On the basis of their differential location within the dermis and of discrete changes in gene and protein expression, two major fibroblast subtypes (papillary and reticular) have traditionally been distinguished. In the last 3 years, a number of research groups have begun to address transcriptomic heterogeneity of human skin cells at the single-cell level by determining mRNA levels of expressed genes through single-cell RNA sequencing technologies. However, the outcome of single-cell RNA sequencing studies is thus far confusing. Very little overlap was found in fibroblast subpopulations, which also varied in number and composition in each dataset. After a careful reappraisal of the transcriptomic data of 13,823 human adult dermal fibroblasts that have been sequenced to date, we show that fibroblasts may robustly be assigned to three major types (axes A‒C), which in turn are composed of 10 major subtypes (clusters), which we denominated A1‒A4, B1 and B2, and C1‒C4. These computationally determined axes and clusters represent the major fibroblast types and subtypes in adult healthy human skin across different datasets, accounting for 92.5% of the sequenced fibroblasts. They thus may provide the basis to improve our understanding of dermal homeostasis and cellular function at the transcriptomic level.

In recent years, the macrophage colony-stimulating factor (M-CSF) and granulocyte-macrophage CSF (GM-CSF) cytokines have been identified as opposing regulators of the inflammatory program. However, the two cytokines are simultaneously present in the inflammatory milieu, and it is not clear how cells integrate these signals. In order to understand the regulatory networks associated with the GM/M-CSF signaling axis, we analyzed DNA methylation in human monocytes. Our results indicate that GM-CSF induces activation of the inflammatory program and extensive DNA methylation changes, while M-CSF-polarized cells are in a less differentiated state. This inflammatory program is mediated via JAK2 associated with the GM-CSF receptor and the downstream extracellular signal-regulated (ERK) signaling. However, PI3K signaling is associated with a negative regulatory loop of the inflammatory program and M-CSF autocrine signaling in GM-CSF-polarized monocytes. Our findings describe the regulatory networks associated with the GM/M-CSF signaling axis and how they contribute to the establishment of the inflammatory program associated with monocyte activation.

Cancer cells can develop a strong addiction to discrete molecular regulators, which control the aberrant gene expression programs that drive and maintain the cancer phenotype. Here, we report the identification of the RNA-binding protein HuR/ELAVL1 as a central oncogenic driver for malignant peripheral nerve sheath tumors (MPNSTs), which are highly aggressive sarcomas that originate from cells of the Schwann cell lineage. HuR was found to be highly elevated and bound to a multitude of cancer-associated transcripts in human MPNST samples. Accordingly, genetic and pharmacological inhibition of HuR had potent cytostatic and cytotoxic effects on tumor growth, and strongly suppressed metastatic capacity in vivo. Importantly, we linked the profound tumorigenic function of HuR to its ability to simultaneously regulate multiple essential oncogenic pathways in MPNST cells, including the Wnt/β-catenin, YAP/TAZ, RB/E2F, and BET pathways, which converge on key transcriptional networks. Given the exceptional dependency of MPNST cells on HuR for survival, proliferation, and dissemination, we propose that HuR represents a promising therapeutic target for MPNST treatment.

Aging is often associated with a loss of cell type identity that results in an increase in transcriptional noise in aged tissues. If this phenomenon reflects a fundamental property of aging remains an open question. Transcriptional changes at the cellular level are best detected by single-cell RNA sequencing (scRNAseq). However, the diverse computational methods used for the quantification of age-related loss of cellular identity have prevented reaching meaningful conclusions by direct comparison of existing scRNAseq datasets. To address these issues we created Decibel, a Python toolkit that implements side-to-side four commonly used methods for the quantification of age-related transcriptional noise in scRNAseq data. Additionally, we developed Scallop, a novel computational method for the quantification of membership of single cells to their assigned cell type cluster. Cells with a greater Scallop membership score are transcriptionally more stable. Application of these computational tools to seven aging datasets showed large variability between tissues and datasets, suggesting that increased transcriptional noise is not a universal hallmark of aging. To understand the source of apparent loss of cell type identity associated with aging, we analyzed cell type-specific changes in transcriptional noise and the changes in cell type composition of the mammalian lung. No robust pattern of cell type-specific transcriptional noise alteration was found across aging lung datasets. In contrast, age-associated changes in cell type composition of the lung were consistently found, particularly of immune cells. These results suggest that claims of increased transcriptional noise of aged tissues should be reformulated.

Feature selection is a relevant step in the analysis of single-cell RNA sequencing datasets. Most of the current feature selection methods are based on general univariate descriptors of the data such as the dispersion or the percentage of zeros. Despite the use of correction methods, the generality of these feature selection methods biases the genes selected towards highly expressed genes, instead of the genes defining the cell populations of the dataset. Triku is a feature selection method that favors genes defining the main cell populations. It does so by selecting genes expressed by groups of cells that are close in the k-nearest neighbor graph. The expression of these genes is higher than the expected expression if the k-cells were chosen at random. Triku efficiently recovers cell populations present in artificial and biological benchmarking datasets, based on adjusted Rand index, normalized mutual information, supervised classification, and silhouette coefficient measurements. Additionally, gene sets selected by triku are more likely to be related to relevant Gene Ontology terms and contain fewer ribosomal and mitochondrial genes. Triku is developed in Python 3 and is available at https://github.com/alexmascension/triku.

Skin is a complex and heterogeneous organ at the cellular level. This complexity is beginning to be understood through the application of single-cell genomics and computational tools. A large number of datasets that shed light on how the different human skin cell types interact in homeostasis—and what ceases to work in diverse dermatological diseases—have been generated and are publicly available. However, translation of these novel aspects to the clinic is lacking. This review aims to summarize the state-of-the-art of skin biology using single-cell technologies, with a special focus on skin pathologies and the translation of mechanistic findings to the clinic. The main implications of this review are to summarize the benefits and limitations of single-cell analysis and thus help translate the emerging insights from these novel techniques to the bedside.