Targeting Unc5b in macrophages drives atherosclerosis regression and pro-resolving immune cell function

Significance

Our findings identify the netrin-1 receptor Unc5b as a therapeutic target to ameliorate advanced atherosclerosis. Myeloid-specific deletion of Unc5b significantly reduced plaque burden and complexity in mice with advanced atherosclerosis, beyond reversal of hypercholesterolemia alone. This reduction was associated with a beneficial rewiring of monocyte/macrophage trafficking within the plaque and increased efferocytosis, which promoted a pro-resolving environment characterized by an increase in local atheroprotective T cells. These results suggest that targeting pro-inflammatory plaque macrophage function by decreasing Unc5b is a promising therapeutic strategy to complement standard-of-care cholesterol lowering therapies for established atherosclerosis and advance our understanding of atherosclerotic plaque regression.

Abstract

Atherosclerosis results from lipid-driven inflammation of the arterial wall that fails to resolve. Imbalances in macrophage accumulation and function, including diminished migratory capacity and defective efferocytosis, fuel maladaptive inflammation and plaque progression. The neuroimmune guidance cue netrin-1 has dichotomous roles in inflammation partly due to its multiple receptors; in atherosclerosis, netrin-1 promotes macrophage survival and retention via its receptor Unc5b. To minimize the pleiotropic effects of targeting netrin-1, we tested the therapeutic potential of deleting Unc5b in mice with advanced atherosclerosis. We generated Unc5b^fl/flCx3cr1^creERT2/WT mice, which allowed conditional deletion of Un5b (∆Unc5b^MØ) in monocytes and macrophages by tamoxifen injection. After inducing advanced atherosclerosis by hepatic PCSK9 overexpression and western diet feeding for 20 wk, Unc5b was deleted and hypercholesterolemia was normalized to simulate clinical lipid management. Deletion of myeloid Unc5b led to a 40% decrease in atherosclerotic plaque burden and reduced plaque complexity compared to Unc5b^fl/flCx3cr1^WT/WT littermate controls (Ctrl^MØ). Consistently, plaque macrophage content was reduced by 50% in ∆Unc5b^MØ mice due to reduced plaque Ly6C^hi monocyte recruitment and macrophage retention. Compared to Ctrl^MØ mice, plaques in ∆Unc5b^MØ mice had reduced necrotic area and fewer apoptotic cells, which correlated with improved efferocytotic capacity by Unc5b-deficient macrophages in vivo and in vitro. Beneficial changes in macrophage dynamics in the plaque upon Unc5b deletion were accompanied by an increase in atheroprotective T cell populations, including T-regulatory and Th2 cells. Our data identify Unc5b in advanced atherosclerosis as a therapeutic target to induce pro-resolving restructuring of the plaque immune cells and to promote atherosclerosis regression.

Atherosclerotic cardiovascular disease (ASCVD) and its complications, myocardial infarction and stroke, remain a leading cause of morbidity and mortality worldwide (1). Increased understanding of ASCVD over the last decades has led to a steady decline in deaths due to CVD, which can be attributed to the promotion of healthy lifestyle changes, advances in diagnostics, and the effective management of risk factors, particularly hypercholesterolemia (1). Yet, the residual risk of a cardiovascular event remains high even in patients who achieve very low levels of circulating cholesterol through pharmacological intervention (2, 3). Furthermore, the global rise in obesity and its attendant cardiovascular risk factors (e.g., diabetes, hypertension, dyslipidemia) represent a burgeoning health crisis worldwide. Indeed, declines in ASCVD mortality rates in the United States reversed in 2020 and continue to increase (4), highlighting the need to develop orthogonal approaches for ASCVD management.

Inflammation is an important mediator of ASCVD pathogenesis, from its initiation to the emergence of complications (5). Atherogenesis is triggered by the accumulation of cholesterol-rich low-density lipoproteins (LDL) and remnant lipoprotein particles in the subendothelial space of the arterial wall at focal regions of disturbed blood flow (6). Activation of the overlying endothelium and the ensuing inflammatory response orchestrated by innate and adaptive immune cells leads to the formation of fibrofatty plaques. Macrophages are central players in this inflammatory response and contribute to the initiation, progression, and ultimately, rupture of atherosclerotic plaques (7). Tissue-resident macrophages and recruited monocyte-derived macrophages internalize the retained lipoproteins resulting in the formation of lipid-laden foam cells, a hallmark of atherosclerotic plaques (8). Accumulating macrophages fuel vascular inflammation by producing pro-inflammatory mediators including; cytokines and chemokines that recruit and activate smooth muscle cells and other immune cells (e.g., neutrophils, dendritic cells, T and B cells) (8, 9); reactive oxygen and nitrogen species that further modify retained lipoproteins and elicit tissue damage, and; proteases that degrade extracellular matrix leading to plaque instability (10). Notably, lipid-laden macrophages have a decreased ability to migrate, resulting in their persistence in the artery wall and impeding resolution of inflammation (11, 12). Failure to resolve atherosclerotic inflammation leads to plaque progression characterized by macrophage cell death and dysfunctional efferocytosis (clearance of dying cells), necrotic core formation, and thinning of the smooth muscle cell fibrous cap, which can increase plaque vulnerability leading to erosion or rupture (13, 14). In humans, atherosclerotic plaques can progress slowly over decades, but this process is accelerated by local and/or systemic inflammation (15) increasing the risk of atherothrombotic events such as myocardial infarction and stroke.

Identification of the factors that contribute to macrophage accumulation and failed inflammation resolution in plaques offers promise for the targeted treatment of atherosclerotic inflammation. We and others have shown that neuronal guidance proteins, primarily studied for their role in guiding developing axons during embryogenesis, can participate in the pathogenesis of atherosclerosis (16). For example, semaphorins and ephrins can promote vascular inflammation, endothelial dysfunction, and vascular smooth muscle cell proliferation (17, 18). In addition, netrin-1 has been reported to have both pro- and antiatherogenic roles by altering the migration and survival of various cell types in the atherosclerotic vasculature (19, 20). Netrin-1 expression by endothelial or epithelial cells reduces leukocyte adhesion via the adenosine 2b receptor (A2BAR) and consequently, intravascular infusion of netrin-1 inhibits inflammation by reducing leukocyte infiltration into tissues (21, 22). Conversely, netrin-1 expression by macrophages promotes their chemostasis, accumulation, and survival via the receptor Unc5b, contributing to sustained inflammation in the artery wall, as well as adipose tissue in obesity (12, 23). Ntn1 and Unc5b are transcriptionally upregulated by hypoxia inducible factor 1 alpha in response to hypoxic conditions, such as those found within atherosclerotic plaques (24). Indeed, both netrin-1 and Unc5b are upregulated in macrophages present in human and mouse atherosclerotic plaques (25, 26), suggesting that autocrine/paracrine signaling may prevent macrophage egress from plaques. Netrin-1 secreted by macrophages can also promote smooth muscle cell migration via the receptor neogenin leading to plaque progression (12). Notably, macrophage-specific deletion of netrin-1 in mice with established atherosclerosis promotes plaque regression via reduced retention of plaque macrophages and smooth muscle cells within the lesion core, but maintains the ACTA2-rich fibrous cap, allowing a pro-resolving reorganization of the plaque immune landscape (27). Together, these findings implicate netrin-1 in preventing inflammation resolution in the plaque, however, whether selectively targeting its receptor Unc5b holds potential for the treatment of established atherosclerosis remains unexplored.

The Canakinumab Anti-Inflammatory Thrombosis Outcome Study (CANTOS) trial has demonstrated the clinical relevance of anti-inflammatory therapies by blocking interleukin-1β (IL-1β) with the canakinumab antibody, resulting in a significant reduction in major adverse cardiovascular events. However, an unintended and undesirable consequence was an increase in fatal infections, underscoring the potential risks associated with targeting inflammatory pathways. Similar concerns arise with targeting netrin-1, a molecule with divergent roles in acute and chronic inflammation, mediated by its interactions with different cell types and receptors. Netrin-1 binding to Neogenin and DCC receptors mediates chemoattraction, while its interaction with Unc5b-expressing cells induces chemorepulsion (28, 29). In cancer research, inhibiting the netrin-1-Unc5b axis has been shown to be a promising strategy to prevent epithelial-to-mesenchymal transition and reduce tumor mass (30, 31). Hypoxia-induced upregulation of both netrin-1 and Unc5b in atherosclerotic plaques suggests Unc5b as a potential target for atherosclerosis treatment (24). Therefore, we aimed to evaluate whether selective inhibition of the Unc5b receptor on macrophages can provide therapeutic benefits in established atherosclerosis, potentially offering a more specific and safer alternative to broad netrin-1 inhibition.

Previous studies have shown that while cholesterol lowering through diet or statin use reduces cardiovascular disease risk, it is insufficient to appreciably regress existing atherosclerotic plaques. Notably, we observed that netrin-1 levels in plaque remain high even after lipid lowering (27), suggesting that netrin-1 supports macrophage persistence and sustained inflammation after normalization of hypercholesterolemia. Here, we tested whether genetically deleting macrophage expression of the netrin-1 receptor Unc5b, coincident with lipid lowering, could induce inflammation resolution and plaque regression in mice with established atherosclerosis. Using a mouse model with a tamoxifen-inducible Cre expressed under the macrophage-specific Cx3cxr1 promoter, we show that deleting myeloid Unc5b in advanced plaques reduces atherosclerotic plaque burden and complexity consistent with atherosclerosis regression, compared to reducing hypercholesteremia alone.

Monocyte-macrophage tracking techniques revealed that deleting Unc5b expression decreased the existing plaque macrophage population by enhancing their egress but also by mitigating the influx of pro-inflammatory Ly6C^hi monocytes into the plaque. Unc5b-deficient macrophages displayed increased efferocytic capacity, which corresponded to reduced apopototic cell accumulation and necrotic area in plaques. Alongside these beneficial changes in macrophage dynamics, we observed increased accumulation of anti-inflammatory T cell subsets in the plaque after Unc5b deletion, including Th2 and T regulatory (Treg) cells, consistent with a shift toward a pro-resolving environment. Collectively, our data indicate that targeting macrophage expression of Unc5b can reverse chronic inflammation in the plaque and promote atherosclerosis regression.

Results

Myeloid Cells express Unc5b in Atherosclerotic Plaque in the Absence of Hypercholesterolemia.

Reversing hypercholesterolemia is an important first step to halt atherosclerosis progression but is considered insufficient to drive significant regression of advanced plaque. To investigate the expression of Unc5b in plaque immune cells, we interrogated single-cell RNA-sequencing (scRNA-seq) of aortic arch leukocytes of Ldlr^–/– mice with advanced atherosclerosis induced by western diet feeding (WD; 14 to 16 wk) followed by normalization of hypercholesterolemia induced by switching to chow diet for 4 wk (Fig. 1A). Within plaque immune cells, Unc5b expression was largely specific to the myeloid lineage (Fig. 1B). Differential gene expression analysis indicated that in mice with advanced atherosclerosis, Unc5b⁺ myeloid cells not only persist (Fig. 1C), but are significantly increased following the normalization of hypercholesterolemia (total cholesterol: baseline group 1,403 ± 758 mg/dL, halted progression group 124 ± 74 mg/dL; P < 0.0001). Comparison of Unc5b-expressing versus nonexpressing myeloid cells identified 902 differentially expressed genes (Padj ≤ 0.05; SI Appendix, Fig. S1), which pathway analysis identified as strongly enriched for genes associated with T cell–mediated inflammation, including Th1 activation, and IL-33, IL-7, and STAT3 signaling (Fig. 1D). Based on these data, we hypothesized that Unc5b-expressing myeloid cells may sustain inflammatory processes that prevent atherosclerosis regression following lipid lowering.

Fig. 1.

Myeloid Unc5b expression impairs atherosclerosis regression. (A and B) Uniform Manifold Approximation and Projection visualization of scRNA-seq of CD45⁺ cells isolated from the aortic arch of male Ldlr^–/– mice fed a WD for 16 wk (A), with dot plot showing Unc5b normalized gene expression (B). DN, double negative; DP, double positive; NK, Natural Killer; n = 13,397 cells. (C) Violin plot showing normalized Unc5b expression in Unc5b⁺ myeloid cells at baseline and 4 wk after mice were switched to chow diet. (D) Pathway enrichment analysis of genes differentially expressed in Unc5b⁺ and Unc5b^– myeloid cells. (E) Schematic of interventional study design in mice; Unc5b^fl/fl-Cx3cr1^WT/WT(Ctrl^BL) and Unc5b^fl/fl-Cx3cr1^CreERT2/WT (Unc5b^BL) mice were injected with PCSK9-AAV and fed western diet for 20 wk to establish atherosclerosis (baseline), and then injected with tamoxifen and put on chow diet for 4 wk (TAM + chow) to induce Unc5b deletion (∆Unc5b^MØ) or not (Ctrl^MØ). (F) Unc5b mRNA levels in thioglycolate-elicited peritoneal macrophages (Left, pMØ) and bone marrow–derived macrophages (Right; BMDM) isolated from ∆Unc5b^MØ and Ctrl^MØ mice (n = 3 per group). (G) Representative hematoxylin and eosin (H&E) staining of aortic root plaques (outlined in dashed-line; Scale bar, 200 µm), and quantification of plaque area along the aortic root of mice at baseline and after TAM + chow diet. (H and I) Representative images of aortas En Face (H), and quantification (I) of plaque area in the whole aorta, aortic arch, and abdominal aorta of mice at baseline and after TAM + chow diet. Scale bar, 2 mm. Data are mean ± SEM. P-values by Student’s T-test (C and F) or one-way ANOVA with Tukey’s multiple comparison test (G and I).

Macrophage-Specific Deletion of Unc5b Favors Plaque Regression.

To test the role of Unc5b in preventing inflammation resolution and plaque regression, we devised an interventional study in mice in which advanced atherosclerotic plaques were induced by overexpression of gain-of-function murine PCSK9-D377Y (PCSK9-AAV) and WD feeding (27), followed by therapeutic cholesterol lowering in the presence or absence of Unc5b deletion in macrophages (Fig. 1E and SI Appendix, Fig. S2). The temporal deletion of Unc5b was achieved using Unc5b^fl/fl mice (32) expressing a tamoxifen-inducible CreER^T2 fusion protein under control of the Cx3cr1 promoter, so that Unc5b is deleted in Cx3cr1-expressing monocytes and macrophages in Cre-positive mice (Unc5b^fl/flCx3cr1^CreERT2/WT) or remains expressed in Cre-negative littermate control mice (Unc5b^fl/flCx3cr1 ^WT/WT) upon administration of tamoxifen (Fig. 1E). To assess atherosclerosis parameters prior to intervention, a cohort of Unc5b^fl/flCx3cr1^CreERT2/WT mice (Unc5b^BL) and Unc5b^fl/flCx3cr1^WT/WT littermates (Ctrl^BL) was analyzed 20 wk after treatment with PCSK9-AAV and WD feeding, and the remaining mice were switched to chow diet to lower plasma cholesterol levels and injected with tamoxifen intraperitoneally (i.p.) for 5 d to either delete Unc5b in macrophages (∆Unc5b^MØ) or not (Ctrl^MØ) (Fig. 1E). Deletion of Unc5b after tamoxifen injection was confirmed in both thioglycolate-elicited peritoneal macrophages and in bone marrow–derived macrophages (BMDM) of ∆Unc5b^MØ mice (Fig. 1F).

Morphometric assessment of atherosclerosis revealed no differences in plaque burden at baseline between the two groups of mice (Unc5b^BL and Ctrl^BL) (Fig. 1 G and H), and as expected, chow diet combined with tamoxifen treatment did not alter plaque size in Ctrl^MØ mice, confirming that reverting hypercholesterolemia alone is insufficient for plaque regression (Fig. 1 G and H). By contrast, plaque area in the aortic root was reduced by 30% in mice with induced myeloid Unc5b-deficiency (∆Unc5b^MØ) compared to control mice (Ctrl^MØ) postintervention (Fig. 1G), despite equivalent body weight, plasma cholesterol levels, and lipoprotein cholesterol distribution (SI Appendix, Fig. S2 A–C). Furthermore, ∆Unc5b^MØ mice showed 40% lower aortic plaque burden (Fig. 1 H and I), including in both the aortic arch and abdominal aorta (Fig. 1I) compared with baseline mice and Ctrl^MØ mice postintervention.

Therapeutic Deletion of Unc5b Promotes Favorable Cellular Remodeling within Plaque.

To examine changes in plaque composition from baseline to postintervention, we measured macrophage, lipid, and smooth muscle actin (SMA) content by immunofluorescence staining in cross-sections of the aortic root (Fig. 2A). Consistent with a role for Unc5b in macrophage persistence, ∆Unc5b^MØ mice showed a 40% decrease in plaque CD68⁺ macrophage staining compared to their baseline counterparts and Ctrl^MØ mice, even after adjusting for plaque size (Fig. 2B). Furthermore, ∆Unc5b^MØ mice displayed a 60% reduction in plaque neutral lipid accumulation compared to their baseline counterpart and Ctrl^MØ mice, as assessed by BODIPY staining (Fig. 2C). By contrast, there was no difference in total SMA content in ∆Unc5b^MØ and Ctrl^MØ plaques at baseline or postintervention, whether in absolute terms (Fig. 2D) or normalized by plaque size (SI Appendix, Fig. S3). However, focusing our analysis on the fibrous cap region where smooth muscle cell–mediated plaque stabilization occurs in humans (33), we found an increase in SMA content within the 40 µm-subendothelial region of the plaque in ∆Unc5b^MØ mice compared to Ctrl^MØ mice (Fig. 2E). Furthermore, staining for collagen revealed a threefold increase in plaque collagen content in ∆Unc5b^MØ mice compared to baseline, a trend that is less pronounced in Ctrl^MØ littermate controls (Fig. 2F). Together, our data indicate that therapeutic deletion of Unc5b in macrophages induces beneficial plaque remodeling suggestive of reduced inflammation and lipid burden, coincident with fibrous cap thickening.

Fig. 2.

Macrophage-specific deletion of Unc5b reduces plaque macrophage and lipid content. (A) Representative immunofluorescence images of aortic root plaque of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and after TAM + chow diet costained for macrophages (CD68; red) and neutral lipid content (BODIPY, green), SMA (ACTA2, white). White dashed-line indicates plaque area and yellow dashed-line outlines plaque 40 µm-subendothelial space (Scale bar, 200 µm.) (B–F) Quantification of CD68⁺ (B); BODIPY⁺ (C); total ACTA2⁺ (D); fibrous cap ACTA2⁺ (E); and collagen (F) content of plaques shown in A. Data are mean ± SEM. P-values by one-way ANOVA with Tukey’s multiple comparison test (B, C, E, and F) or Kruskal–Wallis test with Dunn’s multiple comparison test (D).

Unc5b Regulates Macrophage Dynamics within the Plaque.

To understand the mechanisms responsible for the observed decrease in lesional macrophages in ∆Unc5b^MØ mice, we measured monocyte-macrophage dynamic processes in the plaque including recruitment, proliferation, retention, and death. To assess monocyte recruitment, we pulse-labeled circulating pro-inflammatory Ly6C^hi monocytes with EdU and patrolling Ly6C^lo monocytes with fluorescent microbeads 72 h prior to harvest, allowing us to trace their trafficking into the plaque (27, 34, 35). While there were no differences in monocyte recruitment between groups at baseline, EdU⁺ pro-inflammatory Ly6C^hi monocyte recruitment was significantly reduced in the aortic root plaque of ∆Unc5b^MØ mice postintervention compared to baseline and Ctrl^MØ mice (Fig. 3A). By contrast, recruitment of bead-labeled patrolling Ly6C^lo monocytes was reduced in both ∆Unc5b^MØ and Ctrl^MØ mice postintervention (Fig. 3B). Assessment of Ly6C^hi and Ly6C^lo monocyte presence in plaques by digestion of the aortic arch and flow cytometric analysis showed reduced accumulation of Ly6C^hi monocytes in plaques of both Ctrl^MØ and ∆Unc5b^MØ mice postintervention, whereas Ly6C^lo monocyte content was increased, particularly in ∆Unc5b^MØ mice (Fig. 3C). The increase in Ly6C^lo macrophage content postintervention may represent an increase in Ly6C^hi-to-Ly6C^lo monocyte differentiation in the plaque, since trafficking studies in the aortic root showed reduced Ly6C^hi and Ly6C^lo monocyte recruitment compared to baseline (36) (Fig. 3 A and B). To assess the impact of deleting Unc5b on monocyte and macrophage proliferation, we measured CD68 and Ki67 coexpression by immunofluorescence staining in aortic root plaques (Fig. 3D). We found that Ki67⁺CD68⁺ cells decreased in plaques of both genotypes postintervention, albeit to a greater degree in ∆Unc5b^MØ mice (Fig. 3D). To evaluate macrophage retention within plaques, we pulse-labeled circulating monocytes with red beads immediately before initiating the tamoxifen intervention and assessed the number of bead-labeled macrophages remaining in plaques 4 wk later. We observed a decrease in plaque macrophage retention in both groups postintervention, although this decrease was much greater in ∆Unc5b^MØ mice compared to Ctrl^MØ (−60% versus −30%) (Fig. 3E). Finally, we quantified the accumulation of apoptotic cells (AC) in plaques via TdT-mediated dUTP-biotin nick end labeling (TUNEL) staining and found a 75% reduction in ACs in ∆Unc5b^MØ mice compared to baseline and a 55% reduction compared to Ctrl^MØ mice postintervention (Fig. 3F). These data suggest that targeting Unc5b alters monocyte-macrophage dynamics in the plaque to promote plaque regression.

Fig. 3.

Unc5b controls macrophage dynamics within the plaque. (A and B) Quantification EdU⁺ Ly6C^hi monocyte recruitment (A) and green bead-labeled Ly6C^lo monocyte recruitment (B) in the aortic root plaques of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and after TAM + chow diet intervention. (C) Flow cytometric quantification of Ly6C^hi (Left) and Ly6C^lo (Right) monocyte accumulation in aortic arch plaques of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and postintervention. (D) Representative immunofluorescence staining and quantification of Ki67⁺CD68⁺ proliferating macrophages (Ki67, green; CD68, red) in aortic root plaques. Arrows indicate colocalization and white dashed-line indicates lesion border. Scale bar, 100 µm, Inset Scale bar, 20 µm. (E) Quantification of red bead-labeled macrophages retained in aortic root plaques of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and postintervention. (F) Quantification of TUNEL⁺ cells in aortic root plaques of of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and postintervention. Data are mean ± SEM. P-values by one-way ANOVA with Tukey’s multiple comparison test (A–D and F) or Kruskal–Wallis test with Dunn’s multiple comparison test (E).

Therapeutic Deletion of Unc5b Increases Efferocytosis and Reduces Plaque Complexity.

To understand whether the reduction in plaque ACs upon Unc5b deletion was due to decreased cell death or increased clearance of ACs by surrounding macrophages, we quantified intralesional efferocytosis as the ratio of free-to-phagocytosed ACs in aortic root plaques. We observed a striking eightfold increase in efferocytosis in ∆Unc5b^MØ mice postintervention compared to baseline (Fig. 4A), whereas efferocytosis was not significantly upregulated in Ctrl^MØ mice (Fig. 4A). To test the role of Unc5b in efferocytosis in vitro, we generated BMDMs from Unc5b^fl/flLyz2^Cre/WT mice, in which Unc5b is constitutively deleted in macrophages (∆Unc5b^BMDM) and compared them to Unc5b competent littermate control mice lacking Cre expression (Ctrl^BMDM) (SI Appendix, Fig. S4). First, we assessed whether deletion of Unc5b impacts phagocytosis, a crucial step in the efferocytic process, by challenging ∆Unc5b^BMDM and Ctrl^BMDM with bovine serum albumin (BSA)-coated polystyrene beads 7 µm in diameter. Using time-lapse differential interference contrast microscopy coupled with fluorescence microscopy, we measured a significant increase in the ability of ∆Unc5b^BMDM to internalize beads after 30 and 60 min, compared to Ctrl^BMDM (Fig. 4B). Interestingly, ∆Unc5b^BMDM were able to internalize several beads per cell, despite their size, suggesting an ability for extensive cytoskeletal rearrangement (Fig. 4B). However, whether the uptake of multiple ACs would be observed in Unc5b-deleted macrophages remained uncertain, given that macrophages degrade the efferocytic load for anaplerotic purposes. To test this, we performed a two-step experiment in which we first loaded macrophages with an excess of apoptotic Jurkat T cells labeled with a transferrable dye for 1 h, then washed away unbound cells, and allowed the macrophages to digest the efferocytic load for 24 h (Fig. 4C). While ∆Unc5b^BMDM and Ctrl^BMDM internalized equal amounts of ACs at 1 h, there was a decrease in the amount of efferocytic debris left in ∆Unc5b^BMDM after 24 h, measured as the amount of fluorescence transferred from labeled ACs to the macrophage (Fig. 4D). This increased degradation of efferocytic load could also be appreciated visually as a redistribution of the fluorescent signal in ∆Unc5b^BMDM compared to Ctrl^BMDM, indicating improved efferocytic processing (Fig. 4D). Increased efferocytic capacity is an important characteristic of pro-resolving macrophages that contributes to plaque regression through clearance of dying cells and debris (13, 14, 37). Consistent with increased efferocytosis upon Unc5b deletion, we observed reductions in necrotic area in the aortic root plaques of ∆Unc5b^MØ compared to Ctrl^MØ mice postintervention and baseline (Fig. 4E). In addition, ∆Unc5b^MØ mice showed a decrease in overall plaque complexity, as measured by Stary scoring, with an increase in class II lesions compared to Ctrl^MØ mice post intervention, at the expense of class IV and V lesions (Fig. 4F). Together, these data suggest that therapeutic targeting of Unc5b expression increases macrophage phagocytic and efferocytic capacity, allowing beneficial remodeling of the plaque.

Fig. 4.

Deletion of Unc5b increases macrophage efferocytosis and reduces plaque complexity. (A) Quantification of the ratio of phagocytosed-to-free TUNEL⁺ apoptotic cells (AC) in aortic root plaques of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and after tamoxifen + chow diet intervention. (B) Representative imaging of latex bead engulfment and quantification of the number of beads engulfed per cell in Unc5b-deleted (∆Unc5b^BMDM) or WT (Ctrl^BMDM) BMDMs. Scale bar, 40 µm. (C) Schematic of experimental design. Cells were pulsed with green-labeled apoptotic Jurkat T cells for 1 h, washed, and incubated for an additional 24 h chase. (D) Representative images and quantification of phagocytosed AC uptake at 1 h and AC-derived fluorescence intensity (relative fluorescence units) remaining at 24 h in ∆Unc5b^BMDM and Ctrl^BMDM. Scale bar, 40 µm. (E) Quantification of plaque necrotic area (outlined in black) in H&E stained aortic root plaques. (F) Histological classification of aortic root plaques according to Stary’s grading (38); II, moderate lesions (foam cells, SMCs, intracellular lipid accumulation); III, preatheroma (foam cells, SMCs, pools of extracellular lipids); IV, atheroma (foam cells, SMCs, large pools of extracellular lipids, necrotic core); V, fibroatheroma (foam cells, SMCs, large pools of extracellular lipids, large irregular necrotic core). (B) Dots are average data per experiment (pool of n = 3); and (D) dots are individual fields of view from three individual experiments (n = 3). Data are mean ± SEM. P-values by one-way ANOVA with Tukey’s multiple comparison test (A and E), two-way ANOVA with Sidak’s (B) or Tukey’s (F) multiple comparison test, or Student’s t test (D).

Pro-resolving T Cells Are Enriched in Plaques after Macrophage-Specific Deletion of Unc5b.

Macrophages play important roles in orchestrating the pro-resolving adaptive immune response during atherosclerosis regression, including by increasing immunoregulatory T cell populations, such as Th2 and Treg cells (39). As our scRNA-seq analysis of the aortic arch identified an enrichment in genes associated with T cell differentiation pathways in Unc5b^– compared to Unc5b⁺ myeloid cells within the plaque (Fig. 1C), we investigated whether induced deletion of myeloid Unc5b altered the accumulation of CD4⁺ T cell populations in the aortic arch of atherosclerotic mice by flow cytometry. Compared to baseline mice, both groups showed an upregulation in the proportion of aortic CD4⁺ T cells postlipid lowering, but this increase was significantly more pronounced in ∆Unc5b^MØ compared to Ctrl^MØ mice (Fig. 5A). Within the CD4⁺ T cell compartment, we identified a decrease in T-bet-expressing Th1 cells and an increase in GATA3⁺ Th2 cells in both ∆Unc5b^MØ mice and Ctrl^MØ mice compared to baseline (Fig. 5 B and C). However, ∆Unc5b^MØ mice had significantly higher levels of aortic GATA3⁺ Th2 cells compared to Ctrl^MØ mice postintervention (Fig. 5C). Furthermore, we observed an increase in aortic FOXP3⁺ Tregs uniquely in ∆Unc5b^MØ mice postintervention (Fig. 5D). Immunostaining of the aortic root, a site of more advanced lesion, showed no difference in CD3⁺GATA3⁺ Th2 cells between the groups (Fig. 5E), but a significant increase in FOXP3⁺ Treg abundance specifically in ∆Unc5b^MØ mice postintervention, confirming our aortic arch flow cytometry data (Fig. 5 F and G). Moreover, FOXP3⁺ Tregs were increased in deep cervical and mediastinal lymph nodes of ∆Unc5b^MØ mice postintervention, which are in close proximity to the atherosclerotic plaque, but not in mesenteric lymph nodes or spleen, suggestive of local priming in the plaque (Fig. 5 H–J). Together, our results show that targeted deletion of Unc5b in macrophage induces the accumulation of pro-resolving T cells that have been shown to promote plaque regression (39, 40).

Fig. 5.

Myeloid Unc5b deletion increases accumulation of pro-resolving T cells in plaques. (A–D) Flow cytometric quantification of aortic arch CD4⁺ T cells (A), T-bet⁺ (Th1) cells (B), GATA3⁺ (Th2) cells (C), and FOXP3⁺ (Treg) cells (D) in Ctrl^MØ and ∆Unc5b^MØ mice at baseline and after TAM + chow diet intervention. (E–G) Representative immunofluorescence staining and quantification of CD3⁺GATA3⁺ Th2 cells (E); and FOXP3⁺ Treg (F and G) in aortic root plaques of Ctrl^MØ and ∆Unc5b^MØ mice at baseline and postintervention. Arrows indicate positive cells and white dashed-line indicates lesion borders. Scale bar, 100 µm, Inset Scale bar, 20 µm. (H–J) Flow cytometric quantification of FOXP3⁺ Treg abundance in the deep cervical and mediastinal lymph nodes (H), intestinal lymph nodes (I), and spleen (J). Data are mean ± SEM. P-values by one-way ANOVA with Tukey’s multiple comparison test (A–I), or Kruskal–Wallis with Dunn’s multiple comparison test (J).

Discussion

Over the last decades, significant reductions in major adverse cardiovascular events associated with atherosclerosis have been achieved through lifestyle modifications and intensive lipid-lowering therapies (41–43). Furthermore, coronary plaque regression and stabilization have been demonstrated following high intensity statins (44–47), ezetimibe (48), or PCSK9 inhibitors (49). Yet, despite their efficacy, such cholesterol-lowering strategies have achieved only minimal reductions in plaque size (50), highlighting the need for novel approaches to address residual plaque burden. Similar to humans, we and others have shown that switching hypercholesterolemic mice with advanced atherosclerosis to a low cholesterol chow diet mirrors lipid-lowering therapies in patients, halting atherosclerosis progression, but is insufficient to appreciably regress plaque size (27, 34, 39, 51). We leveraged this paradigm to uncover factors that contribute to the persistence of plaques after cholesterol normalization in mice and identified Unc5b, a receptor for the neuroimmune guidance molecule netrin-1. Netrin-1 supports inflammation in various tissues, including bone (52–54), adipose tissue (23, 55), and vasculature (12), by fostering macrophage chemostasis and survival, and like Unc5b, its expression in atherosclerotic plaques is maintained after lipid lowering (27), implicating this axis in hindering inflammation resolution. Using a translational approach in mice with advanced atherosclerosis that simulates therapeutic Unc5b intervention in conjunction with lipid management, we showed that targeting Unc5b in myeloid cells reduces atherosclerotic inflammation and reverses plaque burden and complexity. Favorable plaque remodeling occurred independently of changes in plasma lipoprotein cholesterol distribution between the groups. We observed no significant change in the distribution of HDL cholesterol, although HDL enrichment in apolipoprotein E and A1 have also been reported to drive plaque stabilization during regression (56). By tracing monocyte-macrophage dynamics in the plaque, we establish that targeting Unc5b reduces processes contributing to macrophage accumulation (e.g., monocyte influx and macrophage retention) and enhances pathways that enable tissue repair (e.g., macrophage efferocytosis, regulatory T cells). Our findings identify the expression of Unc5b by macrophages as a factor that impedes the resolution of inflammation in plaques after lipid lowering and demonstrate the promise of targeting the netrin-1-Unc5b axis to regress atherosclerosis.

Targeting inflammation has emerged as a promising adjunct therapeutic approach to reduce cardiovascular risk, as observed in the CANTOS trial of IL-1β antibody treatment (57). However, the increased risk of fatal infection associated with blocking IL-1β limits its clinical use, highlighting the need for approaches that target maladaptive inflammation in the plaque, while still protecting the host from pathogens (50). Netrin-1 is a secreted factor that can be recognized by distinct receptors on a variety of cell types (58). As a result, netrin-1 can have dichotomous roles in acute and chronic inflammatory settings. For example, while netrin-1 expression by macrophages promotes inflammation in atherosclerotic plaques and adipose tissue in obesity, netrin-1 expression by endothelial or epithelial cells during acute inflammation, or its intravenous infusion, can protect tissues from leukocyte infiltration through receptors A2BAR and DCC (reviewed in ref. 58). Moreover, in atherosclerotic plaques, secreted netrin-1 can also alter the behavior of smooth muscle cells via the receptor neogenin (12). Thus, strategies targeting specific netrin-1 receptors may offer more tailored approaches to selectively block its deleterious effects, as shown herein. We previously showed that Unc5b is the primary netrin-1 receptor on macrophages responsible for mediating macrophage chemostasis and survival (12, 23, 24, 54), and in the current study, we show that Unc5b also inhibits macrophage efferocytosis thus impairing inflammation resolution within the plaque. As netrin-1 is the only known ligand of Unc5b, these findings position Unc5b as a compelling target for therapeutic intervention in the plaque.

At the cellular level, plaque regression requires the coordinated action of numerous pathways to reshape the plaque’s immune landscape to decrease the burden of pro-inflammatory macrophages and enrich for inflammation dampening T cell subsets (7, 39, 59). Why this pro-resolving reorganization fails to occur in atherosclerosis, despite normalization of cholesterol levels, remains an area of intense study. Efferocytosis, initially considered a basic feature of pro-resolving macrophages, has emerged as a central regulator of the regressing plaque not only through the clearance of dead and dying cells but also by fueling the production of anti-inflammatory cytokines that promote tissue repair and Treg expansion (60, 61). Here, we show that Unc5b decreases macrophage efferocytosis and particle engulfment in vitro and that its targeted deletion in vivo enhances plaque efferocytosis to reduce AC accumulation. These findings are consistent with the role of the netrin-1-Unc5b axis as an inhibitor of the cytoskeletal machinery through downregulation of RAC1 activity (20), which is required for the engulfment process and rapid particle internalization (62). Ultimately, the increased efferocytic ability of Unc5b-deficient macrophages was not solely attributed to phagocytic capacity but also faster processing of the efferocytic load. Interestingly, efferocytic load digestion provides long-chain fatty acids required to fuel oxidative phosphorylation, which enables the production of the anti-inflammatory cytokine IL-10 (63). In parallel, AC-derived nucleotides trigger signaling events that induce the proliferation of IL-10- and TGFβ-producing macrophages, reinforcing pro-resolving pathways (37). Our results indicate that sustained expression of Unc5b in plaques impairs the efferocytic process, thus interfering with macrophage-centric pro-resolving actions that enable tissue repair and plaque regression (60, 61).

Integrated approaches combining cytometry and scRNA-seq have revealed that human and mouse atherosclerotic plaques also contain a highly diverse T cell compartment (64, 65). Macrophages are important regulators of T cell differentiation through their roles in antigen-presentation and cytokine secretion (66, 67). Pro-resolving macrophages can direct T cell differentiation toward the alternatively activated Th2 lineage through IL-4 signaling (67, 68) and toward the immunomodulatory Treg lineage via IL-10 and TGFβ (67–70). Tregs in particular, are essential to moderate the immune response and maintain immune tolerance, including in atherosclerosis where they are required to induce the regression of established plaques (39, 40, 71). By leveraging scRNA-seq of immune cells isolated from aortic arch plaques of atherosclerotic mice with either elevated or normalized cholesterol levels, we show that Unc5b expression in myeloid cells is associated with enriched expression of genes associated with T cell differentiation pathways, suggestive of an important cross talk. Notably, tamoxifen-induced deletion of Unc5b in macrophages in vivo resulted in an increase in the number of Tregs, and to a lesser extent Th2 cells, in plaques. Plaque-specific cross talk mechanisms may be responsible for these observations, as the increase in pro-resolving macrophages and efferocytosis upon Unc5b deletion would be expected to stimulate local production of IL-10 and TGFβ (37, 63), which can promote Treg differentiation (66–70). Consistently, we found that Tregs appear to be locally primed by Unc5b-deficient macrophages within the vicinity of the lesion, as their frequency is only increased within the plaque and the adjunct lymph nodes but not systemically in the thymus, spleen, or mesenteric lymph nodes. Our data showing upregulated efferocytosis and increased Treg presence in the plaque following macrophage-specific deletion of Unc5b suggest that therapeutic targeting of the netrin-1-Unc5b axis drives a macrophage-orchestrated reorganization of the plaque immune cell landscape that favors a pro-resolving environment in the plaque.

Taken together, our studies identify Unc5b as a macrophage-specific driver of inflammation in atherosclerosis that persists after lipid lowering. In mice, a combinatorial approach of reverting hypercholesterolemia and deleting Unc5b, specifically in monocytes and macrophages, drives a macrophage-mediated reorganization of the plaque immune cell landscape that promotes atherosclerosis regression. Notably, targeting Unc5b in macrophages stimulated beneficial remodeling of advanced plaques, including reductions in macrophage burden, AC accumulation, necrotic area, and plaque complexity. These favorable changes are indicative of a more stable plaque phenotype, however, it should be noted that mice do not typically exhibit plaque rupture and myocardial infarction (72). Future studies in other models of atherosclerosis, such as the tandem stenosis model that better replicates plaque instability and rupture in humans (73) will be key to establishing the translational relevance of targeting Unc5b for human treatment.

Material and Methods

Expanded methods are found in SI Appendix.

Data Analysis of scRNA-seq of Halted Atherosclerosis Progression in Mice.

scRNA-seq of CD45⁺ cells isolated from the aortic arch plaques of Ldlr^–/– mice fed a western diet for 14 to 16 wk (baseline) and subsequently switched to a diet for 4 wk to halt atherosclerosis progression (halted progression) was performed as reported (27). Data were extracted from GSE253555 for baseline and GSE246316 and GSE161494 for halted progression. Mice from those studies that received additional pro-resolving treatments were removed from the analysis. Data alignment, preprocessing, filtering, normalization, dimensionality reduction, and visualization were performed according to protocols established in the respective publications (27, 74). A total of seven distinct populations were identified using Louvain clustering on 13,397 cells, with further subclustering on myeloid cells (monocytes, macrophages, dendritic cells, and neutrophils). Atherosclerotic plaque immune cells were annotated based on expression of canonical gene markers identified in mouse atherosclerosis (65). The presto package (v1.0.0) was used to identify differentially expressed genes between cell types and Ingenuity Pathway Analysis (Qiagen) was used to perform pathway enrichment analyses.

Mouse Studies.

Experimental procedures were in accordance with the US Department of Agriculture Animal Welfare Act and the US Public Health Service Policy on Humane Care and Use of Laboratory Animals and approved by NYU’s Institutional Animal Care and Use Committee. Mice were housed in a specified pathogen-free facility. Analyses of mouse experiments were blinded through numerical coding of samples. Unc5b floxed (Unc5b^fl/fl) mice on a C57BL6 background were provided by the University of Utah and Dean Li (current address: Merck Research Laboratories) and crossed with C57BL6 Cx3cr1^CreERT2/WT mice (The Jackson Laboratory, Stock# 020940) to generate Cre-positive experimental Unc5b^fl/flCx3cr1^CreERT2/WT (Unc5b^BL) and Cre-negative Unc5b^fl/flCx3cr1^WT/WT littermate control mice (Ctrl^BL). In the treatment cohort, administration of tamoxifen results in Cre-mediated deletion of Unc5b in monocytes and macrophages of Unc5b^fl/flCx3cr1^CreERT2/WT mice (hereafter ∆Unc5b^MØ) and wild type levels of Unc5b in monocytes and macrophages of Unc5b^fl/flCx3cr1^WT/WT mice (hereafter Ctrl^MØ).

Atherosclerosis Analyses.

Advanced atherosclerosis was induced in 8-wk-old male mice by adenoviral delivery of D377Y-mPCSK9 by adeno-associated virus (AAV-mPCSK9; 5 × 10¹¹; Penn Vector core, PA, USA) to reduce hepatic low density lipoprotein receptor (LDLR) expression, followed by WD feeding (40% fat kcal, 0.3% cholesterol) for 20 wk. As AAV-mPCSK9 transduction of the liver is inconsistent in female mice as previously described (75), only male mice were included in this study. Mice with total plasma cholesterol levels above 600 mg/dL 3 wk after study initiation were included in the study. Mice that failed to maintain total cholesterol levels above 475 mg/dL after 20 wk of WD feeding were excluded from the analysis. After 20 wk, mice were randomly assigned to either the baseline group or the tamoxifen treated group. The treatment cohort of mice was injected i.p. for 5 d with 75 mg/kg body weight tamoxifen and put onto a chow diet (13% fat kcal, 0% cholesterol) for 4 wk. Deletion of Unc5b in macrophages of mice posttamoxifen was confirmed by quantitative RT-PCR of messenger RNA (mRNA) extracted from BMDMs (27) or elicited peritoneal macrophages isolated by peritoneal lavage 4 d after i.p. injection of 3 mL of thioglycolate, as described (76). At the study endpoint, mice were killed, exsanguinated by cardiac puncture, and perfused with PBS. The heart was separated from the aorta at the root and embedded in optimal cutting temperature (OCT) compound and snap-frozen at −80 °C. The aorta and the tissues were further processed as described below. A detailed list of diets and compounds used is provided in SI Appendix, Table S1.

Flow Cytometric Analyses.

Single-cell suspensions were prepared from aortic arches via digestion in Liberase, hyaluronidase, and DNase I for 15 min at 37 °C using the GentleMacs dissociator (Milteny). Briefly, cells were washed, filtered, and stained for viability and extracellular antigens after blocking of unspecific binding with CD16/32. Cells were stained intracellularly after permeabilization and acquired on an LSRII HTS cytometer (BD Biosciences). For all flow cytometric experiments, compensation beads (ThermoFisher) were used to set compensation. Unstained cells and FMO controls were used as negative controls to set gates. A detailed list of antibodies used for flow cytometry is provided in SI Appendix, Table S2.

Immunohistochemical Staining.

OCT embedded hearts were sectioned throughout the aortic root (6 μm) and stained with hematoxylin and eosin (H&E). For morphometric analyses of lesion area, six sections per mouse, spanning the entire aortic root, were quantified using ImageJ software (https://fiji.sc/). Necrotic core area was quantified using the HALO image analysis platform and HALO AI (v3.6.4134, Indica Labs). The HALO AI model (DenseNet V2 deep learning classifier) was trained by example to classify acellular regions of tissue and measure their area. For immunostaining and histochemical staining, slides were fixed with 4% formaldehyde, permeabilized with 0.1% triton and blocked with 5% BSA or DAKO blocking solution (Agilent). Plaque composition was assessed by staining with BODIPY for neutral lipid, PicroSirius Red for collagen, or with primary antibodies against Ki67 to detect proliferating cells, CD68 for monocytes/macrophages or ACTA2 for smooth muscle cells, followed by appropriate secondary antibodies as described (27). To detect T cell subpopulations, sections were stained with antibodies against either FOXP3 or CD3 and GATA3 and detected using appropriate secondary antibodies before autofluorescence quenching using the TruView kit (VectorLab) according to the manufacturer’s protocol. ACs were stained using TUNEL. Sections processed for immunohistofluorescence analyses were counterstained for nuclei with DAPI. Images were acquired using a BZX-810 all-in-one fluorescence microscope (Keyence), and aortic root plaques were manually measured, with investigators blinded to treatment assignment as described (27). Quantification of staining area was performed based on control, and positive area/cells above threshold were counted in ImageJ. Single-cell counting was performed on all three leaflets within the aortic root and normalized to average plaque size. Representative images were selected to represent the mean value of each condition. A detailed list of primary and secondary antibodies, suppliers and concentrations are provided in SI Appendix, Table S3.

Labeling and Tracking of Blood Monocytes.

In vivo labeling of circulating Ly6C^lo and Ly6C^hi monocytes was performed to assess their recruitment and retention within the atherosclerotic plaque, as previously described (12, 23, 36). Briefly, circulating Ly6C^lo monocytes were pulse-labeled 72 h prior to the end of the experiment by i.v. administration of 250 µL diluted Fluoresbrite polychromatic green-dyed plain microspheres following the manufacturer’s instructions. Similarly, Ly6C^hi monocytes were pulsed-labeled via i.p. injection of dissolved 5-Ethinyl-2′-deoxyuridin (EdU) 72 h prior to the end of the experiment. To assess the retention of macrophages within the plaque, circulating monocytes were labeled with red-dyed Fluoresbrite polychromatic plain microspheres 72 h prior to Tamoxifen injection. The bead and EdU labeling efficiency were confirmed 24 h postinjection by flow cytometry. At the end of the experiment, EdU-positive cells in the aortic roots were stained using the Click-iT EdU Imaging Kit Alexa Fluor 647 nm-azide (Invitrogen) according to the manufacturer’s instructions. The aortic roots were imaged on a Leica SCN400F slide scanner, and EdU-positive and green and red bead-labeled cells within the plaque were quantified using ImageJ.

Phagocytosis and Efferocytosis In Vitro.

Unc5b-deficient BMDM (∆Unc5b^BMDM) and control (Ctrl^BMDM) BMDM were prepared from bone marrow of Unc5b^fl/flLyz2^Cre/WT mice and Unc5b^fl/flLyz2^WT/WT littermates, respectively, as we described (39, 55). To assess phagocytosis, 7.32 µm beads were coated with BSA following the manufacturer’s protocol and BMDMs were incubated with a 1% bead solution for 0 to 60 min, washed extensively, fixed with 4% paraformaldehyde, and counterstained with DAPI. Macrophage bead uptake was counted automatically using ImageJ and normalized to cell number, with five fields of view per time point from a pool of three mice. To measure efferocytosis and efferocytic load degradation, Jurkat T cells were labeled with 10 µM of transferable CellTracker Green 5-Chloromethylfluorescein diacetate (CMFDA) dye (Life Technologies, #C2925) for 30 min and made apoptotic by treatment with 1 µM staurosporine (16 h). BMDMs were cocultured with apoptotic Jurkat cells at a 1:5 ratio (BMDM:Jurkat) for 1 h, washed extensively, and allowed to degrade their efferocytic load for an additional 24 h before being fixed with 4% paraformaldehyde and counterstained with DAPI. Macrophage uptake and degradation of apoptotic Jurkat cells were measured by intact Jurkat cell counting at 1 h and total fluorescence content per macrophage at 24 h from a triplicate of eight fields of view in three individual animals and normalized to macrophage cell number using ImageJ.

Statistics.

Statistical analyses were performed using GraphPad Prism 9.5. Gaussian distribution was determined using the Shapiro–Wilk or, respectively, D’Agostino–Pearson normality test. Data determined to be parametric were analyzed by unpaired two-tailed Student’s t test to compare two independent groups or more than two groups, by two-way ANOVA test followed by Tukey’s HSD or Sidak’s post hoc multiple comparison analysis. Data determined to be nonparametric were analyzed by Mann–Whitney U test for two group comparison or a Kruskal–Wallis test with post hoc Dunn’s test for more multiple comparison on more than two groups. Outliers were detected and removed using the ROUT outlier test with an false discovery rate (FDR) of 1%. Experimental data are reported as mean ± SEM, and statistical significance was established at P ≤ 0.05.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

scRNA-seq datasets used in this study have been deposited in the Gene Expression Omnibus and are available under accession number GSE253555 (77) and GSE246316 (78). All other data are included in the article or SI Appendix.