AAPS Journal最新文献

Stepwise covariate modeling (SCM) has a high computational burden and can select the wrong covariates. Machine learning (ML) has been proposed as a screening tool to improve the efficiency of covariate selection, but little is known about how to apply ML on actual clinical data. First, we simulated datasets based on clinical data to compare the performance of various ML and traditional pharmacometrics (PMX) techniques with and without accounting for highly-correlated covariates. This simulation step identified the ML algorithm and the number of top covariates to select when using the actual clinical data. A previously developed desipramine population-pharmacokinetic model was used to simulate virtual subjects. Fifteen covariates were considered with four having an effect included. Based on the F1 score (an accuracy measure), ridge regression was the most accurate ML technique on 200 simulated datasets (F1 score = 0.475 ± 0.231), a performance which almost doubled when highly-correlated covariates were accounted for (F1 score = 0.860 ± 0.158). These performances were better than forwards selection with SCM (F1 score = 0.251 ± 0.274 and 0.499 ± 0.381 without/with correlations respectively). In terms of computational cost, ridge regression (0.42 ± 0.07 seconds/simulated dataset, 1 thread) was ~20,000 times faster than SCM (2.30 ± 2.29 hours, 15 threads). On the clinical dataset, prescreening with the selected ML algorithm reduced SCM runtime by 42.86% (from 1.75 to 1.00 days) and produced the same final model as SCM only. In conclusion, we have demonstrated that accounting for highly-correlated covariates improves ML prescreening accuracy. The choice of ML method and the proportion of important covariates (unknown a priori) can be guided by simulations.

The aim of this study was to investigate the in-use compatibility of eight commercially available closed system transfer device brands (CSTDs) with a formulated model antibody drug conjugate (ADC). Overall, in-use simulated dosing preparation applying the CSTD systems investigated raised concerns for several product quality attributes. The incompatibilities observed were mainly associated with increased visible and subvisible particles formation as well as significant changes in holdup volumes. Visible and subvisible particles contained heterogeneous mixtures of particle classes, with the majority of subvisible particles associated with silicone oil leaching from CSTD systems during simulated dose preparation upon contact with the ADC formulation. These observations demonstrate that CSTD use may adversely impact product quality and delivered dose which could potentially lead to safety and efficacy concerns during administration. Other product quality attributes measured including turbidity, color, ADC recovery, and purity by size exclusion HPLC, did not show relevant changes. It is therefore strongly recommended to test and screen the compatibility of CSTDs with the respective ADC, in a representative in-use simulated administration setting, during early CMC development, i.e., well before the start of clinical studies, to include information about compatibility and to ensure that the CSTD listed in the manuals of preparation for clinical handling has been thoroughly assessed before human use.

Subcutaneous (SC) administration of therapeutic proteins is perceived to pose higher risk of immunogenicity when compared with intravenous (IV) route of administration (RoA). However, systematic evaluations of clinical data to support this claim are lacking. This meta-analysis was conducted to compare the immunogenicity of the same therapeutic protein by IV and SC RoA. Anti-drug antibody (ADA) data and controlling variables for 7 therapeutic proteins administered by both IV and SC routes across 48 treatment groups were analyzed. RoA was the primary independent variable of interest while therapeutic protein, patient population, adjusted dose, and number of ADA samples were controlling variables. Analysis of variance was used to compare the ADA incidence between IV and SC RoA, while accounting for controlling variables and potential interactions. Subsequently, 10 additional therapeutic proteins with ADA data published for both IV and SC administration were added to the above 7 therapeutic proteins and were evaluated for ADA incidence. RoA had no statistically significant effect on ADA incidence for the initial dataset of 7 therapeutic proteins (p = 0.55). The only variable with a significant effect on ADA incidence was the therapeutic protein. None of the other controlling variables, including their interactions with RoA, was significant. When all data from the 17 therapeutic proteins were pooled, there was no statistically significant effect of RoA on ADA incidence (p = 0.81). In conclusion, there is no significant difference in ADA incidence between the IV and SC RoA, based on analysis of clinical ADA data from 17 therapeutic proteins.

Drug clearance in obese subjects varies widely among different drugs and across subjects with different severity of obesity. This study investigates correlations between plasma clearance (CLp) and drug- and patient-related characteristics in obese subjects, and evaluates the systematic accuracy of common weight-based dosing methods. A physiologically-based pharmacokinetic (PBPK) modeling approach that uses recent information on obesity-related changes in physiology was used to simulate CLp for a normal-weight subject (body mass index [BMI] = 20) and subjects with various severities of obesity (BMI 25-60) for hypothetical hepatically cleared drugs with a wide range of properties. Influential variables for CLp change were investigated. For each drug and obese subject, the exponent that yields perfect allometric scaling of CLp from normal-weight subjects was assessed. Among all variables, BMI and relative changes in enzyme activity resulting from obesity proved highly correlated with obesity-related CLp changes. Drugs bound to α1-acid glycoprotein (AAG) had lower CLp changes compared to drugs bound to human serum albumin (HSA). Lower extraction ratios (ER) corresponded to higher CLp changes compared to higher ER. The allometric exponent for perfect scaling ranged from -3.84 to 3.34 illustrating that none of the scaling methods performed well in all situations. While all three dosing methods are generally systematically accurate for drugs with unchanged or up to 50% increased enzyme activity in subjects with a BMI below 30 kg/m², in any of the other cases, information on the different drug properties and severity of obesity is required to select an appropriate dosing method for individuals with obesity.

The standard errors (SE) of the maximum likelihood estimates (MLE) of the population parameter vector in nonlinear mixed effect models (NLMEM) are usually estimated using the inverse of the Fisher information matrix (FIM). However, at a finite distance, i.e. far from the asymptotic, the FIM can underestimate the SE of NLMEM parameters. Alternatively, the standard deviation of the posterior distribution, obtained in Stan via the Hamiltonian Monte Carlo algorithm, has been shown to be a proxy for the SE, since, under some regularity conditions on the prior, the limiting distributions of the MLE and of the maximum a posterior estimator in a Bayesian framework are equivalent. In this work, we develop a similar method using the Metropolis-Hastings (MH) algorithm in parallel to the stochastic approximation expectation maximisation (SAEM) algorithm, implemented in the saemix R package. We assess this method on different simulation scenarios and data from a real case study, comparing it to other SE computation methods. The simulation study shows that our method improves the results obtained with frequentist methods at finite distance. However, it performed poorly in a scenario with the high variability and correlations observed in the real case study, stressing the need for calibration.

Hepatocytes are one of the most physiologically relevant in vitro liver systems for human translation of clearance and drug-drug interactions (DDI). However, the cell membranes of hepatocytes can limit the entry of certain compounds into the cells for metabolism and DDI. Passive permeability through hepatocytes can be different in vitro and in vivo, which complicates the human translation. Permeabilized hepatocytes offer a useful tool to probe mechanistic understanding of permeability-limited metabolism and DDI. Incubation with saponin of 0.01% at 0.5 million cells/mL and 0.05% at 5 million cells/mL for 5 min at 37°C completely permeabilized the plasma membrane of hepatocytes, while leaving the membranes of subcellular organelles intact. Permeabilized hepatocytes maintained similar enzymatic activity as intact unpermeabilized hepatocytes and can be stored at -80°C for at least 7 months. This approach reduces costs by preserving leftover hepatocytes. The relatively low levels of saponin in permeabilized hepatocytes had no significant impact on the enzymatic activity. As the cytosolic contents leak out from permeabilized hepatocytes, cofactors need to be added to enable metabolic reactions. Cytosolic enzymes will no longer be present if the media are removed after cells are permeabilized. Hence permeabilized hepatocytes with and without media removal may potentially enable reaction phenotyping of cytosolic enzymes. Although permeabilized hepatocytes work similarly as human liver microsomes and S9 fractions experimentally requiring addition of cofactors, they behave more like hepatocytes maintaining enzymatic activities for over 4 h. Permeabilized hepatocytes are a great addition to the drug metabolism toolbox to provide mechanistic insights.

Selective chemical inhibitors are critical for reaction phenotyping to identify drug-metabolizing enzymes that are involved in the elimination of drug candidates. Although relatively selective inhibitors are available for the major cytochrome P450 enzymes (CYP), they are quite limited for the less common CYPs and non-CYPs. To address this gap, we developed a multiplexed high throughput screening (HTS) assay using 20 substrate reactions of multiple enzymes to simultaneously monitor the inhibition of enzymes in a 384-well format. Four 384-well assay plates can be run at the same time to maximize throughput. This is the first multiplexed HTS assay for drug-metabolizing enzymes reported. The HTS assay is technologically enabled with state-of-the-art robotic systems and highly sensitive modern LC-MS/MS instrumentation. Virtual screening is utilized to identify inhibitors for HTS based on known inhibitors and enzyme structures. Screening of ~4600 compounds generated many hits for many drug-metabolizing enzymes including the two time-dependent and selective aldehyde oxidase inhibitors, erlotinib and dibenzothiophene. The hit rate is much higher than that for the traditional HTS for biological targets due to the promiscuous nature of the drug-metabolizing enzymes and the biased compound selection process. Future efforts will focus on using this method to identify selective inhibitors for enzymes that do not currently have quality hits and thoroughly characterizing the newly identified selective inhibitors from our screen. We encourage colleagues from other organizations to explore their proprietary libraries using a similar approach to identify better inhibitors that can be used across the industry.