By Doug Macron
As it nears the end of its second stage, the Broad Institute's RNAi Consortium is already preparing for its next phase of work developing RNAi technologies for biological research.
And while the exact goals of the so-called TRC3 have yet to be defined, the initiative expects that these will include the expansion of its scope to include reagents targeting non-coding RNAs, TRC Project Leader David Root told Gene Silencing News last week.
"We are … learning a lot more about what really constitutes the full mammalian transcriptome," he explained, noting that deep sequencing and other methodologies have revealed "a large number" of previously unidentified ncRNAs, while work at Broad and elsewhere "has shown that these … can be really important functionally, as many people suspected."
While miRNAs have received much attention lately amid growing evidence that they play key roles in numerous biological processes and disease, there are a variety of other ncRNAs that have been discovered, including ones dubbed large intergenic non-coding RNAs, Root noted.
In 2009, researchers from Broad reported in Nature on the use of "chromatin-state maps to discover discrete transcriptional units intervening known protein-coding loci … [which] identified approximately 1,600 large multi-exonic RNAs across four mouse cell types."
As part of that work, the team identified "a diverse range of roles for lincRNAs in processes from embryonic stem cell pluripotency to cell proliferation," they wrote.
In light of this and other work, TRC intends to extend its growing library of lentivirus-expressed shRNAs against human and mouse genes to also target the non-coding genome, while also exploring other technologies for ncRNA suppression where appropriate, Root said.
"The predominant tool for suppressing transcripts [are] short-hairpin RNAs," he said. "We have found those to work and be effective against non-coding RNAs, so that will [remain] the core of the library. But certainly we are looking at other types of tools to suppress gene function," as in the case of miRNAs, for example.
In a more significant departure from its primary focus of using RNAi for gene perturbation, TRC3 is also expected to begin building a library of reagents for gain-of-function experiments, Root said. At Broad, "we've been working on [open reading frame] libraries outside of The RNAi Consortium, but we want to make that part of the [consortium's] efforts," he said.
"Complementary to suppressing transcripts is to over-express them," he noted. "People have been doing that for some time, but I think there is still a lot of room to improve the resources for gain-of-function cell-based screening using over-expression constructs. We are working on that and will make it part of phase three."
As of right now, however, TRC3's mission has not yet been fully defined.
Recognizing that there are "a number of things that will make the RNAi tools yet more powerful for biological discovery … we're in the process of putting together that [next stage] and its membership, [while] defining and refining the goals," Root said. As of right now, TRC3 "is not locked in, so to speak, with what exactly it will be."
The next stage in the initiative is, however, expected to begin immediately after The RNAi Consortium's currently ongoing second stage ends on April 1.
Building a Consortium
In early 2004, Broad established TRC, a public/private consortium of RNAi researchers focused on the development and public release of genome-scale sets of virally expressed shRNAs targeting mouse and human genes (GSN 4/9/2004).
[ pagebreak ]
During that first, three-year phase, TRC created a library of about 160,000 shRNAs against roughly 32,000 human and mouse genes, which is available from Sigma-Aldrich and Thermo Fisher Scientific's Open Biosystems unit. Concurrently, TRC developed methods to use the library for loss-of-function genetic screens.
Members of TRC1 included Harvard Medical School, Massachusetts Institute of Technology, Dana-Farber Cancer Institute, Massachusetts General Hospital, the Whitehead Institute for Biomedical Research, Bristol-Myers Squibb, the Ontario Institute of Cancer Research, Sigma-Aldrich, and Academia Sinica in Taiwan. Novartis and Eli Lilly also participated.
In 2007, TRC began a second, four-year phase focused on improving the "power of that library as a screening resource," Root said last week.
To do so, TRC2 began expanding the library. "Of course, more shRNAs per gene give you more independent measurements per gene and more of a way to monitor off-target effects," he said. "And it just gives you more chances to come up with highly effective hairpins … that have strong knockdown and distribution."
TRC2 is in the process of doubling the size of its shRNA library, not only increasing the average number of shRNAs per gene from five to eight, but also boosting the overall number of genes targeted.
"In the time between TRC1 and TRC2, there were refinements of the human and mouse genomes and there was new information about the [most] reliable genes and transcript sequences to target," Root noted. "So we added 8,000 new genes, and the TRC2 library will have 40,000 mouse and human targets covered" once TRC2 concludes in April.
At the same time, TRC2 took steps to comprehensively validate the shRNAs in its library.
"We actually have obtained [knockdown measurement] data for nearly 100,000 of the shRNAs" targeting nearly 16,000 different genes, Root said. By the end of the consortium's second phase, close to 150,000 shRNAs against roughly 20,000 genes will have been validated.
"The value of having these knockdown measurements on such a large number of shRNAs is multifold," he said. "Of course, we have incredibly good statistics on the library's performance overall, but for a large proportion of the genes that we've targeted, we have a gene-by-gene assessment of how well our library covers those genes.
"Beyond that, because we have a quantitative assessment of the knockdown of the shRNAs … when you actually go to use these libraries for biological discovery, you can compare the phenotype you see with the degree of knockdown of the target gene across different shRNAs on a gene-by-gene basis," he added.
Although improving the shRNA library was its main goal, TRC2 also has ancillary projects centering around improving the overall understanding of shRNA design and processing, which "obviously feeds into being able to make the best possible library," Root said.
"To do that … we created over 1,000 shRNAs that used less-biased design in order to investigate the best rules for designing shRNAs, with the main focus being on the target sequence one should aim for, but also looking at some other design parameters like the loop and the length of the stem."
TRC2 investigators also examined how lentiviral-delivered shRNAs are processed, infecting cells and then creating libraries of small RNAs from them that can be matched to the hairpins, he added.
In its second phase, the consortium has also been developing better methods for applying its shRNA library for screening experiments, including pooled screening methodology.
[ pagebreak ]
"That's been worked on for some time by other groups for shRNAs … but we've been trying to push the frontiers in terms of being able to efficiently, effectively screen large pools of shRNAs and get very robust answers from those pooled screens," Root said.
In 2008, TRC2 published a paper in the Proceedings of the National Academy of Sciences detailing the work and results obtained in different cancer cell lines, and consortium researchers have recently submitted a new paper on the screening methodology using larger shRNA pools, he noted.
Also expected to be published in the near future are data on an inducible version of the shRNA vector used in the TRC's library, which can be used for "follow-up experiments and special purposes," Root added. "It has worked quite well … and is quite tight, so you can turn it off nearly completely," and has been tested both in vitro and in mice.
Harvard Medical School, MIT, Dana-Farber, Massachusetts General Hospital, Whitehead, Bristol-Myers Squibb, the Ontario Institute of Cancer Research, Sigma-Aldrich, and Academia Sinica, in addition to new member Johnson & Johnson, are members of TRC2. Novartis and Lilly did not take part in the second phase.
With the conclusion of TRC2 less than a year away, data from these projects and feedback from consortium members are expected to help draft the blueprint for TRC3.
Ultimately, "we'll do what's most important to make RNAi as powerful as possible for the kinds of biological discovery experiments that we and the rest of the biological research community want to do," Root said.