Deep sequencing blood transcriptomes

In this issue of Blood, five satellite manuscripts from the Functional Annotation of the Mammalian Genome (FANTOM5) consortium show the power of deep transcriptome sequencing to identify genes, promoters, and enhancers specific to different hematopoietic subpopulations.^1-5  The papers by Motakis et al¹  and Rönnerblad et al³  and 2 papers by Schmidl et al^4,5  examine the transcriptomes of mast cells, 4 cell populations in granulopoiesis, naïve and memory/regulatory and conventional T-cell populations, and CD14/CD16 monocyte subpopulations, respectively, whereas the paper by Prasad et al²  identifies key cell type specific epigenetic regulators of hematopoiesis.

The FANTOM5 consortium consisting of >250 researchers from 20 countries and coordinated at RIKEN Yokohama, Japan, has used cap analysis of gene expression (CAGE) to map transcription start sites and generate a comprehensive expression atlas across a collection of 975 human and 399 mouse samples.⁶  Of particular interest to Blood readers, primary cells covering the majority of hematopoietic lineages have been profiled. Additionally, 50 different blood cancer-derived cell lines were included, providing a rich data source for comparison between normal and disease states. The atlas provides expression profiles for known and novel, and coding and noncoding transcripts alike, and as revealed below, even identifies active enhancer elements.

A particular highlight of the Blood papers is the analyses on enhancers. One outcome of the FANTOM5 project was the realization that active enhancers could be mapped based on the observation that they generate CAGE tags in both directions.⁷  With such a map of active enhancers, 3 of the Blood papers, Rönnerblad et al³  and the 2 papers by Schmidl et al,^4,5  are able to examine for the first time lineage-specific differences at these active enhancers.

Rönnerblad et al³  combine DNA methylome profiles with CAGE and microarray expression profiles to examine changes during granulopoiesis. The authors conclude that the differentially methylated sites are preferentially located in areas distal to CpG islands and shores and enriched at enhancer elements that become active during differentiation.

The first paper by Schmidl et al⁵  examines differences between human monocyte subsets: classical (CD14⁺⁺CD16⁻), intermediate (CD14⁺CD16⁺), and nonclassical (CD14dimCD16⁺). Using a combination of CAGE and genome-wide chromatin immunoprecipitation data (ChIP-seq) for histone modifications associated with enhancers, the authors examine the landscape of transcripts (coding, and noncoding) and transcribed enhancers used in each monocyte population. Nonclassical monocytes appear to have higher oxidative and mitochondrial activities and interestingly express higher levels of the interferon induced transmembrane proteins (IFITM1, IFITM2, and IFITM3), all linked to viral resistance.⁸ 

Similarly in the second paper by Schmidl et al,⁴  naïve and memory conventional and regulatory T cells are profiled by CAGE and ChIP-seq. The enhancer-specific marks were also tested here, but this time the authors examine the genome-wide binding of the transcription factors forkhead box P3 (FOXP3), runt-related transcription factor 1 (RUNX1), v-ets avian erythroblastosis virus E26 oncogene homolog 1 (ETS1), and signal transducer and activator of transcription 5B (STAT5). They find that STAT5 was enriched at enhancers active in regulatory T cells, whereas RUNX1 and ETS1 were enriched at conventional T-cell enhancers. Surprisingly FOXP3, the key determinant of regulatory T cells, was at enhancers of both regulatory and conventional T cells.

By profiling freshly isolated skin mast cells and expanded and stimulated mast cells, Motakis et al challenge 2 commonly held theories.¹  The data suggest that native mast cells and basophils are more distantly related than previously thought and that they express significantly fewer typical immune-related genes than previous reports suggest. The data also find a diverse collection of proteins previously not associated with mast cell function, including functional bone morphogenetic protein receptors. The authors demonstrate bone morphogenetic proteins promote survival of mast cells.

Last, Prasad et al²  overviews the expression of 199 epigenetic regulators in hematopoiesis. The work is the first of its kind and identifies epigenetic regulators specific to different blood lineages and with enriched expression in leukemia lines. Of note, the authors identify factors previously implicated in leukemia (lysine [K]-specific methyltransferase 2A [KMT2A], chromobox homolog 8 [CBX8], and enhancer of polycomb homolog 1 [EPC1]).

Together, these 5 papers highlight the power of deep transcriptome profiling in isolated primary cell populations in the context of a large comparative atlas. Further studies on rarer populations will follow in the months to come. On a final note, genome-wide analyses of cell type-specific transcripts revealed that a large fraction of the mammalian genome is specifically expressed in the hematopoietic lineages, the majority of which appear to be previously unrecognized transcripts including novel long noncoding and enhancer RNAs.⁶  Additionally, an evolutionary study of protein domain architectures points to early evolution of myeloid and later evolution of lymphoid (in particular, T) cells.⁹  There is much more to find in the FANTOM5 data.