AAPS Journal最新文献

GT-14, identified as [(E)-4-((1-(1-methyl-1H-indol-2-yl) ethylidene)amino)phenol], is a novel inhibitor targeting the Gα_i2 protein, which is crucial in facilitating cell migration and invasion in prostate, ovarian, and breast cancer cells. therefore a valuable target for treating metastatic castration-resistant prostate cancer (mCRPC). In this study, GT-14's physicochemical properties, permeability, metabolic behavior, and tissue distribution were assessed. The results showed that GT-14 exhibited very slight aqueous solubility at room temperature (0.11 mg/mL) but was soluble in solvents including dimethyl sulfoxide and dimethyl acetamide, and sparingly or slightly soluble in several cosolvents. GT-14 exhibited a distinct pH-dependent solubility profile, being stable across a broad pH range (1.2-7.4) but degrading in strongly basic conditions. It exhibited high permeability (1.3 x 10^-5 cm/s) in Caco-2 cell culture models and therefore identified as a BCS II compound. Hepatic microsomal studies revealed that GT-14 underwent Phase I metabolism, with more than 90% remaining in 60 min incubation in rat liver microsomes. A stable co-solvent formulation was developed to enable intravenous administration for pharmacokinetic studies. Previous pharmacokinetic studies showed that GT-14 exhibited biphasic disposition with a terminal plasma elimination half-life of 268.07 minutes (> 4 hours). Tissue distribution analysis indicated the highest concentration of GT-14 in the prostate, followed by the kidneys, lungs, heart, and liver. Our study presents an early-stage preclinical drug development roadmap that integrates modern technologies for efficiency and success, using GT-14 as a model compound. It showed promising characteristics, reinforcing its potential as a new therapeutic agent for mCRPC.

Model-informed drug development (MIDD) plays an important role in pharmacometrics by leveraging mathematical models to optimize drug dosing strategies. Traditional methods such as nonlinear mixed effects modeling (NONMEM) have long been the gold standard in population pharmacokinetic (PPK) modeling. However, the development of artificial intelligence (AI) presents a potential improvement in predictive performance and computational efficiency. This study evaluates the effectiveness of AI-based MIDD methods for PPK analysis by comparing them against traditional nonlinear mixed-effects (NLME)-based methods (e.g., NONMEM). We tested five machine learning (ML) models, three deep learning (DL) models, and a neural ordinary differential equations (ODE) model on both simulated and real clinical datasets under different scenarios, assessing predictive performance with metrics such as root mean squared error (RMSE), mean absolute error (MAE), and coefficient of determination (R²). Simulated datasets with known ground truth were created using a two-compartment model, while the real clinical dataset included data from 1,770 patients pooled from multiple clinical trials. Results indicate that AI/ML models often outperform NONMEM, with variations in performance depending on model type and data characteristics. Neural ODE models showed good performance, providing strong performance and explainability with large datasets. These findings underscore the potential of AI/ML methodologies to complement or enhance traditional PPK modeling approaches in MIDD, highlighting their applicability in future pharmacometrics workflows.

Understanding a drug's plasma half-life is essential in guiding dosage regimens and optimizing therapeutic outcomes, particularly in the early stages of drug development. By using published pharmacokinetic data from Food Animal Residue Avoidance Databank, we collected 560 data points of plasma half-lives for different drugs in dogs following intravenous administration. The dataset was then preprocessed and the mean elimination half-life for each drug was selected in the final clean dataset for model training and testing. Five types of chemical descriptors and four types of supervised machine learning (ML) algorithms were employed to build ML-empowered Quantitative Structure-Activity Relationship (QSAR) models. Model performances were assessed by determination coefficient (R²) and root mean square error values. The results showed that the Deep Neural Networks model with all-combined descriptor type had the best performance with R² = 0.80 for the fivefold cross-validation set and R² = 0.57 for the testing set. Furthermore, the applicability domains of the well-trained models are shown via Williams plots. This study reports an ML-based QSAR tool in predicting the elimination half-lives of drugs in dogs based on only chemical structures. This approach can be used to support drug development in dogs and provides a basis for potential interspecies extrapolation.

While T-cell Bispecifics (TCBs) are promising molecules for cancer treatment, their clinical development remains challenging due to Cytokine Release Syndrome (CRS). There is currently no method to accurately predict the doses expected to trigger CRS from pre-clinical data, resulting in the selection of non optimal First-In-Human (FIH) doses, far from the doses expected to show clinical benefit, to address safety concerns. In this work, a retrospective analysis aiming to identify which in vitro assay (focusing on tumor cell killing and cytokine release) is predictive of CRS at the first administration of TCBs developed for oncology indications at Roche is presented. A concentration threshold of C_max 25-times the in vitro tumor cell killing EC50 from the most sensitive assay (or 96% of the maximum killing effect) has been identified and can be used to guide the selection of clinical dose associated with no CRS. Above this concentration threshold, any CRS without distinction from grades 1-3 have been observed. This work demonstrates that in vitro tumor cell killing assay can be used to determine a FIH for TCBs at which no CRS is expected (at least for TCBs similar to the one used in this analysis). It also suggests that CRS events are connected to the intended pharmacological effect (tumor cell killing). This work can guide the selection of TCB FIH clinical doses, allowing higher starting dose compared to the minimum anticipated biological effect approach, therefore increasing patients' benefit.

Micafungin is an important echinocandin for invasive fungal infections, but its pharmacokinetics (PK) in critically ill patients is highly variable. This study aimed to characterize micafungin population PK in intensive care unit (ICU) patients, identify clinical covariates influencing exposure, and evaluate practical PK/pharmacodynamic (PD) indices, steady-state area under the curve (AUCss) and trough concentration (Cmin), as alternatives to the traditional AUC/minimum inhibitory concentration (MIC) ratio for therapeutic drug monitoring (TDM). Sixty critically ill adults received 100 mg/day intravenous micafungin, with rich PK sampling on Days 0, 4, and 14. Data were analyzed using a two-compartment model with first-order elimination in Monolix, and covariates were identified via forward-backward stepwise selection. The typical clearance was 1.56 L/h with high interindividual variability (coefficient of variation 59.9%), and total bilirubin was a significant predictor of clearance. Among 30 evaluable patients, the AUC/MIC ratio was a poor predictor of microbiological eradication (receiver operating characteristic [ROC] AUC = 0.579), whereas AUCss (ROC AUC = 0.656) and Cmin (ROC AUC = 0.641) demonstrated better performance. Time-to-event analysis showed that achieving AUCss > 90.15 mg·h/L (p = 0.045) or Cmin > 1.22 mg/L (p = 0.036) was significantly associated with faster eradication. In this heterogeneous ICU cohort, AUCss and Cmin were associated with microbiological outcomes and may offer pragmatic, MIC-independent monitoring metrics. These exploratory findings warrant prospective validation but incorporating simplified, exposure-guided dosing strategies may improve micafungin optimization and clinical outcomes in critically ill populations.

Generic drugs offer cost-effective alternatives to brand-name medications while ensuring comparable safety and efficacy. However, narrow therapeutic index drugs (NTIDs), which require precise dosing due to narrow margins between therapeutic and toxic concentrations, present additional regulatory challenges. Concerns regarding the interchangeability of generic NTIDs are amplified by international variation in definitions, bioequivalence (BE) standards, and regulatory approaches. This systematic review compares NTID-related regulatory frameworks across major authorities to inform the development of the ICH M13C guideline and foster global harmonization of evaluation standards for generic NTIDs. A comprehensive comparative analysis was conducted of NTID-related regulatory frameworks in five ICH member countries (United States [US], European Union [EU], Japan, Canada, and South Korea), with Egypt, Jordan, and Saudi Arabia included as reference countries. Data were obtained from literature searches and official regulatory sources, focusing on NTID definitions, BE standards, and NTID lists. Marked regulatory divergence was observed. South Korea uniquely incorporates quantitative pharmacological and toxicological criteria into NTID definitions. The US employs the most stringent NTID BE standards, utilizing a fully replicated design, reference-scaled average bioequivalence (RSABE), and variability assessment. Only cyclosporine and tacrolimus are classified as NTIDs by all five core countries. Variability in NTID lists and evaluation criteria complicates global harmonization efforts. Achieving consistent evaluation and safe international use of generic NTIDs requires global alignment on definitions, BE criteria, and NTID lists. This review supports integrating real-world data into regulatory decision-making and advances the ICH M13C guideline.

This study aimed to investigate the dissolution profiles of ionizable drugs in biorelevant bicarbonate buffer (BCB) at the intermediate gastrointestinal pH level. For the pH maintenance tests, BCB was prepared by adjusting the pH and the ionic strength (I) of NaHCO₃/Na₂CO₃ solutions using HCl and NaCl (BCB: 5-20 mM, pH 3.0-5.0, I = 0.14 M). The floating lid method was used to prevent CO₂ loss. For the dissolution tests, febuxostat (FBX), dipyridamole (DPM), dantrolene Na (DNT Na), pioglitazone HCl (PIO HCl), and tosufloxacin tosylate monohydrate (TFLX TS) were employed. The dissolution profiles were measured at pH 4.5 (10 mM BCB, I = 0.14 M). Compendial citrate-phosphate buffer (CPB) and acetate buffer (ACB) were used for comparison. In the pH maintenance test, the pH change was ≤  + 0.11 for 2 h in all conditions. The dissolution rates of FBX and DPM were slower in BCB than in CPB and ACB. DNT Na showed slightly less supersaturation in CPB than in BCB and ACB. In contrast, PIO HCl showed markedly higher supersaturation in BCB than in CPB and ACB. TFLX TS showed higher and lower supersaturation in the absence and presence of Cl^-, respectively. The hemi-hydrochloride salt formed in the latter case. The dissolution profiles of ionizable drugs in BCB differed from those in CPB and ACB, especially in the case of the salt-form drugs with an acidic counterion. The floating lid method enables dissolution testing using BCB in the intermediate pH range.

Physiologically based pharmacokinetics (PBPK) models are increasingly being used in pediatric drug development and with the conduct of clinical studies in specific countries, the development of such models to describe both age and ethnicity related differences is a logical step forward. This study described the development, verification, and application of a Chinese pediatric PBPK (p-PBPK) model. Chinese pediatric physiological systems parameters and clinical data was derived from public databases and the literature, the PBPK model was assembled so that demographic and physiological outputs such as height, cardiac output, and liver size with age represented the Chinese pediatric population. The model was tested using two drugs predominately metabolized by CYP3A4 (fentanyl and midazolam), one dual CYP3A4/CYP2C9 substrate (ruxolitinib), two by other CYPs (efavirenz and theophylline), and two by renal elimination (ceftazidime and vancomycin). Overall, 79% of all pharmacokinetic parameters were predicted within 0.8 to 1.25-fold, and 100% within 0.67 to 1.5-fold of the observed data. The application of the Chinese p-PBPK model is demonstrated with two bridging scenarios, by investigating whether recommended dosing regimens for efavirenz and theophylline are suitable for Chinese pediatric subjects. Given the increased regulatory use of pediatric PBPK models in drug development, expanding these models to other ethnic groups is important. There is a need to further develop the current model across a wider range of drugs with different elimination pathways, to increase model confidence, this should involve academia, industry, model providers, and regulatory agencies.

The past decades have witnessed a growing interest in patient-centric sampling, including dried capillary blood microsampling. Despite its advancements, a key issue lies in interpreting dried capillary blood results against existing plasma-based reference ranges. To address this challenge, various methodologies to convert dried capillary blood concentrations to plasma equivalents have been proposed. This study systematically evaluates and compares different methodologies for converting dried capillary blood results to venous plasma results using clinical pharmacokinetic (PK) data of paracetamol and metabolites as a case study. Paired capillary volumetric absorptive microsampling (cVAMS) and venous plasma samples were collected from 55 patients. For each analyte, cVAMS results were converted using multiple approaches, including conversion based on the median (time-dependent) capillary-to-plasma ratio, Passing-Bablok regression analysis, linear mixed-effects modeling and the hematocrit (Hct). Performance was assessed by comparing the agreement between converted cVAMS concentrations and actual plasma results to pre-defined analytical and clinical acceptance criteria. All approaches, except Hct-based conversion, yielded acceptable results for all analytes, with minor variations in analytical performance. Estimated C_max and AUC_6h values (and corresponding 90% confidence intervals) calculated based on the converted cVAMS results were within bio-equivalence criteria for all conversion approaches, except Hct-based conversion for one analyte. Both from analytical and clinical perspectives, this study demonstrated the reliability of different approaches to convert capillary blood microsampling-based to plasma-based results. The framework utilized in this study may guide future microsampling (PK) studies aiming at switching from collecting samples from the reference matrix to an alternative capillary blood matrix.

Bevacizumab is a humanized monoclonal antibody (mAb) approved to treat various cancers in adults and was investigated in pediatric patients. While the drug development of small molecules in pediatrics has greatly benefited from the use of physiologically-based pharmacokinetic (PBPK) modeling for pharmacokinetic (PK) extrapolation, such application remains relatively limited in mAbs. In this study, our objective was to evaluate the applicability of PBPK modeling in characterizing the age-dependent PK of bevacizumab. A minimal PBPK model was developed incorporating bevacizumab-specific drug parameters, with age-dependent physiological changes such as tissue volume, blood and lymphatic flow, and endogenous immunoglobulin G (IgG) levels, and validated using observed PK data in 786 adult and 141 pediatric patients from 23 bevacizumab clinical studies. The final model was applied to predict the exposure of bevacizumab in pediatric patients ranging from six months to 18 years old. Clinically observed bevacizumab PK data in adults following single or multiple dosing of 5, 10, and 15 mg/kg were generally within the 95% model prediction intervals. In pediatrics, the individually simulated bevacizumab concentrations were consistent with the individual observed data, including the pediatric patients as young as six months old. Sensitivity analysis revealed that endogenous IgG concentration and neonatal Fc receptors abundance play critical roles in bevacizumab PK in children. Overall, the PBPK model successfully bridges the bevacizumab PK from adult to pediatric patients by incorporating age-dependent physiological changes. This work represents a significant step forward in advancing the application of PBPK modeling of mAbs in children.