Towards the human cellular microRNAome

Matthew N. McCall, Min‐Sik Kim, Mohammed Adil, Arun H. Patil, Yin Lu, Chris J. Mitchell, Pamela Leal, Jinchong Xu, Manoj Kumar, Valina L. Dawson, Ted M. Dawson, Alexander S. Baras, Avi Z. Rosenberg, Dan E. Arking, Kathleen H. Burns, Akhilesh Pandey, Marc K. HalushkabioRxiv 2017 · DOI

Abstract

MicroRNAs (miRNAs) are short RNAs that serve as master regulators of gene expression and are essential components of normal development as well as modulators of disease. MicroRNAs generally act cell autonomously and thus their localization to specific cell types is needed to guide our understanding of microRNA activity. Current tissue-level data has caused considerable confusion and comprehensive cell-level data does not yet exist. Here we establish the landscape of human cell-specific microRNA expression. This project evaluated 8 billion small RNA-seq reads from 46 primary cell types, 42 cancer or immortalized cell lines, and 26 tissues. It identified both specific and ubiquitous patterns of expression that strongly correlate with adjacent super-enhancer activity. Analysis of unaligned RNA reads uncovered 207 unknown minor strand (passenger) microRNAs of known microRNA loci and 2,632 novel putative microRNA loci. Although cancer cell lines generally recapitulated the expression patterns of matched primary cells, their isomiR sequence families exhibited increased disorder suggesting Drosha and Dicer-dependent microRNA processing variability. Cell-specific patterns of microRNA expression were used to deconvolute variable cellular composition of adipose tissue samples highlighting one use of this cell-specific microRNA expression data. Characterization of cellular microRNA expression across a wide variety of cell types provides a new understanding of this critical regulatory RNA species.

Introduction: The Need for Cellular Resolution

MicroRNAs function as critical regulators of gene expression, acting as "dimmer switches" that modulate protein production. While their biological importance is well-established, our understanding of their spatial distribution has been limited by the use of whole-tissue samples. Tissues are complex mixtures of diverse cell types; consequently, tissue-level data often obscures whether a change in miRNA expression reflects a change in cellular state or a shift in the tissue's cellular composition (e.g., immune cell infiltration). This study addresses this gap by establishing a high-resolution "cellular microRNAome," providing a definitive map of which specific cells produce which miRNAs.

Methods and Datasets

The researchers performed a massive meta-analysis of 530 small RNA-seq runs, totaling over 8 billion reads. This dataset includes new sequencing for 39 primary cell types, augmented with high-quality data from the Sequence Read Archive (SRA).

Summary of the extensive RNA-seq datasets analyzed, categorizing primary cells, cancer/immortalized lines, and tissues.

To process this vast amount of data, the team utilized miRge, a tool optimized for identifying isomiRs (miRNA variants). Technical noise and batch effects across different platforms were mitigated using RUV (Remove Unwanted Variation) and the DEXUS algorithm. For novel discovery, miRDeep2 was employed to analyze 1.2 billion unmapped reads, identifying previously unknown miRNA loci.

Key Results and Insights

The analysis detected 2,319 known miRNAs and revealed highly specific expression patterns. Only six miRNAs were found to be truly ubiquitous across all 46 primary cell types, while 94 miRNAs were strictly specific to a single cell class.

The miRNA analysis pipeline (A), t-SNE clustering of cell types (B), cell-specific miRNA expression heatmap (C), and tissue deconvolution application (F).

Major Findings:

  • Canonical Sequence Discrepancies: For approximately 37% of known miRNAs, the "canonical" sequence listed in miRBase is not the most abundant version found in human cells. This finding has significant implications for the design of PCR primers and miRNA inhibitors.
  • Processing Disorder in Cancer: Cancer and immortalized cell lines showed significantly higher "disorder" (Shannon entropy) in miRNA processing compared to primary cells, suggesting that the Drosha and Dicer cleavage machinery becomes less precise in malignancy.
  • Super-Enhancer Correlation: miRNA expression levels were found to correlate strongly with nearby super-enhancers, providing a mechanistic explanation for the high abundance of specific miRNAs in certain cell types.
  • Novel miRNA Discovery: The analysis uncovered 207 unknown passenger strands and 2,632 novel putative miRNA loci. One such discovery, JHU_ID_23828, is a primate-specific miRNA highly abundant in endothelial cells.
  • Tissue Deconvolution: By leveraging cell-specific miRNA signatures, the researchers demonstrated that it is possible to mathematically "unmix" complex tissue samples, such as calculating the exact proportions of blood and vascular cells within adipose tissue.

Discussion and Limitations

This atlas provides a foundational resource for understanding miRNA regulation at the cellular level. However, some limitations persist. Much of the cell-specific data is derived from in vitro cultures, which may not fully capture the behavior of cells in their native in vivo environment. Furthermore, the thousands of newly identified miRNAs are currently classified as "putative" and require further functional validation. Future work will expand the atlas to include missing critical cell types, such as hepatocytes and neutrophils, further refining our map of the human microRNAome.

Summary updated

Abstract

Abstract microRNAs are short RNAs that serve as master regulators of gene expression and are essential components of normal development as well as modulators of disease. MicroRNAs generally act cell autonomously and thus their localization to specific cell types is needed to guide our understanding of microRNA activity. Current tissue-level data has caused considerable confusion and comprehensive cell-level data does not yet exist. Here we establish the landscape of human cell-specific microRNA expression. This project evaluated 8 billion small RNA-seq reads from 46 primary cell types, 42 cancer or immortalized cell lines, and 26 tissues. It identified both specific and ubiquitous patterns of expression that strongly correlate with adjacent super-enhancer activity. Analysis of unaligned RNA reads uncovered 207 unknown minor strand (passenger) microRNAs of known microRNA loci and 2,632 novel putative microRNA loci. Although cancer cell lines generally recapitulated the expression patterns of matched primary cells, their isomiR sequence families exhibited increased disorder suggesting Drosha and Dicer-dependent microRNA processing variability. Cell-specific patterns of microRNA expression were used to deconvolute variable cellular composition of adipose tissue samples highlighting one use of this cell-specific microRNA expression data. Characterization of cellular microRNA expression across a wide variety of cell types provides a new understanding of this critical regulatory RNA species.
BibTeX
@misc{mccall2017towards,
  author = {Matthew N. McCall and Min‐Sik Kim and Mohammed Adil and Arun H. Patil and Yin Lu and Chris J. Mitchell and Pamela Leal and Jinchong Xu and Manoj Kumar and Valina L. Dawson and Ted M. Dawson and Alexander S. Baras and Avi Z. Rosenberg and Dan E. Arking and Kathleen H. Burns and Akhilesh Pandey and Marc K. Halushka},
  title = {{Towards the human cellular microRNAome}},
  year = {2017},
  doi = {10.1101/120394},
  howpublished = {bioRxiv},
}
Figure