RNA-dependent Protein Complexes

The function of RNAs - both coding mRNAs as well as non-coding RNAs (ncRNAs) - is intricately linked to their ability to interact with DNA, RNA or proteins. Already in the early days of lncRNA research, we hance developed a string interest in Ribonucleoprotein (RNP) complexes and RNA-binding proteins in cancer.

We uncovered the important oncogenic function of IGF2BP1 in liver cancer (Hepatology 2014) as well as interaction partners of the mRNA of the immune activating ligand MICB (Nat Commun 2014).
We also established state-of-the-art techniques in our lab for the identification of RNA-protein complexes in vitro and in vivo (Nature 2015).

Currently, we are most excited about a new concept developed in our lab: RNA dependence (Mol Cell 2019, Nat Protoc 2020). We define a protein as RNA-dependent if its interactome depends on RNA. This functional definition has a lot of important implications overcoming the questions of specificity and quantitation of the common definition of RNA-binding proteins. We have recently performed a successful proteome-wide screen defining RNA-dependent proteins - and make it available to the scientific community in our R-DeeP database.


Long non-coding RNAs in Cancer

70% of the human genome are transcribed - but only 2% encode for proteins!

One major field of our research aims at the elucidation of the molecular and cellular function as well as expression profiles and regulatory mechanisms of long non-coding RNAs (lncRNA) in cancer. Recent studies suggest that a large part (up to 70%) of the human genome is transcribed, while only 2% of the human genome encodes for proteins. However, for most of the ncRNA transcripts, their functions remain to be discovered. Interestingly, many of these ncRNAs are differentially expressed in cancer compared to normal tissue and hence could serve as biomarkers or therapeutic targets in the future. Additionally, many ncRNAs are highly conserved during evolution which provides another indication of potential functional importance of this diverse class of molecules.

Initially, we carried out microarray-based profiling of 17000 ncRNAs in three major tumor entities: lung, liver and breast cancer. This analysis has identified hundreds of differentially expressed non-coding RNAs specifically expressed or silenced in human cancer. Additionally, we have performed extensive strand-specific transcriptome sequencing of rRNA-depleted RNA of lung adenocarcinoma, which provided valuable information about all classes of expressed RNAs.
Our research focuses on differentially expressed ncRNAs in lung adenocarcinoma, hepatocellular carcinoma and mammary carcinoma. For these, we determine the cellular functions by studying gain- and loss-of-function models. To reveal the molecular function of these novel long non-coding RNAs, we use RNA affinity purifications to elucidate the ncRNA-protein networks (Nature 2015).

As an important novel technique for studying loss-of function phenotypes in human tumors, we developed a method to knockout ncRNAs in human cancer cell lines using Zinc Finger Nucleases (Genome Res 2011). Today, we are using the CRISPR / Cas9 system for gain- and loss-of-function models of lncRNAs - however, we also realize the additional challenges and the required measures of caution when working with lncRNAs in complex loci (Nucleic Acids Res 2017).

The prime example of our research is the study of the lncRNA MALAT1: Sven Diederichs already discovered MALAT1 during his PhD thesis as a marker associated with metastasis development in lung cancer (Oncogene 2003). We then established a loss-of-function model for MALAT1 in human lung cancer cells (Genome Res 2011). This enabled us to demonstrate that MALAT1 was not only a marker, but an active and essential player in lung cancer metastasis. It could serve as a therapeutic target for metastasis prevention therapy. It functioned as an epigenetic regulator of a pro-migratory, pro-invasive gene signature (Cancer Res 2013). Surprisingly and despite the strong impact on lung cancer cell migration, knockout of MALAT1 in the mouse did not result in any observable phenoptype (RNA Biol 2012).

Additionally, we unraveled the regulation, function and mechanism of additional lncRNAs like HULC (Hepatology 2013), LIMT (EMBO Mol Med 2016), TP53TG1 (PNAS 2016), VELUCT (Nucleic Acids Res 2017), linc00152 (Sci Rep 2017), NOP14-AS1 (Nucleic Acids Res 2017), CASC9 (Hepatology 2018) and discovered lincNMR as a regulator of nucleotide metabolism in cancer (Nat Commun 2020).

To study lncRNA function using high-throughput technologies, we developed and screened our own custom-designed siRNA library targeting 638 lncRNAs upregulated in lung, liver and breast cancer (Nucleic Acids Res 2017, Sci Rep 2017, Hepatology 2018). To further extend the spectrum, we have designed and generated a library for CRISPR / Cas9-based screens targeting 2100 genes (mRNAs and lncRNAs) linked to lung adenocarcinoma by 42000 sgRNAs. The successful and comprehensive CRISPRi screens in a panel of lung adenocarcinoma cells has provided important novel insights into the active pathways, mechanisms, vulnerabilities and synthetic lethalities in this deadly disease.


The non-canonical Cancer Mutome

Nonstop Extension Mutations

In the past decade, deep sequencing approaches have revolutionized cancer genome sequencing has identified millions of somatic mutations that a tumor has acquired, but which are not present in healthy cells of same individual. However, the functional relevance of the vast majority of these mutations is unknown. We investigate the impact of mutations in cancer with an emphasis on discovering novel types of relevant mutations which we collectively refer to as "non-canonical mutations". Beyond missense and nonsense mutations altering the protein sequence, we are convinced that many other alterations can significantly effect tumorigenesis (EMBO Mol Med 2019).

Nonstop extension mutations convert a stop codon into a sense codon and thereby extending a protein at its C-terminus - and had never been studied in cancer before. We systematically compiled nonstop mutations in human cancer and provide a comprehensive database (NonStopDB). For the tumor suppressor gene SMAD4, nonstop mutations in pancreatic and colon cancer abrogate protein expression by inducing proteasomal degradation via a novel degron of ten amino acids in the added extension (Nat Cell Biol 2020).

Synonymous mutations alter the gene and its mRNA sequence but not the protein due to the degeneracy of the genetic code. Traditionally, they have thus been viewed as "silent" mutations without a functional relavance. In a pan-cancer analysis, we have cataloged 659194 synonymous mutations in cancer at SynMICdb. Importantly, we identified relevant functions for synonymous in the strong oncogene KRAS affecting its expression and mRNA secondary structure (Nat Commun 2019).

In summary, our research thus uncovered that also so far neglected types of non-canonical mutations can have fundamental impact on cancer genes. The pioneering characterization of novel types of mutations in cancer will hence be the focus of our future research - with one main aim to dramatically increase the throughput, with which we can systematically characterize mutations in human tumor cells as endogenous model systems - to go from cancer genomics to mutomics.

to top
powered by webEdition CMS