Decreased miR-143-3p in Advanced Carotid Atherosclerosis

miRNAs in extracellular vesicles were isolated from plasma samples of patients with advanced carotid atherosclerosis and donors without manifested atherosclerosis to assess the relative levels of miR-143-3p by real-time PCR. Median miR-143-3p levels were 70.6% lower in the plasma of patients with carotid stenosis (Fig. 1a). No significant difference was observed in terms of sex and diabetes (Fig. S1). A receiver operating characteristic curve analysis showed an area under the curve of 0.79 (p = 0.0027) and a cut-off value < 0.34 produced a sensitivity of 71.4% and a specificity of 84.6% for advanced disease (Fig. 1b). This confirms that miR-143-3p could be a potential biomarker of carotid stenosis and atherosclerosis in humans.

Fig. 1

Plasma detection of miR-143-3p in carotid stenosis patients. (a) miRNAs in extracellular vesicles were isolated from the plasma of patients with advanced atherosclerosis and the plasma from healthy donors without manifest atherosclerosis. The relative expression of miR-143-3p was measured by real-time PCR (Mann-Whitney test, healthy donors n = 13, carotid stenosis n = 28, log2-scaled y-axis) (b) A receiver operating characteristic curve analysis illustrates how miR-143-3p can be used as a biomarker for carotid atherosclerosis

Altered miRNAs With Age and Atherosclerosis Severity in Mice

By comparing young (11 weeks) and middle-aged (46 weeks) female HuBL mice we were able to assess atheroprogression (Fig. 2a). The median atherosclerotic burden in the en-face-prepared aortic arches was 28.4 times higher in the 46-week group (Fig. 2b). Oil Red O staining of the aortic root showed 3.9 times higher lipid-stained lesions in the 46-week group (Fig. 2c). Body weight, spleen weight, splenocyte count, and blood leukocytes were not significantly altered with age in this cohort of mice (Table S4).

Fig. 2

Dysregulation of a set of miRNAs in mouse carotid arteries. (a) Experimental setup comparing atheroprogression in HuBL mice. (b) Micrographs of Sudan-IV-stained aortic arches with a 2 mm scale bar and quantification of the atherosclerotic plaques (orange color) divided by total aortic arch area. (c) Micrographs of Oil Red O-stained sections 300 µm from the aortic root with a 500 µm scale bar. Mean aortic root lipids were quantified (red color) and divided by aortic cross-section area. (d) Sudan-IV stained aortic arches with branches to depict microdissection of common carotids and carotid bifurcation. (e) Micrograph of a carotid artery from an 11-week-old HuBL mouse, the arrow indicates Sudan-IV+ plaque in the carotid bifurcation. (f-g) Relative quantification of miRNAs in the carotid bifurcation (bright colors) and common carotids (pale colors) from the 11-week (pink colors, n = 6) and 46-week (purple colors, n = 4) old female HuBL mice. (h) Linear regression between lipid deposition in the aortic root and the levels of miR-143-3p in the carotid bifurcation and between lipid deposition in the aortic root and the levels of miR-155-5p and miR-143-3p in the common carotids. (i) Plasma concentration of cholesterol and triglycerides

The carotid bifurcation and common carotids were microdissected as displayed in a set of Sudan-IV-stained arteries (Fig. 2d). Initial lesion formation was observed in the carotid bifurcation at 11 weeks (Fig. 2e). Relative levels of miR-15a-5p, miR-143-3p, miR-155-5p, and miR-199a-3p were assessed but only the expression of miR-143-3p was significantly downregulated in the carotid bifurcation of the middle-aged HuBL mice (Fig. 2f). In the common carotids, miR-155-5p levels were significantly increased with age in addition to the downregulated miR-143-3p (Fig. 2g). Further supporting these changes, carotid miR-143-3p levels showed a negative association with disease progression as measured by Oil Red O staining in the aortic roots, and miR-155-5p levels in the common carotids showed a positive association with disease progression (Fig. 2h). The age difference between the two groups did not significantly impact plasma cholesterol and triglyceride levels (Fig. 2i). This would indicate that the disease process itself, and not hypercholesterolemia, drives these local miRNA changes at predilection sites in the vasculature.

To assess whether there was site-specific miRNA dysregulation in either early or advanced atherosclerosis, we compared the carotids separately for the 11- and 46-week age groups (Fig. 3a-b). The most striking changes were observed when comparing miRNA levels between the common carotids and the carotid bifurcation in early atherosclerosis: miR-15a-5p, miR-143-3p, and miR-199a-3p were significantly downregulated, and miR-155-5p was upregulated (Fig. 3a). In advanced atherosclerosis, there was no significant difference in miRNA expression between the carotid bifurcation and common carotids (Fig. 3b), likely due to the extension of the disease to the common carotids at this time point.

Fig. 3

Site-specific miRNA level changes in early and late carotid atherosclerosis. (a-b) Relative quantification of miRNAs in the common carotids (pale colors) and carotid bifurcation (bright colors) from the 11-week (pink colors, n = 6) and 46-week (purple colors, n = 4) old female HuBL mice. (c) mRNA levels for genes of interest in 11-week-old mice. (d) Linear regression between Ccl2 mRNA and miR-155-5p (left) and miR-199a-3p (right) in carotids with early atherosclerosis. (e) mRNA levels for genes of interest in 46-week-old mice. (f) Linear regression between Cd68 mRNA (left) and Vcam1 mRNA (right), respectively, and miR-143-3p in carotids with advanced atherosclerosis

mRNA levels for a handful of genes related to atherosclerosis progression were investigated in a site-specific manner in the carotids. In early atherosclerosis, only Ccl2 mRNA, encoding monocyte chemoattractant protein 1, was overexpressed in the carotid bifurcation compared with the common carotid (Fig. 3c). Ccl2 mRNA showed a strong positive relationship with miR-155-5p expression and was inversely correlated with miR-199a-3p levels (Fig. 3d). In advanced atherosclerosis, Vcam1 mRNA, encoding vascular cell adhesion molecule 1, was increased in the common carotids and Cd68, a highly expressed mRNA in macrophages, was increased in the carotid bifurcation (Fig. 3e), indicating that these sites were in different phases of atheroprogression. Cd68 mRNA levels were negatively associated with miR-143-3p whereas Vcam1 mRNA levels were positively associated with miR-143-3p (Fig. 3f). Although these associations were very weak, they support the notion that miR-143-3p is dysregulated in advanced atherosclerosis.

Since there was a site-specific alteration of mRNA levels, we wanted to assess whether the expression was altered by disease severity. When comparing the mRNA levels of the selected genes of interest in the carotid bifurcation of 11 vs. 46-week mice, Nos2 mRNA, encoding inducible nitric oxide synthase, and Ccl2 mRNA were found to be increased (Fig. S2a). However, these transcript levels did not correlate with any of the measured miRNAs. In the common carotids, there was an increase in mRNA levels of Vcam1, Ccl2, and Il1b, the latter encoding the interleukin-1β cytokine precursor, in middle-aged mice (Fig. S2b). The expression of these three transcripts showed a positive relationship with miR-155-5p levels (Fig. S2c). At the same time, Ccl2 and Vcam1 mRNA were inversely correlated with miR-143-3p expression (Fig. S2d). This further denotes the differential expression of miR-143-3p and miR-155-5p in the common carotids, which is a vascular site that typically has laminar blood flow and is affected late in the atherosclerotic process.

Altered miRNAs in Immune Cells of Atherosclerotic Mice

We investigated whether the selected miRNAs were altered in immune cells during early and advanced carotid atherosclerosis since previous studies mainly focused on their role in resident vascular cells. CD3+ T cells and CD3− non-T cells were isolated from the spleen of young and middle-aged mice (Fig. 4a-b). When studying the miRNA expression in the splenic CD3+ T cells, miR-143-3p was selectively increased in mice with advanced atherosclerosis, and miR-199a-3p was non-detectable in those cells (Fig. 4c). A handful of transcripts related to the T-cell phenotype were analyzed. In mice with advanced atherosclerosis, Foxp3, Ifng, Il21, and Pdcd1 mRNA levels were increased in the splenic CD3+ cells indicating that atheroprogression drives T-cell activation and regulatory transcriptional programs (Fig. 4d). Pdcd1 mRNA, encoding the immune checkpoint programmed cell death protein 1 (PD1), and Tbx21 mRNA, encoding the T-box transcription factor expressed in T-helper type 1 cells, were positively correlated with miR-143-3p expression in those cells, implicating this miRNA in the context of immune regulation (Fig. 4e).

Fig. 4

miRNA levels in splenic immune cells during atheroprogression. (a) Experimental setup of CD3+ and CD3− cell isolation. (b) Cell fractions before and after isolation of CD3+ cells from the spleen. (c) miRNA levels in CD3+ cells from the 11-week (pink, n = 6) and 46-week (purple, n = 4) old female HuBL mice. (d) mRNA levels for genes of interest in CD3+ cells. (e) Linear regression between miR-143-3p in CD3+ cells and Pdcd1 and Tbx21 mRNA. (f) miRNA levels in CD3− cells. (g) mRNA levels in CD3− cells. (h-i) miRNA levels in the CD3+ and CD3− cells from the 11- and 46-week groups. (j-k) Germinal center B cells as quantified using flow cytometry from the mediastinal and renal lymph nodes and their correlation to miR-199a-3p in splenic CD3− cells. (l-m) PD1high TEM cells in the mediastinal and renal lymph nodes and their correlation to miR-143-3p in splenic CD3+ cells

When analyzing the relative expression of miRNAs in the CD3− cells, only miR-199a-3p was increased during atheroprogression (Fig. 4f). A few transcripts involved in inflammatory pathways were analyzed, and mRNA from the MHC class II encoding gene H2ab1 was found to be significantly decreased in the non-T cell fraction (Fig. 4g), indicating that antigen presentation could be differentially regulated during atheroprogression in the spleen. To observe specific enrichments in the immune cell compartment during atheroprogression, we compared if there was an alteration in miRNAs between splenic CD3− and CD3+ cells in early and advanced atherosclerosis. In early atherosclerosis, significant enrichments of miR-15a-5p and miR-155-5p were found in CD3+ cells (Fig. 4h), while this enrichment was subdued in advanced atherosclerosis (Fig. 4i).

Specifically, cellularity in the aorta-associated mediastinal and renal lymph nodes was increased, while cellularity remained at a similar level in the spleen (Fig. S3a-b and Table S4). An expansion of germinal center B cells and plasma cells was observed in the mediastinal and renal lymph nodes as analyzed by flow cytometry (Fig. 4j and S3c-f). The expansion of plasma cells was, at this time point and location, limited to the absolute cell counts (Fig. S3d-e). As miR-199a-3p is known to be upregulated in germinal center B cells [19], this miRNA correlated with the germinal center reactions (Fig. 4k, Fig. S3n).

Blood lymphocytes as well as total T cell number in the mediastinal and renal lymph nodes were not significantly different between groups (Fig. S3g-h and Table S4). However, there was an increase in the total and relative number of T effector/memory (TEM) cells (Fig. S3i-j). Confirming the upregulated Pdcd1 mRNA levels, PD1high TEM cells were found to accumulate in the aorta-associated lymph nodes during atheroprogression (Fig. 4l and S3l). Speculatively, persistent presentation of atherosclerosis-associated antigens may be driving these effects. The TEM cell with intermediate PD1 expression remained unaltered (Fig. S3k-m). Expression of miR-143-3p in T cells correlated with the number of PD1high TEM cells (Fig. 4m). Since the correlation to Tbx21 mRNA was stronger, this might reflect an upregulation in activated T cells or Th1 cells more specifically. However, this could not be distinguished using our experimental design.

Altered miRNAs in Middle-aged Mice

Finally, we investigated whether the dysregulation of miRNAs occurs during atheroprogression without the influence of age differences. Age-matched one-year-old male Ldlr−/− and HuBL mice fed a standard chow diet were used for this purpose (Fig. 5a). The atherosclerotic burden was 7.0-fold increased in the aorta of HuBL mice as driven by more severe hypercholesterolemia induced by transgenic overproduction of APOB100 (Fig. 5b-c). No difference in body weight was recorded but HuBL mice had higher blood counts of lymphocytes, monocytes, and granulocytes (Table S5) as well as plasma cholesterol and triglyceride levels (Fig. 5c). To analyze miRNA levels, the iliac lymph nodes in proximity to the aorta were used. Levels of miR-15a-5p, miR-143-3p, and miR-199a-3p were lower in HuBL compared to Ldlr−/− mice (Fig. 5d). This is similar to our findings in the vasculature of female HuBL mice at different ages but in the iliac lymph nodes of those mice, only miR-143-3p was found to be significantly decreased (Fig. S4).

Fig. 5

Dysregulated transcripts in iliac lymph nodes of middle-aged mice. (a) Experimental setup. (b) Micrographs of Sudan-IV-stained aortas with a 4 mm scale bar. (c) Total plasma cholesterol and triglycerides concentrations. (d) Relative quantification of miRNAs in the iliac lymph nodes from male Ldlr−/− (pink color, n = 8) and HuBL mice (purple color, n = 8). (e–f) Linear regression between miR-15a-5p and miR-199-3p with plasma cholesterol and triglycerides. (g) mRNA levels for genes of interest in the iliac lymph nodes. (h) Linear regression between Pdcd1 and Ifng mRNA, respectively, and miR-15a-5p levels in the iliac lymph nodes

miR-15a-5p showed a negative association with plasma lipid levels, indicating its inverse association with hypercholesterolemia (Fig. 5e). A weaker association was observed between miR-199a-3p and plasma cholesterol (Fig. 5f). miR-143-3p did not correlate with hypercholesterolemia levels, indicating that its dysregulation is more associated with disease progression than dyslipidemia. However, atheroprogression is tightly linked to hypercholesterolemia in these mice, and separating these entities is challenging.

A few selected transcripts related to T-cell phenotype were analyzed. Gata3 mRNA, encoding a transcription factor involved in T-helper type 2 responses, Ifng mRNA, encoding interferon-γ, Il21 mRNA, encoding the T follicular helper-related cytokine interleukin-21, Rorc mRNA, encoding retinoic acid receptor-related orphan receptor C associated with T-helper 17 cells, and Pdcd1 mRNA were all observed to be lower in the HuBL mice (Fig. 5g). Pdcd1 and Ifng correlated with miR-15a-5p expression (Fig. 5h). Taken together, severe dyslipidemia was identified as a contributing mediator of dysregulation of miR-15a-5p and, to a lesser extent, miR-143-3p.