Abstract
Patients with hyperlipidemia are at risk of atherosclerosis, but not all develop cardiovascular disease, highlighting the importance of other risk factors such as inflammation. Both the innate and adaptive arms of the immune system have been suggested in the initiation and propagation of plaque formation. Tri-partite motif (TRIM) 21 is a regulator of tissue inflammation and pro-inflammatory cytokine production, and has been implicated in chronic inflammatory disease. Here, we investigate a potential role for TRIM21 in coronary artery disease.
Trim21-deficient or wild-type bone marrow was transplanted into Ldlr-/- mice fed a hypercholesterolemic diet. The Trim21-/-->Ldlr-/- mice developed larger atherosclerotic plaques, with significantly higher collagen content compared to mice transplanted with wild-type cells. High collagen content of the atheroma is stabilizing, and has recently been linked to IL-17. Interestingly, Trim21-/-->Ldlr-/- mice had elevated CD4 and IL-17 mRNA expression in plaques, and increased numbers of activated CD4+ T cells in the periphery. An increased differentiation of naïve T cells lacking Trim21 into Th17 cells was confirmed in vitro, with transcriptomic analysis revealing upregulation of genes of a non-pathogenic Th17 phenotype. Also, decreased expression of matrix metalloproteinases (MMPs) was noted in aortic plaques. Analysis of human carotid plaques confirmed that TRIM21 expression negatively correlates with the expression of key Th17 genes and collagen, but positively to MMPs also in patients, linking our findings to a clinical setting.
In this study, we demonstrate that TRIM21 influences atherosclerosis via regulation of Th17 responses, with TRIM21 deficiency promoting IL-17 expression and a more fibrous, stable, phenotype of the plaques.
1. Introduction
Cardiovascular disease is the leading cause of mortality worldwide,1 with atherosclerosis recognized as a major culprit behind most cardiovascular events.2 Atherosclerosis is a complex process in which dyslipidemia, accumulation of cholesterol-rich low-density lipoprotein (LDL) particles, and chronic inflammation all contribute to the formation of atheromatous plaques.3 However, not all individuals with dysregulated lipid metabolism and signs of atherosclerosis develop cardiovascular events, indicating that other factors, both intrinsic and environmental, play important roles in the progression to clinical disease. Notably, innate and adaptive immune responses both contribute to the pathophysiology of atherosclerosis and are strongly influenced by host factors. While fat-loaded macrophages drive the initial local activation of innate immunity and recruitment of inflammatory cells, T cells are also rapidly recruited to the inflamed tissue and developing atheromatous plaques.3,4 The pro-atherogenic effect of Th1 cells is well documented.5 By contrast, the role of Th17 cells remains unclear. Although Th17 cells exert a pathogenic effect in several autoimmune and inflammatory diseases, they have also recently been proposed to play a protective role in the progression towards cardiovascular disease. Indeed, IL-17 has been linked to increased collagen content and stabilization of atherosclerotic plaques,6 thereby decreasing the risk of plaque rupture and subsequent myocardial infarction.
TRIM21 is an ubiquitin E3 ligase predominantly expressed in haematopoietic cells. It was recently shown to play a critical role in controlling innate immune responses and tissue inflammation via the regulation of cytokine production.7,8 Here, we show that Trim21 deficiency in the haematopoietic cell compartment leads to enlarged plaques with enriched collagen content through increased T cell-mediated IL-17 responses in the Ldlr-/- mouse model of atherosclerosis. Further, we for the first time show that Trim21 deficiency leads to an increased propensity of Th17 differentiation of a non-pathogenic phenotype. These findings are supported by the reverse correlation of TRIM21 expression and the expression of key Th17 molecules as well as collagen in human carotid plaques.
2. Methods
2.1 Patients and gene expression analysis
Gene-gene expression associations were analysed in the BiKE cohort, which was described previously.9,10 In brief, mRNA from carotid plaques (n = 127 patients) was collected from patients undergoing surgical endarterectomies due to severe carotid stenosis. Gene expression was analysed by Affymetrix HG-U133 Plus 2.0 Genechip array. The study was conducted according to the Helsinki declaration and approved by the regional Ethics committee. All subjects provided written informed consent.
2.2 Animals and bone marrow transplantations
Trim21-/- mice were generated on a C57BL/6 J background as previously described.7 For the atherosclerosis experiments, bone marrow from femur and tibia was extracted from female Trim21-/- and +/+ littermates, and injected intravenously into lethally irradiated female Ldlr-/- mice on the C57BL/6 J background (two doses of 700 rad, 3 h apart). Each recipient received 5 x 106 cells in 50–100 μL sterile PBS. Mice were treated with sulfadiazine and trimethoprim for the first 3 weeks and left to reconstitute for a total of 6 weeks. After recovery, mice were fed a high fat diet (D12108C containing 40 kcal% fat, 1.25% cholesterol, 0% cholic acid; Research Diets, USA). For the Th differentiation assays, Trim21-/- and +/+ littermates were used. Mice were euthanized by CO2 asphyxiation. The local Region North Ethics committee in Stockholm, Sweden, approved the animal studies and all experimental procedures were performed in accordance with the Directive 2010/63/EU of the European Parliament.
2.3 Tissue preparation, lesion analysis, and immunohistochemistry
Blood for plasma preparation was retrieved from live mice via tail vein puncture 2 days prior to euthanization. Blood from euthanized mice was collected by cardiac puncture for serum preparation and the vasculature was thereafter perfused with sterile RNase-free PBS. The aorta (with adventitia) and para-aortic lymph nodes were dissected and snap frozen for later RNA isolation. Hearts, including the aortic root up to the brachiocephalic artery, were collected and preserved by snap-freezing for immunohistochemistry and quantitative lesion analysis. Spleen, inguinal and axillar lymph nodes were retrieved and kept in PBS on ice until preparation for flow cytometric analyses.
Aortic lesions were analysed as previously described.6,11 In brief, hearts were cryosectioned from the proximal 1 mm of the aortic root. Lesion size was evaluated in eight consecutive sections at 100 μm distance with haematoxylin and Oil red O. Lesion surface area and total vessel area were quantified using Image J software (NIH). Relative lesion area was calculated as 100x mean lesion area/total vessel area, where the latter is defined as the area inside the external elastic lamina–lumen area.11,12 Thus, relative lesion area compensates for oblique sectioning, which may be a confounding factor in quantitative lesion analysis.21
Acetone-fixed sections were stained with primary rat anti-mouse antibodies to CD4, CD8, CD68 and VCAM-1 (BD Biosciences) followed by biotinylated rabbit anti-rat IgG or anti-mouse I-Ab (directly biotinylated, BD Biosciences). Mouse α-smooth muscle actin was detected by primary mouse IgG2a antibodies (Sigma-Aldrich) followed by a mouse-on-mouse kit (Vector Laboratories) with mouse IgG2a (BD Biosciences) as isotype control. Bound antibodies were visualized using the Vectastain ABC kit (Vector Laboratories). The size of the necrotic core was measured in haematoxylin stained sections and was defined as the area with <5 nucleated cells, divided by the total plaque size.13 A 3000 μm2 minimum threshold was used to avoid inclusion of regions that do not represent necrotic core areas. Stainings were quantified by Leica QWin Standard Y 2.8 software (Leica Microsystem Imaging Solutions). Collagen content in plaques was assessed in Picrosirius Red (PSR) stained sections under linear polarized light as previously described,14 allowing for the evaluation of total collagen content, as well as the amount of mature collagen fibres identified as thick, orange-red coloured fibres, and thin immature fibres of green colour. The fibrous cap was defined as the collagenous and SMC containing structure overlaying the necrotic core. Thickness was measured in PSR stained sections, as previously described.15,16
Due to technical issues, not all mice could be analysed for all immunohistochemical markers. The number of mice analysed for respective marker is defined in the figure legends.
2.4 Flow cytometry
Single cell-suspensions from spleen and lymph nodes were incubated with mouse Fc-receptor block (BD Biosciences) for 15 min at 4 °C and subsequently stained for 20 min at 4 °C with the following fluorescent antibodies (all from Biolegend, except if otherwise indicated): CD4-PcP, CD3-APC, CD62L-PE (BD Biosciences), CD44-APC, CD19-PE, CD11b-PcP, IFNγ-PE (BD Biosciences), IL4-PE, IL17-PE (BD Biosciences). Samples were run on a Dako CyAn flow cytometer (Beckman Coulter) and analysed with FlowJo 7.6 software.
2.5 Cholesterol measurement, oxLDL ELISA
Total plasma cholesterol and triglycerides were determined using an enzymatic colorimetric kit according to the manufacturer’s description (Randox Laboratories). Antibody titers to oxLDL were determined in serum by ELISA, as described previously.17 Briefly, wells were coated with oxLDL 10 μg/mL over night, whereupon sera diluted 1:10 in 1% gelatin were added and incubated for 2 h. Bound IgG was detected with an ALP-conjugated anti-mouse IgG antibody (Dako).
2.6 RNA isolation, cDNA synthesis and quantitative real-time PCR analysis
RNA was isolated from indicated cells and tissues using the RNeasy Lipid Mini kit for aortic RNA, and the RNeasy Mini kit for all other samples (Qiagen), according to the manufacturer’s instructions. cDNA was generated by reverse transcription using Superscript-II and random hexamers (Invitrogen). Amplification was performed by SYBR green real-time PCR (Invitrogen) for indicated genes, apart from Mmp-8, Mmp-9 and Spp1 for which TaqMan Assays-on-Demand primers and probes were used (Applied Biosystems). Primer sequences are available upon request. The ΔΔCt method adjusted for amplification efficiency was used to calculate expression relative to the housekeeping gene Hprt (BioRad).
2.7 In vitro T helper cell differentiation
Spleens and lymph nodes were collected from Trim21+/+ and Trim21-/- mice upon euthanization by CO2 asphyxiation and organs were subjected to mechanical dissociation and red blood cell lysis. Naïve T cells were first purified by negative magnetic bead separation using CD4+ T cell isolation kit II (Milteyi), and thereafter by FACS-sorting for CD4+CD62L+CD44- cells. Cells were cultured in 96-well plates pre-coated with anti-CD3 1 μg/mL (BD) at 105 cells/well in RPMI medium (Sigma) supplemented with 10% FSC, 1% HEPES, 1% sodium pyruvate, 1% L-glutamine, 1% penicillin-streptomycin and 50 μM β-Mercaptoethanol. The cells were stimulated with soluble anti-CD28 (BD) at 1 μg/mL and with different combinations of cytokines: For Th1 conditions, IL-12 (10 ng/mL) and anti-IL-4 (50 ng/mL); for Th2 conditions, IL-4 (10 ng/mL) and anti-IFNγ (10 μg/mL); for Th17 conditions, murine TGF-β1 (2 ng/mL), murine IL-6 (20 ng/mL), anti-IFNγ (10 μg/mL), and anti-IL-4 (10 μg/mL). Cells were re-plated on non-coated plates day 3 and IL-2 (5 ng/mL) was added to the Th1 and 2 cultures, and IL-23 (20 ng/mL) was added to the Th17 cultures. On day 5, cells were prepared for flow cytometric analysis or qPCR. For flow cytometeric analysis, cells were treated with GolgiPlug (BD Biosciences), PMA (50 ng/mL), and Ionomycin (1 μg/mL) for 4 h. Cell surface staining with CD4-PcP and intracellular cytokine staining was performed for IFNγ-PE, IL-4-PE or IL-17-PE (Biolegend). For qPCR analysis, cells were pelleted, resuspended in RLT buffer and stored at –70 °C until mRNA extraction.
2.8 Gene expression analysis of in vitro differentiated Th17 cells
Total mRNA from sorted CD4+CD62L+CD44- T cells of Trim21+/+ and Trim21-/- mice, and cells taken 48 h following in vitro Th17 differentiation, was extracted using the RNAeasy Micro Kit (Qiagen, Venlo, The Netherlands) and expression was analysed with a custom made Th17 signature code set (NanoString Technologies, Seattle, WA).18 Data was expressed as relative to the mean of included reference genes and gene expression below the negative controls was considered as no expression and given a value of 0. Values were transformed into log2 Z-score. Batch effects were minimized by randomized placement of samples across the arrays and all Nanostring cartridges were from the same lot.
2.9 Statistical analysis
For comparison of groups the Mann-Whitney U test was used, and a two-way ANOVA with Bonferroni post-test was performed when comparing difference in lesion size. Values are expressed as mean ± SEM unless otherwise indicated. Genes differentially regulated (P < 0.05) between Trim21-/- and +/+ animals measured using the Nanostring Th17 custom codeset were identified using the ComparativeMarkerSelection (CMS) Module in the Broad Institute GenePattern program. Pearson correlation coefficient was used for calculation correlations. For analysis of differences in the expression of Th17-associated genes in subjects expressing the highest and lowest levels of TRIM21 in plaques (n = 8 in each group), differentially regulated genes (P < 0.05) were identified using CMS and tested for enrichment of Th17 regulators utilizing the defined gene set in Ciofani et al.19 by hypergeometric means test. Prism GraphPad 7 was used for all other statistical analyses. P-values <0.05 were considered significant.
3. Results
3.1 Trim21 deficiency leads to larger atherosclerotic plaques with increased collagen content in Ldlr-/- mice
TRIM21 is predominantly expressed in haematopoietic cells and has been shown to control innate immune responses and tissue inflammation.7 We therefore hypothesized that TRIM21 may influence the development of atherosclerosis by modulating the inflammatory component. To investigate a potential involvement of TRIM21 in atherosclerosis, we transferred bone marrow from Trim21+/+ or Trim21-/- mice into lethally irradiated Ldlr-/- mice, which develop hypercholesterolemia and subsequent atherosclerosis when fed a high-fat diet. We observed that, in the absence of Trim21, Ldlr-/- mice developed significantly larger atherosclerotic plaques in the aorta after 6 weeks of high-fat diet (Figure 1A and B). Our previous studies have demonstrated that Trim21 is a negative regulator of pro-inflammatory responses, and that long-term inflammation will lead to more pronounced differences in the absence of Trim21.7 We therefore quantified plaque formation in bone marrow chimeric mice maintained on a high-fat diet during a prolonged period of 12 weeks. Here, Trim21-/- bone marrow chimeras developed even larger plaques, with substantially increased relative and actual lesion size (Figure 1C and D). Interestingly, we also observed a significant increase in plaque collagen content as visualized by Pircosirius Red (Figure 1E). As increased collagen content is associated with plaque stability, we therefore further assessed the necrotic core and fibrous cap. The fibrous cap was significantly increased in Trim21-/-->Ldlr-/- chimeric mice (Figure 1F), however no significant differences were observed in the size of the necrotic core (Figure 1G). Further, the ratio of mature collagen of total collagen was increased in the mice that had received Trim21-/- bone marrow (Figure 1H). In all, we observed larger plaques developing in Trim21-/-->Ldlr-/- chimeric mice, however with multiple signs of increased plaque stability, such as increased collagen content and thicker fibrous cap.
Trim21 deficiency increases atherogenesis and fibrous content of plaques. (A) Lethally irradiated Ldlr-/- mice were transplanted with bone marrow (BM) from Trim21+/+ or Trim21-/- mice, and fed a high-fat diet for 6 weeks (n = 8 in each group). Trim21+/+->Ldlr-/- depicted with open circles and Trim21-/-->Ldlr-/- with black boxes. Trim21-/-. Micrographs of representative Oil red O stains. (B) Quantification of lesion size from eight consecutive sections, 100 to 800 μm from the aortic sinus. (C, D) Same as A and B, but Trim21+/+ and Trim21-/- BM transplanted mice were kept on high-fat diet for 12 weeks (n = 8 and 8, respectively). (E) Collagen content visualized in polarized light after Picrosirius Red staining (n = 8; 9), (F) the thickness of the fibrous cap quantified by anti-Smooth Muscle actin staining (n = 8; 9) (G) area of the necrotic core (n = 7; 9), and (H) the ratio of mature of total collagen (n = 8; 9) in plaques of mice fed high-fat diet for 12 weeks. Stains displayed at x10 magnification.*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. (Mann-Whitney U-test, ANOVA with Bonferroni post-test; mean ±SEM).
3.2 An attenuated Th17 response exacerbates atherogenesis in Trim21-/-->Ldlr-/- chimeric mice
To understand how Trim21 deficiency affects atherogenesis, we next assessed both tissue specific and systemic responses in Ldlr-/- chimeric mice. Locally, we observed higher levels of CD4 and IL-17 A mRNA transcripts in the aorta of Trim21-/- transplanted mice, whereas no difference in IFNγ expression was seen (Figure 2A). Decreased expression of I-Ab, an IFNγ induced MHC class II protein, was also detected in Trim21-/-->Ldlr-/- chimeric mice, further supporting that development of larger plaques in these mice was not primarily driven by IFNγ-producing Th1 cells (Figure 2B). In concordance with the increased CD4 mRNA levels in the aortas, more infiltrating CD4+ cells were also observed (Figure 2C), but not CD8+ cells (Figure 2D). Of note, despite the larger plaques in Trim21-/-->Ldlr-/- mice, no difference in the abundance of the adhesion molecule VCAM-1 or CD68+ macrophages was found in the lesions compared to Trim21+/+->Ldlr-/- mice (Figure 2E and F). At a systemic level, we observed increased frequencies of CD3+, CD4+ and activated CD4+CD62Llow T cells in the lymph nodes and spleen of Trim21-/-->Ldlr-/- chimeras compared to Trim21+/+->Ldlr-/- mice (Figure 3A and B). By contrast, the frequency of CD11b+ cells, i.e. mostly macrophages, was not increased in Trim21-/- transplanted mice. Also, anti-oxidized LDL IgM and IgG antibody levels reached similar levels in Trim21-/- as in Trim21+/+->Ldlr-/- chimeras (Figure 3C). No significant differences in plasma cholesterol and triglyceride concentrations or body weight were detected (Figure 3D), further emphasizing that transplantation of Trim21-deficient bone marrow primarily influences the immune component of the atherosclerotic process in Ldlr-/- mice. In summary, our data indicate a Th17-specific effect of Trim21 deficiency in atherosclerotic plaque formation.
Increased atherosclerosis associates with local T cell activation and IL-17 expression in Trim21-deficient mice. Mice fed a high-fat diet for 12 week (n = 8 and 10 respectively) were analysed. Open circles depict Trim21+/+ and black boxes Trim21-/- transferred mice. (A) Gene expression of CD4, IL-17 A, and IFNγ in the aorta assessed by RT-PCR relative to HPRT. (B–F) Quantification and representative stains of (B) I-Ab (n = 5; 9), (C) CD4 (n = 5; 9), (D) CD8 stains (n = 5; 9) (x40 magnification), (E) CD68 (n = 7; 9) and (F) VCAM-1 (n = 7; 9) (x10 magnification). *P < 0.05, ****P < 0.0001. (Mann-Whitney U-test; mean ±SEM).
Systemic activation of T cells in Trim21-deficient mice during atherogenesis. Mice fed a high-fat diet for 12 week (n = 8 and 10, respectively); open circles depict Trim21+/+ and black boxes Trim21-/- transferred mice. (A–B) Frequencies of lymphocytes determined by flow cytometry in spleen (A) and lymph nodes (B). T cells were identified as all CD3+ cells in lymphocyte gate, T-helper cells as CD4+ of the latter and activated T helper cells as CD62L low expressers. (C) Titers of anti-oxidized LDL antibodies measured with optical density. (D) Quantification of plasma cholesterol and triglycerides (n = 8; 8), and weight (n = 8; 10). *P < 0.05, ***P < 0.001. (Mann-Whitney U-test; mean ±SEM).
3.3 An augmented intrinsic capacity for Th17 differentiation in Trim21-deficient naïve T cells
We next asked whether the expansion of CD4+ T cells and increased IL-17 A expression we observed in the Trim21-/-->Ldlr-/- mice was due to an intrinsic propensity of Trim21 deficient T cells to differentiate into Th17 cells. To address this question, we cultured purified naïve CD4+CD62L+CD44- T cells in the presence of Th-differentiating cytokines. Trim21 deficiency did not lead to any significant differences in either the Th1 or Th2 differentiation (Figure 4A and B). By contrast, we found that a larger proportion of Trim21-/- T cells produced IL-17 compared to Trim21+/+ cells when cultured in Th17-differentiating conditions, as detected by both intracellular cytokine staining and quantitative RT-PCR (Figure 4C). Our data thus reveal that Trim21 negatively regulates generation of Th17 cells in a cell-intrinsic manner.
Enhanced Th17 differentiation of naïve Trim21-/- CD4+ T cells. CD4+CD44- CD62L+ naïve T cells were cultured under conditions promoting T helper cell differentiation for 5 days. Cells were analysed by flow cytometry and RT-PCR. For flow cytometric analysis, live CD3+ T cells were stained and analysed for the expression of lineage specific cytokines. Trim21+/+ open circles and Trim21-/- black boxes (n = 6 in each group), representative data of three independent experiments. (A) Cells differentiated into Th1, (B) Th2 cells, and (C) Th17 cells. (D) Gene expression analysis of T cells undergoing Th17 differentiation, based on genetic signatures of pathogenic and non-pathogenic Th17 cells. (n = 6 Trim21+/+ and 5 Trim21-/-) *P < 0.05, **P < 0.01. (Mann-Whitney U-test; mean ±SEM).
IL-17 producing T cells are commonly found in chronic inflammatory diseases, and are thought to propagate autoimmune disease. However, in recent years regulatory, non-pathogenic Th17 cells have been described.20,21 These cells are identified through specific gene expression pattern and differentiation through TGFβ1 rather than TGFβ3.21 Interestingly, TGFβ1 is the most prominent TGFβ isoform in the vasculature, expressed by all cell types in the plaque and associated with collagen production.22 We therefore investigated the gene expression of Trim21-/- and Trim21+/+ cells during in vitro Th17 differentiation using a Th17 gene signature Nanostring code set to understand which Th17 subset that dominated.18 Notably, a distinct, TGFβ1 driven non-pathogenic phenotype could be observed in Trim21-/- cells, which, via its role in promoting collagen production, can be linked to the more stable plaques observed in the Trim21-/-->Ldlr-/- chimeras fed high-fat diet (Figure 4D).
3.4 TRIM21 expression inversely correlates with regulators of Th17 differentiation in human atherosclerotic lesions
Our findings show that, under hyperlipidemic conditions, loss of Trim21 in the haematopoietic compartment leads to the formation of atherosclerotic plaques with increased IL-17 expression and higher collagen content. Mechanistically, this was linked to a greater propensity of Trim21-deficient T cells to differentiate into IL-17-producing cells. To understand if these findings are relevant to human cardiovascular disease, we analysed mRNA expression of TRIM21 and Th17-associated genes in human atherosclerotic plaques, investigating 127 carotid plaques collected in the Biobank of Karolinska Endarterectomy (BiKE) cohort.23 We found that TRIM21 expression was inversely correlated with that of IRF4 and RORC, two transcription factors essential for the differentiation of Th17 cells.18,19 In addition, IL-23 and GM-CSF, cytokines crucial for the expansion and stabilization of Th17 cells, and the signature cytokine IL-17 A also inversely correlated with TRIM21 expression, whereas no association with IFNγ was detected (Figure 5A). Indeed, comparing Th17-related transcriptional profiles19 between patients with the highest and lowest TRIM21 plaque expression, we observed striking differences between the two groups in key positive (Figure 5B) and negative (Figure 5C) Th17 regulators. Analysis by hypergeometric means confirmed enrichment in differences of the Th17-related gene expression (P = 6.2 x 10−16). Our data thus reveal a significant negative association between TRIM21 and multiple genes associated with Th17 differentiation and function also in patient-derived atherosclerotic plaques, supporting a role for TRIM21 modulating atherosclerosis and cardiovascular disease in humans.
TRIM21 expression correlates to the Th17 pathway in human coronary artery disease. Gene expression in atherosclerotic plaques obtained from 127 patients undergoing carotid endarterectomy. (A) Correlation between expression of TRIM21 mRNA and interferon-regulatory factor 4 (IRF4), RORγt, GM-CSF, IL-23, IL-17 A, and IFNγ, respectively. (B, C) Expression profiles of key positive (B) and negative (C) Th17 regulators comparing patients with the highest and lowest TRIM21 expression levels (TRIM21low (n = 8) and TRIM21hi (n = 8)). Genes with P < 0.05 are depicted (57/71 genes included in the analysis). (Pearson correlation, permutation-based t-test with FDR control).
3.5 Low Trim21 expression associates with altered collagen turnover
Hypercholesterolemic Trim21-/-->Ldlr-/- chimeric mice developed larger plaques with higher collagen content, related to the increased expression of the pro-fibrogenic cytokine IL-17.6 To further investigate indicators of plaque stability, we assessed signs of both altered production and degradation of collagen. Increased production is associated with smooth-muscle cells of a collagen-producing, synthetic phenotype, which are characterized by high expression of osteopontin.24–26 We therefore analysed the expression of osteopontin using mRNA from the aortic root from Trim21-/-->Ldlr-/- and Trim21±/±->Ldlr-/- chimeric mice. Interestingly, we observed a significantly increased expression of osteopontin in Trim21-deficient mice (Figure 6A). Matrix metalloproteinases (MMPs) are central to collagen degradation, in mice primarily represented by Mmp-8 and -9.27 Quantification of Mmp-8 and -9 expression in the aortic root revealed significantly lower expression of Mmp-9 in mice transplanted with Trim21-/- bone marrow, while Mmp-8 expression was low in both groups (Figure 6B).
Higher collagen content in plaques of Trim21-deficient conditions is associated with altered collagen turnover. Gene expression assessed by RT-PCR in aorta from lethally irradiated Ldlr-/- mice transplanted with bone marrow (BM) from Trim21+/+ or Trim21-/- mice. Quantification of (A) Osteopontin and (B) Mmp-8 and Mmp-9 expression relative to Hprt (n = 8 and 7, respectively). (C–D) Gene expression in atherosclerotic plaques obtained from 127 patients undergoing carotid endarterectomy. (C) Correlation between TRIM21 and Collagen type 1. Patients with the highest (open circles) and lowest TRIM21 expression (black boxes) in carotid plaques (n = 10 in each group) were compared for the expression of (D) matrix metalloproteinases MMP-1, MMP-8, MMP-9, and MMP-13.
Lastly, we wanted to confirm our data on altered collagen turnover in the murine atherosclerosis model in human carotid plaques. In the plaques, we observed a strong and inverse correlation between collagen type 1 and TRIM21 expression (Figure 6C), as well as a significant difference in MMPs –1, –8, –9 expression, that have all been implicated in human atherosclerotic plaque development,28–30 between TRIM21hi and TRIM21low patients (Figure 6D).
In all, these data indicate that the increased collagen content in the plaques of Trim21-deficient mice relates to both increased production by SMCs and decreased degradation of the collagen. In all, our observations confirm and expand the basis for a more stable plaque phenotype associated with TRIM21 deficiency.
4. Discussion
The cardiovascular disease burden is rapidly growing, and is currently the number one cause of mortality worldwide.1 Several risk factors, such as smoking, obesity, hypertension, and hyperlipidemia have been identified. Chronic inflammation has repeatedly been linked to increased risk of atherosclerosis and cardiovascular disease.3,31 Both the innate and the adaptive immunity have been implicated in plaque formation, and monocytes and T cells are instrumental in the atheromatous process. In this study, we demonstrate that loss of the immune-regulatory factor Trim21 influences the atheromatous process. Our data reveal that Trim21-/- bone marrow chimeric Ldlr-/- mice develop larger atherosclerotic lesions with a higher fibrous content and thicker fibrous caps than Ldlr-/- mice transplanted with Trim21+/+ bone marrow. We also observed an increased presence of activated CD4+ T cells and aggravated IL-17 production accompanying the larger atherosclerotic lesions observed in Trim21-/--> Ldlr-/- mice. Importantly, such plaques are considered stable as opposed to plaques with low fibrous content, which are more prone to rupture,32,33 suggesting that TRIM21 may influence the development of myocardial infarction by modulating plaque content and stability.
Trim21 is mainly expressed in leucocytes and is an E3 ligase, the effector ligase in the ubiquitination cascade.7,34 The role of Trim21 has primarily been implicated in immune processes, in which it executes regulatory functions. It was recently found to control interferon regulating factors and proteins in the NF-kB cascade, thereby regulating the production of pro-inflammatory cytokines.7,35,36 In the present study, Trim21-deficient mice express higher levels of IL-17, lower levels of IFNγ induced proteins and display increased numbers of activated CD4+ cells, implicating IL-17 and Th17 cells in disease development. Also, non-inflammatory risk factors such as fatty acid levels and weight gain did not differ between the groups, further supporting the immune-mediated effector functions of Trim21.
In this study we, for the first time, demonstrate that Trim21 directly regulates the generation of Th17 cells negatively, but does not affect the development of other T helper cell lineages. Th17 cells have for long been implicated in driving pathogenicity in chronic inflammation, especially in autoimmunity. Recently, a more comprehensive picture of the effector function of Th17 cells has emerged identifying both pathogenic and non-pathogenic IL-17 producing T cells,21,37 and we observe that the Th17 cells generated by Trim21-/- cells display rather a non-pathogenic phenotype based on gene expression profiling.
The role of Th17 cells in atherosclerosis is currently unclear. IL-17 A mRNA transcripts have been detected in human plaques and in serum of patients with coronary artery disease. Increased levels of circulating IL-17 A and IL-17-producing CD4+ T cells have been reported in patients and in mouse models. Further, blockade or deletion of molecules crucial in the IL-17 pathway attenuates disease development in several disease models. However, T cell-derived IL-17 A has recently been found to promote a more stable plaque phenotype with increased collagen fibre content, suggesting that IL-17 may protect against cardiovascular events by preventing plaque rupture.6,38–43 Notably, we find that low TRIM21 expression correlates with high IL-17 A levels in human atherosclerotic plaques, giving further support to a protective role for IL-17 in atherosclerosis and progression to myocardial infarction through promoting a stable plaque phenotype.
Further analysis of the increased collagen content correlating with low TRIM21 expression revealed an association with increased levels of synthetic smooth-muscle cells (SMCs) and decreased expression of matrix metalloproteinases (MMPs). The differentiation of smooth-muscle cell into synthetic SMCs is central in the atherosclerotic process and pro-inflammatory molecules, including VCAM-1, have been shown to promote it.25,44 Interestingly, IL-17 has repeatedly been shown to induce VACM-1 expression,45,46 which could explain the increased amount of collagen-producing synthetic SMCs in the Trim21-/--≥ Ldlr-/- mice. The expression of MMPs has been associated to collagen content in plaques, contributing to the degradation of the stabilizing collagen.27,47 In Trim21-deficient mice and human plaques, we observed a decreased expression of MMPs indicating that a contributing factor for the increased collagen content may also be a decreased degradtion. In all, our data indicate that TRIM21 affects collagen content in the plaques, not only by promoting production via IL-17, but also by modulating collagen turnover. However, additional effects of Trim21 deficiency contributing to the plaque formation and increased collagen content at this stage cannot be rule out.
In conclusion, we here demonstrate a novel T cell intrinsic role for TRIM21 as a negative regulator of Th17 differentiation, and show that plaques of a more stable phenotype with higher collagen content develop under Trim21-deficiency. These findings provide a mechanistic basis for how TRIM21 may influence the development of atherosclerosis and coronary artery disease, and identify IL-17 as a possible plaque-stabilizing therapeutic target.
Time for primary review: 23 days
Acknowledgements
We thank Vijole Ottosson for excellent technical help, and gratefully acknowledge Dr Aurélie Ambrosi, Karolinska Institutet, for contributing to the writing of the paper.
Conflict of interest: none declared.
Funding
This work was supported by the Swedish Research Council; the China Scholarship Council; the Swedish Heart-Lung Foundation; Stockholm County Council; Karolinska Institutet; the Swedish Rheumatism Association; and the King Gustaf the V: th 80-year Foundation.
References
