Uncovering cancer growth genes that have each other’s back when the other is down

From the Berger Lab, Human Biology Division

When graduate student Phoebe Parrish was deciding what type of PhD project to pursue, she not only considered the type of science she wanted to work on, but how she would pursue it. Parrish knew she wanted to work on a functional genomics project, one that she built from the ground up, following it from beginning to end. Importantly, she wanted to conduct her research in a highly collaborative manner while making her research publicly accessible, and ideally under the mentorship of a female principal investigator. As a rotation project in the lab of Dr. Alice Berger, part of the Human Biology Division, Parrish started a CRISPR-Cas9 based approach to identify synthetic lethal paralog genes - or gene pairs that are able to functionally compensate for each other when one is knocked out. Working with a female PhD mentor on a functional genomics project using a CRISPR-based approach (a gene editing tool discovered by two women – Dr. Jennifer Doudna and Dr. Emmanuelle Charpentier), provided Parrish a research opportunity that combined her scientific passions and aligned with her values of promoting women in science. She knew Dr. Berger’s lab was the perfect fit for her PhD career. This work that began as a rotation project was recently published in Cell Reports. Here, Parrish et al. identified synthetic lethal paralogs essential for lung cancer cell growth in addition to uncovering tumor suppressor paralog pairs.

This project was inspired by a peculiar observation in the Berger Lab where they found that in a genome-wide CRISPR screen in lung cancer cells, paralog genes were less likely to be essential for cell growth compared to non-paralog genes, despite paralogs consisting of two-thirds of protein coding genes in the human genome. Additionally, they found that individual knockout of paralogs MEK1 or MEK2 in lung cancer cells had no effect, but treatment with a dual inhibitor of MEK1 and MEK2 resulted in cell death. Given that many large-scale or genome-wide screens seek to identify cancer-essential genes that might make effective therapeutic targets, the Berger Lab realized the importance of uncovering these paralog blind spots. Furthermore, “synthetic lethal therapies for cancer are really attractive for cases where cancer cells have mutations in one paralogous gene”, but normal cells do not. In this case, specifically targeting the non-mutated gene paralog would affect cancer cells but prevent off-target effects in healthy cells, according to Parrish. Parrish goes on to explain that she “found out that this actually was not such a novel idea,” as other groups started publishing work identifying synthetic lethal paralogs in various contexts, “but being at the Hutch, with excellent resources and collaborators, we were able to get this work out quickly, while bringing a unique perspective to it.”

Schematic of CRISPR-Cas9 based method, pgPEN, used to identify synthetic lethal and tumor suppressor paralogs.
Schematic of CRISPR-Cas9 based method, pgPEN, used to identify synthetic lethal and tumor suppressor paralogs. Image taken from original article

Collaborative science was something Parrish strived for, as it “literally makes my science better” she exclaims. However, “so many collaborations are transactional. I wanted to take that aspect out of it, making it truly collaborative and fun,” Parrish remarks. To that end, she worked together with Dr. James Thomas, a postdoc in the Bradley Lab, part of the Basic and Public Health Sciences Divisions, to develop a CRISPR-Cas9 method to directly probe paralog compensation on a genomic scale. The research team sought to specifically identify synthetic lethal paralogs that could serve as potential lung cancer drug targets and focused on duplicated genes- paralog families consisting of only two genes. Parrish et al. then knocked out over 1,000 human paralog families (>2,000 genes), targeting each gene paralog individually or in combination with its paralog pair, generating paired guide RNAs for a paralog gene interaction mapping (pgPEN) library consisting of over 33,000 paired guide RNAs (pgRNAs). The researchers tested their pgPEN library in lung adenocarcinoma cells and identified paralogs that were essential for cell growth and viability by using next-generation sequencing to assess the abundance of pgRNAs after a 21-day competitive growth screen. Here, pgRNAs in low abundance would be indicative of genes promoting cancer cell growth. Among these top hits were important cell cycle regulators and MEK1/MEK2, confirming that their previous observed discrepancy between genetic and drug data was indeed due to paralog redundancy. The researchers then experimentally validated top synthetic lethal paralog pairs by developing a fluorescent reporter-based competitive fitness assay in lung cancer cells, where they verified that cells with both paralog genes knocked out were less competitive than those lacking one paralog gene.

To understand which synthetic lethal paralogs are shared across cell types and which are cell lineage-specific, Parrish et al. repeated this pgPEN screen in HeLa cervical cancer cells. The researchers found that while many of the top synthetic lethal pairs were shared between cervical and lung cancer cells, others were cell-type specific. Interestingly, the cell lineage-specific synthetic lethal interactions could not be explained by differences in expression between the two lines, demonstrating that paralog dependencies are modified by cellular context. Furthermore, they identified several gene paralog pairs that when knocked out, promoted cell growth, indicating these paralog families likely had tumor suppressive functions that require loss of the family to reveal the cellular phenotype. Interestingly, they identified four tumor suppressor paralogs in lung cancer cells, none of which were shared with the six tumor suppressor paralog pairs identified in cervical cancer cells. Collectively, this work identified putative paralog gene targets for lung cancer therapies in addition to uncovering tumor suppressor paralogs. More generally, this study highlights the importance of duplicate paralog genes, encouraging scientists to consider how pervasive these functionally redundant genes may be in their own research and providing a method for researchers to test this question in their system.

In addition to choosing to publish this work in an open access journal, Parrish is working to further enhance the accessibility of her research by now working to publish the computational methods and code generated for the data analysis of this project. Apart from science, Parrish describes that one of the amazing aspects about working on this project was conducting this research in the currently all-female Berger Lab, which fosters an inclusive, diverse environment that Parrish felt was critical in advancing her own science. This research experience further emphasized the “importance of funding basic science research, which really fuels discoveries,” but also “funding and supporting diverse individuals within science,” Parrish notes. Beyond the Berger Lab, Parrish has worked to promote diversity and inclusion by establishing a diversity program within UW Genome Sciences, which she credits as being a source of motivation throughout the highs and lows of her own research.

This work was supported by the Lung Cancer Research Foundation, the National Science Foundation, the National Institutes of Health, National Cancer Institute, National Institute of Diabetes and Digestive and Kidney Diseases, National Lung, Heart and Blood Institute, the Prevent Cancer Foundation, the Stephen H. Petersdorf Lung Cancer Research Award, the Seattle Translational Tumor Research Program, the Innovators Endowed Chair, the Edward P/ Evans Foundation, the Leukemia and Lymphoma Society, Mark Foundation for Cancer Research, the Department of Defense Breast Cancer Research Program, Paul G. Allen Frontiers Group, and the McIlwain Family Endowed Chair in Data Science.

Fred Hutch/University of Washington/Seattle Children's Cancer Consortium members Alice Berger and Robert Bradley contributed to this research.

Parrish PCR, Thomas JD, Gabel AM, Kamlapurkar S, Bradley RK, Berger AH. Discovery of synthetic lethal and tumor suppressor paralog pairs in the human genome. Cell Rep. 2021 Aug 31;36(9):109597. doi: 10.1016/j.celrep.2021.109597. PMID: 34469736; PMCID: PMC8534300.