Rapid RNAi in mice unveils novel cancer genes

Modern genomics is uncovering hundreds of mutations associated with cancer, but mining this data for novel therapies is predicated on weeding out “bystander” alterations and identifying “driver” mutations as well as components responsible for tumorigenesis and metastasis. In this study, we devised and employed a novel direct in vivo RNAi screening methodology to simultaneously test function of hundreds of putative cancer genes within a single mouse and found novel tumors suppressors with hitherto unknown function – the results were recently published in Science.

HFSP Long-Term Fellows Daniel Schramek, Ataman Sendoel and Slobodan Beronja and colleagues
authored on Mon, 13 October 2014

In the hunt for genetic mutations that cause cancer, there is a lot of white noise. Human sequencing efforts have identified hundreds of genetic alterations in cancer, and the challenge now is to identify which ones are actually responsible for cancerous growth and metastasis. HFSP Long-Term Fellow Daniel Schramek and colleagues at the Rockefeller’s Laboratory of Mammalian Cell Biology and Development have created a new technique that can weed out random genetic and epigenetic changes in expression— the bystanders — and identify the ones that are critical for cancer.  Applying it to Head and Neck Cancers, the 6th most life-threatening cancer world-wide, they discovered seven new tumor-suppressor genes whose role in cancer was previously unknown.

Figure: Direct in vivo shRNA screen for Head and Neck Cancer tumor suppressors. (A)Schematic of pooled shRNA screen. Tumor-prone TGFβRII conditional knock-out mice are infected in utero with a lentiviral pool of up to 2000 shRNAs targeting 400 different putative cancer genes (5shRNAs/gene represented by a specific color) generating patches of knock-down tissue, where the targeted gene is repressed. These chimeric mice are then followed-up into adulthood for tumor development. shRNAs and therefore the corresponding gene responsible for tumor development can be identified and quantified by deep sequencing of genomic DNA from the tumor. (B) Representative immunofluorescence analyses of tumor sections from adult TβRII-cKO mice that had been infected with the lentiviral shRNA library at E9.5 in utero. Basal epithelial markers Keratin 5 and β4-integrin are shown in red and green, respectively, and show that the tumor invades deep into the dermis.

The new technique is particularly useful because it takes a fraction of the resources and much less time than the traditional method for determining gene function — breeding genetically modified animals to study the impact of missing genes.

“Using knockout mice, which are model organisms bred to have a particular gene missing, is not feasible when there are hundreds of potential genes to sort through,” says Daniel Schramek, a postdoctoral fellow in the lab headed by Rebecca C. Lancefield Professor Elaine Fuchs. “It can take about two years per gene. Our method can assess about 300 genes in a single mouse, in as little as five weeks.”

The researchers made use of RNA interference, a natural process whereby RNA molecules inhibit gene expression. They took short pieces of RNA which are able to turn off the function of specific genes, attached them to highly concentrated viruses, and then using ultrasound to guide the needle without damaging surrounding tissue, they injected the viruses into the sac of each mouse embryo.

“The virus is adsorbed and integrated into the chromosomes of the single layer of surface cells that cover the tiny embryo,” explains Fuchs, “as the embryo develops, this layer of cells makes the skin, mammary glands and oral tissue, enabling us to efficiently, selectively and quickly eliminate the expression of any desired gene in these tissues. The non-invasive method avoids triggering a wound or inflammatory response that is typically associated with conventional methods to knockdown a gene in cultured cells and then engraft the cells onto a mouse.”  When the mice grew, the researchers determined which genes, when turned off, were promoting tumor growth, and what they found was surprising. 

“Among the seven novel tumor suppressor genes we found, our strongest hit was Myh9, which codes for the protein myosin IIa, a motor protein with well-known function in cell structure and cell migration,” says Schramek. “Through further functional studies we found that myosin IIa is also required for activation of the main guardian of the genome — a tumor suppressor protein called p53.” When the myosin IIa gene is mutated, p53 is not able to build up in the cell nucleus, and chaos ensues: genes responsible for repairing damaged cells and killing off tumor cells are not activated, and invasive carcinomas spread within three months.

The researchers devised a strategy to reactivate p53 in these cells, and showed in vitro that tumor suppression was restored. “Interestingly, Myh9 is also mutated in human Head and Neck Cancers, and low expression of myosin IIa correlates with poor prognosis for the patient,” says Schramek. Hence the group hopes to examine the effect in clinical trials in the future. They also plan to look at the function of the other six genes their study identified. 

“We’ve demonstrated that this method of RNA interference is highly useful in the rapid discovery, validation and characterization of tumor suppressor genes that might otherwise be missed in a genetic screen,” says Schramek. “It can now be applied to many kinds of cancers, such as breast and lung.”


Direct in vivo RNAi screen unveils myosin IIa as a tumor suppressor of squamous cell carcinomas. Schramek D, Sendoel A, Segal JP, Beronja S, Heller E, Oristian D, Reva B, Fuchs E. Science.2014 Jan 17;343(6168):309-13. doi: 10.1126/science.1248627. PMID: 24436421

Science link

Pubmed link

Cancer Discovery Commentary

Front Oncol Commentary

F1000 link