Novel Reversible Model of Atherosclerosis and Regression Using Oligonucleotide Regulation of the LDL Receptor

Abstract

Rationale

Animal models have been used to explore factors that regulate atherosclerosis. More recently, they have been used to study the factors that promote loss of macrophages and reduction in lesion size after lowering of plasma cholesterol levels. However, current animal models of atherosclerosis regression require challenging surgeries, time-consuming breeding strategies, and/or methods that block liver lipoprotein secretion.

Objective

We sought to develop a more direct and time-effective method to create and then reverse hypercholesterolemia as well as atherosclerosis via transient knockdown of the hepatic LDL receptor (LDLR) followed by its rapid restoration.

Methods and Results

We used antisense oligonucleotides directed to LDLR mRNA to create hypercholesterolemia in wild type C57BL/6 mice fed an atherogenic diet. This led to the development of lesions in the aortic root, aortic arch, and brachiocephalic artery. Use of a sense oligonucleotide replicating the targeted sequence region of the LDLR mRNA rapidly reduced circulating cholesterol levels due to recovery of hepatic LDLR expression. This led to a decrease in macrophages within the aortic root plaques and brachiocephalic artery, i.e. regression of inflammatory cell content, after a period of 2–3 weeks.

Conclusion

We have developed an inducible and reversible hepatic LDLR knockdown mouse model of atherosclerosis regression. While cholesterol reduction decreased early en-face lesions in the aortic arches, macrophage area was reduced in both early and late lesions within the aortic sinus after reversal of hypercholesterolemia. Our model circumvents many of the challenges associated with current mouse models of regression. The use of this technology will potentially expedite studies of atherosclerosis and regression without use of mice with genetic defects in lipid metabolism.

INTRODUCTION

With the development of mice with genetic loss of apolipoprotein (apo) E or the LDL receptor (LDLR)^1,2, investigators have been able to create atherosclerosis in mice, a species that normally has relatively low circulating levels of apoB-containing lipoproteins. These models proved the LDL hypothesis, as a single mutation changing lipid metabolism led to atherosclerotic disease in previously disease-free animals. ApoE-deficient and LDLR-deficient mice have become the standard models for investigating how lipoprotein alterations, diet, and genetic changes in white blood cells and other arterial cells modulate atherosclerosis.

Although vascular biologists have primarily studied atherogenesis, increasingly, both clinical medicine and investigative models now focus on changes in lesion size and loss of inflammatory macrophages that occur with cholesterol lowering. The conversion of inflamed atherosclerotic lesions into a more benign fibrotic scar-like histology is often referred to as regression. In humans, intravascular ultrasound studies found that ~2/3 of patients have some lesion regression with high-dose statin therapy³, and more recently, with use of PCSK9 antibodies⁴. Additionally, simvastatin treatment reduced plaque inflammation in humans, measured by ¹⁸F-fluorodeoxyglucose (¹⁸F-FDG) accumulation which corresponds to the macrophage-rich area of the plaque⁵. The biology of this remodeling has also been studied in mice, and investigators have delineated many of the cellular requirements for regression⁶. For example, we and others have reproduced the defects in atherosclerosis regression found with diabetes^{7, 8} and have shown the importance of changes in macrophage number and gene expression with cholesterol reduction.

Atherosclerosis regression requires reduced arterial exposure to high circulating lipoprotein levels. Marked reductions in circulating apoB-cholesterol have been achieved using the following methods: 1) arteries have been transplanted from hypercholesterolemic Apoe^−/− or Ldlr^−/− mice to normolipidemic wild type (WT) C57BL/6 mice^9–12; 2) Apoe, Ldlr, or very low-density lipoprotein receptor (Vldlr) have been reintroduced using adenoviruses^13–18; 3) microsomal triglyceride transfer protein (MTP) has been deleted in the “Reversa” mouse or pharmacologically inhibited^19–21; or 4) apoB synthesis has been blocked using antisense oligonucleotides (ASOs)²². These models, reviewed in detail elsewhere ^{23, 24}, have certain limitations. Arterial transplantation is technically demanding, while reduction of hepatic production of apoB-containing lipoproteins results in fatty liver and altered triglyceride-rich lipoprotein metabolism. Finally, the need to use Apoe^−/− or Ldlr^−/− mice to induce lesions is costly and time-consuming because genes of interest have to be crossed onto these mouse models of atherosclerosis. Recently, adeno-associated virus (AAV)-mediated overexpression of proprotein convertase subtilisin kexin-like 9 (PCSK9) or inducible degrader of the LDL receptor (IDOL) have been used to induce atherosclerosis^25–28. However, reversal of hypercholesterolemia requires MTP inhibition or very long periods of chow diet (CD) feeding²⁹.

Here, we describe the use of an antisense oligonucleotide (ASO) targeted to LDLR mRNA to create hypercholesterolemia in WT C57BL/6 mice. As expected, this resulted in atherosclerosis development in the aortic arch, aortic sinus, and brachiocephalic artery (BCA). We then describe methods to reduce cholesterol by gradual decay of LDLR ASO after its withdrawal or by more rapid reversal with injection of LDLR sense oligonucleotides (SOs). This novel methodology is expected to markedly facilitate mechanistic studies on atherosclerosis regression.

METHODS

The authors declare that all supporting data are available within the article [and its online supplementary files].

Mice and diets

All procedures were approved by the Institutional Animal Care and Use Committees at New York University Langone Health and University Clinic Freiburg, Germany. 12–14 week old, male and female C57BL/6J mice were maintained in a temperature controlled (25°C) facility with a 12-h light/dark cycle. Male mice were used for a majority of experiments to avoid introducing variability due to hormonal changes in female mice. Mice were given free access to water and food, except when fasting blood specimens were obtained. Mice were either fed a rodent CD or an atherogenic Western diet (WD) (Dyets Inc. catalog no. 101977, 0.3% cholesterol), as indicated. For the data shown in Online Figure IV, modified Western Diet (Research Diet D12108, 1.25% cholesterol) was used.

Oligonucleotide treatments

GalNAc-conjugated Gen 2.5 ASO targeting mouse LDLR was developed and provided by Ionis Pharmaceuticals. Previous reports have demonstrated up to 30-fold improvements in liver ASO potency with GalNAc conjugation together with Gen 2.5 (cET) ASO modifications ^{30, 31}. LDLR ASO was injected intraperitoneally at a dose of 5 mg/kg body weight or 2.5 mg/kg body weight (Online Figure IV) once a week for nine or 16 weeks, as indicated. GalNAc-conjugated SO designed to bind and inactivate the LDLR ASO was injected once intraperitoneally at a dose of 20 mg/kg body weight.

Blood sampling

All blood samples were collected after 4 hours of fasting. Blood was collected from the retro-orbital plexus of mice using heparinized micro capillary tubes. Blood was centrifuged at 10,000 g for 10 minutes for cell removal and collection of the plasma, which was then used for lipid measurements and/or frozen at −80 °C.

Tissue collection

Mice were deeply anesthetized with xylazine (10 mg/kg) and ketamine (100 mg/kg) and then perfused by heart puncture with 10 ml of phosphate buffered saline (PBS) or until the livers blanched. Tissues were rapidly excised and snap frozen in liquid nitrogen unless otherwise noted.

Lipid measurements

Total cholesterol (TC) was measured using Infinity Total Cholesterol Reagent (#TR13521, Thermo Scientific, Waltham, MA).

Lipoprotein fractionation

Equal amounts of mouse plasma (70–100 μl) were used for sequential density ultracentrifugation to separate VLDL (d < 1.006 g/ml), LDL (d 1.006–1.063 g/ml), and HDL (d 1.063–1.21 g/ml) in a TLA 100 rotor (Beckmann Instruments, Palo Alto, CA). Fractions were used to measure TC, as described above.

Western blot analysis

Snap-frozen liver samples were homogenized using a RIPA buffer with protease inhibitors. Protein concentrations were determined using the Pierce BCA Protein kit (#23227 Thermo Scientific, Rockford, IL). 40 μg samples of total protein were subjected to Western blot analysis with a mouse LDLR antibody (R&D AF2255). GAPDH antibody (# 9484 Abcam) was used for control of protein loading.

Tissue gene expression

Total RNA was prepared using a GeneJET RNA Purification Kit (Thermo Scientific). 1 μg of RNA was used for reverse transcription using the Verso cDNA synthesis kit (Thermo Scientific). Real-time quantitative PCR was performed using an ABI 7700 (Applied Biosystems). Amplification was performed using SYBR Green PCR Master Mix (Applied Biosystems). Primers used for PCR amplification were obtained from Primer Bank. Analysis was performed using Sequence Detection Software (Applied Biosciences). Data were normalized to Rn18s.

Atherosclerotic lesion analysis

Each anesthetized mouse was perfused with PBS. The aorta was then exposed and fat was carefully cleaned under a binocular microscope. Pictures of the aortic arch and BCA were taken using a camera fitted to the binocular microscope. The BCA was collected in 10% formalin, kept overnight at 4°C, and then stored in 70% ethanol at 4°C for further processing. The root of the heart was cut and embedded in TissueTek Optimal Cutting Temperature (OCT), frozen, and stored at −80°C.

Histological and morphometric analysis of aortic roots

Serial sections (6 μm) were obtained by cryosectioning the frozen aortic roots that were embedded in OCT. The sections were then stained with Picrosirius Red and analyzed for lesion size using brightfield microscopy. Collagen content was measured under polarized light. Mac-2 or CD68 immunostaining was used to determine macrophage content in the aortic root sections, as described previously^32–34.

Staining and analysis of BCAs

BCAs embedded in paraffin were sectioned (5 μm). Every 5th cross section was stained using the Movat’s pentachrome method to visualize atherosclerotic lesions. The maximal lesion site was determined, and adjacent sections were immunostained using a Mac-2 antibody to detect lesion macrophages, as described previously³⁵.

Morphometric measurements

Morphometric measurements were performed on digitized images of stained serial sections by using IMAGEPRO PLUS software. At least 12 BCA sections per mouse were analyzed (60 sections/mouse with every 5^th section used for Movat’s stain and morphometry), and the maximal lesion area value for each mouse was used as the summary parameter. At least five sections per root were analyzed and the mean value used as the summary parameter.

Statistical analysis

Data are presented as means ± SD. Statistical differences were assessed via unpaired Student’s t-test or a one-way or two-way ANOVA using GraphPad Prism 7 where appropriate. A p value of <0.05 was considered as significant.

RESULTS

Effect of LDLR ASO on plasma cholesterol in WT mice

Male C57BL/6J mice were treated with intraperitoneal injections of LDLR ASO or control ASO once a week for 9 weeks. The mice were fed an atherogenic WD throughout the study (Fig. 1A). Plasma TC levels increased within a week of ASO treatment and stayed high (~400–800mg/dl) throughout the 9 weeks (Fig. 1B) compared to control ASO-treated mice. LDLR ASO treatment led to hypercholesterolemia in both male (Fig. 1B) and female mice (Online Figure I). Hepatic LDLR knockdown was confirmed (Fig. 1D) at the end of the study. One group of mice, denoted ‘baseline’, was sacrificed after 9 weeks of LDLR ASO treatment. The rest of the mice were divided into two groups. The first group, denoted ‘SO’, was treated with SOs antagonizing the residual LDLR ASO, while the second group, denoted ‘None’, received no further intervention. Mice in both the groups were kept on WD to monitor changes in plasma TC as a function of hepatic LDLR recovery. Plasma TC dropped substantially in both SO and None groups by 2 weeks (week 11); ~ 135 mg/dl in SO group and ~430 mg/dl in None group. After another week (week 12), both SO and None groups had similar TC levels (~200 mg/dl) (Fig. 1C). LDLR expression recovered in both SO and None groups, with higher levels in the SO group, as shown by western blot (Fig. 1D) and ELISA (Online Figure II).

Figure 1. Model of atherosclerosis regression using oligonucleotide-mediated hepatic LDLR modulation.

(A) Experimental outline; (B) time course of plasma TC in 12 week old male mice treated with Control ASO and LDLR ASO. All mice were fed with Western diet (WD) throughout, n=18; (C) plasma TC levels after 9 weeks of ASO treatment and at week 11 (2 weeks after SO) and week 12 (3 weeks after SO), n=4–5/group; (D) western blot of hepatic LDLR at the end of the study. Controls for the western blot are: positive control- WT (C57BL/6J) mice on chow diet and negative control- KO (Ldlr^−/−) mouse on chow diet; (E) representative images of aortic arch lesions in the three groups, n=4–6/group; (F) quantification of aortic root lesion area after staining with Picrosirius Red; (G) quantification of Oil Red O⁺ lesion area and (H) quantification showing percentage of macrophage content in aortic root lesions, visualized by Mac-2 immunofluorescence in baseline mice and two regression groups of SO and None n=4–5/group, *p<0.05, **p<0.01, ***p<0.001, data presented as means ± SD.

Atherosclerosis in LDLR ASO treated WT mice

Direct visualization of aortae in the baseline group of mice showed fatty-streak type early lesions in the aortic arch and adjacent BCA (Fig. 1E, baseline, n=4–5). Remarkably, mice in both SO and None groups showed drastic declines in en-face visible lesions in the aortic arch and adjacent BCA (Fig. 1E, SO and None, n=4–5/group). Next, lesions in the aortic roots were analyzed. While total plaque areas were not different among the groups (Fig. 1F), Oil Red-O stained lesion areas were significantly reduced with SO treatment and showed a trend towards a decrease in None, compared to baseline (Fig. 1G). Importantly, there was an extensive reduction of macrophage content within the aortic root lesions of both SO and None groups (Fig. 1H).

These initial studies confirm that LDLR ASO-driven cholesterol increases (~800mg/dl) lead to atherosclerosis, which can then be reduced by lowering cholesterol levels (~200mg/dl).

Effect of LDLR SO and diet on complex atherosclerotic lesions

In our previous studies, we used longer hypercholesterolemic exposure to create advanced lesions relevant to human disease ^{17, 19, 36}. This allowed us to better assess the mechanisms responsible for monocyte recruitment and macrophage accumulation, and to quantify reduced regression with diabetes^{7, 8, 17}. For the current experiment, LDLR ASO treatment was extended to 16 weeks prior to initiating cholesterol reduction. The effect of CD versus WD on lesion regression was also assessed (Fig. 2A). As before, plasma TC increased to ~500 mg/dl within 2 weeks of LDLR ASO treatment and was maintained at high levels (~800 mg/dl) until 16 weeks (Fig. 2B). One group of mice was sacrificed for baseline analysis at the 16-week time point when LDLR ASO treatments were stopped. From our initial study (Fig. 1), we learned that including LDLR SO in our regression model allowed us to shorten the regression time. In this long-term study, cholesterol reduction was therefore induced by SO in all the mice, as we predicted that SO treatment would restore hepatic LDLR levels and lower plasma cholesterol faster than no treatment. TC levels were significantly lowered in both the CD- and WD-fed groups by one week of SO treatment; these levels remained low until the end of the study at week 18 (Fig. 2C). Plasma lipoprotein fractionation showed that decreased plasma TC after SO treatment was mainly due to a reduction in LDL cholesterol (Fig. 2D). The VLDL cholesterol fraction was also significantly reduced in both regression groups, whereas HDL cholesterol was unaffected (Fig. 2D). In contrast to the previous short-term study, no obvious changes in en-face aortic arch or BCA plaque density were found between the three groups, upon direct visualization (Fig. 2E).

Figure 2. Regression of advanced atherosclerotic lesions.

(A) Experimental outline; (B) time course of plasma TC in 12–14 week old male mice treated with LDLR ASO and fed with WD; (C) plasma TC levels before SO treatment (week 16), after 1 week (week 17) and after 2 weeks of SO (week 18); one group of mice were kept on WD and another was switched to CD after SO treatment (D) cholesterol content in VLDL, LDL, and HDL fractions of plasma in baseline group and the two regression groups of chow diet (CD) and western diet (WD) at the end of the study; (E) representative images of aortic arch lesions in the three groups. n = 15–16/group ***p<0.001, data presented as means ± SD.

The continued opacity of the aorta despite cholesterol reduction has been shown in a prior study¹⁷, and is most likely explained by the conversion of the plaques to fibrous lesions. To test this, we assessed the histology of atherosclerotic lesions in the aortic sinus. Picrosirius Red brightfield microscopy showed that total lesion area was not significantly altered in the baseline and the two regression groups (Fig. 3A, 3B). Lesion collagen content, an indicator of vascular remodeling towards stable lesions, tended to increase in the regression groups (Online Figure III), but this did not reach statistical significance. However, macrophage content of lesions, as measured by Mac-2 immunofluorescence, was significantly reduced in the lesions of both the regression groups compared to the baseline (Fig 3C, 3D).

Figure 3. Assessment of atherosclerosis lesions in aortic roots of mice at baseline and after 2 weeks of cholesterol reduction.

(A) Morphology of aortic root lesions in baseline mice and two regression groups of CD and WD visualized by Picrosirius Red brightfield microscopy and its (B) quantification; (C) macrophage content of aortic root lesions visualized by Mac-2 immunofluorescence and its (D) quantification showing Mac-2 positive lesion area (%). Representative images are shown. * p < 0.05, ** p<0.01. n=15–16/group, data presented as means ± SD.

Lesions within the BCA were also characterized, as this site develops complex lesions more representative of certain human plaque characteristics. Similar to the observations in aortic root lesions, there were significantly fewer macrophages within BCA lesions of both CD and WD groups, compared to baseline (Fig. 4C, 4D). Total lesion area, visualized by Movat’s staining, was unchanged among all three groups (Fig 4A, 4B).

Figure 4. Assessment of atherosclerosis lesions in BCA at baseline and after 2 weeks of cholesterol reduction.

(A) Lesions in the BCA visualized by Movat’s stain, (B) quantification of total lesion area; (C) Mac-2 positive area showing lesion macrophages and (D) its quantification, in the baseline group mice and the two regression groups of CD and WD. Representative pictures of Movat’s staining and Mac-2 staining are shown. * p < 0.05, ** p<0.01. n=15–16/group, data presented as means ± SD.

Effect of a lower dose of LDLR ASO on atherosclerosis

Finally, to determine if the current protocol could be altered to allow use of less LDLR ASO, C57BL/6J WT mice were treated with half the dose of LDLR ASO (2.5 mg/kg) for 16 weeks (Online Figure IV A). This dose was sufficient to raise plasma TC to about 500 mg/dl when the mice were fed a modified WD. When these mice were treated with SO and switched to CD, plasma cholesterol dropped below 200 mg/dl (Online Figure IV B). Aortic root lesion analysis showed no changes in total plaque size and lipid content between baseline and SO groups (Online Figure IV C, D, and E). However, macrophage content of lesions was reduced in the SO treated mice compared to baseline (Online Figure IV F), similar to what was seen in our previous experiments.

DISCUSSION

The development of atherosclerosis-prone mice nearly 25 years ago led to major advances in our understanding of the roles of circulating lipoproteins and inflammatory factors in the creation of arterial lesions. More recently, a number of investigators have focused on the regression of lesions in an effort to understand the biology that occurs when patients with atherosclerosis are treated with potent cholesterol-reducing medications. Clinical studies have shown that this intervention reduces the incidence of cardiovascular events ³⁷, even though the changes in atherosclerosis burden found using intravascular ultrasound is less than 2% ⁴. This disconnect between events and lesion size has been explained by studies in mice that have shown that cholesterol reduction causes a more dramatic reduction in plaque macrophages and increase in collagen content than overall lesion size ⁶. Thus, what others and we have observed is more properly the regression of the inflammatory state of the plaque, which in humans would be expected to decrease plaque vulnerability and increase stability. Defining the many cellular and circulating factors that mediate vascular remodeling, and determining how they are defective in diseases such as diabetes, is an active area of investigation.

In this report, we describe methods to both create atherosclerosis and then assess the biology of regression after cholesterol reduction. Using the ASO approach, hepatic LDLR expression was reduced followed by its reversal with introduction of LDLR SOs. This protocol allowed for the creation of atherosclerosis and its resolution, without the time-intensive and costly process of crossing mice onto a genetically hypercholesterolemic background. Early fatty streak lesions developed after 9 weeks, whereas more advanced lesions were observed in mice after 16 weeks. Regression was induced without the need for transplantation or additional genetic or pharmacologic treatments to reduce cholesterol. Although cessation of use of the LDLR ASO led to reduced cholesterol, the cholesterol reduction was more rapid using the SOs. Moreover, using the SOs we could even achieve cholesterol lowering and regression of lesional macrophages while the mice continued to eat a high-cholesterol atherogenic diet (WD). We analyzed atherosclerosis lesions in two different vascular beds, namely aortic root and BCA, both of which showed decreased macrophage content within lesions after cholesterol reduction.

We did not compare the effects of the LDLR ASO to the atherosclerosis that occurs in LDLR knockout (Ldlr^−/−) mice, though grossly, the characteristics of the lesions appeared similar to our historical experience with these mice. We and others have noted the marked variability in cholesterol levels of Ldlr^−/− mice as a function of strain, diet, and perhaps even the microbiome ³⁸. We found similar variability in plasma cholesterol and lesions between LDLR ASO and Ldlr^−/− mice³⁹. As has been found in humans ⁴⁰, it is possible that genetic deficiency of LDLRs might lead to increased atherosclerosis over and above that which is expected from adult levels of LDL cholesterol levels. This might occur either because even on chow the LDLR deficiency is associated with small increases in cholesterol above normal, or because the LDLR mutation leads to additional vascular toxicity. Also, reports have shown a significant contribution of macrophage LDLR towards lesion development, so our model described in this study would be expected to have less atherosclerosis compared to Ldlr^−/− mice ^{41, 42}. At the doses used (5 mg/kg and 2.5 mg/kg), we would not have expected to see much activity of LDLR ASO outside the liver. Use of two different doses of LDLR ASO pointed to dose-dependent effect on cholesterol and possibly lesion size as well.

In summary, we describe a new method to create both atherosclerosis and regression in mice that do not have genetic alterations in lipoprotein metabolism. This method is readily transferable to use for assessment of atherosclerotic plaques and their repair in animals, and will allow investigators to more conveniently test whether changes in leukocytes, smooth muscle cells, endothelial cells, or non-arterial cells affect atherogenesis.

NOVELTY AND SIGNIFICANCE.

What Is Known?

• Atherosclerosis regression is defined as the reduction of macrophage and lipid content of plaques when plasma LDL-cholesterol (LDL-c) is markedly decreased using drugs like statins.

• Studies of regression entails technically difficult aortic transplants or time-consuming breeding strategies.

• LDL receptor (LDLR) is the major receptor responsible for the removal of LDL-c from circulation..

What New Information Does This Article Contribute?

• A new method to create atherosclerosis and its regression in C57Bl/6 mice is described, where hepatic LDLRs were knocked down with ASOs, leading to plaques in the aortic root and brachiocephalic artery (BCA).

• The expression of the receptors was restored by sense oligonucleotide (SO) treatment, which led to reversal of hypercholesterolemia and fewer plaque macrophages in the aortic root and BCA.

• This model of regression can be used with any genetically modified mouse strain and obviates the need for extensive breeding.

Several genetic mouse models and a surgical model have been used to study the pathophysiology of atherosclerosis regression in mice. However, these involve time-consuming breeding strategies or technically demanding surgeries. To overcome these challenges, we have created hypercholesterolemia and atherosclerosis in mice by transient knockdown of hepatic LDLR using an ASO. Early fatty streak lesions developed after 9 weeks, whereas more advanced lesions were observed in mice after 16 weeks. Subsequently, plasma LDL-c was reduced using SO to block the residual ASO and restore hepatic LDLR availability. Regression of lesional macrophages was observed in two different vascular beds, namely aortic root and BCA. This novel in vivo strategy is readily transferable and should expedite studies of atherosclerosis and regression.

Non-standard Abbreviations and Acronyms

ASO

antisense oligonucleotides

SO

sense oligonucleotides

BCA

brachiocephalic artery