Date: 26th January 2021
A better understanding of the genetic evolution of metastatic disease has the potential to reveal the therapeutic vulnerabilities of tumours. As every step of the process leaves a clear phylogenetic signature, it can reveal critical insights such as the relationship between primary and metastasised tumours, the timing, frequency and destination of the metastatic seeding. As metastasis is the major cause of cancer-related mortality, the ability to translate this knowledge into halting its progression would be highly valuable. However, up until now we haven’t had the technology to tracks these rare, transient, and stochastic events. Now, scientists have developed a Cas9-based, single-cell lineage tracer to study the rates, routes, and drivers of metastasis in a lung cancer xenograft mouse model.
Classical retrospective lineage tracing strategies infer tumour ancestry from the pattern of shared sequence variations across tumour subpopulation. However, the resolution of these approaches is somewhat limited. In contrast, prospective lineage tracing approaches, where cells are statically labelled with a tag or DNA barcode and tracked, can measure gross population dynamics at a clonal resolution, but lack fine detail such as evolution and the rate, order and directionality of the metastatic event.
However, recent advances in technologies, such as combing Cas9-enabled lineage tracing techniques with single-cell RNA-sequencing readouts could potentially be used to track cancer progression on a larger scale with finer resolution than has been previously possible.
Now, a team of researchers led by Jonathan Weissman, Trever Bivona and Nir Yosef, from the University of California, Princeton University, Whitehead Institute, Massachusetts Institute of Technology and Harvard University, US, have used Cas9-induced indels at target sites to produce heritable markers of lineage. Then using computational approaches to reconstruct a phylogenetic tree, they have explored the subclonal dynamics of metastatic dissemination in a mouse cancer model. Revealing stark heterogeneity in metastatic capacity, arising from pre-existing and heritable differences in gene expression, identifying genes that can drive invasiveness.
The team started by engineering an aggressive lung adenocarcinoma cell line so that it contained a reporter gene for live imaging (luciferase), Cas9 for generating heritable edits, ~10 uniquely barcoded copies of the target site (to record lineage information captured by single-cell RNA-sequencing), and a triple-single guide RNAs (sgRNA) which would direct the Cas9 to the target sites.
They then implanted ~5,000 engineered cells, embedded in matrigel, into the left lung of immunodeficient mice. The team followed the fluorescent cells by live imaging, seeing tumour growth and metastasis. After 54 days, tumours were found in the five lung lobes, throughout the mediastinal lymph tissue, and on the liver, samples were taken from 6 areas, and processed for single-cell RNA-sequencing and phylogenetic reconstruction was performed.
Analysis showed that the initial 5,000 cells contained 2,150 distinguishable clones, from which only around 100 clonal populations engrafted and survived. The clonal populations exhibited distinct distributions across the six tissues, for example some clones remained in the primary tissue, whilst others could be detected in all tissues, whilst others were over represented in a tissue. This suggested to the team that there was broad metastatic heterogeneity across the tumour populations.
Whilst, this information gave an overview of metastasis, the team wanted to analyse the data in higher resolution and not on a clonal level. So they reconstructed high-resolution phylogenetic trees, which would describe the phylogenetic relationships between all cells within the clonal population and summarised their history of metastatic dissemination between tissues. This revealed stark heterogeneity in metastatic capacity, both between and within clonal populations.
When the team layered this information onto the transcriptional profiles of the cells, the data suggested that the different metastatic phenotypes manifested in characteristic differences in gene expression. They then investigated these potential drivers of differential metastatic phenotypes, finding many previously identified metastasis-associated factors, and also identifying many interesting and reproducible gene positive and negative candidates.
To determine whether these candidates were indeed drivers and not merely associated with metastatic behaviour, the team used CRISPR-inhibition (CRISPRi) or CRISPR–activation (CRISPRa) to investigate this further in the aggressive lung cancer cell line. They chose 5 gene targets, and found that CRISPRi knock-down resulted in decreased invasiveness for positive metastasis-associated genes and increased invasiveness for negative metastasis-associated genes. Conversely, CRISPRa caused the opposite results for all the candidates.
One unexpected discovery was that the gene Keratin 17 (KRT17) had a suppressive role in metastasis, it was strongly expressed in low metastatic tumours when compared to highly metastatic tumours. The CRISPR studies revealed this gene was anti-correlated and inhibited tumour invasiveness, as previous work implicated KRT17 in promoting invasiness, this role was unanticipated.
Conclusions and future applications
The team here have applied next generation Cas9-based lineage tracing to determine relevant and significant features of metastatic biology, that could only be determined at subclonal lineage level. Among these key insights were the broad range of metastatic rates for different tumour populations, the pre-existence and stable heritability of these heterogeneous metastatic phenotypes, and the complex, multidirectional tissue routes by which cancer cells disseminated in this model.
The ability to reconstruct deeply resolved and accurate cell phylogenies will be a valuable tool in directing research efforts. For example here they discovered that one tissue, the mediastinal lymph tissue, appeared to be a hub from which the cancer cells could then seed other areas. Therapeutically this could provide crucial information of where to focus treatments, halting or even eradicating metastasis before it can happen. Another example is that those metastatic cells that return to the primary tumour site could contribute to resistance of treatment, again an area of therapeutic interest.
The next step for the team is to extend the technology beyond merely observing the cells, and to start to predict their behaviour. With advances in digital technology such as artificial intelligence and deep learning, this should be achievable and a highly informative task. Indeed, recent work using a deep learning algorithm enabled the automated detection of metastases at the level of single cells in whole mice. One consideration is to extend the model, to determine other metastatic seeding patterns in different types of cancer, and the likelihood that patient-specific patterns may impact the outcomes.
Essentially, constructing of catalogue of metastatic drivers will reveal novel targets for developing an innovative wave of therapeutics to treat cancer, and we are now starting to see novel tools emerging to fight metastases. We recently reported the creation a ‘stealth bomber’ virus that evaded the innate immune system and could suppress both primary and disseminated aggressive lung tumour growth, resulting in a significant prolonging of life. It is hoped that these types of nascent studies, including the one here, will reveal previously unmapped facets of cancer biology, drive new treatments, and offer hope to those patients with aggressive, higher mortality-associated cancers.
For more information please see the press release from Massachusetts Institute of Technology
Quinn, J. J., M. G. Jones, R. A. Okimoto, S. Nanjo, M. M. Chan, N. Yosef, T. G. Bivona and J. S. Weissman (2021). “Single-cell lineages reveal the rates, routes, and drivers of metastasis in cancer xenografts.” Science: eabc1944.
https://doi.org/10.1126/science.abc1944
