
Staphylococcus aureus (S. aureus) is the leading cause of skin infections worldwide. Rising antimicrobial resistance limits the effectiveness of current treatments, highlighting the need for new therapeutic strategies. In this context, Rifabutin (RFB), could be repurposed for skin infections based on its strong anti-staphylococcal activity. However, its low oral bioavailability, and the rapid capacity of bacteria to develop drug tolerance, makes necessary to find novel platforms to ensure antimicrobial efficacy.This work proposes the development of Nano-in-Gel drug delivery systems, integrating RFB-loaded nanostructured lipid carriers (NLCs) into gelatin hydrogels for topical administration.These composite materials combine the controlled release capacity of hydrogels with NLCs’ ability to maximize RFB clinical potential. RFB-loaded NLCs were prepared and characterized in terms of physicochemical properties and cytocompatibility in human dermal fibroblasts (HDF). Moreover, their antibacterial activity against two different S. aureus strains-including mutants generated after RFB exposure was evaluated using a combination of conventional quantification methods and droplet digital PCR (ddPCR).Following the incorporation of NLCs into the gels, the mechanical properties of the Nano-in-Gel systems, and their ability to modulate nanoparticle release were assessed. Results demonstrated that the RFB-loaded NLCs possessed optimal physicochemical characteristics and high biocompatibility. Furthermore, the nanocarriers exhibited enhanced anti-staphylococcal activity compared to the free drug and successfully overcame RFB-resistance in mutant strains. The Nano-in-Gel platforms showed excellent mechanical performance for topical application and achieved a controlled NLCs release profile. These findings suggest that the Nano-in-Gel systems constitute a promising strategy for the advanced management of S. aureus skin infections.
Background Drug-induced immune thrombocytopenia (DITP) is a rare but potentially life-threatening adverse drug reaction, with evidence largely limited to case reports. Piperacillin-tazobactam (PTZ) appears to carry a relatively higher risk, particularly in older patients. This study aims to elucidate the clinical characteristics and risk factors associated with piperacillin-tazobactam (PTZ)-induced thrombocytopenia in older patients, to develop and validate a risk prediction model. Methods In this retrospective cohort study, we screened 5454 older inpatients (≥60 years) treated with PTZ for pulmonary infections at Xiangya Hospital (2020–2022) for model development, with a separate 2024 cohort for temporal validation. After strict exclusion, 258 patients entered the model development phase (202 non-cases and 56 thrombocytopenia cases). Among the 56 cases, 40 were confirmed as PTZ-related and analyzed by multivariate logistic regression to identify independent risk factors, leaving 242 patients in the final development cohort. Model performance was evaluated using the separate temporal validation cohort (n = 151) via receiver operating characteristic (ROC) curve analysis. Results PTZ-related thrombocytopenia occurred in 16.5% (40/242) of the development cohort. The independent risk factors identified included: baseline platelet count ≤150×10⁹/L (OR=8.065, 95% CI: 3.510–18.530), serum creatinine ≥80 μmol/L (OR=3.925, 95% CI: 1.752–8.797), aspartate aminotransferase ≥30 U/L (OR=3.154, 95% CI: 1.415–7.028), and hemoglobin ≤90 g/L (OR=3.172, 95% CI: 1.319–7.629). The prediction model achieved an area under the curve (AUC) of 0.819 (95% CI: 0.753–0.885) in the development set and 0.713 (95% CI: 0.568–0.858) in the temporal validation set. Conclusion A predictive model incorporating baseline platelet count, serum creatinine, aspartate aminotransferase, and hemoglobin demonstrated good discriminative performance for PTZ-induced thrombocytopenia in older patients, supporting its potential utility for individualized risk stratification.
OBJECTIVES Process digitalization through high-fidelity modelling is increasingly used to support Quality by Design (QbD) and Operational Excellence in pharmaceutical manufacturing. In this study, an industrial data-calibrated mechanistic modelling framework was developed for an industrial pharmaceutical batch fluid bed drying process, integrating a mechanistic model calibrated using real industrial data. The objective was not only to predict final moisture content, but also to characterise process dynamics and the effects of key process parameters, transforming the validated model into a decision-support tool for defining robust operating regions within a QbD framework. METHODS The mechanistic core incorporated two-phase fluidisation hydrodynamics, granular moisture diffusion and convective heat transfer, while model parameters were calibrated using data from eight full-scale industrial batches and evaluated using four independent batches. The validated framework enabled virtual experimentation across inlet air temperature, airflow rate and drying time, together with Monte Carlo-based uncertainty propagation. RESULTS Validation against four independent batches showed agreement for final residual moisture within the experimentally represented range, with a root mean square error of approximately 0.3% on a wet basis. Simulation results revealed nonlinear interactions among critical process parameters and identified operating regions satisfying residual moisture criteria while maintaining favourable relative energy-performance conditions. Monte Carlo-based uncertainty propagation showed that central operating regions exhibited lower sensitivity to disturbances, whereas boundary conditions amplified variability in critical quality attributes. CONCLUSIONS Overall, the framework extends mechanistic process simulation towards probabilistic design-space construction and risk-informed decision support in pharmaceutical manufacturing.
The role of oxaliplatin-based hyperthermic intraperitoneal chemotherapy (HIPEC) in colorectal peritoneal metastases remains debated after the PRODIGE 7 trial showed no overall survival benefit for the widely used 30-min high-dose regimen. This has raised interest in whether alternative oxaliplatin schedules, particularly lower-dose and longer-duration regimens, may improve tumor exposure and pharmacodynamic (PD) activity. In this study, we present a mechanistic framework integrating pharmacokinetics with patient-derived organoid (PK–PDO) data to compare two established HIPEC regimens for colorectal peritoneal metastases: 460 mg/m2 over 30 min and 200 mg/m2 over 120 min. A population-based PK model was coupled with a spatiotemporal transport model to simulate oxaliplatin distribution within tumor nodules up to 2 mm. Exposure metrics were integrated with published PDO sensitivity data, including IC50 and Hill coefficients, to evaluate response under dynamic concentration profiles.The long-duration regimen increased tumor-to-plasma ultrafiltrate exposure. However, overall PD efficacy at the population level, quantified using a Hill-type exposure–response model, was not significantly different between regimens. Substantial interpatient variability was predicted, with some tumors favoring one regimen over the other depending on drug sensitivity. Threshold analysis indicated that the 120-min regimen is advantageous only when its IC50 is at least fivefold lower than that of the 30-min regimen. In conclusion, despite differences in PK profiles, both regimens yield comparable overall PD effects. These findings suggest that switching oxaliplatin HIPEC protocols may not consistently improve treatment efficacy and that outcomes depend on the sensitivity of tumor nodules to treatment duration, as reflected by PDOs.
Arginine deiminase (ADI) is a promising therapeutic enzyme for targeting arginine-auxotrophic cancers, but its clinical application is limited by insufficient catalytic efficiency and poor stability. In this study, we developed a rational protein engineering strategy integrating folding free energy (ΔΔG) prediction and evolutionary conservation analysis to optimize ADI from Pseudomonas mendocina. A comprehensive site-saturation mutational scan was performed using the Pythia framework, followed by ConSurf-based filtering to exclude highly conserved residues, thereby efficiently identifying candidate beneficial mutations. Experimental evaluation showed that several single mutations enhanced enzymatic activity, and combinatorial optimization further revealed synergistic effects. Notably, the triple mutant K91L/N95E/Y346E exhibited the most significant improvement, achieving a 2.7-fold increase in catalytic efficiency (kcat/KM) compared to the wild type. In addition, this variant demonstrated enhanced thermal stability and improved plasma stability, with an extended half-life and higher residual activity. Importantly, in vitro cytotoxicity assays confirmed that the engineered mutant displayed superior antitumor activity against colorectal cancer cells. Overall, this study highlights the effectiveness of integrating thermodynamic and evolutionary constraints for enzyme engineering and provides a promising strategy for developing ADI-based anticancer therapeutics.