As evidenced by numerous examples scattered across the various areas of biology, a cell phenotype is not solely determined by its genotype but is rather shaped by a multitude of dynamic non-genetic mechanisms. These include DNA and histones modifications, high-order chromatin architecture, gene expression dynamics and RNA-protein interactions, amongst others; all of them acting in concert to grant cells with the capability to dynamically adapt within an ever-changing environment. Despite the increasing recognition of the relevance of non-genetically encoded information for many biological processes, including cancer development, the underlying molecular details remain largely unknown.
Thus, our research focuses on the study of non-genetic information and its potential carriers, as we aim to unravel their genesis, dynamics, and inheritance as well as their role in responses to biological cues in models of epithelial-to-mesenchymal transition, oncogene-induced transformation and resistance to anticancer drugs. We believe that integrating these two crucial biological concepts – namely genetic and non-genetic information - and deciphering their interplay will drive our understanding of cancer evolution forward to ultimately translate our discoveries into more effective anticancer therapeutic paradigms.
Understanding of the role of non-genetic factors in cancer evolution
Accumulating evidence, including previous work from our group, has shown that even genetically identical cells (i.e. in vitro generated clonal cells), can be highly heterogeneous at the non-genetic level, displaying major differences in RNA and/or protein expression. Strikingly, this non-genetic heterogeneity could be linked to dramatically different cellular phenotypes, thus providing direct evidence for its biological relevance. Interestingly, it is widely accepted that extracellular cues from the environment can trigger dynamic intracellular changes at the non-genetic level which is often associated with phenotypic changes linked to adaptation to a given stimulus. Similarly, genomic mutations, which are often observed during cancer development, can be considered as intracellular effectors that result in changes in non-genetic compartments and thus modify phenotypic output (i.e. mutations might lead to the transformation of a normal cell into a malignant one). However, despite decades of research, the detailed molecular events that underlie malignant transformation are yet to be determined (Figure 1A).
Notably, and in spite of recent technical advances, the evolutionary path that tumours take are mainly inferred in a retrospective manner, mainly due to missing sampling from early stages of the disease and thus rendering the actual contribution of genetic and non-genetic components to cancer initiation unanswered. In order to fill this void in cancer evolution knowledge, our lab is currently developing a high-throughput platform with the aim to shed light on the molecular mechanisms of cancer onset promoted by the expression of hundreds of oncogenes and at single cell level (Figure 1B). Our experimental/computational pipeline enables us to combine information on the genetic level with non-genetic information encoded in the transcriptome, representing an important step to bridge the gap in knowledge on how these two frameworks interact and influence each other in the context of cancer evolution.
Importantly, our innovative approach can be applied to a variety of cell types allowing us to determine particular transforming oncogenes relevant for different cancer types in a quantitative manner. Indeed, by taking advantage of its quantitative nature we can rank transforming oncogenes – in a given cellular system of choice – based on their transforming potential, where the oncogenes displaying the most significant differences are further selected to be challenged in more complex models such as organoids or animal experimental settings. Moreover, the non-genetic information gained from the transcriptome data for each cell can be linked to individual oncogenes, thus allowing us to establish a direct comparison between non-genetic characteristics (defined as transcriptomic states) and different oncogenes or oncogene families. Notably, changes at the level of transcriptomic states present before and after transformation can be identified and informatically mined to reveal intracellular pathways that are being activated or downregulated upon the cellular response to oncogene expression and/or the transforming event. Moreover, the resulting dataset can be further explored to gain a better understanding of the molecular mechanisms driving cellular transformation as well as potential starting points for the development of novel therapy strategies.
Having built and set up all the components of our platform, during the past year we have successfully completed our first round of pilot experiments comprising a reduced-complexity oncogene library of 26 RAS variants, a family of oncogenes commonly mutated in various cancer types. Notably, the transcriptome data of individual cells could be successfully assigned to each respective RAS variant and clustering of the data into the identified transcriptomic states was highly reproducible. Most strikingly, drastic differences in the transforming potential between the different RAS variants could be detected. These encouraging initial results demonstrate the capability of our platform to provide crucial insights into the dynamic mechanisms promoting cancer onset. Therefore, we are currently expanding our platform to investigate hundreds of oncogenes simultaneously and now include further experimental improvements which will allow us to explore the step-wise evolution following malignant transformation in a time-resolved manner. Importantly, our approach minimises the number of animal experiments dramatically by using an in vitro approach to select for successfully transformed cells and by pre-selecting the most relevant oncogenes before validating potential findings in vivo. Due to its relevance from the 3Rs perspective (reduce, replace, refine animal experiments) our efforts have been recognised at the joint CRUK MI-AstraZeneca 3Rs' poster event last year where we were awarded first prize.
Although many levels of non-genetic traits co-existing within a cell have been described, very little is known about the molecular rules underlying their establishment, dynamics and inheritance. One of the main aims of our lab is to shed light on these underlying mechanisms in order to better understand the role of non-genetic information in the establishment of biologically relevant phenotypes and their plasticity (the ability of the cellular phenotype to dynamically change/adapt). By using our in-lab developed lineage tracing approach, we are able to determine the transcriptomic states of thousands of individual cells and their direct progeny and thus can follow the dynamic evolution of metastable states along their lineages in a time resolved manner (Figure 1C). Strikingly, we uncovered for both transformed and pre-malignant (immortalised) clonal cell populations that cellular plasticity in terms of transcriptomic states is not random but instead is governed by molecular effectors that seem to restrict the dynamic transition between states. Our current efforts are directed towards deciphering the underlying molecular rules constraining non-genetic plasticity of cell populations and identifying key molecular players in this process. Along these lines, we are currently pursuing promising leads that suggest that non-coding RNAs and IDR-containing proteins (Intrinsically Disordered Proteins) segregate into phase-separated compartments and participate in the genesis and dynamic behaviour of transcriptomic states.
Finally, we have implemented a multimodal single cell analysis toolkit that enables us to explore non-genetic elements such as histone modifications or transcription factor binding concomitantly with the measurement of transcriptomic states (Figure 1C). Ultimately, we believe that by fusing together information about the various non-genetic determinants with classical genetics (i.e. oncogenes) we will create a comprehensive picture that will better reflect the complex intricacies of cancer evolution and its dynamics.