Another string to your bow: machine learning prediction of the pharmacokinetic properties of small molecules.

INTRODUCTION: Prediction of pharmacokinetic (PK) properties is crucial for drug discovery and development. Machine-learning (ML) models, which use statistical pattern recognition to learn correlations between input features (such as chemical structures) and target variables (such as PK parameters), are being increasingly used for this purpose. To embed ML models for PK prediction into workflows and to guide future development, a solid understanding of their applicability, advantages, limitations, and synergies with other approaches is necessary. AREAS COVERED: This narrative review discusses the design and application of ML models to predict PK parameters of small molecules, especially in light of established approaches including in vitro-in vivo extrapolation (IVIVE) and physiologically based pharmacokinetic (PBPK) models. The authors illustrate scenarios in which the three approaches are used and emphasize how they enhance and complement each other. In particular, they highlight achievements, the state of the art and potentials of applying machine learning for PK prediction through a comphrehensive literature review. EXPERT OPINION: ML models, when carefully crafted, regularly updated, and appropriately used, empower users to prioritize molecules with favorable PK properties. Informed practitioners can leverage these models to improve the efficiency of drug discovery and development process.

A randomized, double-blind phase 1b study evaluating the safety, tolerability, pharmacokinetics and pharmacodynamics of the NLRP3 inhibitor selnoflast in patients with moderate to severe active ulcerative colitis.

BACKGROUND: The NLRP3 inflammasome drives release of pro-inflammatory cytokines including interleukin (IL)-1ß and IL-18 and is a potential target for ulcerative colitis (UC). Selnoflast (RO7486967) is an orally active, potent, selective and reversible small molecule NLRP3 inhibitor. We conducted a randomized, placebo-controlled Phase 1b study to assess the safety, tolerability, pharmacokinetics (PK) and pharmacodynamics (PD) of selnoflast. METHODS: Nineteen adults with previous diagnosis of UC and current active moderate to severe disease were randomized 2:1 to selnoflast or placebo for 7 days. A dose of 450 mg QD (once daily) was selected to achieve 90% IL-1ß inhibition in plasma and colon tissue. Consecutive blood, sigmoid colon biopsies and stool samples were analyzed for a variety of PD markers. Safety and PK were also evaluated. RESULTS: Selnoflast was well-tolerated. Plasma concentrations increased rapidly after oral administration, reaching Tmax 1 h post-dose. Mean plasma concentrations stayed above the IL-1ß IC90 level throughout the dosing interval (mean Ctrough on Day 1 and Day 5: 2.55 µg/mL and 2.66 µg/mL, respectively). At steady state, post-dose selnoflast concentrations in sigmoid colon (5-20 µg/g) were above the IC90 . Production of IL-1ß was reduced in whole blood following ex vivo stimulation with lipopolysaccharide (LPS) (in the selnoflast arm). No changes were observed in plasma IL-18 levels. There were no meaningful differences in the expression of an IL-1-related gene signature in sigmoid colon tissue, and no differences in the expression of stool biomarkers. CONCLUSIONS: Selnoflast was safe and well-tolerated. Selnoflast 450 mg QD achieved plasma and tissue exposure predicted to maintain IL-1ß IC90 over the dosing interval. However, PD biomarker results showed no robust differences between treatment arms, suggesting no major therapeutic effects are to be expected in UC. The limitations of this study are its small sample size and indirect assessment of the effect on IL-1ß in tissue. TRIAL REGISTRATION: ISRCTN16847938.

Causal inference in drug discovery and development.

To discover new drugs is to seek and to prove causality. As an emerging approach leveraging human knowledge and creativity, data, and machine intelligence, causal inference holds the promise of reducing cognitive bias and improving decision-making in drug discovery. Although it has been applied across the value chain, the concepts and practice of causal inference remain obscure to many practitioners. This article offers a nontechnical introduction to causal inference, reviews its recent applications, and discusses opportunities and challenges of adopting the causal language in drug discovery and development.

Discovery of a first-in-class orally available HBV cccDNA inhibitor.

BACKGROUND & AIMS: The persistence of covalently closed circular DNA (cccDNA) in infected hepatocytes is the major barrier preventing viral eradication with existing therapies in patients with chronic hepatitis B. Therapeutic agents that can eliminate cccDNA are urgently needed to achieve viral eradication and thus HBV cure. METHODS: A phenotypic assay with HBV-infected primary human hepatocytes (PHHs) was employed to screen for novel cccDNA inhibitors. A HBVcircle mouse model and a uPA-SCID (urokinase-type plasminogen activator-severe combined immunodeficiency) humanized liver mouse model were used to evaluate the anti-HBV efficacy of the discovered cccDNA inhibitors. RESULTS: Potent and dose-dependent reductions in extracellular HBV DNA, HBsAg, and HBeAg levels were achieved upon the initiation of ccc_R08 treatment two days after the HBV infection of PHHs. More importantly, the level of cccDNA was specifically reduced by ccc_R08, while it did not obviously affect mitochondrial DNA. Additionally, ccc_R08 showed no significant cytotoxicity in PHHs or in multiple proliferating cell lines. The twice daily oral administration of ccc_R08 to HBVcircle model mice, which contained surrogate cccDNA molecules, significantly decreased the serum levels of HBV DNA and antigens, and these effects were sustained during the off-treatment follow-up period. Moreover, at the end of follow-up, the levels of surrogate cccDNA molecules in the livers of ccc_R08-treated HBVcircle mice were reduced to below the lower limit of quantification. CONCLUSIONS: We have discovered a small-molecule cccDNA inhibitor that reduces HBV cccDNA levels. cccDNA inhibitors potentially represent a new approach to completely cure patients chronically infected with HBV. IMPACT AND IMPLICATIONS: Covalently closed circular DNA (cccDNA) persistence in HBV-infected hepatocytes is the root cause of chronic hepatitis B. We discovered a novel small-molecule cccDNA inhibitor that can specifically reduce cccDNA levels in HBV-infected hepatocytes. This type of molecule could offer a new approach to completely cure patients chronically infected with HBV.

Discovery of Novel cccDNA Reducers toward the Cure of Hepatitis B Virus Infection.

Chronic hepatitis B virus (HBV) infection is a worldwide disease that causes thousands of deaths per year. Currently, there is no therapeutic that can completely cure already infected HBV patients due to the inability of humans to eliminate covalently closed circular DNA (cccDNA), which serves as the template to (re)initiate an infection even after prolonged viral suppression. Through phenotypic screening, we discovered xanthone series hits as novel HBV cccDNA reducers, and subsequent structure optimization led to the identification of a lead compound with improved antiviral activity and pharmacokinetic profiles. A representative compound 59 demonstrated good potency and oral bioavailability with no cellular toxicity. In an HBVcircle mouse model, compound 59 showed excellent efficacy in significantly reducing HBV antigens, DNA, and intrahepatic cccDNA levels.

Cross-reactive immunity potentially drives global oscillation and opposed alternation patterns of seasonal influenza A viruses.

Several human pathogens exhibit distinct patterns of seasonality and circulate as pairs. For instance, influenza A virus subtypes oscillate and peak during winter seasons of the world's temperate climate zones. Alternation of dominant strains in successive influenza seasons makes epidemic forecasting a major challenge. From the start of the 2009 influenza pandemic we enrolled influenza A virus infected patients (n = 2980) in a global prospective clinical study. Complete hemagglutinin sequences were obtained from 1078 A/H1N1 and 1033 A/H3N2 viruses. We used phylodynamics to construct high resolution spatio-temporal phylogenetic hemagglutinin trees and estimated global influenza A effective reproductive numbers (R) over time (2009-2013). We demonstrate that R oscillates around R = 1 with a clear opposed alternation pattern between phases of the A/H1N1 and A/H3N2 subtypes. Moreover, we find a similar alternation pattern for the number of global viral spread between the sampled geographical locations. Both observations suggest a between-strain competition for susceptible hosts on a global level. Extrinsic factors that affect person-to-person transmission are a major driver of influenza seasonality. The data presented here indicate that cross-reactive host immunity is also a key intrinsic driver of influenza seasonality, which determines the influenza A virus strain at the onset of each epidemic season.

Optimization of the TeraTox Assay for Preclinical Teratogenicity Assessment.

Current animal-free methods to assess teratogenicity of drugs under development still deliver high numbers of false negatives. To improve the sensitivity of human teratogenicity prediction, we characterized the TeraTox test, a newly developed multilineage differentiation assay using 3D human-induced pluripotent stem cells. TeraTox produces primary output concentration-dependent cytotoxicity and altered gene expression induced by each test compound. These data are fed into an interpretable machine-learning model to perform prediction, which relates to the concentration-dependent human teratogenicity potential of drug candidates. We applied TeraTox to profile 33 approved pharmaceuticals and 12 proprietary drug candidates with known in vivo data. Comparing TeraTox predictions with known human or animal toxicity, we report an accuracy of 69% (specificity: 53%, sensitivity: 79%). TeraTox performed better than 2 quantitative structure-activity relationship models and had a higher sensitivity than the murine embryonic stem cell test (accuracy: 58%, specificity: 76%, and sensitivity: 46%) run in the same laboratory. The overall prediction accuracy could be further improved by combining TeraTox and mouse embryonic stem cell test results. Furthermore, patterns of altered gene expression revealed by TeraTox may help grouping toxicologically similar compounds and possibly deducing common modes of action. The TeraTox assay and the dataset described here therefore represent a new tool and a valuable resource for drug teratogenicity assessment.

Besca, a single-cell transcriptomics analysis toolkit to accelerate translational research.

Single-cell RNA sequencing (scRNA-seq) revolutionized our understanding of disease biology. The promise it presents to also transform translational research requires highly standardized and robust software workflows. Here, we present the toolkit Besca, which streamlines scRNA-seq analyses and their use to deconvolute bulk RNA-seq data according to current best practices. Beyond a standard workflow covering quality control, filtering, and clustering, two complementary Besca modules, utilizing hierarchical cell signatures and supervised machine learning, automate cell annotation and provide harmonized nomenclatures. Subsequently, the gene expression profiles can be employed to estimate cell type proportions in bulk transcriptomics data. Using multiple, diverse scRNA-seq datasets, some stemming from highly heterogeneous tumor tissue, we show how Besca aids acceleration, interoperability, reusability and interpretability of scRNA-seq data analyses, meeting crucial demands in translational research and beyond.

Tissue heterogeneity is prevalent in gene expression studies.

Lack of reproducibility in gene expression studies is a serious issue being actively addressed by the biomedical research community. Besides established factors such as batch effects and incorrect sample annotations, we recently reported tissue heterogeneity, a consequence of unintended profiling of cells of other origins than the tissue of interest, as a source of variance. Although tissue heterogeneity exacerbates irreproducibility, its prevalence in gene expression data remains unknown. Here, we systematically analyse 2 667 publicly available gene expression datasets covering 76 576 samples. Using two independent data compendia and a reproducible, open-source software pipeline, we find a prevalence of tissue heterogeneity in gene expression data that affects between 1 and 40% of the samples, depending on the tissue type. We discover both cases of severe heterogeneity, which may be caused by mistakes in annotation or sample handling, and cases of moderate heterogeneity, which are likely caused by tissue infiltration or sample contamination. Our analysis establishes tissue heterogeneity as a widespread phenomenon in publicly available gene expression datasets, which constitutes an important source of variance that should not be ignored. Consequently, we advocate the application of quality-control methods such as BioQC to detect tissue heterogeneity prior to mining or analysing gene expression data.

Ten simple rules for doing a postdoc in pharma.

Postdoctoral programs in the pharmaceutical and life science industry offer opportunities for personal and professional development, if you know why to join, what to expect, and how to prepare.

Focus on germ-layer markers: A human stem cell-based model for in vitro teratogenicity testing.

Human induced pluripotent stem cells (hiPSC) were used to develop an assay format that may deliver information on teratogenicity of drugs. A human pluripotent stem cell scorecard panel was used to monitor the expression of 96 marker genes that are indicative of the stem cell state or differentiation into the ectoderm, mesoderm and endoderm lineages. We selected a human episomal iPS cell line for the assay based on karyotype stability, initial pluripotency, differentiation capacity and overall gene expression variability. The assay is based on embryoid body formation and was developed to be simply automated. In this proof of concept study, we used eight reference compounds (valproic acid, all-trans-retinoic acid, thalidomide, methotrexate, hydroxyurea, ascorbic acid, penicillin G and ibuprofen) to test the technical performance of the assay (readout stability) in concentration-response and time-course experiments. We also found that each compound affected marker gene expression in a different way. Various forms of data analysis identified 19 out of 96 early developmental genes as potential predictive markers for teratogenicity. Machine-learning models were run to exemplify how the assay will be developed further. The preliminary results from these analyses suggest that the assay could be suitable for the pre-screening of candidate pharmaceutical compounds. The approach presented here points a way towards development of a human cell-based assay that could replace the murine EST currently used to screen for early indications of potential teratogenicity of drug candidates.

Inducers of the endothelial cell barrier identified through chemogenomic screening in genome-edited hPSC-endothelial cells.

The blood-retina barrier and blood-brain barrier (BRB/BBB) are selective and semipermeable and are critical for supporting and protecting central nervous system (CNS)-resident cells. Endothelial cells (ECs) within the BRB/BBB are tightly coupled, express high levels of Claudin-5 (CLDN5), a junctional protein that stabilizes ECs, and are important for proper neuronal function. To identify novel CLDN5 regulators (and ultimately EC stabilizers), we generated a CLDN5-P2A-GFP stable cell line from human pluripotent stem cells (hPSCs), directed their differentiation to ECs (CLDN5-GFP hPSC-ECs), and performed flow cytometry-based chemogenomic library screening to measure GFP expression as a surrogate reporter of barrier integrity. Using this approach, we identified 62 unique compounds that activated CLDN5-GFP. Among them were TGF-ß pathway inhibitors, including RepSox. When applied to hPSC-ECs, primary brain ECs, and retinal ECs, RepSox strongly elevated barrier resistance (transendothelial electrical resistance), reduced paracellular permeability (fluorescein isothiocyanate-dextran), and prevented vascular endothelial growth factor A (VEGFA)-induced barrier breakdown in vitro. RepSox also altered vascular patterning in the mouse retina during development when delivered exogenously. To determine the mechanism of action of RepSox, we performed kinome-, transcriptome-, and proteome-profiling and discovered that RepSox inhibited TGF-ß, VEGFA, and inflammatory gene networks. In addition, RepSox not only activated vascular-stabilizing and barrier-establishing Notch and Wnt pathways, but also induced expression of important tight junctions and transporters. Taken together, our data suggest that inhibiting multiple pathways by selected individual small molecules, such as RepSox, may be an effective strategy for the development of better BRB/BBB models and novel EC barrier-inducing therapeutics.

Large-Scale Production of Human iPSC-Derived Macrophages for Drug Screening.

Tissue-resident macrophages are key players in inflammatory processes, and their activation and functionality are crucial in health and disease. Numerous diseases are associated with alterations in homeostasis or dysregulation of the innate immune system, including allergic reactions, autoimmune diseases, and cancer. Macrophages are a prime target for drug discovery due to their major regulatory role in health and disease. Currently, the main sources of macrophages used for therapeutic compound screening are primary cells isolated from blood or tissue or immortalized or neoplastic cell lines (e.g., THP-1). Here, we describe an improved method to employ induced pluripotent stem cells (iPSCs) for the high-yield, large-scale production of cells resembling tissue-resident macrophages. For this, iPSC-derived macrophage-like cells are thoroughly characterized to confirm their cell identity and thus their suitability for drug screening purposes. These iPSC-derived macrophages show strong cellular identity with primary macrophages and recapitulate key functional characteristics, including cytokine release, phagocytosis, and chemotaxis. Furthermore, we demonstrate that genetic modifications can be readily introduced at the macrophage-like progenitor stage in order to interrogate drug target-relevant pathways. In summary, this novel method overcomes previous shortcomings with primary and leukemic cells and facilitates large-scale production of genetically modified iPSC-derived macrophages for drug screening applications.

Signaling pathways predisposing to chronic kidney disease progression.

The loss of functional nephrons after kidney injury triggers the compensatory growth of the remaining ones to allow functional adaptation. However, in some cases, these compensatory events activate signaling pathways that lead to pathological alterations and chronic kidney disease. Little is known about the identity of these pathways and how they lead to the development of renal lesions. Here, we combined mouse strains that differently react to nephron reduction with molecular and temporal genome-wide transcriptome studies to elucidate the molecular mechanisms involved in these events. We demonstrated that nephron reduction led to 2 waves of cell proliferation: the first one occurred during the compensatory growth regardless of the genetic background, whereas the second one occurred, after a quiescent phase, exclusively in the sensitive strain and accompanied the development of renal lesions. Similarly, clustering by coinertia analysis revealed the existence of 2 waves of gene expression. Interestingly, we identified type I interferon (IFN) response as an early (first-wave) and specific signature of the sensitive (FVB/N) mice. Activation of type I IFN response was associated with G1/S cell cycle arrest, which correlated with p21 nuclear translocation. Remarkably, the transient induction of type I IFN response by poly(I:C) injections during the compensatory growth resulted in renal lesions in otherwise-resistant C57BL6 mice. Collectively, these results suggest that the early molecular and cellular events occurring after nephron reduction determine the risk of developing late renal lesions and point to type I IFN response as a crucial event of the deterioration process.

An Introduction to Machine Learning.

In the last few years, machine learning (ML) and artificial intelligence have seen a new wave of publicity fueled by the huge and ever-increasing amount of data and computational power as well as the discovery of improved learning algorithms. However, the idea of a computer learning some abstract concept from data and applying them to yet unseen situations is not new and has been around at least since the 1950s. Many of these basic principles are very familiar to the pharmacometrics and clinical pharmacology community. In this paper, we want to introduce the foundational ideas of ML to this community such that readers obtain the essential tools they need to understand publications on the topic. Although we will not go into the very details and theoretical background, we aim to point readers to relevant literature and put applications of ML in molecular biology as well as the fields of pharmacometrics and clinical pharmacology into perspective.

Multiscale modelling of drug mechanism and safety.

Here, we introduce models at three levels-molecular level, cellular and omics level, and organ and system level-that study drug mechanism and safety in preclinical drug discovery. The models differ in both their scope of study and technical details, but are all rooted in mathematical descriptions of complex biological systems, and all require informatics tools that handle large-volume, heterogeneous, and noisy data. We present principles and recent developments with examples at each level and highlight the synergy by a case study. We proffer a multiscale modelling view of drug discovery, call for a seamless flow of information in the form of models, and examine potential impacts.

Alzheimer's Risk Gene TREM2 Determines Functional Properties of New Type of Human iPSC-Derived Microglia.

Microglia are key in the homeostatic well-being of the brain and microglial dysfunction has been implicated in neurodegenerative disorders such as Alzheimer's disease (AD). Due to the many limitations to study microglia in situ or isolated for large scale drug discovery applications, there is a high need to develop robust and scalable human cellular models of microglia with reliable translatability to the disease. Here, we describe the generation of microglia-like cells from human induced pluripotent stem cells (iPSC) with distinct phenotypes for mechanistic studies in AD. We started out from an established differentiation protocol to generate primitive macrophage precursors mimicking the yolk sac ontogeny of microglia. Subsequently, we tested 36 differentiation conditions for the cells in monoculture where we exposed them to various combinations of media, morphogens, and extracellular matrices. The optimized protocol generated robustly ramified cells expressing key microglial markers. Bulk mRNA sequencing expression profiles revealed that compared to cells obtained in co-culture with neurons, microglia-like cells derived from a monoculture condition upregulate mRNA levels for Triggering Receptor Expressed On Myeloid Cells 2 (TREM2), which is reminiscent to the previously described disease-associated microglia. TREM2 is a risk gene for AD and an important regulator of microglia. The regulatory function of TREM2 in these cells was confirmed by comparing wild type with isogenic TREM2 knock-out iPSC microglia. The TREM2-deficient cells presented with stronger increase in free cytosolic calcium upon stimulation with ATP and ADP, as well as stronger migration towards complement C5a, compared to TREM2 expressing cells. The functional differences were associated with gene expression modulation of key regulators of microglia. In conclusion, we have established and validated a work stream to generate functional human iPSC-derived microglia-like cells by applying a directed and neuronal co-culture independent differentiation towards functional phenotypes in the context of AD. These cells can now be applied to study AD-related disease settings and to perform compound screening and testing for drug discovery.

Modeling the Effects of Severe Metabolic Disease by Genome Editing of hPSC-Derived Endothelial Cells Reveals an Inflammatory Phenotype.

The kinase AKT2 (PKB) is an important mediator of insulin signaling, for which loss-of-function knockout (KO) mutants lead to early onset diabetes mellitus, and dominant active mutations lead to early development of obesity and endothelial cell (EC) dysfunction. To model EC dysfunction, we used edited human pluripotent stem cells (hPSCs) that carried either a homozygous deletion of AKT2 (AKT2 KO) or a dominant active mutation (AKT2 E17K), which, along with the parental wild type (WT), were differentiated into ECs. Profiling of EC lines indicated an increase in proinflammatory and a reduction in anti-inflammatory fatty acids, an increase in inflammatory chemokines in cell supernatants, increased expression of proinflammatory genes, and increased binding to the EC monolayer in a functional leukocyte adhesion assay for both AKT2 KO and AKT2 E17K. Collectively, these findings suggest that vascular endothelial inflammation that results from dysregulated insulin signaling (homeostasis) may contribute to coronary artery disease, and that either downregulation or upregulation of the insulin pathway may lead to inflammation of endothelial cells. This suggests that the standard of care for patients must be expanded from control of metabolic parameters to include control of inflammation, such that endothelial dysfunction and cardiovascular disorders can ultimately be prevented.

Comprehensive evaluation of transcriptome-based cell-type quantification methods for immuno-oncology.

MOTIVATION: The composition and density of immune cells in the tumor microenvironment (TME) profoundly influence tumor progression and success of anti-cancer therapies. Flow cytometry, immunohistochemistry staining or single-cell sequencing are often unavailable such that we rely on computational methods to estimate the immune-cell composition from bulk RNA-sequencing (RNA-seq) data. Various methods have been proposed recently, yet their capabilities and limitations have not been evaluated systematically. A general guideline leading the research community through cell type deconvolution is missing. RESULTS: We developed a systematic approach for benchmarking such computational methods and assessed the accuracy of tools at estimating nine different immune- and stromal cells from bulk RNA-seq samples. We used a single-cell RNA-seq dataset of ∼11 000 cells from the TME to simulate bulk samples of known cell type proportions, and validated the results using independent, publicly available gold-standard estimates. This allowed us to analyze and condense the results of more than a hundred thousand predictions to provide an exhaustive evaluation across seven computational methods over nine cell types and ∼1800 samples from five simulated and real-world datasets. We demonstrate that computational deconvolution performs at high accuracy for well-defined cell-type signatures and propose how fuzzy cell-type signatures can be improved. We suggest that future efforts should be dedicated to refining cell population definitions and finding reliable signatures. AVAILABILITY AND IMPLEMENTATION: A snakemake pipeline to reproduce the benchmark is available at https://github.com/grst/immune_deconvolution_benchmark. An R package allows the community to perform integrated deconvolution using different methods (https://grst.github.io/immunedeconv). SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.