The AAV-PCSK9 murine model of atherosclerosis and metabolic dysfunction.

Aims: Mouse models with genetic modifications are required to investigate atherogenesis and associated metabolic syndrome. Adeno-associated virus-8 (AAV8)-mediated overexpression of PCSK9 (AAV8-PCSK9) induces hyperlipidaemia and promotes atherosclerosis in C57BL/6 mice. We aimed to assess whether AAV8-PCSK9-injected C57BL/6 mice fed high-fat diet with added cholesterol (HFD-C) would serve as a model of combined metabolic syndrome and atherosclerosis. Methods and results: C57BL/6 mice received i.v. injection of AAV-PCSK9 and sex- and age-matched Ldlr-/- and C57BL/6 control mice were placed on HFD-C or chow diet for 20 weeks (B6-PCSK9-HFD-C, Ldlr-/- HFD-C, B6-HFD-C, and B6-Chow, respectively). High-fat diet with added cholesterol feeding led to insulin resistance and impaired glucose clearance in B6-PCSK9-HFD-C mice compared with B6-Chow controls. This decrease in metabolic health in B6-PCSK9-HFD-C mice as well as the development of atherosclerosis was similar to Ldlr-/- HFD-C mice. Importantly, HFD-C feeding induced pancreatic islet hyperplasia in B6-PCSK9-HFD-C and B6-HFD-C compared with B6-Chow controls. In line with alterations in the metabolic phenotype, there was an increase in the number of pro-inflammatory Ly6Chigh/med monocytes within the adipose tissues of B6-PCSK9-HFD-C and B6-HFD-C compared with B6-Chow controls.
Conclusion: High-fat diet with added cholesterol-fed AAV-PCSK9-injected C57BL/6 mice can serve as a useful model of integrated metabolic syndrome and atherosclerosis that does not require genetic manipulations.

Introduction

The coincidence of cardiovascular disease and obesity-related Type-2 diabetes (T2D) is widely known.[1] Recently, a high-fat diet with added cholesterol (HFD-C) was reported as a suitable model for insulin resistance, glucose intolerance, and atherosclerosis in low-density lipoprotein receptor-deficient (Ldlr/) mice.[2,3] While the HFD-C-fed Ldlr mouse model is a great tool to study atherogenesis and insulin resistance, crossing these mice with other transgenic mice requires substantial time and resources. A novel model of atherosclerosis can bypass such crossing via a single injection of recombinant adeno-associated virus (AAV) encoding a gain-of-function mutant form of proprotein convertase subtilisin/kexin type 9 (PCSK9) that targets hepatic LDLR and causes a rapid increase of circulating cholesterol and atherogenesis when paired with high-fat diet.[4] In humans, PCSK9 levels significantly correlate with several parameters of metabolic syndrome including obesity, hypertension, fasting blood glucose, and serum insulin levels.[5] Current research aims to understand whether PCSK9 levels are a result of insulin resistance, or if PCSK9 plays a contributory role in glucose metabolism. To date, the AAV-PCSK9 model of atherosclerosis has not been evaluated as a model of metabolic syndrome and atherosclerosis. Here, we hypothesize that AAV-PCSK9-induced hyperlipidaemia and HFD-C feeding will contribute to increased glucose intolerance and insulin resistance providing a suitable model for the investigation of concomitant metabolic syndrome and atherosclerosis.

Methods

AAV-PCSK9 vector production

The PCSK9 adeno-associated virus (rAAV8/D377Y-mPCSK9) was produced at the University of North Carolina (Chapel Hill, NC, USA) using the PCSK9 gain-of-function plasmid (Addgene, #58376).[4]

Animals and study design

Eight-week-old male C57Bl/6J mice (n = 10), were intravenously injected via tail vein with a single dose of 1 × 1011 AAV-PCSK9 particles and placed on a high-fat diet with 0.15% added cholesterol (HFD-C; fat: 60% kcal; carbohydrate: 26% kcal, Bio-Serv, F3282). Age- and sex-matched C57BL/6 and Ldlr mice (Jackson, #000664, #002207, respectively) were placed on either HFD-C or chow diet for 20 weeks. Glucose tolerance (GTT, 2 g/kg) and insulin tolerance (ITT, 5 mU/g) tests were performed at 20 weeks HFD-C feeding as described previously.[3] Mice were given 2 days recovery between ITT and GTT testing, and tissues were collected 1 week later. Aortas were excised and stained for atherosclerotic lesions using Oil Red O and the percent area occupied by lesions was determined using ImageJ.[3] All animal protocols were approved by the EVMS Institutional Animal Care and Use Committee.

Plasma lipid analysis

Concentrations of total cholesterol (TC, Wako, #999-02601), LDL (LDL-C, Crystal Chem, #79980), triglycerides (TG, Cayman Chemical, #10010303) were measured in plasma of non-fasted or 16 h-fasted mice. HDL (HDL-C) level was calculated based on the TC, LDL, and TG levels using the Friedewald equation. B6-PCSK9-HFD-C mice that achieved <300 mg/dL TC were omitted from the study (n = 3).

Histological analysis of islet architecture

Formalin-fixed pancreas was stained with anti-insulin antibodies (Cell Signaling Technology #4590S), then with secondary antibodies (ThermoFisher, #A21206), and islet size/number was determined using Olympus cellSens software. Two sections of the pancreas greater than 200 µm apart (average 15–25 islets/section) were used for analysis.

Flow cytometry of digested adipose tissues

Epididymal adipose tissues (AT) were digested as previously described,[6] and single-cell suspension was stained for CD45 and a Ly6C marker for pro-inflammatory Ly6Chigh/mid monocytes (Biolegend, #103130, #128006), analyzed via FACSCalibur (BD Biosciences/Cytek Inc.), and FlowJo analysis software (Tree Star Inc.). B6-Chow AT data represent two mice pooled together.

Statistics

Data are reported as mean ± SEM. One-way ANOVA with Sidak’s multiple comparisons test was used for statistical comparisons between the groups. All data sets were tested and confirmed for Gaussian distribution via the Shapiro–Wilk test. Outliers were determined as mean ± 1.5×SD. Statistical significance was set at P < 0.05.

Results

To initially establish the effects of HFD-C feeding on metabolic parameters, C57BL/6 mice underwent ITT and GTT testing after 20 weeks of either chow (B6-Chow) or HFD-C diet feeding. B6-HFD-C mice showed a significant insulin resistance and impaired glucose tolerance compared with B6-Chow mice (). Next, we investigated the concomitant impact of HFD-C feeding and hypercholesterolaemia on the development of metabolic dysfunctions in HFD-C fed AAV-PCSK9 (B6-PCSK9-HFD-C) mice. Our data demonstrate that achieved levels of insulin resistance and glucose intolerance in B6-PCSK9-HFD-C mice were comparable to levels of metabolic dysfunctions in HFD-C-fed Ldlr mice (). Interestingly, B6-PCSK9-HFD-C and Ldlr HFD-C mice showed some improvement in glucose homeostasis compared with B6-HFD-C at later time points (); however, the area under curve was not statistically significant (). Thus, combinatory effects of elevated hypercholesteraemia and HFD-C feeding are enough to induce metabolic phenotype in C57BL/6 mice without the addition of LDLR-deficiency. High-fat diet with added cholesterol feeding induced some small lesions in B6-HFD-C mice, and the plaque development was further significantly increased in B6-PCSK9-HFD-C mice (). While HFD-C feeding significantly increased the body weight of B6-HFD-C mice, further elevation of TC in B6-PCSK9-HFD-C and Ldlr HFD-C mice had minimal effect on body weight (). Total cholesterol, LDL, and TG levels were significantly elevated in 16 h fasted B6-PCSK9-HFD-C and Ldlr HFD-C mice vs. B6-HFD-C mice (), which is typically associated with decreased insulin sensitivity. HDL levels were the highest in the plasma of Ldlr HFD-C mice (). Additionally, 16-h fasting glucose was increased in B6-HFD-C vs. B6-chow mice but was not further elevated in the B6-PCSK9-HFD-C and Ldlr HFD-C groups (). Non-fasting TC levels were significantly increased in B6-HFD-C in comparison with B6-chow mice with further increase in B6-PCSK9-HFD-C mice (159 ± 7, 361 ± 20, and 500 ± 56 mg/dL, respectively, P < 0.05), whereas non-fasting TG levels were not different between these groups (112 ± 19, 125 ± 9, and 122 ± 19 mg/dL, respectively). These findings agree with previous reports whereby fasting increases TC and TG levels in mice with PCSK9 over-expression.[7] As B6-PCSK9-HFD-C mice showed a promisingly similar metabolic phenotype to Ldlr HFD-C, we further investigated changes in pancreatic islet architecture as a reflection of metabolic syndrome in the PCSK9-HFD-C model.

High-fat diet with added cholesterol-fed AAV-PCSK9-injected C57BL/6 mice represents a combined model of metabolic syndrome and atherosclerosis. (A–D) Time course data of blood glucose for insulin tolerance and glucose tolerance tests with area under the curve measurements. (E–G) Fasting total cholesterol, LDL, and HDL levels. (H) Representative images of en face aortas and (I) quantification of plaque area. (J–K) Fasting plasma triglyceride and glucose levels. (L) Body weight at 20 weeks high-fat diet with added cholesterol feeding. One-way ANOVA with Sidak’s multiple comparisons test was used for statistical analysis.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n = 6–10 mice/group.

At the initial stages of pre-diabetic conditions, insulin resistance can cause a compensatory increase in insulin production within the islets, which results in larger islet size and mass in mice.[8] In line with the metabolic profile of B6-PCSK9-HFD-C mice, an increased average islet size was observed in B6-PCSK9-HFD-C and B6-HFD-C compared with the B6-chow mice (). Interestingly, increased levels of TC had no further effects on islet size in B6-PCSK9-HFD-C mice compared with B6-HFD-C, yet we did observe decreased islet numbers in PCSK9-HFD-C mice (). These data suggest that high cholesterol alone has no additional effects on islet size and affects the overall number of islets.

Increased islet size in B6-HFD-C and B6-PCSK9-HFD-C mice. (A) Representative islet images from whole pancreas sections stained for insulin. Scale bar = 20µm. (B) Average islet area and (C) islet numbers per section. One-way ANOVA with Sidak’s multiple comparisons test was used for statistical analysis. *P < 0.05, **P < 0.01. n = 7–8 mice/group.

Adipose tissues inflammation is a key contributor to peripheral insulin resistance and emigrated peripheral leucocytes are involved in inflammation within AT.[9] Therefore, we examined epididymal AT for the presence of infiltrated leucocytes and particularly monocytes/macrophages. High-fat diet with added cholesterol feeding elevated the epididymal AT mass () and total number of CD45+ leucocytes in B6-HFD-C mice in comparison with the B6-Chow controls (), while the content of CD45+ leucocytes was not significantly elevated in B6-PCSK9-HFD-C mice likely due to normalization of number of CD45+ cells to the significantly elevated AT mass. In line with demonstrated insulin resistance and glucose intolerance in HFD-C-fed mice, we found an increased number of pro-inflammatory Ly6Chigh/med monocytes/macrophage within the AT of B6-PCSK9-HFD-C and B6-HFD-C mice compared with B6-Chow mice ().

Increased numbers of pro-inflammatory monocytes/macrophages in AT of B6-HFD-C and PCSK9-B6-HFD-C mice. (A) Representative FACS plots of CD45+ leucocytes and CD45+ gated Ly6C + cells from digested adipose tissue. (B) Epididymal adipose tissue weight. (C–D) Total CD45+ and CD45+ Ly6C+ cells normalized to adipose tissues weight. One-way ANOVA with Sidak’s multiple comparisons test was used for statistical analysis. *P < 0.05, ***P < 0.001, ****P < 0.0001. n = 7–8 mice/group/

Discussion

Rodent models of diet-induced obesity are widely used to simulate metabolic syndrome, yet limited models provide the paired conditions of atherosclerosis and metabolic syndrome.[1] Besides the HFD-C fed Ldlr mice, there are only a few mouse models such as leptin- or leptin receptor-deficient mice on atherogenic background that represent a model for both T2D and atherosclerosis. Here, we show that HFD-C feeding induces a clear insulin resistance and glucose intolerance in atherosclerotic AAV-PCSK9-injected C57BL/6 mice. We further demonstrate that B6-PCSK9-HFD-C mice successfully mimicked the metabolic phenotype of HFD-C fed Ldlr mice despite having lower plasma cholesterol in comparison to HFD-C fed Ldlr mice. Interestingly, elevated TC in PCSK9-B6-HFD-C mice had little effect on metabolic functions compared with HFD-C-B6 mice. Collectively, our results suggest an existence of saturation levels for plasma cholesterol or the need for other lipid pathways that drive metabolic dysfunction in the conditions of HFD-C diet. Alternatively, liver-targeted modulation of PCSK9 has a limited negative impact on pancreatic islets as the local PCSK9 expression but not circulating PCSK9 regulates levels of LDLR and cholesterol content within the islets.[10] In summary, combination of HFD-C diet with AAV-PCSK9 can be used to model concomitant insulin resistance and glucose intolerance under atherosclerotic conditions.