Prospective de novo drug design with deep interactome learning

De novo drug design aims to generate molecules from scratch that possess specific chemical and pharmacological properties. We present a computational approach utilizing interactome-based deep learning for ligand- and structure-based generation of drug-like molecules. This method capitalizes on the unique strengths of both graph neural networks and chemical language models, offering an alternative to the need for application-specific reinforcement, transfer, or few-shot learning. It enables the "zero-shot" construction of compound libraries tailored to possess specific bioactivity, synthesizability, and structural novelty. In order to proactively evaluate the deep interactome learning framework for protein structure-based drug design, potential new ligands targeting the binding site of the human peroxisome proliferator-activated receptor (PPAR) subtype gamma are generated. The top-ranking designs are chemically synthesized and computationally, biophysically, and biochemically characterized. Potent PPAR partial agonists are identified, demonstrating favorable activity and the desired selectivity profiles for both nuclear receptors and off-target interactions. Crystal structure determination of the ligand-receptor complex confirms the anticipated binding mode. This successful outcome positively advocates interactome-based de novo design for application in bioorganic and medicinal chemistry, enabling the creation of innovative bioactive molecules.

Integrating QSAR modelling and deep learning in drug discovery: the emergence of deep QSAR

Quantitative structure-activity relationship (QSAR) modelling, an approach that was introduced 60 years ago, is widely used in computer-aided drug design. In recent years, progress in artificial intelligence techniques, such as deep learning, the rapid growth of databases of molecules for virtual screening and dramatic improvements in computational power have supported the emergence of a new field of QSAR applications that we term 'deep QSAR'. Marking a decade from the pioneering applications of deep QSAR to tasks involved in small-molecule drug discovery, we herein describe key advances in the field, including deep generative and reinforcement learning approaches in molecular design, deep learning models for synthetic planning and the application of deep QSAR models in structure-based virtual screening. We also reflect on the emergence of quantum computing, which promises to further accelerate deep QSAR applications and the need for open-source and democratized resources to support computer-aided drug design.

Enabling late-stage drug diversification by high-throughput experimentation with geometric deep learning

Late-stage functionalization is an economical approach to optimize the properties of drug candidates. However, the chemical complexity of drug molecules often makes late-stage diversification challenging. To address this problem, a late-stage functionalization platform based on geometric deep learning and high-throughput reaction screening was developed. Considering borylation as a critical step in late-stage functionalization, the computational model predicted reaction yields for diverse reaction conditions with a mean absolute error margin of 4-5%, while the reactivity of novel reactions with known and unknown substrates was classified with a balanced accuracy of 92% and 67%, respectively. The regioselectivity of the major products was accurately captured with a classifier F-score of 67%. When applied to 23 diverse commercial drug molecules, the platform successfully identified numerous opportunities for structural diversification. The influence of steric and electronic information on model performance was quantified, and a comprehensive simple user-friendly reaction format was introduced that proved to be a key enabler for seamlessly integrating deep learning and high-throughput experimentation for late-stage functionalization.

Exploring protein-ligand binding affinity prediction with electron density-based geometric deep learning

Rational structure-based drug design relies on accurate predictions of protein-ligand binding affinity from structural molecular information. Although deep learning-based methods for predicting binding affinity have shown promise in computational drug design, certain approaches have faced criticism for their potential to inadequately capture the fundamental physical interactions between ligands and their macromolecular targets or for being susceptible to dataset biases. Herein, we propose to include bond-critical points based on the electron density of a protein-ligand complex as a fundamental physical representation of protein-ligand interactions. Employing a geometric deep learning model, we explore the usefulness of these bond-critical points to predict absolute binding affinities of protein-ligand complexes, benchmark model performance against existing methods, and provide a critical analysis of this new approach. The models achieved root-mean-squared errors of 1.4-1.8 log units on the PDBbind dataset, and 1.0-1.7 log units on the PDE10A dataset, not indicating significant advantages over benchmark methods, and thus rendering the utility of electron density for deep learning models context-dependent. The relationship between intermolecular electron density and corresponding binding affinity was analyzed, and Pearson correlation coefficients r > 0.7 were obtained for several macromolecular targets.

Baricitinib and tofacitinib off-target profile, with a focus on Alzheimer's disease

INTRODUCTION

Janus kinase (JAK) inhibitors were recently identified as promising drug candidates for repurposing in Alzheimer's disease (AD) due to their capacity to suppress inflammation via modulation of JAK/STAT signaling pathways. Besides interaction with primary therapeutic targets, JAK inhibitor drugs frequently interact with unintended, often unknown, biological off-targets, leading to associated effects. Nevertheless, the relevance of JAK inhibitors' off-target interactions in the context of AD remains unclear.

METHODS

Putative off-targets of baricitinib and tofacitinib were predicted using a machine learning (ML) approach. After screening scientific literature, off-targets were filtered based on their relevance to AD. Targets that had not been previously identified as off-targets of baricitinib or tofacitinib were subsequently tested using biochemical or cell-based assays. From those, active concentrations were compared to bioavailable concentrations in the brain predicted by physiologically based pharmacokinetic (PBPK) modeling.

RESULTS

With the aid of ML and in vitro activity assays, we identified two enzymes previously unknown to be inhibited by baricitinib, namely casein kinase 2 subunit alpha 2 (CK2-α2) and dual leucine zipper kinase (MAP3K12), both with binding constant (K d) values of 5.8 μM. Predicted maximum concentrations of baricitinib in brain tissue using PBPK modeling range from 1.3 to 23 nM, which is two to three orders of magnitude below the corresponding binding constant.

CONCLUSION

In this study, we extended the list of baricitinib off-targets that are potentially relevant for AD progression and predicted drug distribution in the brain. The results suggest a low likelihood of successful repurposing in AD due to low brain permeability, even at the maximum recommended daily dose. While additional research is needed to evaluate the potential impact of the off-target interaction on AD, the combined approach of ML-based target prediction, in vitro confirmation, and PBPK modeling may help prioritize drugs with a high likelihood of being effectively repurposed for AD.

Highlights

This study explored JAK inhibitors' off-targets in AD using a multidisciplinary approach.We combined machine learning, in vitro tests, and PBPK modelling to predict and validate new off-target interactions of tofacitinib and baricitinib in AD.Previously unknown inhibition of two enzymes (CK2-a2 and MAP3K12) by baricitinib were confirmed using in vitro experiments.Our PBPK model indicates that baricitinib low brain permeability limits AD repurposing.The proposed multidisciplinary approach optimizes drug repurposing efforts in AD research.

Identifying opportunities for late-stage C-H alkylation with high-throughput experimentation and in silico reaction screening

Enhancing the properties of advanced drug candidates is aided by the direct incorporation of specific chemical groups, avoiding the need to construct the entire compound from the ground up. Nevertheless, their chemical intricacy often poses challenges in predicting reactivity for C-H activation reactions and planning their synthesis. We adopted a reaction screening approach that combines high-throughput experimentation (HTE) at a nanomolar scale with computational graph neural networks (GNNs). This approach aims to identify suitable substrates for late-stage C-H alkylation using Minisci-type chemistry. GNNs were trained using experimentally generated reactions derived from in-house HTE and literature data. These trained models were then used to predict, in a forward-looking manner, the coupling of 3180 advanced heterocyclic building blocks with a diverse set of sp3-rich carboxylic acids. This predictive approach aimed to explore the substrate landscape for Minisci-type alkylations. Promising candidates were chosen, their production was scaled up, and they were subsequently isolated and characterized. This process led to the creation of 30 novel, functionally modified molecules that hold potential for further refinement. These results positively advocate the application of HTE-based machine learning to virtual reaction screening.

Designing molecules with autoencoder networks

Autoencoders are versatile tools in molecular informatics. These unsupervised neural networks serve diverse tasks such as data-driven molecular representation and constructive molecular design. This Review explores their algorithmic foundations and applications in drug discovery, highlighting the most active areas of development and the contributions autoencoder networks have made in advancing this field. We also explore the challenges and prospects concerning the utilization of autoencoders and the various adaptations of this neural network architecture in molecular design.

Compression of Molecular Fingerprints with Autoencoder Networks

Several binary molecular fingerprints were compressed using an autoencoder neural network. We analyzed the impact of compression on fingerprint performance in downstream classification and regression tasks. Classifiers trained on compressed fingerprints were negligibly affected. Regression models benefitted from compression, especially of long fingerprints (Morgan, RDK). However, their performance dropped rapidly for compression levels exceeding 90%. Property co-learning positively influenced the predictive power of the compressed fingerprints, with a mean score improvement up to 20%, suggesting that autoencoder compression with property co-learning biases the molecular representation toward the predicted target, facilitating downstream training.

Structure-based drug design with geometric deep learning

Structure-based drug design uses three-dimensional geometric information of macromolecules, such as proteins or nucleic acids, to identify suitable ligands. Geometric deep learning, an emerging concept of neural-network-based machine learning, has been applied to macromolecular structures. This review provides an overview of the recent applications of geometric deep learning in bioorganic and medicinal chemistry, highlighting its potential for structure-based drug discovery and design. Emphasis is placed on molecular property prediction, ligand binding site and pose prediction, and structure-based de novo molecular design. The current challenges and opportunities are highlighted, and a forecast of the future of geometric deep learning for drug discovery is presented.

Allosteric targeting resolves limitations of earlier LFA-1 directed modalities

Integrins are a family of cell surface receptors well-recognized for their therapeutic potential in a wide range of diseases. However, the development of integrin targeting medications has been impacted by unexpected downstream effects, reflecting originally unforeseen interference with the bidirectional signalling and cross-communication of integrins. We here selected one of the most severely affected target integrins, the integrin lymphocyte function-associated antigen-1 (LFA-1, α_Lβ₂, CD11a/CD18), as a prototypic integrin to systematically assess and overcome these known shortcomings. We employed a two-tiered ligand-based virtual screening approach to identify a novel class of allosteric small molecule inhibitors targeting this integrin's αI domain. The newly discovered chemical scaffold was derivatized, yielding potent bis-and tris-aryl-bicyclic-succinimides which inhibit LFA-1 in vitro at low nanomolar concentrations. The characterisation of these compounds in comparison to earlier LFA-1 targeting modalities established that the allosteric LFA-1 inhibitors (i) are devoid of partial agonism, (ii) selectively bind LFA-1 versus other integrins, (iii) do not trigger internalization of LFA-1 itself or other integrins and (iv) display oral availability. This profile differentiates the new generation of allosteric LFA-1 inhibitors from previous ligand mimetic-based LFA-1 inhibitors and anti-LFA-1 antibodies, and is projected to support novel immune regulatory regimens selectively targeting the integrin LFA-1. The rigorous computational and experimental assessment schedule described here is designed to be adaptable to the preclinical discovery and development of novel allosterically acting compounds targeting integrins other than LFA-1, providing an exemplary approach for the early characterisation of next generation integrin inhibitors.

Machine Learning for Fast, Quantum Mechanics-Based Approximation of Drug Lipophilicity

Lipophilicity, as measured by the partition coefficient between octanol and water (log P), is a key parameter in early drug discovery research. However, measuring log P experimentally is difficult for specific compounds and log P ranges. The resulting lack of reliable experimental data impedes development of accurate in silico models for such compounds. In certain discovery projects at Novartis focused on such compounds, a quantum mechanics (QM)-based tool for log P estimation has emerged as a valuable supplement to experimental measurements and as a preferred alternative to existing empirical models. However, this QM-based approach incurs a substantial computational cost, limiting its applicability to small series and prohibiting quick, interactive ideation. This work explores a set of machine learning models (Random Forest, Lasso, XGBoost, Chemprop, and Chemprop3D) to learn calculated log P values on both a public data set and an in-house data set to obtain a computationally affordable, QM-based estimation of drug lipophilicity. The message-passing neural network model Chemprop emerged as the best performing model with mean absolute errors of 0.44 and 0.34 log units for scaffold split test sets of the public and in-house data sets, respectively. Analysis of learning curves suggests that a further decrease in the test set error can be achieved by increasing the training set size. While models directly trained on experimental data perform better at approximating experimentally determined log P values than models trained on calculated values, we discuss the potential advantages of using calculated log P values going beyond the limits of experimental quantitation. We analyze the impact of the data set splitting strategy and gain insights into model failure modes. Potential use cases for the presented models include pre-screening of large compound collections and prioritization of compounds for full QM calculations.

Leveraging molecular structure and bioactivity with chemical language models for de novo drug design

Generative chemical language models (CLMs) can be used for de novo molecular structure generation by learning from a textual representation of molecules. Here, we show that hybrid CLMs can additionally leverage the bioactivity information available for the training compounds. To computationally design ligands of phosphoinositide 3-kinase gamma (PI3Kγ), a collection of virtual molecules was created with a generative CLM. This virtual compound library was refined using a CLM-based classifier for bioactivity prediction. This second hybrid CLM was pretrained with patented molecular structures and fine-tuned with known PI3Kγ ligands. Several of the computer-generated molecular designs were commercially available, enabling fast prescreening and preliminary experimental validation. A new PI3Kγ ligand with sub-micromolar activity was identified, highlighting the method's scaffold-hopping potential. Chemical synthesis and biochemical testing of two of the top-ranked de novo designed molecules and their derivatives corroborated the model's ability to generate PI3Kγ ligands with medium to low nanomolar activity for hit-to-lead expansion. The most potent compounds led to pronounced inhibition of PI3K-dependent Akt phosphorylation in a medulloblastoma cell model, demonstrating efficacy of PI3Kγ ligands in PI3K/Akt pathway repression in human tumor cells. The results positively advocate hybrid CLMs for virtual compound screening and activity-focused molecular design.

Machine Learning and Computational Chemistry for the Endocannabinoid System

Computational methods in medicinal chemistry facilitate drug discovery and design. In particular, machine learning methodologies have recently gained increasing attention. This chapter provides a structured overview of the current state of computational chemistry and its applications for the interrogation of the endocannabinoid system (ECS), highlighting methods in structure-based drug design, virtual screening, ligand-based quantitative structure-activity relationship (QSAR) modeling, and de novo molecular design. We emphasize emerging methods in machine learning and anticipate a forecast of future opportunities of computational medicinal chemistry for the ECS.

[Tumor predisposition syndromes and nephroblastoma : Early diagnosis with imaging]

CLINICAL/METHODICAL ISSUE

The Beckwith-Wiedemann spectrum (BWSp) as well as the WT1-related syndromes, Denys-Drash syndrome (DDS) and WAGR spectrum (Wilms tumor, Aniridia, genitourinary anomalies and a range of developmental delays) are tumor predisposition syndromes (TPS) of Wilms tumor (WT). Patients with associated TPS are at higher risk of developing chronic kidney disease and bilateral and metachronous tumors as well as nephrogenic rests.

STANDARD RADIOLOGICAL METHODS

Standard imaging diagnostics for WT include renal ultrasound and magnetic resonance imaging (MRI). In the current renal tumor studies Umbrella SIOP-RTSG 2016 and Randomet 2017, thoracic computed tomography (CT) is also recommended as standard. Positron emission tomography (PET)-CT and whole-body MRI, on the other hand, are not part of routine diagnostics.

METHODOLOGICAL INNOVATIONS

In recent publications, renal ultrasound is recommended every 3 months until the age of 7 years in cases of clinical suspicion or molecularly proven TPS.

PERFORMANCE

Patients with TPS and regular renal ultrasounds have smaller tumor volumes and lower tumor stages at WT diagnosis than patients without such a screening. This allows a reduction of therapy intensity and facilitates the performance of nephron sparing surgery, which is prognostically relevant especially in bilateral WT.

ACHIEVEMENTS

Early diagnosis of WT in the context of TPS ensures the greatest possible preservation of healthy and functional renal tissue. Standardized screening by regular renal ultrasounds should therefore be firmly established in clinical practice.

PRACTICAL RECOMMENDATIONS

The initial diagnosis of TPS is clinical and requires a skilled and attentive examiner in the presence of sometimes subtle clinical manifestations, especially in the case of BWSp. Clinical diagnosis should be followed by genetic testing, which should then be followed by sonographic screening.

QMugs, quantum mechanical properties of drug-like molecules

Machine learning approaches in drug discovery, as well as in other areas of the chemical sciences, benefit from curated datasets of physical molecular properties. However, there currently is a lack of data collections featuring large bioactive molecules alongside first-principle quantum chemical information. The open-access QMugs (Quantum-Mechanical Properties of Drug-like Molecules) dataset fills this void. The QMugs collection comprises quantum mechanical properties of more than 665 k biologically and pharmacologically relevant molecules extracted from the ChEMBL database, totaling ~2 M conformers. QMugs contains optimized molecular geometries and thermodynamic data obtained via the semi-empirical method GFN2-xTB. Atomic and molecular properties are provided on both the GFN2-xTB and on the density-functional levels of theory (DFT, ωB97X-D/def2-SVP). QMugs features molecules of significantly larger size than previously-reported collections and comprises their respective quantum mechanical wave functions, including DFT density and orbital matrices. This dataset is intended to facilitate the development of models that learn from molecular data on different levels of theory while also providing insight into the corresponding relationships between molecular structure and biological activity.

Measurement(s)	Quantum Mechanics
Technology Type(s)	density functional theory

TBIO-08. The molecular basis for rational targeting of FGFR-driven growth and invasiveness in pediatric brain tumors

The oncogenic activation of receptor tyrosine kinases (RTK) promotes growth, survival and dissemination in pediatric tumors including glioma, ependymoma and medulloblastoma (MB). Direct targeting of either the RTK or of downstream kinases can effectively block tumor promoting pathway functions. However, emergence of resistance is common. We hypothesized that alternative interference strategies that target protein-protein interactions (PPIs) instead of enzymatic activities could overcome the emergence of resistance. We characterized the molecular interactions downstream of the FGFR that regulate relevant growth and invasion-promoting mechanisms in MB cells, to identify potentially druggable PPIs. We found that the FRS2 protein is an essential up-stream effector of FGFR signaling towards invasiveness. Using a proteomics approach, we furthermore identified the Striatin 3 protein as a novel oncogenic effector of the FGFR pathway downstream of FRS2, as it integrates antagonistic growth and invasion signals downstream of FGFR. Mechanistically, Striatin 3 interacts with the Ser/Thr kinase MAP4K4, couples it to the protein phosphatase 2A, and thereby inactivates growth repressing activities of MAP4K4. In parallel, Striatin 3 enables MAP4K4-mediated phosphorylation of PKC-theta and VASP, which combined are necessary to promote tissue invasion. To selectively repress pro-invasive FGFR functions, we identified and functionally validated small molecule ligands of FRS2, that prevent FRS2 activation and downstream signaling. We demonstrate efficacy of these compounds in inhibiting invasion and growth promoting activities in vitro and in vivo, and identified potential off-target activities of the ligand using a proteome-wide interaction analysis. We propose inhibition of FRS2 by a small molecular PTB domain ligand as a strategy to repress FGF signaling in FGFR-driven tumors. The development of this ligand, and the de novo design of functional analogs thereof bear promise for further pre-clinical evaluation of these structures as anti-growth promoting and anti-metastatic therapeutics applicable to FGFR-driven tumors.

Generating Bioactive Natural Product-inspired Molecules with Machine Intelligence

The computer-assisted design of new chemical entities has made a leap forward with the development of machine learning models for automated molecule generation. The overarching goal of this conceptual approach is to augment the creativity of medicinal chemists with a machine intelligence. In this Perspective we highlight prospective applications of “de novo” drug design and target prediction, aiming to generate natural product-inspired bioactive compounds from scratch. A virtual chemist transforms pharmacologically active natural products into new, easily synthesizable small molecules with desired properties and activity. Computational activity prediction and automated compound generation offer the possibility to systematically transfer the wealth of pharmaceutically active natural products to synthetic small molecule drug discovery. We present selected prospective examples and dare a forecast into the future of natural product-inspired drug discovery.

Translating from proteins to ribonucleic acids for ligand-binding site detection

Identifying druggable ligand-binding sites on the surface of the macromolecular targets is an important process in structure-based drug discovery. Deep-learning models have been shown to successfully predict ligand-binding sites of proteins. As a step toward predicting binding sites in RNA and RNA-protein complexes, we employ three-dimensional convolutional neural networks. We introduce a dataset splitting approach to minimize structure-related bias in training data, and investigate the influence of protein-based neural network pre-training before fine-tuning on RNA structures. Models that were pre-trained on proteins considerably outperformed the models that were trained exclusively on RNA structures. Overall, 71% of the known RNA binding sites were correctly located within 4 Å of their true centres with a structural overlap of at least 25%.

Δ-Quantum machine-learning for medicinal chemistry

Many molecular design tasks benefit from fast and accurate calculations of quantum-mechanical (QM) properties. However, the computational cost of QM methods applied to drug-like molecules currently renders large-scale applications of quantum chemistry challenging. Aiming to mitigate this problem, we developed DelFTa, an open-source toolbox for the prediction of electronic properties of drug-like molecules at the density functional (DFT) level of theory, using Δ-machine-learning. Δ-Learning corrects the prediction error (Δ) of a fast but inaccurate property calculation. DelFTa employs state-of-the-art three-dimensional message-passing neural networks trained on a large dataset of QM properties. It provides access to a wide array of quantum observables on the molecular, atomic and bond levels by predicting approximations to DFT values from a low-cost semiempirical baseline. Δ-Learning outperformed its direct-learning counterpart for most of the considered QM endpoints. The results suggest that predictions for non-covalent intra- and intermolecular interactions can be extrapolated to larger biomolecular systems. The software is fully open-sourced and features documented command-line and Python APIs.

X-ray Fourier transform holography with beam shaping optical elements

Holography is a powerful method for achieving 3D images of objects. Extending this method to short wavelengths potentially offers significantly higher resolution than visible light holography. However, current X-ray holography setups employ nanoscale pinholes to form the reference wave. This approach is relatively inefficient and limited to very small sample size. Here, we propose a new setup for X-ray holography based on a binary diffractive optical element (DOE), which forms at the same time the object illumination and the reference wave. This optic is located separately from the sample plane, which permits investigation of larger sample areas. Using an extended test sample, we demonstrate a resolution of 90 nm (half-pitch) at an undulator beamline at BESSY II. The new holography setup can be directly transferred to free electron laser sources enabling time-resolved nanoscale imaging for ultra-fast processes.

Perplexity-Based Molecule Ranking and Bias Estimation of Chemical Language Models

Chemical language models (CLMs) can be employed to design molecules with desired properties. CLMs generate new chemical structures in the form of textual representations, such as the simplified molecular input line entry system (SMILES) strings. However, the quality of these de novo generated molecules is difficult to assess a priori. In this study, we apply the perplexity metric to determine the degree to which the molecules generated by a CLM match the desired design objectives. This model-intrinsic score allows identifying and ranking the most promising molecular designs based on the probabilities learned by the CLM. Using perplexity to compare "greedy" (beam search) with "explorative" (multinomial sampling) methods for SMILES generation, certain advantages of multinomial sampling become apparent. Additionally, perplexity scoring is performed to identify undesired model biases introduced during model training and allows the development of a new ranking system to remove those undesired biases.

De Novo Molecular Design with Chemical Language Models

Artificial intelligence (AI) offers new possibilities for hit and lead finding in medicinal chemistry. Several instances of AI have been used for prospective de novo drug design. Among these, chemical language models have been shown to perform well in various experimental scenarios. In this study, we provide a hands-on introduction to chemical language modeling. A technique based on recurrent neural networks is discussed in detail, together with a step-by-step guide to applying this AI method for focused compound library design. The program code is freely available at URL: github.com/ETHmodlab/de_novo_design_RNN .

Reflection on the proposed changes to dose quantities-an industrial perspective

In 2021, the ICRP initiated the revision of the general recommendations of the system of radiation protection, and part of it will focus on dose quantities. The recently published ICRP Publication 147 and ICRU Report 95 have described the extent of the proposed modifications and paved the way for the strategy to be adopted. These revisions would seek to simplify, improve the accuracy and extend the field of use of dose quantities. While the Radiological Protection Working Group of the World Nuclear Association recognises the notable improvement in the estimation of the protection quantities and the usefulness of such changes for the medical and research sector, the benefits of the proposed new system seem very limited for the nuclear industry and industries involving naturally occurring radioactive materials. The complexity associated with changing a long-standing and robust system and the risk incurred by the human factor seem unjustified, bearing in mind the likely cost.

Computer-Aided Design and Synthesis of a New Class of PEX14 Inhibitors: Substituted 2,3,4,5-Tetrahydrobenzo[F][1,4]oxazepines as Potential New Trypanocidal Agents

African and American trypanosomiases are estimated to affect several million people across the world, with effective treatments distinctly lacking. New, ideally oral, treatments with higher efficacy against these diseases are desperately needed. Peroxisomal import matrix (PEX) proteins represent a very interesting target for structure- and ligand-based drug design. The PEX5-PEX14 protein-protein interface in particular has been highlighted as a target, with inhibitors shown to disrupt essential cell processes in trypanosomes, leading to cell death. In this work, we present a drug development campaign that utilizes the synergy between structural biology, computer-aided drug design, and medicinal chemistry in the quest to discover and develop new potential compounds to treat trypanosomiasis by targeting the PEX14-PEX5 interaction. Using the structure of the known lead compounds discovered by Dawidowski et al. as the template for a chemically advanced template search (CATS) algorithm, we performed scaffold-hopping to obtain a new class of compounds with trypanocidal activity, based on 2,3,4,5-tetrahydrobenzo[f][1,4]oxazepines chemistry. The initial compounds obtained were taken forward to a first round of hit-to-lead optimization by synthesis of derivatives, which show activities in the range of low- to high-digit micromolar IC₅₀ in the in vitro tests. The NMR measurements confirm binding to PEX14 in solution, while immunofluorescent microscopy indicates disruption of protein import into the glycosomes, indicating that the PEX14-PEX5 protein-protein interface was successfully disrupted. These studies result in development of a novel scaffold for future lead optimization, while ADME testing gives an indication of further areas of improvement in the path from lead molecules toward a new drug active against trypanosomes.

Bioaffinity Screening with a Rapid and Sample-Efficient Autosampler for Native Electrospray Ionization Mass Spectrometry

Fast and efficient handling of ligands and biological targets are required in bioaffinity screening based on native electrospray ionization mass spectrometry (ESI-MS). We use a prototype microfluidic autosampler, called the "gap sampler", to sequentially mix and electrospray individual small molecule ligands together with a target protein and compare the screening results with data from thermal shift assay and surface plasmon resonance. In a first round, all three techniques were used for a screening of 110 ligands against bovine carbonic anhydrase II, which resulted in five mutual hits and some false positives with ESI-MS presumably due to the high ligand concentration or interferences from dimethyl sulfoxide. In a second round, 33 compounds were screened in lower concentrations and in a less complex matrix, resulting in only true positives with ESI-MS. Within a cycle time of 30 s, dissociation constants were determined within an order of magnitude accuracy consuming only 5 pmol of ligand and less than 15 pmol of protein per screened compound. In a third round, dissociation constants of five compounds were accurately determined in a titration experiment. Thus, the gap sampler can rapidly and efficiently be used for high-throughput screening.

Reflections on low-dose radiation, the misconceptions, reality and moving forward

Low dose radiation has been widely accepted by the radiation protection community as presenting a very low risk to human health, if any. Over-conservatism in optimisation principles and regulations have resulted in a disproportionate fear of radiation amongst the general public and government authorities alike, overlooking the great benefits nuclear science and techniques have brought to society as a whole. As such, the World Nuclear Association advocates for a recontextualisation of the radiation hazards with regards to low dose radiation, and a greater awareness as to the absence of any discernible effects associated with it.

Beam Search for Automated Design and Scoring of Novel ROR Ligands with Machine Intelligence*

Chemical language models enable de novo drug design without the requirement for explicit molecular construction rules. While such models have been applied to generate novel compounds with desired bioactivity, the actual prioritization and selection of the most promising computational designs remains challenging. Herein, we leveraged the probabilities learnt by chemical language models with the beam search algorithm as a model-intrinsic technique for automated molecule design and scoring. Prospective application of this method yielded novel inverse agonists of retinoic acid receptor-related orphan receptors (RORs). Each design was synthesizable in three reaction steps and presented low-micromolar to nanomolar potency towards RORγ. This model-intrinsic sampling technique eliminates the strict need for external compound scoring functions, thereby further extending the applicability of generative artificial intelligence to data-driven drug discovery.

A critical overview of computational approaches employed for COVID-19 drug discovery

COVID-19 has resulted in huge numbers of infections and deaths worldwide and brought the most severe disruptions to societies and economies since the Great Depression. Massive experimental and computational research effort to understand and characterize the disease and rapidly develop diagnostics, vaccines, and drugs has emerged in response to this devastating pandemic and more than 130 000 COVID-19-related research papers have been published in peer-reviewed journals or deposited in preprint servers. Much of the research effort has focused on the discovery of novel drug candidates or repurposing of existing drugs against COVID-19, and many such projects have been either exclusively computational or computer-aided experimental studies. Herein, we provide an expert overview of the key computational methods and their applications for the discovery of COVID-19 small-molecule therapeutics that have been reported in the research literature. We further outline that, after the first year the COVID-19 pandemic, it appears that drug repurposing has not produced rapid and global solutions. However, several known drugs have been used in the clinic to cure COVID-19 patients, and a few repurposed drugs continue to be considered in clinical trials, along with several novel clinical candidates. We posit that truly impactful computational tools must deliver actionable, experimentally testable hypotheses enabling the discovery of novel drugs and drug combinations, and that open science and rapid sharing of research results are critical to accelerate the development of novel, much needed therapeutics for COVID-19.

Learning from Nature: From a Marine Natural Product to Synthetic Cyclooxygenase-1 Inhibitors by Automated De Novo Design

The repertoire of natural products offers tremendous opportunities for chemical biology and drug discovery. Natural product-inspired synthetic molecules represent an ecologically and economically sustainable alternative to the direct utilization of natural products. De novo design with machine intelligence bridges the gap between the worlds of bioactive natural products and synthetic molecules. On employing the compound Marinopyrrole A from marine Streptomyces as a design template, the algorithm constructs innovative small molecules that can be synthesized in three steps, following the computationally suggested synthesis route. Computational activity prediction reveals cyclooxygenase (COX) as a putative target of both Marinopyrrole A and the de novo designs. The molecular designs are experimentally confirmed as selective COX-1 inhibitors with nanomolar potency. X-ray structure analysis reveals the binding of the most selective compound to COX-1. This molecular design approach provides a blueprint for natural product-inspired hit and lead identification for drug discovery with machine intelligence.

POS0091 OFF-TARGET PROFILING OF JANUS KINASE (JAK) INHIBITORS IN RHEUMATOID ARTHRITIS: A COMPUTER-BASED APPROACH FOR DRUG SAFETY STUDIES AND REPURPOSING

Background:

The JAK inhibitors (JAKi’s) tofacitinib and baricitinib are new alternatives for treating rheumatoid arthritis. Safety concerns associated with JAKi’s, such as the increased risk for thrombosis and viral infections, have emerged worldwide. The underlying explanatory mechanisms remain unknown, suggesting the elevated risk is likely due to underlying confounding or an off-target binding effect. Computational approaches can explore the potential for a small molecule drug to interact with previously unknown biological targets and identify potential safety-related concerns, and open doors for potential drug repurposing.

Objectives:

To identify and characterize the off-target binding effects of baricitinib and tofacitinib, with a focus on targets related to thrombosis and viral infection

Methods:

Potential targets of baricitinib and tofacitinib were predicted using two neural-network-based systems (TIGER[1] and SPiDER[2]). Targets were considered relevant if they had (1) a SPiDER confidence with p<0.05, or (2) a TIGER score >1. Selected targets related to the outcome of interest were experimentally evaluated at Eurofins Cerep (France-Celle L’Evescault, www.eurofins.com) if commercial available. Compounds were tested at (1) single concentration (30 µM) with technical replicates, using radioligand or enzymatic assays, or (2) multiple concentrations (30 µM highest concentration; dilution factor in a log-scale) with technical replicates, using calcium flux or inhibition of [cAMP] assays. Observed activity of ≥50% inhibition or stimulation on the target was considered active, between 25 to 50% inhibition (or a dissociation constant [Kd] from 1 to 10 µM) was considered as moderate activity, and lower than 25% was considered inactive. Dose-response curve were performed on active and moderate targets for IC50 / EC50 (half maximal inhibitory / effective concentration) determination.

Results:

TIGER and SPiDER suggested a total of 99 off-target binding effects (baricitinib n=41; tofacitinib n=58), of which 17 targets had potential impact on thrombosis or viral infection (baricitinib n=5 and 4, respectively; tofacitinib n=5 and 3, respectively). Commercial testing was available on 11 targets (Adenosine Receptor A2A [AA2AR], Epidermal growth factor receptor, induclible NOS, PI3 Kinase (p110b/p85a), Phosphodiesterase 10A2 [PDE10A2] and Protein Kinase N2 [PKN2] for baricitinib; and Adenosine receptor A3, 15-Lipoxygenase [15-LO], PKN2, Transient receptor potential cation channel [TRPM6] and AA2AR for tofacitinib). Of these, 5 targets showed active or moderately active binding activity (baricitinib n=2; tofacitinib n=3), and were tested for dose-response curves. Test results confirmed ligand-binding activity with IC50 on nanomolar (PKN2), and micromolar ranges (PDE10A2 and TRPM6).

Conclusion:

The results suggest both baricitinib and tofacitinib are promiscuous binders with effects on several families. Although it may lead to side effects, off-target binding also represents a potential opportunity for drug repurposing. Besides on-target effects, both drugs are under clinical investigation for the treatment of COVID-19 due to off-target interactions. The proposed pharmacological off-target effects of those with active binding include attenuation of pulmonary vascular remodeling, anti-fibrotic and anti-psychotic activities (PDE10A2), modulation of viral response (PKN2), and hypomagnesaemia (TRPM6), which is involved in cardiovascular diseases. This study supports tofacitinib and baricitinib as candidates for drug repurposing (e.g., in COVID-19, Hepatitis C virus, and pulmonary hypertension). We did not identify active off-target interactions linked to thrombosis to explain the elevated risk observed in clinical practice. Further research is required to elucidate the underlying patient-specific factors (confounders) that could explain this safety concern.

References:

[1]Schneider P et al. Angew Chem Int Ed 2017;56:11520–4.

[2]Reker D et al. PNAS 2014;111:4067–72.

Disclosure of Interests:

None declared

Combining generative artificial intelligence and on-chip synthesis for de novo drug design

Automating the molecular design-make-test-analyze cycle accelerates hit and lead finding for drug discovery. Using deep learning for molecular design and a microfluidics platform for on-chip chemical synthesis, liver X receptor (LXR) agonists were generated from scratch. The computational pipeline was tuned to explore the chemical space of known LXRα agonists and generate novel molecular candidates. To ensure compatibility with automated on-chip synthesis, the chemical space was confined to the virtual products obtainable from 17 one-step reactions. Twenty-five de novo designs were successfully synthesized in flow. In vitro screening of the crude reaction products revealed 17 (68%) hits, with up to 60-fold LXR activation. The batch resynthesis, purification, and retesting of 14 of these compounds confirmed that 12 of them were potent LXR agonists. These results support the suitability of the proposed design-make-test-analyze framework as a blueprint for automated drug design with artificial intelligence and miniaturized bench-top synthesis.

Artificial intelligence in drug discovery: recent advances and future perspectives

Introduction: Artificial intelligence (AI) has inspired computer-aided drug discovery. The widespread adoption of machine learning, in particular deep learning, in multiple scientific disciplines, and the advances in computing hardware and software, among other factors, continue to fuel this development. Much of the initial skepticism regarding applications of AI in pharmaceutical discovery has started to vanish, consequently benefitting medicinal chemistry.Areas covered: The current status of AI in chemoinformatics is reviewed. The topics discussed herein include quantitative structure-activity/property relationship and structure-based modeling, de novo molecular design, and chemical synthesis prediction. Advantages and limitations of current deep learning applications are highlighted, together with a perspective on next-generation AI for drug discovery.Expert opinion: Deep learning-based approaches have only begun to address some fundamental problems in drug discovery. Certain methodological advances, such as message-passing models, spatial-symmetry-preserving networks, hybrid de novo design, and other innovative machine learning paradigms, will likely become commonplace and help address some of the most challenging questions. Open data sharing and model development will play a central role in the advancement of drug discovery with AI.

Coloring Molecules with Explainable Artificial Intelligence for Preclinical Relevance Assessment

Graph neural networks are able to solve certain drug discovery tasks such as molecular property prediction and de novo molecule generation. However, these models are considered "black-box" and "hard-to-debug". This study aimed to improve modeling transparency for rational molecular design by applying the integrated gradients explainable artificial intelligence (XAI) approach for graph neural network models. Models were trained for predicting plasma protein binding, hERG channel inhibition, passive permeability, and cytochrome P450 inhibition. The proposed methodology highlighted molecular features and structural elements that are in agreement with known pharmacophore motifs, correctly identified property cliffs, and provided insights into unspecific ligand-target interactions. The developed XAI approach is fully open-sourced and can be used by practitioners to train new models on other clinically relevant endpoints.

Interaction analysis of glycoengineered antibodies with CD16a: a native mass spectrometry approach

Minor changes in the quality of biologically manufactured monoclonal antibodies (mAbs) can affect their bioactivity and efficacy. One of the most important variations concerns the N-glycosylation pattern, which directly affects an anti-tumor mechanism called antibody-dependent cell-meditated cytotoxicity (ADCC). Thus, careful engineering of mAbs is expected to enhance both protein-receptor binding and ADCC. The specific aim of this study is to evaluate the influence of terminal carbohydrates within the Fc region on the interaction with the FcγRIIIa/CD16a receptor in native and label-free conditions. The single mAb molecule comprises variants with minimal and maximal galactosylation, as well as α2,3 and α2,6-sialic acid isomers. Here, we apply native electrospray ionization mass spectrometry to determine the solution-phase antibody-receptor equilibria and by using temperature-controlled nanoelectrospray, a thermal stability of the complex is examined. Based on these, we prove that the galactosylation of a fucosylated Fc region increases the binding to CD16a 1.5-fold when compared with the non-galactosylated variant. The α2,6-sialylation has no significant effect on the binding, whereas the α2,3-sialylation decreases it 1.72-fold. In line with expectation, the galactoslylated and α2,6-sialylated mAb:CD16a complex exhibit higher thermal stability when measured in the temperature gradient from 20 to 50°C. The similar binding pattern is observed based on surface plasmon resonance analysis and immunofluorescence staining using natural killer cells. The results of our study provide new insight into N-glycosylation-based interaction of the mAb:CD16a complex.

Molecular Scaffold Hopping via Holistic Molecular Representation

Molecular descriptors encode a variety of molecular representations for computer-assisted drug discovery. Here, we focus on the Weighted Holistic Atom Localization and Entity Shape (WHALES) descriptors, which were originally designed for scaffold hopping from natural products to synthetic molecules. WHALES descriptors capture molecular shape and partial charges simultaneously. We introduce the key aspects of the WHALES concept and provide a step-by-step guide on how to use these descriptors for virtual compound screening and scaffold hopping. The results presented can be reproduced by using the code freely available from URL: github.com/ETHmodlab/scaffold_hopping_whales .

MODL-14. SMALL MOLECULE TARGETING OF ONCOGENIC FGF2-FGFR SIGNALING IN BRAIN TUMORS

FGF2, the ligand of FGF receptors (FGFRs), is expressed in the developing and adult brain. FGF2-FGFR1 signaling causes the induction and maintenance of cancer stem cells through ERK-dependent up-regulation of ZEB1 and Olig2 in glioblastoma. In SHH medulloblastoma, Olig2 triggers tumor initiation from GCPs, maintains quiescent stem-like cells during the disease and contributes to tumor outgrowth at recurrence. We found that FGF2-FGFR signaling causes increased growth and tissue invasion through the FGFR adaptor protein FRS2 in SHH and group-3 medulloblastoma ¹. Thus, targeting of FGFR-FRS2 signaling could abrogate brain tumor growth and spread by repressing tumor-promoting functions that are induced by microenvironmental FGF2. Using virtual screening combined with functional validation, we identified protein-protein interaction inhibitors (F2i) that bind FRS2 and abrogate FGFR signaling to the MAP-ERK pathway. Consistent with the requirement of FRS2 for pro-invasive signaling downstream of FGFR1 in medulloblastoma, F2i also efficiently block FGF2-induced migration and invasion in medulloblastoma-derived cells. Selected F2i display excellent binding kinetics with a similar Kd as the natural ligand domain of FGFR and cause steric alterations in the targeted protein domain. On-target activity was confirmed by thermal proteome profiling. Neither in silico screening nor empirical testing revealed significant off-target activity of the compounds. No toxicity of F2i was observed in cell-based models with confirmed functional activity on invasion and MAPK activation. Thus, we identified novel, low molecular weight pharmacological protein-protein interaction inhibitors with an excellent potential to specifically block FGFR functions relevant for brain tumor progression. 1. Santhana Kumar et al., CellReports23, 3798–3812.e8 (2018).

The burden in chronic prurigo: patients with chronic prurigo suffer more than patients with chronic pruritus on non-lesional skin: A comparative, retrospective, explorative statistical analysis of 4,484 patients in a real-world cohort

BACKGROUND

Chronic prurigo (CPG) is known as a high burdensome disease characterized by severe pruritus and multiple pruriginous lesions. Interestingly, the disease-specific burden is not well established and there are no data which compare the impact of CPG with chronic pruritus (CP) on non-lesional skin (CP-NL).

OBJECTIVES

To address this issue, we analysed datasets from 4484 patients with either CPG or CP-NL.

METHODS

Demographic medical data and additional information collected by validated patient reported outcome tools were analysed. The visual analogue scale and numerical rating scale (NRS) were used for assessing the pruritus intensity, the ItchyQoL for patients' quality of life, the Hospital Anxiety and Depression Scale and the Patient Needs Questionnaire' as a part of Patient Benefit Index for Pruritus for measuring the importance of 27 patient needs in terms of treatment goals. The Neuroderm questionnaire was used to assess the history of pruritus characteristics and the impact on sleep.

RESULTS

Patients with CPG suffered longer and with a higher intensity from pruritus [NRS worst the last 24 h, CPG 6.0 (4.0;8.0) vs. CP-NL 3.0 (5.0;7.0), P < 0.001]. In them, pruritus occurred more often and the whole day and night which led to more loss in sleeping hours [CPG 3.0 h (2.0;4.0) vs. CP-NL 2.0 h (1.0;4.0), P < 0.001]. Patients with CPG showed higher scores for depression [HADS-D, CPG 6.0 (3.0;10.0) vs. CP-NL 5.0 (2.0;8.0), P < 0.001], more impaired quality of life [ItchyQol; CPG: 72.6 (61.6;83.6) vs. CP-NL 59.4 (48.4;70.4), P < 0.001] and higher weighted needs in the predefined treatment goals.

DISCUSSION

Not only the presence of severe pruritus and pruriginous lesions but also sleep disorders and other mental symptoms may contribute to a higher burden in patients with CPG when compared with patients with CP-NL.

Engineering of a functional γ-tocopherol transfer protein

α-tocopherol transfer protein (TTP) was previously reported to self-aggregate into 24-meric spheres (α-TTP_S) and to possess transcytotic potency across mono-layers of human umbilical vein endothelial cells (HUVECs). In this work, we describe the characterisation of a functional TTP variant with its vitamer selectivity shifted towards γ-tocopherol. The shift was obtained by introducing an alanine to leucine substitution into the substrate-binding pocket at position 156 through site directed mutagenesis. We report here the X-ray crystal structure of the γ-tocopherol specific particle (γ-TTP_S) at 2.24 Å resolution. γ-TTP_S features full functionality compared to its α-tocopherol specific parent including self-aggregation potency and transcytotic activity in trans-well experiments using primary HUVEC cells. The impact of the A156L mutation on TTP function is quantified in vitro by measuring the affinity towards γ-tocopherol through micro-differential scanning calorimetry and by determining its ligand-transfer activity. Finally, cell culture experiments using adherently grown HUVEC cells indicate that the protomers of γ-TTP, in contrast to α-TTP, do not counteract cytokine-mediated inflammation at a transcriptional level. Our results suggest that the A156L substitution in TTP is fully functional and has the potential to pave the way for further experiments towards the understanding of α-tocopherol homeostasis in humans.

Central hemodynamic effects in patients with chronic coronary syndrome after long-term ivabradine therapy

Objectives

We sought to assess central hemodynamic effects in 23 patients (18 male, 5 female) with a resting heart rate (HR) of ≥70 beats per minute (bpm) and chronic coronary syndrome after long-term ivabradine therapy (6 months) by cardiac magnetic resonance (CMR).

Methods and results

In a cross-over design, all patients were treated with ivabradine (Iva, 7.5 mg bid) and placebo for 6 months each. CMR was performed three times (at baseline, after 6 and 12 months) to determine left ventricular (LV) function parameters, including end-diastolic and end-systolic volumes (EDVi, ESVi), stroke volume (SVi) and ejection fraction (EF) as well as volume-time curve (VTC) parameters, including peak ejection rate (PER), peak ejection time (PET), peak filling rate (PFR), peak filling time from ES (PFT), peak ejection rate normalized to EDV (PER/EDV) and peak filling rate normalized to EDV (PFR/EDV) for global LV function (systolic and diastolic) assessment. Flow measurements of the ascending aorta were performed with phase-contrast velocity imaging.

Treatment with Iva led to a HR reduction of 11.4 bpm (Iva 58.8±8.2 bpm vs placebo 70.2±8.3 bpm, p&lt;0.0001).There was no difference in LVEF (%) (Iva 57.4±11.2 vs placebo 53.0±10.9, p=0.18), EDVi or ESVi. SVi (ml/m2) remained comparatively unchanged after long-term treatment with Iva (Iva 40.6±9.6 vs placebo 35.7±8.8, p=0.08). VTC parameters reflecting systolic LV function (PER, PET) were unaffected by Iva, while both PFR and PFR/EDV were significantly increased (PFR/EDV (s-1) Iva 2.4±0.4 vs placebo 2.1±0.4, p=0.03). There was a trend to longer PFT during treatment with Iva, though not reaching statistical significance. Medium and maximum aortic flow were not affected by treatment with Iva, while mean velocity (cm/s) was significantly reduced (Iva 6.7±2.7 vs placebo 9.0±3.4, p=0.01). Aortic flow parameters were correlated to aortic distensibility (AD), as surrogate parameter for arterial stiffness. AD was significantly correlated to both aortic flow and flow velocity, whereby mean velocity showed the strongest correlation to AD (r=0.74 [0.61 to 0.83], p&lt;0.0001).

Conclusion

Systolic LV function was unaffected by treatment with Iva, while the filling during diastole was significantly improved. While medium and maximum aortic flow were not affected by Iva, mean velocity was significantly reduced. Aortic distensibility as surrogate parameter for arterial stiffness was significantly correlated to aortic mean velocity. This study confirms the underlying physiological principle of the If-current inhibitor Ivabradine.

Funding Acknowledgement

Type of funding source: Public Institution(s). Main funding source(s): This work was supported by the Deutsche Herzstiftung (German Heart Foundation) (F/14/11 to F.C.) and the Deutsche Forschungsgemeinschaft (DFG KFO 196 to U.L., S.H.S and M.B. and SFB TTR 219, S-01 to M.B.). The Saarland University Medical Center has received an unrestricted grant from Servier (France).

Morphing of Amphipathic Helices to Explore the Activity and Selectivity of Membranolytic Antimicrobial Peptides

Naturally occurring membranolytic antimicrobial peptides (AMPs) are rarely cell-type selective and highly potent at the same time. Template-based peptide design can be used to generate AMPs with improved properties de novo. Following this approach, 18 linear peptides were obtained by computationally morphing the natural AMP Aurein 2.2d2 GLFDIVKKVVGALG into the synthetic model AMP KLLKLLKKLLKLLK. Eleven of the 18 chimeric designs inhibited the growth of Staphylococcus aureus, and six peptides were tested and found to be active against one resistant pathogenic strain or more. One of the peptides was broadly active against bacterial and fungal pathogens without exhibiting toxicity to certain human cell lines. Solution nuclear magnetic resonance and molecular dynamics simulation suggested an oblique-oriented membrane insertion mechanism of this helical de novo peptide. Temperature-resolved circular dichroism spectroscopy pointed to conformational flexibility as an essential feature of cell-type selective AMPs.

Scanning acoustic microscopy investigation of weld lines in injection-molded parts manufactured from industrial thermoplastic polymer

Scanning acoustic microscopy (SAM) is used to characterize welds in a thermoplastic polymer (ABS) manufactured by injection-molding, particularly at the locations of weld-lines known to form as unavoidable significant defects. Acoustic micrographs obtained at 420 MHz clearly resolve the weld lines with morphological deformations and microelastic heterogenity. This is also where terahertz (THz) measurements, carried out in support of the SAM study, reveal enhanced birefringence corresponding to the location of these lines enabling verification of the SAM results. Rayleigh surface acoustic waves (RSAW), quantified by V(z) curves (with defocusing distance of 85 μm), are found to propagate slower in regions close to the weld lines than in regions distant from these lines. The discrepancy of about 100 m/s in the velocity of RSAW indicates a large variation in the micro-elastic properties between areas close to and distant from the weld lines. The spatial variations in velocity (V_R) of RSAWs indicate anisotropic propagation of the differently polarized ultrasonic waves.