The ENL YEATS epigenetic reader domain critically links MLL-ENL to leukemic stem cell frequency in t(11;19) Leukemia.

MLL (KMT2a) translocations are found in ~10% of acute leukemia patients, giving rise to oncogenic MLL-fusion proteins. A common MLL translocation partner is ENL and associated with a poor prognosis in t(11;19) patients. ENL contains a highly conserved N-terminal YEATS domain that binds acetylated histones and interacts with the PAF1c, an epigenetic regulator protein complex essential for MLL-fusion leukemogenesis. Recently, wild-type ENL, and specifically the YEATS domain, was shown to be essential for leukemic cell growth. However, the inclusion and importance of the YEATS domain in MLL-ENL-mediated leukemogenesis remains unexplored. We found the YEATS domain is retained in 84.1% of MLL-ENL patients and crucial for MLL-ENL-mediated leukemogenesis in mouse models. Mechanistically, deletion of the YEATS domain impaired MLL-ENL fusion protein binding and decreased expression of pro-leukemic genes like Eya1 and Meis1. Point mutations that disrupt YEATS domain binding to acetylated histones decreased stem cell frequency and increased MLL-ENL-mediated leukemia latency. Therapeutically, YEATS containing MLL-ENL leukemic cells display increased sensitivity to the YEATS inhibitor SGC-iMLLT compared to control AML cells. Our results demonstrate that the YEATS domain is important for MLL-ENL fusion protein-mediated leukemogenesis and exposes an "Achilles heel" that may be therapeutically targeted for treating t(11;19) patients.

Identification of Mycobacterium tuberculosis Antigens with Vaccine Potential Using a Machine Learning-Based Reverse Vaccinology Approach.

Tuberculosis (TB) is the leading cause of death of any single infectious agent, having led to 1.4 million deaths in 2019 alone. Moreover, an estimated one-quarter of the global population is latently infected with Mycobacterium tuberculosis (MTB), presenting a huge pool of potential future disease. Nonetheless, the only currently licensed TB vaccine fails to prevent the activation of latent TB infections (LTBI). These facts together illustrate the desperate need for a more effective TB vaccine strategy that can prevent both primary infection and the activation of LTBI. In this study, we employed a machine learning-based reverse vaccinology approach to predict the likelihood that each protein within the proteome of MTB laboratory reference strain H37Rv would be a protective antigen (PAg). The proteins predicted most likely to be a PAg were assessed for their belonging to a protein family of previously established PAgs, the relevance of their biological processes to MTB virulence and latency, and finally the immunogenic potential that they may provide in terms of the number of promiscuous epitopes within each. This study led to the identification of 16 proteins with the greatest vaccine potential for further in vitro and in vivo studies. It also demonstrates the value of computational methods in vaccine development.