The Mendell laboratory investigates fundamental aspects of post-transcriptional gene regulation, noncoding RNA regulation and function, and the roles of these pathways in normal physiology, cancer, and other diseases. In addition, we employ high-throughput screening approaches to interrogate diverse problems in RNA and cancer biology.
To briefly summarize some of our most important discoveries, we uncovered the first example of a vertebrate transcription factor that regulates miRNA expression (O’Donnell et al., Nature, 2005). This study, which demonstrated that the MYC oncogenic transcription factor directly transactivates the pro-tumorigenic miR-17-92 cluster, was important for establishing the principle that miRNAs have been functionally integrated into core cancer pathways. Subsequent work from my laboratory further defined the roles of miRNAs in several critical oncogenic and tumor suppressor pathways. Our laboratory has been at the forefront of elucidating miRNA functions in vivo (e.g. Chivukula et al., Cell, 2014) and translating these findings into novel therapeutic approaches, most notably through our demonstration that systemic delivery of miRNAs potently suppresses tumorigenesis in mouse cancer models without toxicity (e.g. Kota et al, Cell, 2009). We have also advanced our understanding of miRNA regulation, identifying examples of regulated miRNA biogenesis, decay, and target engagement (e.g. Hwang et al., Science, 2007; Golden et al., Nature, 2017). Our laboratory has also discovered important long noncoding RNAs (lncRNAs) and dissected their functions in cell and animal models (e.g. Lee et al., Cell, 2016). Most recently, we have employed high-throughput approaches to interrogate RNA biology and post-transcriptional regulation (Golden et al., Nature, 2017), a strategy that we are now applying to diverse problems in the laboratory.
Examples of ongoing areas of research in the laboratory are highlighted below.
miRNAs are an abundant and diverse family of ~20- to 23-nucleotide RNAs that recognize sites of complementarity in target mRNAs, resulting in decay and reduced translation of target transcripts. Approximately 500-1000 miRNAs are encoded in mammalian genomes. Our work on the miRNA pathway focuses mainly on three broad questions: What functions do miRNAs perform in normal physiologic states? How does aberrant miRNA activity contribute to diseases such as cancer? How is miRNA abundance and activity regulated in normal physiology and disease?
In addressing these questions, we emphasize the use of in vivo models to robustly dissect miRNA function. The importance of studying miRNAs in vivo is exemplified by our analysis of the miR-143/145 cluster (Chivukula et al., Cell, 2014), one of the most widely studied anti-tumorigenic miRNA loci. Downregulation of miR-143/145 in colorectal tumors was one of the first reported miRNA abnormalities in human cancer and over 200 subsequent papers implicated these miRNAs as epithelial-intrinsic tumor suppressors. Nevertheless, the natural functions of these miRNAs in epithelial tissues had not been studied. We generated miR-143/145 conditional knockout mice to address this question. Our studies of these animals unexpectedly revealed that this miRNA cluster is exclusively expressed in the mesenchymal compartment of intestine, where it performs an essential function in promoting epithelial repair following injury (Figure 1). These findings provided important new insights into mechanisms of injury repair and clarified the role of this miRNA cluster in cancer. Our laboratory has studied other important miRNAs using in vivo models, including miR-122 (Hsu et al., JCI, 2012) and the miR-26 family (Zeitels et al., Genes Dev, 2014) (Figure 2). CRISPR-mediated genome editing has dramatically enhanced our ability to generate gain- and loss-of-function models to study miRNA biology, representing an important aspect of our ongoing research program.
In addition to miRNAs, it is now appreciated that many other types of noncoding RNAs have important roles in development and disease. Analogous to our earlier studies of the miRNA pathway, we are currently exploring the biology of long noncoding RNAs (lncRNAs) and their diverse functions. For instance, we recently discovered a remarkable lncRNA that we termed Noncoding RNA Activated by DNA Damage or NORAD (Lee et al., Cell, 2016). NORAD is highly conserved and abundant in mammals, with expression levels of approximately 500-1,000 copies per cell. Surprisingly, inactivation of NORAD triggers dramatic aneuploidy in previously karyotypically-stable cell lines. NORAD maintains genomic stability by sequestering PUMILIO proteins, which repress the stability and translation of messenger RNAs to which they bind. In the absence of NORAD, PUMILIO proteins drive chromosomal instability by hyperactively repressing mitotic, DNA repair, and DNA replication factors (Figure 3). These findings introduced a new mechanism that regulates the activity of a deeply conserved and highly dosage-sensitive family of RNA binding proteins and revealed unanticipated roles for a lncRNA and PUMILIO proteins in the maintenance of genomic stability. We are continuing to study the function of NORAD and other lncRNAs using cellular and animal models as well as performing screens to identify new lncRNAs with important roles in normal physiology and disease.
CRISPR-mediated genome editing is revolutionizing all aspects of biological inquiry. One of the most exciting applications of this technology is its use in high-throughput screens to interrogate phenotypes and pathways in an unbiased manner. Our laboratory has developed the complete experimental and computational infrastructure to employ this methodology to dissect broad questions in RNA biology and, more generally, cancer biology. We are currently applying this strategy to perform genome-wide loss- and gain-of-function screens to investigate noncoding RNA mechanisms and functions as well as to dissect diverse post-transcriptional regulatory pathways.
As an example of how we utilize genome-wide CRISPR screens to dissect regulatory pathways of interest, we recently applied this method to study the miRNA pathway (Golden et al., Nature, 2017). By generating a sensitive GFP reporter that is tightly regulated by a miRNA, we were able to perform a comprehensive genome-wide screen to identify new factors that are essential for miRNA-mediated repression (Figure 4). Unexpectedly, this revealed a critical role for phospho-regulation of Argonaute (AGO) proteins, the key effectors in the miRNA pathway. Our studies demonstrated that when AGO engages target mRNAs, it becomes a substrate for the kinase CSNK1A. Phosphorylation on a set of highly conserved residues triggers target release, whereupon the ANKRD52-PPP6C phosphatase complex rapidly dephosphorylates AGO. We believe that this phosphorylation cycle allows AGO to efficiently move from target to target, enabling miRNAs to efficiently navigate the target landscape in order to productively silence a vast excess of target mRNAs. Accordingly, interrupting this phosphorylation cycle by mutating the kinase or phosphatase, or by mutating the AGO phosphorylation sites, significantly impairs miRNA-mediated silencing. We are intrigued by the possibility that this phosphorylation cycle may also be used to control the activity of the miRNA pathway in physiologic settings or in disease states, a possibility that we are currently investigating.