A molecular model of anti-cancer protein p53 binding to DNA. One of the applications of Sánchez-Rivera's sensor editing technique is to study mutations in p53 that prevent it from activating anti-tumour genes.Credit: Juan Gaertner/Science Photo Library
Cancer is fundamentally a genetic disease. Improving its treatment demands a deeper understanding of individual genetic mutations and how they influence disease progression. However, modelling cancer’s complex underpinnings is no easy feat. Francisco J. Sánchez-Rivera runs a lab at the Massachusetts Institute of Technology (MIT) dedicated to this challenge, using precision genome editing to generate single-nucleotide mutations and then studying their impact on cancer development. The lab’s precision approach is made possible by sophisticated ‘sensor sequences’ that monitor editing efficiency, including assessing whether the base editor has caused any undesired mutations. Sánchez-Rivera and colleagues collaborated with Agilent Technologies using the SurePrint High Fidelity Oligonucleotide Library Synthesis (HiFi OLS) platform to develop sensors reliable enough for this task. The cancer biologist explains how HiFi oligonucleotide synthesis, developed by Agilent for labs in a wide range of fields, is helping his group push the boundaries of cancer research.
Why study cancer at the level of specific mutations?
Every gene can be mutated in hundreds or even thousands of ways, and every mutation can have its own biological effect. We need to know how all these mutations function individually to understand how cancer develops and progresses. CRISPR-based technologies that inactivate or alter the expression of entire genes can identify therapeutic targets and biological processes in cancer. But they don’t model the effects of specific mutations that are seen in tumours, which can vary significantly among cancer patients.
What is your experimental approach?
We use CRISPR-based tools called base editors to generate precise single-nucleotide changes in genetic sequences. A CRISPR Cas9 nickase protein makes a single-strand DNA nick, and then a base-converting enzyme, physically attached to Cas9, makes a cytosine-to-thymine switch. We typically deliver base editors using viral vectors, and then study their effects in cell lines or lab mice. For instance, we might determine if a given mutation affects cancer cell proliferation, or whether it promotes metastasis or drug resistance.
How can you be sure a base editor introduces the correct mutation?
This can be a tricky problem: you don’t want your base editor to mutate other nearby sequences unintentionally. The way we get around that is by embedding what we call a ‘sensor sequence’ into the viral vector backbone. The sensor comprises a guide RNA coupled to a synthetic version of the endogenous target site. Both the synthetic and endogenous DNA sequences are mutated simultaneously. By sequencing the sensor, we can determine the precision with which a given base editor makes the endogenous change.
How does high-fidelity oligonucleotide synthesis enable this research?
During preliminary research, we found that nearly half of our sensors had errors that arose during synthesis. Because of the way our sensors are designed, as well as their significant length, we need to get the highest fidelity oligonucleotide synthesis possible. Fortunately, we were able to tap into Agilent’s SurePrint HiFi OLS platform. We were beta testers before it became commercially available, and that was really the only way we could get our sensor constructs synthesized. These types of sensors now represent a central pillar of our lab armamentarium.
What have you done to validate this approach?
We recently used it to investigate thousands of mutations identified in human tumours, gathered by researchers at the Memorial Sloan Kettering Cancer Center in New York. We showed that sensor editing accurately reported the edits at their corresponding endogenous sites.
Has your method led to discoveries in cancer biology?
After receiving this validation, we developed panels of guide RNAs for generating targeted mutations in the p53 tumour-suppressor gene. More than half of all cancers have mutations in p53 that prevent it from activating anti-tumour genes. We identified five mutations that had not been studied before, and introduced them into pancreatic epithelial cells harbouring wild-type copies of the p53 gene. When we implanted these edited cells into mice, all the animals developed pancreatic tumours. Based on our experiments, we also now know that different mutations produce tumours with varying levels of aggression in the body. These studies demonstrate that not all mutations are the same, and we need to understand how each of them drives different biology in patients.
How could your research improve the understanding and treatment of cancer?
If we can functionally stratify patients by the mutations in their tumours, then we can better understand and study how individual mutations alter the response to therapy — including treatments that might target mutant p53, for instance. We’re currently developing more precise cell-based and mouse models that closely mimic cancer biology in patients.
What is next for your lab?
Our research shows that it is possible to interrogate thousands of genetic variants at once using SurePrint HiFi oligonucleotide libraries. But with base editing, we can only engineer transition mutations, such as thymine substituting for cytosine. We therefore miss roughly half the mutations seen in cancer patients. We have begun to incorporate newer ‘prime editing’ technology that allows us to systematically engineer all possible mutations in a given gene, and the results so far have been spectacular. These approaches will help us define all the mechanisms that govern complex genetic interactions, which will impact all areas of biology and medicine.