Hyperlipidemia induces hippocampal inflammation and loss of vascularity and can be rescued by silencing RIPK1

Abstract

Vascular cognitive impairment (VCI) is a common form of dementia associated with cerebrovascular dysfunction and chronic inflammation. Impaired cerebral perfusion is a result of atherosclerotic plaque build-up in the medium and large arteries in the brain and is exacerbated by numerous inflammatory triggers, including aging, high blood pressure, high cholesterol, and smoking. Previously, we demonstrated that receptor-interacting protein 1 kinase (RIPK1) promotes the progression of aortic atherosclerosis and NFκB activation in mice, and that atherosclerotic vascular disease was decreased by therapeutically inhibiting RIPK1 with antisense oligonucleotides. Given the relationship between atherosclerosis and cognitive impairment, we hypothesize that RIPK1 also contributes to the development of cerebrovascular disease and neuroinflammation in a mouse model of hyperlipidemia. Male and female Apoe^−/− were fed a chow or Western diet (WD) for 16 weeks and subsequently treated for an additional 8 weeks with either scramble (control) or anti-RIPK1 antisense oligonucleotides (ASO) to systemically knockdown RIPK1 expression (RIPK1^KD). WD feeding induced significant carotid atherosclerotic lesion burden and reduced cerebral blood flow compared to chow-fed mice. These vascular changes were associated with impaired spatial learning and memory, reduced hippocampal vascularity, and altered expression of neurovascular and inflammatory markers. Systemic RIPK1 silencing significantly reduced carotid artery lesion size, restored arterial stiffness and blood flow in the brain. Furthermore, treatment with RIPK1^KD preserved vascularity in the hippocampus, restored blood flow in the brain and prevented the impairment in spatial learning and memory. Overall, our study identifies RIPK1 as a mediator of hyperlipidemia-induced cerebrovascular dysfunction and suggests that targeting RIPK1 may represent a therapeutic approach to limit cognitive impairment.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-54533-w.

Introduction

The prevalence of dementia, neurodegenerative diseases, and cognitive impairment has risen alongside the increasing age of the global population¹. Alzheimer’s disease (AD) is the leading cause of dementia worldwide, accounting for 60–80% of cases, while vascular dementia (VaD) accounts for up to 20% of cases^2,3. Before the onset of overt dementia, individuals may experience cognitive deficits such as memory loss, learning difficulties, reduced focus, and impaired ability to perform routine tasks, which are all prominent features of working memory impairment³. Although these deficiencies are typically modest and not clinically noticeable, they progressively worsen until dementia is evident^4,5. Nearly 55 million people worldwide are affected by dementia, and this is estimated to reach 139 million by 2050⁶. There are currently no direct treatments that halt or reverse VaD, and symptom management and primary prevention remain the focus of treatment.

Atherosclerosis of the cerebral arteries, also known as intracranial atherosclerotic disease (ICAD), is a major pathological contributor to the development of cerebrovascular disease (CBVD) and VaD^7,8. ICAD is characterized by atherosclerotic plaque formation within the middle cerebral, basilar, intracranial vertebral, and intracranial carotid arteries, and represents a leading cause of dementia worldwide^9–11. The development of ICAD is initiated by endothelial dysfunction driven by hypercholesterolemia, oxidative stress, and chronic inflammation, resulting in increased endothelial permeability, lipoprotein retention, and immune cell recruitment within the arterial wall¹². As atherosclerotic lesions progress in intracranial arteries, luminal narrowing, altered shear stress, and arterial stiffening impair cerebral blood flow (CBF), leading to chronic hypoperfusion and heightened susceptibility to ischemic injury^13,14. These¹³ ICAD-associated vascular abnormalities compromise downstream microvascular function, disrupt neurovascular coupling¹⁵ and blood–brain barrier integrity¹⁶, and promote neuronal dysfunction¹⁷, thereby contributing to progressive cognitive impairment and increasing the risk of VaD^5,18. Despite the critical role of ICAD in stroke and vascular dementia^19,20, current therapeutic methods are still mostly focused on altering systemic risk factors, and there are no disease-modifying medications that target cerebral atherosclerotic pathology.

RIPK1 is a critical mediator of inflammatory and cell death pathways²¹. Via its scaffolding function, RIPK1 can act as an activator of the NFκB pathway, or, via its kinase activity can regulate apoptotic and necroptotic cell death. In brain ischemia models, treatment with Necrostatin-1, a broad and non-specific inhibitor of RIPK1, reduces neuroinflammation^22,23. RIPK1 activation is evident in brain samples from humans with AD²³, and inhibiting RIPK1 kinase activity in mouse models of AD reduces inflammatory microglia and cognitive deficits^24,25. Our team previously showed that the expression of receptor-interacting protein 1 kinase (RIPK1) in human atherosclerotic arteries was highly associated with the expression of NFκB and proinflammatory pathways²⁶. In a mouse model of atherosclerosis, we demonstrated that silencing the RIPK1 gene decreased the progression of atherosclerotic lesions, which was accompanied by reductions in NFkB activation in both macrophages and vascular endothelial cells²⁷. This led to the hypothesis that blocking RIPK1 in a model of atherosclerosis-induced VaD may similarly reduce lesion burden and reduce CBVD-associated cognitive decline.

Here, we describe that feeding Apoe^−/− mice a Western diet leads to reduced cerebral blood flow, loss of vascularity in the hippocampus, increased expression of inflammatory markers, and a deficit in learning and memory. We find that subcutaneous administration of antisense oligonucleotides (ASO) targeting RIPK1 in mice with established atherosclerosis reduces carotid lesion burden, restores blood flow and vascularity in the hippocampus, and rescues cognitive deficits. These data suggest that therapeutic mechanisms aimed at reducing RIPK1 expression are beneficial in atherosclerosis-associated vascular dementia.

Results

Systemic silencing of RIPK1 reduces carotid atherosclerosis and vascular stiffness in Western-diet fed Apoe^−/− mice

We have previously demonstrated that knockdown of RIPK1 using antisense oligonucleotides (ASOs, RIPK1^KD) during early-stage atherosclerosis progression led to reduced atherosclerotic lesions in the aortic root and aorta²⁶. As atherosclerotic remodelling of the common carotid artery is closely linked to cerebrovascular dysfunction underlying vascular cognitive impairment, we sought to evaluate whether treatment with RIPK1^KD would similarly reduce carotid lesion area at later stages of lesion development. To this end, we fed female and male Apoe^−/− mice a Western-type diet, containing high -fat, high -sucrose and high cholesterol (Western diet, WD) for 16 weeks to induce atherosclerosis, after which we randomized WD-fed mice to receive a subcutaneous administration of control ASO (WD-scramble ASO: Scr, corresponding to a scrambled non-targeting sequence) or ASOs targeting RIPK1 to knock down RIPK1 expression (WD-RIPK1^KD, Fig. 1A). We have previously confirmed that, compared to non-treated controls, scrambled oligonucleotides do not affect atherosclerosis development, weight gain, or any other parameters evaluated, compared to non-treated mice²⁶. As expected, mice fed the Western diet gained more weight than chow-fed mice (Fig. 1B & Supplemental Fig. 1A, male and females). After treatment with ASOs and continued WD feeding for a subsequent 8 weeks, we observed reduced weight gain in RIPK1^KD-treated animals compared to the scramble counterparts, in both females and males (Fig. 1B & Supplemental Fig. 1B). All WD-fed animals showed a significant increase in total cholesterol during the course of the study (Fig. 1C) compared to chow-fed animals. To confirm RIPK1 knockdown, we performed Western blotting (WB) and immunofluorescent (IF) analysis of liver samples and observed a significantly lower expression of RIPK1 in animals treated with RIPK1^KD compared to scramble in both sexes (Fig. 1D & Supplemental Fig. 1C-D). Effective hepatic knockdown of RIPK1 confirms target engagement of the antisense strategy. As Apoe-/- mice develop carotid atherosclerosis upon Western diet feeding^28–30, we next evaluated atherosclerotic lesion burden in the common carotid arteries of female and male mice. After 8 weeks of treatment, there was a significant decrease in lesion burden in RIPK1^KD mice compared to scramble controls (82.5% reduction in females; 37.8% reduction in males, Fig. 1E-F), despite ongoing hypercholesterolemia. Minimal lesions were detectable in the chow-fed mice, as expected. Immunofluorescence analysis revealed an accumulation of macrophages in carotid lesions in WD-scramble treated, which was largely abrogated in RIPK1^KD lesions (Fig. 1G & Supplemental Fig. 1E). We then assessed the common carotid artery lumen diameter³⁰ by M-mode and Doppler ultrasound³¹ and found a significant decrease in lumen diameter in both the left and right common carotid arteries (LCCA and RCCA) in WD-fed mice compared to chow-fed control during both diastole and systole (Fig. 1H-I, Supplemental Fig. 1F-H). The lumen diameter of both LCCA and RCCA of RIPK1^KD-treated animals was significantly increased compared to Scr-treated mice, and similar to that of chow-fed mice in both females and males. We then used Doppler ultrasound to determine Pulse Wave Velocity (PWV)²⁸ to assess arterial stiffness in chow, WD-scramble and WD-RIPK1^KD mice. Compared to chow-fed mice, WD-scramble mice demonstrated a significantly higher PWV in both sexes, which was decreased significantly in WD-RIPK1^KD mice (Fig. 1J). Similar to our previous reports^26,27, mice fed a WD-scramble treated showed an increase in systemic pro-inflammatory markers such as IL-1β and IL-6, compared to those treated with WD-RIPK1^KD (Fig. 1K). This implies that systemically suppressing RIPK1 significantly decreases circulating pro-inflammatory cytokines despite continued hyperlipidemia. Together, these findings demonstrate that, as expected, a high cholesterol diet induces plaque buildup in the carotid arteries, increasing pressure and vascular stiffness during the progression of atherosclerosis. However, even in the setting of a continued high-cholesterol diet, intervention with RIPK1^KD reduces carotid lesion size and improves vascular function possibly by decreasing systemic inflammation.

Fig. 1.

Intervention with RIPK1^KD in mice with established atherosclerosis reduces carotid lesion burden. (A) Outline of the experimental strategy. (B) Weight gain at 24 weeks as a percentage of baseline (16 weeks). (C) Cholesterol levels at 24 weeks. (D) WB of liver lysates for RIPK1 and GAPDH. (E) Cross-sections H&E staining of LCC artery. (F) Plaque area in LCC arteries. (G) IF staining of macrophages (CD68, green) and endothelial cells (CD31, red) in carotids arteries. (H) LCC diastolic and (I) systolic lumen diameter (mm²). (J) Pulse wave velocity (PWV, m/s). (K) Elisa quantification of IL-1β and IL-6. Scale bars, 100 μm. F chow = 9-15, F scr = 9-15, F RIPK1^KD = 9-15, M chow = 9-15, M scr = 9-15, M RIPK1^KD = 9-15. All data was tested for normality by D’Agostino-Pearson test. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by Two-Way ANOVA.

Hyperlipidemia impairs spatial learning and memory, which is rescued by silencing RIPK1

We evaluated the behaviour of female and male mice after 24 weeks of chow or WD-feed treated with scramble or RIPK1^KD to investigate the relationship between carotid atherosclerosis and cognition. Using the novel object recognition test (NOR) to assess recognition memory, we found no difference between chow diet, WD-scramble and WD-RIPK1^KD mice (Supplemental Fig. 2A-B). There was no difference between groups when anxiety-like behaviours were assessed using the open field test (OF, Supplemental Fig. 2C-D). Similarly, in the elevated plus maze test (EPM, Supplemental Fig. 2E-F). Next, we evaluated spatial memory and learning using a five-day Morris Water Maze test (MWM, Fig. 2A). Compared to chow-fed mice, both female and male mice WD-scramble had an increased latency to reach the platform over the course of 5 days, indicating impaired spatial learning. In contrast, WD-RIPK1^KD mice showed the same latency to the platform as chow-fed mice, suggesting similarities in spatial learning and memory (Fig. 2B-C). Mice were rested for two days, then probed in the absence of the platform to assess spatial memory. WD-scramble mice spent less time in the target quadrant, and they also showed random swimming compared to chow-fed and the WD-RIPK1^KD mice (Fig. 2D) which spent significantly more time in the target quadrant and showed a more target-oriented swimming pattern (Fig. 2E-F). The groups did not differ in the distance travelled within the pool (data not shown). These data show that Apoe^−/− mice fed a WD show impaired spatial learning and memory, and treatment with RIPK1^KD in conjunction with WD feeding effectively rescues this defect.

Fig. 2.

Intervention with RIPK1^KD rescues atherosclerosis-induced decline in spatial learning & memory. (A) Schematic of Morris Water Maze test (MWM). (B-C) MWM training period, measuring latency to platform days 1 to 5 (B= females, C= males). (D) MWM heatmap. (E-F) Percentage time spend in quadrant on probe day (E= female, F= male). F chow = 13, F scr = 12, F RIKP1^KD = 10, M chow = 15, M scr = 15, M RIPK1^KD = 13. All data was tested for normality by D’Agostino-Pearson test. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by Two-Way ANOVA.

Hyperlipidemia reduces vascularity and increases inflammatory activation in the hippocampus, which can be reversed by RIPK1 silencing

To investigate whether WD might affect cerebral blood flow (CBF), we used functional ultrasound (fUS) imaging, a non-invasive method of real-time cerebral vascularity. We observed that, compared to chow-fed mice, mice fed a WD-scramble had significantly lower cerebral vascularity (Fig. 3A, Supplemental Fig. 3A). Compared to WD-scramble, WD-RIPK1^KD treatment led to a significant improvement in vascularity (Fig. 3A, Supplemental Fig. 3A). This finding suggests that the reduced atherosclerotic plaques in the carotid arteries of RIPK1^KD-treated mice allows for a greater blood flow to the brain. Reduced cerebral perfusion can disrupt cerebral microvessels, potentially leading to cognitive decline. Since the hippocampus is involved in learning and memory processes, which we found to be impaired in WD-scramble mice, we assessed the hippocampal neurovascular unit (NVU) in chow, WD-scrambled and WD-RIPK1^KD mice. We found that endothelial cell length, as measured using immunofluorescence of CD31+, is reduced in both female and male WD-scramble mice compared to chow-fed controls (Fig. 3B-D, Supplemental Fig. 3B). Treatment with RIPK1^KD during WD feeding partially rescued the length of the CD31 + vessels (Fig. 3D, Supplemental Fig. 3C). To investigate the inflammatory status of the hippocampus during hyperlipidemia, we assessed GFAP expression as a marker of astrogliosis and IBA1 expression as a marker of microglia. Interestingly, we found that in female mice the number of GFAP + and IBA1 + cells was very similar between chow and WD-RIPK1^KD-treated mice, whereas WD-scramble were slightly lower (Fig. 3E-F). As for males, we observed that the number of GFAP+ cells was significantly higher in chow and RIPK1^KD compared to scramble (Fig. 3E). However, the number of IBA-1 + cells was significantly lower in both chow and RIPK1^KD compared to scramble (Fig. 3F). To further evaluate NVU and pro-inflammatory marker expression, we used Western blotting on hippocampus lysates. Compared to chow-fed mice, WD-scramble mice had downregulation of Claudin-5 (CLDN5), a critical protein in tight junctions in the brain blood barrier (BBB), and platelet-derived growth factor receptor beta (PDGFRβ), a receptor involved in pericyte recruitment and stability (Fig. 3G, Supplemental Fig. 3D). Similar to this, WD-RIPK1^KD restored CLDN5 and PDGFR expression, which may have stabilized BBB permeability and function. We observed a significantly lower expression of cell death and proinflammatory markers Casp3, IL-1β, NFκB and RIPK1 in chow and WD-RIPK1^KD both sexes compared to WD-scramble mice (Fig. 3H, Supplemental Fig. 3E). To assess vascular permeability, we intravenously administered a fluorescently labeled 10 kDa dextran via retro-orbital injection. Dextran of this size is normally confined to the vascular lumen but extravasates into the brain parenchyma when the BBB is altered and thus serves as tracer to detect changes in vascular permeability^32,33. WD-fed mice treated with scramble ASO exhibited elevated dextran signal within amygdala and the median eminence (data not shown) and in both the dorsal and ventral hippocampus, compared to chow and WD-RIPK1^KD groups (Fig. 3I-J, Supplemental Fig. 3F). Notably, dextran was not only detected in parenchymal spaces, but a significant accumulation of dextran was detected in the larger vessels of WD-fed mice. These findings suggest that chronic hyperlipidemia may promote regional alterations in cerebrovascular integrity, and silencing RIPK1 reduces neurovascular inflammation in the hippocampus possibly via preserving BBB integrity.

Fig. 3.

Western diet reduces vascularity & increases inflammatory activation in the hippocampus, and is rescued with RIPK1^KD. (A) Cerebral blood flow (represented as percentage of vascularity) measured by functional ultrasound (fUS). (B) IF staining of astrocytes (GFAP, green) and endothelial cells (CD31, red) in male hipp. (C) IF staining of microglia (IBA1, green) and endothelial cells (CD31, red) in male hipp. (D) CD31 quantification: from (D) and Supplemental Figure 3 A (female mice). (E) GFAP+ and (F) IBA-1+ cell quantification. (G) WB Claudin-5 and PDGFRb. (H) WB Casp3, IL-1β, NFκB and RIPK1. (I-J) Dextran accumulation in the hippocampus following i.v. delivery of 10 kDa dextran Volume / MFI total amount of dextran Alexa Fluor-555 dye (light grey) in the ventral hippocampus. (J) Assessment of blood-brain barrier (BBB) permeability. Scale bars, 100 μm. F Chow = 8, F Scr = 8, F RIPK1^KD = 8, M Chow = 8, M Scr = 8, M RIPK1^KD = 8. *. All data was tested for normality by D’Agostino-Pearson test. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by Two-Way ANOVA.

Pro-atherogenic stimulation induces inflammation and cell death in brain microvascular endothelial cells

Brain microvascular endothelial cells (BMECs) are a key cellular component of the blood–brain barrier (BBB), which regulates the exchange of molecules between the brain and the peripheral circulation^15,17. To investigate the effects of pro-atherogenic stimuli and RIPK1 signaling in the brain microvasculature, BMECs were isolated from Apoe^−/− mice of similar age to those used in the in vivo experiments. Cells were stimulated in vitro with either TNFα (10 ng/mL) as a pro-inflammatory stimulus or oxidized LDL (oxLDL, 50 µg/mL) as a hyperlipidemic stimulus for 6, 16, and 24 h. BMECs exposed to TNFα or oxLDL displayed morphological changes characterized by increased intercellular gaps, formation of cell islands, and the appearance of star-shaped cellular processes, suggesting cytoskeletal remodeling and reduced intercellular adhesion. Both TNFα and oxLDL significantly increased phosphorylated RIPK1 (pRIPK1) expression at 16 and 24 h (measured by mean fluorescent intensity, MFI; Fig. 4A–B). Similarly, increased expression of Casp3 and TNFα was observed between 16 and 24 h following stimulation (Fig. 4C–D, Supplemental Fig. 4A–B). To model the effects of circulating factors associated with hyperlipidemia, BMECs were exposed for 24 h to serum collected from chow-fed or Western diet–fed Apoe^−/− mice treated with scramble control or RIPK1^KD ASO at the end of the 24-week study. BMECs exposed to serum from WD-scramble mice exhibited increased expression of pRIPK1, Casp3, NFκB, and IL-1β compared with cells treated with serum from chow-fed mice and WD-RIPK1^KD mice (Fig. 4E–J, Supplemental Fig. 4C–D). Therefore, our analyses indicate that hyperlipidemic and proinflammatory stimuli drive activation of RIPK1-dependent inflammatory and apoptotic pathways in BMECs, promoting endothelial stress and potentially compromising BBB integrity. Systemic RIPK1 silencing appears to mitigate these molecular changes, supporting a mechanistic role for RIPK1 in linking systemic atherosclerosis to cerebrovascular endothelial dysfunction (Fig. 5A).

Fig. 4.

Brain microvascular endothelial cells (BMECs) increase inflammatory and cell death markers under pro-inflammatory & pro-atherogenic conditions. (A-D) IF / MFI staining / quantification of pRipk1 and (C-D) Caspase-3 at 6-, 16- and 24-hour treatment with TNFa and oxLDL. (E-F) IF / MFI staining / quantification of pRipk1, (G-H) IF / MFI staining / quantification of Caspase-3. (I) MFI quantification of IL-1β and (J) MFI quantification of NF-κB at 24 h of treatment with serum as indicated in the legend. Scale bars, 50 μm. Chow = 5, F/M WD = 5, F/M anti-Ripk1 = 5/6. All data was tested for normality by D’Agostino-Pearson test. *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001 by Two-Way ANOVA.

Fig. 5.

Graphical abstract. Hyperlipidemia induces carotid atherosclerosis, arterial stiffness and reduces cerebral perfusion in Apoe^-/- mice, which is accompanied by loss of cerebral endothelial vascularity, increased inflammation in the circulation and hippocampus, and impaired learning and memory. Systemic silencing of RIPK1 using anti-sense oligonucleotides reduces atherosclerotic lesions burden in the carotid arteries, improves arterial stiffness and cerebral blood flow, maintains endothelial network function and integrity, and improved learning and memory in the setting of hyperlipidemia.

Discussion

In the present study, we demonstrate that systemic RIPK1 silencing confers protection against hyperlipidemia-induced cerebrovascular dysfunction and cognitive decline in Apoe^−/−−/− mice. We found that in both female and male Apoe^−/− mice, Western diet feeding induced carotid atherosclerosis, reduced cerebral vascularity and perfusion, disrupted neurovascular unit and tight junction protein expression, and impaired cognitive performance. Notably, RIPK1 silencing prevented these vascular and cognitive deficits despite continued Western diet exposure, indicating that RIPK1 is a critical mediator of diet-induced cerebrovascular injury. Moreover, RIPK1 inhibition preserved cerebral vascularity and perfusion as well as maintaining the expression of key neurovascular and tight junction proteins. Together, this work identified RIPK1 as a driver of cerebrovascular disease in mice and suggests that RIPK1 silencing therapies are a promising therapeutic strategy for vascular cognitive impairment.

Systemic hyperlipidemia and inflammation can have profound effects on the cerebrovascular system. In both preclinical mouse models and humans, elevated cholesterol levels are strongly associated with neuroinflammation, enhanced BBB permeability and the development of dementia^19,34,35. In the hippocampus, hypercholesterolemia activates key inflammatory markers and pro-inflammatory cytokines, such as IL-1β and IL-6^36,37 associated with microglia activation³⁸ suggesting that peripheral inflammatory signals can propagate to the brain and disrupt neurovascular homeostasis. Atherosclerosis is a major contributor to VaD, largely through cholesterol-driven activation of inflammatory pathways³⁵. In macrophages, TNFα is released in response to cholesterol crystals, triggering RIPK1-dependent activation of NFkB and the transcription of pro-survival and pro-inflammatory genes³⁹. In addition, cholesterol can induce necroptosis in macrophages, particularly in response to high levels of oxLDL and other DAMPs, which can be reduced by treatment with Necrostatin-1, an inhibitor of RIPK1 necroptotic activity^21,22. We have previously demonstrated that RIPK1 promotes the development of atherosclerosis in the aorta, and that RIPK1 expression in the vascular endothelium and macrophages activates NFkB and induces the expression of pro-atherogenic cytokines^26,27.

We have previously shown that silencing RIPK1 expression by an RNA-targeting anti-sense therapy protects from a multitude of cardiometabolic risk factors. In the current, study, the delayed-intervention design allowed us to examine whether RIPK1 silencing is effective not only during early atherosclerotic lesion initiation, as we had seen previously in the aorta²⁷, but also in established atherosclerosis. Indeed, the present results confirm that RIPK1 silencing can protect from progression of established disease when administered at later stages of disease. Similar to what we observed in our previous studies in male C57Bl6 mice, RIPK1 silencing in Apoe^−/− mice protected against weight gain over the 8-week treatment period, and some mice even lost weight^26,27. Our previous work also demonstrated that RIPK1 silencing reduces adipose inflammation, enhances the metabolism of carbohydrates, and reduces levels of circulating pro-inflammatory cytokines²⁷. Obesity and metabolic syndrome are characterized by higher levels of systemic inflammation, and obesity has been linked to late-life dementia and Alzheimer’s disease (AD)⁴. Because adiposity induces systemic inflammation, and circulating inflammatory markers can compromise the BBB⁴⁰, this may contribute to the improved inflammatory markers and vascularity in the hippocampus in RIPK1^KD-treated mice^41,42.

Arterial wall stiffness, as measured by increased PWV^31,43, has been linked to poor cardiovascular outcomes, and higher PWV is associated with progression of dementia and cognitive decline over time^44–48. In Apoe^−/− mice fed a Western diet, we observed an increase in PWV compared to chow-fed animals, but this was mitigated by RIPK1^KD, despite the continued feeding of Western diet. The pronounced reduction in lesion area particularly in females suggests that RIPK1 contributes to lesion progression and remodeling at later disease stages³¹. Because carotid plaques are major contributors to cerebral hypoperfusion, attenuation of lesion burden in this vascular bed has important implications for cerebrovascular health^19,33. We did not measure blood pressure in our animals; therefore, we cannot exclude the possibility that the decrease in PWV in RIPK1^KD mice is associated with lower diastolic and/or systolic blood pressure. However, the significant reduction in carotid plaque size in RIPK1^KD-treated animals and a wider carotid lumen diameter likely contribute significantly to the decrease in PWV compared to WD-scrambled controls. Combined with the observation that RIPK1^KD preserved and/or restored vascularity in the hippocampus, these data indicate that silencing RIPK1 is a strategy for reducing vascular stiffness while preserving brain blood flow.

In the brain, RIPK1 signaling promotes neuroinflammation, neuron and oligodendrocyte cell death, and its inhibition using pharmacological or genetic approaches in mice leads to preservation of neurological function in central nervous system (CNS) pathologies²⁵. RIPK1-dependent necroptosis has been implicated in cerebral ischemia and models of ischemic stroke, and blocking RIPK1 in these models is neuroprotective and reduces neuroinflammation^49–51. In human Multiple Sclerosis (MS), RIPK1 and its downstream necroptosis targets RIPK3 and MLKL are phosphorylated, and inhibition of RIPK1 rescues MS phenotype in mice. The majority of studies of RIPK1 in the brain use kinase inhibitors or genetic mutations (D138N, K45A or K584R knock-in mice), implicating the kinase activity of RIPK1 in these diseases^21,39,52. Notably, in mouse models of Alzheimer’s Disease (AD), RIPK1 mediates lysosomal dysfunction in microglia and the pathology of AD in mice, which may be independent of its role in cell death^24,25. In the present study, we used anti-sense inhibitors that reduce RIPK1 mRNA expression in target tissues. Unlike kinase inhibition, RNA silencing reduces all RIPK1 activity in the cell⁵³. However, unlike full genetic loss of RIPK1, which leads to unrestrained cell death in many tissues, we have previously shown that the RIPK1 ASO approach dampens RIPK1-associated inflammation but still enables sufficient expression of RIPK1 to allow for its regulation of cell death^26,27. We confirmed loss of RIPK1 in the liver, as expected given that ASOs predominantly target the liver, due to natural physiological processes that favor hepatic accumulation and acts as a primary clearance organ, but we also noted a decrease in RIPK1 expression in the hippocampus in RIPK1^KD-treated mice (albeit variable and not statistically significant). The type of ASOs used in the present study does not generally cross the BBB⁵⁴, although it is possible that with Western diet-induced loss of endothelization that we observed in Apoe^−/− mice, some ASOs may indeed accumulate in brain parenchymal spaces. A side effect with RIPK1 genetic loss is an increase in skin lesions, due to the unrestrained cell death with complete absence of RIPK1⁵⁵. While some mice receiving RIPK1^KD did have skin lesions, this also occurred in mice receiving non-targeting scramble ASO and was likely caused by the combination of Western diet feeding and MWM testing environment. Nonetheless, future therapeutics aimed at reducing RIPK1 expression must consider careful titration of RIPK1 expression to avoid unwanted inflammatory lesions.

Loss of BBB integrity and reduced brain vascularity are key contributors of neuroinflammation and cognitive decline. Hypercholesterolemia has been shown to promote neuronal and vascular injury in the brain; for example, in Ldlr- and Apoe-double knockout mice, prolonged exposure to a high-cholesterol diet increases the expression of apoptotic markers in the hippocampus and striatum, which worsens as the duration of diet exposure lengthens^39,56. In both wild-type and Ldlr^−/− mice, high-cholesterol diet feeding increases BBB permeability in the hippocampus and prefrontal cortex⁵⁶. Consistent with these studies we observed that Caspase-3 was elevated in hippocampus lysates in Apoe^−/− mice fed a Western diet, which was accompanied by loss of EC tight junction CLDN5 and the pericyte marker PDGFRβ. Together, these findings suggest that endothelial and perivascular cells within the NVU may undergo apoptosis in response to chronic hyperlipidemia. Death of cells within the NVU would result in an overall loss of vascularity, which is supported by our functional ultrasound imaging data demonstrating reduced cerebral vascularity and shortened CD31⁺ endothelial structures in the hippocampus, indicating impaired microvascular integrity in WD-scramble mice⁵⁷ in comparison to chow-fed controls. Remarkably, RIPK1^KD-treated mice maintained cerebral perfusion and vascular content, showed reduced caspase-3 expression and preserved expression of CLDN5 and PDGFRβ, indicating improved survival of NVU cells and stabilization of BBB structure.

Although hyperlipidemia is commonly associated with robust astrogliosis and microglial activation, we observed modest changes in GFAP⁺ astrocytes and IBA1⁺ microglia, with subtle sex-dependent differences⁵⁸. In females, GFAP⁺ astrocyte and IBA1⁺ microglial cell numbers remained relatively stable between chow and RIPK1^KD conditions compared to WD-scr mice, suggesting a more resilient or tightly regulated neuroimmune environment. This aligns with prior evidence that female brains may exhibit attenuated microglial reactivity and enhanced astrocytic support under metabolic and inflammatory stress, potentially mediated by sex hormones and intrinsic transcriptional programs⁵⁹. In contrast, males displayed a more pronounced inflammatory phenotype under WD-scr conditions, characterized by increased IBA1⁺ microglia and reduced GFAP⁺ astrocytic coverage, both of which were normalized by RIPK1 silencing. This pattern is consistent with reports that microglia in males are more reactive to metabolic and vascular insults, contributing to heightened neuroinflammation and vulnerability to cognitive decline⁶⁰. Also, the progressive loss of NVU components, including endothelial cells and pericytes, may limit the capacity of glial cells to mount a classical inflammatory response, particularly in regions with reduced vascular density⁶¹. Indeed, pericytes play a critical role in maintaining BBB stability and regulating endothelial tight junction formation, and their loss has been associated with vascular hypoperfusion and increased BBB permeability⁶². In agreement with these structural changes, tracer experiments revealed increased accumulation of 10 kDa dextran in hippocampal vessels of WD-scramble mice, suggesting dysregulated vascular function in WD mice. Breakdown of the BBB permits entry of circulating inflammatory mediators and plasma proteins into the brain parenchyma, which can further amplify neuroinflammatory signaling and neuronal dysfunction⁶³. Crucially, inhibition of RIPK1 signaling preserves NVU integrity and stabilizes BBB function in the context of chronic hyperlipidemia. This is because systemic RIPK1 silencing is consistent with both improved perfusion and restored barrier function, rather than just decreased permeability, allowing preserved cerebral vascularity and preventing dextran accumulation.

We noted that isolated BMECs treated with pro-atherogenic stimuli TNFα and oxLDL or with mouse serum from WD-fed mice exhibited elevated phosphorylated RIPK1, caspase-3 and inflammatory marker including IL1β and NFkB. These findings suggest that circulating factors associated with hyperlipidemia can directly activate inflammatory and apoptotic pathways in brain endothelial cells. Importantly, serum from RIPK1^KD-treated mice attenuated these responses, indicating that systemic RIPK1 inhibition may reduce the pro-inflammatory circulating environment that contributes to BBB dysfunction. While we cannot rule out that other mechanisms may lead to reduced vascularity in the hippocampus, our data and that of previous studies support a model in which chronic hyperlipidemia promotes carotid atherosclerosis and systemic inflammation, leading to impaired cerebral perfusion, BBB disruption, and loss of NVU cells. The resulting exposure of the brain to circulating inflammatory mediators may further amplify neuroinflammation and contribute to cognitive decline. By reducing systemic inflammation and preserving cerebrovascular integrity, RIPK1 silencing appears to interrupt this pathological cascade linking peripheral atherosclerosis to neurovascular dysfunction.

While the use of a systemic approach to silence RIPK1 expression does not allow for the dissection of which cell type or tissue may be more responsible for the observed benefits in the brain, and further studies using cell-specific genetic loss may lead to more mechanistic insights. Restoration of vascular function, including improved cerebral blood flow and hippocampal vascularity, indicates that the vascular system retains a capacity for functional recovery even under sustained Western diet–induced metabolic stress. Nevertheless, although we cannot fully uncouple the cell-type specific effect of RIPK1^KD on the improvement in vascularity and cognitive performance, our current data and previous data suggest silencing RIPK1 during metabolic stress overall has a positive impact on obesity, atherosclerosis and, as shown here, cognitive function, making it an attractive therapeutic approach. Collectively, these results highlight RIPK1 as a promising therapeutic target, with the vasculature emerging as a key mediator of its beneficial effects.

Materials and methods

Animals

All experiments involving live animals were performed according to protocols reviewed and approved the University of Ottawa Animal Care and Use Committee, in accordance with the international standards established by the Canadian Council on Animal Care. Females and males (between 3 and 15 each) C57BL/6 Apoe^−/− mice (B6.129P2-Apoe^tm1Unc/J, Strain #:002052, RRID: IMSR_JAX:002052), between 8 and 12 weeks old were used for experiments. All mice were fed either a regular chow diet (Teklad 2018 S/2018SC) or a Western-type diet containing 21% fat and 34% sucrose by weight, and 0.2% total cholesterol (WD, Teklad TD.88137) containing at ad libitum with water, in a 12/12-h light/dark cycle. Out of the 24 weeks fed with either diet, the last 8 weeks were randomly assigned to simultaneously receive weekly subcutaneous injections of 50 mg/kg scrambled non-targeting ASO (CCTTCCCTGAAGGTTCCTCC) or RIPK1 ASO (TCAGCCACTTCTGAAGCATT, Ionis Pharmaceuticals)^26,27. Body weight was measured weekly throughout the last 8 weeks of the study and at the end of the study was: Chow (M) = 36.7 ± 4.2 g, (F) = 27.8 ± 2.7 g; WD Scr (M) = 50.4 ± 6.8 g, (F) = 39.2 ± 3.6 g; WD-RIPK1^KD (M) = 37.1 ± 2.7 g, (F) = 29.6 ± 3.6 g. Mice were fasted for 4 h and anesthetized with 100 mg/kg of pentobarbital before cardiac puncture, then total removal of the brain was performed and subsequently perfused with PBS. Tissues were snap frozen and kept at -80 °C until they were processed further. All animal experiments were performed in accordance with the ARRIVE guidelines.

Mouse behaviour assessment

Behavioural tests were completed at the University of Ottawa’s Behaviour Core Facility between 9 am and 4 pm. All animals were left to acclimatize in a light/dark cycle housing room for 8 days. On testing day, animals were habituated to the testing room for 30 min. Behaviour tests were performed with all mice in the following order: open field test (OF, 1 day); elevate plus maze test (EPM, 1 day), novel object recognition test (NOR, 2 days) and Morris Water Maze test (MWM, 8 days). All tests were conducted according to published protocols.

Open field test (OF)

Locomotor and anxiety-like behaviour was examined using a 10-minute open field test as described⁶⁴. This test was completed in bright light. Using a light meter, the light intensity was set to 100 lx for anxiety measurement. Each mouse was placed in the centre of an empty open field box (45 cm × 45 cm × 45 cm) and left to explore for 10 min. Three major arena zones were set up: 4 corner zones (each at 10 cm × 10 cm); a large center zone (25 cm × 25 cm); a small center zone (15 cm × 15 cm). Total distance travelled (cm), velocity (cm/s) and time spent in each respective zone were recorded with a camera mounted on the ceiling directly above the cage and analyzed by a computerized tracking system Ethovision 14 XT software (Noldus).

Elevated plus maze test (EPM)

Standard test of anxiety-like behaviour as described⁶⁵. Using a light meter, the light intensity was set to 100 lx. Each mouse was placed in the center of a maze composed of two arms crossed perpendicularly, each measuring 6 cm wide and 75 cm long. One arm consists of an open platform, while the second arm is enclosed by 20 cm high walls. The mouse was left to explore the maze for 10 min. Ethovision 14 XT software (Noldus) was used to record the number of entries and time spent in open arms with a camera mounted on the ceiling directly above the cage.

Novel object recognition test (NOR)

A 2-day NOR was as described⁶⁶. On day one, habituation day, each animal was habituated to an empty arena (45 cm × 45 cm × 45 cm) for 5 min. Following habituation, each mouse was removed from the empty arena and placed in a clean cage for 5 min. Two identical objects (a red cup or a white funnel) were placed in the arena, and the mouse was returned to the arena for a 5-minute familiarization period. The mouse was removed from the arena and placed in their cage. The following day (day 2), experimental day, mice were placed in the cage with the object recognition test, which consisted of one clean familiar object and one clean novel object (red cup or white funnel switched). The mouse was returned to their cage after a 5-minute recognition period. All interactions with the objects were recorded with a camera mounted on the ceiling directly above the cage and analyzed by a computerized tracking system, Ethovision 14 XT software (Noldus). Object recognition was scored as the time during which the nose of the animal was located within 2 cm of the object. A discrimination index was calculated as: time spent interacting with novel object/ (time spent interacting with novel object + time spent interacting with familiar object)].

Morris water maze test (MWM)

For the MWM test⁶⁷, a round pool with a diameter of 90 cm was filled with warm water (21 °C) stained with an opaque dye and a moving platform 10 cm in diameter was placed in the Right-Back quadrant (R-B) in the pool, 1–2 cm below the water surface. There were extra-maze cues on the room wall that remained constant throughout testing. Animals were trained for 5 days after modelling the damaging factor. The training cycle consisted of 4 attempts lasting 60 s each. Putting the animal into the pool on the 4 different quadrants trained them to find the platform. If the animal finds the platform, it is permitted to remain on it for 5 s. If it fails to find the platform for the testing time, it is placed on the platform forcibly for 15 s. All trials were videotaped with a camera mounted on the ceiling directly above the pool and analyzed by a computerized tracking system, Ethovision 14 XT software (Noldus). The latency to find the platform was calculated. Further, the reconsolidation of the long-term memory of the animal was assessed 48 h after the last training session. This parameter was assessed by testing animals in a pool without a platform for 1 min. The duration of the animal’s stay in the area where the platform was previously located, to the total testing time (percentage of time in the quadrant), was estimated.

Brain blood flow & vascularity measurements

One to two days before the ultrasound procedure, hair was removed from the skull (for cerebral blood flow) and neck (for carotids). On the day of imaging, animals were brought to the imaging room and allowed one hour for adaptation to the environment. After 5 to 8 min under anesthesia (isoflurane 2–3%, oxygen 1-1.5%) and with the mouse fully sedated, we placed the mouse on a heating platform to maintain stable body temperature. Left and right carotid lumen diameter during systole and diastole was measured with M-mode imaging and Doppler ultrasound. Only segments with visually circular or near-circular lumens were included for analysis³¹. Images were acquired with FUJIFILM VisualSonics Vevo 3100 LAZR-X photoacoustic imaging system. The MX550D transducer (Bandwidth: 25–55 MHz, Axial Resolution: 40 μm) was used to image carotids, and the MX250 transducer (Bandwidth: 15–30 MHz, Axial Resolution: 75 μm) was used for the blood-brain flow, providing sufficient spatial resolution for quantification of cerebral vascular flow and microvascular architecture. Colours indicate blood flow direction and velocity: red color represents blood flow moving toward the transducer. Blue color represents blood flow moving away from the transducer. Darker colors denote slower movement, while lighter/brighter colors imply faster flow.

Brain immunohistochemistry

Immunohistochemistry of mouse brain tissue was performed as previously described⁶⁸. Mice were placed under anesthesia (pentobarbital, 100 mg/kg, i.p.) prior to brain isolation. The whole brain was removed from the skull and postfixed in 4% paraformaldehyde (PFA) overnight at 4 °C. After fixation, brains were cryoprotected with 20% (2–3 days) and then 30% sucrose. Tissue was sectioned (25 μm) from embedded brains in optimal cutting temperature (OCT, #23-730-571) on a coronal side. Slides were rinse 3x with PBS 1x and permeabilized with PBS-T 5% (5% Triton-X in PBS) solution 3x for 5 min and blocked for 1.5 h in a humid slide chamber with blocking solution (10% normal donkey serum and 5% fish gelatin in PBS-T 5%), followed by primary antibodies in blocking solution for 24 h at 4 °C overnight, and finally with Alexa Fluor-conjugated secondary antibodies (1:500, Invitrogen) in blocking solution for 2 h at room temperature in a humid slide chamber. Slides were incubated for 30 s with TrueBlack^® (Lipofuscin Autofluorescence Quencher # 23007) and mounted with aqueous mounting medium (Dako #s3023). 14–16 tailed Z-stack images (Program Zen 2.6 pro) along the hippocampus and the DG section per animal were imaged with a Zeiss Axio Imager M2 microscope equipped with a digital camera (Axiocam 506 mono) and the ApoTome 2 module, with a 10× and 20× objective lens. The images were analyzed with Fiji ImageJ2 software Version 2.16.0 (Cambridge Astronomical Survey Unit). The following primary antibodies were used: guinea pig anti-GFAP (1:400, #173 004, Synaptic System), goat anti-IBA1 (1:300, #ab289874, Abcam) and rat anti-CD31 (1:300, #553370, BD Biosciences).

BBB integrity assessment

Mice were anesthetized with a mixture of ketamine (100 mg/kg of body weight) and xylazine (10 mg/kg of body weight) and administered 0.1 mL of 1 mg/mL solution of fixable Alexa Fluor 555-dextran (D34679, mW = 10 kDa, ThermoFisher Scientifics) through retro-orbital injection. After 30 min, mice were perfused with ice-cold 0.1 M PBS for 5 min, followed by 5 min with 4% paraformaldehyde (PFA). Brains were post-fixed overnight in 4% PFA at 4 °C in the dark, then sliced on a vibratome (Leica) at 40 μm thickness. Free-floating sections were washed in 1x PBS and incubated for 2 h in blocking solution (5% NDS in 1x M PBS with 5% fish gelatin). Sections were then incubated in rat anti-CD31 (1:300, #553370, BD Biosciences) and rabbit anti-aquaporin 4 (1:300, #AB3594, Sigma-Aldrich). Sections were washed again three times in PBS, cover slipped and mounted with aqueous mounting medium (Dako #s3023). 14-16-stack images of the median eminence, amygdala, ventral and dorsal hippocampus, with a Zeiss Axio Imager M2 microscope equipped with a digital camera (Axiocam 506 mono) and the ApoTome 2 module, with a 10× objective lens. For analysis of MFI volume, 5 planes from 5 non-consecutive sections were acquired from each animal (3). The volume and intensity of Alexa Fluor 555-dextran signal in the total number of vessels were analyzed with Fiji ImageJ2 software Version 2.16.0 (Cambridge Astronomical Survey Unit).

Carotid immunohistochemistry

Atherosclerotic plaque burden was assessed by histological processing of both carotid arteries. The carotid arteries, including the upper aortic arch and carotid bifurcations, were harvested, fixed in 4% PFA, and preserved in a 30% sucrose solution. Tissues were embedded in Optimal Cutting Temperature (OCT, #23-730-571). Serial cross-sections Ten µm-thick sections were sectioned with a Thermo Cryostat (Scientific #HM550). Slides were rinsed 3x with PBS 1x and permeabilized with PBS-T 5% (5% Triton-X in PBS) solution 3x for 5 min and block for 1.5 h in a humid slide chamber with blocking solution (10% normal donkey serum and 5% fish gelatin in PBS-T 5%), followed by primary antibodies in blocking solution for 24 h at 4 °C overnight, and finally with Alexa Fluor-conjugated secondary antibodies (1:500, Invitrogen) in blocking solution for 2 h at room temperature in a humid slide chamber. Slides were mounted with aqueous mounting medium (Dako #s3023). 10 tailed Z-stack images (Program Zen 2.6 pro) per animal were imaged with a Zeiss Axio Imager M2 microscope equipped with a digital camera (Axiocam 506 mono) and the ApoTome 2 module, with a 20× objective lens. The images were analyzed with Fiji ImageJ2 software Version 2.16.0 (Cambridge Astronomical Survey Unit). The following primary antibodies were used: guinea pig anti-CD68 (1:400, #MCA1957, Bio Rad), goat anti-MAC2 (1:200, #CL8942AP, Cedarlane) and rat anti-CD31 (1:300, #553370, BD Biosciences). Histopathologic evaluation was performed by staining with hematoxylin and eosin. All images were acquired on a slide scanner (Leica Biosystems) Aperio VERSA Scanning System on a brightfield scanning. Scan times for the Aperio were approximately 206 s for a 15 × 15 mm area at 10× or 20× magnification.

Primary mouse brain endothelial cell isolation

Isolation of primary mouse brain macrovascular endothelial cells (BMECs) was performed as previously described⁶⁹. All mice were euthanized by cervical dislocation. The cortex from Apoe^−/− mice (34–36 weeks old) was dissected in cold HBSS without calcium and magnesium using autoclaved tools submerged in 100% ethanol 30 min before dissection. The cerebral cortex was minced into small 2–3 mm pieces and dissociated using Neural Tissue Dissociation Kit P compounds (Miltenyi Biotec, #130-092-628) to obtain a cell suspension. Cell isolation procedures were completed according to the manufacturer’s instructions. Cell suspension underwent a positive selection to isolate BMECs (CD31 Microbeads, Mouse Miltenyi Biotec #130-097-418), using a magnetic MACs separator. ECs were seeded in an ibidi 8-well slide (#80807, ibidi) and after 7–8 days without passages, pure BMECs were obtained.

Primary mouse brain endothelial cell treatments

BMECs were isolated, seeded and fully grown on ibidi slides. Cells were treated for 6, 16 and 24 h with 10 ng/ml of TNFα (Thermo Fisher Scientific #315–01 A) or 50 µg/ml of oxLDL according to previous methods⁷⁰(Cu₂SO₄ oxidation 6 h) or with 10% serum extracted from mice on chow, WD with scrambled ASO and WD with anti-Ripk1 ASO for 24 weeks. After treatment slides were rinsed 3x with PBS 1x, fixed with 4% PFA for 20 min, rinse 2x with PBS 1x and permeabilized with PBS-T 2% (2% Triton-X in PBS) solution 3x for 5 min and block for 2 with blocking solution (10% normal donkey serum and 5% fish gelatin in PBS-T 2%) room temperature, followed by primary antibodies in blocking solution for 1 h room temperature, and with Alexa Fluor-conjugated secondary antibodies (1:500, Invitrogen) in blocking solution for 2 h at room temperature. Z-stack images (Program Zen 2.6 pro) were taken with a Zeiss Axio Imager M2 microscope equipped with a digital camera (Axiocam 506 mono) and the ApoTome 2 module, with a 20× objective lens. The images were analyzed with Fiji ImageJ2 software Version 2.16.0 (Cambridge Astronomical Survey Unit). The following primary antibodies were used: rat anti-CD31 (1:300, #553370, BD Biosciences), rabbit anti-IL-1β (1:300, #ab234437, Abcam), rabbit anti- NF-κB (1:300, #ab16502, Abcam), rabbit anti-Casp3 (1:300, #9664S, Cell Signaling), rabbit anti-pRipk1 (1:300, #38662, Cell Signaling).

Western blot analysis

The hippocampus was dissected immediately upon sacrifice and snap frozen until protein extraction. Tissues were thawed and incubated with lysis buffer (RIPA; Tris-HCl (e.g., 250 mM, pH 7.4); NaCl (e.g., 1.5 M); NP-40 (e.g., 10% v/v); Sodium deoxycholate (e.g., 0.5% w/v); SDS (e.g., 0.1% w/v); EDTA or EGTA (e.g., 1 mM) and supplemented with protease inhibitor cocktail (Roche #04 693 123 001) and phosphatase inhibitor cocktail (Roche #04 906 873 001) on ice (added immediately before use). Proteins (30 µg) were subjected to 4–15% SDS-PAGE and transferred to PVDF membranes (Bio-Rad # 1620177) for immunoblot analysis. PVDF membranes were blocked with 50% TBS-T 1 × (1% Tween 20 in TBS) and 50% blocking solution (Bio-Rad #12010020), incubated overnight with antibodies against rabbit anti-IL-1β (1:1000, ab234437, Abcam), rabbit anti- NF-κβ (1:1000, #ab16502, Abcam), rabbit anti-Casp3 (1:1000, 9664 S, Cell Signaling), rabbit anti-Rip (1:1000, #3493, Cell Signaling), rabbit anti-Claudin 5 (1:1000, #34-1600, Thermo Fisher Scientific) and goat anti-PDGF R (1:1000, #AF1042, RnD Systems). Goat anti-rabbit and donkey anti-goat IRDye^® secondary antibodies were used for detection on ChemiDoc™ Imaging Systems (Bio-Rad).

Statistical analysis

We performed power calculations based on our primary endpoint of atherosclerotic lesion size, as we had previously conducted a similar study in Apoe^−/− mice using RIPK1 ASOs, thus used this for our power calculations. Based on a group size of n = 12, we had > 95% power to detect a 20% change in lesion size (α = 0.05), with a standard deviation of 10% using a One-Way ANOVA (this variation in lesion size is what we typically observe^71,72. We added 3 mice per group to account for possible attrition during the long-term diet studies. Mice were allocated to different groups for certain end-point experimental analyses depending on different sample processing needs. Data shown are a representative mean ± SEM experiment of at least 3 experiments. For comparison between all groups, 1-way ANOVA (P ≤ 0.05) with multiple comparisons tests was performed using GraphPad Prism 10 Software. Data distribution was assumed to be normal, as we tested for normality by either D’Agostino-Pearson or Shapiro-Wilk. A P < 0.05 indicates the data deviates significantly from a Gaussian distribution. If any group within a comparison did not meet normality assumptions, non-parametric analyses were performed for that dataset. A one-way or two-way ANOVA, followed by a post-hoc test (Tukey’s), was used for multigroup comparisons. p < 0.05 was considered significant.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

KJR and JSL conceived the study, with input from BL, RS and MFA. JSL, CS and KJR wrote the manuscript, with input from all authors. Experiments were conducted by JSL with help from MFA, VGP, SS, AH, CS, NS and MG. Funding for the studies were obtained by KJR and BL. All authors reviewed and approved the manuscript.

Funding

These studies were supported by the Canadian Institutes for Health Research (to KJR, BL), the European Research Area Network on Cardiovascular Diseases (to KJR) and from the Brain-Heart Interconnectome Canada First Research Excellence Fund (to JSL, RS, BL, KJR).

Data availability

The experimental data that support the findings of this study are available in FigShare https://figshare.com/articles/dataset/Figures_1_to_4_raw_data/30911765/2.

Declarations

Competing interests

The authors declare no competing interests.