Eva Karlöf’s research while affiliated with Karolinska Institutet and other places

Objective Carotid endarterectomy for symptomatic carotid stenosis is recommended for patients with >70% stenosis, but not in those with <50%. Because non-significant, low-degree stenoses may still cause strokes, refined risk stratification is necessary, which could be improved by assessing biological features of plaque instability. To challenge risk-stratification based on luminal narrowing, we compared biological features of carotid plaques from symptomatic patients with low-degree (<50%) vs high-degree (>70%) stenosis and explored potential mechanisms behind plaque instability in low-degree stenoses. Methods Endarterectomy specimens were taken from symptomatic patients with high-degree (n = 204) and low-degree (n = 34) stenosis, all part of the Biobank of Karolinska Endarterectomies. Patient demographics, image-derived plaque morphology, and gene expression analyses of extracted lesions were used for comparisons. Plaque biology was assessed by transcriptomics using dimensionality reduction, differential gene expression, and gene-set enrichment analyses. Immunohistochemistry was used to study proteins corresponding to upregulated genes. Results The demographics of the two groups were statistically similar. Calcification, lipid-rich necrotic core, intraplaque hemorrhage, plaque burden, and fibrous cap thickness were similar in both groups, whereas the sum of lipid-rich necrotic core and intraplaque hemorrhage was higher (P = .033) in the high-degree stenosis group. Dimensionality reduction analysis indicated poor clustering separation of plaque gene expression in low-compared with high-degree stenosis lesions, whereas differential gene expression showed upregulation of hypoxia-inducible factor 3A (log2 fold change, 0.7212; P = .0003), and gene-set enrichment analyses identified pathways related to tissue hypoxia and angiogenesis in low-degree stenoses. Hypoxia-inducible factor 3-alpha protein was associated with smooth muscle cells in neo-vascularized plaque regions. Conclusions Plaques from symptomatic patients with non-significant low-degree carotid stenoses showed morphologic and biological features of atherosclerotic plaque instability that were comparable to plaques from patients with high-degree stenoses, emphasizing the need for improved stroke risk stratification for intervention in all patients with symptomatic carotid stenosis irrespective of luminal narrowing. An increased expression of hypoxia-inducible factor 3A in low-degree stenotic lesions suggested mechanisms of plaque instability associated with tissue hypoxia and plaque angiogenesis, but the exact role of hypoxia-inducible factor 3A in this process remains to be determined. Clinical relevance Carotid plaques from symptomatic patients with <50% stenosis show morphologic and biological features of plaque instability, comparable to high-degree stenosis, which emphasizes the need for improved stroke risk stratification beyond stenosis severity.

The identification of carotid atherosclerotic lesion at risk for plaque rupture, eventually resulting in cerebral embolism and stroke, is of paramount clinical importance. High stress in the fibrous plaque cap has been proposed as risk factor. However, among others, residual strains influence said stress predictions, but quantitative and qualitative implications of residual strains in this context are not well explored. We therefore propose a multiplicative kinematics-based Growth and Remodeling (G&R) framework to predict residual strains from homogenizing tissue stress and then investigate its implication on plaque stress. Carotid vessel morphology of four patients was reconstructed from clinical Computed Tomography-Angiography (CT-A) images and equipped with heterogeneous tissue constitutive properties assigned through a histology-based artificial intelligence image segmentation tool. As compared to a purely elastic analysis and depending on patient-specific morphology and tissue distributions, the incorporation of residual strains reduced the maximum wall stress by up to $30\%$ 30 % and resulted in a fundamentally different distribution of stress across the atherosclerotic wall. Regardless residual strains homogenized tissue stresses, the fibrous plaque cap may persistently be exposed to spots of high stress. In conclusion, the incorporation of residual strains in biomechanical studies of atherosclerotic carotids may be important for a reliable assessment of fibrous plaque cap stress.

Background: Calcification, a key feature of advanced human atherosclerosis, is positively associated with vascular disease burden and adverse events. We showed that macrocalcification can be a stabilizing factor for carotid plaque molecular biology, due to inverse association with immune processes. Mast cells (MCs) are important contributors to plaque instability, but their relationship with macrocalcification is unexplored. With a hypothesis that MC activation negatively associates with carotid plaque macrocalcification, we aimed to investigate the link between MCs and carotid plaque vulnerability, and study MC role in plaque calcification via smooth muscle cells (SMCs). Methods: Pre-operative computed tomography angiographies of patients (n = 40) undergoing surgery for carotid stenosis were used to characterize plaque morphology. Plaque microarrays (n = 40 and n = 126) were used for bioinformatic deconvolution of immune cell populations. Tissue microarrays (n = 103) were used to histologically validate the contribution of activated and resting MCs in plaques. Results: Activated MCs and their typical markers were negatively correlated with macrocalcification. The ratio of activated vs. resting MCs was increased in low-calcified plaques from symptomatic patients. There was no modulating effect of medication on MC ratios. In vitro experiments showed that SMC calcification attenuated MC activation, while both active and resting MCs stimulated SMC calcification and induced dedifferentiation towards a pro-inflammatory-, osteochondrocyte-like phenotype, without modulating their migro-proliferative function. Conclusions: Integrative analyses from human plaques showed that MC activation is inversely associated with macrocalcification and positively with parameters of plaque vulnerability. Mechanistically, MCs induce SMC osteogenic reprograming, while matrix calcification in turn attenuates MC activation, offering new therapeutic avenues for exploration.

Intimal calcification and vascular stiffening are predominant features of end-stage atherosclerosis. However, their role in atherosclerotic plaque instability and how the extent and spatial distribution of calcification influence plaque biology remain unclear. We recently showed that extensive macro calcification can be a stabilizing feature of late-stage human lesions, associated with a reacquisition of more differentiated properties of plaque smooth muscle cells (SMCs) and extracellular matrix (ECM) remodeling. Here, we hypothesized that biomechanical forces related to macro-calcification within plaques influence SMC phenotype and contribute to plaque stabilization. We generated a finite element modeling (FEM) pipeline to assess plaque tissue stretch based on image analysis of preoperative computed tomography angiography (CTA) of carotid atherosclerotic plaques to visualize calcification and soft tissues (lipids and extracellular matrix) within the lesions. Biomechanical stretch was significantly reduced in tissues in close proximity to macro calcification, while increased levels were observed within distant soft tissues. Applying this data to an in vitro stretch model on primary vascular SMCs revealed upregulation of typical markers for differentiated SMCs and contractility under low stretch conditions but also impeded SMC alignment. In contrast, high stretch conditions in combination with calcifying conditions induced SMC apoptosis. Our findings suggest that the load bearing capacities of macro calcifications influence SMC differentiation and survival and contribute to atherosclerotic plaque stabilization.

Proprotein convertase subtilisin/kexins (PCSKs) constitute a family of nine related proteases: PCSK1-7, MBTPS1, and PCSK9. Apart from PCSK9, little is known about PCSKs in cardiovascular disease. Here, we aimed to investigate the expression landscape and druggability potential of the entire PCSK family for CVD. We applied an integrative approach, combining genetic, transcriptomic and proteomic data from three vascular biobanks comprising carotid atherosclerosis, thoracic and abdominal aneurysms, with patient clinical parameters and immunohistochemistry of vascular biopsies. Apart from PCSK4, all PCSK family members lie in genetic regions containing variants associated with human cardiovascular traits. Transcriptomic analyses revealed that FURIN, PCSK5, MBTPS1 were downregulated, while PCSK6/7 were upregulated in plaques vs. control arteries. In abdominal aneurysms, FURIN, PCSK5, PCSK7, MBTPS1 were downregulated, while PCSK6 was enriched in diseased media. In thoracic aneurysms, only FURIN was significantly upregulated. Network analyses of the upstream and downstream pathways related to PCSKs were performed on the omics data from vascular biopsies, revealing mechanistic relationships between this protein family and disease. Cell type correlation analyses and immunohistochemistry showed that PCSK transcripts and protein levels parallel each other, except for PCSK9 where transcript was not detected, while protein was abundant in vascular biopsies. Correlations to clinical parameters revealed a positive association between FURIN plaque levels and serum LDL, while PCSK6 was negatively associated with Hb. PCSK5/6/7 were all positively associated with adverse cardiovascular events. Our results show that PCSK6 is abundant in plaques and abdominal aneurysms, while FURIN upregulation is characteristic for thoracic aneurysms. PCSK9 protein, but not the transcript, was present in vascular lesions, suggesting its accumulation from circulation. Integrating our results lead to the development of a novel ‘molecular’ 5D framework. Here, we conducted the first integrative study of the proprotein convertase family in this context. Our results using this translational pipeline, revealed primarily PCSK6, followed by PCSK5, PCSK7 and FURIN, as proprotein convertases with the highest novel therapeutic potential.

Background Rupture of unstable atherosclerotic plaques with a large lipid-rich necrotic core and a thin fibrous cap cause myocardial infarction and stroke. Yet it has not been possible to assess this for individual patients. Clinical guidelines still rely on use of luminal narrowing, a poor indicator but one that persists for lack of effective means to do better. We present a case study demonstrating the assessment of biomechanical indices pertaining to plaque rupture risk non-invasively for individual patients enabled by histologically validated tissue characterization. Methods Routinely acquired clinical images of plaques were analyzed to characterize vascular wall tissues using software validated by histology (ElucidVivo, Elucid Bioimaging Inc.). Based on the tissue distribution, wall stress and strain were then calculated at spatial locations with varied fibrous cap thicknesses at diastolic, mean and systolic blood pressures. Results The von Mises stress of 152 [131, 172] kPa and the equivalent strain of 0.10 [0.08, 0.12] were calculated where the fibrous cap thickness was smallest (560 μm) (95% CI in brackets). The stress at this location was at a level predictive of plaque failure. Stress and strain at locations with larger cap thicknesses were calculated to be lower, demonstrating a clinically relevant range of risk levels. Conclusion Patient specific tissue characterization can identify distributions of stress and strain in a clinically relevant range. This capability may be used to identify high-risk lesions and personalize treatment decisions for individual patients with cardiovascular disease and improve prevention of myocardial infarction and stroke.

Introduction: Proprotein convertase subtilisin/kexins (PCSKs) constitute a family of 7 related proteases (PCSK1-7) and 2 distant ones (MBTPS1, PCSK9), with unexplored role in cardiovascular disease (CVD), apart from PCSK9 in lipid metabolism. Here, we aimed to investigate the expression landscape and therapeutic targeting potential of the entire PCSK family for ameliorating CVD. Methods: An integrative approach was applied with genetic, transcriptomic and proteomic data mining from public databases and three independent vascular biobanks comprising carotid atherosclerosis, thoracic and abdominal aneurysms. Omics analyses were followed by gene expression association with patient clinical parameters and immunohistochemistry in vascular biopsies. Results: Genetic studies revealed that, apart from PCSK4 , all PCSK family members lie in regions containing variants associated with human cardiovascular traits. Transcriptomic analyses showed that FURIN, PCSK5, MBTPS1 were downregulated, while PCSK6/7 were upregulated in atherosclerotic plaques vs. normal arteries. In abdominal aneurysms, FURIN, PCSK5, PCSK7, MBTPS1 were downregulated, while PCSK6 was enriched in diseased media. In thoracic aneurysms, only FURIN was significantly upregulated. To understand the mechanistic relationships of this protein family with the disease, network analyses of the upstream and downstream pathways related to PCSKs were done on the omics data from vascular biopsies. Immunohistochemistry indicated that PCSK protein levels correspond to the mRNA expression, except in the case of PCSK9 protein that was abundant in vascular biopsies. Correlation to clinical parameters in a carotid endarterectomy cohort revealed positive associations for PCSK5/6/7 with adverse cardiovascular events. Conclusions: Our results show that PCSK6 is the most enriched PCSK in plaques and abdominal aneurysms, while FURIN upregulation is characteristic for thoracic aneurysms. PCSK9 protein but not the transcript, was present in vascular lesions, suggesting its accumulation from circulation. Overall evaluation revealed that PCSK6 is the most attractive protease from this family to target for drug development in CVD.

Unstable carotid stenosis is an important cause of ischemic stroke, yet the basis of disease pathophysiology remains largely unknown. We hypothesized that integrated analyses of symptomatic carotid stenosis patients at increased stroke risk stratified by clinical scores CAR and ABCD2, with transcriptomic and clinical data, could improve identification of molecular pathways and targets for instability. We show that high CAR score reflects plaque instability processes related to intra-plaque hemorrhage, angiogenesis, inflammation and foam cell differentiation, while ABCD2 associates with neutrophil mediated immunity, foam cell differentiation, cholesterol transport and coagulation. Repressed processes in plaques from high risk patients were ossification, chondrocyte differentiation, SMC migration and ECM organization. ABCB5 gene was found as the top upregulated in high-risk patient’s plaques, localized to macrophages in areas with neovascularization and intra-plaque hemorrhage. The link between ABCB5 and intra-plaque hemorrhage suggests its key role for plaque instability that warrants further exploration.

Rationale: Vascular calcification is a prominent feature of late-stage diabetes, renal and cardiovascular disease (CVD), and has been linked to adverse events. Recent studies in patients reported that plasma levels of osteomodulin (OMD), a proteoglycan involved in bone mineralisation, associate with diabetes and CVD. We hypothesised that OMD could be implicated in these diseases via vascular calcification as a common underlying factor and aimed to investigate its role in this context. Methods and results: In patients with chronic kidney disease, plasma OMD levels correlated with markers of inflammation and bone turnover, with the protein present in calcified arterial media. Plasma OMD also associated with cardiac calcification and the protein was detected in calcified valve leaflets by immunohistochemistry. In patients with carotid atherosclerosis, circulating OMD was increased in association with plaque calcification as assessed by computed tomography. Transcriptomic and proteomic data showed that OMD was upregulated in atherosclerotic compared to control arteries, particularly in calcified plaques, where OMD expression correlated positively with markers of smooth muscle cells (SMCs), osteoblasts and glycoproteins. Immunostaining confirmed that OMD was abundantly present in calcified plaques, localised to extracellular matrix and regions rich in α-SMA+ cells. In vivo, OMD was enriched in SMCs around calcified nodules in aortic media of nephrectomised rats and in plaques from ApoE-/- mice on warfarin. In vitro experiments revealed that OMD mRNA was upregulated in SMCs stimulated with IFNγ, BMP2, TGFβ1, phosphate and β-glycerophosphate, and by administration of recombinant human OMD protein (rhOMD). Mechanistically, addition of rhOMD repressed the calcification process of SMCs treated with phosphate by maintaining their contractile phenotype along with enriched matrix organisation, thereby attenuating SMC osteoblastic transformation. Mechanistically, the role of OMD is exerted likely through its link with SMAD3 and TGFB1 signalling, and interplay with BMP2 in vascular tissues. Conclusion: We report a consistent association of both circulating and tissue OMD levels with cardiovascular calcification, highlighting the potential of OMD as a clinical biomarker. OMD was localised in medial and intimal α-SMA+ regions of calcified cardiovascular tissues, induced by pro-inflammatory and pro-osteogenic stimuli, while the presence of OMD in extracellular environment attenuated SMC calcification.

Objective Ischaemic strokes can be caused by unstable carotid atherosclerosis, but methods for identification of high risk lesions are lacking. Carotid plaque morphology imaging using software for visualisation of plaque components in computed tomography angiography (CTA) may improve assessment of plaque phenotype and stroke risk, but it is unknown if such analyses also reflect the biological processes related to lesion stability. Here, we investigated how carotid plaque morphology by image analysis of CTA is associated with biological processes assessed by transcriptomic analyses of corresponding carotid endarterectomies (CEAs). Methods Carotid plaque morphology was assessed in patients undergoing CEA for symptomatic or asymptomatic carotid stenosis consecutively enrolled between 2006 and 2015. Computer based analyses of pre-operative CTA was performed to define calcification, lipid rich necrotic core (LRNC), intraplaque haemorrhage (IPH), matrix (MATX), and plaque burden. Plaque morphology was correlated with molecular profiles obtained from microarrays of corresponding CEAs and models were built to assess the ability of plaque morphology to predict symptomatology. Results Carotid plaques (n = 93) from symptomatic patients (n = 61) had significantly higher plaque burden and LRNC compared with plaques from asymptomatic patients (n = 32). Lesions selected from the transcriptomic cohort (n = 40) with high LRNC, IPH, MATX, or plaque burden were characterised by molecular signatures coupled with inflammation and extracellular matrix degradation, typically linked with instability. In contrast, highly calcified plaques had a molecular signature signifying stability with enrichment of profibrotic pathways and repressed inflammation. In a cross validated prediction model for symptoms, plaque morphology by CTA alone was superior to the degree of stenosis. Conclusion The study demonstrates that CTA image analysis for evaluation of carotid plaque morphology, also reflects prevalent biological processes relevant for assessment of plaque phenotype. The results support the use of CTA image analysis of plaque morphology for risk stratification and management of patients with carotid stenosis.