T-Cell-Derived miRNA-214 Mediates Perivascular Fibrosis in Hypertension

Abstract

Rationale: Despite increasing understanding of the prognostic importance of vascular stiffening linked to perivascular fibrosis in hypertension, the molecular and cellular regulation of this process is poorly understood.

Objectives: To study the functional role of microRNA-214 (miR-214) in the induction of perivascular fibrosis and endothelial dysfunction driving vascular stiffening.

Methods and results: Out of 381 miRs screened in the perivascular tissues in response to Ang II (angiotensin II)-mediated hypertension, miR-214 showed the highest induction (8-fold, P=0.0001). MiR-214 induction was pronounced in perivascular and circulating T cells, but not in perivascular adipose tissue adipocytes. Global deletion of miR-214^-/- prevented Ang II-induced periaortic fibrosis, Col1a1, Col3a1, Col5a1, and Tgfb1 expression, hydroxyproline accumulation, and vascular stiffening, without difference in blood pressure. Mechanistic studies revealed that miR-214^-/- mice were protected against endothelial dysfunction, oxidative stress, and increased Nox2, all of which were induced by Ang II in WT mice. Ang II-induced recruitment of T cells into perivascular adipose tissue was abolished in miR-214^-/- mice. Adoptive transfer of miR-214^-/- T cells into RAG1^-/- mice resulted in reduced perivascular fibrosis compared with the effect of WT T cells. Ang II induced hypertension caused significant change in the expression of 1380 T cell genes in WT, but only 51 in miR-214^-/-. T cell activation, proliferation and chemotaxis pathways were differentially affected. MiR-214^-/- prevented Ang II-induction of profibrotic T cell cytokines (IL-17, TNF-α, IL-9, and IFN-γ) and chemokine receptors (CCR1, CCR2, CCR4, CCR5, CCR6, and CXCR3). This manifested in reduced in vitro and in vivo T cell chemotaxis resulting in attenuation of profibrotic perivascular inflammation. Translationally, we show that miR-214 is increased in plasma of patients with hypertension and is directly correlated to pulse wave velocity as a measure of vascular stiffness.

Conclusions: T-cell-derived miR-214 controls pathological perivascular fibrosis in hypertension mediated by T cell recruitment and local profibrotic cytokine release.

Keywords: blood pressure; collagen; fibrosis; hypertension; inflammation.

Figure 1.

MiR-214 is increased in the vasculature and immune cells in Ang II (angiotensin II)-induced hypertension. A, Principal component analysis of miRs in perivascular adipose tissue (PVAT; top) and volcano plot (bottom) in sham and Ang II-treated mice. B, Heatmap presenting z-scores calculated from individual miR −ΔCt values of all miRs significantly altered (t-test FDR adj P<0.05; n=5–6). C, Levels of miR-214/199 cluster (top) and pri-miR-214 (bottom) in PVAT (n=4–9). D, In situ hybridization of miR-214 in the vascular wall. Scale bar=300 and 150 μm. E, Levels of miR-214 in isolated PVAT, vessel wall layers (n=3 replicates of 3 pulled mice—9 mice/group), kidney, and selected immune organs (n=3–6). F, Levels of miR-214 and pri-miR-214 in PVAT of mice treated with placebo (PBO) or hydralazine/hydrocholrothiazide (Hyd/HCT; n=6/group). Data presented as mean±SEM; t-test with FDR correction (C, P values adjusted for 4 comparisons; miR clusters; E, P values adjusted for 8 comparisons; various organs) or by 2-way ANOVA (for miR-214 level; P^AngII×Drug=3.7×10⁻⁶, P^AngII=4.3×10⁻⁷, P^Drug=1.4×10⁻⁵; for pri-miR-214 level; P^AngII×Drug=0.018, P^AngII=1.2×10⁻³, P^Drug=0.12) with Tukey multiple comparisons test (F; P values adjusted for 6 comparisons).

Figure 2.

Pivotal role of miR-214 in the regulation of perivascular fibrosis in Ang II (angiotensin II)-dependent hypertension. A, Tail-cuff BP in sham and Ang II infused miR-214^−/− and WT littermates (n=9/group). B, Systolic BP measured by telemetry at baseline and during Ang II infusion (average, left; example readings, right; n=5). C, Perivascular collagen accumulation in picrosirius red (top) and Masson trichrome (bottom) staining (representative of n=9; scale bar=300 μm). D, Quantitative analysis of perivascular collagen deposition in Masson trichrome (n=9/group). E, Collagen 1, 3, and 5 mRNA in miR-214^−/− and WT littermates infused with buffer or Ang II (n=11/group). F, Aortic collagen quantification by hydroxyproline assay (n=7 in Sham and n=9 in Ang II). Data presented as mean±SEM; repeated measures 2-way ANOVA (A and B), Kruskal-Wallis test with FDR correction (D and F; P values adjusted for 6 comparisons) or 2-way ANOVA with Tukey test (E; P values adjusted for 6 comparisons). Overall P values for repeated measures 2-way ANOVA; A (P^Time=0.0051, P^Group=1.8×10⁻⁸, P^Time×Group=0.1378); B (P^Group=0.06, P^Time=2.6×10⁻⁷, P^Group×Time=0.83); for 2-way ANOVA; E, for Col1a1 (P^{AngII×Genotype}=0.003, P^AngII=1.3×10⁻⁵, P^Genotype=0.0035), Col3a1 (P^{AngII×Genotype}=0.037, P^AngII=0.0034, P^Genotype=0.033) Col5a1 (P^{AngII×Genotype}=0.028, P^AngII=0.0053, P^Genotype=0.042), for Kruskal-Wallis; D (P=0.0003) and F (P=0.0006).

Figure 3.

Role of miR-214 in the regulation of vascular stiffening in mouse and human hypertension. A, Thoracic aorta stiffness studied by pressure myography (n=5–8). B, Elastic modulus of adventitial stiffness measured by atomic force microscopy (right) with representative force-indentation curves (left; n=5–6/group). C, Circulating miR-214 serum levels in normotensive (Ctr, n=49) and patients with hypertension (HTN, n=51). D, Correlation between plasma miR-214 levels and pulse wave velocity (PWV) in humans with and without hypertension (n=100). E, TGF-β (transforming growth factor β) and fibronectin mRNA expression in control and hypertensive miR-214^−/− and WT mice (n=4–9/group). F, Protein level of TGF-β and fibronectin in aortas studied by WB (n=5–9). Data presented as mean±SEM repeated measures 2-way ANOVA with Bonferroni correction (A), Kruskal-Wallis test with FDR correction (B; P-adjusted for 6 comparisons) or 2-way ANOVA with Tukey test (E and F; P values adjusted for 6 comparisons) or t-test (C). Spearman rank test was used to assess correlations (D). Overall P values for 2-way ANOVA; A (P^AngII=1.8×10⁻⁵, P^Genotype=0.042), P^{AngII×Genotype}=0.097), E for Tgfb1 (P^{AngII×Genotype}=0.0059, P^AngII=0.0031, P^Genotype=0.225), FN1 (P^AngII=1.2×10⁻⁵, P^Genotype=0.266, P^{AngII×Genotype}=0.218); F for TGF-β (P^{AngII×Genotype}=0.0034, P^AngII=5.1×10⁻⁷, P^Genotype=0.0087), for FN1 (P^AngII=0.0015, P^Genotype=0.243, P^{AngII×Genotype}=0.589) and Kruskal Wallis B (P=0.0368).

Figure 4.

Essential role of miR-214 in endothelial dysfunction and vascular oxidative stress in hypertension. A, Relationship between plasma miR-214 levels and endothelial function (flow mediated dilatation [FMD]) or nonendothelium-dependent relaxations to nitroglycerin (NMD) in humans (n=100). B, Isometric tension studies of endothelium dependent (acetylcholine; ACh) and independent (sodium nitroprusside [SNP]) vasorelaxations (n=5–7/group) using wire myography in sham buffer and Ang II (angiotensin II) infused WT and miR-214^−/− mice. C, eNOS (endothelial nitric oxide synthase) protein level in mouse aortas (Western blotting; n=9–12/group). D, Aortic superoxide production measured by lucigenin (5 μM) enhanced chemiluminescence (n=7–9/group). E, Aortic Nox2 and Nox4 mRNA expression (n=5/group) and (F) protein levels in aortas of sham and Ang II infused miR-214^−/− and WT mice (n=9–12/group). Data presented as mean±SEM and analyzed by Spearman rank correlation test (A), repeated measures 2-way ANOVA with Bonferroni correction (B) or 2-way ANOVA with Tukey post hoc test (C–F; P values adjusted for 6 comparisons). Overall P values for repeated measures 2-way ANOVA; B for Ach (P^Response=0.0009, P^Group=0.044, P^{Response×Group}=0.016) and SNP (P^Response=04.1×10⁻⁷, P^Group=0.049, P^{Response×Group}=0.217). Two-way ANOVA; C (P^AngII=0.949, P^Genotype=0.862, P^{AngII×Genotype}=0.235); D (P^{AngII×Genotype}=0.061, P^AngII=0.0042, P^Genotype=0.051); E for Nox2 (P^{AngII×Genotype}=0.0062, P^AngII=0.193, P^Genotype=0.203) and Nox4 (P^{AngII×Genotype}=0.331, P^AngII=0.002, P^Genotype=0.005); F for Nox2 (P^AngII=0.0096, P^Genotype=0.683, P^{AngII×Genotype}=0.116) and Nox4 (P^AngII=0.0093, P^Genotype=0.924, P^{AngII×Genotype}=0.109).

Figure 5.

Crucial role of miR-214 in regulation of perivascular inflammation in hypertension. A, Total number of leukocytes and T cells (B) with representative density plots (n=6–9/group). C, Number of perivascular macrophages (F4/80+), dendritic cells (CD11c+), B cells (B220+), and NK cells (NK1.1+) per mg of tissue, studied by flow cytometry (n=5–7/group). Data presented on volcano plots depicting median (^__) and quartiles (^….). D, miR-214 induction in T cells (CD3+) and remaining leukocytes (CD3−) in Ang II hypertension measured in cells from peripheral blood (left, PBMC n=5–6/group) and from perivascular adipose tissue (PVAT; right, n=3/group of 3 pulled mice). E, MiR-214 level in PVAT of WT and RAG1^−/− animals (n=5–7/group). F, miR-214 in quiescent (direct lysis; control mAb) and anti-CD3 mAb activated T cells (n=3–4/group). Data presented as mean±SEM and analyzed by Kruskal-Wallis test (A–C), 2-way ANOVA with Tukey post hoc test (D; P values adjusted for 6 comparisons), t-test (E), and by Kruskal-Wallis test with FDR (F; P values adjusted for 3 comparisons). Overall P values for Kruskal-Wallis; A (P=0.017), B (P=0.011), C (P=0.06), D (P=0.05), and F (0.0001); 2-way ANOVA; D for PBMC (P^CD3×AngII=0.0006, P^AngII=0.0026, P^CD3=0.0006) and PVAT (P^AngII=0.024, P^CD3=0.115, P^CD3×AngII=0.115).

Figure 6.

T cell miR-214 mediates perivascular fibrosis and profibrotic T cell accumulation in PVAT (perivascular adipose tissue) in Ang II (angiotensin II)- induced hypertension. A, Vascular collagen accumulation in response to Ang II-hypertension assessed using picrosirius red (left) and Massons trichrome (right; scale bar=300 µm, representative of n=4–5/group) with quantification of adventitial collagen (B) in RAG1^−/− mice and upon adoptive transfer of WT and miR-214^−/− T cells (n=4–5/group). C, Ang II-dependent induction of Col 1, 3, and 5 mRNA in aortas of RAG1^−/− transferred with WT or miR-214^−/− T cells (n=4–5). D, Levels of circulating T cells in RAG1^−/− mice and upon adoptive transfer of WT or miR-214^−/− T cells (left) with representative dot plots (right; n=4–5/group). E, CD3 mRNA in PVAT of RAG1^−/− mice and upon adoptive transfer of WT or miR-214^−/− T cells (n=4–5/group). Data presented as mean±SEM and analyzed by 1-way ANOVA with Tukey post hoc t test (B, D, E; P values adjusted for 6 comparisons), t-test with FDR (C; P values adjusted for 3 comparisons). Overall P values for 1-way ANOVA; B (P=6.3×10⁻⁵), D (P=1.5×10⁻⁸), and E (P=7.8×10⁻⁶).

Figure 7.

Essential role of miR-214 in T cell responses in Ang II (angiotensin II)-induced hypertension. A, Genes with absolute log₂FC>1, significantly (FDR adj P<0.05 calculated using Wald test in DESeq2) changed in T cells following 2-week AngII infusion in vivo in WT (left) and mir214^−/− (right) mice (n=3). Venn diagram indicates total numbers of genes changed in each strain of mice in response to Ang II. B, Top 30 pathways significantly upregulated in WT T cells following 2-weeks AngII infusion in vivo in WT mice. C, Top 50 genes (based on interaction term estimate) preferentially induced by Ang II infusion in WT when compared with mir214^−/− mice (FDR adj P_interaction <0.05 for all presented). D, Top 50 genes preferentially repressed by AngII infusion in WT as compared with miR-214^−/− mice. Predicted targets of miR-214 for miRWalk (# red) and TargetScan 7.1 (# green). Data on heat maps are presented as z-scores of log transformed, normalized gene read counts (Log NRC) calculated with DESeq2.

Figure 8.

MiR-214 modulates T cell proinflammatory cytokine production, chemokine receptors expression and T cells migration in Ang II (angiotensin II)-induced hypertension. A, Heatmap of selected profibrotic cytokine mRNA in T cells upon Ang II-induced hypertension in WT (left) and miR-214^−/− (right) mice. Gene expression was studied upon ex vivo activation by anti-CD3 and anti-CD28 for 24 h (n=3–4/group) and presented as fold change, relative to average of sham. B, Cytokine production by T cells in sham buffer and Ang II infused WT and miR-214^−/− mice (n=4–6/group) upon 6-h stimulation with PMA studied by flow cytometry. C, T cell chemokine surface receptors in sham and Ang II-infused WT and miR-214^−/− mice studied by flow cytometry with representative histograms (n=4–6/group). D, In vitro chemotaxis of T cells obtained from hypertensive animals toward CXCL10 and CCL5 in Boyden chambers (n=6/group) with representative dot plots reflecting control and chemokine induced migration. E, Air pouch model of T cell chemotaxis (top) showing T cell recruitment toward membranes surrounding air-pouch filled with carrageenan and recombinant proteins (CXCL10 and CCL5) in hypertensive WT and miR-214^−/− mice with representative density plots (n=3–4/group). Data presented as mean±SEM and analyzed by t-test with FDR (All genes show FDR adj P<0.05 in WT but not in miR-214^−/−), Mann-Whitney U test (D) or t-test (E). Two-way ANOVA with Tukey (B and C; P values adjusted for 6 comparisons). Overall P values for 2-way ANOVA; B for IFN (P^{AngII×Genotype}=0.0073, P^AngII=0.0068, P^Genotype=0.0988), TNF (P^{AngII×Genotype}=0.0475, P^AngII=0.0001, P^Genotype=0.0042), IL17a (P^AngII=0.0007, P^Genotype=0.1316, P^{AngII×Genotype}=0.0814) and IL9 (P^AngII=0.0045, P^Genotype=0.0245, P^{AngII×Genotype}=0.0994); C for CCR1 (P^{AngII×Genotype}=0.0413, P^AngII=0.0198, P^Genotype=0.2508), CCR2 (P^{AngII×Genotype}=0.0467, P^AngII=0.0452, P^Genotype=0.1193), CCR4 (P^{AngII×Genotype}=0.0081, P^AngII=0.0022, P^Genotype=0.0013), CCR5 (P^{AngII× Genotype}=0.0142, P^AngII=0.0047, P^Genotype=0.0186), CCR6 (P^{AngII×Genotype}=0.0445, P^AngII=0.0066, P^Genotype=0.0596), CXCR3 (P^AngII=0.0167, P^Genotype=0.0771, P^{AngII×Genotype}=0.163).