Drug efficacy prediction remains a cornerstone of drug development and precision therapy. However, integrating heterogeneous biomedical data, including multi-omics profiles, pathological imaging, electronic health records, and pharmacokinetic-pharmacodynamic (PK/PD) time-series, faces three fundamental barriers, namely cross-domain distribution shifts between preclinical and clinical data, relational mismatches between isolated vector representations and biological networks, and feature heterogeneity across disparate modalities. To address these challenges, three AI paradigms have emerged, transfer learning for cross-domain alignment, graph neural networks for structured relational modeling, and Transformers for global cross-modal feature interaction. Importantly, these techniques form a many-to-many complementary system rather than a one-to-one correspondence, a key insight that this review explicitly formalizes. We further elaborate encoding workflows for PK/PD data to bridge static molecular signatures with dynamic in vivo exposure trajectories. Four graded clinical applications are outlined, including personalized monotherapy, combination optimization, drug repurposing, and preclinical-to-clinical evaluation of novel candidates. We also dissect persistent bottlenecks such as data harmonization, model interpretability, and prospective validation, and propose five actionable directions, namely privacy-preserving benchmarks, causally interpretable models, temporal dynamic frameworks, cross-domain generalization, and lightweight clinical tools. By integrating theoretical rationales, methodological synergies, and hierarchical translational scenarios, this review provides a unified roadmap to accelerate the clinical deployment of multimodal drug response prediction.

The gut-brain axis is increasingly recognized as a critical pathway in Parkinson's disease (PD), but the gut-targeted mechanisms underlying the therapeutic effects of herbal formulas remain poorly understood. In this study, an enterogenic PD-like mouse model was established by long-term rotenone gavage and treated with Duzhong Fang (DZF) to investigate whether DZF acts through intestinally retained components to influence gut-brain α-synuclein pathology. DZF markedly improved motor dysfunction, increased nigral tyrosine hydroxylase (TH)-positive cell density, alleviated colonic inflammation, and reduced α-Syn accumulation in the vagus nerve and brain, while colonic α-Syn showed a decreasing trend. Mechanistically, DZF downregulated colonic TPH1 and 5-HT3A expression, reshaped gut microbiota composition, and modulated fecal metabolites related to tryptophan metabolism and the 5-HT biosynthetic pathway. Molecular docking predicted potential interactions between TPH1 and gut-retained gingerol-related compounds, including 6-shogaol, 10-gingerdione, 10-gingerol, and 4-gingerol. These findings suggest that DZF ameliorates rotenone-induced PD-like phenotypes and may act by modulating the intestinal TPH1/5-HT-related pathway and α-Syn pathology along the gut-brain axis. This study supports a gut-targeted, neural-regulated therapeutic paradigm for traditional Chinese medicine and provides potential lead compounds for gut-directed anti-PD intervention.

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) main protease (Mpro) is a crucial therapeutic target for anti-coronavirus disease 2019 (COVID-19) drug development, as it is essential for viral replication. However, mutations within the active site have compromised the efficacy of current competitive inhibitors, prompting the exploration of alternative inhibition strategies. In this study, we systematically investigated the allosteric inhibition mechanism of SARS-CoV-2 Mpro by pelitinib and leveraged this insight for new inhibitor discovery. Through extensive molecular dynamics simulations, we showed that pelitinib exerts allosteric inhibition via the L141-S144-C145-H41 interaction network: it restricts the flexibility of L141 through CH-π interactions, transmits this effect to C145 via S144, stabilizes the hydrogen bond between C145 and H41, and thereby reduces the flexibility of the S3 helix (residues 40–60). This series of conformational changes induces the contraction of the Mpro catalytic pocket from ~1200 ų to ~800 ų, impairs substrate binding, and ultimately appears to impair Mpro activity. Based on this mechanism, we performed structure-based virtual screening and identified a novel compound (Cpd-1). Biological evaluations showed that Cpd-1 exhibits superior Mpro inhibitory activity compared to pelitinib, with negligible off-target binding to human EGFR and Myt1 kinase, low cytotoxicity (cell viability > 60% at 200 μM), and predicted inhibitory activity against clinically relevant Mpro-resistant mutants based on computational analysis. Our findings provide mechanistic insights into a key allosteric mechanism for Mpro inhibition but also provide a promising chemical scaffold for further development as an Mpro-targeting inhibitor.

Belzutifan and cabozantinib are oral drugs for advanced renal cell carcinoma, which display exposure-dependent toxicities. No validated method exists for simultaneously quantifying these drugs in human plasma. In this study, we developed and validated a high-performance liquid chromatography–ultraviolet method for simultaneously quantifying belzutifan and cabozantinib in human plasma. Plasma samples (50 μL) were processed via simple protein precipitation. Chromatographic separation on a C18 column was performed isocratically with 0.5% KH₂PO₄ (pH 4.5)/acetonitrile (47/53, v/v) at 1.0 mL/min, with UV detection at 230 nm. Performance characteristics (calibration, precision, accuracy, recovery, stability, selectivity, and interference by concomitant drugs) were systematically evaluated. Linear calibration ranges were 50–4,000 ng/mL for belzutifan and 25–4,000 ng/mL for cabozantinib with excellent linearity (r² ≥ 0.9998). Intra- and interday coefficients of variation were ≤ 5.41% for belzutifan and ≤ 12.25% for cabozantinib. Both analytes were stable under bench-top, short-term, long-term, and freeze–thaw conditions. No interference from endogenous plasma components or six commonly co-administered medications was observed. The newly developed assay for the simultaneous quantification of belzutifan and cabozantinib in human plasma is rapid, requires minimal sample volumes, and demonstrates robust analytical performance.

The nicotinic acetylcholine receptor is an ion channel receptor that is highly expressed in the brain and can cooperatively propagate several signals with other ion channel receptors in neurons. However, little is known about how nicotinic acetylcholine and kainate receptors, both of which enable cation influx into neurons, cooperatively regulate immediate early genes. In this study, we administrated both ligands, nicotine and kainate, into mice and investigated alterations of the immediate early genes activity-regulated cytoskeleton-associated protein (Arc), junB, c-fos, and egr1 in the cerebral cortex, hippocampus, olfactory bulb, and cerebellum. Kainate alone showed a tendency to induce all four immediate early genes in all four brain regions, whereas nicotine alone did not induce them strongly. However, pretreatment of nicotine before administration of kainate showed differential effects of nicotine on mRNA expression across different brain regions. Nicotine selectively inhibited kainate-induced Arc and junB mRNA expression in the cerebral cortex. These findings suggest that nicotine influences specific brain regions via the induction of certain immediate early genes.

Overdose incidence has increased among young individuals, particularly involving psychotropic and over-the-counter (OTC) drugs. However, there is limited information regarding how the physical form of medications—such as tablet size or thickness—affects the ability to take large doses. This study aims to determine which tablet characteristics (size, thickness, and volume) can discourage overdose by using sensory testing methods. The diameters and thicknesses of tablets commonly used in overdosing were measured. Additionally, dummy tablets were three-dimensionally (3D) printed, after which sensory tests were conducted to evaluate how many tablets could be comfortably consumed in large quantities, as well as the ease of consumption, rated on a 5-point scale. The respondents were students (average age: 21.3 ± 2.1 years, n = 253). Results indicate that OTC drugs typically have larger and thicker tablets compared with prescription drugs. Sensory tests of 3D-printed dummy tablets found that smaller tablets are easier to ingest in large numbers, whereas larger and thicker tablets are harder to consume in bulk. For tablets with the same diameter, a 5-mm thickness tends to make overdose more difficult compared with a 2-mm thickness. Importantly, tablets with a combined diameter and thickness of 13 mm or more and a volume of at least 190 mm³ are less likely to be overdosed. These findings can assist pharmacists in selecting appropriate dosage forms for patients at risk of overdose and provide pharmaceutical companies with guidance during product development.

Urinary retention is a known adverse effect of opioids, but limited information is available on how this adverse event varies among individual opioids. This study evaluated disproportionality signals between major opioids and urinary retention using the Japanese Adverse Drug Event Report database (JADER). We performed a disproportionality analysis of reports submitted between April 2004 and October 2025. Signals were assessed using reporting odds ratios (RORs), 95% confidence intervals (CIs), and the Information Component (IC). Adverse events were identified using the Medical Dictionary for Regulatory Activities/Japanese version (MedDRA/J) term "urinary retention." A sensitivity analysis was conducted using a broader definition of urinary dysfunction that also included "Dysuria." Time to onset was calculated from the start of opioid administration to event onset, and daily doses were converted to oral morphine equivalents. Significant signals were detected for morphine, oxycodone, tramadol, and codeine. Transdermal fentanyl was the only opioid without a detectable signal, even under the broader definition of urinary dysfunction. Among opioids with detected signals and available dose information, urinary retention was reported even at relatively low morphine-equivalent doses, suggesting the need for careful monitoring at the start of opioid therapy. The absence of a signal for transdermal fentanyl may reflect a lower reporting frequency in JADER, but not a lower clinical risk, of urinary retention. Given the limitations of spontaneous reporting systems, these findings are hypothesis-generating and are insufficient to guide opioid selection or switching.

Alzheimer's disease (AD) still lacks scalable oral strategies for early intervention. This article examines acitretin, a systemic retinoid used for severe psoriasis, as a clinically characterized scaffold for central nervous system (CNS) medicinal chemistry—not as a parent drug for direct repurposing or as a predefined candidate. Retinoic acid receptor/retinoid X receptor signaling can induce a disintegrin and metalloprotease 10 (ADAM10), the principal neuronal α-secretase of amyloid precursor protein, thereby encouraging non-amyloidogenic cleavage and soluble amyloid precursor protein-α (sAPPα) production. Supporting evidence includes mechanistic studies, murine blood-brain barrier penetration, functional observations in an amyloid mouse model, and a small randomized human study showing a short-term increase in cerebrospinal fluid sAPPα. These findings provide a human biomarker anchor but do not establish adequate unbound brain exposure, durable target engagement, clinical efficacy, or a safe chronic therapeutic window. The resulting concept is an evidence-gated scaffold-redesign strategy: future work may test whether retinoid-ADAM10-APP signaling can be retained while systemic retinoid burden and metabolite persistence are reduced. Further progression would require convergence of unbound CNS exposure, bounded sAPPα modulation, and acceptable safety within the same concentration range.

Hypercholesterolemia, and particularly elevated low-density lipoprotein cholesterol (LDL-C), is a major causal factor in atherosclerotic cardiovascular disease (ASCVD) and contributes substantially to coronary heart disease, ischemic stroke, and peripheral artery disease. Statins remain the cornerstone of lipid-lowering therapy, while ezetimibe and injectable proprotein convertase subtilisin/kexin 9 (PCSK9) targeted agents are added when further LDL-C reduction is required. However, concerns about injections, cost, accessibility, and treatment inertia have limited the use of existing PCSK9 therapies, creating a need for potent oral alternatives. Enlicitide, approved by the US Food and Drug Administration in July 2026, is the first approved oral PCSK9 inhibitor. This macrocyclic peptide blocks the interaction between circulating PCSK9 and hepatic LDL receptors, thereby increasing receptor recycling and LDL-C clearance. In the phase III CORALreef Lipids and CORALreef HeFH trials, once-daily enlicitide reduced LDL-C by approximately 60%, and the effect was maintained for up to 52 weeks. Adverse event rates were generally similar to those observed with a placebo. Nevertheless, those trials primarily evaluated LDL-C reduction rather than cardiovascular outcomes, and their duration was insufficient to establish long-term safety. The ongoing CORALreef Outcomes trial will determine whether enlicitide-mediated LDL-C reduction translates into fewer major cardiovascular events and will help define its long-term clinical role.

Narcolepsy type 1 (NT1) is a chronic neurological disorder caused by selective loss of orexin-producing neurons, resulting in unstable sleep-wake regulation and characteristic symptoms including excessive daytime sleepiness (EDS), cataplexy, sleep paralysis, hypnagogic or hypnopompic hallucinations, and disrupted nocturnal sleep. Current treatments mainly provide partial symptomatic relief and do not correct the underlying orexin deficiency, highlighting the need for mechanism-based therapies. Oveporexton (TAK-861), an orally administered selective orexin receptor 2 (OX2R) agonist developed by Takeda Pharmaceuticals, represents the first approved therapy targeting the orexin pathway in NT1. By activating preserved OX2R signaling, oveporexton restores wakefulness regulation and alleviates multiple core symptoms of NT1. Its approval was aided by two global phase 3 trials which demonstrated significant improvements in wakefulness and cataplexy compared with placebo. Although its long-term safety, accessibility, and real-world efficacy need to be evaluated further, oveporexton represents a major advance toward disease mechanism-based treatment for NT1.