Application of Generative Artificial Intelligence in Predicting Membrane Partitioning of Drugs: Combining Denoising Diffusion Probabilistic Models and MD Simulations Reduces the Computational Cost to One-Third.

The optimal interaction of drugs with plasma membranes and membranes of subcellular organelles is a prerequisite for desirable pharmacology. Importantly, for drugs targeting the transmembrane lipid-facing sites of integral membrane proteins, the relative affinity of a drug to the bilayer lipids compared to the surrounding aqueous phase affects the partitioning, access, and binding of the drug to the target site. Molecular dynamics (MD) simulations, including enhanced sampling techniques such as steered MD, umbrella sampling (US), and metadynamics, offer valuable insights into the interactions of drugs with the membrane lipids and water in atomistic detail. However, these methods are computationally prohibitive for the high-throughput screening of drug candidates. This study shows that applying denoising diffusion probabilistic models (DDPMs), a generative AI method, to US simulation data reduces the computational cost significantly. Specifically, the models used only partial (one-third) data from the US simulations and reproduced the complete potential of mean force (PMF) profiles for three FDA-approved drugs (ß2-adrenergic agonists) and ∼20 biologically relevant chemicals with known experimentally characterized bilayer locations. Intriguingly, the model can predict the solvation-free energies for partitioning and crossing the bilayer, preferred bilayer locations (low-energy well), and orientations of the ligands with high accuracy. The results indicate that DDPMs can be used to characterize the complete membrane partitioning profile of drug molecules using fewer umbrella sampling simulations at select positions along the bilayer normal (z-axis), irrespective of their amphiphilic-lipophilic-cephalophilic characteristics.

Promiscuity and Quantitative Contribution of UGT2B17 in Drug and Steroid Metabolism Determined by Experimental and Computational Approaches.

Uridine 5'-diphospho-glulcuronosyltransferase 2B17 (UGT2B17) is important in the metabolism of steroids and orally administered drugs due to its high interindividual variability. However, the structural basis governing the substrate selectivity or inhibition of UGT2B17 remains poorly understood. This study investigated 76 FDA-approved drugs and 20 steroids known to undergo glucuronidation for their metabolism by UGT2B17. Specifically, we assessed the substrate selectivity for UGT2B17 over other UGT enzymes using recombinant human UGT2B17 (rUGT2B17), human intestinal microsomes, and human liver microsomes. The quantitative contribution of intestinal UGT2B17 in the glucuronidation of these compounds was characterized using intestinal microsomes isolated from UGT2B17 expressors and nonexpressors. In addition, a structure-based pharmacophore model for UGT2B17 substrates was built and validated using the studied pool of substrates and nonsubstrates. The results show that UGT2B17 could metabolize 23 out of 96 compounds from various chemical classes, including alcohols and carboxylic acids, particularly in the intestine. Interestingly, amines were less susceptible to UGT2B17 metabolism, though they could inhibit the enzyme. Three main pharmacophoric features of UGT2B17 substrates include (1) the presence of an accessible -OH or -COOH group near His35 residue, (2) a hydrophobic functional group at ∼4.5-5 Šfrom feature 1, and (3) an aromatic ring ∼5-7 Šfrom feature 2. Most of the studied compounds inhibited UGT2B17 activity irrespective of their substrate potential, indicating the possibility of multiple mechanisms. These data suggest that UGT2B17 is promiscuous in substrate selectivity and inhibition and has a high potential to produce significant variability in the absorption and disposition of orally administered drugs.

Human beta defensin-3 mediated activation of ß-catenin during human respiratory syncytial virus infection: interaction of HBD3 with LDL receptor-related protein 5.

Respiratory Syncytial Virus (RSV) is a non-segmented negative-sense RNA virus belonging to the paramyxovirus family. RSV infects the respiratory tract to cause pneumonia and bronchiolitis in infants, elderly, and immunocompromised patients. Effective clinical therapeutic options and vaccines to combat RSV infection are still lacking. Therefore, to develop effective therapeutic interventions, it is imperative to understand virus-host interactions during RSV infection. Cytoplasmic stabilization of ß-catenin protein results in activation of canonical Wingless (Wnt)/ß-catenin signaling pathway that culminates in transcriptional activation of various genes regulated by T-cell factor/lymphoid enhancer factor (TCF/LEF) transcription factors. This pathway is involved in various biological and physiological functions. Our study shows RSV infection of human lung epithelial A549 cells triggering ß-catenin protein stabilization and induction of ß-catenin mediated transcriptional activity. Functionally, the activated ß-catenin pathway promoted a pro-inflammatory response during RSV infection of lung epithelial cells. Studies with ß-catenin inhibitors and A549 cells lacking optimal ß-catenin activity demonstrated a significant loss of pro-inflammatory chemokine interleukin-8 (IL-8) release from RSV-infected cells. Mechanistically, our studies revealed a role of extracellular human beta defensin-3 (HBD3) in interacting with cell surface Wnt receptor LDL receptor-related protein-5 (LRP5) to activate the non-canonical Wnt independent ß-catenin pathway during RSV infection. We showed gene expression and release of HBD3 from RSV-infected cells and silencing of HBD3 expression resulted in reduced stabilization of ß-catenin protein during RSV infection. Furthermore, we observed the binding of extracellular HBD3 with cell surface localized LRP5 protein, and our in silico and protein-protein interaction studies have highlighted a direct interaction of HBD3 with LRP5. Thus, our studies have identified the ß-catenin pathway as a key regulator of pro-inflammatory response during RSV infection of human lung epithelial cells. This pathway was induced during RSV infection via a non-canonical Wnt-independent mechanism involving paracrine/autocrine action of extracellular HBD3 activating cell surface Wnt receptor complex by directly interacting with the LRP5 receptor.

Molecular basis for the recognition of 24-(S)-hydroxycholesterol by integrin αvß3.

A growing body of evidence suggests that oxysterols such as 25-hydroxycholesterol (25HC) are biologically active and involved in many physiological and pathological processes. Our previous study demonstrated that 25HC induces an innate immune response during viral infections by activating the integrin-focal adhesion kinase (FAK) pathway. 25HC produced the proinflammatory response by binding directly to integrins at a novel binding site (site II) and triggering the production of proinflammatory mediators such as tumor necrosis factor-α (TNF) and interleukin-6 (IL-6). 24-(S)-hydroxycholesterol (24HC), a structural isomer of 25HC, plays a critical role in cholesterol homeostasis in the human brain and is implicated in multiple inflammatory conditions, including Alzheimer's disease. However, whether 24HC can induce a proinflammatory response like 25HC in non-neuronal cells has not been studied and remains unknown. The aim of this study was to examine whether 24HC produces such an immune response using in silico and in vitro experiments. Our results indicate that despite being a structural isomer of 25HC, 24HC binds at site II in a distinct binding mode, engages in varied residue interactions, and produces significant conformational changes in the specificity-determining loop (SDL). In addition, our surface plasmon resonance (SPR) study reveals that 24HC could directly bind to integrin αvß3, with a binding affinity three-fold lower than 25HC. Furthermore, our in vitro studies with macrophages support the involvement of FAK and NFκB signaling pathways in triggering 24HC-mediated production of TNF. Thus, we have identified 24HC as another oxysterol that binds to integrin αvß3 and promotes a proinflammatory response via the integrin-FAK-NFκB pathway.