If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Center for Molecular Medicine, MaineHealth Institute for Research, MaineHealth, Scarborough, ME, 04074, USAGraduate School of Biomedical Science and Engineering, University of Maine, USA
Center for Molecular Medicine, MaineHealth Institute for Research, MaineHealth, Scarborough, ME, 04074, USAGraduate School of Biomedical Science and Engineering, University of Maine, USA
Conditional deletion of Spry1 in VSMC increased Sca1 and PDGFRb, reduced leukocyte infiltration, and reduced VSMC migration.
•
Loss of Spry1 in VSMC increased advanced plaque burden, decreased collagen, and decreased markers of plaque stability.
•
Knockdown of Spry1 increased the expression of pro-inflammatory ICAM1, Cox2 and PAI1, and mesenchymal markers KLF4 and PDGFRb in human VSMC.
Abstract
Background and aims
Sprouty1 (Spry1) regulates the differentiation of vascular smooth muscle cells (VSMC), and our aim was to determine its role in atherogenesis. A significant proportion of cells within atherosclerotic lesions are derived from migration and pathological adaptation of medial VSMC.
Methods
We used global Spry1 null mouse, and Myh11-CreERT2, ROSA26-STOPfl/fl-tdTomato-Spry1fl/fl mice to allow for lineage tracing and conditional Spry1 deletion in VSMC. Atherosclerosis was induced by injection of a mutant form of mPCSK9D377Y-AAV followed by Western diet. Human aortic VSMC (hVSMC) with shRNA targeting of Spry1 were also analyzed.
Results
Global loss of Spry1 increased inflammatory markers ICAM1 and Cox2 in VSMC. Conditional deletion of Spry1 in VSMC had no effect on early lesion development, despite increased Sca1high cells. After 26 weeks of Western diet, mice with VSMC deletion of Spry1 had increased plaque burden, with reduced collagen content and smooth muscle alpha actin (SMA) in the fibrous cap. Lineage tracing via tdTomato marking Cre-recombined cells indicated that VSMC with loss of Spry1 had decreased migration into the lesion, noted by decreased proportions of tdTomato+ and tdTomato+/SMA + cells. Loss-of-function of Spry1 in hVSMC increased mesenchymal and activation markers, including KLF4, PDGFRb, ICAM1, and Cox2. Loss of Spry1 enhanced the effects of PDGFBB and TNFa on hVSMC.
Conclusions
Loss of Spry1 in VSMC aggravated plaque formation at later stages, and increased markers of instability. Our results indicate that Spry1 suppresses the mesenchymal and inflammatory phenotype of VSMC, and its expression in VSMC is protective against chronic atherosclerotic disease.
Atherosclerosis is a complex chronic inflammatory disease initiated by endothelial cell damage, leukocyte infiltration into the vessel wall, and activation and migration of vascular smooth muscle cells (VSMC) from the medial layer into the growing plaque [
]. Mortality is due to rupture of unstable atherosclerotic lesions, which can lead to thrombosis, acute myocardial infarctions, and related coronary syndromes [
VSMC have a unique ability to alter their phenotype in adaptation to cytokines or oxidative stress. As a result, they may lose their contractile characteristics and revert to mesenchymal stem cell-like, macrophage-like foam cells, or myofibroblasts [
]. VSMC contribute to 30%–70% of all plaque cells within an atherosclerotic lesion, and are a key determinant of the stability of advanced atherosclerotic plaques [
]. Despite many studies defining the importance and lineage of plaque cells during atherogenesis, there are still new discoveries related to upstream modifiers of VSMC that promote variant phenotypes during disease.
Sprouty1 (Spry1) is a member of the family of feedback inhibitors of receptor tyrosine kinases that are important in vascular development and diseases [
]. Spry1 is required to maintain the VSMC contractile phenotype in vitro, and plays an important role in regulation of injury-induced artery restenosis [
]. This study investigates the role of Spry1 in hypercholesterolemia-induced VSMC inflammatory phenotypic transition, and in the pathology of atherosclerosis.
2. Materials and methods
2.1 Mouse procedures and tissue processing
All procedures involving animals were approved by the Maine Medical Center Institutional Animal Care and Use Committee (IACUC), and conducted in compliance with ethical and safe research practices. Spry1LacZ (Spry1-/-) global null or wild type (Spry1+/+) mice on an FVB background were from the Mouse Mutant Regional Resource Center (UC, Davis) [
]. Mice were fed a standard chow diet (2018 Teklad global 18% rodent diet, Envigo). At the time of tissue harvest, mice were euthanized with isoflurane inhalant. Aortae or carotid arteries were carefully dissected, and processed for histological analysis, RT-qPCR or immunoblotting analysis.
Spry1fl/fl mice, in which the Spry1 open reading frame was flanked by loxP sequences, were on a C57BL6/J background [
]. Genotyping primers are listed in the Major Resources Table. To obtain conditional Spry1 null mice, a Spry1fl/fl female was bred to a male Myh11-CreERT2 [
], and littermate breeding of Spry1fl/+with Spry1fl/+;Myh11-CreERT2 was performed to obtain Spry1fl/fl;Myh11-CreERT2 or Spry1+/+;Myh11-CreERT2. For ease of reading, we will refer to these groups as Spry1WT (Spry1+/+;Myh-CreERT2) and Spry1VSMCKO (Spry1fl/fl;Myh-CreERT2).
To obtain conditional Spry1 null with simultaneous VSMC lineage-tracing, Spry1fl/fl mice were bred with ROSA26-STOPfl/fl-tdTomato to obtain Spry1fl/+;ROSA26+/−, and then littermate Spry1fl/+;ROSA26 ± mice were bred to obtain Spry1fl/fl;ROSA26-/- or Spry1+/+;ROSA26-/-. Female Spry1fl/fl were bred with male Myh11-CreERT2 mice to obtain Spry1fl/+;Myh-CreERT2, then littermates were bred to obtain Spry1+/+;Myh-CreERT2 or Spry1fl/fl;Myh-CreERT2. Female Spry1fl/fl;ROSA26-/- or Spry1+/+;ROSA26-/- were bred with male Spry1fl/fl;Myh11-CreERT2 or Spry1+/+;Myh11-CreERT2. Because the Myh11-CreERT2 transgene is located on the Y chromosome, all conditional null studies were performed on male mice. For ease of reading, we will refer to these groups as tdTomato-Spry1WT (Spry1+/+;ROSA26-/-;Myh-CreERT2) and tdTomato-Spry1VSMCKO (Spry1fl/fl;ROSA26-/-;Myh-CreERT2).
Cre-lox mediated recombination of ROSA26-tdTomato and Spry1 deletion was induced in ∼7-week-old mice after intraperitoneal injections of 1 mg of tamoxifen (T-5648, Sigma) per mouse/day for 5 days. Two days after the last tamoxifen injection, mice were injected with 1011 gc/mouse of adeno-associated viral vector (AAV) expressing a mutant form of murine proprotein convertase subtilisin/kexin type 9 [
] (mPCSK9D377Y-AAV, Harvard) via retroorbital injection. Mice were then fed a Western diet (D2079bi, Research Diet) containing 0.15% cholesterol for 6, 12, 18 or 26 weeks. Body weight was measured weekly.
Blood samples were collected under isofluorane anesthesia after 6 hours of fasting using hematocrit capillary tubes (22-362566, Thermofisher Scientific) for serum/plasma cholesterol measurement using a colorimetric cholesterol assay (STA-384, Cell Biolabs, Inc). On the day of harvest, mice were fasted for 6 hours and euthanized by overdose of isoflurane, and then perfused with 10 mL of PBS. Hearts, aortae, and carotid arteries were carefully dissected, fixed, and embedded in OCT or processed for further analyses.
2.2 Cell culture
Human aortic vascular smooth muscle cells (hVSMC) from at least three individual donors were obtained from Lonza or Invitrogen and cultured in SmGM2 complete medium (Lonza). Cells were used at passage 6–8 for all experiments. For knockdown studies, hVSMC were plated into 6 or 12-well plates (Falcon) at ∼40% confluence and transduced with shRNA lentiviruses targeting human Spry1 or non-targeting control (NT, Open Biosystems) for ∼8 hours, and then replaced with fresh SmGM2 medium. To test cellular responses to PDGFBB or TNFa, cells were deprived of serum and growth factors for 24 hours in smooth muscle cell basal medium (SmBM, Lonza), then stimulated with 20 ng/ml PDGFBB or TNFa (Peprotech) for indicated time periods.
2.3 RNA sequencing
Spry1LacZ (Spry1-/-) global null or wild type (Spry1+/+) mice were retroorbitally injected with 1 × 1011 gc/mouse mPCSK9D377Y-AAV, then fed a Western diet for 4 weeks. Dorsal aortae were separated from the adventitia and used for total RNA extraction. RNA quantity and quality were assessed using the Agilent Bioanalyzer RNA 6000 Pico Chip. The RNASeq library was prepared using Stranded Illumina TruSeq, and sequenced through HiSeq Sequencing-by-Synthesis at the Vermont Integrative Genomics Resource, University of Vermont. Sequencing reads were first trimmed for quality then aligned to the most current mouse genome (mm10) using the STAR aligner. Reads were then quantified using an Expectation-Maximization model. Differential gene expression was statistically assessed and genes annotated using DESeq2 under a Negative Binomial model.
2.4 RT-PCR
Total RNA from hVSMC or mouse arteries was extracted using RNeasy plus mini kit (Qiagen) according to the manufacturer's protocol. RNA quality and quantity were evaluated using agarose gel electrophoresis and Nanodrop spectrophotometry (Thermo Fisher), respectively. cDNA was generated using the ProtoScript First-Strand cDNA Synthesis kit (Bio-Rad), and qPCR analysis was performed using AzuraQuant Green Fast qPCR master mix (Azura). Relative mRNA levels were calculated based on the comparative Cτ (the number of amplification cycles reaching the threshold fluorescence) method using Cypa or Gapdh as internal controls. Primer sequences are listed in the Major Resources Table.
2.5 Immunoblotting
For protein assays, cells or arteries were lysed in 2X SDS sample buffer or in HNTG buffer [50 mM HEPES, pH 7.4, 150 mM NaCl, 10% (vol/vol) glycerol, 1.5 mM MgCl2, 1 mM EGTA] containing protease inhibitor cocktail (Roche), phosphatase inhibitors and 0.2% Triton X-100. Lysates were subjected to SDS-PAGE and immunoblot analysis using antibodies against ICAM-1, SPRY1, Cox2, Lrp1 (Cell Signaling Technology), SMA, and tubulin (Sigma). Antibodies are described in detail in the Major Resources Table.
2.6Fluorescence-activated cell sorting (FACS)
Dorsal aortae (perivascular adipose tissue removed) from tdTomato-Spry1wt and tdTomato-Spry1VSMCKO mice fed a Western diet for 6 weeks were collected, digested, and processed for FACS analysis. Antibodies used were: APC/Cy7-Rat-anti-mCD68, BV421-Rat-anti-mLy6A/E(Sca1), PE-Rat-anti-mCD140b, PE-Rat-anti-mCD140 from Biolegend; FITC-Rat-anti-mCD45, FITC-Rat-anti-mCD31 from BD Bioscience.
2.7 Histology and mutiplex RNAscope analysis
For morphological analysis, sections of aortic valves or brachiocephalic arteries were stained with hematoxylin and eosin (H&E) or Masson's trichrome stain. Immunofluorescence staining was performed by incubation with rabbit anti-CD45 (Abcam,1:100), anti-ICAM-1 (Biolegend, 1:100), anti-PDGFRb (Santa Cruz, 1:50) followed by FITC-anti-rabbit or rat IgG. FITC-anti-SMA antibody (Sigma) staining was used for tdTomato tracing immunofluorescent assays. Details of antibodies can be found in the Major Resources Table. In situ RNAscope assays for Pdgfrb, Sca1, KLF4, and Spry1 were performed according to RNAscope fluorescent multiplex assay instructions (ACDBio). Images were captured using Canon EOS camera and remote imaging software (Canon), or a Leica SP8 confocal microscope. Quantification was performed using ImageJ.
2.8 Statistical analysis
GraphPad Prism 7.0 was used to perform the statistical analyses. Data were analyzed using either 1-way or 2-way ANOVA. When justified, multiple-comparisons tests were used to determine p value. Unpaired t-tests were used for comparison of two groups of data. Differences were considered significant at p < 0.05, and all results are reported as means ± standard deviation.
3. Results
3.1 Global loss of Spry1 alters aortic gene expression and enhanced atherogenesis in vivo
To further our understanding of Spry1 function in VSMC physiology, we performed bulk RNA sequencing on dorsal aortae from age matched wild type (Spry1+/+) and Spry1 global null (Spry1-/-) mice (n = 4 per genotype, FVB background) that were injected with mPCSK9D377Y-AAV and fed with Western diet for four weeks to induce hypercholesterolemia [
]. A total of 22,001 genes were sequenced. Principal Component Analysis did not show a strong change in the biological signal of aortae from Spry1-/- mice, likely due to a high variation within groups. However, using traditional thresholds, 276 genes passed the false discovery rate (FDR≤0.2) and were significantly differentially expressed (p ≤ 0.05). The forty top differentially expressed genes are displayed in a heatmap in Supplementary Fig. 1A. Gene Ontology (GO) enrichment analysis showed the most enriched genes were associated with: 1) protein refolding (HSP90AA1, HSPA8 and HSPH1), 2) fibroblast growth factor receptor signaling pathway (CTGF, FAT4 and FGF2), 3) glycosaminoglycan binding (SFRP1, CTGF, FGF2, ITIH4), 4) branching morphogenesis, and 5) protein kinase B activation. Heat shock proteins (HSP90AA1, HSPA8 and HSPH1) are multifunctional, and their expression is positively correlated with cardiovascular diseases [
]. Conversely, HSP90 has also been reported to bind oxidized phospholipids and prevent their further oxidation to more pathogenic and reactive products [
], aortae from Spry1-/- mice had a significant reduction of calponin1 (CNN1) mRNA compared to controls by RNAseq analysis. Further, SMA immunofluorescence combined with Spry1 in situ RNAscope analysis showed decreased intensity of SMA protein in the Spry1-/- blood vessel (Supplementary Fig. 1B). Similarly decreased SMA mRNA was also observed in carotid arteries from Spry1-/- mice compared to wild type controls, correlating to the absence of Spry1 protein in the arteries (Supplementary Figs. 1C–D). Because of the link of inflammation to vascular disease, we analyzed several inflammatory markers including ICAM-1, which is involved in leukocyte trafficking, and Cox2, which is induced during inflammatory pathologies. We also analyzed Lrp-1, which is protective against atherosclerosis [
]. Loss of Spry1 resulted in increased levels of ICAM-1 mRNA compared to controls (Supplementary Fig. 1E). In addition, aortae from Spry1-/- mice had increased Cox2, but decreased Lrp-1 protein compared to controls (Supplementary Figs. 1F–G).
To determine whether these gene expression changes in Spry1-/- mice correlated to changes in vascular physiology, we performed an initial study of experimental atherosclerosis using age matched Spry+/+ and Spry1-/- male mice (n = 4/genotype). After a single injection of mPCSK9D377Y-AAV (1 × 1011 gc/mouse), mice were fed a Western diet for 12 weeks. We found increased plaque formation in dorsal aortae and aortic valves in Spry1-/- mice compared to controls (Supplementary Figs. 2A–D). Immunohistochemistry to detect macrophages using F4/80 showed increased macrophage infiltration in aortic lesions from Spry1-/- mice (Supplementary Fig. 2E). Interestingly, one out of the four Spry1-/- mice had severe coronary artery plaques (Supplementary Fig. 2F), which is uncommon in atherosclerosis models in mice. The ratio of heart weight to body weight was not different based on genotype (Supplementary Fig. 2G).
3.2 Early effects of loss of Spry1 in VSMC in mice
Mice with global loss of Spry1 on an FVB background are viable unlike the lethal phenotype on a C57BL/6 background. However, we observed multiple partially penetrant abnormalities in Spry1-/- mice including kidney cysts (∼25–50%), enlarged uterus (>80%), and enlarged abdominal adipose tissue [
], which may contribute to atherosclerotic lesion formation. To address the role of Spry1 in VSMC in atherosclerosis more precisely, we generated tamoxifen-inducible VSMC-specific conditional Spry1 null mice (Myh11-CreERT2;Spry1fl/fl) or control mice (Myh11-CreERT2;Spry1+/+), hereafter referred to as Spry1VSMCKO or Spry1WT mice, respectively (Fig. 1A). Cre recombinase was induced with five consecutive tamoxifen intraperitoneal injections (1mg/mouse/day). Hypercholesterolemia was induced by a single dose of mPCSK9D377Y-AAV (10 [
]gc/mouse) administered via retroorbital injection. Mice were fed a high fat, Western diet. Atherosclerosis is a heterogeneous process in humans; similarly, we observed heterogeneous plaque formation in the aortae in mice. At 12 weeks after Western diet feeding, there were no significant differences in aortic plaque area (Fig. 1B–C), and total serum cholesterol levels (Fig. 1D) and body weight were similar between groups. We also analyzed lesions along the aortic root (Fig. 1E), and assessed lipid accumulation and plaque size using oil Red O staining (Fig. 1F). The lesion area in aortic valves in Spry1VSMCKO mice showed a trend towards smaller plaques compared to those in Spry1WT mice (Fig. 1F–G).
Fig. 1Loss of Spry1 does not affect overall atherosclerotic plaque, but increases Sca1+ cells in lesions.
(A) Overview of mouse model and experimental design. Spry1WTand Spry1VSMCKO mice received tamoxifen injections at 7 weeks of age, followed by one mPCSK9D377Y-AAV retroorbital injection (1011 gc/mouse), and western diet (WD) feeding for 12 wks. (B) En face Sudan IV staining of lipid species in aortae was used as a measure of plaque area, which is quantified in (C). (D) Mouse serum was collected after fasting for 6 hours at the end of the experiment for quantification of total cholesterol. (E) Schematic of the aortic root region for sectioning and histologic analysis. (F) Representative Oil Red O stained images from four sequential segments (∼170μm per segment) of aortic roots. Staining is quantified in (G). (H) Representative images of multicolor RNAScope analysis of Spry1, Sca1 (white arrows), and PDGFRb mRNA in aortic roots regions without plaque, with quantification of these areas outside of the lesion in (I–J). (K) Representative images of multicolor RNAScope analysis of Spry1, Sca1 (white arrows), and PDGFRb mRNA in aortic root plaques, with quantification in (L–M). Graphs show means ± SD and p values were determined using either a 2-way ANOVA with alpha = 0.05 followed by multiple comparisons test (L–M), or t-test (C, D, G, I-J). **p < 0.01.
]. Thus, we performed in situ multiplex RNAScope analysis to examine expression of Sca1, PDGFRb and Spry1 mRNA. Spry1 mRNA in wild type mice was detected in aortic VSMC, adventitial cells, and some plaque cells, and was specifically deleted in VSMC in aortic arteries in Spry1VSMCKO mice (Fig. 1H, K). Sca1 positive cells were observed in the medial VSMC in lesions from both genotypes and in non-lesion areas in Spry1VSMCKO mice (Fig. 1I, L). It is worth stressing that Sca1 positive cells were distributed in a scattered pattern, particularly in medial layers of the aortae. Nonetheless, Spry1VSMCKO mice had significantly elevated Sca1 positive cells in plaques compared to controls (Fig. 1K-L) whereas there were no differences in the aortic media. The mRNA of KLF4, another key regulator of VSMC mesenchymal transition, was also increased in aortic sections from Spry1VSMCKO mice (Supplementary Fig. 3A). FACS analysis on aortic cells also suggested a trend of increased Sca1+, CD61 and CD140a + cell populations in Spry1VSMCKO mice compared to Spry1WT mice, although this did not reach statistical significance (Supplementary Fig. 3B). We also assessed PDGFRb mRNA expression in aortic valve medial layer cells compared to controls (Fig. 1H, J, K, and M), and levels between groups were similar.
3.3 Spry1 deficiency reduces CD45+ cell infiltration and VSMC migration into the plaque
The migration of VSMC from the media to the intima plays an important role in plaque formation. To better quantify VSMC populations and trace the movement of medial VSMC, we generated tamoxifen-inducible VSMC-specific conditional Spry1 null mice with simultaneous VSMC lineage-tracing in Myh11-CreERT2ROSA26-STOPfl/fl-tdTomatoSpry1fl/fl or Spry1+/+ mice, hereafter referred as tdTomato-Spry1VSMCKO or tdTomato-Spry1WT mice, respectively. These double transgenic mice were treated with tamoxifen to induce Cre, followed by a single injection of mPCSK9D377Y-AAV and 6 weeks of Western diet (Fig. 2A). Upon tissue harvest, the adipose tissue was removed from the aorta, and vessel wall cells were dissociated by collagenase digestion and subjected to flow cytometry to analyze the proportions of tdTomato, Sca1 and PDGFRb-producing cells. We observed that loss of Spry1 in VSMC decreased overall tdTomato positive cells, while the proportion of Sca1+ and PDGFRb/Sca1 double positive cells did not change (Fig. 2B).
Fig. 2Loss of Spry1 alters migration pattern of VSMC and reduces CD45+ cells in plaques.
(A) Schematic of the mouse model and experimental design. Myh11-CreERT2;tdTomato-ROSA26-STOPfl/fl-Spry1fl/fl or Spry1WT mice received 5 intraperitoneal (IP) tamoxifen injections in corn oil at 7 weeks of age to label smooth muscle cells with tdTomato (tdTomato-Spry1VSMCKO or tdTomato-Spry1WT, respectively). Mice received a single mPCSK9-AAV retroorbital injection (1011gc/mouse), followed by Western diet (WD) feeding for 6 or 9 weeks before analysis. (B) After 6 weeks of Western diet, dorsal aortae were processed (n = 6) for FACS analysis, and the proportion of tdTomato+, Sca1+, and PDGFRb + cells quantified. (C) Aortic root sections were stained with Bodipy™493/503 to visualize lipids. Representative confocal images show lipids (green) and tdTomato + cells (red). Scale bar = 50μm. (D) Quantification of tdTomato + cells within lesions (n = 6/genotype). (E) The distance of tdTomato + cells to the inner elastic lamina (white dashed line) was measured using Image J. The frequency distribution of 154 measurements/genotype was analyzed in Prism. (F) Sections from the aortae after 9 weeks of Western diet were stained by Bodipy™493/503 to visualize lipids (green), scale bar = 60μm. (G) Quantification of tdTomato + cells/plaque area (n = 4/genotype). (H) Distribution of tdTomato + cell distances from 210 measurements/genotype. (I) CD45 staining was used to quantify leukocytes in sections from mice fed WD for 9 weeks. Scale bar = 50μm. (J) Plaque area was calculated, and used to normalize number of CD45+ cells/genotype (K). Graphed are means ± SD, p values were determined using t-test. *p < 0.05; **p < 0.01.
Next, we used Bodipy™493/503, a tracer for neutral lipids, to visualize the localization of lipids in VSMC and atherosclerotic lesions. Hypercholesterolemia induces heterogeneous lesion formation with various degrees of tdTomato + VSMC involvement (Fig. 2C). While the overall tdTomato positive cells in the plaques in aortic roots was not different between genotypes (Fig. 2D), it is worth noting that we observed a dim tdTomato signal in plaques from tdTomato-Spry1VSMCKO mice (Fig. 2C and 2F), indicating loss of Spry1 in VSMC may alter tdTomato protein metabolism. The dim tdTomato positive cells likely were excluded in FACS analysis, resulting in overall decreased tdTomato + cells in tdTomato-Spry1VSMCKO mice (Fig. 2B), a discrepancy compared to data from cell counting results (Fig. 2D and 2G). We also quantified the location of these cells in the plaque with relation to the internal elastic lamina (dotted line) to assess distance of VSMC migration. There were more Spry1VSMCKO tdTomato positive cells closer to the internal elastic lamina (Fig. 2E), within 10μm from the internal elastic lamina. In addition, tdTomato cells from Spry1WT mice were throughout the plaque, including >70μm away from the internal elastic lamina, which was not seen in plaques from Spry1VSMCKO mice. There was ∼1.8 fold decrease in the total distance of tdTomato positive cells from the internal elastic lamina in the aortic root lesions of Spry1VSMCKO mice compared to Spry1WTmice.
We used the same analysis in the aortae (Fig. 2F–H). Similarly, the overall number of tdTomato positive cells within plaques did not vary by genotype (Fig. 2G). However, like in the aortic roots, we found increased tdTomato positive cells closer to the internal elastic lamina (within 5μm) in the lesions of Spry1VSMCKO mice compared to Spry1WT mice (Fig. 2H). Additionally tdTomato positive cells from Spry1WT mice were found >50μm from the internal elastic lamina, whereas this was not observed in the plaques of Spry1VSMCKO mice. In the plaques of the aortae, there was ∼1.9 fold decrease in total distance of tdTomato positive cells from the internal elastic lamina in the lesions of Spry1VSMCKO mice compared to Spry1WT mice. These results are consistent with our previous report that loss of Spry1 decreased hVSMC migration in vitro [
Endothelial damage and leukocyte infiltration is an important component of lesion formation with hypercholesterolemia. We performed immunofluorescence staining to detect leukocytes with the pan-leukocyte marker CD45 (Fig. 2I). In the aortic root plaques after nine weeks of high fat diet, overall lesion size did not differ based on genotype (Fig. 2J). However, the number of CD45 positive cells/plaque area was significantly decreased in lesions from Spry1VSMCKO mice compared to Spry1WT mice (Fig. 2K).
3.4 Spry1 deficiency in VSMC leads to increased advanced plaque burden and decreased indices of plaque stability under prolonged hypercholesterolemia
Given the chronic nature of atherosclerosis, which develops over decades in humans, we were interested whether Spry1 deficiency has effect on lesion formation and stability with long-term hypercholesterolemia. Using the same mouse models, we performed a study to extend high fat diet feeding for 26 weeks, until the mice were 34 weeks of age (Fig. 3A). Analysis was performed in the aortic roots, the brachiocephalic arteries, and the aortae (Fig. 3B). There were no differences in plasma cholesterol between the genotypes (Fig. 3C). In contrast to shorter term high fat diet feeding, after 26 weeks, Spry1VSMCKO mice had increased plaque formation visualized by whole mount Sudan IV staining on en face dorsal aortae (Fig. 3D–E). Serum cholesterol is a known risk factor, and levels in the mouse after high fat feeding were variable between individuals. Thus, we performed correlation analysis on cholesterol level vs. aortic lesion area. The plaque area positively correlated with cholesterol levels in Spry1WT mice, but not in Spry1VSMCKO mice (Fig. 3F), indicating the increased plaque burden with loss of Spry1 protein in VSMC is independent of cholesterol level.
Fig. 3Spry1 deficiency in VSMC leads to increased advanced plaque burden and necrosis.
(A and B) Schematic of the mouse model, experimental design and analysis. (C) Six-hour fasting plasma was collected after Western diet (WD) feeding for 10 weeks (n = 12/group). (D) Representative images of aorta en face Sudan IV staining, and quantification of lipid area (E). Spry1VSMCKO mice had more lipid deposition in aortae compared to controls. (F) Correlation of lipid area with plasma cholesterol levels. Sections from three sequential segments (one section per segment) of brachiocephalic arteries (G, ∼130μm per segment) or aortic roots (J, ∼220μm per segment) were stained with H&E. Average areas of plaques and necrotic cores were quantified from three sections from each group from brachiocephalic arteries (H–I) or aortic roots (K–L). Graphed are means ± SD, p values were determined using either 2-way ANOVA with alpha = 0.05 followed by multiple comparisons test (H, K), or t-test (C, E, F, I, and L). *p < 0.05; **p < 0.01; ***p < 0.001; n = 12 for each genotype).
Next, we examined the morphology of plaques by H&E staining. Brachiocephalic arteries and aortic roots from mice of both genotypes formed complex lesions with an imperfectly formed fibrous cap overlying a central necrotic, lipid-laden plaque (Fig. 3G and 3J). Loss of Spry1 in VSMC had no effect on atherosclerotic lesion size in the brachiocephalic arteries (Fig. 3G) or overall necrotic area (Fig. 3H). However, because of the high variability in size of lesions and necrotic area/plaque, we quantified the proportion of necrosis within each plaque. Plaques from Spry1VSMCKO mice had a higher proportion of necrotic area per plaque compared to those from Spry1WT mice (Fig. 3I). In the aortic roots (Fig. 3J), loss of Spry1 in VSMC led to increased plaque area (Fig. 3K), but no changes in overall necrotic area or proportion of necrosis within each plaque (Fig. 3L).
The level of collagen deposition in lesion is a useful indicator of plaque stability. Using Masson's Trichrome stain, we observed non-uniform fibrosis caps, with some overlaid by a layer of xanthoma, particularly in Spry1VSMCKO plaques (Fig. 4A–B). Loss of Spry1 in VSMC decreased overall collagen content in lesions within the brachiocephalic arteries (Fig. 4A) and aortic roots (Fig. 4B). VSMC migration to the fibrous cap and re-expression of genes associated with VSMC differentiation, such as SMA, contribute to plaque stability. We therefore examined the tdTomato + VSMC localization in cap areas as well as SMA protein localization. Using immunofluorescence and confocal microscopy, we observed thicker SMA-stained fibrous caps in both the brachiocephalic artery (Fig. 4C) and aortic root lesions (Fig. 4D) in Spry1WT compared to Spry1VSMCKO mice. The SMA positive fibrous caps were contiguous along the lumenal border of the plaque in Spry1WT mice, whereas the staining in plaques from Spry1VSMCKO mice was discontinuous with less SMA staining (white arrows).
Fig. 4Loss of Spry1 in VSMC leads to decreased indices of advanced plaque stability.
Masson's Trichrome stain of brachiocephalic arteries (A) or aortic roots (B) with quantification of trichrome stain from three sequential segments. Plaque collagen content was quantified by measuring blue collagen intensity, and then normalized to pink cytoplasm intensity. Immunofluorescence staining using FITC-conjugated mouse monoclonal anti-SMA (green) to examine SMA expression in brachiocephalic arteries (C) or aortic roots, scale bar = 200μm. (D) to visualize the fibrous cap. Scale bar = 230 μm. White arrows show SMA-positive fibrous cap regions. (E) Representative images of left coronary cusp lesion cap areas in aortic roots. Double broken white lines indicate a ∼40μm subendothelial fibrotic cap region. L = lumen, I = intima. Scale bar = 63μm. (F) TdTomato+ and SMA + cells in this area were quantitated from three segments. (G) The proportion of tdTomato negative cells was compared between genotypes. (H) Representative images from aortic roots stained to detect CD45 (green). Lower magnification images were represented in a and b, scale bar = 120um. Areas within the double dash lines were imaged at higher magnification, and represented in c and d, respectively. Scale bar = 50μm. The subendothelial region is indicated by dashed lines. L = lumen, I = intima, M = media, and N = necrotic area. Scale bar = 50μm. (I) The proportion of CD45 positive cells was quantified. Average CD45+tdTomato-cells were calculated from three segments. Graphed are means ± SD, p values were determined using either 2-way ANOVA with alpha = 0.05 followed by multiple comparisons test (F, G), or t-test (B). *p < 0.05; **p < 0.01; ***p < 0.001; n = 12 for each genotype.
We then assessed the tdTomato + cell localization within the subendothelial fibrotic cap of coronary cusps of aortic valves (Fig. 4E–G). In lesions from Spry1VSMCKO mice, we observed decreased proportions of tdTomato + cells and tdTomato+, SMA + double positive cells (Fig. 4E–F). Conversely, loss of Spry1 in VSMC resulted in increased proportions of non-tdTomato positive cells (negative) in subendothelial cap regions (Fig. 4G). In addition, CD45+ leukocyte infiltration into the subendothelial compartment was significantly increased in plaques from Spry1VSMCKO mice compared to those from Spry1WT mice (Fig. 4H–I, and Figs. S4A–B). Our results indicate that endogenous Spry1 functions in the pathogenesis of atherosclerosis include promotion of VSMC migration throughout the lesion, particularly to the fibrous cap, where it promotes SMA expression and stabilization of the plaque. This is associated with enhanced collagen deposition within the lesion. Loss of Spry1 in VSMC impairs the differentiated phenotype of fibrous cap cells and leads to increased leukocyte accumulation in the subendothelial region of the lesion.
3.5 Deleting Spry1 in human VSMC in vitro increases mesenchymal and inflammatory marker expression
To determine whether the loss of function phenotype of Spry1 in VSMC in mice correlated to its activity in human VSMC, we studied human VSMC in vitro. Spry1 is expressed in human VSMC, and we utilized shRNA lentivirus targeting Spry1 for loss-of-function studies. Consistent with our previous report [
], suppression of Spry1 promoted a synthetic phenotype in human VSMC with lower levels of the contractile marker calponin and increased levels of the mesenchymal markers KLF4 and PDGFRb (Fig. 5A). Inflammatory activation of medial VSMC correlates to formation of atherosclerotic plaques. Suppression of Spry1 increased ICAM-1, a molecule involved in leukocyte trafficking and inflammation pathology, and CCL2, a chemokine, while not affecting SDF-1 transcript level. Immunoblotting confirmed higher ICAM-1 protein in VSMC with Spry1 knockdown (Fig. 5B–C).
Fig. 5Deletion of Spry1 in hVSMC regulates mesenchymal and inflammatory markers.
(A) Spry1 shRNA lentivirus was used to reduce Spry1 protein in human VSMC (S1kd). Cells were used for RT-qPCR analysis in triplicate to quantify transcripts for calponin, KLF4 and PDGFRb, CCL2, and ICAM-1. (B) Representative immunoblotting of ICAM-1 and C) quantification from three independent experiments. (D) Cells were made quiescent in basal SBM2 medium, and analyzed after PDGFBB (20 ng/ml) stimulation. (E) RT-qPCR analysis for PAI-1 mRNA after PDGFBB stimulation for 8 hours. (F) Human VSMC were serum starved for 24 hours, then stimulated with PDGFBB for 24 hours, and lysates collected for immunoblotting to quantify ICAM-1, Cox2, Spry1 and Tubulin. (G) ICAM-1 and Cox2 protein levels were quantified using ImageJ. (H) Cells were treated with 10 ng/ml TNFa for 24 hours, and subjected to RT-qPCR analysis to evaluate Spry1, ICAM-1, CCL2, IL-6, and PAI-1 mRNA expression. (I) Cell lysates were immunoblotted to quantify Spry1 (J) and Cox2 (K) protein. mRNA or protein values are graphed as means ± SD. p values were determined using either a 2-way ANOVA with alpha = 0.05 followed by multiple comparisons test (A, E, G, H, J), or t-test (C). *p < 0.05; **p < 0.01; ***p < 0.001.
]. We tested whether increased PDGFRb expression in Spry1 knockdown human VSMC affects downstream signaling and altered cytokine expression. PDGFBB stimulation increased CCL2 and CCL3 mRNA, suppression of Spry1 did not alter induction of CCL2, but reduced induction of CCL3 transcript (Fig. 5D). Spry1 mRNA was transiently decreased by PDGFBB stimulation at 2 h, and restored thereafter (Fig. 5D). PAI-1 is another critical factor that contributes to the pathogenesis of cardiovascular diseases [
]. We found that basal mRNA level of PAI-1 in human VSMC was elevated after suppression of Spry1, with the same effect after PDGFBB stimulation (Fig. 5E). In addition, the increased basal ICAM-1 protein in Spry1 knockdown human VSMC was further increased by PDGFBB stimulation (Fig. 5F–G). We also observed that suppression of Spry1 increased Cox2 protein. PDGFBB stimulation decreased Cox2 protein in control cells, and to a less degree in Spry1 knockdown human VSMC (Fig. 5F–G). Interestingly, PDGFBB stimulation markedly decreased Spry1 protein levels, and this effect was sustained up to 24 h, unlike the transient decrease in Spry1 mRNA (Fig. 5D).
Next, we examined human VSMC response to TNFa. Consistent with its pro-inflammatory activity, TNFa markedly increased the mRNA expression of cytokines and factors including CCL2, IL-6, and PAI-1 in human VSMC. Conversely, TNFa stimulation significantly decreased Spry1 mRNA (Fig. 5H) and protein (Fig. 5I–J). The elevated Cox2 protein in Spry1 knockdown human VSMC was further increased by TNFa stimulation (Fig. 5I, K).
4. Discussion
The regulation of the pro-inflammatory phenotype of VSMC in atherosclerotic lesion progression and stability is complex and multi-factorial. In this study, we investigated the effects of Spry1 production in VSMC, and consequences of its loss-of-function in atherogenesis. We used a mouse model targeting Spry1 in VSMC using inducible Myh11CreERT2, and also used a Cre-activated tdTomato lineage tracing method to track the fate of VSMC in the atherosclerotic plaque. Our studies represent the first description of Spry1 function in the vasculature during vascular obstructive disease. We showed that hypercholesterolemia caused a small set of VSMC to convert to a Sca1high mesenchymal stem-like cell, and Spry1 deficiency enhanced this effect. Loss of Spry1 in VSMC resulted in increased CD45+ cell infiltration into the subendothelial plaque layer, and advanced plaque burden and instability after prolonged high fat diet. In addition, Spry1 deficient VSMC had compromised restoration of SMA expression in the fibrous cap of advanced lesions. These studies in the mouse were correlated to loss-of-function studies using human VSMC. In vitro suppression of Spry1 in VSMC increased the expression of ICAM-1, Cox2, PAI1, and PDGFRb inflammatory mesenchymal markers, and promoted, to some extent, PDGFBB or TNFa-induced inflammatory effects. Together, our findings indicate that endogenous Spry1 suppresses the transition of VSMC into a proinflammatory mesenchymal transition, and promotes migration and differentiation of fibrous cap VSMC, suggesting that Spry1 plays a protective role in atherosclerosis. Despite lesion size not changing with the absence of Spry1 in our mouse model of atherogenesis, mortality in human disease is not correlated with plaque size, but rather plaque rupture due to instability [
]. Thus, Spry1 action to promote fibrous cap formation and collagen deposition is predicted to increase plaque stability.
Atherosclerosis is a chronic, progressive condition with a long asymptomatic phase. Most plaques remain subclinical, and vulnerable atherosclerotic plaques containing a thin inflamed fibrous cap over a large lipid core are particularly susceptible to disruption [
]. VSMC investment into and retention within the lesion and fibrous cap contributes to stabilization of atherosclerotic plaques. PDGF/PDGFRb signaling not only drives VSMC to a secretory phenotype, synergizes with hypercholesterolemia to promote the formation of advance atherosclerotic plaques, but also is essential for VSMC investment and transition to a SMA + myofibroblast-like state [
]. Our current results show that Spry1 deficiency increased PDGFRb expression, and to some extent synergized with PDGF- and TNFa-mediated VSMC inflammation, suggesting a negative role for Spry1 on PDGF and TNFa-initiated VSMC inflammation. Interestingly, PDGFBB and TNFa markedly suppressed Spry1 protein expression in VSMC. We previously reported that Spry1 is a key molecule to maintain VSMC in a contractile phenotype [
]. Therefore, VSMC-specific deletion of Spry1 may mediate, at least in part, some effects mediated by PDGF and TNFa signaling. However, VSMC-specific Spry1 deletion resulted in a compromised restoration of SMA expression in the fibrous cap. This and other decreased stability indices in advanced plaques indicates that Spry1 has a role in maintaining fibrous cap stability independent of PDGFR signaling.
Spry1 mRNA was only transiently down-regulated at 2 h following PDGF stimulation, and thereafter was restored in VSMC in vitro. The different dynamics of steady state transcripts versus protein is common in mammalian cells. Therefore, it is possible that Spry1 mRNA transcription and later translation is necessary for PDGF/PDGFRb signaling mediated SMA expression in cap VSMC. The gap between Spry1 protein down-regulation and restoration maybe also necessary to ensure PDGF-induced VSMC mesenchymal transition. Indeed, we have reported that growth factor enriched medium (SmGM2) transient down-regulates Spry1 protein in hVSMC in vitro [
]. Further investigation is needed to address the role of the dynamic expression of Spry1 protein downstream of PDGF/PDGFRb signaling. Currently, there is no available anti-SPRY1 specific antibody for immunofluoresence staining that could allow for more focused study of SPRY1 protein dynamics in VSMC.
Cellular clonal expansion contributing to human atherosclerotic lesions was first identified by Benditt et al. [
]. The concept of VSMC monoclonality in atherosclerotic lesions was consolidated using X-linked inactivation assays and polymerase chain reaction techniques [
]. Recently, use of multicolor reporter mice demonstrated that VSMCs in murine plaques are derived by the clonal expansion of a very few cells in the medial arterial wall [
Extensive proliferation of a subset of differentiated, yet plastic, medial vascular smooth muscle cells contributes to neointimal formation in mouse injury and atherosclerosis models.
]. The hypercholesterolemia-induced appearance of scattered Sca1high VSMC within the media in wild type mice may also be due to expansion and migration of clonal VSMC that contribute to the atherogenic lesion. Spry1 deficiency increased Sca1high cells in the medial layer of vessels, indicating that loss of Spry1 may increase conversion of VSMC into atherogenic prone cells or clonal cell expansion of plaque cells.
The remodeling and turnover of VSMC, which is achieved through processes such as selective cell proliferation/apoptosis or matrix synthesis/degradation, is very low in a normal adult artery [
] (production of Spry1 antibody used in the previous study was discontinued), the scattered Spry1 mRNA expression in the normal medial layer may result from low VSMC remodeling and turnover. Similarly, we observed a steady high tdTomato in the normal medial layer and a decreased/punctuated tdTomato signal in VSMC next to the atherosclerotic lesion, suggesting that loss of Spry1 increases VSMC remodeling and turnover.
VSMC-specific Spry1 deletion resulted in reduced fatty streak formation during early atherosclerosis despite an increased inflammatory phenotypic transition of VSMC. Atherosclerosis is a progressive process, in which endothelial cell dysfunction induced by oxidative stress initiates leukocyte infiltration and following fatty streak formation [
]. A single intravenous PCSK9DY-AAV injection followed by western diet feeding is an effective and relatively rapid approach that doesn't require multiple genetic mutations for atherogenesis; this model results in rapid and long-term sustained hyperlipidemia and atherosclerosis [
]. Using this method, we consistently generated hypercholesterolemia and atherogenesis. We observed slightly lower serum cholesterol levels in VSMC-specific Spry1 null mice compared to controls, without noticeable differences in food consumption or body weight gain. It is possible that loss of Spry1 in VSMC affects lipid metabolism, and future studies will address this possibility. Also in VSMC-specific Spry1 null mice, there was less CD45+ cell infiltration and smaller fatty streak formation compared to controls. In summary, our findings indicate that endogenous Spry1 suppresses the mesenchymal-like transition of VSMC, and thus is protective in chronic atherosclerotic disease.
Financial support
This project was supported by 2R01HL141149 (Liaw, PI), and 1P20GM121301 (Liaw, PI, Friesel pilot project PI; Histopathology and Microscopy Core Facility). Institutional core facilities at MaineHealth Institute for Research are also supported in part by 2U54GM115516 (Rosen and Stein, PIs).
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We would like to thank our institutional Histopathology and Microscopy Core (Armie Mangoba and Drs. Volkhard Lindner and Igor Prudovsky) for assistance with histology and microscopy.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Extensive proliferation of a subset of differentiated, yet plastic, medial vascular smooth muscle cells contributes to neointimal formation in mouse injury and atherosclerosis models.
00177-6&id=ga1.jpg){kind=link}
00177-6&id=gr1.jpg){kind=link}
00177-6&id=gr2.jpg){kind=link}
00177-6&id=gr3.jpg){kind=link}
00177-6&id=gr4.jpg){kind=link}
00177-6&id=gr5.jpg){kind=link}