Sclerostin, which has three loops, inhibits bone formation and impairs whole-body lipid and glucose metabolism. The marketed therapeutic sclerostin antibody for postmenopausal osteoporosis (POP) mainly targeting loop2 promotes bone formation and improves whole-body lipid and glucose metabolism. However, FDA/EMA warns of its cardiovascular risk. We previously demonstrate that sclerostin loop3 contributes to the inhibitory effect of sclerostin on bone formation but not its cardioprotective effect. Here we find elevated serum sclerostin levels in both POP-T2DM patients and newly-diagnosed T2DM patients and further demonstrate that sclerostin loop3 participates in the impairment effect of sclerostin on whole-body lipid and glucose metabolism in vivo. Mechanistically, specific blockade of adipocytic sclerostin loop3-LRP4 interaction attenuates the impairment effect of sclerostin on lipid and glucose metabolism in vitro and in vivo. This study provides an innovative strategy, blocking adipocytic sclerostin loop3-LRP4 interaction, to normalize lipid and glucose metabolism in POP-T2DM patients, in cardiovascular safety.

Dickkopf-1 (DKK1) is a secreted glycoprotein that traditionally acts as an antagonist of canonical Wnt/β-catenin signaling. Although it functions as a tumor suppressor in some specific biological background and disease stages, growing evidence links DKK1 to tumor progression, immune evasion, and therapy resistance in a variety of multiple malignancies. This review provides a comprehensive bench-to-bedside overview of DKK1 in cancer. We first delineate how DKK1 regulates both Wnt-dependent and Wnt-independent pathways. From a clinical perspective, we evaluate the application potential of DKK1 as a diagnostic and prognostic biomarker. We further discuss the progress of DKK1-targeted interventions, ranging from monoclonal antibodies in clinical trials to next-generation therapeutic modalities. Finally, we discuss the challenges in clinical translation and suggest future directions for DKK1-based precision medicine. In summary, by integrating preclinical insights with current clinical data, this review provides a strategic roadmap for advancing DKK1-targeted therapies in cancer.

Conventional chemotherapy frequently causes uneven drug distribution and systemic toxicity due to inadequate in vivo selectivity. Passive targeting strategies, such as the enhanced permeability and retention effect, can no longer guarantee effective drug accumulation within tumor tissues. The success of active targeted chemotherapy cannot be simplified to a basic “ligand-drug” binding model, but relies on the coordinated regulation of multiple processes, including drug distribution, tumor penetration, cellular internalization, and intracellular processing. From a design-oriented perspective, this review elaborates on the core design logic of targeted chemotherapy delivery and highlights that target expression and accessibility, pharmacokinetic characteristics of targeting moieties, payload compatibility, and drug release mechanisms serve as critical determinants of therapeutic outcomes. Furthermore, this review summarizes and evaluates current mainstream targeted chemotherapy delivery platforms, including antibody-based, aptamer-based, small-molecule, peptide, and hybrid delivery systems. Finally, future research directions for targeted chemotherapy development are prospected. Rather than pursuing a single optimal delivery platform, a problem-oriented design strategy is proposed. The design of next-generation delivery systems requires an optimal balance among targeting construct size, in vivo drug exposure, and tissue permeability, as well as payload release profiles, based on the dominant transport limitations of specific tumor microenvironments.

Myostatin (Mstn), a well-characterized member of the transforming growth factor-β (TGF-β) superfamily, serves as a key negative regulator of skeletal muscle mass. Its overactivation is closely associated with the pathogenesis of various musculoskeletal and metabolic disorders. Over the past decades, inhibiting Mstn has emerged as a promising therapeutic strategy to promote muscle growth. A range of Mstn-targeted inhibitors has been developed, yielding encouraging preclinical and clinical outcomes. These include small molecules, monoclonal antibodies, peptibodies, and gene therapy-based approaches. This review summarizes the biological structure and function of Mstn, provides a comprehensive overview of recent advances in Mstn-targeted therapeutics, and offers critical insights into future directions for drug development and clinical translation.

Background:The reliability of meta-analyses on Chinese herbal medicine (CHM) largely depends on the methodological quality of the included randomized controlled trials (RCTs). This study assessed the risk of bias (RoB) among Chinese-language CHM RCTs embedded in meta-analyses (MAs) published between 2021 and 2022. Methods:Chinese-language CHM MAs were identified through literature searches across seven databases (three international and four Chinese databases). RCTs were extracted from the eligible MAs and their methodological quality was appraised using the Cochrane Risk of Bias (RoB) tool. Logistic regression analyses were conducted to explore associations between bibliographical characteristics and quality. Results:Of 2060 RCTs (published between 1997-2021), 93.0% (n = 1916) had low RoB in incomplete data, but only 43.98% (n = 906) in random sequence generation. No RCTs had low RoB in selective outcome reporting. Non-patent CHM RCTs performed poorly in sequence generation, allocation concealment, and outcome assessment blinding but well in incomplete data. Government-funded RCTs showed better performance in sequence generation and outcome assessment but worse in incomplete data. No analysis was conducted for selective reporting due to lack of low RoB RCTs. Conclusion:The methodological quality of Chinese-language RCTs embedded in recently published MAs was found to be unsatisfactory, indicating their limited utility in informing clinical decision-making. There is an urgent need for enhancing clinical trial training for researchers, reviewers, and editors in China.

Triple-negative breast cancer (TNBC) urgently requires promising therapeutic targets. This study identifies sclerostin, an osteocyte-derived secretory protein traditionally linked to bone homeostasis, as an unexpected intracellular oncogenic driver in TNBC. Although genetic ablation of sclerostin markedly suppresses tumor progression and lung metastasis, neither its antibody nor recombinant protein exerts any effects, excluding the role of extracellular sclerostin in TNBC. Genetic and pharmacological approaches (sclerostin aptamer-based proteolysis-targeting chimera with potent intracellular sclerostin-degrading activity, Apc101) show the emerging role of intracellular sclerostin in promoting TNBC progression and metastasis. Notably, in both TNBC cell-derived and patient-derived xenograft models, Apc101 significantly suppresses tumor progression. Mechanistically, intracellular sclerostin interacts with caprin1 to stabilize CDK1 and Cyclin B1 mRNAs. Collectively, this study reveals an oncogenic function of intracellular sclerostin in TNBC and proposes that targeting it represents a promising therapeutic strategy.

Previously, we reported a dual combination based on 4-hydroxycoumarin and dodecanedioic acid that could synergistically bind to human serum albumin (HSA). However, optimizing this combination remains challenging and could often be guided by empirical selection and extensive experimental screening, which may limit the chemical diversity and suboptimal affinity. In this study, we established a systematic artificial intelligence framework that integrates computational optimization with wet-lab synthesis and experimental validation, enabling improvement of the dual combination while preserving the core chemotypes. We first trained a machine learning classifier on curated HSA binding data and used it as an external scoring function to guide reinforcement learning-driven scaffold decoration with LibINVENT, enabling goal-directed generation of coumarin derivatives and fatty acid derivatives. Candidate molecules were prioritized through multiparameter filtering and diversity-aware selection, followed by synthesis and experimental validation using surface plasmon resonance. The optimized representatives show nanomolar HSA binding and enhanced affinity compared to the original ligands. Molecular docking and molecular dynamics simulations further provide a mechanistic rationale for the affinity improvements by revealing additional stabilizing interactions and more favorable binding energetics at the corresponding HSA sites. Besides, the optimized coumarin derivative (CD1) is a warfarin-derived coumarin analogue, yet it did not show detectable anticoagulant activity in an acute clotting time assay, whereas warfarin did. Overall, this work demonstrates a practical AI-guided route to expand chemical diversity and improve affinity for a synergistic HSA binding combination.

Background Mechanical unloading leads to bone loss and cardiovascular deconditioning, accompanied by elevated sclerostin expression. Genetic Sost knockout or pharmacologic sclerostin antibody treatment was reported to counteract bone loss during mechanical unloading in mice. However, severe cardiovascular events were reported in postmenopausal osteoporotic patients treated with commercially available sclerostin antibody targeting loop2. It is desirable to develop a precise sclerostin inhibition strategy to counteract unloading-induced bone loss, without increasing cardiovascular risk. Methods and results In a previously published rodent studies under normal loading condition, it was found that sclerostin loop3 participated in the inhibitory effect of sclerostin on bone formation, while the preventive action of sclerostin against cardiovascular events was independent of sclerostin loop3. Nevertheless, whether and how sclerostin loop3 contributes to bone formation reduction and bone loss under mechanical unloading condition remains unclear. In this study under mechanical unloading condition, either sclerostin loop3-specific deficiency in Sostloop3−/− mice or sclerostin loop3-specific inhibition by our tailor-made aptamer Apc001 counteracted unloading-induced bone loss without increasing arterial stiffness, whereas either Sost knockout or romosozumab treatment significantly increased unloading-induced arterial stiffness in mice. These findings indicated sclerostin loop3 as a therapeutic target with cardiovascular safety against unloading-induced bone loss. Mechanistically, we identified that sclerostin loop3 bound to LRP4 in osteoblasts under mechanical unloading condition. Osteoblast-specific Lrp4 knockout counteracted unloading-induced bone formation reduction and bone loss in OB. Lrp4−/− mice. Further, blocking the interaction of sclerostin loop3 with LRP4 via mutation of the interaction residues (Lrp4m) or pharmacologic inhibition with LRP4 peptide tool (LRP4-Pep) dramatically attenuated binding of sclerostin to LRP6, counteracted decrease of Wnt/β-catenin signaling activity and osteogenic potential in osteoblasts under mechanical unloading condition in vitro. Consistently, Lrp4m counteracted unloading-induced bone formation reduction and bone loss in mice in vivo. In Lrp4m/OB-Lrp4 mice, osteoblast-conditional correction of Lrp4m to wild-type Lrp4 attenuated the counteractive effect of Lrp4m on unloading-induced bone loss. Pharmacologically, osteoblasts-targeted LRP4-Pep counteracted bone formation reduction and bone loss during mechanical unloading in wild-type mice. Conclusion Sclerostin loop3-mediated anchoring of sclerostin to LRP4 facilitated its binding to LRP6 in osteoblasts, contributing to bone formation reduction and bone loss under mechanical unloading condition. The translational potential of this article Specifically blocking the interaction of sclerostin loop3 with LRP4 in osteoblasts would offer a precise strategy with cardiovascular safety for treatment of unloading-induced bone loss.

Background The reliability of meta-analyses on Chinese herbal medicine (CHM) largely depends on the methodological quality of the included randomized controlled trials (RCTs). This study assessed the risk of bias (RoB) among Chinese-language CHM RCTs embedded in meta-analyses (MAs) published between 2021 and 2022. Methods Chinese language CHM MAs were identified through literature searches across seven databases (three international and four Chinese databases). RCTs were extracted from the eligible MAs and their methodological quality was appraised using the Cochrane Risk of Bias (RoB) tool. Logistic regression analyses were conducted to explore associations between bibliographical characteristics and quality. Results Of 2,060 RCTs (published between 1997-2021), 93.01% (n=1,916) had low RoB in incomplete data, but only 43.90% (n=906) in random sequence generation. No RCTs had low RoB in selective outcome reporting. Non-patent CHM RCTs performed poorly in sequence generation, allocation concealment, and outcome assessment blinding but well in incomplete data. Government-funded RCTs showed better performance in sequence generation and outcome assessment but worse in incomplete data. No analysis was done for selective reporting due to lack of low RoB RCTs. Conclusion The methodological quality of Chinese language RCTs embedded in recently published MAs was found to be unsatisfactory, indicating their limited utility in informing clinical decision. There is an urgent need for enhancing clinical trial training for researchers, reviewers, and editors in China.

Duchenne muscular dystrophy (DMD), a fatal X-linked disorder, features progressive muscle fibrosis as a key driver of mortality. While CTGF represents a therapeutic target for DMD, its VWC-domain-targeting antibody (FG-3019) failed in clinical trials. Through experimental validation, we identified CT-domain as a superior target domain, as it contributed more fibrosis activity of CTGF than VWC-domain without elevating compensatory TGF-β1 level. Aptamers are synthetic oligonucleotides identified through SELEX, which can specifically bind to flexible protein domains through their unique 3D conformations. Their small molecular size enables effective tissue penetration while maintaining high target specificity, making them ideal CT-domain inhibitors. Nevertheless, conventional SELEX involves time-consuming and inefficient multiple screening rounds. Here, we employed our generative AI model, AptGEN, to rapidly discover a potent CT-domain specific aptamer within 42 days. This chemically modified aptamer (Apc003OA) distributed and remained in muscle tissues for an extended period, whereas FG-3019 could not. Importantly, it demonstrated better fibrosis inhibitory activity in vitro and in mdx mice when compared to FG-3019. Furthermore, Apc003OA demonstrated a favorable safety profile in mdx mice. Within 10 months, we progressed from target domain discovery, aptamer drug discovery, and then obtained both Orphan Drug Designation and Pediatric Rare Disease Designation by US Food and Drug Administration.

Nucleic acid aptamers are synthetic, single-stranded oligonucleotides that bind targets with high affinity and specificity, enabling precise and programmable functional modulation. However, low biostability remains a key druggability bottleneck of therapeutic aptamers. In vivo enzymatic lability constitutes a primary pharmacokinetic barrier, impeding systemic exposure, tissue distribution, and sustained target occupancy, all critical determinants of in vivo efficacy. Consequently, rational chemical stabilization via enzyme-resistant modifications has emerged as a cornerstone strategy in aptamer drug development. In this review, we systematically categorize current stabilization approaches across two complementary design dimensions: local-level modifications (including terminal modification, sugar modification, backbone modification, and base modification) and global structural-level engineering (including Spiegelmers, cyclization modification, and multivalent assembly). Furthermore, we discuss persistent translational challenges to illuminate a coherent framework for designing aptamers with high biostability that achieve enhanced nuclease resistance while concurrently exhibiting favorable pharmacokinetics, high target binding affinity and specificity.

Abstract Sclerostin negatively regulates bone formation. The marketed antibody against sclerostin loop2 promoted bone formation but may have caused severe cardiovascular events in clinical use. In our published studies, sclerostin loop3 was found to be involved in inhibitory effects of sclerostin on bone formation, whereas cardiovascular protective effects of sclerostin in mice were independent of loop3. It is necessary to investigate how sclerostin loop3 participates in the inhibitory effects of sclerostin on bone formation to facilitate developing precise strategies that promote bone formation without increasing cardiovascular risk. In this study, sclerostin loop3 was identified to bind to LRP4, thereby facilitating binding of sclerostin to LRP6 in osteoblasts. Blockade of sclerostin loop3-LRP4 interaction by both Lrp4 mutation (Lrp4m) and blocking peptide (LRP4-Pep) diminished the antagonistic effect of sclerostin on Wnt/β-catenin signaling in osteoblasts in vitro. Consistently, Lrp4m promoted bone formation in Lrp4m mice in vivo. Mechanistically, osteoblast-conditional correction of Lrp4m to wild-type Lrp4 resulted in significantly lower bone formation than Lrp4m mice, indicating that the promotive effects of Lrp4m on bone formation acted in osteoblasts in vivo. Moreover, re-expression of sclerostin dramatically inhibited bone formation in sost −/− mice, whilst the inhibitory effects of sclerostin were significantly weaker in sost −/− .Lrp4m mice. Pharmacologically, LRP4-Pep diminished the inhibitory effects of sclerostin on bone formation in SOST ki mice. Taken together, osteoblastic sclerostin loop3-LRP4 interaction, as an anchor, was required by sclerostin to bind to LRP6, thereby inhibiting bone formation. Translationally, blockade of sclerostin loop3-LRP4 interaction in osteoblasts would provide precise therapeutic strategies to promote bone formation without increasing cardiovascular risk.

Paclitaxel resistance and poor tumor selectivity remain significant challenges in epithelial ovarian cancer therapy. To overcome these challenges, we engineered PSaA360, a structurally constrained aptamer-drug conjugate with a dual-functional molecular lock that simultaneously rigidifies the AS1411 aptamer and delivers potent telomerase inhibition. Unlike the conformational flexibility of conventional aptamer-drug conjugates, PSaA360 employs the G-quadruplex stabilizer 360A to simultaneously rigidify the AS1411 aptamer into a high-affinity conformation and deliver potent telomerase inhibition. This structure-constrained and therapy-integrated strategy improved nucleolin binding, enhanced cellular internalization, and counteracted paclitaxel chemoresistance. In vivo, PSaA360 exhibited marked tumor inhibition with minimal systemic toxicity. By transforming a therapeutic agent into a structural stabilizer, PSaA360 establishes a new paradigm for mechanism-guided aptamer engineering in chemotherapy-resistant malignancies.

Nucleic acid aptamers are emerging as powerful “chemical antibodies,” namely synthetic oligonucleotide ligands that emulate antibody-like molecular recognition while retaining the programmability and scalability of chemical synthesis, for precision medicine. Yet their clinical translation has been historically bottlenecked by the stochastic nature of classical discovery methods like SELEX. This trial-and-error paradigm is labor-intensive and often fails to identify binders with optimal therapeutic properties. The integration of Artificial Intelligence (AI) is currently driving a fundamental paradigm shift from empirical screening to rational engineering. This review synthesizes the rapid evolution of this landscape, framing High-Throughput SELEX not merely as a selection tool, but as a critical data engine fueling computational discovery. We categorize current AI frameworks into two distinct paradigms: predictive architectures that decipher complex sequence-structure-function rules from noisy datasets, and generative models (e.g., VAEs, diffusion models, and inverse folding) that enable the de novo design of binders targeting specific geometries. Furthermore, we highlight translational breakthroughs in AI-designed biosensors and chemically optimized therapeutics. Finally, we discuss current challenges in data sparsity and interpretability, outlining a roadmap toward closed-loop ecosystems, that is, iterative design–test–refine workflows in which computational design, experimental validation, and model updating are continuously linked, to accelerate the development of next-generation molecular tools.

Proteolysis targeting chimeras (PROTACs) have emerged as a groundbreaking class of anticancer therapeutics. These bifunctional molecules harness the endogenous ubiquitin-proteasome system to facilitate the degradation of targeted proteins of interest (POIs). Notably, the clinical translation of PROTACs has gained substantial momentum, with many PROTAC candidates targeting various cancers currently undergoing clinical trials (Phase I-III). However, the rational design of high-efficacy PROTAC compounds remains a significant challenge. In this review, we presented a comprehensive overview of POI ligands, E3 ligands, and their interconnected linkers in PROTAC design, including their generation, structural optimization, and contribution to degradation efficiency and selectivity. Particularly, we analyzed the distinct preferences of various types of POI ligands (small molecule, nucleic acid, and peptide) toward specific targets. Furthermore, we emphasized the significant role of artificial intelligence technology in PROTAC design, including POI/E3 ligands discovery and linkers generation or optimization. We also summarized the applications and challenges of PROTACs in cancer therapy. Finally, we discussed the future development of PROTAC by combining multidisciplinary technologies and novel modalities for cancer therapy. Overall, this review aims to provide valuable insights for advancing PROTAC design strategies for cancer therapy.

The binding affinity of aptamers to targets has a crucial role in the pharmaceutical and biosensing effects. Despite diverse post-systematic evolution of ligands by exponential enrichment (post-SELEX) modifications explored in aptamer optimization, accurate prediction of high-affinity modification strategies remains challenging. Sclerostin, which antagonizes the Wnt signaling pathway, negatively regulates bone formation. Our screened sclerostin aptamer was previously shown to exert bone anabolic potential. In the current study, an interactive methodology involving the exchange of mutual information between experimental endeavors and machine learning was initially proposed to design a high-affinity post-SELEX modification strategy for aptamers. After four rounds of interactive training (a total of 422 modified aptamer-target affinity datasets with diverse modification types and sites), an antifcial intelligence model with high predictive accuracy with a correlation coefficient of 0.82 between the predicted and actual binding affinities was obtained. Notably, the machine learning-powered modified aptamer selected from this work exhibited 105-fold higher affinity (picomole level KD value) and a 3.2-folds greater Wnt-signal re-activation effect compared to naturally unmodified aptamers. This approach harnessed the power of machine learning to predict the most promising high-affinity modification strategy for aptamers.

Therapeutic antibody against sclerostin loop2 promoted bone formation in postmenopausal osteoporosis but caused severe cardiovascular events in clinical applications. The studies of atherosclerosis and aortic aneurysm in SOST ki .ApoE −/− mice and sost −/− . ApoE −/− mice collectively indicated the cardiovascular protective action of sclerostin. However, how sclerostin exerts cardiovascular protective action remains unclear. In this study, ApoER2 (LRP8) is notably identified as a novel transmembrane receptor for sclerostin in macrophages. Mechanistically, blockade of macrophagic sclerostin loop2‐ApoER2 interaction attenuates the suppressive effects of sclerostin on NF‐κB nuclear translocation, phosphorylation, and mRNA expression in macrophages, reduces the promotive effects of sclerostin on macrophage conversion to anti‐inflammatory phenotypes, and inhibits the preventive effects of sclerostin on atherosclerosis and aortic aneurysm in ApoE −/− mice. Together, macrophagic sclerostin loop2‐ApoER2 interaction is required by sclerostin to suppress inflammatory responses, atherosclerosis, and aortic aneurysm in ApoE −/− mice. Sclerostin plays a compensatory protective role in the cardiovascular system when ApoE is absent or mutated. Translationally, it provided critical pre‐clinical evidence regarding the prediction of cardiovascular risk populations (e.g. , APOE variants) for the marketed antibody against sclerostin loop2. Importantly, targeting sclerostin while preserving macrophagic sclerostin loop2‐ApoER2 interaction would offer the next generation of precise sclerostin inhibition strategy without cardiovascular safety concern, while promoting bone formation.

Sclerostin, encoded by the SOST gene, is a novel bone anabolic target for bone diseases. Humanized anti-sclerostin antibody, romosozumab, was approved for treatment of postmenopausal osteoporosis by the US Food and Drug Administration (FDA), but with a black-box warning on cardiovascular risk. The clinical data regarding cardiovascular events from various pre-marketing and post-marketing studies of romosozumab were inconsistent. Overall, the cardiovascular risk of sclerostin inhibition could not be excluded. The restriction of romosozumab in patients with cardiovascular disease history would be necessary. Moreover, genome-wide association study (GWAS) analyses of SOST variants revealed inconsistent results of the association between SOST variations and cardiovascular diseases. Further research incorporating larger sample sizes and functional analyses are necessary. In analyses of serum/tissue sclerostin levels in patients with cardiovascular diseases, the results were controversial but indicated an association between sclerostin and the presence/severity/outcomes of cardiovascular diseases. Nonclinical studies in rodents indicated the inhibitory effect of sclerostin on inflammation, aortic aneurysm, atherosclerosis, and vascular calcification. Sclerostin loop3 participated in the inhibitory effect of sclerostin on bone formation, while the cardiovascular protective effect of sclerostin was independent of sclerostin loop3. Macrophagic sclerostin loop2–apolipoprotein E receptor 2 (ApoER2) interaction participated in the inhibitory effect of sclerostin on inflammation in vitro. Sclerostin in human aortic smooth muscle cells participated in the reduction in calcium deposition. The role of sclerostin in cardiovascular system deserves further investigation.