SAN FRANCISCO — Although RNAi screens based on siRNAs have proven robust, using shRNA-containing lentiviral vectors in such screening experiments can offer considerable benefits in the right situations, according to a researcher from Merck subsidiary Rosetta Inpharmatics.
“The purpose of lentivirus RNAi screening is to provide assay flexibility — to be able to get into hard-to-transfect cell lines including primary, as well as non-dividing, undifferentiated cells,” Richard Klinghoffer, a biology research fellow at Rosetta, said here this week at Cambridge Healthtech Institute’s annual Beyond Genome conference.
“The greatest advantage, though, of a lentivirus platform is that it allows for long-term gene silencing,” he said. “Since the lentivirus … will integrate its genome into the host cell genome … we get stable integration and essentially limitless expression of your shRNAs, [which] allows the scientist to assess the effects of long-term gene silencing on phenotypes of interest.”
Yet the lentiviral approach is not appropriate in all situations, Klinghoffer warned, in part because of variable target gene-silencing efficiency and host cell responses. But with careful target cell selection, these pitfalls can be avoided, he said.
In general, there is a “feeling that [lentiviruses] can be used in all cases where the siRNA approach cannot,” Klinghoffer said. “This is simply not the case [because] the lentivirus approach does not perform equally well in all cell types.”
In cases where lentiviruses perform poorly, experiments may be plagued with variable silencing efficiency and a variety of unintended host cell responses, both of which can impact phenotypes of interest.
As such, “target cell selection is critical for lentivirus RNAi experiments,” he said.
To highlight his point, Klinghoffer cited a set of experiments conducted at Rosetta about three years ago in which “we simply took 12 different cell lines and infected them with a titration of” shRNA-expressing lentiviruses.
There is a “feeling that [lentiviruses] can be used in all cases where the siRNA approach cannot. This is simply not the case [because] the lentivirus approach does not perform equally well in all cell types.”
“In most cases, we could see, upon increasing concentrations of virus, nice target gene silencing,” he explained. “However, there were a handful of cell lines that were simply refractory to lentivirus-mediated RNAi.”
The experiments were repeated in the same cell lines using multiple shRNAs targeting a variety of different genes, and in some cases, with different vectors. Again, in certain cell lines, “we could just never see efficient target silencing,” Klinghoffer said.
One of these cell lines, the colon cancer DLD1 cell line, had “previously been a workhorse for us for a number of our siRNA screens,” he noted. Still, shRNAs failed to yield the same robust silencing results seen with siRNAs.
“Our initial hypothesis … was [that] these [cells] were just hard to infect,” Klinghoffer said. To test this, he and his colleagues infected cells with increasing concentrations of a GFP-expressing lentiviral vector.
Measuring GFP by flow cytometry, the Rosetta team found “nice expression of GFP” in the DLD1 cells, leading them to suggest that certain cell lines “may simply have weak shRNA-processing machinery,” he said.
Klinghoffer added that he and others at Rosetta have not had time enough to follow up on this particular hypothesis. Despite this, he said that the work indicates that “careful cell selection for the lentiviral platform, just as it is for the siRNA platform, is essential.”
Further emphasizing the point was the discovery that transducing cell lines with lentiviral vectors can trigger significant transcriptional responses, he said.
“We came across this conclusion in a roundabout way,” he said. “Initially, we wanted to infect cells — in this case, HeLa cells — with [two distinct] vectors targeting the tumor suppressor p10 … and do some pathway profiling.”
When comparing uninfected HeLa cells to both shRNA-containing lentiviruses and empty controls retaining all the components of the vector but without an shRNA payload, “we saw a large transcriptional response that really dwarfed the intended response of p10 silencing,” Klinghoffer said.
“We wanted to look for way to mitigate this response and … found the best way … to simply to try to use cell lines that didn’t show this response,” he added.
In a follow-up study, “we took two of the [common interferon response] genes that were up-regulated in our [p10] experiment … then titrated in the virus and followed the expression of these two interferon response genes in a number of different cell types,” Klinghoffer said.
”What we found, as expected from our microarray analysis of the HeLa cells, was that the HeLa cells resulted in a dose-dependent manner with increased [target gene] induction,” he said. “However, three other cell lines … did not show a response,” underscoring the necessity of careful target cell selection for any lentivirus-mediated RNAi experiment, particularly RNAi screens.
In the end, though, Klinghoffer stressed that there is a place for both siRNA- and shRNA-based approaches in screening experiments.
“The lentivirus platform, in my eyes, is [not] meant as a platform that will replace siRNA screening, but as a complement” to it, he said. “And we, indeed, run both platforms at Rosetta.”