Regulatory T cells (Tregs) play a central role in counteracting inflammation and autoimmunity. A more complete understanding of cellular heterogeneity and the potential for lineage plasticity in human Treg subsets may identify markers of disease pathogenesis and facilitate the development of optimized cellular therapeutics. To better elucidate human Treg subsets, we conducted direct transcriptional profiling of CD4⁺FOXP3⁺Helios⁺ thymic-derived Tregs and CD4⁺FOXP3⁺Helios⁻ T cells, followed by comparison with CD4⁺FOXP3⁻Helios⁻ T conventional cells. These analyses revealed that the coinhibitory receptor T cell Ig and ITIM domain (TIGIT) was highly expressed on thymic-derived Tregs. TIGIT and the costimulatory factor CD226 bind the common ligand CD155. Thus, we analyzed the cellular distribution and suppressive activity of isolated subsets of CD4⁺CD25⁺CD127^lo/− T cells expressing CD226 and/or TIGIT. We observed TIGIT is highly expressed and upregulated on Tregs after activation and in vitro expansion, and is associated with lineage stability and suppressive capacity. Conversely, the CD226⁺TIGIT⁻ population was associated with reduced Treg purity and suppressive capacity after expansion, along with a marked increase in IL-10 and effector cytokine production. These studies provide additional markers to delineate functionally distinct Treg subsets that may help direct cellular therapies and provide important phenotypic markers for assessing the role of Tregs in health and disease.

Statistical analysis and modeling of mass spectrometry (MS) data have a long and rich history with several modern MS-based applications using statistical and chemometric methods. Recently, machine learning (ML) has experienced a renaissance due to advents in computational hardware and the development of new algorithms for artificial neural networks (ANN) and deep learning architectures. Moreover, recent successes of new ANN and deep learning architectures in several areas of science, engineering, and society have further strengthened the ML field. Importantly, modern ML methods and architectures have enabled new approaches for tasks related to MS that are now widely adopted in several popular MS-based subdisciplines, such as mass spectrometry imaging and proteomics. Herein, we aim to provide an introductory summary of the practical aspects of ML methodology relevant to MS. Additionally, we seek to provide an up-to-date review of the most recent developments in ML integration with MS-based techniques while also providing critical insights into the future direction of the field.

Senescence is a cellular state linked to ageing and age-onset disease across many mammalian species^1,2. Acutely, senescent cells promote wound healing^3,4 and prevent tumour formation⁵; but they are also pro-inflammatory, thus chronically exacerbate tissue decline. Whereas senescent cells are active targets for anti-ageing therapy^6–11, why these cells form in vivo, how they affect tissue ageing and the effect of their elimination remain unclear^12,13. Here we identify naturally occurring senescent glia in ageing Drosophila brains and decipher their origin and influence. Using Activator protein 1 (AP1) activity to screen for senescence^14,15, we determine that senescent glia can appear in response to neuronal mitochondrial dysfunction. In turn, senescent glia promote lipid accumulation in non-senescent glia; similar effects are seen in senescent human fibroblasts in culture. Targeting AP1 activity in senescent glia mitigates senescence biomarkers, extends fly lifespan and health span, and prevents lipid accumulation. However, these benefits come at the cost of increased oxidative damage in the brain, and neuronal mitochondrial function remains poor. Altogether, our results map the trajectory of naturally occurring senescent glia in vivo and indicate that these cells link key ageing phenomena: mitochondrial dysfunction and lipid accumulation.