Global Analysis of Models for Predicting Human Absorption: QSAR, In Vitro, and Preclinical Models

In vitro and in silico evaluation of fluorinated diphenylamine chalcone derivatives as potential antimalarial and anticancer agents

A series of novel diphenylamine fluorinated chalcone derivatives (B1-B10) were synthesized and characterized using 1H and 13C NMR, IR, and MS, and purity was determined using HPLC. The compounds were evaluated for their antimicrobial, antimalarial, and anticancer activities, with Chloramphenicol, Griseofulvin, and 5-Fluorouracil serving as standard reference drugs. Notably, B6 exhibited excellent antifungal activity, comparable to that of the standard drug Griseofulvin. Compounds B3 and B5 showed strong antimalarial effects against Plasmodium falciparum. Both B3 and B5 exhibit substantial cytotoxicity against HeLa cells, with IC50 values of 24.53 µg/ml for B5 and 32.42 µg/ml for B3. These results clearly demonstrate that both compounds outperform the standard drug 5-Fluorouracil, establishing their strong potential as effective alternatives in cancer therapy. Molecular docking studies revealed that B3 and B6 effectively interacted with the active site of Falcilysin, while B5 and B7 showed favourable binding to proteins 6GUE and 2 × 7 F. Molecular dynamics simulations confirmed the stability of B3 and B6 with P. falciparum, while B5 and B3 exhibited promising interactions with 6GUE and 2X7F. These results suggest that compounds B3 and B5 are potential lead candidates for developing novel antimicrobial, antimalarial, and anticancer therapies.

Explainable Machine Learning for ETR and Drug Chameleonicity

Explainable machine learning that identifies molecular "hot spots" for chameleonicity can guide rapid chemical design for oral absorption of beyond-rule-of-five (bRo5) drugs. Traditional in silico methods rely on computationally intensive 3D physics-based modeling or classical descriptors that do not fully explain bRo5 drug behavior. To address this, we introduced the EPSA-to-TPSA ratio (ETR) as a high-throughput measure of polarity reduction, generating data for thousands of macrocycles, PROTACs, and other bRo5s. Using this data set, we developed an explainable deep learning model to predict EPSA and locate polarity-reducing "hot spots" that influence chameleonicity. This first-of-its-kind interpretable model in the bRo5 3D domain guides chemical modifications before synthesis, helping chemists optimize physicochemical properties and design complex bRo5 drugs with improved oral bioavailability. Model insights validated by molecular dynamics enable robust, high-throughput predictions of bRo5 chameleonic behavior, building on Lipinski descriptors to establish new frameworks for complex drug design.

The In Vitro Pharmacokinetics of Medicinal Plants: A Review

Background: This review examines in vitro techniques for characterizing the pharmacokinetics of medicinal plants, focusing on their role in understanding absorption, distribution, metabolism, and excretion (ADME). The diverse bioactive compounds in medicinal plants highlight the need for robust pharmacokinetic evaluations to ensure their safety and efficacy. Objectives: The objectives were to identify and analyze in vitro techniques applied to medicinal plants' pharmacokinetics, addressing a gap in the literature. Methods: Studies were included based on predefined eligibility criteria: in vitro pharmacokinetic studies involving medicinal plants, focusing on ADME stages. Ex vivo, in vivo, and in silico studies were excluded, along with reviews. Data were collected from the PubMed, Web of Science, and Scopus databases in June 2024 using Health Sciences Descriptors (DeCS) and their MeSH synonyms. The data extracted included study location, plant species, bioactive compounds, in vitro protocols, and ADME characteristics. Results: The review included 33 studies, with most focusing on metabolism (60%), absorption (25%), or a combination of ADME aspects. Techniques like Caco-2 cells, human liver microsomes, and simulated gastric and intestinal fluids were widely used. Conclusions: The findings highlight methodological heterogeneity, including variability in extract preparation, compound concentrations, and experimental conditions, which limits the comparability and clinical applicability of results. Key limitations include the lack of standardized protocols and physiological relevance in in vitro models, underscoring the need for multidisciplinary approaches and integration with in vivo studies.

AURA: Accelerating drug discovery with accuracy, utility, and rank-order assessment for data-driven decision making

Biopharmaceutical companies generate a wealth of data, ranging from in silico physicochemical properties and machine learning models to both low and high-throughput in vitro assays and in vivo studies. To effectively harnesses this extensive data, we introduce a statistical methodology facilitated by Accuracy, Utility, and Rank Order Assessment (AURA), which combines basic statistical analyses with dynamic data visualizations to evaluate endpoint effectiveness in predicting intestinal absorption. We demonstrated that various physicochemical properties uniquely influence intestinal absorption on a project-specific basis, considering factors like intestinal efflux, passive permeability, and clearance. Projects within both the "Rule of 5" (Ro5) and beyond "Rule of 5" (bRo5) space present unique absorption challenges, emphasizing the need for tailored optimization strategies over one-size-fits-all approaches. This is corroborated by the improved accuracy of project-specific correlations over global models. The differences in correlations between and within project teams-due to their unique chemical spaces-highlight how complex and nuanced the prediction of intestinal absorption can be. Here, we implement a standardized methodology, AURA, that any organization can incorporate into their workflow to enhance early-stage drug optimization. By automating analytics, integrating diverse data types, and offering flexible visualizations, AURA enables cross-functional teams to make data-driven decisions, optimize workflows, and enhance research efficiency.

Application of Mechanistic Multiparameter Optimization and Large-Scale In Vitro to In Vivo Pharmacokinetics Correlations to Small-Molecule Therapeutic Projects

Computational chemistry and machine learning are used in drug discovery to predict the target-specific and pharmacokinetic properties of molecules. Multiparameter optimization (MPO) functions are used to summarize multiple properties into a single score, aiding compound prioritization. However, over-reliance on subjective MPO functions risks reinforcing human bias. Mechanistic modeling approaches based on physiological relevance can be adapted to meet different potential key objectives of the project (e.g., minimizing dose, maximizing safety margins, and/or minimizing drug-drug interaction risk) while retaining the same underlying model structure. The current work incorporates recent approaches to predict in vivo pharmacokinetic (PK) properties and validates in vitro to in vivo correlation analysis to support mechanistic PK MPO. Examples of use and impact in small-molecule drug discovery projects are provided. Overall, the mechanistic MPO identifies 83% of the compounds considered as short-listed for clinical experiments in the top second percentile, and 100% in the top 10th percentile, resulting in an area under the receiver operating characteristic curve (AUCROC) > 0.95. In addition, the MPO score successfully recapitulates the chronological progression of the optimization process across different scaffolds. Finally, the MPO scores for compounds characterized in pharmacokinetics experiments are markedly higher compared with the rest of the compounds being synthesized, highlighting the potential of this tool to reduce the reliance on in vivo testing for compound screening.

How predictive are isolated perfused liver data of in vivo hepatic clearance? A meta-analysis of isolated perfused rat liver data

Isolated perfused rat liver (IPRL) experiments have been used to answer clearance-related questions, including evaluating the impact of pathological and physiological processes on hepatic clearance (CLH). However, to date, IPRL data has not been evaluated for in vivo CLH prediction accuracy.In addition to a detailed overview of available IPRL literature, we present an in-depth analysis of the performance of IPRL in CLH prediction.While the entire dataset poorly predicted CLH (GAFE = 3.2; 64% within 3-fold), IPRL conducted under optimal experimental conditions, such as in the presence of plasma proteins and with a perfusion rate within 2-fold of physiological liver blood flow and corrected for unbound fraction in the presence of red blood cells, can accurately predict rat CLH (GAFE = 2.0; 78% within 3-fold). Careful consideration of experimental conditions is needed to allow proper data analysis.Further, isolated perfused liver experiments in other species, including human livers, may allow us to address the current in vitro-in vivo disconnects of hepatic metabolic clearance and improve our methodology for CLH predictions.

High-Throughput SFC-MS/MS Method to Measure EPSA and Predict Human Permeability

Permeability is a key factor driving the absorption of orally administered drugs. In early discovery, the efficient evaluation of permeability, particularly for compounds violating Lipinski's Rule of 5, remains challenging. Addressing this, we established a high-throughput method to measure the experimental polar surface area (HT-EPSA) as an in vitro surrogate to measure permeability. Compared to earlier methods, HT-EPSA significantly reduces data acquisition time with enhanced sensitivity, selectivity, and data quality. In the effort of translating EPSA to human in vitro and in vivo passive permeability, we demonstrated the application of EPSA for predicting Caco-2 cell and human intestinal permeability, showing improvements over topological polar surface area and the parallel artificial membrane permeability assay for rank-ordering permeability in a proteolysis targeting chimera case study. The HT-EPSA method is expected to be highly beneficial in guiding early stage compound rank-ordering, faster decision-making, and in predicting in vitro and/or in vivo human intestinal permeability.

Advances in Modeling Approaches for Oral Drug Delivery: Artificial Intelligence, Physiologically-Based Pharmacokinetics, and First-Principles Models

Oral drug absorption is the primary route for drug administration. However, this process hinges on multiple factors, including the drug's physicochemical properties, formulation characteristics, and gastrointestinal physiology. Given its intricacy and the exorbitant costs associated with experimentation, the trial-and-error method proves prohibitively expensive. Theoretical models have emerged as a cost-effective alternative by assimilating data from diverse experiments and theoretical considerations. These models fall into three categories: (i) data-driven models, encompassing classical pharmacokinetics, quantitative-structure models (QSAR), and machine/deep learning; (ii) mechanism-based models, which include quasi-equilibrium, steady-state, and physiologically-based pharmacokinetics models; and (iii) first principles models, including molecular dynamics and continuum models. This review provides an overview of recent modeling endeavors across these categories while evaluating their respective advantages and limitations. Additionally, a primer on partial differential equations and their numerical solutions is included in the appendix, recognizing their utility in modeling physiological systems despite their mathematical complexity limiting widespread application in this field.

Conquering the beyond Rule of Five Space with an Optimized High-Throughput Caco-2 Assay to Close Gaps in Absorption Prediction

Current drug development tends towards complex chemical molecules, referred to as "beyond rule of five" (bRo5) compounds, which often exhibit challenging physicochemical properties. Measuring Caco-2 permeability of those compounds is difficult due to technical limitations, including poor recovery and detection sensitivity. We implemented a novel assay, with optimized incubation and analytics, to measure permeability close to equilibrium. In this setup an appropriate characterization of permeability for bRo5 compounds is achievable. This equilibrated Caco-2 assay was verified with respect to data validity, compound recovery, and in vitro to in vivo correlation for human absorption. Compared to a standard assay, it demonstrated comparable performance in predicting the human fraction absorbed (fa) for reference compounds. The equilibrated assay also successfully characterized the permeability of more than 90% of the compounds analyzed, the majority of which were bRo5 (68%). These compounds could not be measured using the standard assay. Permeability and efflux ratio (ER) were highly predictive for in vivo absorption for a large set of internal bRo5 compounds. Reference cut-offs enabled the correct classification of high, moderate, and low absorption. This optimized equilibrated Caco-2 assay closes the gap for a high-throughput cellular permeability method in the bRo5 chemical space.

Beyond Rule of Five and PROTACs in Modern Drug Discovery: Polarity Reducers, Chameleonicity, and the Evolving Physicochemical Landscape

Developing orally bioavailable drugs demands an understanding of absorption in early drug development. Traditional methods and physicochemical properties optimize absorption for rule of five (Ro5) compounds; beyond rule of five (bRo5) drugs necessitate advanced tools like the experimental measure of exposed polarity (EPSA) and the AbbVie multiparametric score (AB-MPS). Analyzing AB-MPS and EPSA against ∼1000 compounds with human absorption data and ∼10,000 AbbVie tool compounds (∼1000 proteolysis targeting chimeras or PROTACs, ∼7000 Ro5s, and ∼2000 bRo5s) revealed new patterns of physicochemical trends. We introduced a high-throughput "polarity reduction" descriptor: ETR, the EPSA-to-topological polar surface area (TPSA) ratio, highlights unique bRo5 and PROTAC subsets for specialized drug design strategies for effective absorption. Our methods and guidelines refine drug design by providing innovative in vitro approaches, enhancing physicochemical property optimization, and enabling accurate predictions of intestinal absorption in the complex bRo5 domain.

The Comparison of Machine Learning and Mechanistic In Vitro-In Vivo Extrapolation Models for the Prediction of Human Intrinsic Clearance

Accurate prediction of human pharmacokinetics (PK) remains one of the key objectives of drug metabolism and PK (DMPK) scientists in drug discovery projects. This is typically performed by using in vitro-in vivo extrapolation (IVIVE) based on mechanistic PK models. In recent years, machine learning (ML), with its ability to harness patterns from previous outcomes to predict future events, has gained increased popularity in application to absorption, distribution, metabolism, and excretion (ADME) sciences. This study compares the performance of various ML and mechanistic models for the prediction of human IV clearance for a large (645) set of diverse compounds with literature human IV PK data, as well as measured relevant in vitro end points. ML models were built using multiple approaches for the descriptors: (1) calculated physical properties and structural descriptors based on chemical structure alone (classical QSAR/QSPR); (2) in vitro measured inputs only with no structure-based descriptors (ML IVIVE); and (3) in silico ML IVIVE using in silico model predictions for the in vitro inputs. For the mechanistic models, well-stirred and parallel-tube liver models were considered with and without the use of empirical scaling factors and with and without renal clearance. The best ML model for the prediction of in vivo human intrinsic clearance (CLint) was an in vitro ML IVIVE model using only six in vitro inputs with an average absolute fold error (AAFE) of 2.5. The best mechanistic model used the parallel-tube liver model, with empirical scaling factors resulting in an AAFE of 2.8. The corresponding mechanistic model with full in silico inputs achieved an AAFE of 3.3. These relative performances of the models were confirmed with the prediction of 16 Pfizer drug candidates that were not part of the original data set. Results show that ML IVIVE models are comparable to or superior to their best mechanistic counterparts. We also show that ML IVIVE models can be used to derive insights into factors for the improvement of mechanistic PK prediction.

Quantitative Translation of Substrate Intrinsic Clearance from Recombinant CYP1A1 to Humans

CYP1A1 is a cytochrome P450 family 1 enzyme that is mostly expressed in the extrahepatic tissues. To understand the CYP1A1 contribution to drug clearance in humans, we examined the in vitro-in vivo extrapolation (IVIVE) of intrinsic clearance (CLint) for a set of drugs that are in vitro CYP1A1 substrates. Despite being strong in vitro CYP1A1 substrates, 82% of drugs gave good IVIVE with predicted CLint within 2-3-fold of the observed values using human liver microsomes and hepatocytes, suggesting they were not in vivo CYP1A1 substrates due to the lack of extrahepatic contribution to CLint. Only three drugs (riluzole, melatonin and ramelteon) that are CYP1A2 substrates yielded significant underprediction of in vivo CLint up to 11-fold. The fold of CLint underprediction was linearly proportional to human recombinant CYP1A1 (rCYP1A1) CLint, indicating they were likely to be in vivo CYP1A1 substrates. Using these three substrates, a calibration curve can be developed to enable direct translation from in vitro rCYP1A1 CLint to in vivo extrahepatic contributions in humans. In vivo CYP1A1 substrates are planar and small, which is consistent with the structure of the active site. This is in contrast to the in vitro substrates, which include large and nonplanar molecules, suggesting rCYP1A1 is more accessible than what is in vivo. The impact of CYP1A1 on first-pass intestinal metabolism was also evaluated and shown to be minimal. This is the first study providing new insights on in vivo translation of CYP1A1 contributions to human clearance using in vitro rCYP1A1 data.

Challenges in Permeability Assessment for Oral Drug Product Development

Drug permeation across the intestinal epithelium is a prerequisite for successful oral drug delivery. The increased interest in oral administration of peptides, as well as poorly soluble and poorly permeable compounds such as drugs for targeted protein degradation, have made permeability a key parameter in oral drug product development. This review describes the various in vitro, in silico and in vivo methodologies that are applied to determine drug permeability in the human gastrointestinal tract and identifies how they are applied in the different stages of drug development. The various methods used to predict, estimate or measure permeability values, ranging from in silico and in vitro methods all the way to studies in animals and humans, are discussed with regard to their advantages, limitations and applications. A special focus is put on novel techniques such as computational approaches, gut-on-chip models and human tissue-based models, where significant progress has been made in the last few years. In addition, the impact of permeability estimations on PK predictions in PBPK modeling, the degree to which excipients can affect drug permeability in clinical studies and the requirements for colonic drug absorption are addressed.

Physiologically-Based Pharmacokinetic Modeling of PAXLOVID™ with First-Order Absorption Kinetics

PURPOSE

PAXLOVID™ is nirmatrelvir tablets co-packaged with ritonavir tablets. Ritonavir is used as a pharmacokinetics (PK) enhancer to reduce metabolism and increase exposure of nirmatrelvir. This is the first disclosure of Paxlovid physiologically-based pharmacokinetic (PBPK) model.

METHODS

Nirmatrelvir PBPK model with first-order absorption kinetics was developed using in vitro, preclinical, and clinical data of nirmatrelvir in the presence and absence of ritonavir. Clearance and volume of distribution were derived from nirmatrelvir PK obtained using a spray-dried dispersion (SDD) formulation where it is considered to be dosed as an oral solution, and absorption is near complete. The fraction of nirmatrelvir metabolized by CYP3A was estimated based on in vitro and clinical ritonavir drug-drug interaction (DDI) data. First-order absorption parameters were established for both SDD and tablet formulation using clinical data. Nirmatrelvir PBPK model was verified with both single and multiple dose human PK data, as well as DDI studies. Simcyp® first-order ritonavir compound file was also verified with additional clinical data.

RESULTS

The nirmatrelvir PBPK model described the observed PK profiles of nirmatrelvir well with predicted AUC and Cmax values within ± 20% of the observed. The ritonavir model performed well resulting in predicted values within twofold of observed.

CONCLUSIONS

Paxlovid PBPK model developed in this study can be applied to predict PK changes in special populations, as well as model the effect of victim and perpetrator DDI. PBPK modeling continues to play a critical role in accelerating drug discovery and development of potential treatments for devastating diseases such as COVID-19. NCT05263895, NCT05129475, NCT05032950 and NCT05064800.

Today's drug discovery and the shadow of the rule of 5

INTRODUCTION

The rule of 5 developed by Lipinski et al., a landmark and prescient piece of scholarship, focused the minds of drug hunters by systematically characterizing the physical make-up of drug molecules for the first time, noting many sub-optimal compounds identified by high-throughput screening practices. Its profound influence on thinking and practices, whilst providing benefit, perhaps etched the guidelines too strongly in the minds of some drug hunters who applied the bounds too literally without understanding the implications of the underlying statistics.

AREAS COVERED

This opinion is based on recent key developments that take thinking, measurements, and standards beyond those first set out, particularly the influences of molecular weight and the understanding, measurement, and calculation of lipophilicity.

EXPERT OPINION

Techniques and technologies for physicochemical estimations set new standards. It is timely to celebrate the significance and influence of the rule of 5, whilst taking thinking to new levels with better characterizations. The shadow of the rule of 5 may be long, but it is not dark, as new measurements, predictions and principles emerge as guiding lights in the design and prioritization of higher-quality molecules redefining the meaning of beyond the rule of 5.

In Vitro-In Vivo Extrapolation and Scaling Factors for Clearance of Human and Preclinical Species with Liver Microsomes and Hepatocytes

In vitro-in vivo extrapolation ((IVIVE) and empirical scaling factors (SF) of human intrinsic clearance (CL_int) were developed using one of the largest dataset of 455 compounds with data from human liver microsomes (HLM) and human hepatocytes (HHEP). For extended clearance classification system (ECCS) class 2/4 compounds, linear SFs (SF_lin) are approximately 1, suggesting enzyme activities in HLM and HHEP are similar to those in vivo under physiological conditions. For ECCS class 1A/1B compounds, a unified set of SFs was developed for CL_int. These SFs contain both SF_lin and an exponential SF (SF_β) of fraction unbound in plasma (f_u,p). The unified SFs for class 1A/1B eliminate the need to identify the transporters involved prior to clearance prediction. The underlying mechanisms of these SFs are not entirely clear at this point, but they serve practical purposes to reduce biases and increase prediction accuracy. Similar SFs have also been developed for preclinical species. For HLM-HHEP disconnect (HLM > HHEP) ECCS class 2/4 compounds that are mainly metabolized by cytochrome P450s/FMO, HLM significantly overpredicted in vivo CL_int, while HHEP slightly underpredicted and geometric mean of HLM and HHEP slightly overpredicted in vivo CL_int. This observation is different than in rats, where rat liver microsomal CL_int correlates well with in vivo CL_int for compounds demonstrating permeability-limited metabolism. The good CL_int IVIVE developed using HLM and HHEP helps build confidence for prospective predictions of human clearance and supports the continued utilization of these assays to guide structure-activity relationships to improve metabolic stability.

N-Propargylic β-enaminones in breast cancer cells: Cytotoxicity, apoptosis, and cell cycle analyses

Breast cancer is one of the most common cancers worldwide and the discovery of new cytotoxic agents is needed. Enaminones are regarded to be a significant structural motif that is found in a variety of pharmacologically active compounds however the number of studies investigating the anticancer activities of N-propargylic β-enaminones (NPEs) is limited. Herein we investigated the potential cytotoxic and apoptotic effects of 23 different NPEs (1-23) on human breast cancer cells. Cytotoxicity was evaluated via MTT assay. Apoptotic cell death and cell cycle distributions were investigated by flow cytometry. CM-H2DCFDA dye was used to evaluate cellular ROS levels. Expression levels of Bcl-2, Bax, p21, and Cyclin D1 were measured by quantitative real-time PCR. ADME properties were calculated using the ADMET 2.0 tool. NPEs 4, 9, 16, and 21 showed selective cytotoxic activity against breast cancer cells with SI values >2. NPEs induced apoptosis and caused significant changes in Bcl-2 and Bax mRNA levels. The cell cycle was arrested at the G0/G1 phase and levels of p21 and Cyclin D1 were upregulated in both breast cancer cells. ROS levels were significantly increased by NPEs, suggesting that the cytotoxic and apoptotic effects of NPEs were mediated by ROS. ADME analysis revealed that NPEs showed favorable distributions in both breast cancer cell lines, meaning good lipophilicity values, low unfractionated values, and high bioavailability. Therefore, these potential anticancer compounds should be further validated by in vivo studies for their appropriate function in human health with a safety profile, and a comprehensive drug interaction study should be performed.

IDL-PPBopt: A Strategy for Prediction and Optimization of Human Plasma Protein Binding of Compounds via an Interpretable Deep Learning Method

The prediction and optimization of pharmacokinetic properties are essential in lead optimization. Traditional strategies mainly depend on the empirical chemical rules from medicinal chemists. However, with the rising amount of data, it is getting more difficult to manually extract useful medicinal chemistry knowledge. To this end, we introduced IDL-PPBopt, a computational strategy for predicting and optimizing the plasma protein binding (PPB) property based on an interpretable deep learning method. At first, a curated PPB data set was used to construct an interpretable deep learning model, which showed excellent predictive performance with a root mean squared error of 0.112 for the entire test set. Then, we designed a detection protocol based on the model and Wilcoxon test to identify the PPB-related substructures (named privileged substructures, PSubs) for each molecule. In total, 22 general privileged substructures (GPSubs) were identified, which shared some common features such as nitrogen-containing groups, diamines with two carbon units, and azetidine. Furthermore, a series of second-level chemical rules for each GPSub were derived through a statistical test and then summarized into substructure pairs. We demonstrated that these substructure pairs were equally applicable outside the training set and accordingly customized the structural modification schemes for each GPSub, which provided alternatives for the optimization of the PPB property. Therefore, IDL-PPBopt provides a promising scheme for the prediction and optimization of the PPB property and would be helpful for lead optimization of other pharmacokinetic properties.