Mechanisms and consequences of myeloid adhesome dysfunction in atherogenesis

Abstract

Aims

In the context of atherosclerosis, macrophages exposed to oxidized low-density lipoproteins (oxLDLs) exhibit cellular abnormalities, specifically in adhesome functions, yet the mechanisms and implications of these adhesive dysfunctions remain largely unexplored.

Methods and results

This study reveals a significant depletion of Kindlin3 (K3) or Fermt3, an essential component of the adhesome regulating integrin functions, in macrophages located within atherosclerotic plaques in vivo and following oxLDL exposure in vitro. To examine the effects of K3 deficiency, the study utilized hyperlipidaemic bone marrow chimeras devoid of myeloid Kindlin3 expression. The absence of myeloid K3 increased atherosclerotic plaque burden in the aortas in vivo and enhanced lipid accumulation and lipoprotein uptake in macrophages from Kindlin3-null chimeric mice in vitro. Importantly, re-expression of K3 in macrophages ameliorated these abnormalities. RNA sequencing of bone marrow-derived macrophages (BMDM) from K3-deficient mice revealed extensive deregulation in adhesion-related pathways, echoing changes observed in wild-type cells treated with oxLDL. Notably, there was an increase in Olr1 expression [encoding the lectin-like oxidized LDL receptor-1 (LOX1)], a gene implicated in atherogenesis. The disrupted K3–integrin axis in macrophages led to a significant elevation in the LOX1 receptor, contributing to increased oxLDL uptake and foam cell formation. Inhibition of LOX1 normalized lipid uptake in Kindlin3-null macrophages. A similar proatherogenic phenotype, marked by increased macrophage LOX1 expression and foam cell formation, was observed in myeloid-specific Itgβ1-deficient mice but not in Itgβ2-deficient mice, underscoring the critical role of K3/Itgβ1 interaction.

Conclusion

This study shows that the loss of Kindlin3 in macrophages upon exposure to oxLDL leads to adhesome dysfunction in atherosclerosis and reveals the pivotal role of Kindlin3 in macrophage function and its contribution to the progression of atherosclerosis, providing valuable insights into the molecular mechanisms that could be targeted for therapeutic interventions.

Time of primary review: 41 days

1. Introduction

Atherosclerosis is a polyaetiological pathological process resulting in the genesis and progression of cardiovascular disease (CVD). According to the 2023 American Heart Association report, CVD was the underlying cause of 928 741 deaths in the USA in 2020 and remained the number one cause of death regardless of race or sex.¹ Accumulation of oxidized low-density lipoproteins (oxLDLs) in blood² as well as within the arterial wall is a hallmark of the atherosclerotic process.^3–5 OxLDLs have a number of well-documented biological activities impacting cells in the arterial wall as well as in the bloodstream.^6,7 OxLDLs activate the inflammasome in various cells⁸ and up-regulate the adhesion molecules and proinflammatory chemokines in the endothelium, thereby promoting monocyte recruitment within the vascular wall.⁹ For invading monocytes and macrophages, the uptake of oxLDL via scavenger receptors leads to a broad spectrum of consequences, including profound changes in gene expression, phenotypic changes, enhanced monocyte differentiation, foam cell formation, induction of apoptosis, and other.¹⁰ The lipid-laden macrophages lose adhesion and polarity, fail to egress the lesions, and enhance plaque growth.^7,11,12 Plaque regression, on the other side, is accompanied by an increase in the migration of plaque-associated macrophages.^13,14 OxLDLs stimulate the expression of genes involved in the inflammatory and oxidative stress response or in cell cycle regulation and others.^15–18 The transcriptomic studies of the plaque-associated macrophages during plaque regression revealed up-regulation of the genes for contractile proteins and down-regulation of ‘protocadherin’ family members that promote cell–cell adhesion and for cell cycle/cell division gene families.^19,20 The mechanisms of macrophage failure to egress lesions are poorly understood but include oxLDL-induced expression of factors inhibiting migration^13,21 and suppression of LXR-CCR7-dependent migration of plaque-associated CD68⁺ cells.²² Another proposed mechanism is the inhibition of macrophage migration by oxLDL via CD36-mediated signalling to the cytoskeleton.^23,24

Firm adhesion of myeloid cells to vascular walls and subsequent transmigration depend on the activation of cell adhesion receptors—integrins.²⁵ Genetic deletion of the primary myeloid integrin subunit β2 (CD18) in bone marrow-derived cells, on an low-density lipoprotein receptor (LDLR)-null background, accelerated initial lesion formation but restricted their further growth.²⁶ However, the deletion of its key partner, αm (CD11b) subunit on the same background showed no impact.²⁷ The use of antibodies to block integrins in animal models has been shown to delay and/or reduce the adhesion of monocytes to the vascular wall,^28,29 yet human clinical trials aimed at treating myocardial infarction have yielded unsatisfactory results.³⁰ This suggests that different integrin subunits may have varying roles in the development of atherosclerosis, and their functions could differ between the early and late stages of the disease. Furthermore, the apparent inconsistencies in these findings may stem from the role of integrins not just in monocytes and macrophages but also in other immune cells such as neutrophils, lymphocytes, and mast cells, indicating a multifaceted nature of atheroprogression.

Integrin functions on all blood-derived cells are controlled by the direct binding of two main adapters, Talin and Kindlin (FERMT) to the integrin cytoplasmic domain.^31,32 The lack of Kindlin3 in humans results in a devastating disorder characterized by bleeding, immune problems, and bone abnormalities.^32,33 Deletion of Kindlin3 in myeloid cells leads to severe adhesome deficiency—loss of cell adhesion, spreading, polarity, and migration.^34,35 Kindlin3 also regulates membrane-to-cortex attachment and co-ordinates the membrane mechanics.^36,37

Since Kindlin isoforms exhibit tissue specificity—Kindlin2 is predominantly expressed in vascular cells, while Kindlin3 is characteristic of bone marrow-derived cells—the composition and stability of atherosclerotic plaques can be evaluated based on the relative presence of these Kindlins. Specifically, an increased presence of Kindlin2 indicates a plaque enriched with smooth muscle cells (SMCs), whereas a higher level of Kindlin3 correlates with macrophage markers, marking unstable macrophage-rich atherosclerotic plaques.³⁸ Aside from the aforementioned study, research on Kindlins in atherosclerosis as well as cell-specific analyses of Kindlins in proatherosclerotic conditions remains sparse. Nonetheless, a translational study by Kraemer et al. has highlighted that the level of Kindlin3 protein is significantly decreased in the platelets of patients experiencing myocardial infarction, in contrast to control samples due to proteolysis.³⁹

Here, we utilized multiple in vitro approaches as well as several Kindlin3 and integrin deficient and mutant hyperlipidaemic mouse models to address the mechanisms of adhesome dysfunction in atherogenesis.

2. Methods

2.1. Animals

Animal experimental procedures were performed in accordance with National Institutes of Health (NIH) guidelines on animal care, and all protocols were approved by the Animal Care Committee at Cleveland Clinic (protocol#00002290). Euthanasia was conducted by ketamine/xylazine overdose according to the Cleveland Clinic IACUC protocol (ketamine 300 mg/kg/xylazine 30 mg/kg). C57BL/6, Cx3cr1-CreER, Apoe^−/−, Ldlr^−/−, and CD18 (Itgb2) hypomorph mice were obtained from the Jackson Laboratory. K3^int and K3^hypo mice with knockin of mutant Kindlin3 Q597W598/AA have full and low (∼15%) expression of mutant K3 protein, respectively. K3KO (CX3CR1-creK3f/f), integrin β1KO mice were described previously.^34,40,41 SMC lineage-tracing Apoe^−/− mice [Myh11-CreERT2 Rosa-Stop-YFP (yellow fluorescence protein)] were generated and genotyped as previously described.^42,43 In all animal experiments, male and female mice were assigned to groups randomly (if necessary), and the results were assessed in a blinded manner.

2.2. Cells

Raw 264.7 cells were purchased from ATCC. Mouse peritoneal macrophages (MPM) were isolated from the peritoneal cavity 72 h after 4% thioglycolate IP injection. L929 cells were purchased from ATCC. Bone marrow-derived macrophages (BMDM) were isolated from mice bones (humerus, femur, and tibia). Kindlin3 knockout RAW 264.7 cells were generated using CRISPR–Cas9 technology. Overexpression of hKindlin3 in RAW, K3KO, and K3^int BMDM has been done by lentivirus infection as described previously.^35,36,44 Monocyte progenitors were isolated as described previously.⁴⁰ Briefly, bone marrow cells were aspirated from mice femurs. Red cells were lysed by cold water. T cells and B cells were removed by negative magnetic cell sorting using anti-CD90 and anti-CD45R immunomagnetic beads. Cell purity was tested by flow cytometry using the antibody against CD11b and reached >90%.

The lentivirus infection was performed in accordance with the Cleveland Clinic Institutional Biosafety Committee (protocol #1407).

2.3. LDL isolation, oxidation, and labelling

Human LDL was prepared from the human plasma by sequential density gradient ultracentrifugation as described previously,⁴⁵ and oxLDL was generated as described earlier.⁴⁵ Fluorescently labelled LDL and oxLDL were generated using Alexa-Fluor-488-SDP according to the manufacturer’s instructions. Whole blood was drawn from healthy volunteers by venipuncture at the Cleveland Clinic Foundation (CCF), Cleveland, OH, USA, after written informed consent according to the Declaration of Helsinki. The study was approved by the CCF Institutional Review Board (IRB#3982).

2.4. Atherosclerosis studies on APOE^−/− background

2.4.1. Adoptive transfer of wild-type and K3-mutant monocytes to the Apoe^−/− mice

Isolated wild-type (WT) monocyte progenitors were labelled with a near-infrared fluorescent dye Vivo Track 680, and K3^hypo monocytes were labelled with another near-infrared fluorescent dye, DiR (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindotri-carbocyanine Iodide). An equal amount of labelled WT and K3^hypo cells were mixed, and 2.5 × 10⁶ cells per group were injected in the tail vein of recipient Apoe^−/− mice that had been fed a Western diet for 20 weeks. Three days later, mice were sacrificed, and aortas were isolated and analysed using an IVIC Spectrum CT Imaging system (PerkinElmer). The intensity of fluorescent signals was normalized to the number of fluorescently labelled cells incubated in vitro during the experiment. To control for the effect of dye on monocyte migration in the separate experiment, WT monocytes were labelled with DIR and K3^hypo monocytes with Vivo Track 680.

2.4.2. Immunostaining

The brachiocephalic arteries were isolated as previously published.⁴⁶ Briefly, SMC lineage-tracing Apoe^−/− mice (Myh11-CreERT2 Rosa-Stop-YFP) were injected with tamoxifen [1 mg in 100 µL peanut oil (Sigma-Aldrich, St Louis, Missouri)] over 10 days starting at 6 weeks of age to activate SMC lineage tracing, followed by Western diet feeding for 18 weeks. At the end of the experiment, mice were euthanized and perfused with 4% paraformaldehyde (PFA). Brachiocephalic arteries were dissected, embedded in paraffin, and sectioned at 10 µm thickness from the aortic arch to the right subclavian artery. Sections were stained with primary antibodies: Anti-Mac-2, Kindlin3, anti-GFP antibody, and DAPI solution followed by incubation with the corresponding secondary antibody and mounted with a ProLong™ Diamond Antifade Mountant media. The appropriate isotype match non-immune antibodies were used as a negative control. Images were acquired using a Leica DM2500 confocal microscope and analysed with ImageJ.

2.5. Atherosclerosis studies on LDLR^−/− background

2.5.1. Bone marrow transplantation

1 × 10⁷ BMDM from 6-week-old Cx3cr1-Cre^ER mice, 6-week-old Cxcr1-Cre^ERKindlin3^f/f (K3KO) mice, or 6-week-old Itgβ1KO mice were transplanted into 900 rad irradiated sex-matched Ldlr^−/− mice. After 3 weeks, mice were injected with tamoxifen daily for 1 week to excise Kindlin3 in K3KO-derived bone marrow transplants. Tamoxifen was administrated to all experimental groups, including controls. Tamoxifen was shown to reduce serum cholesterol levels and diminish atherosclerotic plaque formation.⁴⁷ As shown in Supplementary material online, FigureS2F, there was no difference in plasma cholesterol levels between our experimental groups. Then, mice were fed a high-fat/high-cholesterol diet for 16 weeks. Mice were then euthanized, and BMDM, MPM, and plasma were collected. Hearts and aortas were harvested after perfusion with 4% PFA.

2.5.2. Atherosclerotic lesion analysis

For en face analysis, the entire aorta, extending 5–10 mm after bifurcation of the iliac arteries and including the subclavian right and left common carotid arteries, was removed, dissected, and stained with 0.3% Oil Red O/60% isopropanol. Lesion formation was evaluated using ImageJ software. Aortic root lesion analysis was performed as previously described.⁴⁸ Briefly, hearts with aortic roots were removed from mice and fixed in the 10% phosphate-buffered formalin. The top half of the heart was embedded in an OCT compound (Tissue-Tek) at −20°C. 10 μm sections through the aortic sinus were cut on a cryostat and placed on microscope slides. The slides were stained with Oil Red O and haematoxylin. Lesion areas were quantified in six sections at 80 μm intervals using Image-Pro software.

2.5.3. Plasma lipoprotein separation by fast protein liquid chromatography

Plasma lipoprotein cholesterol profiles were acquired using fast protein liquid chromatography as previously described.⁴⁹ Briefly, samples diluted in column buffer 1:10 were loaded to tandem Superose 6 HR columns (Healthcare, Piscataway, NJ, USA), and the eluant was incubated with Infinity Cholesterol (Sigma) reagent. Absorbance was read at 505 nm to determine the distribution of cholesterol among lipoprotein fractions. Very low-density lipoprotein (VLDL), intermediate-density lipoprotein (IDL), LDL, and HDL were identified by their coelution with hamster lipoproteins isolated by sequential ultracentrifugation.

2.6. Lipoprotein uptake and neutral lipid accumulation analysis

2.6.1. Foam cell formation

Cells were incubated with 50 μg/mL LDL or 50μg/mL oxLDL at 37°C for 24 h, washed with PBS, then fixed with 4% PFA at 37°C for 30 min, stained with Oil Red O, and counter-stained with Harris-modified haematoxylin as described previously.⁴⁵

2.6.2. LOX1 blocking experiments

MPM were isolated as described above, cultured overnight, and pre-treated with goat anti-mouse LOX1/OLR1 antibody 10 µg/mL or PBS for 30 min, washed and incubated with 50 µg/mL of oxLDL for 24 h at 37°C with 5% CO₂ followed by Oil Red O staining and counter-stained with Harris-modified haematoxylin, or pre-treated with goat anti-mouse LOX1/OLR1 antibody 10 µg/mL or PBS for 30 min, washed, and incubated with DiI-OxLDL 18.7 µg/mL for 3 h, washed, fixed with 4% PFA for 10 min, washed, and blocked in 10% goat serum in PBS for 20 min followed by incubation with wheat germ agglutinin (WGA) for 30 min at +37°C and Hoechst 33342 for 10 min. Images were acquired with an Olympus IX83 coupled with Cellsens Dimension software and analysed with ImageJ.

2.6.3. Lipoprotein uptake

Cells were incubated with 3 μg/mL Alexa-Fluor-488-SDP-LDL or at Alexa-Fluor-488-SDP-oxLDL at 37°C for 2 h. Cells were then washed repeatedly and analysed using a BD FACS flow cytometer (FACSAria II, BD Biosciences, San Jose, California) and FlowJo software.

2.7. RNA-sequencing analysis

BMDM cells were isolated from C57/BL6 mice bone marrow and incubated for 7 days with L929-conditioned media. Cells were incubated with or without 50μg/mL oxLDL at 37°C for 24 h, trypsinized, and pelleted by centrifugation, and cell lysate mRNAs were collected.

BMDM cells were isolated from C57/BL6 mice and K3^int mice bone marrow, differentiated for 7 days using L929-conditioned media, and incubated with and without 50μg/mL oxLDL at 37°C for 24 h.

RNA sequencing was performed by Hiseq 4000, 50 million reads per sample through Quick Biology company (Quick Biology Inc., Pasadena, California). The reads were first mapped to the latest UCSC transcript set using Bowtie2 version 2.1.0, and the gene expression level was estimated using RSEM v1.2.15. Trimmed mean of M-values was used to normalize the gene expression. Differentially expressed genes (DEGs) were identified using the edgeR program. Genes showing altered expression with P < 0.05 and more than 1.5-fold changes were considered differentially expressed.

2.8. Western blotting analysis

Cells were lysed with 1% SDS lysis buffer, proteins resolved on 10% SDS-PAGE gel (Bio-Rad) and transferred to polyvinylidene fluoride membranes (Millipore). After blocking with 5% non-fat dry milk in TTBS (0.2 M Tris,⁵⁰ 1.5 M NaCl, 0.1% thimerosal, and 0.5% Tween 20), membranes were incubated with primary antibody overnight followed by incubation with secondary HRP-linked antibodies. Protein expression was detected by ECL Detection Reagent (ThermoFisher Scientific, Waltham, Massachusets). The following antibodies were used for western blotting: GAPDH (D16H11) rabbit mAb, rabbit Kindlin3, rabbit β-actin, mouse LOX1/OLR1, integrin beta 2/CD18, anti-integrin beta1, mouse dectin-1/CLEC7A, mouse MMR/CD206, mouse CD36/SR-B3, anti-SR-A1, anti-SR-BI, and anti-ABCA1.

2.9. Statistical analysis

All statistical tests were performed in GraphPad Prism 9. The results are represented as the mean ± SEM. The statistical significance of differences between groups was determined using an unpaired two-tailed t-test or two-way ANOVA to compare two groups and the one-way ANOVA followed by Dunnett’s or Tukey’s post hoc analysis to compare more than two groups for normally distributed data. The non-parametric Mann–Whitney U test was used to compare two sample data sets and the Kruskal–Wallis test with Dunn’s post hoc test for 3 or more groups for non-normally distributed data and small sample size (n < 6). Shapiro–Wilk normality and lognormality test was used with n ≥ 6. A P-value <0.05 was considered to be statistically significant.

3. Results

3.1. oxLDLs diminish Kindlin3 levels in vitro and within atherosclerotic lesions in vivo

To evaluate whether the transcriptomic changes of macrophages under oxidative stress conditions include adhesome-related genes, we performed bulk RNA-seq analyses of mouse BMDM exposed to oxLDL. A total of 463 genes were significantly changed in the oxLDL-treated BMDM compared with control cells, among them the adhesome-related genes, including integrins, were dysregulated (Figure 1A). Pathway enrichment analyses revealed that macrophage adhesion, migration, phagocytosis, lipid transport, and inflammatory response were significantly dysregulated in oxLDL-treated cells (Figure 1B left panel). Similar trends were found previously in human macrophages exposed to oxLDL⁵² (see Supplementary material online, FigureS1A). In vivo, pathway analysis of published single-cell RNA-seq data on the human atherosclerotic lesions collected from 46 patients after carotid endarterectomy⁵¹ revealed that macrophage adhesion, migration, phagocytosis, lipid transport, and inflammatory response were significantly dysregulated in patients with recent cardiovascular events (CE)—stroke, transient ischaemic attack—compared with patients without CE (Figure 1B right panel). Interestingly, the expression of the Kindlin3 gene Fermt3 remained unchanged upon oxLDL treatment (Figure 1A). However, oxLDL induced a decrease in the Kindlin3 protein level in a time- and concentration-dependent manner in three types of murine macrophages—BMDM, thioglycolate induced MPM and Raw 264.7 cells (Figure 1C, 1D, Supplementary material online, FigureS1B, S1C). Control native LDL had no effect on Kindlin3 protein levels (Figure 1C). Furthermore, immunostaining of the atherosclerotic brachiocephalic artery cross sections from Apoe^−/− mice fed a Western diet for 18 weeks (advanced stage of atherosclerosis) showed a lower Kindlin3 staining intensity in the LGALS3-positive intraplaque cells compared with the lesion shoulder and 30 µm fibrous cap area. LGALS3 is not an exclusively macrophage marker in that phenotypically modulated SMCs express it within atherosclerotic.^42,53 Therefore, to ensure the macrophage origin of LGALS3-positive cells, we stained brachiocephalic artery cross sections from the SMC lineage-tracing Apoe^−/− mice (Myh11-CreERT² Rosa-stop-YFP Apoe^−/−)⁴³ fed a Western diet for 18 weeks (Figure 1E, 1F, Supplementary Figure 1D). The immunostaining demonstrates that the shoulder LGALS3+ cells expressing high levels of Kindlin3 were non-SMC-origin (presumably macrophages).

Figure 1.

Atherosclerosis-induced diminish of macrophages adhesome and down-regulation of Kindlin-3. (A) RNA-seq analysis of BMDM cells after treatment with 50μg/mL oxLDL vs. non-treated cells. BMDM cells were isolated from C57/BL6 mice bone marrow and induced for 7 days by L929-conditioned media, incubated with and without 50μg/mL oxLDL at 37°C for 24 h. Heatmap shows log₂counts of DEG. (B) Pathway analysis of DEG demonstrates dysregulated adhesion, migration, transport, and inflammatory pathways in oxLDL-treated BMDM (left panel) and plague-associated macrophages of symptomatic vs. asymptomatic patients after carotid endarterectomy⁵¹ (right panel). (C) Western blot analysis of Kindlin3 in BMDM cells incubated with 50μg/mL oxLDL, 50μg/mL LDL, or control buffer at 37°C for 24 h. Quantitative densitometry analysis is shown in the bottom panel. n = 3 biological replicates. (D) Western blot analysis of Kindlin3 in RAW 264.7 macrophages incubated with 50μg/mL oxLDL or control buffer at 37°C for 6 h. Quantitative densitometry analysis is shown in the bottom panel. n = 4 biological replicates. (E) Representative confocal images of brachiocephalic artery of APOE^−/− mice fed with a Western diet for 18 weeks stained with LGALS3, Kindlin3, and DAPI. Scale bar 100 µm. (F) Quantification of fluorescent intensity of Kindlin3 inside the atherosclerotic plaque. n = 5 mice per group. All values are mean ± SEM. Statistical significance was determined by the non-parametric Kruskal–Wallis test followed by Dunn’s post hoc test, Mann–Whitney U test, or two-way ANOVA. *P-value <0.05, ***P-value <0.001.

Overall, these results demonstrate that the dysregulation of adhesome-related processes in macrophages in vitro under oxidative stress conditions or in the settings of atherosclerotic plaque is associated with suppression of Kindlin3 protein level.

3.2. Lack of Kindlin3 in myeloid cells exacerbates atherosclerosis in vivo

Since the Kindlin3 protein was dramatically down-regulated by oxLDL, we next analysed whether the Kindlin3-deficient cells can reach the inflammation sites in artery walls. To this end, we used the adoptive transfer approach in hyperlipidaemic Apoe-knockout mice (Apoe^−/−), a model known to have a high concentration of oxLDL in circulation and lesions.^54,55 As a model of reduced Kindlin3 state, we used K3^hypo mouse strain, a Kindlin3 knockin mice containing a double mutation (Q⁵⁹⁷W⁵⁹⁸/AA) characterized by low level (∼15%) of Kindlin3 expression and diminished integrin binding.^40,56 We compared the accumulation of WT and Kindlin3-hypomorph (K3^hypo) monocytes in the aortas of hyperlipidaemic Apoe^−/− mice, as shown in Supplementary material online, FigureS2A–2D. To this end, CD11b-positive cells were isolated from the bone marrow of WT and K3^hypo mice, labelled with different near-infrared fluorescent dyes VivoTrack 680 (WT) and DiR (K3^hypo), mixed, and injected via the tail vein into Apoe^−/− mice fed a Western diet for 20 weeks. Three days later, the aortas were dissected, and the monocyte infiltration was detected by the IVIC Spectrum CT Imaging system (see Supplementary material online, FigureS2B, S2C). To negate the potential effect of dyes on monocyte migration, in the second set of experiments, the dyes were switched: WT monocytes were labelled with DIR and K3^hypo monocytes with VivoTrack 680 (see Supplementary material online, FigureS2D). Interestingly, the K3^hypo cells, despite a low level of mutant Kindlin3 and defective integrin activation, were still able to reach the inflammation sites in the aortas even though at a rate ≈2.5 times lower than that of control monocytes (see Supplementary material online, FigureS2C). These results are in line with earlier studies on the role of K3 in integrin function and the role of integrins in immune cell extravasation.^33,35,57

To further access the role of Kindlin3 in bone marrow-derived cells in atherogenesis, we used Ldlr^−/− mice, a model with a lipoprotein profile that is closer to humans than Apoe^−/− mice.^58–60 We created the bone marrow chimeric mice by transplanting the bone marrow cells isolated from the inducible myeloid-specific Kindlin3 knockout (Cx3cr1-Cre^ERKindlin3^Flox/Flox) or control (Cx3cr1-Cre^ER) mice into lethally irradiated Ldlr^−/− mice. CX3CR1 rather than LysoM Cre driver was used to minimize the contribution of Kindlin3 on neutrophils and granulocytes.⁶¹ The level of Kindlin3 in Cx3cr1-Cre^ER cells remains unchanged compared with WT (see Supplementary material online, FigureS2E). After 3 weeks of bone marrow reconstitution, mice were injected with tamoxifen daily for 5 days to excise Kindlin3. Mice were then fed a high-fat/high-cholesterol diet for 16 weeks (Figure 2A), and atherosclerotic lesion development in the aortas was assessed. En face analysis of aortas revealed a >2.5-fold increase in the atherosclerotic lesion area in Cx3cr1-Cre^ERKindlin3^Flox/Flox chimeric Ldlr^−/− mice (designated as K3^BMΔ/Δ) compared with Cx3cr1-Cre^ER chimeric Ldlr^−/− mice (designated as K3^BM+/+) (Figure 2B, 2C). Lesion areas in cross sections of the aortic sinuses were also substantially increased (Figure 2D, 2E). The plasma lipid profiles, including plasma cholesterol, VLDL, and IDL levels, were not significantly changed. However, decreased HDL levels were observed in K3^BMΔ/Δ mice and higher LDL/HDL ratios in K3^Δ/Δ mice compared with K3^BM+/+ (see Supplementary material online, FigureS2E).

Figure 2.

Deletion of Kindlin-3 from macrophages promotes atherosclerosis in hyperlipidaemic mouse model due to increased oxLDL uptake. (A) Schematic of the process of bone marrow transplantation. Bone marrow was isolated from the 6-week CX3CR1-Cre^ER (K3^BM+/+) and 6-week CX3CR1-Cre^ER/K3^fl/fl mice (K3^BMΔ/Δ), and 1 × 10⁷ cells were transplanted into 900 rad irradiated sex-matched LDLR^−/− mice. Three weeks later, mice were injected with tamoxifen daily for 5 days to induce Kindling-3 knockout. Then, mice were fed a high-fat/high-cholesterol diet (Envigo Teklad, TD.96121) for 16 weeks. (B, C) Representative images and quantification of lesion areas of en face Oil Red O staining of whole aortas collected from the LDLR^−/− chimeric mice after procedures shown in Figure 1A. Scale bar 5 mm. n = 16 mice for each group. (D) Representative images of Oil Red O and haematoxylin-stained aortic root lesions. Hearts with aortic roots were collected from K3^BM+/+ and K3^BMΔ/Δ mice after procedures in Figure 1A. 10 μm sections were cut through the aortic sinus and stained with Oil Red O and haematoxylin. Lesion areas of six sections at 80 μm intervals were quantified using Image-Pro software. Scale bar 500 mm. (E) Quantification of atherosclerotic lesion area in aortic roots shown in Figure 2D. n = 6 mice for each group. (F) Neutral lipid accumulation in MPM from LDLR^−/− chimeric mice generated as in Figure 2A. Representative images of Oil Red O and haematoxylin and eosin (H&E) staining of the MPM cells that were incubated with 50μg/mL LDL or 50μg/mL oxLDL at 37°C for 24 h. Scale bar 30 mm. (G) Quantification of the ratio of Oil Red O-positive foam cells. n = 10 mice for each group. All values are mean ± SEM Statistical significance was determined by two-way ANOVA, or non-parametric Mann–Whitney U test. **P-value <0.01, ***P-value <0.001.

These results indicate that Kindlin3 in myeloid cells has a surprising atheroprotective role despite being required for monocyte extravasation.

3.3. Kindlin3 deficiency in macrophages promotes lipoprotein uptake

Macrophage lipoprotein uptake contributes to atherogenesis.^62,63 To assess the impact of Kindlin3 deficiency on lipoprotein uptake, we isolated MPM from chimeric mice used in Figure 2A, incubated them for 24 h with LDL or oxLDL, and stained them with Oil Red O to detect neutral lipid accumulation. K3^BMΔ/Δ MPM had significantly higher lipid accumulation compared with K3^BM+/+ MPM (Figure 2F and G). These results were confirmed in BMDM isolated from Cx3cr1-Cre^ER (WT^MϕCreER) and Cxcr1-Cre^ERKindlin3^f/f (K3KO) mice treated with tamoxifen (Figure 3A and B). The re-expression of Kindlin3 in K3KO BMDM using a Ds-red-labelled human Kindlin3-overexpressing vector (Ds-red-hK3) partially normalized the elevated oxLDL uptake (see Supplementary material online, FigureS3A, B, and C). Similarly, Kindlin3 knockout in RAW 264.7 macrophages generated via the CRISPR–Cas9 technology^35,36,44 led to increases in oxLDL but not LDL uptake, and Kindlin3 overexpression using Ds-red-hK3 rescued these effects (Figure 3B and C), thus further solidifying the causative role of Kindlin3 in oxidized lipids uptake.

Figure 3.

Re-expression of Kindlin3 normalizes the lipoprotein uptake in Kindlin3-deficient cells. (A, B) Representative histograms and the quantification of flow cytometry analysis of lipoprotein uptake in CX3CR1-Cre^ER (WT^MϕCreER) and K3KO BMDM after 2h incubation at 37°C with 3 μg/mL Alexa-Fluor-488-SDP-LDL or Alexa-Fluor-488-SDP-oxLDL. n = 4 biological replicates. (C, D) Representative images and the quantification of neutral lipid accumulation in Raw 264.7 cells. WT transduced with Ds-red, WT + Ds-red-transduced K3KO, and K3KO transduced with Ds-red-hK3 Raw 264.7 cells were incubated with 50μg/mL LDL or 50μg/mL oxLDL for 24 h and stained with Oil Red O and HE. n = 16 fields for WT AND K3KO-hK3-oxLDL, 17 for K3KO-LDL, 14 for K3KO-hK3-LDL, 13 for K3KO-oxLDL. Scale bar 10 mm. All values are mean ± SEM. Statistical significance was determined by the non-parametric Mann–Whitney U test or Kruskal–Wallis test followed by Dunn’s post hoc test. *P-value <0.05, ***P-value <0.001.

3.4. Kindlin3/β1 integrin interaction in macrophages regulates the uptake of oxLDL

Kindlin3 plays a crucial role in integrin activation by binding the cytoplasmic domain of several β subunits of integrins.⁵⁷ Considering that Kindlin3 deficiency affects intracellular processes that are not always integrin-dependent,^36,41 we tested whether the Kindlin3–integrin axis is critical for the Kindlin3 role in oxLDL uptake. We used a mouse strain expressing mutant Kindlin3 that lacks a functional integrin-binding site (K3^int). Importantly, the level of mutant Kindlin3 in this strain is similar to the level of Kindlin3 in WT mice.⁴¹ The flow cytometry analysis revealed increased oxLDL uptake in K3^int MPM (see Supplementary material online, FigureS3D and E) as well as in BMDM isolated from the K3^int mice (Figure 4A), a phenotype similar to K3KO cells. Overexpression of hK3 in BMDM isolated from the K3^int mice substantially reversed the phenotype (Figure 4A, 4B). These results suggest that disruption of the Kindlin3/integrin axis is critical for the proatherogenic phenotype.

Figure 4.

The Kindlin3–integrin axis controls lipoprotein uptake in the context of atherosclerosis. (A, B) Representative images and quantification of neutral lipid accumulation in WT transduced with Ds-red BMDM, DS-red-transduced K3^int BMDM, and K3^int BMDM transduced with Ds-red-hK3. Cells were incubated with 50μg/mL LDL or 50μg/mL oxLDL for 24 h and then fixed and stained with Oil red O and H&E. The bar graph shows the per cent of Oil Red O-positive cells. Scale bar 30 mm. n = 4 biological replicates. (C, D) Representative images and quantification of CX3CR1-Cre^ER (WT^MϕCreER), K3KO, Igβ1KO, and Igβ2^hypo MPM. Cells were incubated with 50μg/mL LDL or 50μg/mL oxLDL for 24 h, fixed, and stained with Oil Red O and H&E. The bar graph shows the per cent of Oil Red O-positive cells. Scale bar 10 mm. n = 5 biological replicates for WT^MϕCreER LDL, n = 7 for Igβ1KO oxLDL, and n = 6 for the rest of the groups. (E) Schematic of the process of bone marrow transplantation. Bone marrow was isolated from the 6-week CX3CR1-Cre^ER and 6-week CX3CR1-Cre^ER/Itgβ^fl/fl mice, and 1 × 10⁷ cells were transplanted into 900 rad irradiated sex-matched LDLR^−/− mice. Three weeks later, mice were injected with tamoxifen daily for 5 days to induce Kindling-3 knockout. Then, mice were fed a high-fat/high-cholesterol diet (Envigo Teklad, TD.96121) for 16 weeks. (F, G) Representative images and quantification of aortic lesion areas of en face Oil Red O staining of whole aortas collected from the LDLR^−/− chimeric mice with control and Itgβ1^Δ/Δ bone marrow, after a 1-week tamoxifen treatment and 16 weeks of high-fat/high-cholesterol diet. Scale bar 5 mm. n = 7 mice for each group. All values are mean ± SEM. Statistical significance was determined by the non-parametric Kruskal–Wallis test followed by Dunn’s post hoc test, the non-parametric Mann–Whitney U test, or one-way ANOVA followed by Tukey’s post hoc test. *P-value <0.05, **P-value <0.01, ***P-value <0.001.

We next tested whether and how integrins contribute to the observed phenotype. β1 and β2 integrins are the main integrins in macrophages. Thus, we compared lipid accumulation induced by oxLDL in MPM from WT^MϕCreER, K3KO, Itgβ1KO, and Itgβ2^hypo mice. The Oil Red O staining of oxLDL-treated MPM demonstrates increased lipid accumulation in K3KO compared with WT^MϕCreER MPM as expected. Accumulation of neutral lipids was also significantly increased in Itgβ1KO MPM but not in Itgβ2^hypo MPM (Figure 4C and D). No significant lipid accumulation was induced by native LDL (Figure 4C and D). Interestingly, similar to Kindlin3, the MPM level of Itgβ1 dropped significantly after incubation with relatively low concentrations of oxLDL (see Supplementary material online, FigureS3D). At the same time, levels of Itgβ2 in MPM were not changed by oxLDL.

Based on these data, we assessed the effect of Itgβ1 deficiency in myeloid cells on atherogenesis. Lethally irradiated Ldlr^−/− mice received bone marrow from Cxcr1-CreER^ER (Igβ1^BM+/+) or Cxcr1-Cre^ER-Itgb1^fl/fl mice (Itgβ1^BMΔ/Δ) (Figure 4E). After 3 weeks of bone marrow reconstitution, both control and Itgβ1^Δ/Δ mice were injected with tamoxifen daily for five days to induce excision of Itgb1-Flox allele (Itgβ1^Δ/Δ). Mice were fed with a high-fat/high-cholesterol diet for 16 weeks and then sacrificed, the hearts and aortas isolated, and aortas and aortic roots analysed for the lesion area. Similar to K3^Δ/Δ mice, Itgβ1^Δ/Δ had an increase in aortal lesion coverage compared with control mice (Figure 4E and F), but aortic root staining did not reveal significant differences in the lesion area (not shown). Thus, the deficiency of integrin β1 in macrophages in Ldlr^−/− mice has a significant proatherogenic effect, likely due to increased lipoprotein uptake. However, this effect is less pronounced than that of Kindlin3 deficiency.

3.5. Disruption of the K3/integrin axis leads to dysregulation of multiple adhesome and atherosclerosis-related genes, including Lox1

To assess cellular and molecular mechanisms underlying the anti-atherogenic function of the K3/integrin axis, we evaluated the transcriptomic changes in macrophages lacking Kindlin3-integrin coupling. We performed bulk RNA-seq analysis of BMDM isolated from the WT and K3^int mice in control conditions (WT^Contr and K3^int-Contr) or after incubation with 50 µg/mL oxLDL (WT^oxLDL and K3^int-oxLDL) for 24 h (Figure 5). Analysis of DEGs revealed as expected significant changes in adhesome-, immune-, and inflammation-related genes in the WT^oxLDL/WT^Contr group. These changes were mirrored in the same direction in K3^int-Control/WT^Contr group (Figure 5A). The pathway analysis of DEG in WT^oxLDL/WT^Contr cells also showed a similar to K3^int-Control/WT^Contr dysregulation of focal adhesion-, cell migration-, and cytoskeleton-related pathways (Figure 5B, 5C).

Figure 5.

The transcriptome analysis reveals similarities between WT^oxLDL and K3^int macrophages. (A) The bar graph represents the result of RNA-Seq analysis of BMDM cells after incubation with 50μg/mL oxLDL vs. control cells. BMDM cells were isolated from C57/BL6 and K3^int mice bone marrow and induced for 7 days by L929-conditioned media, incubated with 50μg/mL oxLDL or control buffer at 37°C for 24 h, and analysed. Adhesome-, immune-, and inflammation-related genes changed in the same direction in the WT^oxLDL/WT^Contr and K3^{int−Control}/WT^Contr groups. (B, C) Pathway analysis of WT oxLDL vs. WT untreated cells and K3^int vs. WT cells without treatment. (D) Venn diagram shows overlapping DEG between three groups. (E) Venn diagram shows overlapping pathways between three groups. (F) The bar graph shows relative Class A, B, and E scavenger receptors and transporters Abcg1 and Abca1 mRNA levels in WT^Contr and K3^int-Control cells. (G) Western blot and quantitative densitometry analysis of SR-A1, SR-B1, CD36 (SR-B2), Dectin-1 (SR-E2), MMR/CD206 (SR-E3), and ABCA1 in WT, K3^hypo, and K3^int macrophages. n = 5 biological replicates for CD36 and MMR/CD206; n = 4 biological replicates for SR-B1, dectin-1, SR-A1, ABCA1. All values are mean ± SEM. Statistical significance was determined by the non-parametric Kruskal–Wallis test followed by Dunn’s post hoc test. *P-value <0.05.

Importantly, the RNA-seq analysis revealed that one-third of genes dysregulated in K3^int-Contr cells compared with WT^Contr cells overlapped with genes dysregulated in WT^oxLDL vs. WT^Contr (Figure 5D, Supplementary material online, Table S4). Moreover, gene set enrichment analysis demonstrated up to two-thirds of dysregulated pathways, including adhesome-related pathways, overlapped (Figure 5E), indicating that disruption of Kindlin3/integrin axis mimics, to a certain extent, dysregulation in gene expression and pathways induced in macrophages by oxLDL. These include genes involved in cell adhesion, migration, cell junction, regulation of cytoskeleton organization, immune response, inflammatory response, cytokine production, and transport regulation. Further analysis revealed an overlap of dysregulated genes between K3^int macrophages, WT^oxLDL, and a list of 988 atherosclerosis-related genes.⁶⁴ Strikingly, 94% of genes overlapping between three sets were changed in the same direction in WT^oxLDL and K3^int macrophages. This further supports the conclusion that oxLDL treatment and the K3/integrin axis disruption induce similar gene dysregulation in macrophages. The list included genes involved in regulating inflammation, such as Il1a and Il1r1, intracellular signalling, cell–cell interaction, and other genes known to contribute to atherogenesis (see Supplementary material online, Table S1). Interestingly, the list includes several genes that are reported to change or directly contribute to lipoprotein uptake by macrophages. Expression of Olr1, a gene encoding lectin-like oxidized LDL receptor-1 (LOX1), was significantly increased in K3^int macrophages (see Supplementary material online, Table S2). LOX1 (SR-E1) belongs to the class E scavenger receptor that recognizes and mediates the uptake of modified LDL, including oxLDL.^65,66 In contrast, other SR-E family members, such as Mrs1 (SR-E3) and Clec7a (SR-E2), as well as SR-B and SR-A class members (CD36, Scarb1, Scarb2, Msr1), and transporters Abcg1 and Abca1, did not exhibit significant changes in gene expression in K3^int-Control vs. WT^Contr cells (Figure 5F). We confirmed the RNA-seq data by WB analysis of corresponding proteins level (Figure 5G). Among other proteins, only SR-E member, MMR/CD206 (encoded by the Mrc1 gene) was increased in K3^int cells. MMR/CD206 is known to plays a critical role in the phagocytosis of mannosylated glycoproteins and antigen presentation, but not oxLDL uptake.⁶⁷ The LOX1 protein level was found dramatically up-regulated in K3KO, K3^hypo, and K3^int macrophages (Figure 6A, 6B). Furthermore, a similar increase in LOX1 was observed in Itgβ1KO macrophages (Figure 6C). At the same time, the Itgβ2^hypo cells did not show a statistical difference in LOX1 protein expression compared with WT (Figure 6D).

Figure 6.

Augmented lipid uptake in Kindlin3-deficient and Itgβ1KO macrophages is mediated by LOX1. (A) Western blot and quantitative densitometry analysis of LOX1 in WT, K3^hypo, and K3^int macrophages. n = 6 biological replicates. (B) Western blot and quantitative densitometry analysis of LOX1 in WT^MϕCreER and K3KO macrophages. n = 4 biological replicates. (C) Western blot and quantitative densitometry analysis of LOX1 in WT and Itgβ1KO macrophages. n = 5 biological replicates. (D) Western blot and quantitative densitometry analysis of LOX1 in WT and Itgβ1^hypo macrophages. n = 3 biological replicates. (E, F) Representative images and quantification of WT and K3^int MPM. Cells were pre-treated with anti-LOX-1/OLR1 antibody10 µg/mL or PBS for 30 min and then incubated with 50μg/mL oxLDL for 24 h, fixed, and stained with Oil Red O and H&E. The bar graph shows the relative Oil Red O area per cell. Scale bar 20 µm. n = 18 fields for each group. (G, H) Representative images and quantification of WT and K3^int MPM. Cells were pre-treated with anti-LOX-1/OLR1 antibody10 µg/mL or PBS for 30 min and then incubated with DiI-OxLDL 18.7 µg/mL for 3 h, fixed, blocked, and incubated with WGA for 30 min and Hoechst 33342 10 min. The bar graph shows the relative Dil-oxLDL MFI. Scale bar 20 µm. N = 16 fields for each group. All values are mean ± SEM. Statistical significance was determined by the non-parametric Kruskal–Wallis test followed by Dunn’s post hoc test, the non-parametric Mann–Whitney U test, or one-way ANOVA followed by Holm–Sidak’s post hoc test. *P-value <0.05, **P-value <0.01, ***P-value <0.001.

To assess the role of LOX1 in increased lipoprotein uptake and neutral lipid accumulation in cells lacking the Kindlin3–integrins coupling, we tested an effect of LOX1 blocking antibody on the uptake of DiI-OxLDL and lipid accumulation induced by oxLDL. The Oil Red O staining revealed that LOX1 blockade substantially diminished lipid accumulation by K3^int macrophages (Figure 6E, 6F), indicating that LOX1-mediated uptake of oxLDL is the main or one of the main mechanisms responsible for high lipid uptake in the absence of Kindlin3. At the same time, lipid accumulation in WT macrophages was independent of LOX1 (Figure 6E and F), which is in agreement with low levels of LOX1 in these cells (Figure 6A and B). Immunofluorescent analysis of Dil-oxLDL accumulation mirrored the Oil Red O staining results (Figure 6G and H), thereby confirming the prominent role of LOX1 in Kindlin3-deficient cells.

These findings together with an established role of LOX1 in atherogenesis^66,68 suggest that LOX1 overexpression in macrophages contributes to the proatherogenic phenotype in mice with disrupted K3/Itgβ1 axis.

4. Discussion

Exposure of macrophages to oxLDL in vitro or to oxidized lipids-rich milieu of atherosclerotic lesions induces dysregulation of multiple cellular processes, including adhesome-dependent functions.^12,24 Kindlin3 is indispensable for integrin function and directly involved in all adhesome-dependent processes in haematopoietic cells.^34,35,41,57 In the current study, we show that in vitro exposure to oxLDL and lipid accumulation in vivo in atherosclerotic plaques cause a dramatic reduction in Kindlin3 protein levels in macrophages. Knowing that loss of adhesion and spreading and the overall disturbance of adhesome is one of the hallmarks of lipid-exposed macrophages, we tested the role of the Kindlin3–integrin axis in early and late events of atherogenesis. Despite the reduced rate of initial monocyte recruitment to atherosclerotic plaques, Kindlin3 deficiency in monocytes and macrophages augmented lipoprotein uptake and foam cell formation, thereby promoting the overall proatherogenic phenotype in hyperlipidaemic Ldlr^−/− mice. The absence of Kindlin3 does not significantly affect plasma cholesterol, LDL, or VLDL levels. There is, however, a reduction in HDL levels, which theoretically might also contribute to the overall proatherogenic phenotype of Kindlin3 deficiency (see Supplementary material online, FigureS2F). The examination of RNA-seq data from BMDM isolated from the K3 mutant, which has a disrupted integrin-binding site, uncovered numerous dysregulated genes and pathways related to adhesion. A similar pattern of dysregulation was observed in WT cells following oxLDL incubation. Several of these dysregulated genes, particularly Olr1, which encodes LOX1, are recognized for their direct involvement in atherogenesis. Disruption of the Kindlin3–integrin axis in macrophages resulted in a significant increase in the expression of the LOX1 receptor. Blocking LOX1 normalized the elevated oxLDL uptake, revealing a mechanism behind the high uptake of oxLDL and the formation of foam cells in Kindlin3-deficient cells. A comparable proatherogenic phenotype and elevated macrophage LOX1 were noted in Itgβ1-deficient mice, but not in Itgβ2-deficient mice. This suggests that the absence of Kindlin3/Itgβ1 interaction could be accountable for the observed phenotype.

The reciprocity between lipid uptake and adhesome has been noted in many studies; however, the mechanisms remained largely enigmatic. The lipid-laden macrophages in atherosclerotic lesions lose adhesion and polarity and fail to egress the lesions.^11,13,69 Similar changes were detected in vitro in macrophages exposed to oxLDL.^23,24,70 Diminished macrophage emigration out of atherosclerotic lesions in turn promotes atherogenesis.^11,13,69 The previously suggested mechanism involves a combination of augmented retention¹³ and diminished egression of macrophages.¹¹ Various molecular mechanisms have been suggested, including the inhibition of macrophage migration through oxLDL⁷¹ via CD36-mediated signalling.^23,24 Adhesion and migration of myeloid cells depend critically on Kindlin3, an adaptor protein indispensable for integrin functions.^34,35,41,57 Kindlin3 deficiency in macrophages prevents lamellipodia formation and impairs migration on adhesive substrates, including migration towards the chemoattractant.^34,35

We show that although oxLDL led to various alterations in the macrophage transcriptome, the expression of Fermt3, the gene responsible for Kindlin3, remained largely unchanged in macrophages. This is different from the bulk RNA-seq results of atherosclerotic plaques where Kindlin3 and several macrophage markers such as β2 integrin were up-regulated, thereby reporting on macrophage infiltration.³⁸ OxLDL caused a swift and significant decline in Kindlin3 protein levels, indicating a post-translational nature of this phenomenon. This reduction in Kindlin3 levels was observed in a time- and concentration-dependent manner across three types of murine macrophages, specifically in response to oxLDL. In vivo significance of this finding was confirmed by low Kindlin3 in LGALS3-positive intraplaque cells compared with the lesion shoulder and fibrous cap area in advanced-stage Apoe^−/− mice. Although Kindlin3 functions might be distinct in leukocytes and platelets,⁴⁴ a similar loss of Kindlin3 at the protein level was reported in platelets from patients with myocardial infarction and was attributed to Kindlin3 cleavage.³⁹ Indeed, the cleavage of adhesome components as a mechanism of their regulation has been reported in numerous studies.^72–74

These results highlight that the disturbance in adhesome-related processes in macrophages either exposed to oxLDL in vitro or within the atherosclerotic plaques rich in oxLDL aligns with the decrease in cellular Kindlin3 protein levels. Considering the significant impact of the Kindlin3/integrin axis on all adhesome-dependent functions, our discovery proposes a novel mechanism that interferes with the adhesive and migratory properties of macrophages within atherosclerotic lesions.

The surprising aspect of our findings lies in the protective role of macrophage Kindlin3 against the development of atherosclerotic lesions. Despite its essential role in integrin-dependent trans-endothelial migration—an early step in atherogenesis—we observed a significant decrease in the migration of monocytes with diminished Kindlin3 levels and activity into atherosclerotic lesions (see Supplementary material online, FigureS2A–D). However, the overall impact of an intact Kindlin 3/integrin axis on bone marrow-derived cells proved to be atheroprotective. This was evident in Cx3cr1-CreER Kindlin3^Flox/Flox chimeric Ldlr^−/− mice subjected to a high-fat/high-cholesterol diet for 16 weeks.

Despite the absence of Kindlin3, monocytes were still released into circulation, adhered to the endothelial lining, and accumulated in the lesion—albeit at a significantly reduced rate. This resembles the role of macrophage integrins, particularly β2, which has been demonstrated to control the early stages of macrophage recruitment.^28,75 It appears that the proatherogenic potential of Kindlin3-deficient macrophages compensated for initially reduced adhesion and was sufficient to drive atherosclerotic plaque progression.

The RNA-seq analyses revealed a considerable number of genes displaying significant differential expression in macrophages upon exposure to oxLDL. These genes encompassed various adhesome-related elements, including integrins. Pathway enrichment analyses indicated noteworthy dysregulation in macrophage processes such as adhesion, migration, phagocytosis, lipid transport, and inflammatory response. This aligns with prior findings in human macrophages exposed to oxLDL, where cell adhesion and migration were notably affected pathways,^16,52 and in macrophage foam cells from the LDLRKO mice after the HCHF diet demonstrating suppression of inflammatory-related genes.⁷⁶ Moreover, pathway analysis of published single-cell RNA-seq data from human atherosclerotic lesions obtained post-carotid endarterectomy revealed significant alterations in the adhesive properties of plaque-associated macrophages in patients with recent cardiovascular events compared with asymptomatic individuals.⁵¹ Specifically, macrophage adhesion, migration, phagocytosis, lipid transport, and inflammatory response were substantially dysregulated in patients with cardiovascular events (stroke, transient ischaemic attack) in comparison with those without such events.

The comparison of dysregulated genes and pathways between K3^int and WT^oxLDL macrophages revealed significant parallelism, suggesting that disruption of the K3/integrin axis partially mimics the gene expression and pathway dysregulation induced by oxLDL in macrophages. This includes a multitude of genes implicated in atherogenesis, encompassing cell adhesion, migration, cell junction, regulation of cytoskeleton organization, immune response, inflammatory response, cytokine production, and regulation of transport. Remarkably, 94% of genes overlapping in WToxLDL and K3^int macrophages exhibited changes in the same direction. This strengthens the conclusion that oxLDL and disruption of the K3/integrin axis induce similar gene dysregulation in macrophages. The data suggest that a reduction in Kindlin 3 and disruption of adhesion are responsible for at least a portion of the proatherogenic gene dysregulation observed in lipid-laden macrophages.

The role of integrins in atherosclerosis development has been extensively studied. Notably, the focus on integrin function in macrophages during atherosclerosis was centred on β2 integrins expressed exclusively in leukocytes. While β2 integrin deficiency exhibited a protective effect during early atherosclerosis, it promoted progression at later stages.²⁶

The detailed analysis of α subunits associated with β2 integrin revealed individual effects on atherogenesis. Injection of anti-α_L monoclonal antibodies decreased significantly macrophage adherence and transmigration into the intima of hypercholesterolaemic rats.⁷⁷ α_X deficiency reduced lipid accumulation and macrophage number in atherosclerotic lesions of hyperlipidaemic Apoe^−/− mice.⁷⁸ Integrin α_D-deficiency on the Apoe^−/− background also reduced atherosclerotic lesion formation.⁵⁶ On the other hand, α_M deficiency, at least in female Apoe^−/− mice, led to increased atherosclerotic lesion formation and an augmented number of macrophages in the lesions.⁷⁹ Atherogenesis studies using blocking antibodies and peptides in Apoe^−/− mice demonstrated that integrin α₄ has a proatherogenic role,^80,81 possibly via regulation of macrophage recruitment.⁸² Since Itgb1 knockout mice are not viable, the role of β₁ in atherogenesis is not thoroughly investigated. We show that similar to K3KO, β₁-deficient macrophages accumulated more neutral lipids in foam cell formation assay. Moreover, like K3KO, K3^hypo, and K3^int, Itgβ1KO macrophages up-regulated the major scavenger receptor—LOX1. In contrast to the studies using α₄ subunit blockade, we found that in Ldlr^−/− chimeras with Itgβ1^Δ/Δ bone marrow fed a high-fat/high-cholesterol diet developed more atherosclerotic lesions. Based on these results, we conclude that the intact K3/Itgβ1 axis plays an atheroprotective role at least in part via balanced LOX1 expression.

LOX1 is a scavenger receptor that recognizes modified LDL, including oxLDL.⁶⁵ Under normal conditions, LOX1 is expressed at very low levels in macrophages and, as we showed previously,⁸³ does not significantly contribute to the uptake of modified LDL by MPMs. Nevertheless, during proinflammatory conditions, such as in atherosclerotic lesions, there is an up-regulation in the expression of LOX1, and its role in the uptake of oxLDL by macrophages becomes substantial.^84,85 LOX1 is up-regulated by inflammatory cytokines such as IL-1, IL-6, and TNF-α.⁸⁶ Additionally, oxLDL itself is a potent activator of LOX1 via the NF-κB pathway. Administration of anti-LOX1 antibodies inhibits atherosclerosis,⁶⁶ and LOX1 deficiency is atheroprotective in hyperlipidaemic Ldlr knockout mice.⁸⁷ Its deletion is also associated with a decrease in oxidative stress, inflammatory response, and NO degradation.⁶⁸ In all models with impaired adhesive functions—K3KO, K3^hypo, K3^int, and Itgβ1KO, LOX1 expression was up-regulated, indicating that intact K3/integrin axis and, specifically K3/β₁, restricts LOX1 expression. In our recent study,⁴¹ we demonstrated that the lack of Kindlin3 in myeloid cells changes their adhesive and mechanosensory properties making K3-deficient cells insensitive to the substrate stiffness, thereby leading to up-regulation of TGFβ1 secretion. TGFβ1, in turn, was shown by several studies to promote LOX1 expression in macrophages as well as in vascular cells.^88,89 Since the substrate stiffness is altered in atherosclerotic lesions, stiffness is an important contributing factor in atherogenesis.⁹⁰ It is possible that K3 deficiency causes LOX1 up-regulation via a similar mechanism. Besides, IL-1a, a proinflammatory cytokine involved in the early stages of atherosclerosis,⁹¹ and IL1R1⁹² were up-regulated in K3^int macrophages, suggesting a possibility of an autocrine mechanism. Thus, the disruption of the intact K3/β₁ integrin axis appears to equip macrophages with an extra receptor for the uptake of oxLDL, thereby promoting foam cell formation—a critical early step in atherosclerotic plaque progression.

In summary, our findings indicate that upon migration into the atherosclerotic plaque, macrophages exposed to oxidative conditions may experience a loss of Kindlin3. This loss subsequently disrupts macrophage adhesome functions and induces cytoskeleton changes, augmenting lipid uptake and establishing a positive feedback loop. These results emphasize the pivotal atheroprotective role of Kindlin3 through the Kindlin3–integrin β1 axis, safeguarding myeloid cells from excessive lipid accumulation. This highlights Kindlin3 as a potential novel therapeutic target in CVD.

Translational perspective.

This study uncovers a critical mechanism in atherosclerosis progression, revealing that macrophages exposed to oxidized low-density lipoprotein (oxLDL) lose Kindlin3, a protein essential for cell adhesion. This loss leads to increased lipid accumulation and foam cell formation, contributing to larger atherosclerotic plaques. Reintroducing Kindlin3 restores normal macrophage function, reducing these proatherogenic changes. Kindlin3 deficiency elevates LOX1 receptor levels, enhancing oxLDL uptake. These findings show the potential of targeting the Kindlin3–integrin axis as a therapeutic strategy to prevent macrophage dysfunction and reduce atherosclerosis, offering new avenues for the treatment of cardiovascular disease.

Supplementary material

Supplementary material is available at Cardiovascular Research online.

Authors’ contributions

Conceptualization: T.V.B. and E.A.P.; methodology and investigation: H.L., I.Z., D.G., V.Y., E.A.P., and T.V.B.; data curation: I.Z., H.L., T.D., S.T., O.C., V.Y., and T.V.B.; funding acquisition: T.V.B., E.A.P., and O.C.; resources: D.G., V.Y., E.A.P., and T.V.B.; supervision and writing—review and editing: T.V.B., E.A.P., O.C., and T.D.; and writing—original draft: I.Z., E.A.P., and T.V.B..

Funding

This study was supported by the National Institutes of Health grant R01 HL071625.

Data availability

The RNA-seq data sets are available in the Gene Expression Omnibus GSE254743 and GSE254744.