thank you. Please listen to this article using the above players. ✖
Want to listen to this article for free?
Complete the form below to unlock access to all audio articles.
Researchers and collaborators at Southern Medical University report identifying FDA-approved compounds that can reduce blood lipids by combining computational screening with clinical and experimental validation. The work will be displayed in Acta Pharmacolica Sinica It also addresses the need for alternative therapies in patients who cannot use or respond to standard lipid-lowering drugs.
Screening of existing drugs using multi-model machine learning
Start with 3 430 FDA approved compounds (176 known lipid-lowering agents and 3 254 Other) The team extracted molecular descriptors and fingerprints from smile codes and physicochemical data. Features narrowed through Spearman's correlation and Lasso's regression. A suite of 68 machine learning models, including a combination of random forests, support vector machines, gradient boosts, and elastic nets, was trained and evaluated using AUC accuracy F1 recall specificity metrics. The top performance model reached AUC≈ 0.886 and accuracy≈ 0.888. 29 reuse candidates were generated for drugs predicted to be positive in at least eight of the top 10 models.
Retrospective clinical data supports four candidates
Medical records from nearly 25 The year at Zhujiang Hospital (June 1998 – May 2024) was reviewed for patients treated with model predictive drugs. Four drugs – argatroban, leboxyl (levothyroxine), oseltamivir, and thiamine showed statistically significant lipid changes. Argatroban (n = 63) LDL has decreased by 33 Percentage (2.96 mmol/l to 1.98 mmol/l), total cholesterol 25 Percentages (4.68-3.51 mmol/l) and triglycerides were also reduced (all p < 1 × 10⁻⁸). Leboxil User (n = 87) LDL and TC reduction was 16 Percentage and 12 Percentage each. Oseltamivir and thiamine showed moderate lipid effects.
Mouse studies confirm lipid modulation of multiple drugs
Sixteen drugs selected from computational predictions and clinical evidence were tested in male C57BL/6 mice. Argatroban and Promega reduced total cholesterol to 10 percent. Leboxyl and sulfaphenazole lowered triglycerides by about 27-29 respectively. percent. Prasterone, Alpha Tocopherol Acetate, Sorafenib, Cedazuridine, and Promega significantly increased HDL levels. percent). Unexpectedly, LDL was increased conservatively in some drug groups (such as Procarbazine, Dimenhydrinate, Promega) compared to controls. All findings are based on animal preclinical data and require further investigation.
Molecular docking and dynamics highlight possible mechanisms
Seven promising drugs (argatroban, promega, sulfafenazole, sorafenib, plastrone, leboxyl, and alpha-tocopherolacetic acid) were docked against 12 lipid metabolic targets, including HMG‑CoA reductase, aggregation factor X, and serotonin receptors (Htr2A/C, 5‑HT4R HOM HOM HOM HOM HOM HOM HOM HOM HOMENTORS). TRβ), MTP, RXRα and COX‑ 2. Argatroban bound tightly to coagulation factor X (≈ –7.6 kcal/mol) Hydrophobic interactions and stable hydrogen bond formation in molecular dynamics simulations. Leboxil showed a high affinity for TRα. Sulfaphenazole-binding serotonin receptor subtype. Prasterone was involved in RXRα and COX‑2. Promega related to MTP. Sorafenib showed affinity for HMG‑COA reductase. The results suggest that multiple candidate drugs act through different lipid pathways.
Limitations of research strengths and future directions
This study integrates large-scale predictive modeling with clinical and animal data and mechanistic docking to identify 29 candidate lipid-lowering agents. Argatroban, leboxyl, and sulfate sole appear to be particularly promising based on consistency between methods. Limitations include reliance on retrospective single-centre clinical data and mouse models that may not be translated into humans. Randomized controlled trials and biochemical studies are required to confirm human efficacy and safety. Heterogeneity of positive control drugs suggests that stratified predictive models may improve future accuracy. This framework may apply to drugs that are reused in other therapeutic areas.
reference: The integration of machine learning and experimental validation, including Chen J‑H, Li K‑X, and Fan C‑F, reveals new lipid-lowering drug candidates. Acta Pharmacol Sin. 2025. DOI: 10.1038/S41401‑025‑01539‑1
This content contains text generated with AI support. You can find the AI policy for your technology network here.
