Systems biology approaches for advancing the discovery of effective drug combinations

A Multi-View Feature-Based Interpretable Deep Learning Framework for Drug-Drug Interaction Prediction

Drug-drug interactions (DDIs) can result in deleterious consequences when patients take multiple medications simultaneously, emphasizing the critical need for accurate DDI prediction. Computational methods for DDI prediction have garnered recent attention. However, current approaches concentrate solely on single-view features, such as atomic-view or substructure-view features, limiting predictive capacity. The scarcity of research on interpretability studies based on multi-view features is crucial for tracing interactions. Addressing this gap, we present MI-DDI, a multi-view feature-based interpretable deep learning framework for DDI. To fully extract multi-view features, we employ a Message Passing Neural Network (MPNN) to learn atomic features from molecular graphs generated by RDkit, and transformer encoders are used to learn substructure-view embeddings from drug SMILES simultaneously. These atomic-view and substructure-view features are then amalgamated into a holistic drug embedding matrix. Subsequently, an intricately designed interaction module not only establishes a tractable path for understanding interactions but also directly informs the construction of weight matrices, enabling precise and interpretable interaction predictions. Validation on the BIOSNAP dataset and DrugBank dataset demonstrates MI-DDI's superiority. It surpasses the current benchmarks by a substantial average of 3% on BIOSNAP and 1% on DrugBank. Additional experiments underscore the significance of atomic-view information for DDI prediction and confirm that our interaction module indeed learns more effective information for DDI prediction. The source codes are available at https://github.com/ZihuiCheng/MI-DDI .

A Domain Adaptive Interpretable Substructure-Aware Graph Attention Network for Drug-Drug Interaction Prediction

Accurate prediction of drug-drug interaction (DDI) is essential to improve clinical efficacy, avoid adverse effects of drug combination therapy, and enhance drug safety. Recently researchers have developed several computer-aided methods for DDI prediction. However, these methods lack the substructural features that are critical to drug interactions and are not effective in generalizing across domains and different distribution data. In this work, we present SAGAN, a domain adaptive interpretable substructure-aware graph attention network for DDI prediction. Based on attention mechanism and unsupervised clustering algorithm, we propose a new substructure segmentation method, which segments the drug molecule into multiple substructures, learns the mechanism of drug interaction from the perspective of interaction, and identifies important interaction regions between drugs. To enhance the generalization ability of the model, we improve and apply a conditional domain adversarial network to achieve cross-domain generalization by alternately optimizing the cross-entropy loss on the source domain and the adversarial loss of the domain discriminator. We evaluate and compare SAGAN with the state-of-the-art DDI prediction model on four real-world datasets for both in-domain and cross-domain scenarios, and show that SAGAN achieves the best overall performance. Moreover, the visualization results of the model show that SAGAN has achieved pharmacologically significant substructure extraction, which can help drug developers screen for some undiscovered local interaction sites, and provide important information for further drug structure optimization. The codes and datasets are available online at https://github.com/wyx2012/SAGAN .

Multifunctional Nanoparticles as Radiosensitizers to Overcome Hypoxia-Associated Resistance in Cancer Radiotherapy

Hypoxia, a phenomenon that occurs when the oxygen level in tissues is lower than average, is commonly observed in human solid tumors. For oncological treatment, the hypoxic environment often results in radioresistance and chemoresistance. In this study, a new multifunctional oxygen carrier, carboxymethyl hexanoyl chitosan (CHC) nanodroplets decorated with perfluorohexane (PFH) and superparamagnetic iron oxide (SPIO) nanodroplets (SPIO@PFH-CHC), was developed and investigated. PFH-based oxygen carriers can augment oxygenation within tumor tissues, thereby mitigating radioresistance. Concurrently, oxygenation can cause deoxyribonucleic acid (DNA) damage via oxygen fixation and consequently suppress cancer cell proliferation. Moreover, these pH-sensitive nanodroplets allow higher cellular uptake with minimal cytotoxicity. Two distinctive mechanisms of SPIO@PFH-CHC nanodroplets were found in this study. The SPIO nanoparticles of the SPIO@PFH-CHC nanodroplets can generate hydroxyl radicals (HO•) and other reactive oxygen species (ROS), which is vital to chemodynamic therapy (CDT) via the Fenton reaction. Meanwhile, the higher X-ray absorption among these nanodroplets leads to a local energy surge and causes more extensive deoxyribonucleic acid (DNA) damage via oxygen fixation. This study demonstrates that low cytotoxic SPIO@PFH-CHC nanodroplets can be an efficient radiosensitizer for radiation therapy.

Advancing analytical and graphical methods for binary and ternary mixtures: The toxic interactions of divalent metal ions in human lung cells

Humans are exposed to various environmental chemicals, particles, and pathogens that can cause adverse health outcomes. These exposures are rarely homogenous but rather complex mixtures in which the components may interact, such as through synergism or antagonism. Toxicologists have conducted preliminary investigations into binary mixtures of two components, but little work has been done to understand mixtures of three or more components. We investigated mixtures of divalent metal ions, quantifying the toxic interactions in a human lung model. Eight metals were chosen: heavy metals cadmium, copper, lead, and tin, as well as transition metals iron, manganese, nickel, and zinc. Human alveolar epithelial cells (A549) were exposed to individual metals and sixteen binary and six ternary combinations. The dose-response was modeled using logistic regression in R to extract LC50 values. Among the individual metals, the highest and lowest toxicity were observed with copper at an LC50 of 102 μM and lead at an LC50 of 5639 μM, respectively. First and second-order interaction coefficients were obtained using machine learning-based linear regression in Python. The resulting second-degree polynomial model formed either a hyperbolic or elliptical conic section, and the positive quadrant was used to produce isobolograms and contour plots. The strongest synergism and antagonism were observed in cadmium-copper and iron-zinc, respectively. A three-way interaction term was added to produce full ternary isobologram surfaces, which, to our knowledge, are a significant first in the toxicology literature.

Fast polypharmacy side effect prediction using tensor factorization

MOTIVATION

Adverse reactions from drug combinations are increasingly common, making their accurate prediction a crucial challenge in modern medicine. Laboratory-based identification of these reactions is insufficient due to the combinatorial nature of the problem. While many computational approaches have been proposed, tensor factorization (TF) models have shown mixed results, necessitating a thorough investigation of their capabilities when properly optimized.

RESULTS

We demonstrate that TF models can achieve state-of-the-art performance on polypharmacy side effect prediction, with our best model (SimplE) achieving median scores of 0.978 area under receiver-operating characteristic curve, 0.971 area under precision-recall curve, and 1.000 AP@50 across 963 side effects. Notably, this model reaches 98.3% of its maximum performance after just two epochs of training (approximately 4 min), making it substantially faster than existing approaches while maintaining comparable accuracy. We also find that incorporating monopharmacy data as self-looping edges in the graph performs marginally better than using it to initialize embeddings.

AVAILABILITY AND IMPLEMENTATION

All code used in the experiments is available in our GitHub repository (https://doi.org/10.5281/zenodo.10684402). The implementation was carried out using Python 3.8.12 with PyTorch 1.7.1, accelerated with CUDA 11.4 on NVIDIA GeForce RTX 2080 Ti GPUs.

Reduction of myeloid-derived suppressor cells in prostate cancer murine models and patients following white button mushroom treatment

BACKGROUND

In a previously reported Phase I trial, we observed therapy-associated declines in circulating myeloid-derived suppressor cells (MDSCs) with the administration of white button mushroom (WBM) tablets in prostate cancer (PCa) patients. These observations led us to hypothesise that WBM could mitigate PCa progression by suppressing MDSCs.

METHODS

We performed bidirectional translational research to examine the immunomodulatory effects of WBM consumption in both syngeneic murine PCa models and patients with PCa participating in an ongoing randomised Phase II trial (NCT04519879).

RESULTS

In murine models, WBM treatment significantly suppressed tumour growth with a reduction in both the number and function of MDSCs, which in turn promoted antitumour immune responses mediated by T cells and natural killer (NK) cells. In patients, after consumption of WBM tablets for 3 months, we observed a decline in circulating polymorphonuclear MDSCs (PMN-MDSCs), along with an increase in cytotoxic CD8+ T and NK cells. Furthermore, single immune cell profiling of peripheral blood from WBM-treated patients showed suppressed STAT3/IRF1 and TGFβ signalling in circulating PMN-MDSCs. Subclusters of PMN-MDSCs presented transcriptional profiles associated with responsiveness to fungi, neutrophil chemotaxis, leukocyte aggregation, and regulation of inflammatory response. Finally, in mouse models of PCa, we found that WBM consumption enhanced the anticancer activity of anti-PD-1 antibodies, indicating that WBM may be used as an adjuvant therapy with immune checkpoint inhibitors.

CONCLUSION

Our results from PCa murine models and patients provide mechanistic insights into the immunomodulatory effects of WBM and provide a scientific foundation for WBM as a nutraceutical intervention to delay or prevent PCa progression.

HIGHLIGHTS

White button mushroom (WBM) treatment resulted in a reduction in pro-tumoural MDSCs, notably polymorphonuclear MDSCs (PMN-MDSCs), along with activation of anti-tumoural T and NK cells. Human single immune cell gene expression profiling shed light on the molecular alterations induced by WBM, specifically on PMN-MDSCs. A proof-of-concept study combining WBM with PD-1 blockade in murine models revealed an additive effect on tumour regression and survival outcomes, highlighting the clinical relevance of WBM in cancer management.

A weighted Bayesian integration method for predicting drug combination using heterogeneous data

BACKGROUND

In the management of complex diseases, the strategic adoption of combination therapy has gained considerable prominence. Combination therapy not only holds the potential to enhance treatment efficacy but also to alleviate the side effects caused by excessive use of a single drug. Presently, the exploration of combination therapy encounters significant challenges due to the vast spectrum of potential drug combinations, necessitating the development of efficient screening strategies.

METHODS

In this study, we propose a prediction scoring method that integrates heterogeneous data using a weighted Bayesian method for drug combination prediction. Heterogeneous data refers to different types of data related to drugs, such as chemical, pharmacological, and target profiles. By constructing a multiplex drug similarity network, we formulate new features for drug pairs and propose a novel Bayesian-based integration scheme with the introduction of weights to integrate information from various sources. This method yields support strength scores for drug combinations to assess their potential effectiveness.

RESULTS

Upon comprehensive comparison with other methods, our method shows superior performance across multiple metrics, including the Area Under the Receiver Operating Characteristic Curve, accuracy, precision, and recall. Furthermore, literature validation shows that many top-ranked drug combinations based on the support strength score, such as goserelin and letrozole, have been experimentally or clinically validated for their effectiveness.

CONCLUSIONS

Our findings have significant clinical and practical implications. This new method enhances the performance of drug combination predictions, enabling effective pre-screening for trials and, thereby, benefiting clinical treatments. Future research should focus on developing new methods for application in various scenarios and for integrating diverse data sources.

DSIL-DDI: A Domain-Invariant Substructure Interaction Learning for Generalizable Drug-Drug Interaction Prediction

Drug-drug interactions (DDIs) trigger unexpected pharmacological effects in vivo, often with unknown causal mechanisms. Deep learning methods have been developed to better understand DDI. However, learning domain-invariant representations for DDI remains a challenge. Generalizable DDI predictions are closer to reality than source domain predictions. For existing methods, it is difficult to achieve out-of-distribution (OOD) predictions. In this article, focusing on substructure interaction, we propose DSIL-DDI, a pluggable substructure interaction module that can learn domain-invariant representations of DDIs from source domain. We evaluate DSIL-DDI on three scenarios: the transductive setting (all drugs in test set appear in training set), the inductive setting (test set contains new drugs that were not present in training set), and OOD generalization setting (training set and test set belong to two different datasets). The results demonstrate that DSIL-DDI improve the generalization and interpretability of DDI prediction modeling and provides valuable insights for OOD DDI predictions. DSIL-DDI can help doctors ensuring the safety of drug administration and reducing the harm caused by drug abuse.

Tumor Treating Fields (TTFields) induce homologous recombination deficiency in ovarian cancer cells, thus mitigating drug resistance

Background

Ovarian cancer is the leading cause of mortality among gynecological malignancies. Carboplatin and poly (ADP-ribose) polymerase inhibitors (PARPi) are often implemented in the treatment of ovarian cancer. Homologous recombination deficient (HRD) tumors demonstrate increased sensitivity to these treatments; however, many ovarian cancer patients are homologous recombination proficient (HRP). TTFields are non-invasive electric fields that induce an HRD-like phenotype in various cancer types. The current study aimed to investigate the impact of TTFields applied together with carboplatin or PARPi (olaparib or niraparib) in preclinical ovarian cancer models.

Methods

A2780 (HRP), OVCAR3 (HRD), and A2780cis (platinum-resistant) human ovarian cancer cells were treated in vitro with TTFields (1 V/cm RMS, 200 kHz, 72 h), alone or with various drug concentrations. Treated cells were measured for cell count, colony formation, apoptosis, DNA damage, expression of DNA repair proteins, and cell cycle. In vivo, ID8-fLuc (HRP) ovarian cancer cells were inoculated intraperitoneally to C57BL/6 mice, which were then treated with either sham, TTFields (200 kHz), olaparib (50 mg/kg), or TTFields plus olaparib; over a period of four weeks. Tumor growth was analyzed using bioluminescent imaging at treatment cessation; and survival analysis was performed.

Results

The nature of TTFields-drug interaction was dependent on the drug's underlying mechanism of action and on the genetic background of the cells, with synergistic interactions between TTFields and carboplatin or PARPi seen in HRP and resistant cells. Treated cells demonstrated elevated levels of DNA damage, accompanied by G2/M arrest, and induction of an HRD-like phenotype. In the tumor-bearing mice, TTFields and olaparib co-treatment resulted in reduced tumor volume and a survival benefit relative to olaparib monotherapy and to control.

Conclusion

By inducing an HRD-like phenotype, TTFields sensitize HRP and resistant ovarian cancer cells to treatment with carboplatin or PARPi, potentially mitigating a-priori and de novo drug resistance, a major limitation in ovarian cancer treatment.

Protein tyrosine phosphatases: emerging role in cancer therapy resistance

BACKGROUND

Tyrosine phosphorylation of intracellular proteins is a post-translational modification that plays a regulatory role in signal transduction during cellular events. Dephosphorylation of signal transduction proteins caused by protein tyrosine phosphatases (PTPs) contributed their role as a convergent node to mediate cross-talk between signaling pathways. In the context of cancer, PTP-mediated pathways have been identified as signaling hubs that enabled cancer cells to mitigate stress induced by clinical therapy. This is achieved by the promotion of constitutive activation of growth-stimulatory signaling pathways or modulation of the immune-suppressive tumor microenvironment. Preclinical evidences suggested that anticancer drugs will release their greatest therapeutic potency when combined with PTP inhibitors, reversing drug resistance that was responsible for clinical failures during cancer therapy.

AREAS COVERED

This review aimed to elaborate recent insights that supported the involvement of PTP-mediated pathways in the development of resistance to targeted therapy and immune-checkpoint therapy.

EXPERT OPINION

This review proposed the notion of PTP inhibition in anticancer combination therapy as a potential strategy in clinic to achieve long-term tumor regression. Ongoing clinical trials are currently underway to assess the safety and efficacy of combination therapy in advanced-stage tumors.

MPHGCL-DDI: Meta-Path-Based Heterogeneous Graph Contrastive Learning for Drug-Drug Interaction Prediction

The combinatorial therapy with multiple drugs may lead to unexpected drug-drug interactions (DDIs) and result in adverse reactions to patients. Predicting DDI events can mitigate the potential risks of combinatorial therapy and enhance drug safety. In recent years, deep models based on heterogeneous graph representation learning have attracted widespread interest in DDI event prediction and have yielded satisfactory results, but there is still room for improvement in prediction performance. In this study, we proposed a meta-path-based heterogeneous graph contrastive learning model, MPHGCL-DDI, for DDI event prediction. The model constructs two contrastive views based on meta-paths: an average graph view and an augmented graph view. The former represents that there are connections between drugs, while the latter reveals how the drugs connect with each other. We defined three levels of data augmentation schemes in the augmented graph view and adopted a combination of three losses in the model training phase: multi-relation prediction loss, unsupervised contrastive loss and supervised contrastive loss. Furthermore, the model incorporates indirect drug information, protein-protein interactions (PPIs), to reveal latent relations of drugs. We evaluated MPHGCL-DDI on three different tasks of two datasets. Experimental results demonstrate that MPHGCL-DDI surpasses several state-of-the-art methods in performance.

A Graphical Approach to Visualize and Interpret Biochemically Coupled Biomechanical Models

The last decade has seen the emergence of progressively more complex mechanobiological models, often coupling biochemical and biomechanical components. The complexity of these models makes interpretation difficult, and although computational tools can solve model equations, there is considerable potential value in a simple method to explore the interplay between different model components. Pump and system performance curves, long utilized in centrifugal pump selection and design, inspire the development of a graphical technique to depict visually the performance of biochemically-coupled mechanical models. Our approach is based on a biochemical performance curve (analogous to the classical pump curve) and a biomechanical performance curve (analogous to the system curve). Upon construction of the two curves, their intersection, or lack thereof, describes the coupled model's equilibrium state(s). One can also observe graphically how an applied perturbation shifts one or both curves, and thus how the other component will respond, without rerunning the full model. While the upfront cost of generating the performance curve graphic varies with the efficiency of the model components, the easily interpretable visual depiction of what would otherwise be nonintuitive model behavior is valuable. Herein, we outline how performance curves can be constructed and interpreted for biochemically-coupled biomechanical models and apply the technique to two independent models in the cardiovascular space. The performance curve approach can illustrate and help identify weaknesses in model construction, inform user-applied perturbations and fitting procedures to generate intended behaviors, and improve the efficiency of the model generation and application process.

A Novel Interaction of Slug (SNAI2) and Nuclear Actin

Actin is a protein of central importance to many cellular functions. Its localization and activity are regulated by interactions with a high number of actin-binding proteins. In a yeast two-hybrid (Y2H) screening system, snail family transcriptional repressor 2 (SNAI2 or slug) was identified as a yet unknown potential actin-binding protein. We validated this interaction using immunoprecipitation and analyzed the functional relation between slug and actin. Since both proteins have been reported to be involved in DNA double-strand break (DSB) repair, we focused on their interaction during this process after treatment with doxorubicin or UV irradiation. Confocal microscopy elicits that the overexpression of actin fused to an NLS stabilizes complexes of slug and γH2AX, an early marker of DNA damage repair.

Modeling mixtures interactions in environmental toxicology

In the environment, organisms are exposed to mixtures of different toxicants, which may interact in ways that are difficult to predict when only considering each component individually. Adapting and expanding tools from pharmacology, the toxicology field uses analytical, graphical, and computational methods to identify and quantify interactions in multi-component mixtures. The two general frameworks are concentration addition, where components have similar modes of action and their effects sum together, or independent action, where components have dissimilar modes of action and do not interact. Other interaction behaviors include synergism and antagonism, where the combined effects are more or less than the additive sum of individual effects. This review covers foundational theory, methods, an in-depth survey of original research from the past 20 years, current trends, and future directions. As humans and ecosystems are exposed to increasingly complex mixtures of environmental contaminants, analyzing mixtures interactions will continue to become a more critical aspect of toxicological research.

MSFF-MA-DDI: Multi-Source Feature Fusion with Multiple Attention blocks for predicting Drug-Drug Interaction events

The interaction of multiple drugs could lead to severe events, which cause medical injuries and expenses. Accurate prediction of drug-drug interaction (DDI) events can help clinicians make effective decisions and establish appropriate therapy programs. However, there exist two issues worthy of further consideration. (i) The global features of drug molecules should be paid attention to, rather than just their local characteristics. (ii) The fusion of multi-source features should also be studied to capture the comprehensive features of the drug. This study designs a Multi-Source Feature Fusion framework with Multiple Attention blocks named MSFF-MA-DDI that utilizes multimodal data for DDI event prediction. MSFF-MA-DDI can (i) encode global correlations between long-distance atoms in drug molecular sequences by a self-attention layer based on a position embedding block and (ii) fuse drug sequence features and heterogeneous features (chemical substructure, target, and enzyme) through a multi-head attention block to better represent the features of drugs. Experiments on real-world datasets show that MSFF-MA-DDI can achieve performance that is close to or even better than state-of-the-art models. Especially in cold start scenarios, the model can achieve the best performance. The effectiveness of the model is also supported by the case study on nervous system drugs. The source codes and data are available at https://github.com/BioCenter-SHU/MSFF-MA-DDI.

Personalized tumor combination therapy optimization using the single-cell transcriptome

BACKGROUND

The precise characterization of individual tumors and immune microenvironments using transcriptome sequencing has provided a great opportunity for successful personalized cancer treatment. However, the cancer treatment response is often characterized by in vitro assays or bulk transcriptomes that neglect the heterogeneity of malignant tumors in vivo and the immune microenvironment, motivating the need to use single-cell transcriptomes for personalized cancer treatment.

METHODS

Here, we present comboSC, a computational proof-of-concept study to explore the feasibility of personalized cancer combination therapy optimization using single-cell transcriptomes. ComboSC provides a workable solution to stratify individual patient samples based on quantitative evaluation of their personalized immune microenvironment with single-cell RNA sequencing and maximize the translational potential of in vitro cellular response to unify the identification of synergistic drug/small molecule combinations or small molecules that can be paired with immune checkpoint inhibitors to boost immunotherapy from a large collection of small molecules and drugs, and finally prioritize them for personalized clinical use based on bipartition graph optimization.

RESULTS

We apply comboSC to publicly available 119 single-cell transcriptome data from a comprehensive set of 119 tumor samples from 15 cancer types and validate the predicted drug combination with literature evidence, mining clinical trial data, perturbation of patient-derived cell line data, and finally in-vivo samples.

CONCLUSIONS

Overall, comboSC provides a feasible and one-stop computational prototype and a proof-of-concept study to predict potential drug combinations for further experimental validation and clinical usage using the single-cell transcriptome, which will facilitate and accelerate personalized tumor treatment by reducing screening time from a large drug combination space and saving valuable treatment time for individual patients. A user-friendly web server of comboSC for both clinical and research users is available at www.combosc.top . The source code is also available on GitHub at https://github.com/bm2-lab/comboSC .

MATT-DDI: Predicting multi-type drug-drug interactions via heterogeneous attention mechanisms

The joint use of multiple drugs can result in adverse drug-drug interactions (DDIs) and side effects that harm the body. Accurate identification of DDIs is crucial for avoiding accidental drug side effects and understanding potential mechanisms underlying DDIs. Several computational methods have been proposed for multi-type DDI prediction, but most rely on the similarity profiles of drugs as the drug feature vectors, which may result in information leakage and overoptimistic performance when predicting interactions between new drugs. To address this issue, we propose a novel method, MATT-DDI, for predicting multi-type DDIs based on the original feature vectors of drugs and multiple attention mechanisms. MATT-DDI consists of three main modules: the top k most similar drug pair selection module, heterogeneous attention mechanism module and multi‑type DDI prediction module. Firstly, based on the feature vector of the input drug pair (IDP), k drug pairs that are most similar to the input drug pair from the training dataset are selected according to cosine similarity between drug pairs. Then, the vectors of k selected drug pairs are averaged to obtain a new drug pair (NDP). Next, IDP and NDP are fed into heterogeneous attention modules, including scaled dot product attention and bilinear attention, to extract latent feature vectors. Finally, these latent feature vectors are taken as input of the classification module to predict DDI types. We evaluated MATT-DDI on three different tasks. The experimental results show that MATT-DDI provides better or comparable performance compared to several state-of-the-art methods, and its feasibility is supported by case studies. MATT-DDI is a robust model for predicting multi-type DDIs with excellent performance and no information leakage.

Predicting anticancer synergistic drug combinations based on multi-task learning

BACKGROUND

The discovery of anticancer drug combinations is a crucial work of anticancer treatment. In recent years, pre-screening drug combinations with synergistic effects in a large-scale search space adopting computational methods, especially deep learning methods, is increasingly popular with researchers. Although achievements have been made to predict anticancer synergistic drug combinations based on deep learning, the application of multi-task learning in this field is relatively rare. The successful practice of multi-task learning in various fields shows that it can effectively learn multiple tasks jointly and improve the performance of all the tasks.

METHODS

In this paper, we propose MTLSynergy which is based on multi-task learning and deep neural networks to predict synergistic anticancer drug combinations. It simultaneously learns two crucial prediction tasks in anticancer treatment, which are synergy prediction of drug combinations and sensitivity prediction of monotherapy. And MTLSynergy integrates the classification and regression of prediction tasks into the same model. Moreover, autoencoders are employed to reduce the dimensions of input features.

RESULTS

Compared with the previous methods listed in this paper, MTLSynergy achieves the lowest mean square error of 216.47 and the highest Pearson correlation coefficient of 0.76 on the drug synergy prediction task. On the corresponding classification task, the area under the receiver operator characteristics curve and the area under the precision-recall curve are 0.90 and 0.62, respectively, which are equivalent to the comparison methods. Through the ablation study, we verify that multi-task learning and autoencoder both have a positive effect on prediction performance. In addition, the prediction results of MTLSynergy in many cases are also consistent with previous studies.

CONCLUSION

Our study suggests that multi-task learning is significantly beneficial for both drug synergy prediction and monotherapy sensitivity prediction when combining these two tasks into one model. The ability of MTLSynergy to discover new anticancer synergistic drug combinations noteworthily outperforms other state-of-the-art methods. MTLSynergy promises to be a powerful tool to pre-screen anticancer synergistic drug combinations.

Review of Predicting Synergistic Drug Combinations

The prediction of drug combinations is of great clinical significance. In many diseases, such as high blood pressure, diabetes, and stomach ulcers, the simultaneous use of two or more drugs has shown clear efficacy. It has greatly reduced the progression of drug resistance. This review presents the latest applications of methods for predicting the effects of drug combinations and the bioactivity databases commonly used in drug combination prediction. These studies have played a significant role in developing precision therapy. We first describe the concept of synergy. we study various publicly available databases for drug combination prediction tasks. Next, we introduce five algorithms applied to drug combinatorial prediction, which include traditional machine learning methods, deep learning methods, mathematical methods, systems biology methods and search algorithms. In the end, we sum up the difficulties encountered in prediction models.

Identifying chemicals based on receptor binding/bioactivation/mechanistic explanation associated with potential to elicit hepatotoxicity and to support structure activity relationship-based read-across

The liver is the most common target organ in toxicology studies. The development of chemical structural alerts for identifying hepatotoxicity will play an important role in in silico model prediction and help strengthen the identification of analogs used in structure activity relationship (SAR)- based read-across. The aim of the current study is development of an SAR-based expert-system decision tree for screening of hepatotoxicants across a wide range of chemistry space and proposed modes of action for clustering of chemicals using defined core chemical categories based on receptor-binding or bioactivation. The decision tree is based on ∼ 1180 different chemicals that were reviewed for hepatotoxicity information. Knowledge of chemical receptor binding, metabolism and mechanistic information were used to group these chemicals into 16 different categories and 102 subcategories: four categories describe binders to 9 different receptors, 11 categories are associated with possible reactive metabolites (RMs) and there is one miscellaneous category. Each chemical subcategory has been associated with possible modes of action (MOAs) or similar key structural features. This decision tree can help to screen potential liver toxicants associated with core structural alerts of receptor binding and/or RMs and be used as a component of weight of evidence decisions based on SAR read-across, and to fill data gaps.

Network analyses reveal new insights into the effect of multicomponent Tr14 compared to single-component diclofenac in an acute inflammation model

BACKGROUND

Modifying the acute inflammatory response has wide clinical benefits. Current options include non-steroidal anti-inflammatory drugs (NSAIDs) and therapies that may resolve inflammation. Acute inflammation involves multiple cell types and various processes. We, therefore, investigated whether an immunomodulatory drug that acts simultaneously at multiple sites shows greater potential to resolve acute inflammation more effectively and with fewer side effects than a common anti-inflammatory drug developed as a small molecule for a single target. In this work, we used time-series gene expression profiles from a wound healing mouse model to compare the effects of Traumeel (Tr14), a multicomponent natural product, to diclofenac, a single component NSAID on inflammation resolution.

RESULTS

We advance previous studies by mapping the data onto the "Atlas of Inflammation Resolution", followed by in silico simulations and network analysis. We found that Tr14 acts primarily on the late phase of acute inflammation (during resolution) compared to diclofenac, which suppresses acute inflammation immediately after injury.

CONCLUSIONS

Our results provide new insights how network pharmacology of multicomponent drugs may support inflammation resolution in inflammatory conditions.

Biochemical reaction network topology defines dose-dependent Drug-Drug interactions

Drug combination therapy is a promising strategy to enhance the desired therapeutic effect, while reducing side effects. High-throughput pairwise drug combination screening is a commonly used method for discovering favorable drug interactions, but is time-consuming and costly. Here, we investigate the use of reaction network topology-guided design of combination therapy as a predictive in silico drug-drug interaction screening approach. We focused on three-node enzymatic networks, with general Michaelis-Menten kinetics. The results revealed that drug-drug interactions critically depend on the choice of target arrangement in a given topology, the nature of the drug, and the desired level of change in the network output. The results showed a negative correlation between antagonistic interactions and the dosage of drugs. Overall, the negative feedback loops showed the highest synergistic interactions (the lowest average combination index) and, intriguingly, required the highest drug doses compared to other topologies under the same condition.

Comparing the Effects of Encapsulated and Non-Encapsulated Propolis Extracts on Model Lipid Membranes and Lactic Bacteria, with Emphasis on the Synergistic Effects of Its Various Compounds

Propolis is a resinous compound made by bees with well-known biological activity. However, comparisons between encapsulated and non-encapsulated propolis are lacking. Therefore, the antibacterial activity, effect on the phase transition of lipids, and inhibition of UV-induced lipid oxidation of the two forms of propolis were compared. The results showed that non-encapsulated propolis produces quicker effects, thus being better suited when more immediate effects are required (e.g., antibacterial activity). In order to gain an in-depth introspective on these effects, we further studied the synergistic effect of propolis compounds on the integrity of lipid membranes. The knowledge of component synergism is important for the understanding of effective propolis pathways and for the perspective of modes of action of synergism between different polyphenols in various extracts. Thus, five representative molecules, all previously isolated from propolis (chrysin, quercetin, trans-ferulic acid, caffeic acid, (-)-epigallocatechin-3-gallate) were mixed, and their synergistic effects on lipid bilayers were investigated, mainly using DSC. The results showed that some compounds (quercetin, chrysin) exhibit synergism, whereas others (caffeic acid, t-ferulic acid) do not show any such effects. The results also showed that the synergistic effects of mixtures composed from several different compounds are extremely complex to study, and that their prediction requires further modeling approaches.

An updated gene regulatory network reconstruction of multidrug-resistant Pseudomonas aeruginosa CCBH4851

BACKGROUND

Healthcare-associated infections due to multidrug-resistant (MDR) bacteria such as Pseudomonas aeruginosa are significant public health issues worldwide. A system biology approach can help understand bacterial behaviour and provide novel ways to identify potential therapeutic targets and develop new drugs. Gene regulatory networks (GRN) are examples of in silico representation of interaction between regulatory genes and their targets.

OBJECTIVES

In this work, we update the MDR P. aeruginosa CCBH4851 GRN reconstruction and analyse and discuss its structural properties.

METHODS

We based this study on the gene orthology inference methodology using the reciprocal best hit method. The P. aeruginosa CCBH4851 genome and GRN, published in 2019, and the P. aeruginosa PAO1 GRN, published in 2020, were used for this update reconstruction process.

FINDINGS

Our result is a GRN with a greater number of regulatory genes, target genes, and interactions compared to the previous networks, and its structural properties are consistent with the complexity of biological networks and the biological features of P. aeruginosa.

MAIN CONCLUSIONS

Here, we present the largest and most complete version of P. aeruginosa GRN published to this date, to the best of our knowledge.

Engaging innate immunity for targeting the epidermal growth factor receptor: Therapeutic options leveraging innate immunity versus adaptive immunity versus inhibition of signaling

The epidermal growth factor receptor (EGFR) is a key player in the normal tissue physiology and the pathology of cancer. Therapeutic approaches have now been developed to target oncogenic genetic aberrations of EGFR, found in a subset of tumors, and to take advantage of overexpression of EGFR in tumors. The development of small-molecule inhibitors and anti-EGFR antibodies targeting EGFR activation have resulted in effective but limited treatment options for patients with mutated or wild-type EGFR-expressing cancers, while therapeutic approaches that deploy effectors of the adaptive or innate immune system are still undergoing development. This review discusses EGFR-targeting therapies acting through distinct molecular mechanisms to destroy EGFR-expressing cancer cells. The focus is on the successes and limitations of therapies targeting the activation of EGFR versus those that exploit the cytotoxic T cells and innate immune cells to target EGFR-expressing cancer cells. Moreover, we discuss alternative approaches that may have the potential to overcome limitations of current therapies; in particular the innate cell engagers are discussed. Furthermore, this review highlights the potential to combine innate cell engagers with immunotherapies, to maximize their effectiveness, or with unspecific cell therapies, to convert them into tumor-specific agents.

Evaluating a therapeutic window for precision medicine by integrating genomic profiles and p53 network dynamics

The response variation to anti-cancer drugs originates from complex intracellular network dynamics of cancer. Such dynamic networks present challenges to determining optimal drug targets and stratifying cancer patients for precision medicine, although several cancer genome studies provided insights into the molecular characteristics of cancer. Here, we introduce a network dynamics-based approach based on attractor landscape analysis to evaluate the therapeutic window of a drug from cancer signaling networks combined with genomic profiles. This approach allows for effective screening of drug targets to explore potential target combinations for enhancing the therapeutic window of drug responses. We also effectively stratify patients into desired/undesired response groups using critical genomic determinants, which are network-specific origins of variability to drug response, and their dominance relationship. Our methods provide a viable and quantitative framework to connect genotype information to the phenotypes of drug response with regard to network dynamics determining the therapeutic window.

Cyclooxygenase-2 Upregulated by Temozolomide in Glioblastoma Cells Is Shuttled In Extracellular Vesicles Modifying Recipient Cell Phenotype

Temozolomide (TMZ) resistance is frequent in patients with glioblastoma (GBM), a tumor characterized by a marked inflammatory microenvironment. Recently, we reported that cyclooxygenase-2 (COX-2) is upregulated in TMZ-resistant GBM cells treated with high TMZ concentrations. Moreover, COX-2 activity inhibition significantly counteracted TMZ-resistance of GBM cells. Extracellular vesicles (EV) are considered crucial mediators in orchestrating GBM drug resistance by modulating the tumor microenvironment (TME) and affecting the surrounding recipient cell phenotype and behavior. This work aimed to verify whether TMZ, at low and clinically relevant doses (5-20 µM), could induce COX-2 overexpression in GBM cells (T98G and U87MG) and explore if secreted EV shuttled COX-2 to recipient cells. The effect of COX-2 inhibitors (COXIB), Celecoxib (CXB), or NS398, alone or TMZ-combined, was also investigated. Our results indicated that TMZ at clinically relevant doses upregulated COX-2 in GBM cells. COXIB treatment significantly counteracted TMZ-induced COX-2 expression, confirming the crucial role of the COX-2/PGE2 system in TMZ-resistance. The COXIB specificity was verified on U251MG, COX-2 null GBM cells. Western blotting of GBM-EV cells showed the COX-2 presence, with the same intracellular trend, increasing in EV derived from TMZ-treated cells and decreasing in those derived from COXIB+TMZ-treated cells. We then evaluated the effect of EV secreted by TMZ-treated cells on U937 and U251MG, used as recipient cells. In human macrophage cell line U937, the internalization of EV derived by TMZ-T98G cells led to a shift versus a pro-tumor M2-like phenotype. On the other hand, EV from TMZ-T98G induced a significant decrease in TMZ sensitivity in U251MG cells. Overall, our results, in confirming the crucial role played by COX-2 in TMZ-resistance, provide the first evidence of the presence and effective functional transfer of this enzyme through EV derived from GBM cells, with multiple potential consequences at the level of TME.

Detection of polypharmacy side effects by integrating multiple data sources and convolutional neural networks

The consumption of drug combinations, named polypharmacy, is commonly used for treating patients with several diseases or those with complex conditions. However, the main drawback of polypharmacy is the increased probability of harmful side effects. The polypharmacy side effects are caused by an interaction between two medications. It means that the drug-drug interaction causes changes in their activities due to interfering in each other's performance. Therefore, discovering these side effects is one of the most challenging and important aspects of drug production and consumption as it is associated with human health. In this paper, a method has been introduced for predicting the polypharmacy side effects, called PSECNN. It is a multi-label multi-class deep learning method that combines various basic features of drugs to predict the polypharmacy side effects. Firstly, PSECNN collects five basic features of drugs, such as individual drug's side effects, drug-protein interactions, chemical substructures, targets, and enzymes in order to create a novel combination of drug features. A feature extraction module creates five feature vectors with the same dimension for each drug based on the Jaccard similarity index. Based on the feature vectors, a unique representative is then created for each drug. These representative vectors are given in pairs as input to the deep neural network to predict the occurrence probability of side effects. According to the experimental evaluations, PSECNN could outperform the state-of-the-art polypharmacy side effects prediction methods up to 74%. It has been found that PSECNN has better performance with polypharmacy side effects with a cause of molecular basis due to the novel combination of basic drug features.

Predicting combinations of drugs by exploiting graph embedding of heterogeneous networks

BACKGROUND

Drug combination, offering an insight into the increased therapeutic efficacy and reduced toxicity, plays an essential role in the therapy of many complex diseases. Although significant efforts have been devoted to the identification of drugs, the identification of drug combination is still a challenge. The current algorithms assume that the independence of feature selection and drug prediction procedures, which may result in an undesirable performance.

RESULTS

To address this issue, we develop a novel Semi-supervised Heterogeneous Network Embedding algorithm (called SeHNE) to predict the combination patterns of drugs by exploiting the graph embedding. Specifically, the ATC similarity of drugs, drug-target, and protein-protein interaction networks are integrated to construct the heterogeneous networks. Then, SeHNE jointly learns drug features by exploiting the topological structure of heterogeneous networks and predicting drug combination. One distinct advantage of SeHNE is that features of drugs are extracted under the guidance of classification, which improves the quality of features, thereby enhancing the performance of prediction of drugs. Experimental results demonstrate that the proposed algorithm is more accurate than state-of-the-art methods on various data, implying that the joint learning is promising for the identification of drug combination.

CONCLUSIONS

The proposed model and algorithm provide an effective strategy for the prediction of combinatorial patterns of drugs, implying that the graph-based drug prediction is promising for the discovery of drugs.

Radiosensitizing effect of galunisertib, a TGF-ß receptor I inhibitor, on head and neck squamous cell carcinoma in vitro

Background. Resistance to radiation therapy poses a major clinical problem for patients suffering from head and neck squamous cell carcinoma (HNSCC). Transforming growth factor ß (TGF-ß) has emerged as a potential target. This study aimed to investigate the radiosensitizing effect of galunisertib, a small molecule TGF-ß receptor kinase I inhibitor, on HNSCC cells in vitro. Methods. Three HNSCC cell lines were treated with galunisertib alone, or in combination with radiation. Of those three cell lines, one has a known inactivating mutation of the TGF-ß pathway (Cal27), one has a TGF-ß pathway deficiency (FaDu) and one has no known alteration (SCC-25). The effect on metabolic activity was evaluated by a resazurin-based reduction assay. Cell migration was evaluated by wound-healing assay, clonogenic survival by colony formation assay and cell cycle by FACS analysis. Results. Galunisertib reduced metabolic activity in FaDu, increased in SCC-25 and had no effect on CAL27. Migration was significantly reduced by galunisertib in all three cell lines and showed additive effects in combination with radiation in CAL27 and SCC-25. Colony-forming capabilities were reduced in SCC-25 by galunisertib and also showed an additive effect with adjuvant radiation treatment. Cell cycle analysis showed a reduction of cells in G₁ phase in response to galunisertib treatment. Conclusion. Our results indicate a potential antineoplastic effect of galunisertib in HNSCC with intact TGF-ß signaling in combination with radiation.

Repositioning of Fungal-based Peptides as Modulators of Angiotensin-converting Enzyme-related Carboxypeptidase, SARS-coronavirus HR2 Domain, and Coronavirus Disease 2019 Main Protease

Background and Objectives

Angiotensin-converting enzyme-related carboxypeptidase, SARS-Coronavirus HR2 Domain, and COVID-19 main protease are essential for the cellular entry and replication of coronavirus in the host. This study investigated the putative inhibitory action of peptides form medicinal mushrooms, namely Pseudoplectania nigrella, Russula paludosa, and Clitocybe sinopica, towards selected proteins through computational studies.

Materials and Methods

The respective physicochemical properties of selected peptides were predicted using ProtParam tool, while the binding modes and binding free energy of selected peptides toward proteins were computed through HawkDock server. The structural flexibility and stability of docked protein-peptide complexes were assessed through iMODS server.

Results

The peptides showed an optimum binding afinity with the molecular targets; plectasin from P. nigrella showed the highest binding free energy compared to peptides from R. paludosa and C. sinopica. Besides, molecular dynamic simulations showed all fungal-based peptides could influence the flexibility and stability of selected proteins.

Conclusion

The study revealed fungal-based peptides could be explored as functional modulators of essential proteins that are involved in the cellular entry of coronavirus.

Computational Analysis of Cholangiocarcinoma Phosphoproteomes Identifies Patient-Specific Drug Targets

Cholangiocarcinoma is a form of hepatobiliary cancer with an abysmal prognosis. Despite advances in our understanding of cholangiocarcinoma pathophysiology and its genomic landscape, targeted therapies have not yet made a significant impact on its clinical management. The low response rates of targeted therapies in cholangiocarcinoma suggest that patient heterogeneity contributes to poor clinical outcome. Here we used mass spectrometry-based phosphoproteomics and computational methods to identify patient-specific drug targets in patient tumors and cholangiocarcinoma-derived cell lines. We analyzed 13 primary tumors of patients with cholangiocarcinoma with matched nonmalignant tissue and 7 different cholangiocarcinoma cell lines, leading to the identification and quantification of more than 13,000 phosphorylation sites. The phosphoproteomes of cholangiocarcinoma cell lines and patient tumors were significantly correlated. MEK1, KIT, ERK1/2, and several cyclin-dependent kinases were among the protein kinases most frequently showing increased activity in cholangiocarcinoma relative to nonmalignant tissue. Application of the Drug Ranking Using Machine Learning (DRUML) algorithm selected inhibitors of histone deacetylase (HDAC; belinostat and CAY10603) and PI3K pathway members as high-ranking therapies to use in primary cholangiocarcinoma. The accuracy of the computational drug rankings based on predicted responses was confirmed in cell-line models of cholangiocarcinoma. Together, this study uncovers frequently activated biochemical pathways in cholangiocarcinoma and provides a proof of concept for the application of computational methodology to rank drugs based on efficacy in individual patients. SIGNIFICANCE: Phosphoproteomic and computational analyses identify patient-specific drug targets in cholangiocarcinoma, supporting the potential of a machine learning method to predict personalized therapies.

On the robustness of generalization of drug-drug interaction models

BACKGROUND

Deep learning methods are a proven commodity in many fields and endeavors. One of these endeavors is predicting the presence of adverse drug-drug interactions (DDIs). The models generated can predict, with reasonable accuracy, the phenotypes arising from the drug interactions using their molecular structures. Nevertheless, this task requires improvement to be truly useful. Given the complexity of the predictive task, an extensive benchmarking on structure-based models for DDIs prediction was performed to evaluate their drawbacks and advantages.

RESULTS

We rigorously tested various structure-based models that predict drug interactions using different splitting strategies to simulate different real-world scenarios. In addition to the effects of different training and testing setups on the robustness and generalizability of the models, we then explore the contribution of traditional approaches such as multitask learning and data augmentation.

CONCLUSION

Structure-based models tend to generalize poorly to unseen drugs despite their ability to identify new DDIs among drugs seen during training accurately. Indeed, they efficiently propagate information between known drugs and could be valuable for discovering new DDIs in a database. However, these models will most probably fail when exposed to unknown drugs. While multitask learning does not help in our case to solve the problem, the use of data augmentation does at least mitigate it. Therefore, researchers must be cautious of the bias of the random evaluation scheme, especially if their goal is to discover new DDIs.

Design, Synthesis, and Characterization of 4-Aminoquinazolines as Potent Inhibitors of the G Protein-Coupled Receptor Kinase 6 (GRK6) for the Treatment of Multiple Myeloma

Both previous and additional genetic knockdown studies reported herein implicate G protein-coupled receptor kinase 6 (GRK6) as a critical kinase required for the survival of multiple myeloma (MM) cells. Therefore, we sought to develop a small molecule GRK6 inhibitor as an MM therapeutic. From a focused library of known kinase inhibitors, we identified two hits with moderate biochemical potencies against GRK6. From these hits, we developed potent (IC₅₀ < 10 nM) analogues with selectivity against off-target kinases. Further optimization led to the discovery of an analogue (18) with an IC₅₀ value of 6 nM against GRK6 and selectivity against a panel of 85 kinases. Compound 18 has potent cellular target engagement and antiproliferative activity against MM cells and is synergistic with bortezomib. In summary, we demonstrate that targeting GRK6 with small molecule inhibitors represents a promising approach for MM and identify 18 as a novel, potent, and selective GRK6 inhibitor.

Machine learning liver-injuring drug interactions with non-steroidal anti-inflammatory drugs (NSAIDs) from a retrospective electronic health record (EHR) cohort

Drug-drug interactions account for up to 30% of adverse drug reactions. Increasing prevalence of electronic health records (EHRs) offers a unique opportunity to build machine learning algorithms to identify drug-drug interactions that drive adverse events. In this study, we investigated hospitalizations' data to study drug interactions with non-steroidal anti-inflammatory drugs (NSAIDS) that result in drug-induced liver injury (DILI). We propose a logistic regression based machine learning algorithm that unearths several known interactions from an EHR dataset of about 400,000 hospitalization. Our proposed modeling framework is successful in detecting 87.5% of the positive controls, which are defined by drugs known to interact with diclofenac causing an increased risk of DILI, and correctly ranks aggregate risk of DILI for eight commonly prescribed NSAIDs. We found that our modeling framework is particularly successful in inferring associations of drug-drug interactions from relatively small EHR datasets. Furthermore, we have identified a novel and potentially hepatotoxic interaction that might occur during concomitant use of meloxicam and esomeprazole, which are commonly prescribed together to allay NSAID-induced gastrointestinal (GI) bleeding. Empirically, we validate our approach against prior methods for signal detection on EHR datasets, in which our proposed approach outperforms all the compared methods across most metrics, such as area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC).

Bioinformatics Accelerates the Major Tetrad: A Real Boost for the Pharmaceutical Industry

With advanced technology and its development, bioinformatics is one of the avant-garde fields that has managed to make amazing progress in the pharmaceutical-medical field by modeling the infrastructural dimensions of healthcare and integrating computing tools in drug innovation, facilitating prevention, detection/more accurate diagnosis, and treatment of disorders, while saving time and money. By association, bioinformatics and pharmacovigilance promoted both sample analyzes and interpretation of drug side effects, also focusing on drug discovery and development (DDD), in which systems biology, a personalized approach, and drug repositioning were considered together with translational medicine. The role of bioinformatics has been highlighted in DDD, proteomics, genetics, modeling, miRNA discovery and assessment, and clinical genome sequencing. The authors have collated significant data from the most known online databases and publishers, also narrowing the diversified applications, in order to target four major areas (tetrad): DDD, anti-microbial research, genomic sequencing, and miRNA research and its significance in the management of current pandemic context. Our analysis aims to provide optimal data in the field by stratification of the information related to the published data in key sectors and to capture the attention of researchers interested in bioinformatics, a field that has succeeded in advancing the healthcare paradigm by introducing developing techniques and multiple database platforms, addressed in the manuscript.

Cold-Start Problems in Data-Driven Prediction of Drug-Drug Interaction Effects

Combining drugs, a phenomenon often referred to as polypharmacy, can induce additional adverse effects. The identification of adverse combinations is a key task in pharmacovigilance. In this context, in silico approaches based on machine learning are promising as they can learn from a limited number of combinations to predict for all. In this work, we identify various subtasks in predicting effects caused by drug–drug interaction. Predicting drug–drug interaction effects for drugs that already exist is very different from predicting outcomes for newly developed drugs, commonly called a cold-start problem. We propose suitable validation schemes for the different subtasks that emerge. These validation schemes are critical to correctly assess the performance. We develop a new model that obtains AUC-ROC =0.843 for the hardest cold-start task up to AUC-ROC =0.957 for the easiest one on the benchmark dataset of Zitnik et al. Finally, we illustrate how our predictions can be used to improve post-market surveillance systems or detect drug–drug interaction effects earlier during drug development.

Prediction of Synergistic Drug Combinations for Prostate Cancer by Transcriptomic and Network Characteristics

Prostate cancer (PRAD) is a major cause of cancer-related deaths. Current monotherapies show limited efficacy due to often rapidly emerging resistance. Combination therapies could provide an alternative solution to address this problem with enhanced therapeutic effect, reduced cytotoxicity, and delayed the appearance of drug resistance. However, it is prohibitively cost and labor-intensive for the experimental approaches to pick out synergistic combinations from the millions of possibilities. Thus, it is highly desired to explore other efficient strategies to assist experimental researches. Inspired by the challenge, we construct the transcriptomics-based and network-based prediction models to quickly screen the potential drug combination for Prostate cancer, and further assess their performance by in vitro assays. The transcriptomics-based method screens nine possible combinations. However, the network-based method gives discrepancies for at least three drug pairs. Further experimental results indicate the dose-dependent effects of the three docetaxel-containing combinations, and confirm the synergistic effects of the other six combinations predicted by the transcriptomics-based model. For the network-based predictions, in vitro tests give opposite results to the two combinations (i.e. mitoxantrone-cyproheptadine and cabazitaxel-cyproheptadine). Namely, the transcriptomics-based method outperforms the network-based one for the specific disease like Prostate cancer, which provide guideline for selection of the computational methods in the drug combination screening. More importantly, six combinations (the three mitoxantrone-containing and the three cabazitaxel-containing combinations) are found to be promising candidates to synergistically conquer Prostate cancer.

CellBox: Interpretable Machine Learning for Perturbation Biology with Application to the Design of Cancer Combination Therapy

Systematic perturbation of cells followed by comprehensive measurements of molecular and phenotypic responses provides informative data resources for constructing computational models of cell biology. Models that generalize well beyond training data can be used to identify combinatorial perturbations of potential therapeutic interest. Major challenges for machine learning on large biological datasets are to find global optima in a complex multidimensional space and mechanistically interpret the solutions. To address these challenges, we introduce a hybrid approach that combines explicit mathematical models of cell dynamics with a machine-learning framework, implemented in TensorFlow. We tested the modeling framework on a perturbation-response dataset of a melanoma cell line after drug treatments. The models can be efficiently trained to describe cellular behavior accurately. Even though completely data driven and independent of prior knowledge, the resulting de novo network models recapitulate some known interactions. The approach is readily applicable to various kinetic models of cell biology. A record of this paper's Transparent Peer Review process is included in the Supplemental Information.

Deciphering the combinatorial landscape of immunity

From cellular activation to drug combinations, immunological responses are shaped by the action of multiple stimuli. Synergistic and antagonistic interactions between stimuli play major roles in shaping immune processes. To understand combinatorial regulation, we present the immune Synergistic/Antagonistic Interaction Learner (iSAIL). iSAIL includes a machine learning classifier to map and interpret interactions, a curated compendium of immunological combination treatment datasets, and their global integration into a landscape of ~30,000 interactions. The landscape is mined to reveal combinatorial control of interleukins, checkpoints, and other immune modulators. The resource helps elucidate the modulation of a stimulus by interactions with other cofactors, showing that TNF has strikingly different effects depending on co-stimulators. We discover new functional synergies between TNF and IFNβ controlling dendritic cell-T cell crosstalk. Analysis of laboratory or public combination treatment studies with this user-friendly web-based resource will help resolve the complex role of interaction effects on immune processes.

Programmable Delivery of Synergistic Cancer Drug Combinations Using Bicompartmental Nanoparticles

Delivery of multiple therapeutics has become a preferred method of treating cancer, albeit differences in the biodistribution and pharmacokinetic profiles of individual drugs pose challenges in effectively delivering synergistic drug combinations to and at the tumor site. Here, bicompartmental Janus nanoparticles comprised of domains are reported with distinct bulk properties that allow for independent drug loading and release. Programmable drug release can be triggered by a change in the pH value and depends upon the bulk properties of the polymers used in the respective compartments, rather than the molecular structures of the active agents. Bicompartmental nanoparticles delivering a synergistic combination of lapatinib and paclitaxel result in increased activity against HER2+ breast cancer cells. Surprisingly, the dual drug loaded particles also show significant efficacy toward triple negative breast cancer, even though this cancer model is unresponsive to lapatinib alone. The broad versatility of the nanoparticle platform allows for rapid exploration of a wide range of drug combinations where both their relative drug ratios and temporal release profiles can be optimized.

Individual and mixture toxicity of carbofuran and diuron to the protozoan Paramecium caudatum and the cladoceran Ceriodaphnia silvestrii

The toxicity of the insecticide carbofuran and herbicide diuron (individually and in mixture) to the invertebrates Paramecium caudatum and Ceriodaphnia silvestrii was evaluated. Acute and chronic toxicity tests were carried out with the diuron and carbofuran active ingredients and their commercial products, Diuron Nortox® 500 SC and Furadan® 350 SC, respectively. Individual toxicity tests showed that C. silvestrii was more sensitive to both carbofuran and diuron than P. caudatum. In single exposures, both pesticides caused adverse effects to C. silvestrii in environmentally relevant concentrations (48 h EC₅₀ = 0.001 mg L^-1 and 8 d LOEC = 0.00038 mg L^-1 for formulated carbofuran; 8 d LOEC < 0.05 mg L^-1 for formulated diuron). For P. caudatum, carbofuran and diuron in single exposures were only slightly toxic (24 h IC₅₀ = 5.1 mg L^-1 and 6.9 mg L^-1 for formulated carbofuran and diuron, respectively). Acute and chronic exposures to diuron and carbofuran mixtures caused significant deviations of the toxicity predicted by the Concentration Addition and Independent Action reference models for both test species. For the protozoan P. caudatum, a dose-dependent deviation was verified for mortality, with synergism caused mainly by carbofuran and antagonism caused mainly by diuron. For protozoan population growth, however, an antagonistic deviation was observed when the active ingredient mixtures were tested. In the case of C. silvestrii, antagonism at low concentrations and synergism at high concentrations were revealed after acute exposure to active ingredient mixtures, whereas for reproduction an antagonistic deviation was found. Commercial formulation mixtures presented significantly higher toxicity than the active ingredient mixtures. Our results showed that carbofuran and diuron interact and cause different toxic responses than those predicted by the individually tested compounds. Their mixture toxicity should therefore be considered in risk assessments as these pesticides are likely to be present simultaneously in edge-of-field waterbodies.

An up-to-date overview of computational polypharmacology in modern drug discovery

INTRODUCTION

In recent years, computational polypharmacology has gained significant attention to study the promiscuous nature of drugs. Despite tremendous challenges, community-wide efforts have led to a variety of novel approaches for predicting drug polypharmacology. In particular, some rapid advances using machine learning and artificial intelligence have been reported with great success.

AREAS COVERED

In this article, the authors provide a comprehensive update on the current state-of-the-art polypharmacology approaches and their applications, focusing on those reports published after our 2017 review article. The authors particularly discuss some novel, groundbreaking concepts, and methods that have been developed recently and applied to drug polypharmacology studies.

EXPERT OPINION

Polypharmacology is evolving and novel concepts are being introduced to counter the current challenges in the field. However, major hurdles remain including incompleteness of high-quality experimental data, lack of in vitro and in vivo assays to characterize multi-targeting agents, shortage of robust computational methods, and challenges to identify the best target combinations and design effective multi-targeting agents. Fortunately, numerous national/international efforts including multi-omics and artificial intelligence initiatives as well as most recent collaborations on addressing the COVID-19 pandemic have shown significant promise to propel the field of polypharmacology forward.

Predicting drug synergy for precision medicine using network biology and machine learning

Identification of effective drug combinations for patients is an expensive and time-consuming procedure, especially for in vitro experiments. To accelerate the synergistic drug discovery process, we present a new classification model to identify more effective anti-cancer drug pairs using in silico network biology approach. Based on the hypotheses that the drug synergy comes from the collective effects on the biological network, therefore, we developed six network biology features, including overlap and distance of drug perturbation network, that were derived by using individual drug-perturbed transcriptome profiles and the relevant biological network analysis. Using publicly available drug synergy databases and three machine-learning (ML) methods, the model was trained to discriminate the positive (synergistic) and negative (nonsynergistic) drug combinations. The proposed models were evaluated on the test cases to predict the most promising network biology feature, which is the network degree activity, i.e. the synergistic effect between drug pairs is mainly accounted by the complementary signaling pathways or molecular networks from two drugs.