Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

The cellular-resolution revolution

Few technologies have changed the language and approach of biological research as dramatically and pervasively as single-cell technology, which has joined lipidomics, GFP, ChIP–seq and CRISPR–Cas in the pantheon of biotechnology. Here, we reflect on the influence of single-cell technology on metabolism research, some of which can be found in our new Collection on Single-cell technology in metabolism, featuring articles published in Nature Metabolism.

After the first-published single-cell next-generation-sequenced transcriptomic profile of a single mouse blastomere in 2009, the past 12 years have seen explosive growth in single-cell sequencing technologies. We must first, of course, acknowledge the handful of dedicated research groups that fostered and refined single-cell genomics and single-cell RNA sequencing (scRNA-seq), sorting organs into thousands of 96- and 364-well plates, and braving uncharted territory in R to normalize, analyse and standardize data workflows.

In the early 2010s, the ‘kit-ification’ of scRNA-seq workflow and analysis for the average bench scientist allowed for the widespread application of the technology beyond specialized research groups. In the time since, riffs and variations on scRNA-seq and DNA-seq have abounded, and Drop-seq, scATAC–seq and scDNA methylome profiling have expanded the toolkit to examine heterologous biological processes in cellular populations.

Nature Metabolism has featured several studies that provide excellent examples of how single-cell technologies are being leveraged to improve the understanding and appreciation of metabolism. Whereas bulk RNA-seq of tissues or cell populations can provide an overview of biological activity, the single-cell resolution of scRNA-seq allows for the identification of rare cell types with unique transcriptional profiles. Most recently, Angueira et al. and Shamsi et al., examining human perivascular adipose tissue and mouse interscapular brown adipose tissue, respectively, converged on a previously unidentified thermogenic adipocyte precursor population that expresses the temperature-sensitive cation channel Trpv1.

In addition, scRNA-seq allows for a new granularity of population composition and dynamics, thus facilitating the study of complex tissues and stages of pathogenesis. Newman et al. married bulk sequencing and scRNA-seq of vascular cells, atherosclerotic lesions and lineage-traced cells, and found that smooth muscle cells contribute to atherosclerotic lesion caps and undergo metabolic reprogramming during atherosclerotic-cap development.

Some cell types, such as adipocytes and neurons, are unamenable to sorting or library preparation, but Karunakaran et al. and Ludwig et al. have both successfully used preparations of nuclei from mouse and human tissues, respectively, to circumvent these technical hurdles. Nuclei largely maintain fidelity with transcripts in the cytoplasm; however, when the cytoplasm is removed, isolated nuclei have fewer transcripts available for sequencing on a per-cell basis than intact cells, and this difference must be accounted for when considering sequencing coverage. Therefore, using isolated nuclei as starting material for scRNA-seq enables a more unbiased sampling of cells within tissues that may have dramatically different shapes, sizes or compositions.

With the flurry of publications using scRNA-seq, databases for sequencing data, such as Gene Expression Omnibus, or initiatives such as Tabula Muris, are providing rich resources to advance analysis of existing data and allow for more complex biological queries. Levy et al. have used scRNA-seq profiles of young and old tissues from seven different studies in flies, mice and humans to develop a stochastic model of transcriptional dysregulation in ageing due to DNA mutational load.

Finally, the integration of temporal single-cell gene expression in a ‘homogeneous’ population of cells can now reveal previously unknown developmental stages, such as those identified in in vitro mouse osteoclastogenesis by Tsukasaki et al., as well as regulatory nodes for cell-fate trajectories, such as those identified in the differentiation of human embryonic stem cells into beta-like cells by Wang et al.

Where do we go from here? Currently, researchers attempting to include spatiotemporal data in scRNA-seq analysis, such as studies by Ben-Moshe et al., modelling liver zonation, and Droin et al., modelling circadian rhythms, are applying single-cell technology in three and four dimensions. Of course, a dream for any metabolism researcher would be single-cell metabolomics. Although mass spectrometry technology has not achieved the sensitivity, robustness and throughput of signal necessary for measuring metabolites in single cells as input, proxies have provided a taste of this next era of single-cell metabolic analysis. In this issue, Wu et al. use single-cell imaging of a fluorescence resonance energy transfer–based lactate biosensor as a proxy for glycolytic metabolism. MALDI mass spectrometry imaging, which is often used to extrapolate the mass spectra of metabolites from tissue samples, as used by Choi et al., can identify proteins, lipids and some small metabolites at single-cell or subcellular resolution. A recent study by Levine et al. has come tantalizingly close to single-cell metabolic phenotyping by using mass cytometry (or CyTOF) of 45 known metabolic enzymes simultaneously in populations of CD8+ T cell subsets. Although these methods, together and separately, are entirely sufficient to paint a detailed picture of metabolism at the single-cell level, the possibility of single-cell metabolite tracing or single-cell direct extracellular flux analysis feeds our desire for more.

The field of single-cell study is not yet mature, and debates and differing opinions about standards for intersample and intrasample normalization and batch correction have yet to be fully resolved. These conversations will be necessary, and we can expect more growing pains before we see standardization similar to that currently established for the analysis of bulk RNA-seq data. As a journal, Nature Metabolism aims to serve as a forum for the community to share ideas or propose standards, and we welcome a wider discussion of these important issues as stepping stones towards widespread standardization.

As editors, we see scientists doing the unimaginable in combining old and new technologies to answer some of the most complex metabolic and biological questions. As single-cell technologies continue to diversify over the coming decades, we surely have seen only a glimpse of what is to come.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

The cellular-resolution revolution. Nat Metab 3, 587 (2021). https://doi.org/10.1038/s42255-021-00401-y

Download citation

Search

Quick links