Abstract
The presence of atherosclerotic plaque vessels is a critical factor in plaque destabilization. This may be attributable to the leaky phenotype of these microvessels, although direct proof for this notion is lacking. In this study, we investigated molecular and cellular patterns of stable and hemorrhaged human plaque to identify novel drivers of intraplaque vessel dysfunction. From transcriptome data of a human atherosclerotic lesion cohort, we reconstructed a co-expression network, identifying a gene module strongly and selectively correlated with both plaque microvascular density and inflammation. Spectrin Beta Non-Erythrocytic 1 (sptbn1) was identified as one of the central hubs of this module (along with zeb1 and dock1) and was selected for further study based on its predominant endothelial expression. Silencing of sptbn1 enhanced leukocyte transmigration and vascular permeability in vitro, characterized by an increased number of focal adhesions and reduced junctional VE-cadherin. In vivo, sptbn1 knockdown in zebrafish impaired the development of the caudal vein plexus. Mechanistically, increased substrate stiffness was associated with sptbn1 downregulation in endothelial cells in vitro and in human vessels. Plaque SPTBN1 mRNA and protein expression were found to correlate with an enhanced presence of intraplaque hemorrhage and future cardiovascular disease (CVD) events during follow-up. In conclusion, we identify SPTBN1 as a central hub gene in a gene program correlating with plaque vascularisation. SPTBN1 was regulated by substrate stiffness in vitro while silencing blocked vascular development in vivo, and compromised barrier function in vitro. Together, SPTBN1 is identified as a new potential regulator of the leaky phenotype of atherosclerotic plaque microvessels.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
In the last decades, our understanding of atherosclerosis has vastly increased, and a myriad of novel insights has been gained on the various players that come together in atherosclerosis, i.e. lipid transportation/metabolism, inflammation, and plaque neovascularization. Atherosclerotic plaque angiogenesis is correlated with increased plaque progression, and has been recognized as a double edged sword in the development of advanced, instable atherosclerotic lesions [1]. On the one end, afflux of oxygen to the lesion could alleviate plaque hypoxia, leading to a reduction in e.g. chronic oxidative stress [2]. Yet, in practice, these plaque vessels have also been associated by us and others with increased permeability, leakage of erythrocytes and lipids, and recruitment of inflammatory cells, both in human and murine lesions [3,4,5,6]. We have previously shown that murine plaque-associated microvessels (vasa vasorum) are more permeable than similarly sized control microvessels, displaying more pronounced leukocyte adhesion to the local vessel wall, with a concomitant increase in leukocyte transmigration [7]. The molecular changes that underlie this dysfunction in atherosclerosis are however only poorly understood, especially in human disease, which is not always (completely) reflected by murine studies.
Several factors have long been associated with a leaky phenotype, amongst which growth factors like VEGF [8]. Especially VEGF is well known not only for its role in the angiogenic process, but also in vascular hyperpermeability. In addition, the Tie2/Angiopoietin pathway is well-studied in relation to vessel maturation [9], and also several other growth factors or receptors, e.g. NRP-1 [10] have been implicated in a leaky vessel phenotype. The underlying common mechanisms are local changes in the endothelial cells themselves, promoting permeability and increased leukocyte adhesion.
Recent studies, most notably from the oncology field, have proposed stabilization of unstable, leaky vessels as therapeutic strategy to slow down or prevent disease progression. This may be achieved either by reinstating their barrier function through normalization of their maturation or as recently suggested, by limiting endothelial metabolism [11]. This is an attractive solution to avoid straightforward inhibition of pathological plaque angiogenesis, which is undesirable as evident from the above-described functional ambiguity. Moreover, targeting well-known growth factors, e.g. anti-VEGF treatment, also interferes with the growth factors’ systemic roles, causing variable results in therapeutic interventions [8, 12]. Ideally, therapeutic targets should be more disease-specific, and linked to the local pathological response. It descends from the above, the necessity to identify specific players that are altered in human plaque microvessels.
In this study, we set out to identify potential proteins that are specifically regulated in plaque microvessels in human atherosclerotic lesions. To get a better overview of how the pathological angiogenic response in atherosclerosis is governed, we applied network coexpression analysis of transcriptomic data obtained from well-characterized human atheromas, to pinpoint critical factors in increased angiogenesis and/or the leaky phenotype of plaque microvessels and validate their role in the regulation of microvessel patency in the context of atherosclerosis.
Methods
Patient samples, histology, and immunohistochemistry
Paired stable and unstable segments from human atherosclerotic plaque samples were obtained from carotid artery lesions from 24 patients undergoing endarterectomy (Department of Vascular Surgery, Maastricht University Medical Center, Maastricht, the Netherlands) as part of the Maastricht Pathology Tissue Collection (MPTC). Collection, storage, and use of tissue and patient data were performed in agreement with the Dutch Code for Proper Secondary Use of Human Tissue. Plaque segments were staged by histological analysis based on HE staining according to Virmani et al., where pathological intimal thickening (PIT) was classified as early lesions, thick fibrous cap atheroma (TkFCA) were classified as advanced stable, and intraplaque hemorrhage (IPH) as ruptured (advanced unstable) segments, respectively [13].
Immunohistochemical stainings were performed on consecutive paraffin sections for CD31 (vascular endothelial cells), CD105 (angiogenic marker), αSMA (smooth muscle cell/pericyte), the CD68 (macrophages), and SPTBN1. Appropriate IgG control antibodies were used as a negative control. Double stainings were performed for CD31/CD105, αSMA/CD31, and CD31/SPTBN1 and analyzed using spectral imaging system.
Cell culture
Human umbilical vein endothelial cells (HUVEC) or the human microvascular endothelial cell line (HMEC-1), derived from human foreskin endothelial cells, were cultured on fibronectin (FN) coated plates in respectively EGM2 or RPMI1640 + glutamax. HUVECs were used for experiments between P2 and P4, HMEC-1 cells were used until a maximum of P20. For knockdown of specific genes, cells were treated for 4 h with a targeted siRNA in combination with HiPerfect transfection reagent in Opti-Mem, and cells were used for various assays after 24–48 h post-transfection.
RNA extraction & transcriptomics on patient samples
RNA isolation was performed by Guanidium Thiocyanate lysis followed by Cesium Chloride gradient centrifugation and purified using the Nucleospin RNAII kit. RNA concentrations were measured, and RNA quality and integrity was determined using Lab-on-Chip analysis. Biotinylated cRNA was prepared from 100 ng total RNA using the Illumina TotalPrep RNA Amplification Kit according to the manufacturer’s specifications. 750 ng of cRNA per sample was hybridized to Illumina Human Sentrix-8 V2.0 BeadChip® and washed according to the Illumina standard procedure. Scanning was performed on the Illumina BeadStation 500, image analysis and extraction of raw expression data was performed with default settings and without normalization.
Computational methods
Analyses of transcriptomic data were performed in R [14]. Raw expression data were imported using the package lumi [15], and normalized by robust spline normalization. Co-expression networks were estimated by applying the methods implemented in the package WGCNA [16]. Module-traits association was estimated by Pearson correlation of the expression modules eigenvalues and quantitative traits. Network architecture was visualized using Cytoscape [17] and Db-String. ClueGO was used to enrich the modules of interest for overrepresented Gene Ontology terms and Pathways [18].
In order to reveal the relationships between proteins/peptides and microvessel density (MVD), a ranking list was performed to show the importance of proteins/peptides relating to MVD based on two different measurements, the Pearson Correlation Coefficient and the Maximal Information Coefficient (MIC) [19]. In each ranking list, top-100 high-ranked proteins/peptides were listed, after which overlapping proteins/peptides were marked and cross-referenced with hemorrhage-related proteins.
Zebrafish
Tg(fli1a:eGFP)y1 embryos were injected with control morpholino (Ctrl Mo) or with morpholino against SPTBN1 (7.5ng/microliter) at the one-cell stage (n = 35 per group). Confocal pictures of the caudal vasculature were taken at 48 h post-fertilization (hpf) using a Zeiss confocal microscope. Caudal vein plexus (CVP) thickness and area were analyzed using ImageJ software.
Statistical analysis
Where not explicitly specified otherwise, all data are presented as mean ± SEM. For patient data, groups were compared using a Mann–Whitney rank-sum test for continuous variables. For correlation analyses of human qPCR and protein expression data, Shapiro–Wilk test for normality was performed, after which correlations of Gaussian distributed data were calculated by Pearson and of non-Gaussian data by Spearman correlation test. For group comparisons, data were tested for Gaussian distribution, after which a Student’s t-test (Gaussian) or Mann–Whitney U test (non-Gaussian) was used to compare individual groups; multiple groups were compared by ANOVA or Kruskall-Wallis tests, with Bonferroni or Dunn’s post-hoc test, respectively. Statistics were performed using Graphpad Prism 5.0. A p-value of < 0.05 was considered statistically significant. *, **, and *** denote p < 0.05, p < 0.01, and p < 0.001, RESP.
Full materials and methods are available in the online supplement.
Results
Microvessel density was increased in unstable advanced plaque segments and associated with increased CD105, whereas perivascular coverage was unaltered
To assess plaque and lipid core size, paraffin-embedded, hematoxylin–eosin stained sections were quantified morphometrically, revealing an increased plaque (Fig. 1A) and lipid core size (Fig. 1B) in advanced unstable (with intraplaque hemorrhage) segments compared to segments with an early, stable phenotype of the same symptomatic patient. Staining for the vascular marker CD31 showed both increased microvessel density (MVD) (Fig. 1C) and microvessel hotspots (Fig. 1D) within the plaque. Co-localization of the angiogenic marker CD105 and the vascular marker CD31 showed an augmented percentage of angiogenic plaque microvessels in advanced unstable lesions (Fig. 1E), as assessed in using multispectral analysis (Fig. 1F). Perivascular coverage defined as αSMA-covered microvessels (Fig. 1G) and the amount of perivascular coverage (αSMA-positivity) per microvessel (Fig. 1H) were measured but did not show differences between early and advanced plaque segments. This may in part be due to the high degree of heterogeneity in perivascular coverage observed within lesion segments (Fig. 1I), and/or symptomatic nature of the patient. Analysis of the cross-correlation between plaque traits suggested a strong association of MVD with macrophages and αSMA-coated mature vessel presence, but not with CD105+ angiogenic endothelial cells (EC). (Fig. 1J). Plaque content of CD105+ angiogenic ECs was seen to correlate with plaque size and, at borderline significance, with lipid core size and intraplaque hemorrhage.
Analysis of the gene cluster highly correlating with MVD, but not with angiogenic activity, yielded potential target genes involved in plaque neovascularization
To gain mechanistic insights into the process that link microvessel density to plaque destabilization, we analyzed the microarray dataset from our study population by weighted gene coexpression network analysis (WGCNA). This revealed 38 gene modules overall. The module’s Eigengenes were subsequently correlated to the histological traits e.g. MVD, angiogenic activity, and perivascular coverage of plaque microvessels, after which correlations were visualized in a module-trait heatmap (Fig. 2A). MVD was highly correlated with module L (P = 1.0 × 10–4), and angiogenic activity (CD105+) with module R (P = 0.004). As shown earlier, MVD and lesion macrophage content showed considerable overlap in module correlation pattern, in support of inflammatory regulation and/or influx from leaky microvessels. Hence, we opted for module L. Gene ontology (GO) analysis of this module revealed a clear enrichment of biological processes related to cell- and matrix-adhesion, and wound healing (Fig. 2B), next to overrepresentation of extracellular matrix, focal adhesion, and junctional components (GO: cellular components). To select candidate genes for further study, we next ranked the genes within module L by highest correlation for MVD and highest centrality within the module’s subnetwork (Supplemental Table 1). A protein–protein interaction network constructed for the top-ranking (hub) genes using STRING (Fig. 2C) confirmed the overrepresentation of factors related to endothelial cell function, including VE-cadherin, vinculin, and paxillin. Interrogation of the human plaque single cell RNASeq datasets, showed that SPTBN1 and to a lesser extent also DOCK1 and ZEB1 were mainly expressed by endothelial cell subsets (dataset Alsaigh et al. [20], Fig. 2E; metadata Mosquera et al. [21], Supplemental Fig. 1A). Based on their association with the protein–protein interaction network and the relative ranking, in combination with the single cell expression, we selected Dedicator-Of-Cytokinesis 1 (dock1), Spectrin Beta Non-Erythrocytic 1 (sptbn1), and Zinc Finger E-Box Binding Homeobox 1 (zeb1), for further study of their function in endothelial cells.
Knockdown of sptbn1 and zeb1 both impact endothelial function
First, we confirmed solid and preferential expression of the candidate genes in HMEC-1 (Supplemental Fig. 1B) endothelial cells in vitro compared to (polarized) THP-1 cells and primary vascular smooth muscle cells. We confirmed effective siRNA silencing of the target genes at mRNA and protein level (Supplemental Fig. 1C-D). Subsequently, loss-of-function studies in HMEC1 were performed to assess the impact of the candidates on critical endothelial functions, including EC adhesion, spreading, migration, proliferation and cell cycle analysis, tube formation capacity, vascular permeability, and leukocyte transmigration under flow (Fig. 3). Dock1 knockdown did not result in major differences throughout these assays (Fig. 3A-H), with only cell proliferation being significantly affected (Fig. 3E, F). Zeb1 knockdown also significantly reduced proliferation (Fig. 3E, F), and in addition affected tube formation capacity, with increased tube number (Fig. 3G). Knockdown of sptbn1 however showed a profound reduction of cell adhesion (Fig. 3A), spreading (Fig. 3B), and migration (Fig. 3C, 3D). Moreover, it dramatically altered vascular permeability in a dextran leakage Transwell assay (Fig. 3H), a finding that could be confirmed in the transendothelial electrical resistance (TEER) assay (Fig. 3I). Considering the marginal functional effects of dock1 silencing, it was excluded from subsequent analysis. Silencing of sptbn1 but not zeb1 augmented leukocyte adhesion to the endothelium under flow (Fig. 3J), as well as leukocyte transmigration (Fig. 3K), although the rate of leukocyte transmigration was unaltered (Fig. 3L). This led us to focus on sptbn1 as prime candidate for in vivo validation of its regulatory role in vessel function in vivo. In vivo knockdown of sptbn1 by microinjection of morpholino oligonucleotides into zebrafish embryos led to attenuated development of the caudal vein plexus (CVP) (Figs. 3M-O), in line with the reduced adhesion, spreading, and migration in vitro.
SPTBN1 expression is regulated by tissue stiffness in vitro and in vivo
While expression of sptbn1 was reduced in advanced atherosclerosis, and linked to microvascular permeability in vitro, its regulation in the context of atherosclerosis is unclear. Relevant stimuli, including LPS induced inflammation, oxLDL exposure, or a combination of both, did not significantly alter sptbn1 mRNA expression in HMEC-1. As our network and the network predictions of module L (Fig. 2C) revealed substrate stiffness-responsive genes, we studied the role of this stimulus on sptbn1 expression. As HMEC-1 did not grow well on the stiffness gels, we assessed in HUVECs what effect substrate stiffness would have on sptbn1 expression. Total SPTBN1 protein expression was progressively decreased with increasing substrate stiffness, as judged by immune fluorescence staining (Fig. 4A, red; Fig. 4B), as well as by Western Blotting analysis (Fig. 4C). Furthermore, we could observe a clear inverse correlation between SPTBN1 and the stiffness-sensitive marker DLC1 [22] (Fig. 4D), and in primary tissue, softer venous tissue showed a higher SPTBN1 expression compared to stiffer arterial tissue (Fig. 4E, 4F). Next, we performed multispectral imaging, examining colocalization of SPTBN1 with the vascular marker CD31 in plaque microvessels (Fig. 4G), to dissect the expression of endothelial SPTBN1 during disease progression. Interestingly, the level of colocalization between SPTBN1 and CD31 decreased with disease progression (Fig. 4H). In addition, the local expression levels of SPTBN1 in plaque vessels in more advanced stable and unstable lesions were significantly reduced compared to early lesions (Fig. 4I). As plaque presence lowers the elasticity of the vessel wall and increased local stiffness, especially at later stages, the reduced expression of SPTBN1 in plaque microvessels may be attributable to the biomechanical properties of plaque tissue.
SPTBN1 is involved in gene networks related to cell–cell junctions and cell–matrix interactions
The effects of sptbn1 silencing on permeability and leukocyte transmigration, and its interaction with cell junction and adhesion proteins, point to a role of SPTBN1 in the regulation of EC function. To delineate the mode of action of SPTBN1’s regulatory function, we profiled mRNA expression patterns upon sptbn1 silencing versus control siRNA treated and untreated HMEC-1 and HUVEC. This analysis yielded 514 differentially expressed genes, of which 47 were also contained in the MVD associated module L geneset of the human plaque study cohort. (P = 0.03; Supplemental Fig. 1E). GO analysis and protein–protein interaction network analysis of the top 40 most differential genes after sptbn1 silencing revealed a clear overrepresentation of pathways involved in cell–cell and cell–matrix adhesion, as well as cell–cell junction and focal adhesion regulation e.g. cadherin-1, -2, and -5 (VE-cadherin), paxillin, vinculin, tight junction protein-1, and occludin (Fig. 5A & B), which was confirmed of the BIKE cohort (Supplemental Table 2).
SPTBN1 associates with VE-cadherin at the cell junction and is also involved in focal adhesion regulation
To assess the role of SPTBN1 in endothelial cells, we studied the localization of this protein. Staining of SPTBN1 on endothelial cells revealed clear presence of the protein on stable, linear junctions (Fig. 5C). Also, SPTBN1 was highly prevalent in/near the Golgi, and predominantly localized in the basolateral plane of the cell. To further investigate the association of SPTBN1 with EC junctions, we performed an immunoprecipitation (IP) of SPTBN1, and stained for junctional proteins like VE-cadherin and the tight junction protein ZO-1. VE-cadherin was clearly visible after SPTBN1 pulldown (Fig. 5D), both in HUVEC and HMEC-1, while ZO-1 was not present. These data were confirmed by IP for VE-cadherin and subsequent immunoblotting for SPTBN1 (Supplemental Fig. 2A).
Upon knockdown of sptbn1, we observed less linear junctions, a concomitant increase in reticular and focal adherens junctions (Fig. 5E). This was accompanied by an increased frequency of gaps between endothelial cells (Supplemental Fig. 2B), by increased stress fiber expression (Fig. 5E + F) and basolateral focal adhesions, and by reduced junctional width (Fig. 5G). Of note, we also observed a trend for higher levels of the focal adhesion protein paxillin. These data underpin the causal role of sptbn1 knockdown in junctional instability and altered focal adhesion dynamics, potentially compromising permeability and promoting leukocyte transmigration.
Reduced SPTBN1 expression is associated with increased intraplaque hemorrhage and hemorrhagic event risk
Next, we assessed whether SPTBN1 downregulation contributes to leaky vessel phenotype in human disease. Hereto, we evaluated the correlation of SPTBN1 expression in human plaque with histological features such as intraplaque hemorrhage in our Maastricht Human Plaque Study (MaasHPS) cohort. Sptbn1 expression was not only downregulated in advanced versus early lesions in our MaasHPS cohort (Fig. 2D), but also in BIKE, at mRNA (plaque versus normal artery) (Fig. 6A) and protein level (carotid artery plaque vs adjacent adjacent tissue) (Fig. 6B). Moreover, sptbn1 mRNA expression showed a very significant inverse correlation with plaque hemorrhage size not only for the main isoform (Fig. 6C) but also for the other 2 detectable splicing variants (Supplementary Fig. 2C). This association and its clinical implications could be validated in the Athero-Express Biobank cohort, with significantly lower plaque SPTBN1 protein expression in CVD event-free patient than in patients with an CVD even during follow-up, albeit that this effect was borderline significant (P = 0.06) (Fig. 6D). Validation in the BIKE cohort showed a highly significant 40% lower sptbn1 mRNA expression in symptomatic than asymptomatic patients (Fig. 6E). To confirm the link to hemorrhage and leakage is, we interrogated the peptidomics dataset of the MaasHPS cohort. Interestingly, almost 50% of the top 100 peptides with highest correlation with microvessel density were representing plasma proteins (P = 1.3 × 10–9), but the top 100 peptides were also enriched in focal adhesion and junctional cadherin binding peptides (Supplemental Fig. 2D), again confirming our earlier findings. Mammalian Phenotype Ontology analysis revealed a clear association with hemorrhage (Fig. 6F, red), while GO pathway analysis indicated an overrepresentation of platelet activation and degranulation pathways among the CD31+ plaque vessel correlated peptides (data not shown), consistent with hemorrhage. These observations all support a role for SPTBN1 in microvascular permeability and vulnerability in human plaques.
Discussion
The link between human plaque angiogenesis and enhanced atherosclerotic burden and intraplaque hemorrhage has been well established. However, beyond the prototypic angiogenic factors that have been described in the field of angiogenesis, e.g. VEGF and its receptors, we still lack deeper understanding of how microvessel formation and function in—in particular—human atherosclerotic plaque are regulated. In vitro models and in vivo mouse models have limited value as the regulation of critical genes or proteins that influence vessel function may (in part) depend on the local environment in human plaque, which may not be modelled in murine models. Therefore, we used a genomics-based approach to identify new, unknown regulators of human plaque microvessel function. This novel approach in human atherosclerotic tissue allowed unbiased identification of central mediators of plaque microvascular function. Here we report SPTBN1 as a novel actor in microvessel permeability and leukocyte transmigration in vitro.
SPTBN1, or Spectrin Beta Non-Erythrocytic 1, has previously been implicated in actin crosslinking, and has been described as a scaffold/adaptor protein in actin cytoskeleton anchoring into the membrane [23, 24]. In parallel to similar scaffold proteins, SPTBN1 is therefore thought to be involved in regulating e.g. cell shape, motility, as well as compartmentalization of transmembrane proteins. SPTBN1 functions has been studied in considerable detail in brain and heart physiology. In brain, SPTBN1 found predominantly in Purkinje-cell bodies, and appear to be involved in the assembly of specialized membrane domains in Purkinje neurons [24]. In heart, SPTBN1 has been shown to be a vital part of the membrane-associated cytoskeleton of cardiomyocytes, and was shown to be pivotal for the organization of several key components in cardiomyocyte functioning [25]. More recently, Smith and al showed that lack of cardiac SPTBN1 caused altered localization of the sarcoplasmic reticulum ryanodine receptor 2 (RyR2), altering calcium release and eventually leading to arrhythmias [26].
Taken together, SPTBN1 exerts pleiotropic functions in membrane organization in several cell types. Recently, two reports already described a role of the spectrin cytoskeleton in leukocyte rolling [27] and mechanosensing in endothelial cells [28], in part via anchoring of CD44. In this study, we also show that SPTBN1 was central in our plaque microvessel associated gene network and, in vitro and in situ, exerts important regulatory functions in endothelial cell. In this regard, its activity reminisces of that of its family member SPTAN1, which has previously been described in regulating cell endothelial cell–cell contacts [29]. Based on our data, we propose that SPTBN1 silencing induces two major effects in endothelial cells (Fig. 7). First, knockdown cells show impaired adhesion and spreading capacity upon seeding. This was in conjunction with reduced migration capacity during wound healing and the hampered CVP development in zebrafish in vivo. SPTBN1 in endothelial cells therefore seems to be important for facilitating cell spreading and motion, as also was alluded to by GO term enrichment in the pathway analysis. Second, reduced expression of SPTBN1 increased permeability and leukocyte transmigration over an endothelial monolayer. This phenomenon could be linked to loss of VE-cadherin, the predominant cell junction molecule in vascular endothelial cells, at the cell–cell junction, and less linear endothelial junctions. In addition, focal adhesion numbers where significantly increased upon SPTBN1 knockdown in vitro. As we observed decreased mRNA and protein expression of SPTBN1 in ruptured versus stable advanced plaques and plaque microvessels, respectively, this suggests similar vessel destabilising effects in human disease. Despite reduced SPTBN1 levels, microvessel densities were increased in ruptured plaques. As SPTBN1 levels showed an inverse correlation with the extent of plaque hemorrhage, the predominant effect of lower SPTBN1 in human plaques is most likely microvessel hyperpermeability instead of sprouting.
Gene ontology studies show that SPTBN1 may be acting in (I) ER to Golgi vesicle-mediated transport [30], a notion that is underpinned by the high expression of SPTBN1 we observed in or near the ER and Golgi, and (II) in plasma membrane/cytoskeleton organization and actin binding [25, 26, 30, 31]. The latter suggest that SPTBN1 may be part of the adhesome [32, 33], which forms a crucial link between the plasma membrane and adaptor molecules on the one hand and the actin cytoskeleton on the other hand, and which includes molecules like VE-cadherin [34], or focal adhesion molecules like paxillin and vinculin [35]. The direct link to VE-cadherin that we found in our pull-down experiments was unexpected, even though SPTBN1 had previously been described in localization of E-cadherin to the lateral membrane in epithelial cells [23, 36]. Although the exact mechanism for the increase in focal adhesions remains unclear, it in part could explain the augmented leukocyte transmigration we observed [37]. This contractile endothelial phenotype may also explain the increase in stress fibres we observed upon knockdown of SPTBN1 [38], and is in conjunction with the more instable junctional phenotype.
The adhesome regulatory function of SPTBN1 may underlie the observed effects on vascular permeability in vitro (2.5-fold increase), and in patient tissue, i.e. the significant correlation between SPTBN1 levels and intraplaque hemorrhage in symptomatic lesions in three separate patient cohorts. Moreover, there was a strong enrichment of plasma proteins and hemorrhage GO terms among the MVD correlated peptides in the proteomics study. Moreover, we could attribute the reduction in SPTBN1 both in vitro, as well as in human plaque tissue, to increased tissue stiffness, the latter of which has been implicated in plaque destabilisation [39, 40]. Considering this strong correlation between SPTBN1 and a leaky vessel phenotype, the SPTBN1 axis may therefore be a potential important factor in determining the increased cardiovascular events by promoting hemorrhage, warranting further functional studies in vivo.
In conclusion, using a combination of histological analysis with genomic analyses we could show, that stiffness-dependent expression of SPTBN1 in atherosclerotic lesion microvessels may present a potential central factor involved in the leaky phenotype of these vessels. Intervening in this pathway may therefore be a way of selectively targeting the leaky vessel phenotype of plaque microvessels to prevent hemorrhage and further plaque exacerbation and adverse cardiovascular outcome.
References
Ribatti D, Levi-Schaffer F, Kovanen PT (2008) Inflammatory angiogenesis in atherogenesis–a double-edged sword. Ann Med 40(8):606–621
Sluimer JC, Daemen MJ (2009) Novel concepts in atherogenesis: angiogenesis and hypoxia in atherosclerosis. J Pathol 218(1):7–29
Calcagno C, Fayad ZA, Raggi P (2017) Plaque microvascularization and permeability: Key players in atherogenesis and plaque rupture. Atherosclerosis 263:320–321
Moreno PR, Purushothaman KR, Fuster V, Echeverri D, Truszczynska H, Sharma SK et al (2004) Plaque neovascularization is increased in ruptured atherosclerotic lesions of human aorta: implications for plaque vulnerability. Circulation 110(14):2032–2038
Sluimer JC, Kolodgie FD, Bijnens AP, Maxfield K, Pacheco E, Kutys B et al (2009) Thin-walled microvessels in human coronary atherosclerotic plaques show incomplete endothelial junctions relevance of compromised structural integrity for intraplaque microvascular leakage. J Am Coll Cardiol 53(17):1517–1527
Virmani R, Kolodgie FD, Burke AP, Finn AV, Gold HK, Tulenko TN et al (2005) Atherosclerotic plaque progression and vulnerability to rupture: angiogenesis as a source of intraplaque hemorrhage. Arterioscler Thromb Vasc Biol 25(10):2054–2061
Rademakers T, Douma K, Hackeng TM, Post MJ, Sluimer JC, Daemen MJ et al (2013) Plaque-associated vasa vasorum in aged apolipoprotein E-deficient mice exhibit proatherogenic functional features in vivo. Arterioscler Thromb Vasc Biol 33(2):249–256
Vuorio T, Jauhiainen S, Yla-Herttuala S (2012) Pro- and anti-angiogenic therapy and atherosclerosis with special emphasis on vascular endothelial growth factors. Expert Opin Biol Ther 12(1):79–92
Thomas M, Augustin HG (2009) The role of the Angiopoietins in vascular morphogenesis. Angiogenesis 12(2):125–137
Roth L, Prahst C, Ruckdeschel T, Savant S, Westrom S, Fantin A et al (2016) Neuropilin-1 mediates vascular permeability independently of vascular endothelial growth factor receptor-2 activation. Sci Signal. 9(425):ra42
Cantelmo AR, Conradi LC, Brajic A, Goveia J, Kalucka J, Pircher A et al (2016) Inhibition of the glycolytic activator PFKFB3 in endothelium induces tumor vessel normalization, impairs metastasis, and improves chemotherapy. Cancer Cell 30(6):968–985
Yla-Herttuala S (2006) An update on angiogenic gene therapy: vascular endothelial growth factor and other directions. Curr Opin Mol Ther 8(4):295–300
Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM (2000) Lessons from sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol 20(5):1262–1275
Team RdC (2010) R: A language and environment for statistical computing. In: R foundation for Statistical Computing V, Austria, editor
Du P, Kibbe WA, Lin SM (2008) lumi: a pipeline for processing Illumina microarray. Bioinformatics 24(13):1547–1548
Langfelder P, Horvath S (2008) WGCNA: an R package for weighted correlation network analysis. BMC Bioinform 9:559
Shannon P, Markiel A, Ozier O, Baliga NS, Wang JT, Ramage D et al (2003) Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res 13(11):2498–2504
Bindea G, Mlecnik B, Hackl H, Charoentong P, Tosolini M, Kirilovsky A et al (2009) ClueGO: a Cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks. Bioinformatics 25(8):1091–1093
Reshef DN, Reshef YA, Finucane HK, Grossman SR, McVean G, Turnbaugh PJ et al (2011) Detecting novel associations in large data sets. Science 334(6062):1518–1524
Alsaigh T, Evans D, Frankel D, Torkamani A (2022) Decoding the transcriptome of calcified atherosclerotic plaque at single-cell resolution. Commun Biol 5(1):1084
Mosquera JV, Auguste G, Wong D, Turner AW, Hodonsky CJ, Alvarez-Yela AC et al (2023) Integrative single-cell meta-analysis reveals disease-relevant vascular cell states and markers in human atherosclerosis. Cell Rep 42(11):113380
Schimmel L, van der Stoel M, Rianna C, van Stalborch AM, de Ligt A, Hoogenboezem M et al (2018) Stiffness-induced endothelial DLC-1 expression forces leukocyte spreading through stabilization of the ICAM-1 adhesome. Cell Rep 24(12):3115–3124
Kizhatil K, Yoon W, Mohler PJ, Davis LH, Hoffman JA, Bennett V (2007) Ankyrin-G and beta2-spectrin collaborate in biogenesis of lateral membrane of human bronchial epithelial cells. J Biol Chem 282(3):2029–2037
Chen Y, Yu P, Lu D, Tagle DA, Cai T (2001) A novel isoform of beta-spectrin II localizes to cerebellar Purkinje-cell bodies and interacts with neurofibromatosis type 2 gene product schwannomin. J Mol Neurosci 17(1):59–70
Baines AJ, Pinder JC (2005) The spectrin-associated cytoskeleton in mammalian heart. Front Biosci 10:3020–3033
Smith SA, Sturm AC, Curran J, Kline CF, Little SC, Bonilla IM et al (2015) Dysfunction in the betaII spectrin-dependent cytoskeleton underlies human arrhythmia. Circulation 131(8):695–708
Mylvaganam S, Riedl M, Vega A, Collins RF, Jaqaman K, Grinstein S, Freeman SA (2020) Stabilization of endothelial receptor arrays by a polarized spectrin cytoskeleton facilitates rolling and adhesion of leukocytes. Cell Rep 31(12):107798
Mylvaganam S, Yusuf B, Li R, Lu C-Y, Robinson LA, Freeman SA, Grinstein S (2021) Endothelial integration of mechanosensory signals by the spectrin cytoskeleton. bioRxiv
Benz PM, Blume C, Moebius J, Oschatz C, Schuh K, Sickmann A et al (2008) Cytoskeleton assembly at endothelial cell-cell contacts is regulated by alphaII-spectrin-VASP complexes. J Cell Biol 180(1):205–219
Mohler PJ, Yoon W, Bennett V (2004) Ankyrin-B targets beta2-spectrin to an intracellular compartment in neonatal cardiomyocytes. J Biol Chem 279(38):40185–40193
Lim JA, Baek HJ, Jang MS, Choi EK, Lee YM, Lee SJ et al (2014) Loss of beta2-spectrin prevents cardiomyocyte differentiation and heart development. Cardiovasc Res 101(1):39–47
Schiller HB, Fassler R (2013) Mechanosensitivity and compositional dynamics of cell-matrix adhesions. EMBO Rep 14(6):509–519
Fokkelman M, Balcioglu HE, Klip JE, Yan K, Verbeek FJ, Danen EH, van de Water B (2016) Cellular adhesome screen identifies critical modulators of focal adhesion dynamics, cellular traction forces and cell migration behaviour. Sci Rep 6:31707
Zaidel-Bar R (2013) Cadherin adhesome at a glance. J Cell Sci 126(Pt 2):373–378
Horton ER, Humphries JD, James J, Jones MC, Askari JA, Humphries MJ (2016) The integrin adhesome network at a glance. J Cell Sci 129(22):4159–4163
Kizhatil K, Davis JQ, Davis L, Hoffman J, Hogan BL, Bennett V (2007) Ankyrin-G is a molecular partner of E-cadherin in epithelial cells and early embryos. J Biol Chem 282(36):26552–26561
Yuan SY, Shen Q, Rigor RR, Wu MH (2012) Neutrophil transmigration, focal adhesion kinase and endothelial barrier function. Microvasc Res 83(1):82–88
Livne A, Geiger B (2016) The inner workings of stress fibers - from contractile machinery to focal adhesions and back. J Cell Sci 129(7):1293–1304
Huynh J, Nishimura N, Rana K, Peloquin JM, Califano JP, Montague CR et al (2011) Age-related intimal stiffening enhances endothelial permeability and leukocyte transmigration. Sci Transl Med 3(112):112ra22
Schaefer A, Te Riet J, Ritz K, Hoogenboezem M, Anthony EC, Mul FP et al (2014) Actin-binding proteins differentially regulate endothelial cell stiffness, ICAM-1 function and neutrophil transmigration. J Cell Sci 127(Pt 20):4470–4482
Acknowledgements
The authors are indebted to C. Recarti and P. Namsolleck for their assistance in TEER measurements. A. van der Wal and C. van der Loos are acknowledged for their assistance in setting up the multispectral imaging.
Funding
This study was supported by the CARIM portfolio excellence program “Plaque neoangiogenesis” (TR, SH, MAMJvZ, EB), the Centre for translational Molecular Medicine (CTMM) project Circulating cells (grant number 01C-102; MM, EB), and the China Scholarship Council (CSC) grant 201609120004 (HJ). The BiKE study (LM, UH) is supported by the Swedish Heart and Lung Foundation; Swedish Research Council (K2009-65X-2233–01-3, K2013-65X-06816–30-4, 349–2007-8703); Uppdrag Besegra Stroke (P581/2011–123); Stockholm County Council (ALF2011-0260, ALF-2011–0279). LM is recipient of fellowships and grants from the Swedish Society for Medical Research; Swedish Heart and Lung Foundation, Tore Nilsson’s, Magnus Bergvall’s and Karolinska Institute Foundations.
Author information
Authors and Affiliations
Contributions
TR and EB devised and planned the experiments; TR, LMP, HJMF, TO, PH, JvR, FR, KvK performed and analysed the data; BMEM was involved in acquiring human atherosclerotic tissue, GP provided data on the Athero-Express Biobank cohort; MM and JH performed the computational analyses; TR drafted the manuscript; LMP, SH CJP, SH, MJAPD, GP, MS, MvZ, UH, FD, JvB, JS and EB provided scientific input and revised the manuscript.
Corresponding author
Ethics declarations
Competing interests
No competing financial interests have been reported.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary Information
Below is the link to the electronic supplementary material.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Rademakers, T., Manca, M., Jin, H. et al. Human atherosclerotic plaque transcriptomics reveals endothelial beta-2 spectrin as a potential regulator a leaky plaque microvasculature phenotype. Angiogenesis (2024). https://doi.org/10.1007/s10456-024-09921-z
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s10456-024-09921-z