Search

There is an ongoing aim to replace animal and in vitro laboratory models with in silico methods. Such replacement requires the successful validation and comparably good performance of the alternative methods. We have developed an in silico prediction system for human clinical pharmacokinetics, based on machine learning, conformal prediction and a new physiologically-based pharmacokinetic model, i.e. ANDROMEDA. The objectives of this study were: a) to evaluate how well ANDROMEDA predicts the human clinical pharmacokinetics of a previously proposed benchmarking data set comprising 24 physicochemically diverse drugs and 28 small drug molecules new to the market in 2021; b) to compare its predictive performance with that of laboratory methods; and c) to investigate and describe the pharmacokinetic characteristics of the modern drugs. Median and maximum prediction errors for the selected major parameters were ca 1.2 to 2.5-fold and 16-fold for both data sets, respectively. Prediction accuracy was on par with, or better than, the best laboratory-based prediction methods (superior performance for a vast majority of the comparisons), and the prediction range was considerably broader. The modern drugs have higher average molecular weight than those in the benchmarking set from 15 years earlier ( ca 200 g/mol higher), and were predicted to (generally) have relatively complex pharmacokinetics, including permeability and dissolution limitations and significant renal, biliary and/or gut-wall elimination. In conclusion, the results were overall better than those obtained with laboratory methods, and thus serve to further validate the ANDROMEDA in silico system for the prediction of human clinical pharmacokinetics of modern and physicochemically diverse drugs.

The gastrointestinal uptake of macrocyclic compounds is not fully understood. Here we applied our previously validated integrated system based on machine learning and conformal prediction to predict the passive fraction absorbed (f

Challenges/problems forin vitromethodologies for prediction of human clinical pharmacokinetics include inter- and intra-laboratory variability, and common occurance of high limits of quantification, low recovery, low parameter validity and low reproducibility. In this study, 10 drugs with substantial differences in human hepatocyte intrinsic metabolic clearance (CLint) and fraction unbound in plasma (fu) between laboratories were selected. The average and maximum ratios between highest and lowest reported predictedin vivohepatic metabolic clearance (CLH) for the drugs were 529- and 2436-fold, respectively. Thein vivoCLHwas predicted usingin vitroCLintand fudata from the various highly sources and using ourin silicomethodology. The main aim was to compare the predictive accuracies for thein vitroandin silicomethodologies. Prediction errors forin vitromethodology ranged from 1.1-to 578-fold, with an average of 150-fold for lowest predicted estimates and 16-fold for highest predicted estimates. Thein vitrobased predictions produced 36-to 38-fold higher average and maximum prediction errors than thein silicomethodology, respectively. Mean and maximumin silicoprediction errors were 4.2- and 15-fold, respectively, which is consistent with earlier results. In contrast to thein vitromethodology thein silicomodels did not predict high hepatic extraction ratio for drugs with low CLH. Overall, thein silicomethod clearly outperformedin vitrodata for prediction of CLHin man for 10 drugs with large interlaboratory variability.

The ANDROMEDA software, based on machine learning, conformal prediction and a new physiologically-based pharmacokinetic model, was used to predict and characterize the human clinical pharmacokinetics of 30 selected modern small antibiotic compounds (investigational and marketed drugs). A majority of clinical pharmacokinetic data was missing. ANDROMEDA successfully filled this gap. Most antibiotics were predicted and measured to have limited permeability, good metabolic stability and multiple elimination pathways. According to predictions, most of the antibiotics are mainly eliminated renally and biliary and every other antibiotic is mainly eliminated via the renal route. Mean prediction errors for steady state volume of distribution, unbound fraction in plasma, renal and total clearance, oral clearance, fraction absorbed, fraction excreted renally, oral bioavailability and half-life were 1.3- to 2.3-fold. The overall median and maximum prediction errors were 1.5- and 4.8-fold, respectively, and 92 % of predictions had

The ANDROMEDA toolkit for prediction of human clinical pharmacokinetics, based on machine learning, conformal prediction and a new physiologically-based pharmacokinetic model, was used to predict and characterize the human clinical pharmacokinetics of 12 small anticancer drugs marketed in 2021 and 2022 (molecular weight 355 to 1326 g/mol). The study is part of a series of software validations. A majority of clinical pharmacokinetic data was missing. ANDROMEDA successfully filled this gap. Most drugs were predicted/measured to have relatively complex pharmacokinetics, with limited passive permeability+efflux, high degree of plasma protein binding, significant gut-wall elimination and food interaction, biliary excretion and/or limited dissolution potential. Median, mean and maximum prediction errors for steady state volume of distribution, unbound fraction in plasma, blood-to-plasma concentration ratio, hepatic, renal and total clearance, fraction absorbed, oral bioavailability, half-life and degree of food interaction were 1.6-, 2.4- and 17-fold, respectively. Less than 3-fold errors were found for 78 % of predictions. Results are consistent with those obtained in previous validation studies and are better than with the best laboratory-based prediction methods, which validates ANDROMEDA for predictions of human clinical pharmacokinetics of modern small anticancer drugs with multi-mechanistical and challenging pharmacokinetics.

BackgroundPrevious work has shown considerable laboratory variability of Biopharmaceutics Classification System (BCS) classification, efflux ratio in intestinal cell lines and cytochrome P450 (CYP450)-metabolism pathways. Such variability and inconsistency create uncertainty in predictions of human clinical pharmacokinetics and the pharmacokinetic optimization process and is a problem when developing correspondingin silicomethods.Objectives and MethodologyOne objective of the study was to quantify the degree of laboratory inconsistency for BCS II-classing, MDR-1 and CYP3A4 substrate specificity (substrate/non-substrate). Another objective was to predict BCS II-classing, MDR-1 and CYP3A4 substrate specificity usingin silicomethodology and compare results to laboratory data/classifications.Results and Discussion27 BCS II-classified drugs (with non-contradictory BCS-classing in various sources) were found. 17 (63 %) had anin vivofraction absorbed (fa) of ≥90 % and belong toin vivoBCS I. Within silicomethodology, 74 % correct BCS-classing was reached for the same set of compounds. The mean prediction error for fawas 1.2-fold. MDR-1 and CYP3A4 substrate specificities were collected for 346 and 808 compounds, respectively. For MDR-1, 143 of the compounds had reported data in at least two studies, and out of these, 49 (34 %) and 18 (13 %) had contradictory (reported as both substate and non-substrate) and uncertain substrate specificities, respectively. For CYP3A4, 42 (9.8 %) out of 427 compounds showed inconsistency between laboratories. Within silicomethodology, MDR-1 and CYP3A4 classification predictions were incorrect for 13 and 15 % of compounds.ConclusionThe results show considerable variability/inconsistency for BCS II-classing (63 % inconsistency between BCS II-classing andin vivofa) and MDR-1 (34 % inconsistency between sources) and CYP3A4 (10 % inconsistency between sources) substrate specificities. Corresponding estimates obtained within silicomethodology are 22, 13 and 15 %, respectively, demonstrating the power and applicability of such technology.

IntroductionThe ANDROMEDA software by Prosilico has previously been successfully applied and validated for predictions of absorption characteristics of small drugs in man. The effects of fat food on the gastrointestinal uptake and systemic exposure of drugs have, however, not yet been evaluated with the software.Objective and MethodologyThe main objective was to use ANDROMEDA to predict area under the plasma concentration-time curve ratios in the fed (fat food) and fasted states (AUCfed/AUCfast) for small drugs marketed in 2021 and compare results with corresponding measured clinical estimates. A secondary aim was to make a similar evaluation for a set of selected older small drugs with no and significant food interaction.Results17 new drugs with observed AUCfed/AUCfastbetween 0.4 and 2.0 were found. In addition, 18 older drugs with low (0.5), moderate (1) and high (3.1-5.5) AUCfed/AUCfastwere selected. Predicted AUCfed/AUCfast(for all drugs) had mean, median and maximum errors of 1.4-, 1.1- and 3.9-fold, respectively, and the predictive accuracy (correlation between predicted and observed AUCfed/AUCfast; Q2) was 0.35. For new small drugs, the Q2was 0.41. All 10 new drugs with AUCfed/AUCfastbetween 0.8 and 1.25 were predicted to have AUCfed/AUCfastbetween 0.6 and 1.5 and all 4 new drugs with AUCfed/AUCfastfed/AUCfastfed/AUCfast≥1.25).ConclusionThe results validate ANDROMEDA for prediction of fat food-drug interaction size for small drugs in man. Major advantages with the methodology include that results are produced directly from molecular structure and oral dose and an adequate level of predictive accuracy.

IntroductionSome prodrugs are developed in order to improve gastrointestinal absorption properties such as permeability and solubility/dissolution. Prediction of the uptake of prodrugs and their drugs is challening for reasons including gastrointestinal hydrolysis and active transport.Objective and MethodologyThe objective was to use the ANDROMEDA software by Prosilico to predict absorption characteristics - passive fraction absorbed (fa,passive), dose-adjusted dissolution potential (fdiss) and total fa(fa) - of prodrugs and their drugs (including drugs and their active metabolites), and to evaluate how they differ between prodrugs and drugs and the predictive accuracy of the software.Results70 prodrug-drug pairs were found and selected for the study. The mean predicted fa,passiveand fdissfor the prodrugs were 0.74 and 0.94, respectively. Corresponding estimates for the drugs were 0.72 and 0.98, respectively. For non-hydrolyzed prodrugs, the median relative and absolute prediction errors for fawere 1.17-fold and 0.08, respectively. Corresponding values for drugs were 1.11-fold and 0.07, respectively. The correlation between predicted and observed fafor non-hydrolyzed ester prodrugs and drugs combined (predictive accuracy) was 0.6.ConclusionProdrugs and drugs had similar average predicted fa,passiveand fdiss, and most had or were predicted to have at least 50 % fa. The fafor about 1/3 of non-hydrolyzed prodrugs was higher than for corresponding drugs, showing successful prodrug design. Adequate prediction accuracy validates ANDROMEDA for prediction of prodrug and drug absorption in man.

BackgroundBlood-brain barrier permeability (BBB Pe) and unbound brain-to-plasma concentration ratio (Kp,uu,brain) are relevant parameters describing the brain uptake potential of compounds. BBB efflux by transporter proteins, mainly MDR-1 and BCRP, is an essential factor determining Kp,uu,brain. Kp,uu,brain-values are commonly estimatedin vivoin rats and monkeys and predicted usingin silicomethodology. Such estimates can be used to predict corresponding human clinical values.ObjectiveThe objective of the study was to evaluate the reliability and applicability of human clinical Kp,uu,brain-data for understanding and predictions of brain uptake in man.MethodologyKp,uu,brainin rats, monkeys and humans, measured andin silicopredicted MDR-1 and BCRP substrate specificities andin silicopredicted passive Pewere used for the analysis.In silicopredictions were done using the ANDROMEDA by Prosilico ADME/PK-prediction software.Results and DiscussionRat and monkey Kp,uu,brain-values were highly correlated (R^2=0.74; n=17). Based on this finding a correlation between rat and human Kp,uu,brainwas expected. However, no correlation between rat and human Kp,uu,brainwas found (R^2=0.01; n=13). There was no (as also anticipated) correlation between passive Peand human Kp,uu,brain(R^2=0.04; n=16) and compounds with measured or predicted efflux did not have lower Kp,uu,brainthan compounds without efflux. The compound with highest Kp,uu,brainin man (2.8) is effluxed and predicted to have high passive Peand has no apparent efflux at the rat BBB. The MDR-1 substrate with highest Kp,uu,brainin rat (2.4) has very low Kp,uu,brainin man (0.15) is predicted to have high passive Pe.ConclusionResults indicate that available human Kp,uu,brain-data are too uncertain to be applicable for validation of predictions and understanding of clinical brain uptake of drugs and drug candidates.

IntroductionConformal prediction (CP) methodology sits on top of machine learning methods and produces prediction confidence intervals that depend on how “strange” (non-conforming) test compounds are compared to training set compounds. CP has previously been successfully applied for prediction of steady-state volume of distribution (Vss) in humans, with 69 % of observations within the prediction interval at a 70 % confidence level. We have developed CP models for a variety of human pharmacokinetic (PK) parameters and validated their predictive accuracy (predictedvsobserved estimates), but not validated prediction confidence intervals for them. The main objective of this study was to predict 70 % confidence intervals for Vss, unbound fraction in plasma (fu), intrinsic metabolic clearance (CLint), fraction absorbed passively (fa,passive) and maximum fraction dissolved (fdiss) for a variety of compounds in man and investigate the consistency between prediction intervals and observed/measured values.MethodologyCP models featured in the ANDROMEDA software by Prosilico were used for prediction of 70 % confidence intervals of Vss, fu, CLint, fa,passiveand fdissfor compounds from different chemical classes and with broad physicochemical variety and for small drugs marketed in 2021.Results70 % prediction confidence intervals for 217, 117, 117, 89 and 89 compounds were produced for Vss, fu, CLint, fa,passiveand fdiss, respectively. 78 % (expected 70 %) of observed data were within 70 % confidence intervals for the parameters. 70 % of predictions of Vss, fu, CLintfa,passiveand fdissare expected to have errors of maximally 2-, 4- and 6-fold and 7 and 12 %, respectively, which is in line with prediction errors. These findings validate the CP methodology.ConclusionIn conclusion, the results further validate CP models and confidence intervals of ANDROMEDA for prediction of human PK.

IntroductionIn vitromeasurements and predictions of human clinical pharmacokinetics (PK) are sometimes hindered and made impossible due to factors such as extensive binding to materials, low methodological sensitivity and large variability.MethodsThe objective was to find compounds out of reach forin vitroPK-methods and (if possible) predict corresponding human clinical estimates using the ANDROMEDA by Prosilico software.In vitromethods selected for the investigation were human microsomes and hepatocytes for measuring and predicting intrinsic hepatic metabolic clearance (CLint), Caco-2 cells for measuring apparent intestinal permeability (Papp) for prediction of fraction absorbed (fa), plasma for measurement and estimation of unbound fraction (fu), and water and buffers for measuring solubility (S) for prediction ofin vivodissolution potential (fdiss).Results and Conclusion73 non-quantifiablein vitroPK-measurements for 70 compounds were found in the literature: 40 for CLint, 8 for Papp, 11 for fuand 14 for S. ANDROMEDA was successful in predicting all corresponding clinical PK-estimates for the selection of compounds with non-quantifiablein vitroPK, and predicted estimates were generally in line with observedin vivodata and results/problems atin vitrolaboratories. Thus, ANDROMEDA is applicable for predicting human clinical PK for compounds out of reach for laboratory methods.

BackgroundPassive blood-brain barrier permeability (BBB Pe), fraction bound to brain tissue (fb,brain) and efflux by transport proteins MDR-1 and BCRP are essential determinants for the brain uptake and disposition of drugs.MethodsThe main objective of the study was to use the software ANDROMEDA by Prosilico to predict passive BBB Pe- and fb,brain-classes and MDR-1- and BCRP-specificities for various classes of antidepressants and for CNS-active small drugs marketed during 2020 and 2021, and then to position them according to a new 2-dimensional Brainavailability-Matrix (8 passive BBB Pex 4 fb,brainclasses, where class 11 has highest and 84 lowest values/brainavailability). Predicted estimates were used, except for cases where measured values were available.Results and ConclusionResults for 53 drugs show that adequate CNS uptake and disposition are achieved for compounds placed in the zones for low, moderate and high brainavailability, despite efflux. They also show that high brainavailability and efflux are common for CNS-active drugs and that modern CNS-active drugs generally have lower brainavailability than older antidepressive drugs. Furthermore, they demonstrate that ANDROMEDA by Prosilico and the new Brainavailability-Matrix are applicable for prediction, optimization and positioning of CNS uptake and disposition of drugs and drug candidates in man.

IntroductionVarious in vitro methods are used to measure absorption, distribution, metabolism and excretion/pharmacokinetics (ADME/PK) of candidate drugs and predict and decide whether properties are clinically adequate.MethodsObjectives were to evaluate variability within and between laboratories for commonly used human in vitro ADME/PK methods and to explore whether reliable thresholds may be defined. The literature was searched for in vitro data for intrinsic metabolic clearance (hepatocyte CLint), apparent intestinal permeability (Caco-2 Papp), efflux ratio (Caco-2 ER), solubility (S) and BCS-class, and corresponding clinical estimates. In vitro ADME/PK data for three example drugs (atenolol, diclofenac and gemfibrozil) were used to predict human in vivo ADME/PK and investigate whether these would pass a compound selection process.Results and ConclusionsInterlaboratory variability is considerable, especially for fu, S, ER and BCS-classification, and on average about twice as high as intralaboratory variability. Approximate mean interlaboratory variability for CLint, Papp, ER and fu (3- to 3.5-fold) appears to be about 2- to 3-fold higher than corresponding interlaboratory variability. Mean and maximum interlaboratory range for CLint, Papp, ER, fu and S are approximately 5- to 100-fold and 50- to 4500-fold, respectively, with second largest range for fu and largest range for S. For one drug, laboratories produced almost 1000-fold different CLint • fu-values. It appears difficult/impossible to set clear clinically useful thresholds, especially for CLint, ER and S. Poor in vitro-in vivo consistency for S and BCS-classification and large portions of compounds out of reach for Caco-2 and conventional hepatocyte assays are evident. Predictions for reference compounds are consistent with inadequate in vivo ADME/PK. Ways to improve predictions and compound selection are suggested.

BackgroundA shift towards higher molecular weight (MW) of drug candidates is anticipated to lead to changed pharmacokinetics (PK), including deteriorated absorption.MethodsThe objective of the study was to investigate changes in MW and PK of drugs over time by comparing MW and measured PK of small drugs (here MWin silicopredicted (data produced using the ANDROMEDA by Prosilico software) and of small drugs marketed in 2021 (n=28).ResultsApparently, there has been a shift towards higher MW (from 355 to 551 g/mole on average). This has influenced PK-parameters such as unbound fraction (on average approximately halved), fraction excreted renally (on average approximately halved; markedly decreased contribution by active secretion), bile excretion (almost 4-fold increased appearance; now for more than every other drug) and intrinsic metabolic clearance (increased). The very high percentage of modern drugs with (according toin silicopredictions) significant renal and biliary excretion and gut-wall extraction, metabolic stability, limited passive intestinal permeability+efflux, limited gastrointestinal dissolution/solubility potential and/or a very low fuincreases complexity in predictions and places demands on predictive laboratory and computational methods.ConclusionIncreased MW and changed PK-profiles (increased complexity) with time were observed. This shows the need for updating method set-ups for quantification and prediction of PK-parameters. ANDROMEDA has the capability to predict and optimize human clinical PK-characteristics of modern drug candidates with high accuracy.

BackgroundPharmacokinetics (PK), exposure profiles and doses of candidate drugs in man are commonly predicted using data produced using various in vitro methods, such as hepatocytes (for intrinsic metabolic clearance (CLint)), plasma (for unbound fraction (fu)), Caco-2 (measuring apparent permeability (Papp) for prediction of in vivo fraction absorbed (fa)) and plasma water and buffers (measuring solubility (S) for prediction of in vivo fraction dissolved (fdiss)). For best possible predictions it is required that the clinical relevance of in vitro data is understood (in vitro-in vivo relationships) and that uncertainties have been investigated and considered.MethodsThe aim was to investigate in vitro-in vivo relationships for CLint, Pappvs fa and S vs fdiss and interlaboratory variability for fu, describe the clinical significance and uncertainties at certain levels of in vitro CLint, fu, Papp and S, and (based on the findings) develop a general in vitro-in vivo translation guide.Results and ConclusionIt was possible to finf data for describing how in vivo CLint, fa and fdiss distribute and varies at different levels of in vitro CLint, Papp and S and how fu varies between laboratories and methods at different fu-levels. It is apparent that there are considerable interlaboratory variabilities for CLint, fu and Papp: corresponding to up to 2500-, 700- and 35-fold variability for CLint, fu and fa, respectively. Apparently, S is a poor predictor of fdiss. Proposed S-thresholds do not seem clinically useful (overestimated). It does not seem appropriate to define in vitro CLint of 0.5-2 µL/min/106 cells as good metabolic stability (rather moderate to moderately high). Results shown for CLint, Papp and fu are applicable as general guidelines when internal standard values for reference compounds are unavailable.

BackgroundPROTACs are comparably large and flexible compounds with limited solubility (S) and permeability (Pe). It is crucial to better understand, predict and optimize their human clinical pharmacokinetics (PK).MethodsThe main objective was to use the ANDROMEDA by Prosilico software to predict the human clinicalin vivodissolution potential (fdiss) and fraction absorbed (fa) of 23 PROTACs at a dose level of 50 mg and to explore whether there is any relationship betweenin vitroS andin silicopredictedin vivofdiss.ResultsIn silicopredictions showed that the PROTACs are effluxed by intestinal transporters and have limited fdiss(34 to 98 %), permeability and fa(13 to 58 %) in man. For some PROTACs this may be a major obstacle and jeopardize the clinical development programs, especially in cases of required high oral dose. A modest relationship betweenin vitroS and predictedin vivofdisswas demonstrated (R2=0.26). Predicted human fa(27 %) and oral bioavailability (20 %) of ARV-110 (a PROTAC with some availablein vivoPK data in rodents and man) were consistent with data obtained in rodents (estimated faapproximately 30-40 %; measured oral bioavailability 27-38 %). Laboratories were unable to quantify S for 7 (30 %) of the PROTACs. In contrast, ANDROMEDA could predict parameters for all.ConclusionANDROMEDA predicted fdissand fafor all the chosen PROTACs and showed limited fdiss, Peand faand dose-dependent fdissand fa. One available example shows promise for the applicability of ANDROMEDA for predicting biopharmaceutics of PROTACsin vivoin man. Weak to modest correlations between S and fdissand a considerable portion of compounds with non-quantifiable S limit the use of S-data to predict the uptake of PROTACs.

BackgroundIt is important that pharmacokinetic (PK) prediction methods are validated, and also for compounds with varying physicochemical properties, molecular weights and PK characteristics.MethodsThe objective was to investigate how well the ANDROMEDA by Prosilico software predicts the clinical PK of four compounds of natural origin and with PK obstacles, not yet fully characterized PK, and/or inaccurate lab method-based predictions - colistin (negligible absorption, good metabolic stability, significant excretion), curucumin (low solubility, apparently poor bioavailability), UCN-01 (extremely high degree of plasma protein binding, metabolic stability, long half-life and poor PK prediction) and voclosporin (poorly understood PK).ResultsAll categorial predictions except one were correct, and the median prediction error was 2.5-fold. Largest prediction errors were found for the unbound fraction in plasma (>24-fold), clearance (178-fold) and half-life (90-fold) of UCN-01. Corresponding errors for clearance and half-life obtained with allometry were greater, 5800- and 145-fold, respectively. Extremely high affinity for alpha1-acid glycoprotein could explain these large prediction errors for this compound. A substantial amount of data and knowledge was added with the predictions.ConclusionDespite challenging compounds and PK, predictions were comparably good. The results further validated ANDROMEDA by Prosilico for human clinical PK-predictions.

In vitro-in vivo prediction results for hepatic metabolic clearance (CL

Pharmacokinetic/toxicokinetic (PK/TK) information for chemicals in humans is generally lacking. Here we applied machine learning, conformal prediction and a new physiologically-based PK/TK model for prediction of the human PK/TK of 65 chemicals from different classes, including carcinogens, food constituents and preservatives, vitamins, sweeteners, dyes and colours, pesticides, alternative medicines, flame retardants, psychoactive drugs, dioxins, poisons, UV-absorbents, surfactants, solvents and cosmetics.About 80% of the main human PK/TK (fraction absorbed, oral bioavailability, half-life, unbound fraction in plasma, clearance, volume of distribution, fraction excreted) for the selected chemicals was missing in the literature. This information was now added (from in silico predictions). Median and mean prediction errors for these parameters were 1.3- to 2.7-fold and 1.4- to 4.8-fold, respectively. In total, 59 and 86% of predictions had errors in silico predictions). Median and mean prediction errors for these parameters were 1.3- to 2.7-fold and 1.4- to 4.8-fold, respectively. In total, 59 and 86% of predictions had errors