Insulin receptors in vascular smooth muscle cells regulate plaque stability of atherosclerosis

Abstract

Aims

Increased prevalence of acute myocardial infarction related to diabetes and insulin resistance is associated with an elevated risk of unstable atherosclerotic plaques, which are characterized by reduced vascular smooth muscle cells (VSMCs) and extracellular matrix (ECM) and increased inflammation. Thus, insulin resistance may reduce plaque stability, as deleting insulin receptors (IRs) in VSMCs decreases their proliferation and enhances apoptosis.

Methods and results

Direct effects of insulin on VSMCs to alter plaque composition were studied using mice with double knockout of ApoE and IR genes in VSMCs with SMIRKO/ApoE^−/−, Myh11-CreER^T2EYFP⁺/ApoE^−/−, and Myh11-CreER^T2EYFP⁺IRKO/ApoE^−/− mice, which were also used for lineage tracing studies. Compared with ApoE^−/− mice, SMIRKO/ApoE^−/− mice exhibited more atherosclerotic plaques, which contained less VSMCs and collagen but increased levels of VSMC apoptosis and necrotic areas. Lineage tracing studies showed that Icam1⁺ Vcam1⁺ VSMC was inflammatory, which increased in the aortas of Myh11-CreER^T2EYFP⁺IRKO/ApoE^−/− mice compared with control mice. Isolated VSMCs lacking IRs expressed higher inflammatory cytokines than cells with IRs. Cell-based studies indicated that insulin’s anti-apoptotic and pro-proliferative effects in VSMCs were mediated via activation of the IR/Akt pathway, which were decreased in VSMCs from SMIRKO or high-fat diet mice. An analysis of the IR targets that regulated inflammatory cytokines in VSMCs showed that thrombospondin 1 (Thbs1) and Mmp2 were consistently increased with a loss of IRs. Insulin inhibited Thbs1 expression, but not Mmp2 expression, through p-Akt/p-FoxO1 pathways in VSMCs from ApoE^−/− mice, and was impaired in cells from SMIRKO/ApoE^−/− mice. Thbs1 further induced Icam1 and Mmp2 expressions in VSMCs.

Conclusion

Insulin via IRs has significant actions in VSMCs to decrease inflammation, apoptosis, and ECM turnover via the activation of Akt and FoxO1 pathways. The inhibition of insulin actions and related pathways related to insulin resistance and diabetes may contribute to the formation of unstable atherosclerotic plaques.

Graphical Abstract

Open in new tabDownload slide

Time of primary review: 18 days

1. Introduction

The prognosis of cardiovascular diseases (CVDs) has improved for people with diabetes and insulin resistance, but their CVD-related mortality rate is still two- to seven-fold higher than that of age-matched controls.¹ An important pathology of CVD in diabetes and insulin resistance is the acceleration of atherosclerosis with an increased risk of development of unstable plaque.² Plaque rupture and the subsequent thrombotic events are major risk factors for myocardial infarction and stroke, which occur more frequently with unstable plaques, characterized by thin caps and reduced vascular smooth muscle cells (VSMCs) and extracellular matrix (ECM).³ Mechanisms for the increased propensity for unstable plaques in insulin resistance and diabetes have not yet been elucidated.

Since systemic insulin resistance, with or without diabetes, is strongly associated with CVD,^4–6 insulin actions have been studied in endothelial cells and VSMCs.^7–18 Insulin resistance or deficiency in those with diabetes is associated with a selective loss of insulin signalling of PI3 kinase and Akt pathways and an enhancement of mitogen-activated protein kinase (MAPK)-mediated actions in the vascular tissue.¹¹ Targeted loss of insulin receptors (IRs) and their actions in endothelial cells induce endothelial dysfunction and increase the severity of atherosclerosis,¹⁵ whereas enhanced insulin action in endothelial cells activates Akt pathways and lowers atherosclerosis severity.¹⁴ However, these models of atherosclerosis with diabetes exhibited elevations of VSMCs^19,20 and ECM, suggesting an excessive formation of stable atherosclerotic plaque pathologies.

Changes in VSMC metabolism and composition in insulin resistance and diabetes are important since VSMCs constitute 30–50% of the cellular components of atherosclerotic plaques.^21,22 Lineage tracing studies have suggested that VSMCs could affect atherosclerotic plaque stability, possibly by acquiring inflammatory or macrophage-like properties.²²

Multiple insulin actions on VSMCs have been reported that may accelerate atherosclerosis, such as increased proliferation,^9,12,23 mainly via MAPK pathway activation.⁸ Insulin also exhibits vaso-constrictive effects via its actions on the expressions of endothelin 1, endothelin receptor A, and αl-adrenergic receptors.^17,18 However, the direct consequence of insulin’s multiple actions on VSMCs and its role in atherosclerosis and plaque composition have not been demonstrated. We previously reported that a homogenous IR on VSMCs enhanced its proliferation and exacerbated intimal hyperplasia after wire injury, mainly via Akt pathway activation.¹² Mice with specific IR deletion in VSMCs (SMIRKO mice), but not those with insulin-like growth factor IGF1 receptor deletion in VSMCs (Myh11IGF1RKO mice), showed decreased proliferation and reduced intimal hyperplasia after injury, even with high-fat diet (HFD) and insulin resistance.¹² However, the roles of IRs in VSMCs in atherosclerosis, which have a pathology and pathogenesis that is very different from restenosis, were not studied. The results of the present study provide unexpected evidence that IRs and their actions in VSMCs may have anti-inflammatory and mitogenic actions to regulate both the severity and the stability of atherosclerotic plaques in diabetes and insulin resistance. This study focused on understanding the potential mechanism by which insulin and its receptors mediated their actions to increase the stability and reduce the severity of the atherosclerotic plaque.

2. Experimental procedures

2.1 Animals

IR flox/flox mice were kindly provided by Dr C. Ronald Kahn (Joslin Diabetes Center, Boston, MA, USA). SM22α Cre and ApoE^−/− mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). VSMC lineage tracing mice (Myh11-CreER^T2EYFP⁺/ApoE^−/− mice) were kindly provided by Dr Gary Owens (University of Virginia, Charlottesville, VA, USA). All mice were of C56BL/6J background. IR flox/flox mice were crossed with SM22α Cre mice and ApoE^−/− mice to generate smooth muscle–specific IR-knockout mice (SMIRKO/ApoE^−/− mice). IR flox/flox mice were crossed with Myh11-CreER^T2EYFP⁺/ApoE^−/− mice to generate smooth muscle–specific lineage tracing and IR-knockout mice (Myh11-CreER^T2EYFP⁺IRKO/ApoE^−/− mice). The sample size was calculated based on our previous studies on atherosclerosis. The mice were randomly assigned to groups. All protocols for animal use and euthanasia were reviewed and approved by the Animal Care Committee of the Joslin Diabetes Center, and were in accordance with the National Institutes of Health (NIH) guidelines following the standards established by Animal Welfare Acts and by the documents titled ‘Principles for Use of Animals’ and ‘Guide for the Care and Use of Laboratory Animals’.

2.2 Anaesthesia

The mice were anaesthetized with 100 mg/kg ketamine and 10 mg/kg body weight xylazine by intraperitoneal injection. The mice were euthanized with 200 mg/kg body weight sodium pentobarbital by intraperitoneal injection. The level of anaesthesia was monitored by using the pedal reflex (firm toe pinch).

Detailed methods of genotyping of mice and information on anaesthesia, blood pressure measurement, immunohistochemistry, haematoxylin and eosin (H&E) staining, Sudan IV staining, VSMC culture, flow cytometry, glucose and insulin tolerance tests, lipid measurement, western blotting, TUNEL staining, EdU incorporation, cell death enzyme-linked immunosorbent assay (ELISA), adenovirus and lentivirus transfection, bone marrow (BM) transplantation, streptozotocin-induced diabetes, and qPCR are provided in Supplementary material online, Methods.

2.3 Statistical analysis

No data were excluded from the analysis. The data were analysed for normality by using GraphPad Prism software. The Shapiro–Wilk normality test was used to test the normal distribution of data. For analysing the normal distribution of data, comparisons between two groups were made by using an unpaired Student’s t-test and between more than two groups with two variables by using two-way analysis of variance (ANOVA) with Tukey’s multiple comparison test. The Mann–Whitney U test was used for a non-parametric analysis of non-normally distributed data. Statistical significance was defined as P < 0.05. In the text and graphs, data are presented as mean ± standard error. The data were analysed using GraphPad Prism software version 9.0.2.

3. Results

3.1 IR expression and signalling in the aortas of ApoE^−/− and SMIRKO/ApoE^−/− mice

The expression of IRs in the aortas of ApoE^−/− mice on normal chow (NC) and HFD and SMIRKO/ApoE^−/− mice on NC was characterized. The expression of the IR protein significantly decreased by 48% in HFD-fed vs. NC-fed ApoE^−/− mice (Figure 1A and B), and IR protein levels were undetectable in SMIRKO/ApoE^−/− mice (Figure 1D). Insulin signalling, assessed by Akt phosphorylation (p-Akt) in the aorta, showed insulin-induced p-Akt reductions of 50% in ApoE^−/− mice on HFD (Figure 1A and C, P < 0.01) and 73% in SMIRKO/ApoE^−/− mice (Figure 1D and E, P < 0.01) compared with the NC-fed ApoE^−/− mice.

3.2 Characterization of metabolic phenotypes and atherosclerosis

The mice were fed with NC for 11 months. With regard to body weight, blood pressure, plasma cholesterol, triglyceride, VLDL, LDL, and HDL levels, and intraperitoneal glucose tolerance test (IPGTT) and intraperitoneal insulin tolerance test (IPITT) results, there were no significant differences between SMIRKO/ApoE^−/− and ApoE^−/− mice (see Supplementary material online, Figure S1). In measurements of aortic plaques, the extent of en face aortic lipid deposition, quantified by Sudan IV staining, was 89% higher in SMIRKO/ApoE^−/− than in ApoE^−/− mice (Figure 2A, P < 0.05). VSMC content in aortic plaques, per SM22α⁺ immunostaining, decreased by 46% (P < 0.05), whereas no difference was found with regard to Mac2⁺ immunostaining, which measures macrophage content, between SMIRKO/ApoE^−/− and ApoE^−/− mice (Figure 2B). The necrotic area increased by 41% (P < 0.05) in the plaques of SMIRKO/ApoE^−/− mice compared with those of ApoE^−/− mice (Figure 2C). Collagen content, measured by Sirius red staining, decreased by 29% (P < 0.05) in the plaques of SMIRKO/ApoE^−/− mice compared with those of ApoE^−/− mice (Figure 2D and E).

Inflammatory cell compositions of aortic plaques were analysed by flow cytometry. Gating schema are shown in Supplementary material online, Figure S2. The total leucocytes, as measured by CD45⁺ cells, were increased in the aorta of SMIRKO/ApoE^−/− mice compared with that of ApoE^−/− mice (Figure 2F). There was an increase in the proportion of CD45⁺ cells within the total cell count in the aortas of SMIRKO/ApoE^−/− mice (see Supplementary material online, Figure S3); however, this increase did not reach statistical significance. It is worth noting that there was red blood cell contamination during aorta isolation despite phosphate-buffered saline perfusion. Consequently, the calculated percentage of CD45⁺ cells within the total cell count could be affected by red blood cell contamination and may not accurately represent the true quantity of CD45⁺ cells within the aorta. Dendritic cells (CD45⁺CD11C⁺), monocytes (CD45⁺F4/80⁻CD3⁻CD19⁻Ly6G⁻CD11C⁻CD11b⁺), and inflammatory Ly6C^hi monocytes were all increased by 265, 106, and 220%, respectively, in the aorta of SMIRKO/ApoE^−/− mice compared with that of ApoE^−/− mice (Figure 2F, P < 0.05). Ideally, the cell number in the aorta should be normalized to the weight of the aorta. However, accurately measuring the weight of the aorta required it to be dried, which might affect cell viability. Therefore, the cell number was not normalized by the weight of aorta. No differences in the circulating leucocyte levels in the blood, BM, and spleen were found between these two groups of mice (see Supplementary material online, Figure S4).

Since the SM22 Cre promoter is also expressed in myeloid cells,²⁴ CD45⁺CD11b⁺ cells were sorted by flow cytometry from the aorta to determine IR expression in aortic CD45⁺CD11b⁺ cells. IR mRNA expression was reduced by 70% in CD45⁺CD11b⁺ cells from the aorta of SMIRKO/ApoE^−/− mice compared with control mice (see Supplementary material online, Figure S5). To definitively determine whether the increased atherosclerosis in SMIRKO/ApoE^−/− mice was due to a loss of IRs in VSMCs, but not due to IR changes in myeloid cells, the BM of ApoE^−/− mice was transplanted into lethally irradiated ApoE^−/− or SMIRKO/ApoE^−/− mice. After 4 months of feeding HFD, no differences in body weight, plasma cholesterol and plasma triglyceride levels, and IPITT results were noticed between the two groups of mice (see Supplementary material online, Figure S6). Glucose tolerance was impaired in the BM-transplanted ApoE^−/− mice compared with the BM-transplanted SMIRKO/ApoE^−/− mice (see Supplementary material online, Figure S6D). Sudan IV staining of the aorta showed that the extent of atherosclerotic plaques was still significantly greater in the BM-transplanted SMIRKO/ApoE^−/− mice compared with the BM-transplanted ApoE^−/− mice (see Supplementary material online, Figure S7). IR protein levels were comparable in the macrophages from these two types of mice (see Supplementary material online, Figure S7). These data suggested that the deficiency of IRs in the VSMCs of SMIRKO/ApoE^−/− mice contributed to the accelerated atherosclerosis formation in the aorta of these mice.

3.3 VSMC proliferation and apoptosis in atherosclerotic plaques and in vitro

Changes in apoptosis and cellular proliferation of VSMCs were studied because the pathology of the aortic plaques from SMIRKO/ApoE^−/− mice differed from that of most rodent models of diabetes and the plaques had reduced VMSC content. The proliferation index of VSMCs in the atherosclerotic plaques, determined by dual SM22α and Ki67 staining, was decreased by 69% in the aorta of SMIRKO/ApoE^−/− mice compared with ApoE^−/− mice (Figure 3A and B, P < 0.05). TUNEL and SM22α dual labelling, which estimated apoptosis in aortic VSMCs, was increased by 2.8-fold in the atherosclerotic plaques of SMIRKO/ApoE^−/− mice compared with ApoE^−/− mice (Figure 3C and D, P < 0.05). Moreover, TUNEL and SM22α double-positive cells were also increased in the artery wall outside the plaque (see Supplementary material online, Figure S8).

The effects of IR deletion on cellular replication in cultured aortic VSMCs, as measured by EdU incorporation and flow cytometry, showed that insulin (5 nM) increased VSMC proliferation by 30% in control cells, but only by 17% in IR-deficient VSMCs (Figure 3E, P < 0.05). Since the activation of Akt (p-Akt), a major signalling pathway for insulin, was reduced in cells from SMIRKO/ApoE^−/− mice, its role in VSMC proliferation was studied. GDF-0068, an Akt inhibitor,²⁵ had no effect on the basal state, but it inhibited insulin-stimulated DNA synthesis significantly in VSMCs cultured from the aortas of ApoE^−/− mice (Figure 3F, P < 0.01). Conversely, an overexpression of constitutively activated Akt (AdCA-Akt) in VSMCs cultured from the aortas of ApoE^−/− mice using adenoviral vectors increased the expression of the proliferation marker Ki67 by 100% (Figure 3G, P < 0.01). The levels of apoptosis, measured by DNA fragmentation induced by withdrawing growth factors, were reduced by insulin (31%, P < 0.01) in VSMCs cultured from the aortas of ApoE^−/− mice, but were unaffected in IR-deficient VSMCs cultured from the aortas of SMIRKO/ApoE^−/− mice (Figure 3H). In contrast, constitutively activated Akt (AdCA-Akt) reduced apoptosis significantly in the VSMCs of SMIRKO/ApoE^−/− mice (Figure 3I). In summary, insulin promoted VSMC proliferation and inhibited VSMC apoptosis through the Akt pathway.

3.4 Expression of inflammatory cytokines in aortic VSMCs

Macrophage-like VSMCs have been reported,²² but cell surface markers for isolating macrophage-like VSMCs have not been identified. We hypothesized that VSMCs expressing Icam1 and Vcam1 were inflammatory in nature. To definitely demonstrate that the source of cells is from VSMCs, aortae from VSMC lineage tracing mice (MyH11EYFP⁺/ApoE^−/− mice) on HFD for 5 months were digested into single cells. EYFP⁺Icam1⁻Vcam1⁻ and EYFP⁺Icam1⁺Vcam1⁺ VSMCs were sorted by flow cytometry, using gating schema as shown in Supplementary material online, Figure S9. Gene expression levels in these cells were analysed by qPCR. Contractile genes, Acta2 and Myh11, were decreased in EYFP⁺Icam1⁺Vcam1⁺ VSMCs compared with EYFP⁺Icam1⁻Vcam1⁻ VSMCs (Figure 4A). Interestingly, IR expressions were also significantly lower in Icam1⁺ Vcam1⁺VSMCs than in Icam1⁻ Vcam1⁻VSMCs (Figure 4A). However, inflammatory cytokines such as Icam1, Vcam1, IL-6, Ccl2, Cxcl1, Cxcl12, thrombospondin 1 (Thbs1), C3, and CD45 were markedly increased in EYFP⁺Icam1⁺Vcam1⁺ VSMCs compared with EYFP⁺Icam1⁻Vcam1⁻ VSMCs. Next, we investigated whether inflammatory VSMCs (Icam1⁺ VSMCs) were increased in the aortas of SMIRKO/ApoE^−/− mice. The media layer of thoracic aortas contained mainly VSMCs. VSMCs (PI⁻Ter119⁻CD45⁻ CD31⁻ cells) isolated from the media layer of the thoracic aortas of SMIRKO/ApoE^−/− mice had significantly more Icam1⁺ VSMC content than in the aortas of ApoE^−/− mice (Figure 4B). Cre driven by the Myh11 promoter specifically knocked out floxed genes in VSMCs but not in myeloid cells. Therefore, VSMC lineage tracing mice and VSMC IR-knockout mice were generated by cross-breeding Myh11-CreER^T2EYFP⁺/ApoE^−/− and IR^flox/flox mice (Myh11-CreER^T2EYFP⁺IRKO/ApoE^−/− mice, henceforth referred to as Myh11EYFP⁺IRKO/ApoE^−/− mice). These mice were fed with HFD to mimic a Western pattern of diet in humans. There were no differences in body weight, plasma cholesterol level, and IPITT and IPGTT results between the two groups of mice after HFD for 5 months (see Supplementary material online, Figure S10). In addition, there were no differences in IR expressions in the monocytes in Myh11EYFP⁺IRKO/ApoE^−/− and control ApoE^−/− mice. The extent of atherosclerosis, as measured by en face Sudan IV staining, was increased in the aorta of Myh11EYFP⁺IRKO/ApoE^−/− mice compared with that of control mice (Figure 4C and D). The aortas of Myh11EYFP⁺/ApoE^−/− and Myh11EYFP⁺IRKO/ApoE^−/− mice were digested into single cells and inflammatory VSMCs (DAPI⁻EYFP⁺Icam1⁺Vcam1⁺ VSMCs) were determined by flow cytometry. The ratio of DAPI⁻EYFP⁺Icam1⁺Vcam1⁺ VSMCs in a DAPI⁻EYFP⁺ VSMC was increased in the aorta of MyH11EYFP⁺IRKO/ApoE^−/− mice compared with that of Myh11EYFP⁺/ApoE^−/− mice (Figure 4E). The EYFP⁺ VSMC was sorted from the aorta of MyH11EYFP⁺/ApoE^−/− mice or that of MyH11EYFP⁺IRKO/ApoE^−/− mice, and the gene expression levels in these cells were analysed by qPCR. As expected, IR mRNA expression was significantly lower in the EYFP⁺VSMC from MyH11EYFP⁺IRKO/ApoE^−/− mice than from MyH11EYFP⁺/ApoE^−/− mice. Ccl2 was significantly increased, whereas Cxcl1 and Thbs1 had a tendency to be increased in the EYFP⁺VSMC from MyH11EYFP⁺IRKO/ApoE^−/− mice compared with that from MyH11EYFP⁺/ApoE^−/− mice (Figure 4F). There were no differences in the expression of contractile genes SM22α, Acta2, and Myh11 in the EYFP⁺ VSMC from the two groups of mice (Figure 4F).

3.5 Characterization of genes related to unstable plaques

These findings have shown that a reduction of IRs by HFD or IR deletion in VSMCs induces these cells to acquire more inflammatory properties. To identify genes in VSMCs that could be regulated by insulin and contribute to the composition of unstable plaques, we reanalysed the database provided by Wirka et al.,²⁶ who reported gene profiles derived from a single-cell analysis of VSMCs from the atherosclerosis plaques of ApoE^−/− mice on a Western diet. Differentially expressed genes between Icam1⁺ and Icam1⁻ VSMCs are shown in the volcano plot (see Supplementary material online, Figure S11). Contractile genes, including Myh11, Dstn, and Cnn1, were decreased in VSMCs expressing high levels of Icam1 cells. Supplementary material online, Table S1 identifies top genes that are increased by 1.8-fold and P < 10⁻⁹ in the Icam1⁺ VSMC, which expressed higher levels of inflammatory genes, including Mmp2, C1ra, Vcam1, Thbs1, Ccl19, Tnfsf11b, and C4b. These genes were screened for their potential of being regulated by IRs using the media layer of thoracic aortas from Apoe^−/− and SMIRKO/ApoE^−/− mice. The expressions of five genes, namely, Mmp2, Sfrp3, Comp, Igfbp4, and Thbs1, were consistently elevated in the media layer of the thoracic aortas of SMIRKO/ApoE^−/− mice (Figure 5A). The expressions of these five genes were further quantified in the media layer of the thoracic aortas of control and diabetic mice with HFD-injected (hyperinsulinaemic and insulin-resistant) or streptozotocin-injected (STZ, insulin deficient) mice. The expression levels of Mmp2, Sfrp3, and Thbs1 were increased in the aorta media of HFD mice and STZ mice compared with control mice (Figure 5B).

It has been reported that Mmp2 mRNA is elevated in the arteries of people with diabetes compared with non-diabetic individuals,²⁷ and Thbs1 is up-regulated in tissues of diabetic patients.^28,29 Thus, subsequent studies focused on the regulation of Mmp2 and Thbs1 in VSMCs by insulin and diabetes. To confirm whether Thbs1 and Mmp2 expressions were also associated with IR deletion, we showed that the protein levels of both Thbs1 (3.1-fold) and Mmp2 (1.6-fold) were significantly elevated in the aorta of SMIRKO/ApoE^−/− vs. ApoE^−/− mice (Figure 5C).

3.6 Regulation of Thbs1 and Mmp2 expressions

The up-regulation of Mmp2, Sfrp3, Comp, Igfbp4, and Thbs1 gene expressions in the aortic media of SMIRKO/ApoE^−/− mice could be attributed to the direct or indirect effects of insulin on VSMCs. In cultured VSMCs from ApoE^−/− mice, treatment with 100 nM insulin for 24 h inhibited 30% of Thbs1 gene expression, but not of Mmp2, sFRP3, Comp, and Igfbp4 (Figure 6A). Insulin inhibited Thbs1 in a dose- and time-dependent manner, with inhibition rates of 50 and 62% by 10 and 100 nM insulin, respectively, after treatment for 7 days (see Supplementary material online, Figure S12). The regulation of Thbs1 expression in VSMCs was studied further since the loss of this IR function could provide mechanistic information for the up-regulation of Thbs1 expression in many mouse models of diabetes and in arteries from clinical studies.^30,31 Insulin (5 nM) inhibited Thbs1 protein levels by 33% in culture media of aortic VSMCs from ApoE^−/− mice (P < 0.05), but not from SMIRKO/ApoE^−/− mice (Figure 6B). Basal Thbs1 expression in VSMCs from SMIRKO/ApoE^−/− mice was elevated compared with those from ApoE^−/− mice (Figure 6B, P < 0.05).

To investigate whether Thbs1 can affect the expression of various inflammatory genes in VSMCs, the VSMCs were exposed to Thbs1 protein at 1 µg/mL for 4 h. The addition of Thbs1 increased Icam1 and Mmp2 expressions by 87 and 16%, respectively (P < 0.05), in VSMCs cultured from the aortas of ApoE^−/− mice (Figure 6C). Mmp2 basal expression levels were increased in VSMCs from the aortas of SMIRKO/ApoE^−/− mice, but were not further increased after Thbs1 stimulation (Figure 6C, P < 0.05). The addition of Thbs1 increased IL6 and Ccl2 expressions in VSMCs from ApoE^−/− mice, but the difference was not significant. Thbs1 was further knocked down by adenovirus-mediated shRNA transfer in VSMCs from ApoE^−/− mice. The knockdown of Thbs1 reduced the gene expression levels of Thbs1, IL6, and C3 (Figure 6D). Ccl2 was also reduced in Thbs1 knockdown VSMCs, but the difference was not significant. Therefore, the expression of Thbs1 in VSMCs was inhibited by insulin via Akt activation, leading to a down-regulation of inflammatory genes such as IL6, C3, and ccl2.

It has been reported that Thbs1 has both pro-proliferation³² and anti-proliferation^33,34 effects in VSMCs. The proliferation was increased by Thbs1 after 48 h of treatment, as measured by EdU incorporation in the VSMC (see Supplementary material online, Figure S13A). Insulin’s pro-proliferative effects on VSMCs were not associated with Thbs1 since insulin inhibited Thbs1 expression. For apoptosis, per DNA fragmentation, Thbs1 did not have any effects (see Supplementary material online, Figure S13B).

3.7 Mechanism of insulin’s regulation of Thbs1 expression

Insulin decreased Thbs1 gene expression in a dose-dependent manner (Figure 6A and see Supplementary material online, Figure S11). To characterize the mechanism of insulin’s action on decreasing Thbs1 expression, we studied whether the p-Akt and FoxO1 (p-FOXO1), major signalling pathways for insulin and IRs, could regulate Thbs1 expression. IR levels were significantly decreased in VSMCs from SMIRKO/ApoE^−/− mice compared with those from ApoE^−/− mice (Figure 7A). In cultured VSMCs from SMIRKO/ApoE^−/− mice compared with those from ApoE^−/− mice, p-Akt induced by insulin treatment for 3 h was decreased by 58% (P = 0.006), whereas insulin-induced p-FoxO1(Thr24) was decreased by 46% (P = 0.047) after insulin treatment for 3 h and by 40% (P = 0.025) after insulin treatment for 5 h.

To determine the role of p-Akt on insulin-induced inhibition of Thbs1 expression, Akt expression was knocked down by shRNA lentivirus in VSMCs from ApoE^−/− mice. Insulin significantly reduced Thbs1 mRNA expression in control VSMCs but not in Akt shRNA-treated VSMCs, suggesting that insulin inhibited Thbs1 gene expression through the Akt pathway (Figure 7B). ApoE^−/− VSMCs, overexpressing AdCA-Akt, exhibited 65% significant reductions of Thbs1 mRNA (Figure 7C). In contrast, overexpressing FoxO1 by adenovirus in the ApoE^−/− VSMC significantly increased Thbs1 mRNA expression by 1.83-fold (Figure 7D). Similarly, AdCA-Akt overexpression significantly increased Akt protein level but inhibited Thbs1 protein significantly in culture media by 70% (Figure 7E and F). In contrast, overexpressing FoxO1 significantly increased both FoxO1 and Thbs1 protein levels in cell culture media by 2.88-fold (Figure 7E and F). Both insulin and a FoxO1 inhibitor AS1842856 independently suppressed the expression of the Thbs1 gene. However, when insulin was combined with a FoxO1 inhibitor, resulting in maximal inhibition of FoxO1, there was a further reduction in Thbs1 gene expression (Figure 7G).

4. Discussion

Our findings show that deletion or reduction of IRs in VSMCs exacerbated inflammation and cellular apoptosis, while also reducing VSMC proliferation and ECM in the arterial wall. These changes resulted in augmenting the severity of atherosclerosis and the unstable properties of atherosclerotic plaques. The finding that IR deletion in VSMCs promotes unstable properties of arterial plaques in mouse models of atherosclerosis with insulin-resistant diabetes is unusual, as most studies have reported that diabetes or insulin resistance accelerates atherosclerosis with corresponding increases of VSMCs and ECM in the plaques, without affecting plaque stability.^19,20,35 ApoE^−/− or LDLR^−/− mice on HFD with insulin resistance, hyperglycaemia and hyperinsulinaemia have more severe atherosclerotic plaques but they do not exhibit decreases in ECM and VSMCs in the plaque cap. We previously reported that IR deletion targeting endothelial cells in ApoE^−/− mice accelerated atherosclerotic plaque severity with elevated VSMCs, ECM, inflammatory cells, and lipid deposits.¹⁵ Similarly, an overexpression of Glut1 in the VSMCs of mice also increased macrophage and VSMC content with a greater amount of atherosclerotic plaques.³⁶ For diabetes with insulin deficiency, and severe hyperglycaemia without insulin resistance induced by streptozotocin in ApoE^−/− mice, plaque severity was increased with parallel elevations of VSMCs, ECM, and inflammatory cells.^20,35

It is interesting to note that only SMIRKO/ApoE^−/− mice manifested this discordance, but not HFD-fed ApoE^−/− mice, although both exhibited insulin resistance in the arterial wall. This difference in pathology could be due to the lack of hyperinsulinaemia and the loss of insulin actions on Akt/MAPK pathways in the SMIRKO/ApoE^−/− mice. In ApoE^−/− mice on HFD, hyperinsulinaemia compensated for the selective loss of insulin activation of the IRS/PI3K/Akt pathway to prevent decreases in VSMC proliferation and increases in apoptosis. Our previous report on VSMCs from either IRKO or IGF1RKO mice showed that a homozygous IR can mediate greater VSMC proliferation after intimal injury than IGF1R,¹² which paralleled the enhancement of insulin’s activation of the PI3K/Akt pathway. This finding that physiological levels of insulin can activate the P13K/Akt pathway to induce mitogenic actions is very important because most previous studies have attributed insulin mitogenic actions to the activation of MAPK, which requires supra-physiological insulin levels that are not observed in people with insulin resistance or Type 2 diabetes.³⁷ IGF1’s mitogenic actions have also been postulated to have atherogenic actions. However, Sukhanov et al.³⁸ reported that IGF1R deletion–targeted VSMCs with the SM22α promoter in ApoE^−/− mice exhibited more atherosclerotic plaques. However, findings in the IGF1R deletion mouse were confounded by systemic changes such as the growth hormone/IGF1 axis that led to stunted growth and small arterial vasculature. Further, our previous study that compared IR and IGF1R deletion in the VSMC demonstrated that the homozygous IR had much greater effects in cultured VSMCs and in vivo on the proliferation of VSMCs¹² than IGF1R.

Loss of IRs led to increased inflammatory VSMCs in arterial walls and plaques. This was documented by lineage tracing studies of Myh11EYFP⁺/ApoE^−/− mice on HFD or IRKO, which showed that VSMCs had a greater expression of Icam1 and Vcam1, indicating that the loss of IRs shifted VSMCs from subsets of cells expressing low levels of inflammatory cytokines to those expressing high levels of these cytokines. The shift to an inflammatory VSMC profile is due to IR deletion since the aorta from SMIRKO/ApoE^−/− mice or MyH11EYFP⁺IRKO/ApoE^−/− mice contains a greater percentage of cells with higher expressions of Icam1. Similarly, VSMCs from ApoE^−/− mice, which have more Icam1 and Vcam1, expressed significantly less IR than cells with low expressions of these inflammatory cytokines. These findings provide the first direct evidence that IR deletion may change or induce VSMCs into a more inflammatory profile similar to those of macrophages. Insulin resistance is already present in the arterial walls of individuals with insulin resistance and Type 2 diabetes prior to the onset of atherosclerosis. Atherosclerosis is a chronic inflammatory disease. It is well-established that inflammation contributes to insulin resistance. In the late stage of atherosclerosis, severe inflammation within atherosclerotic plaques can greatly exacerbate insulin resistance in VSMCs, even in non-diabetic individuals. This exacerbation inhibits VSMC proliferation, increases VSMC apoptosis, and enhances ECM turnover, ultimately promoting the development of unstable plaques. Therefore, insulin resistance can contribute to unstable plaques in both diabetic and non-diabetic statuses.

The existence of macrophage-like VSMCs has been well-documented previously²² and can contribute significantly to the severity of atherosclerosis. The mechanism that could be mediating insulin’s actions to enhance inflammatory VSMCs and reduction of ECM was explored by analysing differential gene expressions between Icam1-positive and Icam1-negative cells using the RNA database of the single-cell analyses of atherosclerosis plaques in ApoE^−/− mice provided by Wirka et al.²⁶ After further validation by changes induced by HFD, diabetes, and insulin, the strongest candidates for mediating the exacerbation of inflammation and loss of ECM are the increased expressions of Thbs1 and Mmp2. Increased expressions of both Thbs1 and Mmp2 have been reported to be elevated consistently in atherosclerotic plaques.³⁹

Thbs1 expression has been reported to be affected by FoxO1 in endothelial cells.⁴⁰ However, our findings provided the first report to show that insulin via IR and Akt/FoxO1 pathway can down-regulate Thbs1 expression in VSMCs. With the selective inhibition of the Akt pathway in insulin resistance in Type 2 diabetes and obesity, these findings provide a mechanistic explanation for the elevated Thbs1 expression in insulin-resistant states. Thbs1 is up-regulated at sites of inflammation, atherosclerosis, and tissue injury.^39,41 With multiple binding sites for cell surface receptors such as CD36 and CD47,^42,43 Thbs1 could regulate the function of various cells, including VSMCs⁴⁴ with a spectrum of inflammatory and pro-atherogenic or thrombotic^45–47 effects. Our findings showed that Thbs1 can regulate the expressions of inflammatory cytokines, including Icam1, Mmp2, and, possibly, IL6, in VSMCs. Thus, the loss of insulin actions in VSMCs as observed in SMIRKO/ApoE^−/− mice and in insulin resistance may lead to an elevated expression of Thbs1, followed by increases in inflammatory cytokines and macrophage-like VSMCs that attract more inflammatory cells to the arterial wall.

The reduction of ECM, another feature of unstable plaque, is likely related to the increased expressions of Mmp2. Many studies have reported on Mmp2 elevations in the aorta of diabetic animals and in the arteries of people with diabetes.^27,48 Its increases have been related to unstable plaques and rupture of aortic aneurysms.⁴⁹ However, the mechanism for the elevation of Mmp2 in insulin resistance and diabetes has not been elucidated. Our findings suggest that the increased expressions of Mmp2 in the aorta of SMIRKO/ApoE^−/− and other diabetic mice are not directly attributed to IR but are indirectly related to the increased expression of Thbs1, which is due to the loss of insulin’s inhibitory action via the Akt/FoxO pathway in VSMCs.

A comparison between SMIRKO/ApoE^−/− mice and control mice was done using mice fed with NC, which was appropriate for elucidating the role of IRs in an unstressed state. To simulate human insulin resistance induced by a Western pattern of diet, Myh11EYFP⁺IRKO/ApoE^−/− mice and control mice were fed with HFD. The HFD induced hyperglycaemia, hyperinsulinaemia, hyperlipidaemia, and changes in hormone levels that mimic conditions of Type 2 diabetes, a strong risk factor associated with unstable plaques.

It has been reported that ionizing radiation promoted HFD-induced insulin resistance by affecting the skeletal muscle and adipose progenitor cells.⁵⁰  Supplementary material online, Figure S6D shows that ionizing radiation reduced glucose tolerance in ApoE^−/− mice as others have reported, but it had less effects on SMIRKO/ApoE^−/− mice. In addition, the IPITT and IPGTT results showed no differences between Myh11EYFP⁺/ApoE^−/− and Myh11EYFP⁺IRKO/ApoE^−/− mice (see Supplementary material online, Figure S10C and D), indicating that knockout of IRs in VSMCs did not alter systemic insulin sensitivity. The IPITT result was also similar in irradiated ApoE^−/− mice and SMIRKO/ApoE^−/− mice (see Supplementary material online, Figure S6C). Therefore, the disparity in IPGTT outcomes between irradiated ApoE^−/− mice and irradiated SMIRKO/ApoE^−/− mice may be attributed to variations in glucose disposal in various tissues and possibly to insulin secretion.

There are several limitations to this study. The mouse models are of ApoE^−/− background, which are not commonly encountered in clinical settings. The mechanistic studies used cultured VSMCs, which are known to de-differentiate in culture. Thus, these findings will need to be replicated using human arteries with and without atherosclerosis and with and without insulin resistance and diabetes. The potential mechanisms of insulin regulation of VSMC differentiation will need to be investigated using single-cell RNA sequencing in future studies for a clear demonstration of the various subsets of a VSMC.

In summary, this study provides novel findings that insulin’s signalling and actions in VSMCs can decrease inflammation, ECM turnover, and apoptosis with increased cellular proliferation. The selective loss of insulin signalling via Akt activation in VSMCs as found in insulin resistance and diabetes could be responsible for the unstable plaque pathologies observed in these cardio-metabolic diseases.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Acknowledgements

The authors are grateful to Dr Gary Owens from the University of Virginia, who provided them VSMC lineage tracing mice (Myh11-CreER^T2EYFP⁺/ApoE^−/− mice), and to Dr C. Ronald Kahn from the Joslin Diabetes Center, who provided them IR^flox/flox mice. The authors thank Dr Mingming Ma for editing the manuscript. They also acknowledge the expert technical assistance provided by Bioinformatic Core, Advanced Microscopy Core, Flow Cytometry Core, and Mouse Physiology Core.

Funding

This work was supported by the National Institutes of Diabetes and Digestive and Kidney Diseases, grant R01-DK-053105; Beatson Foundation Gift; and Dianne Nunnally Hoppes Fund. J.F. was supported by a Mary K. Iacocca research fellowship. Biostatistic and Bioinformatic Core, Advanced Microscopy Core, Flow Cytometry Core, Mouse Physiology Core, and Genomic Core are supported by NIH grants 5P30DK036836 and S10OD021740.

Data availability

Data are available from the corresponding author upon reasonable request.

References

Author notes