CD4 + effector lymphocytes (Teff) are traditionally classified by the cytokines they produce. To determine the states that Teff actually adopt in frontline tissues in vivo , we applied single-cell transcriptome and chromatin analysis on colonic Teff cells, in germ-free or conventional mice, or after challenge with a range of phenotypically biasing microbes. Subsets were marked by expression of interferon-signature or myeloid-specific transcripts, but transcriptome or chromatin structure could not resolve discrete clusters fitting classic T H subsets. At baseline or at different times of infection, transcripts encoding cytokines or proteins commonly used as T H markers distributed in a polarized continuum, which was also functionally validated. Clones derived from single progenitors gave rise to both IFN-γ and IL17-producing cells. Most transcriptional variance was tied to the infecting agent, independent of the cytokines produced, and chromatin variance primarily reflected activity of AP1 and IRF transcription factor families, not the canonical subset master regulators T-bet, GATA3, RORγ.

A notable number of acute lymphoblastic leukemia (ALL) patients develop CD19-positive relapse within 1 year after receiving chimeric antigen receptor (CAR) T cell therapy. It remains unclear if the long-term response is associated with the characteristics of CAR T cells in infusion products, hindering the identification of biomarkers to predict therapeutic outcomes. Here, we present 101,326 single-cell transcriptomes and surface protein landscape from the infusion products of 12 ALL patients. We observed substantial heterogeneity in the antigen-specific activation states, among which a deficiency of T helper 2 function was associated with CD19-positive relapse compared with durable responders (remission, >54 months). Proteomic data revealed that the frequency of early memory T cells, rather than activation or coinhibitory signatures, could distinguish the relapse. These findings were corroborated by independent functional profiling of 49 patients, and an integrative model was developed to predict the response. Our data unveil the molecular mechanisms that may inform strategies to boost specific T cell function to maintain long-term remission.

Single-cell multi-omics technologies and methods characterize cell states and activities by simultaneously integrating various single-modality omics methods that profile the transcriptome, genome, epigenome, epitranscriptome, proteome, metabolome and other (emerging) omics. Collectively, these methods are revolutionizing molecular cell biology research. In this comprehensive Review, we discuss established multi-omics technologies as well as cutting-edge and state-of-the-art methods in the field. We discuss how multi-omics technologies have been adapted and improved over the past decade using a framework characterized by optimization of throughput and resolution, modality integration, uniqueness and accuracy, and we also discuss multi-omics limitations. We highlight the impact that single-cell multi-omics technologies have had in cell lineage tracing, tissue-specific and cell-specific atlas production, tumour immunology and cancer genetics, and in mapping of cellular spatial information in fundamental and translational research. Finally, we discuss bioinformatics tools that have been developed to link different omics modalities and elucidate functionality through the use of better mathematical modelling and computational methods.

No abstract

Writing in Nature communications, Seo and collaborators presented PICASSO as a method to achieve 15-color imaging of spatially overlapping proteins using no reference emission spectra in a single staining and imaging round. This accessible tool has the potential to be applied to diverse applications within the spatial biology field without neglecting accuracy. Mixing it upMultiplexed biomolecular imaging of heterogeneous tissues is essential for a variety of biological and biomedical research. In particular, high-plex immunofluorescence imaging such as CODEX 1 , CyCIF 2 , 4i 3 , immunoSABER 4 , and NanoString CosMx 5 can detect tens of protein markers to identify heterogeneous cell types and cell-cell interactions at cellular or even subcellular resolution, enabling the study of architectural and spatial relationships of tissue in situ. Central to most of these approaches is the cyclic imaging which is time consuming and may result in signal loss due to photobleaching or possible loss of tissue morphology during the washing and/or quenching steps. Although it has been demonstrated to achieve much higher plex protein imaging using metal isotope labeling for mass spectrometry 6,7 or organic dyes for Raman microscopy 8 , these techniques require complex antibody tagging and different imaging modalities that are not readily accessible.When spectrally overlapping fluorophores are used to label biomolecules such as proteins for multicolor fluorescence imaging, it may become difficult to distinguish real signals from false positive signals due to bleed-through. Thus, techniques to accurately and efficiently tell apart the signal from each fluorophore become indispensable. Researchers had developed approaches to mitigate signal noise and overlap in multiplexed imaging studies. The signal in each channel is modeled as a linear combination of the contributing fluorophores 9 . By employing a mixing matrix, linear unmixing can produce an unmixed image. The major drawback with the linear unmixing approach is that a reference spectrum is needed. For example, in a multiplexed immunofluorescence experiment with two or three fluorophores that have overlapping emission spectra, to quantify the signal unique to each fluorophore, a reference, which could be a region of the tissue section of interest with 'pure' fluorophore can be used. This can be problematic in highly heterogeneous tissues such as the brain. Recently developed algorithms like LUMoS employ unsupervised machine learning approaches to identify the spectral signatures unique to each fluorophore in a method termed as blind unmixing precludes the need to have a reference spectrum 10 . However, challenges still remain with this approach making higher-level multiplexing difficult.