Bayesian Machine Learning Enables Identification of Transcriptional Network Disruptions Associated with Drug-Resistant Prostate Cancer

NetActivity enhances transcriptional signals by combining gene expression into robust gene set activity scores through interpretable autoencoders

Grouping gene expression into gene set activity scores (GSAS) provides better biological insights than studying individual genes. However, existing gene set projection methods cannot return representative, robust, and interpretable GSAS. We developed NetActivity, a machine learning framework that generates GSAS based on a sparsely-connected autoencoder, where each neuron in the inner layer represents a gene set. We proposed a three-tier training that yielded representative, robust, and interpretable GSAS. NetActivity model was trained with 1518 GO biological processes terms and KEGG pathways and all GTEx samples. NetActivity generates GSAS robust to the initialization parameters and representative of the original transcriptome, and assigned higher importance to more biologically relevant genes. Moreover, NetActivity returns GSAS with a more consistent definition and higher interpretability than GSVA and hipathia, state-of-the-art gene set projection methods. Finally, NetActivity enables combining bulk RNA-seq and microarray datasets in a meta-analysis of prostate cancer progression, highlighting gene sets related to cell division, key for disease progression. When applied to metastatic prostate cancer, gene sets associated with cancer progression were also altered due to drug resistance, while a classical enrichment analysis identified gene sets irrelevant to the phenotype. NetActivity is publicly available in Bioconductor and GitHub.

Emerging role of MYB transcription factors in cancer drug resistance

Decades ago, the viral myeloblastosis oncogene v-myb was identified as a gene responsible for the development of avian leukemia. However, the relevance of MYB proteins for human cancer diseases, in particular for solid tumors, remained basically unrecognized for a very long time. The human family of MYB transcription factors comprises MYB (c-MYB), MYBL2 (b-MYB), and MYBL1 (a-MYB), which are overexpressed in several cancers and are associated with cancer progression and resistance to anticancer drugs. In addition to overexpression, the presence of activated MYB-fusion proteins as tumor drivers was described in certain cancers. The identification of anticancer drug resistance mediated by MYB proteins and their underlying mechanisms are of great importance in understanding failures of current therapies and establishing new and more efficient therapy regimens. In addition, new drug candidates targeting MYB transcription factor activity and signaling have emerged as a promising class of potential anticancer therapeutics that could tackle MYB-dependent drug-resistant cancers in a more selective way. This review describes the correlation of MYB transcription factors with the formation and persistence of cancer resistance to various approved and investigational anticancer drugs.

Hub gene ELK3-mediated reprogramming lipid metabolism regulates phenotypic switching of pulmonary artery smooth muscle cells to develop pulmonary arterial hypertension induced by PM2.5

Fine particulate matter (PM2.5) as an environmental pollutant is related with respiratory and cardiovascular diseases. Pulmonary arterial hypertension (PAH) was characterized by incremental pulmonary artery pressure and pulmonary arterial remodeling, leading to right ventricular hypertrophy, and finally cardiac failure and death. The adverse effects on pulmonary artery and the molecular biological mechanism underlying PM2.5-caused PAH has not been elaborated clearly. In the current study, the ambient PM2.5 exposure mice model along with HPASMCs models were established. Based on bioinformatic methods and machine learning algorithms, the hub genes in PAH were screened and then adverse effects on pulmonary artery and potential mechanism was studied. Our results showed that chronic PM2.5 exposure contributed to increased pulmonary artery pressure, pulmonary arterial remodeling and right ventricular hypertrophy in mice. In vitro, PM2.5 induced phenotypic switching in HPASMCs, which served as the early stage of PAH. In mechanism, we investigated that PM2.5-mediated mitochondrial dysfunction could induce phenotypic switching in HPASMCs, which was possibly through reprogramming lipid metabolism. Next, we used machine learning algorithm to identify ELK3 as potential hub gene for mitochondrial fission. Besides, the effect of DNA methylation on ELK3 was further detected in HPASMCs after PM2.5 exposure. The results provided novel directions for protection of pulmonary vasculature injury, against adverse environmental stimuli. This work also provided a new idea for the prevention of PAH, as well as provided experimental evidence for the targeted therapy of PAH.

Discovery of NLRP3 inhibitors using machine learning: Identification of a hit compound to treat NLRP3 activation-driven diseases

NLRP3 is vital in developing many human diseases as one of the most critical inflammasomes. Developing related inhibitors has been instrumental in advancing the development of therapies for associated diseases. To date, there are no NLRP3 inhibitors on the market. This study identified a series of NLRP3 inhibitors using the self-developed machine learning model. Among them, CSC-6 was validated as the hit molecule with optimal activity and significantly inhibited IL-1β secreted by PMA-THP-1 cells (IC50 = 2.3 ± 0.38 μM). The results show that CSC-6 specifically binds NLRP3 and inhibits NLRP3 activation by blocking ASC oligomerization during NLRP3 assembly. In vivo experiments have demonstrated that CSC-6 effectively reduces the symptoms of NLRP3 overactivation-mediated sepsis and Gout in mouse models. Importantly, CSC-6 has lower cytotoxicity and exhibits better stability in human-derived liver microsomes, which is more favorable for the drug to maintain its efficacy in vivo for longer. The discovery of CSC-6 may contribute to the design and discovery of related NLRP3 inhibitors.