Binding Activity Classification of Anti-SARS-CoV-2 Molecules using Deep Learning Across Multiple Assays

Background: The coronavirus disease-2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2), has urgently necessitated effective therapeutic solutions, with a focus on rapidly identifying and classifying potential small-molecule drugs. Given traditional methods' labor-intensive and time-consuming nature, deep learning has emerged as an essential tool for efficiently processing and extracting insights from complex biological data. Aims: To utilize deep learning techniques, particularly deep neural networks (DNN) enhanced with the synthetic minority oversampling technique (SMOTE), to enhance the classification of binding activities in anti-SARS-CoV-2 molecules across various bioassays. Methods: We used 11 bioassay datasets covering various SARS-CoV-2 interactions and inhibitory mechanisms. These assays ranged from spike-ACE2 protein-protein interaction to ACE2 enzymatic activity and 3CL enzymatic activity. To address the prevalent class imbalance in these datasets, the SMOTE technique was employed to generate new samples for the minority class. In our model-building approach, we divided the dataset into 80% training and 20% test sets, reserving 10% of the training set for validation. Our approach involved employing a DNN that integrates ReLU and sigmoid activation functions, incorporates batch normalization, and uses Adam optimization. The hyperparameters and architecture of the DNN were optimized through various tests on layers, minibatch sizes, epoch sizes, and learning rates. A 40% dropout rate was incorporated to mitigate overfitting. For model evaluation, we computed performance metrics, such as balanced accuracy (BACC), precision, recall, F1 score, Matthews' correlation coefficient (MCC), and area under the curve (AUC). Results: The performance of the DNN across 11 bioassay test sets revealed varying outcomes, significantly influenced by the ratios of active-to-inactive compounds. Assays, such as AlphaLISA and CoV-PPE, demonstrated robust performance across various metrics, including BACC, precision, recall, and AUC, when configured with more balanced ratios (1:3 and 1:1, respectively). This suggests the effective identification of active compounds in both cases. In contrast, assays with higher imbalance ratios, such as 3CL (1:38) and cytopathic effect (1:15), demonstrated higher recall but lower precision, highlighting challenges in accurately identifying active compounds among numerous inactive compounds. However, even in these challenging settings, the model achieved favorable BACC and recall scores. Overall, the DNN model generally performed well, as indicated by the BACC, MCC, and AUC values, especially when considering the degree of dataset imbalance in each assay. Conclusion: This study demonstrates the significant impact of deep learning, particularly DNN models enhanced with SMOTE, in improving the identification of active compounds in bioassay datasets for COVID-19 drug discovery, outperforming traditional machine learning models. Furthermore, this study highlights the efficacy of advanced computational techniques in addressing high-throughput screening data imbalances.

Effects of prasugrel on membrane potential and contractile activity of rat ventricular myocytes.

BACKGROUND: Though prasugrel is one of the important P2Y12 inhibitors currently in use for antiplatelet therapy, its potential effects on contractility and electrical activity of ventricular myocytes have not yet been investigated. Hence this study was designed to study the impact of prasugrel on contractile function and membrane potential of isolated ventricular myocytes. METHODS: Freshly isolated rat ventricular myocytes were used in this study. Myocyte contraction was measured during electrical stimulation of cardiomyocytes and the action potential (AP) recordings were obtained with current clamp mode of the patch-clamp amplifier. RESULTS: AP duration and fractional shortening of ventricular myocytes did not show any change with the administration of 1µM of prasugrel. However, remarkable depolarization of resting membrane potential followed by apparent fibrillation episodes was detected in the cardiomyocytes. Similar events were observed in the contractile activity of myocytes during field stimulation. Also, a higher concentration of prasugrel (10µM) elicited repeated fibrillations, which disappeared after washout or nitric oxide synthase (NOS) inhibition with L-NAME. In contrast, the same concentration of ticagrelor, another P2Y12 inhibitor did not induce fibrillation events though it decreased the contractility of ventricular myocytes significantly. The perfusion of ventricular myocytes with L-NAME did not alter the negative inotropic effect of ticagrelor. CONCLUSIONS: Prasugrel, a widely used antithrombotic agent, may induce depolarization in the membrane potential of myocytes as well as fibrillation via NO mediated pathway.

Rho-kinase inhibition reverses impaired Ca2+ handling and associated left ventricular dysfunction in pressure overload-induced cardiac hypertrophy.

Recent studies have implicated a relationship between RhoA/ROCK activity and defective Ca2+ homeostasis in hypertrophic hearts. This study investigated molecular mechanism underlying ROCK inhibition-mediated cardioprotection against pressure overload-induced cardiac hypertrophy, with a focus on Ca2+ homeostasis. Cardiac hypertrophy model was established by performing transverse aortic constriction (TAC) in 8-week-old male rats. Groups were assigned as SHAM, TAC and TAC+Fas (rats undergoing TAC and treated with fasudil). Rats in the TAC+Fas group were administered fasudil (5mg/kg/day), and rats in the SHAM and TAC groups were treated with vehicle for 10 weeks. Electrophysiological recordings were obtained from isolated left ventricular myocytes and expression levels of proteins were determined using western blotting. Rats in the TAC group showed remarkable cardiac hypertrophy, and fasudil treatment significantly reversed this alteration. TAC+Fas myocytes showed significant improvement in reduced contractility and Ca2+ transients. Moreover, these myocytes showed restoration of slow relaxation rate and Ca2+ reuptake. Although L-type Ca2+ currents did not change in TAC group, there was a significant reduction in the triggered Ca2+ transients which was reversed either by long-term fasudil treatment or incubation of TAC myocytes with fasudil. The hearts of rats in the TAC group showed a significant decrease in ROCK1, ROCK2, RyR2 protein levels and p-PLBS16/T17/SERCA2 ratio and increase in RhoA expression and MLC phosphorylation. However, fasudil treatment largely reversed TAC-induced alterations in protein expression. Thus, our findings indicate that upregulation of the RhoA/ROCK pathway is significantly associated with cardiac hypertrophy-related Ca2+ dysregulation and suggest that ROCK inhibition prevents hypertrophic heart failure.

The Effect of a High-Protein Diet and Exercise on Cardiac AQP7 and GLUT4 Gene Expression.

High-protein (HP) diets are commonly consumed by athletes despite their potential health hazard, which is postulated to enforce a negative effect on bone and renal health. However, its effects on heart have not been known yet. Aquaporin-7 (AQP7) is an aquaglyceroporin that facilitates glycerol and water transport. Glycerol is an important cardiac energy production substrate, especially during exercise, in conjunction with fatty acids and glucose. Glucose transporter 4 (GLUT4) is an insulin-sensitive glucose transporter in heart. We aimed to investigate the effect of HPD on AQP7 and GLUT4 levels in the rat heart subjected to exercise. Male Sprague-Dawley rats were divided into control (n = 12), exercise (E) training (n = 10), HPD (n = 12), and HPD-E training (n = 9) groups. The HPD groups were fed a 45 % protein-containing diet 5 weeks. The HPD-E and E groups were performed the treadmill exercise during the 5-week study period. Real-time polymerase chain reaction and immunohistochemistry techniques were used to determine the gene expression and localization of AQP7 and GLUT4 in heart tissue. Results of relative gene expression were calculated by the 'Pfaffl' mathematical method using the REST program. Differences in AQP7 and GLUT4 gene expression were expressed as fold change compared to the control group. Heart weight/tibia ratio and ventricular wall thickness were evaluated as markers of cardiac hypertrophy. Further, serum glucose, glycerol, and insulin levels were also measured. AQP7 gene expression was found to be increased in the E (3.47-fold, p < 0.001), HPD (5.59-fold, p < 0.001), and HPD-E (3.87-fold, p < 0.001) groups compared to the control group. AQP7 protein expression was also increased in the HPD and HPD-E groups (p < 0.001). Additionally, cardiac mRNA expression levels of GLUT4 showed a significant increase in the E (2.16-fold, p < 0.003), HPD (7.14-fold, p < 0.001), and HPD-E (3.43-fold, p < 0.001) groups compared to the control group. GLUT4 protein expression was significantly increased in the E, HPD, and HPD-E groups compared to the control group (p = 0.024, p < 0.001, and p < 0.001, respectively). Furthermore, Serum glucose levels were significantly different between groups (p < 0.005). This difference was observed between the HPD groups and normal-protein diet groups (C and E). Serum insulin levels were higher for HPD groups compared with the normal-protein diet groups (p < 0.001), whereas no differences were observed between the exercise and sedentary groups (p = 0.111). Serum glycerol levels were significantly increased in the HPD groups compared with control and E groups (p < 0.05 and p < 0.05, respectively). Consumption of HPD supplementation caused the increased effects on AQP7 and GLUT4 expression in rat heart.

2.1 GHz electromagnetic field does not change contractility and intracellular Ca2+ transients but decreases ß-adrenergic responsiveness through nitric oxide signaling in rat ventricular myocytes.

PURPOSE: Due to the increasing use of wireless technology in developing countries, particularly mobile phones, the influence of electromagnetic fields (EMF) on biologic systems has become the subject of an intense debate. Therefore, in this study we investigated the effect of 2.1 GHz EMF on contractility and beta-adrenergic (ß-AR) responsiveness of ventricular myocytes. MATERIALS AND METHODS: Rats were randomized to the following groups: Sham rats (SHAM) and rats exposed to 2.1 GHz EMF for 2 h/day for 10 weeks (EM-10). Sarcomere shortening and Ca(2+) transients were recorded in isolated myocytes loaded with Fura2-AM and electrically stimulated at 1 Hz, while L-type Ca(2+) currents (I(CaL)) were measured using whole-cell patch clamping at 36 ± 1°C. Cardiac nitric oxide (NO) levels were measured in tissue samples using a colorimetric assay kit. RESULTS: Fractional shortening and amplitude of the matched Ca(2+) transients were not changed in EM-10 rats. Although the isoproterenol-induced (10(-6) M) I(CaL) response was reduced in rats exposed to EMF, basal I(CaL) density in myocytes was similar between the two groups (p < 0.01). Moreover, EMF exposure led to a significant increase in nitric oxide levels in rat heart (p < 0.02). CONCLUSIONS: Long-term exposure to 2.1 GHz EMF decreases ß-AR responsiveness of ventricular myocytes through NO signaling.

Calcitonin gene related peptide gene polymorphism in migraine patients.

OBJECTIVE: Calcitonin gene related peptide (CGRP), which has a vasodilator effect, is held responsible for neurogenic inflammation and vasodilatation of the cranial vessels in migraine pathophysiology. In this study, we investigated the association between alpha CGRP gene polymorphism (CALCA T-692C) and migraine. MATERIAL AND METHODS: One hundred and thirty-four female migraineurs and 96 healthy female cases were enrolled in the study. The patient group was further subdivided into migraine with and without aura groups. The CALCA T-692C gene polymorphism was identified using polymerase chain reaction (PCR) technique and restriction fragment length polymorphism (RFLP). RESULTS: The genotype and allele frequencies of CALCA T-692C gene polymorphism did not differ between the migraine and control groups. Between the migraine with and without aura subgroups, there was no difference. No association was seen between the CALCA T-692C gene polymorphisms and migraine attack severity and frequency. CONCLUSION: Our study did not show any association between CALCA T-692C gene polymorphism and migraine.