DiffPROTACs is a deep learning-based generator for proteolysis targeting chimeras.

PROteolysis TArgeting Chimeras (PROTACs) has recently emerged as a promising technology. However, the design of rational PROTACs, especially the linker component, remains challenging due to the absence of structure-activity relationships and experimental data. Leveraging the structural characteristics of PROTACs, fragment-based drug design (FBDD) provides a feasible approach for PROTAC research. Concurrently, artificial intelligence-generated content has attracted considerable attention, with diffusion models and Transformers emerging as indispensable tools in this field. In response, we present a new diffusion model, DiffPROTACs, harnessing the power of Transformers to learn and generate new PROTAC linkers based on given ligands. To introduce the essential inductive biases required for molecular generation, we propose the O(3) equivariant graph Transformer module, which augments Transformers with graph neural networks (GNNs), using Transformers to update nodes and GNNs to update the coordinates of PROTAC atoms. DiffPROTACs effectively competes with existing models and achieves comparable performance on two traditional FBDD datasets, ZINC and GEOM. To differentiate the molecular characteristics between PROTACs and traditional small molecules, we fine-tuned the model on our self-built PROTACs dataset, achieving a 93.86% validity rate for generated PROTACs. Additionally, we provide a generated PROTAC database for further research, which can be accessed at https://bailab.siais.shanghaitech.edu.cn/service/DiffPROTACs-generated.tgz. The corresponding code is available at https://github.com/Fenglei104/DiffPROTACs and the server is at https://bailab.siais.shanghaitech.edu.cn/services/diffprotacs.

Leukotriene B4 receptor 1 (BLT1) does not mediate disease progression in a mouse model of liver fibrosis.

MASH is a prevalent liver disease that can progress to fibrosis, cirrhosis, hepatocellular carcinoma (HCC), and ultimately death, but there are no approved therapies. Leukotriene B4 (LTB4) is a potent pro-inflammatory chemoattractant that drives macrophage and neutrophil chemotaxis, and genetic loss or inhibition of its high affinity receptor, leukotriene B4 receptor 1 (BLT1), results in improved insulin sensitivity and decreased hepatic steatosis. To validate the therapeutic efficacy of BLT1 inhibition in an inflammatory and pro-fibrotic mouse model of MASH and fibrosis, mice were challenged with a choline-deficient, L-amino acid defined high fat diet and treated with a BLT1 antagonist at 30 or 90 mg/kg for 8 weeks. Liver function, histology, and gene expression were evaluated at the end of the study. Treatment with the BLT1 antagonist significantly reduced plasma lipids and liver steatosis but had no impact on liver injury biomarkers or histological endpoints such as inflammation, ballooning, or fibrosis compared to control. Artificial intelligence-powered digital pathology analysis revealed a significant reduction in steatosis co-localized fibrosis in livers treated with the BLT1 antagonist. Liver RNA-seq and pathway analyses revealed significant changes in fatty acid, arachidonic acid, and eicosanoid metabolic pathways with BLT1 antagonist treatment, however, these changes were not sufficient to impact inflammation and fibrosis endpoints. Targeting this LTB4-BLT1 axis with a small molecule inhibitor in animal models of chronic liver disease should be considered with caution, and additional studies are warranted to understand the mechanistic nuances of BLT1 inhibition in the context of MASH and liver fibrosis.

Enhancing clinical decision-making: A novel nomogram for stratifying cancer-specific survival in middle-aged individuals with follicular thyroid carcinoma utilizing SEER data.

Background: Thyroid cancer (TC) is the most common malignant tumor in the endocrine system, is also one of the head and neck tumor. Follicular Thyroid Carcinoma (FTC) plays an important role in the pathological classification of thyroid cancer. This study aimed to develop an innovative predictive tool, a nomogram, for predicting cancer specific survival (CSS) in middle-aged FTC patients. Methods: We collected patient data from the Surveillance, Epidemiology, and End Results (SEER) database. The data from patients between 2004 and 2015 were used as the training set, and the data from patients between 2016 and 2018 were used as the validation set. To identify independent risk factors affecting patient survival, univariate and multivariate Cox regression analyses were performed. Based on this, we developed a nomogram model aimed at predicting CSS in middle-aged patients with FTC. The consistency index (C-index), the area under the receiver operating characteristic (ROC) curve (AUC), and the calibration curve were used to evaluate the accuracy and confidence of the model. Results: A total of 2470 patients were enrolled in this study, in which patients from 2004 to 2015 were randomly assigned to the training cohort (N = 1437) and validation cohort (N = 598), and patients from 2016 to 2018 were assigned to the external validation cohort (N = 435) in terms of time. Univariate and multivariate Cox regression analysis showed that marriage, histological grade and TNM stage were independent risk factors for survival. The C-index for the training cohort was 0.866 (95 % CI: 0.805-0.927), for the validation cohort it was 0.944 (95 % CI: 0.903-0.985), and for the external validation cohort, it reached 0.999 (95 % CI: 0.997-1.001). Calibration curves and AUC suggest that the model has good accuracy. Conclusions: We developed an innovative nomogram to predict CSS in middle-aged patients with FTC. Our model after a rigorous internal validation and external validation process, based on the time proved that the high level of accuracy and reliability. This tool helps healthcare professionals and patients make informed clinical decisions.

Conformational Space Profiling Enhances Generic Molecular Representation for AI-Powered Ligand-Based Drug Discovery.

The molecular representation model is a neural network that converts molecular representations (SMILES, Graph) into feature vectors, and is an essential module applied across a wide range of artificial intelligence-driven drug discovery scenarios. However, current molecular representation models rarely consider the three-dimensional conformational space of molecules, losing sight of the dynamic nature of small molecules as well as the essence of molecular conformational space that covers the heterogeneity of molecule properties, such as the multi-target mechanism of action, recognition of different biomolecules, dynamics in cytoplasm and membrane. In this study, a new model named GeminiMol is proposed to incorporate conformational space profiles into molecular representation learning, which extracts the feature of capturing the complicated interplay between the molecular structure and the conformational space. Although GeminiMol is pre-trained on a relatively small-scale molecular dataset (39290 molecules), it shows balanced and superior performance not only on 67 molecular properties predictions but also on 73 cellular activity predictions and 171 zero-shot tasks (including virtual screening and target identification). By capturing the molecular conformational space profile, the strategy paves the way for rapid exploration of chemical space and facilitates changing paradigms for drug design.

Discovery and characterization of novel potent BCR-ABL degraders by conjugating allosteric inhibitor.

The oncogenic fusion protein BCR-ABL is the driving force of leukemogenesis in chronic myeloid leukemia (CML). Despite the great advance in CML treatment through the application of tyrosine kinase inhibitors (TKIs) against BCR-ABL, disease recurrence after TKI discontinuation and clinical resistance mainly due to BCR-ABL mutations continue to be an issue. Herein we report our efforts to synthesize a novel series of CRBN-recruiting proteolysis-targeting chimeras (PROTACs) targeting BCR-ABL based on the allosteric inhibitor asciminib. Our efforts have led to the discovery of compound 30 (SIAIS100) through extensive SAR studies by the optimization of linker parameters as well as linker attachment points of both target-binding warhead and CRBN ligands, which exhibited the most potent degradative activity with a DC50 value of 2.7 nM and Dmax of 91.2% against BCR-ABL and has an IC50 value of 12 nM in BCR-ABL + K562 cells. The binding model and the stability evaluation of 30-induced ternary complex formation were also elucidated through computational simulations. Furthermore, 30 induced sustained and robust BCR-ABL degradation and maintained the efficacy for 96 h post-washout. Moreover, the proteomics analysis showed that 30 degraded BCR-ABL and three CRBN's neo-substrates, including IKZF1, IKZF3, and ZFP91. Additionally, 30 also exerted degradative activity against a panel of clinically relevant resistance-conferring mutations of BCR-ABL, including gatekeeper mutation T315I, several single mutations associated with TKI resistance, and certain highly resistant compound mutations. Our study provided a deeper understanding of the development of PROTACs targeting BCR-ABL and novel potential therapeutic agents for CML treatment.