henhaobio
Sign in

Nature CommunicationsPathologyPeer-reviewed journal article

TWIST1 drives endothelial-to-mesenchymal-transition to stabilize atherosclerotic plaques

Blanca Tardajos AyllóniD, Mannekomba R. DiagbougaiD, Ankita Das, Siyu Tian, Andreas EdsfeldtiD, Joanna Maria KaluckaiD, Jovana Serbanovic‐CaniciD, Emily V. ChambersiD, Jiangming SuniD, Chrysostomi GialeliiD, Mark DunningiD, Sheila FrancisiD, Xiuying LiiD, Akiko MammotoiD, Michael SimonsiD, Helle Færk JørgenseniD, Isabel GonçalvesiD, Suowen XuiD, Paul Charles EvansiD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

Atherosclerosis progresses from fatty streaks to complex plaques, with rupture leading to life-threatening complications. Endothelial-to-mesenchymal transition (EndMT) is associated with advanced atherosclerotic plaques, but its role in plaque progression remains unclear. To investigate this, we analyzed the role of TWIST1, a key EndMT-driving transcription factor, in plaque development. Using single-cell RNA sequencing of atherosclerotic plaques from hypercholesterolemic mice with inducible deletion of Twist1 from endothelial cells (Twist1ECKO ApoE-/-) we demonstrate that Twist1 regulates endothelial cell heterogeneity by promoting cell states expressing EndMT markers. Cell tracking confirmed that Twist1 induces EndMT in advanced plaques. Mechanistically, we found that TWIST1 contributes to EndMT by promoting endothelial migration and proliferation through the transcriptional coactivator PELP1. Additionally, TWIST1 induces AEBP1 transcription, which upregulates COL4A1 to drive endothelial proliferation. Analyses of murine brachiocephalic plaques show that endothelial Twist1 promotes plaque growth, collagen deposition, and ACTA2-positive cell accumulation, hallmarks of plaque stability, while reducing necrosis and macrophage infiltration, features of plaque instability. Moreover, TWIST1 expression was associated with asymptomatic human carotid atherosclerosis and predicts favourable clinical outcomes. These findings challenge the prevailing view that EndMT destabilizes plaques, by suggesting that TWIST1-driven EndMT can promote plaque stability, offering new insights into atherosclerosis pathophysiology and therapeutic potential.

In-depth summary

Claude Opus 5.5 · xhigh effort · 5 Nov 2026 · from the full text of version 1 (24 May 2025)

Endothelial Twist1 drives partial and advanced endothelial-to-mesenchymal transition via PELP1 and AEBP1-COL4A1 signaling, increasing collagen and ACTA2+ cells to stabilize atherosclerotic plaques in male mice and human carotid arteries.

Endothelial-to-mesenchymal transition (EndMT) occurs in advanced atherosclerotic lesions and has been widely viewed as a destabilizing, pathogenic driver of plaque rupture. Inducible endothelial deletion of Twist1 in hypercholesterolemic mice and transcriptomic profiling of human carotid endarterectomies demonstrate that TWIST1-driven EndMT instead promotes plaque stability features and predicts fewer postoperative cardiovascular events.

Key results

  1. ShownInducible endothelial deletion of Twist1 in hypercholesterolemic mice reduces EndMT-enriched endothelial clusters and decreases lineage-traced endothelial-derived ACTA2+ cells inside advanced plaques.

    SORT-seq scRNA-seq of 2,139 aortic CD31+CD45- cells (1,173 control, 966 Twist1^ECKO ApoE-/-) identified 11 clusters, where Twist1 promoted EndMT-enriched clusters 2, 7 (partial EndMT), and 8 (advanced EndMT) and suppressed clusters 4 and 5; in Rosa26^tdTomato lineage-traced brachiocephalic plaques, Twist1^ECKO (n=10 vs n=11 controls) significantly reduced plaque tdTomato+ EndMT cells. · Twist1 regulates EC clusters enriched for EndMT…; Twist1 controls EndMT in advanced…; Bioinformatics; Figure legends

  2. ShownUnder disturbed flow, TWIST1 drives human aortic endothelial cell (HAEC) migration and proliferation by directly activating PELP1 and an AEBP1-COL4A1 transcriptional axis.

    Silencing TWIST1 in disturbed-flow HAECs altered 293 genes in bulk RNA-seq (n=4), overlapped with murine plaque scRNA-seq on 45 positively regulated genes, and impaired scratch migration and Ki67+ proliferation; ChIP-qPCR (n=4) confirmed TWIST1 binding to PELP1, AEBP1, and COL4A1 loci, PELP1 regulated 34% (99/288) of TWIST1 target genes, and exogenous collagen IV rescued proliferation in TWIST1-silenced cells. · Molecular mechanisms of TWIST1-driven EndMT; Figure legends

  3. ShownEndothelial Twist1 promotes brachiocephalic plaque growth and stability features in male ApoE-/- mice, but has no effect on EndMT or plaque composition in female mice.

    In male Twist1^ECKO ApoE-/- mice (n=9-11) versus controls (n=14-15) after 14 weeks of Western diet, Twist1 deletion reduced total plaque burden, fibrillar collagen (Picrosirius Red), and ACTA2+ cell content while increasing MAC3+ macrophages, necrotic core area, and elastin breaks without altering plasma lipids; female mice showed no differences in EndMT or plaque parameters. · Twist1 promotes features of plaque stability; Sex differences in Twist1 control of EndMT and…; Figure legends

  4. ShownHigher TWIST1 mRNA expression in human carotid endarterectomy plaques associates with asymptomatic disease and lower postoperative cardiovascular event risk.

    In the Carotid Plaque Imaging Project (CPIP) biobank, bulk RNA-seq of the most stenotic plaque region showed significantly higher TWIST1 expression in asymptomatic (n=27) than symptomatic (n=51) patients, and patients in the 2nd and 3rd TWIST1 expression tertiles had reduced cardiovascular events over up to 5 years of follow-up compared to the 1st tertile. · TWIST1 predicts stabilization of human carotid…; Human carotid plaque clinical data, bulk…; Figure legends

Methods, data and reproducibility

Induced endothelial Twist1 deletion with tamoxifen at week 8 of a 14-week Western diet in Twist1^ECKO ApoE-/- mice and analyzed aortic ECs by SORT-seq scRNA-seq and brachiocephalic plaques by histology. Performed siRNA screening, bulk RNA-seq, and TWIST1-FLAG ChIP-qPCR in disturbed-flow HAECs, and analyzed bulk RNA-seq and outcomes in human carotid plaques.

  • FACS-sorted aortic endothelial cells (CD31+CD45-) from male Twist1^ECKO ApoE-/- and control mice after 14 weeks of Western diet (SORT-seq scRNA-seq) · mouse · n = 2,139 filtered cells (1,173 control, n=4 mice; 966 Twist1^ECKO, n=5 mice) · GSE293412
  • Human aortic endothelial cells (HAECs) under disturbed flow with siRNA silencing of TWIST1, PELP1, or AEBP1 (bulk RNA-seq, ChIP-qPCR, migration/proliferation) · human · n = 4 donors (3 male, 1 female; n=3-4 per experiment) · GSE293412
  • Brachiocephalic artery plaques from male and female Twist1^ECKO ApoE-/- and Rosa26^tdTomato lineage-traced mice · mouse · n = n=9-12 Twist1^ECKO and n=11-15 control male mice
  • Human carotid endarterectomy plaques from the Carotid Plaque Imaging Project (CPIP, Malmo, Sweden) biobank (bulk RNA-seq and follow-up) · human · n = 51 symptomatic and 27 asymptomatic patients (RNA-seq); n=3 IF
Methods & tools
STAR, Seurat, BBrowserX, AUCell, nf-core/rnaseq, Salmon, DESeq2, tximport, edgeR, MetaCore, Metascape, QuPath
Reagents & environment
R package Seurat (v3.1); nf.core rna-seq pipeline (version 3.6.0); GENCODE transcriptome release 27 (S15, S19, S25).
Reproducibility gaps
Mouse and HAEC sequencing data are deposited in GEO (GSE293412), whereas human carotid plaque data from the CPIP biobank are under GDPR-controlled access and available only in group form upon request (S28).
Limitations (2)
  • Endothelial Twist1 deletion affected EndMT, plaque burden, and plaque stability features only in male mice and had no effect in female mice (S7, S9). · stated by the authors
  • While TWIST1 promoted protective EndMT clusters (clusters 2, 7, and 8), EndMT cluster 3 was insensitive to Twist1 deletion, leaving the transcription factors driving other potentially pathogenic EndMT subsets unidentified (S3, S9). · stated by the authors

Read this if you study endothelial-to-mesenchymal transition, shear-stress mechanotransduction, or cellular mechanisms of atherosclerotic plaque stability.

Citations
10
FWCI
26.49
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISWatts '26Slenders '25Peng '24Weiser-Evan '23Li '26Luo '26
68% semantic matchGenomics
Kelsey M. Watts, Katyayani Sukhavasi, Rebecca Hernandez+10 more·bioRxiv·Feb 9, 2026