Fig 1). Intriguingly, IGF1R deficiency in macrophages, although it should free-up InsR hemi-dimers to form holoreceptors, did not increase insulin signaling. This finding is relevant because macrophage InsR deficiency has been shown to modulate atherosclerosis development (although results have been contradictory^{30, 31}). Since we found that MΦ-IGF1R-KO did not alter insulin signaling in macrophages, it is unlikely that changes in insulin action on macrophages plays a significant role in the phenotype of MΦ-IGF1R-KO mice.

Macrophages become activated as they infiltrate into a target tissue and are exposed to stimuli, expressing a highly pro-inflammatory phenotype or a less inflammatory but phagocytic and antigen-presenting phenotype. The former status was termed “classical activation” or M1 activation whereas the latter was referred to as “alternative activation” or M2 activation. Recent investigations, however, indicate that macrophage activation status likely encompasses a broad spectrum where M1 vs. M2 activation are likely two extremes^82–84. Recognizing that macrophage activation represents a continuum, we tested M1 and M2 activation, and also Mox activation, which was described as a unique activation status found in plaque macrophages³⁹, in order to provide insights into the effects of IGF1R deficiency. Our results indicate that IGF1R deficiency resulted in macrophages acquiring a highly inflammatory (i.e. M1) phenotype (Fig 3), whereas M2 marker gene expression levels were not altered (Fig 4F–H). However, MΦ-IGF1R-KO macrophages are not simply skewed to a more inflammatory phenotype; these cells were also shifted to a phenotype induced by oxLDL, characterized by upregulation of the anti-oxidant genes, Hmox1 and Txnrd1, described as Mox activation³⁹ (Fig 5AB). Reports of the effects of oxLDL and its specific lipid moieties on macrophage activation are variable and include enhancement of the inflammatory phenotype^{36, 38, 85}, induction of M2 activation⁸⁶ or a unique activation status³⁹. In summary, our results suggest that IGF1R deficiency skews macrophages to a unique activation state, which can be characterized by enhanced pro-inflammatory response and elevated anti-oxidant system.

The effect of IGF1R deficiency on macrophage activation status prompted us to examine how IGF1R deficiency influences uptake of modified LDL, cholesterol efflux as well as efferocytosis. It has been shown that M1 activated macrophages demonstrate reduced foam cell formation in response to oxLDL^{36, 87} and that Mox activated macrophages have attenuated phagocytosis and efferocytosis³⁹. It would thus have been reasonable to speculate that IGF1R deficiency, which enhances the proinflammatory phenotype as well as Mox marker expression levels, could lead to reduced foam cell formation. Our results, nevertheless, showed no evidence of altered acLDL or oxLDL uptake by IGF1R-deficient macrophages (Fig 6AB). We also assessed expression levels of genes which are functional in efferocytosis, however there was no apparent alteration caused by IGF1R deficiency (Suppl. Fig 5). Thus, despite enhanced inflammatory responses and redox gene expression, IGF1R deficient macrophages do not appear to have impairment in modified lipid uptake or efferocytosis. However, further assessment of lipid handling by IGF1R-deficient macrophages indicated that cholesterol efflux is impaired by IGF1R deficiency (Fig 6CD), potentially promoting foam cell formation. Cholesterol efflux is mediated by the lipid transporters, ABCA1, ABCG1, and SRB1⁸⁸. IGF1R deficiency caused lower expression levels of ABCG1 (Fig 7), which should, at least in part, account for lowered cholesterol efflux. Intriguingly, the LXR-dependent regulation of ABCG1 expression was compromised in MΦ-IGF1R-KO macrophages. LXR is a major regulator of ABCG1 expression^{89, 90}, in fact, LXR drives ABCG1 expression upon lipid-loading by modified LDL (which causes accumulation of oxysterols) ^{89, 90}. Thus impaired LXR regulation of ABCG1 expression could be an important mechanism underlying lowered cholesterol efflux in MΦ-IGF1R-KO macrophages. MΦ-IGF1R-KO increased atherosclerosis burden as assessed by oil red o staining of en face aortae and by histological analysis of aortic root cross-sections (Fig 1). These mice had an increase in lesional macrophages and in recruitment of monocytes to atherosclerosis plaques (Fig 8). Consistent with the increased recruitment of monocytes, chemokine expression levels were upregulated in MΦ-IGF1R-KO macrophages (i.e. MCP-1 and Fractalkine, Fig 3–5). Moreover, MΦ-IGF1R-KO increased features of plaque vulnerability, as evidenced by histological features of intraplaque hemorrhage and medial elastin breaks in lesions from brachiocephalic artery (Suppl. Fig 4 and Table 1). These findings were in accordance with the changes in plaque composition, notably, the increased population of macrophages, the reduced population of smooth muscle cells particularly within the plaque cap, and decreased collagen (Fig 2). To obtain insights into underlying mechanisms, we assessed MMP levels in peritoneal macrophages from MΦ-IGF1R-KO mice and showed significant upregulation of MMPs (Fig 2D) which have previously shown to be relevant to atherogenesis and plaque vulnerability^{91, 92}. In addition, we found significant upregulation of MMP-1, -2, -8, and -9 in lysates of ascending aorta from MΦ-IGF1R-KO animals, consistent with data obtained using peritoneal macrophages. A major source of collagen matrix synthesis and deposition in atherosclerotic plaques is smooth muscle cells, and thus one can speculate that the enhanced inflammatory milieu induced by MΦ-IGF1R-KO altered smooth muscle cell homeostasis. In fact, it is noteworthy that MΦ-IGF1R-KO robustly enhanced macrophage production of pro-inflammatory cytokines, such as IL-1α, TNFα, and IL-6 (Fig 3), and it has been reported that cytokines such as TNFα reduce IGF-1 and increase IGF binding protein-3 in vascular smooth muscle cells, leading to a reduction in bioactive IGF-1⁹³. IGF-1 positively regulates collagen synthesis by smooth muscle cells⁹⁴. Indeed, we have recently shown that increased IGF-1 signaling in vascular smooth muscle cells increases features of plaque stability, as determined by increased fibrous cap area, α-smooth muscle actin-positive smooth muscle cells and collagen content without affecting plaque burden⁶, potentially mediated by IGF-1 induction of smooth muscle differentiation⁶ and collagen synthesis⁹⁴. Thus, it seems reasonable to speculate that increased production of inflammatory cytokines in MΦ-IGF1R-KO mice disrupts normal IGF-1 signaling in smooth muscle cells and reduces collagen deposition in plaques of MΦ-IGF1R-KO mice. Taken together with our previous results, our current study provides insights into potential interactions between macrophages and smooth muscle cells, i.e. the enhanced inflammatory milieu by macrophages suppresses IGF-1 signaling in smooth muscle cells leading to a reduction of plaque stability.

Our results are consistent with the growing body of experimental evidence that IGF1 has anti-atherogenic effects. Infusion of IGF-1 in Apoe^−/− mice reduces atherosclerotic burden⁵, and overexpression of IGF-1 in vascular smooth muscle cells increases plaque collagen content, smooth muscle cell levels and reduces necrotic core size⁶. IGF-1 has been shown to have anti-oxidant effects on the endothelium, via upregulation of glutathione peroxidase levels⁹⁵. Our current findings demonstrate that IGF-1 signaling has a major effect on macrophage biology which is critical for atherogenesis. However, a limitation of experimental studies to date has been that they have been performed in murine models and studies in larger animals phylogenetically closer to humans are lacking, and will be important for development of IGF-1 based therapeutic strategies.

In summary, we have shown that IGF1R-deficiency in macrophages of Apoe^−/− mice increases atherosclerotic burden and changes plaque composition to one of lowered smooth muscle cell and collagen content. Our data suggest that the loss of IGF-1 signaling skews macrophage activation to a pro-inflammatory status and promotes lipid accumulation in macrophages by lowering lipid efflux. There is increasing evidence linking low IGF-1 to multiple cardiovascular risk factors, including metabolic syndrome and aging^{57, 96}. In fact, it has been reported that decreased IGF-1 bioavailability is an adverse prognostic factor for coronary heart disease (reviewed in^{97, 98}). Our findings herein suggesting that IGF-1 regulates macrophage inflammatory responses and lipid metabolism may provide the basis for a novel therapeutic approach for the treatment of atherosclerotic vascular disease development and progression.

Clinical Perspectives.

Atherosclerosis is an inflammatory disease and acute coronary events result largely from erosion or rupture of unstable plaques with increased inflammatory cells and a relative reduction in vascular smooth muscle cells. Macrophages play a major role in atherogenesis by scavenging and accumulating lipids to become lipid-laden foam cells. Further, pro-inflammatory macrophages induce smooth muscle cell death by secreting cytokines, and degrade extracellular matrix by producing enzymes such as matrix metalloproteinases (MMPs), weakening the tensile strength of plaques and predisposing them to rupture. In the past decade, Insulin-Like Growth Factor-1 (IGF-1) has demonstrated anti-atherogenic effects in experimental models, however mechanisms are poorly elucidated. In this study, we investigated macrophage-IGF-1 receptor deficiency in a murine model of atherosclerosis. We found that IGF-1 receptor deficiency increased monocyte/macrophage recruitment to lesions, skewed macrophage activation to a pro-inflammatory status, promoted lipid accumulation in macrophages by lowering lipid efflux, and upregulated MMP production, resulting in an increase in atherosclerotic burden and a decrease in features of plaque stability. Our findings are consistent with epidemiological studies suggesting low circulating IGF-1 is a predictor of ischemic heart disease and mortality. There is also evidence linking low IGF-1 to multiple cardiovascular risk factors, including metabolic syndrome. Our findings suggest that IGF-1 regulation of macrophage inflammatory responses and lipid metabolism may be the basis for novel approaches to reduce atherosclerotic lesion progression and promote plaque stability.