The cells in question are senescent vascular smooth muscle cells, and recent single-cell lineage-tracing studies have pushed a striking reframing of atherosclerosis as a tumor-like disease driven by these very cells. Within plaques, vascular smooth muscle cells undergo clonal expansion, dedifferentiation, and metabolic reprogramming, forming pathological structures that resemble solid tumors. The senescent subset behaves like drug-resistant dormant tumor cells: they persist indefinitely, resist apoptosis, and remodel their surroundings through the senescence-associated secretory phenotype, a cocktail of inflammatory signals that drives necrotic core expansion and thins the fibrous cap that keeps plaques from rupturing. Selectively killing these cells, an approach known as senolysis, has already shown promise for extending lifespan in animal studies, but translating that promise into cardiovascular medicine has been hampered by toxicity concerns with existing drug candidates.

The new study began with a deceptively simple observation. When the researchers modeled senescence in human aortic smooth muscle cells by exposing them to palmitate, a lipid that mimics the lipotoxic stress of atherosclerosis, the cells entered a stable senescent state marked by elevated beta-galactosidase activity and upregulation of the cell-cycle inhibitors p16 and p21. Crucially, the cells also ramped up production of heat shock protein 90 alpha, or HSP90α, a chaperone molecule famous in oncology as a scaffold that supports the survival of tumor cells. Given the parallels between tumorigenesis and atherogenesis, the team hypothesized that senescent smooth muscle cells might exploit the same HSP90α machinery to maintain homeostasis and evade death in the hostile plaque microenvironment.

Proteomic analysis after HSP90α knockdown revealed what that machinery was protecting. Silencing HSP90α specifically activated the PERK/eIF2α/ATF4/CHOP arm of the endoplasmic reticulum stress response, suggesting that high HSP90α levels actively restrain this stress pathway in senescent cells. The team then turned to human tissue, integrating four published single-cell RNA sequencing datasets covering 40 samples of human atherosclerotic lesions from carotid endarterectomy and coronary arteries. In senescent smooth muscle cells identified by co-expression of p21 and p16, HSP90α was highly expressed, but the CHOP stress marker was conspicuously absent. Immunofluorescence of human carotid plaques confirmed the same spatial dissociation, and cultured senescent cells showed suppression of the PERK branch specifically, while the ATF6 and IRE1 branches of the unfolded protein response remained untouched.

The mechanism behind that selective suppression proved to be an elegant piece of molecular scaffolding. Prior work had shown that the kinase AKT can bind to the autophosphorylation site of PERK and block its activation. Using protein-protein docking, co-immunoprecipitation, and domain mapping, the researchers demonstrated that HSP90α holds AKT and PERK together in a ternary complex, with the middle region of HSP90α, amino acids 284 through 620, sufficient to bind both partners. In senescent cells this interaction intensifies, stabilizing AKT and keeping PERK dormant. When HSP90α was removed, AKT degraded rapidly through the proteasome, PERK signaling surged, and the senescent cells began to die. Reactivating the pathway with thapsigargin, a specific inducer of endoplasmic reticulum stress, collapsed mitochondrial membrane potential and pushed the cells into apoptosis while reducing their senescence markers.

The downstream killing mechanism runs through calcium. PERK activation is known to tighten physical contacts between the endoplasmic reticulum and mitochondria, and the team documented exactly that: after HSP90α knockdown or thapsigargin treatment, transmission electron microscopy showed expanded mitochondria-associated ER membranes, fluorescent probes confirmed elevated mitochondrial calcium, and the calcium flux funneled through the mitochondrial calcium uniporter, MCU. Blocking mitochondrial calcium uptake with ruthenium red, or silencing MCU directly, rescued the cells from apoptosis. In other words, the senescent cells survive only as long as PERK stays suppressed; once the brake is released, calcium floods the mitochondria, membrane potential collapses, and the cell self-destructs.

Having defined the target, the team went hunting for a drug. They compiled a dataset of 112 natural products with documented anti-atherosclerotic activity from a literature search spanning 1947 to 2024, then docked each compound against the crystal structure of HSP90α using three independent docking programs and a robust rank aggregation method. The top hit was salvianolic acid A, or SAA, a water-soluble polyphenol from Salvia miltiorrhiza, the red sage root used in traditional Chinese medicine. Its docking score rivaled that of 17-DMAG, a well-characterized synthetic HSP90 inhibitor. Two independent binding assays sealed the case: a cellular thermal shift assay showed that SAA stabilizes HSP90α against heat denaturation, and a biotinylated SAA probe pulled down endogenous HSP90α from cell lysates, confirming direct physical binding.

SAA’s effects on senescent cells were strikingly selective. At concentrations of 20 to 100 micromolar, it triggered apoptosis in senescent smooth muscle cells, reduced p16 and p21, and suppressed secretion of the inflammatory SASP factors IL-1α and IL-6, while leaving normal smooth muscle cells unharmed. Mechanistically, SAA disrupted the HSP90α-AKT-PERK ternary complex, shifted AKT ubiquitination from the K63-linked form toward K48-linked chains that mark proteins for proteasomal destruction, and thereby reactivated PERK signaling. Overexpressing HSP90α reversed all of these effects, and knocking down PERK rescued the cells, confirming that the pathway runs through the complex the team had identified. A cross-linking assay even captured the intact 340-kilodalton ternary complex and watched it dissolve after SAA treatment.

The in vivo results were equally compelling. In ApoE-deficient mice fed a high-fat diet for 20 weeks, intraperitoneal SAA from week 9 onward significantly reduced plaque area, shrank necrotic cores, thickened fibrous caps, and cut lipid accumulation in the aorta, while lowering serum IL-6. Using adeno-associated viruses to manipulate HSP90α specifically in vascular smooth muscle cells, the researchers showed that VSMC-specific knockdown increased p21-CHOP colocalization in plaques, whereas overexpression blunted SAA’s effects and caused a rebound of p21-positive cells. Co-immunoprecipitation from mouse aortas confirmed that SAA weakened the HSP90α-AKT-PERK interaction in living tissue, closing the loop between the cell culture mechanism and the animal outcome.

What makes this work resonate beyond the lab bench is SAA’s clinical head start. Unlike geldanamycin-derived HSP90 inhibitors, whose toxicity has blocked every approval attempt in oncology, SAA is already in a phase II trial for diabetic microangiopathy and two phase I trials for angina pectoris and diabetes, suggesting a tolerability profile that could accelerate translation. The authors caution that mouse-to-human translation is complex and that whether SAA can actually reverse plaque progression will require endpoint clinical trials. Still, the study delivers something the senolytics field has lacked: a senolytic strategy grounded in a defined survival mechanism of the target cells, delivered by a natural compound with human safety data, and validated from molecular docking all the way to plaque stabilization. If the findings hold, the humble red sage root may offer a way to evict the zombie cells that make aging arteries fragile.

Subject of Research: Targeting the HSP90α-AKT-PERK complex in senescent vascular smooth muscle cells as a senolytic strategy against atherosclerosis

Article Title: Salvianolic Acid a Disrupts the HSP90α‐AKT‐PERK Ternary Complex to Alleviate Atherosclerosis by Activating Endoplasmic Reticulum Stress of Senescent Vascular Smooth Muscle Cells

Article References: Guan, X., Lu, M., Cui, X., Zhang, D., Zhang, L., Lin, L., Li, Y., Hu, Y., & Li, C. (2026). Salvianolic Acid a Disrupts the HSP90α‐AKT‐PERK Ternary Complex to Alleviate Atherosclerosis by Activating Endoplasmic Reticulum Stress of Senescent Vascular Smooth Muscle Cells. Advanced Science, Article e77891. https://doi.org/10.1002/advs.77891

Image Credits: AI Generated

DOI: 10.1002/advs.77891

Keywords: atherosclerosis, senescence, senolytics, HSP90α, PERK, endoplasmic reticulum stress, vascular smooth muscle cells, salvianolic acid A, SASP, mitochondrial calcium, AKT, plaque stability