WADDAICA: A webserver for aiding protein drug design by artificial intelligence and classical algorithm

Redefining a new frontier in alkaptonuria therapy with AI-driven drug candidate design via in- silico innovation

A rare metabolic condition called alkaptonuria (AKU) is caused by a decrease in homogentisate 1,2 dioxygenase (HGO) activity due to a mutation in homogentisate dioxygenase (HGD) gene. Homogentisic acid is a byproduct of the catabolism of tyrosine and phenylalanine that darkens the urine and accumulates in connective tissues which causes an agonizing arthritis. Employing the use of deep learning artificial intelligence (AI) drug design, this study aims to alleviate the current toxicity of the AKU drugs currently in use, particularly nitisinone, by utilizing the natural flavanol kaempferol molecule as a 4-hydroxyphenylpyruvate dioxygenase inhibitor. Kaempferol was employed to generate three effective de novo drug candidates targeting the enzyme 4-hydroxyphenylpyruvate dioxygenase using an AI drug design tool. We present novel AIK formulations in the present study. The AIK's (Artificial Intelligence Kaempferol) examination of drug-likeliness among the three led to its choice as a possible target. The toxicity assessment research of AIK demonstrates that it is not only safer to use than other treatments, but also more efficient. The docking of the AIGT with 4-hydroxyphenylpyruvate dioxygenase, which revealed a binding affinity of around -9.099 kcal/mol, highlights the AIK's potential as a therapeutic candidate. An innovative approach to deal with challenging circumstances is thus presented in this study by new formulations kaempferol that have been meticulously designed by AI. The results of the in vitro tests must be confirmed in vivo, even though AI-designed AIK is effective and sufficiently safe as computed.

Discovery of novel PARP-1 inhibitors using tandem in silico studies: integrated docking, e-pharmacophore, deep learning based de novo and molecular dynamics simulation approach

Cancer accounts for the majority of deaths worldwide, and the increasing incidence of breast cancer is a matter of grave concern. Poly (ADP-ribose) polymerase-1 (PARP-1) has emerged as an attractive target for the treatment of breast cancer as it has an important role in DNA repair. The focus of the study was to identify novel PARP-1 inhibitors using a blend of tandem structure-based screening (Docking and e-pharmacophore-based screening) and artificial intelligence (deep learning)-based de novo approaches. The scrutiny of compounds having good binding characteristics for PARP-1 was carried out using a tandem mode of screening along with parameters such as binding energy and ADME analysis. The efforts afforded compound Vab1 (PubChem ID 129142036), which was chosen as a seed for obtaining novel compounds through a trained artificial intelligence (AI)-based model. Resultant compounds were assessed for PARP-1 inhibition; binding affinity prediction and interaction pattern analysis were carried out using the extra precision (XP) mode of docking. Two best hits, Vab1-b and Vab1-g, exhibiting good dock scores and suitable interactions, were subjected to 100 nanoseconds (ns) of molecular dynamics simulation in the active site of PARP-1 and compared with the reference Protein-Ligand Complex. The stable nature of PARP-1 upon binding to these compounds was revealed through MD simulation.Communicated by Ramaswamy H. Sarma.

Study on the molecular mechanism of dihydromyricetin in alleviating liver cirrhosis based on network pharmacology

Dihydromyricetin (DHM) is a bioactive flavonoid extracted from Hovenia dulcis, which has various activities. In the present study, the molecular mechanism of dihydromyricetin (DHM) in relieving liver cirrhosis was investigated through network pharmacology and experimental verification. The cell model was induced by TGF-β1 activating the human hepatic stellate cell line (HSC; LX-2). The protein levels of α-SMA, collagen I, and collagen III and pathway-related proteins within LX-2 cells were detected using Western blot. EdU staining was conducted to detect cell proliferation. Immunofluorescence staining was performed to detect the expression levels of α-SMA and collagen I. Next, the drug targets of DHM were screened from the PubChem database. The differentially expressed genes in the liver cirrhosis dataset GSE14323 were identified. The expression of the identified drug targets in LX-2 cells was verified using qRT-PCR. The results showed that TGF-β1 treatment notably increased LX-2 cell viability, promoted cell proliferation, and elevated α-SMA, collagen I, and collagen III protein contents. DHM treatment could partially eliminate TGF-β1 effects, as evidenced by the inhibited cell viability and proliferation and reduced α-SMA, collagen I, and collagen III contents. After network pharmacology analysis, nine differentially expressed target genes (MMP2, PDGFRB, PARP1, BCL2L2, ABCB1, TYR, CYP2E1, SQSTM1, and IL6) in liver cirrhosis were identified. According to qRT-PCR verification, DHM could inhibit the expression of MMP2, PDGFRB, PARP1, CYP2E1, SQSTM1, and IL6, and enhance ABCB1 expression levels within LX-2 cells. Moreover, DHM inhibited mTOR and MAPK signaling pathways in TGF-β1-induced HSCs. In conclusion, DHM could inhibit HSC activation, which may be achieved via acting on MMP2, PDGFRB, PARP1, CYP2E1, SQSTM1, IL6, and ABCB1 genes and their downstream signaling pathways, including mTOR and MAPK signaling pathway.

Artificial Intelligence Assisted Pharmacophore Design for Philadelphia Chromosome-Positive Leukemia with Gamma-Tocotrienol: A Toxicity Comparison Approach with Asciminib

BCR-ABL1 is a fusion protein as a result of a unique chromosomal translocation (producing the so-called Philadelphia chromosome) that serves as a clinical biomarker primarily for chronic myeloid leukemia (CML); the Philadelphia chromosome also occurs, albeit rather rarely, in other types of leukemia. This fusion protein has proven itself to be a promising therapeutic target. Exploiting the natural vitamin E molecule gamma-tocotrienol as a BCR-ABL1 inhibitor with deep learning artificial intelligence (AI) drug design, this study aims to overcome the present toxicity that embodies the currently provided medications for (Ph+) leukemia, especially asciminib. Gamma-tocotrienol was employed in an AI server for drug design to construct three effective de novo drug compounds for the BCR-ABL1 fusion protein. The AIGT’s (Artificial Intelligence Gamma-Tocotrienol) drug-likeliness analysis among the three led to its nomination as a target possibility. The toxicity assessment research comparing AIGT and asciminib demonstrates that AIGT, in addition to being more effective nonetheless, is also hepatoprotective. While almost all CML patients can achieve remission with tyrosine kinase inhibitors (such as asciminib), they are not cured in the strict sense. Hence it is important to develop new avenues to treat CML. We present in this study new formulations of AIGT. The docking of the AIGT with BCR-ABL1 exhibited a binding affinity of −7.486 kcal/mol, highlighting the AIGT’s feasibility as a pharmaceutical option. Since current medical care only exclusively cures a small number of patients of CML with utter toxicity as a pressing consequence, a new possibility to tackle adverse instances is therefore presented in this study by new formulations of natural compounds of vitamin E, gamma-tocotrienol, thoroughly designed by AI. Even though AI-designed AIGT is effective and adequately safe as computed, in vivo testing is mandatory for the verification of the in vitro results.

DTiGNN: Learning drug-target embedding from a heterogeneous biological network based on a two-level attention-based graph neural network

MOTIVATION

In vitro experiment-based drug-target interaction (DTI) exploration demands more human, financial and data resources. In silico approaches have been recommended for predicting DTIs to reduce time and cost. During the drug development process, one can analyze the therapeutic effect of the drug for a particular disease by identifying how the drug binds to the target for treating that disease. Hence, DTI plays a major role in drug discovery. Many computational methods have been developed for DTI prediction. However, the existing methods have limitations in terms of capturing the interactions via multiple semantics between drug and target nodes in a heterogeneous biological network (HBN).

METHODS

In this paper, we propose a DTiGNN framework for identifying unknown drug-target pairs. The DTiGNN first calculates the similarity between the drug and target from multiple perspectives. Then, the features of drugs and targets from each perspective are learned separately by using a novel method termed an information entropy-based random walk. Next, all of the learned features from different perspectives are integrated into a single drug and target similarity network by using a multi-view convolutional neural network. Using the integrated similarity networks, drug interactions, drug-disease associations, protein interactions and protein-disease association, the HBN is constructed. Next, a novel embedding algorithm called a meta-graph guided graph neural network is used to learn the embedding of drugs and targets. Then, a convolutional neural network is employed to infer new DTIs after balancing the sample using oversampling techniques.

RESULTS

The DTiGNN is applied to various datasets, and the result shows better performance in terms of the area under receiver operating characteristic curve (AUC) and area under precision-recall curve (AUPR), with scores of 0.98 and 0.99, respectively. There are 23,739 newly predicted DTI pairs in total.

Developing novel antifungals: lessons from G protein-coupled receptors

Up to 1.5 million people die yearly from fungal disease, but the repertoire of antifungal drug classes is minimal and the incidence of drug resistance is rising rapidly. This dilemma was recently declared by the World Health Organization as a global health emergency, but the discovery of new antifungal drug classes remains excruciatingly slow. This process could be accelerated by focusing on novel targets, such as G protein-coupled receptor (GPCR)-like proteins, that have a high likelihood of being druggable and have well-defined biology and roles in disease. We discuss recent successes in understanding the biology of virulence and in structure determination of yeast GPCRs, and highlight new approaches that might pay significant dividends in the urgent search for novel antifungal drugs.

Learning with uncertainty to accelerate the discovery of histone lysine-specific demethylase 1A (KDM1A/LSD1) inhibitors

Machine learning including modern deep learning models has been extensively used in drug design and screening. However, reliable prediction of molecular properties is still challenging when exploring out-of-domain regimes, even for deep neural networks. Therefore, it is important to understand the uncertainty of model predictions, especially when the predictions are used to guide further experiments. In this study, we explored the utility and effectiveness of evidential uncertainty in compound screening. The evidential Graphormer model was proposed for uncertainty-guided discovery of KDM1A/LSD1 inhibitors. The benchmarking results illustrated that (i) Graphormer exhibited comparative predictive power to state-of-the-art models, and (ii) evidential regression enabled well-ranked uncertainty estimates and calibrated predictions. Subsequently, we leveraged time-splitting on the curated KDM1A/LSD1 dataset to simulate out-of-distribution predictions. The retrospective virtual screening showed that the evidential uncertainties helped reduce false positives among the top-acquired compounds and thus enabled higher experimental validation rates. The trained model was then used to virtually screen an independent in-house compound set. The top 50 compounds ranked by two different ranking strategies were experimentally validated, respectively. In general, our study highlighted the importance to understand the uncertainty in prediction, which can be recognized as an interpretable dimension to model predictions.

New Treatment for Type 2 Diabetes Mellitus Using a Novel Bipyrazole Compound

2',3,3,5'-Tetramethyl-4'-nitro-2'H-1,3'-bipyrazole (TMNB) is a novel bipyrazole compound with unknown therapeutic potential in diabetes mellitus. This study aims to investigate the anti-diabetic effects of TMNB in a high-fat diet and streptozotocin-(HFD/STZ)-induced rat model of type 2 diabetes mellitus (T2D). Rats were fed HFD, followed by a single low dose of STZ (40 mg/kg). HFD/STZ diabetic rats were treated orally with TMNB (10 mg/kg) or (200 mg/kg) metformin for 10 days before terminating the experiment and collecting plasma, soleus muscle, adipose tissue, and liver for further downstream analysis. TMNB reduced the elevated levels of serum glucose in diabetic rats compared to the vehicle control group (p < 0.001). TMNB abrogated the increase in serum insulin in the treated diabetic group compared to the vehicle control rats (p < 0.001). The homeostasis model assessment of insulin resistance (HOMA-IR) was decreased in the diabetic rats treated with TMNB compared to the vehicle controls. The skeletal muscle and adipose tissue protein contents of GLUT4 and AMPK were upregulated following treatment with TMNB (p < 0.001, < 0.01, respectively). TMNB was able to upregulate GLUT2 and AMPK protein expression in liver (p < 0.001, < 0.001, respectively). LDL, triglyceride, and cholesterol were reduced in diabetic rats treated with TMNB compared to the vehicle controls (p < 0.001, 0.01, respectively). TMNB reduced MDA and IL-6 levels (p < 0.001), and increased GSH level (p < 0.05) in diabetic rats compared to the vehicle controls. Conclusion: TMNB ameliorates insulin resistance, oxidative stress, and inflammation in a T2D model. TMNB could represent a promising therapeutic agent to treat T2D.

Resveratrol-Loaded Glutathione-Coated Collagen Nanoparticles Attenuate Acute Seizures by Inhibiting HMGB1 and TLR-4 in the Hippocampus of Mice

Epilepsy is a relatively complicated neurological disorder that results in seizures. The use of resveratrol in treating seizures has been reported in recent studies. However, the low bioavailability of resveratrol and the difficulty of reaching the targeted location in the brain reduce its efficacy considerably. The side effects due to the higher concentration of drugs are another matter of concern. The purpose of the present study is to enhance the antiepileptic potential of resveratrol by delivering it to the brain's targeted location by encapsulating it in glutathione-coated collagen nanoparticles. The collagen nanoparticles increase the bioavailability of resveratrol, while the transport of resveratrol to its target location in the brain is facilitated by glutathione. By encapsulating resveratrol in glutathione-coated collagen nanoparticles, the concentration also substantially decreases. Resveratrol encapsulated in synthesized nanoparticles is referred to as nanoresveratrol. In the present study, nanoresveratrol effectiveness was studied through PTZ-induced seizures (PTZ-IS) and the increasing current electroshock (ICES) test. The efficacy of nanoresveratrol was further established using biochemical analysis, histopathological examinations, ELISA and real-time-PCR tests, and immunohistochemistry examination of the hippocampus of mice. Hence, this study is unique in the sense that it synthesized nanoresveratrol by using glutathione-coated collagen nanoparticles, followed by its application to treating seizures. On the basis of the study results, nanoresveratrol was found to be effective in preventing cognitive impairment in the mice and controlling epilepsy seizures to a greater extent than resveratrol. The proposed nanoformulation also reduces the concentration of resveratrol considerably. The present study results show that even 0.4 mg/kg of nanoresveratrol outperforms 40 mg/kg of resveratrol.

Computationally prioritized drugs inhibit SARS-CoV-2 infection and syncytia formation

The pharmacological arsenal against the COVID-19 pandemic is largely based on generic anti-inflammatory strategies or poorly scalable solutions. Moreover, as the ongoing vaccination campaign is rolling slower than wished, affordable and effective therapeutics are needed. To this end, there is increasing attention toward computational methods for drug repositioning and de novo drug design.

Here, multiple data-driven computational approaches are systematically integrated to perform a virtual screening and prioritize candidate drugs for the treatment of COVID-19. From the list of prioritized drugs, a subset of representative candidates to test in human cells is selected. Two compounds, 7-hydroxystaurosporine and bafetinib, show synergistic antiviral effects in vitro and strongly inhibit viral-induced syncytia formation. Moreover, since existing drug repositioning methods provide limited usable information for de novo drug design, the relevant chemical substructures of the identified drugs are extracted to provide a chemical vocabulary that may help to design new effective drugs.

Machine learning & deep learning in data-driven decision making of drug discovery and challenges in high-quality data acquisition in the pharmaceutical industry

Predicting novel small molecule bioactivities for the target deconvolution, hit-to-lead optimization in drug discovery research, requires molecular representation. Previous reports have demonstrated that machine learning (ML) and deep learning (DL) have substantial implications in virtual screening, peptide synthesis, drug ADMET screening and biomarker discovery. These strategies can increase the positive outcomes in the drug discovery process without false-positive rates and can be achieved in a cost-effective way with a minimum duration of time by high-quality data acquisition. This review substantially discusses the recent updates in AI tools as cheminformatics application in medicinal chemistry for the data-driven decision making of drug discovery and challenges in high-quality data acquisition in the pharmaceutical industry while improving small-molecule bioactivities and properties.

Structure based prediction of a novel GPR120 antagonist based on pharmacophore screening and molecular dynamics simulations

•

Hypothesis of the important residues in conserving the GPR120S “ionic-lock”.

•

Computational model targeting W277 and N313 for virtual screening of GPR120S ligands.

•

Cpd 9 emerged as a potential GPR120S antagonist and anti-cancer treatment.

The G-protein coupled receptor, GPR120, has ubiquitous expression and multifaceted roles in modulating metabolic and anti-inflammatory processes. Recent implications of its role in cancer progression have presented GPR120 as an attractive oncogenic drug target. GPR120 gene knockdown in breast cancer studies revealed a role of GPR120-induced chemoresistance in epirubicin and cisplatin-induced DNA damage in tumour cells. Higher expression and activation levels of GPR120 is also reported to promote tumour angiogenesis and cell migration in colorectal cancer. Some agonists targeting GPR120 have been reported, such as TUG891 and Compound39, but to date development of small-molecule inhibitors of GPR120 is limited.

Herein, following homology modelling of the receptor a pharmacophore hypothesis was derived from 300 ns all-atomic molecular dynamics (MD) simulations on apo, TUG891-bound and Compound39-bound GPR120S (short isoform) receptor models embedded in a water solvated lipid bilayer system. We performed comparative MD analysis on protein–ligand interactions between the two agonist and apo simulations on the stability of the “ionic lock” – a Class A GPCRs characteristic of receptor activation and inactivation. The detailed analysis predicted that ligand interactions with W277 and N313 are critical to conserve the “ionic-lock” conformation (R136 of Helix 3) and prevent GPR120S receptor activation. The results led to generation of a W277 and N313 focused pharmacophore hypothesis and the screening of the ZINC15 database using ZINCPharmer through the structure-based pharmacophore.

100 ns all-atomic molecular dynamics (MD) simulations were performed on 9 small molecules identified and Cpd 9, (2-hydroxy-N-{4-[(6-hydroxy-2-methylpyrimidin-4-yl) amino] phenyl} benzamide) was predicted to be a small-molecule GPR120S antagonist. The conformational results from the collective all-atomic MD analysis provided structural information for further identification and optimisation of novel druggable inhibitors of GPR120S using this rational design approach, which could have future potential for anti-cancer drug development studies.