Genentech last week presented data demonstrating target gene knockdown in cell culture and in cancer animal models using antibody-siRNA conjugates.
Despite the promising findings, which stemmed from Genentech’s one-time interest in therapeutic RNAi, the company is no longer actively pursuing drugs based on the gene-silencing technology, the Genentech researcher who presented the data told Gene Silencing News.
Still, she and her colleagues have been able to work on the approach as a side project, she added.
Genentech had been working on RNAi-based drugs since at least 2008, when it inked a deal to gain access to Thermo Fisher Scientific’s siRNAs to develop drugs for undisclosed indications (GSN 11/6/2008).
The company made little public about its efforts in RNAi since then, and questions arose about its commitment to the technology after Roche, which fully acquired Genentech in 2009, announced in late 2010 that it was shutting down its in-house RNAi research and development activities as part of a corporate-wide cost-savings initiative (GSN 11/18/2010).
Nonetheless, scientists within Genentech have been permitted to continue working on RNAi, according to Trinna Cuellar, a molecular biology researcher at the firm who presented data from the project at the GTC Nucleic Acid Summit last week in San Francisco. And as part of such efforts, Cuellar and colleagues have been investigating the use of antibodies to carry therapeutic siRNA payloads into solid tumors, she said at the conference.
“One of the key challenges of developing RNAi as a therapeutic is delivery,” she said during a presentation at the event. When it comes to getting siRNAs into solid tissue, this means that the oligos need to “survive circulation, enter the desired cells, escape the endosomes, and then incorporate into active RISC complexes.”
Given Genentech’s expertise in antibodies — the company earlier this year received US Food and Drug Administration approval for an antibody-drug conjugate that delivers chemotherapy specifically to tumor cells as a treatment for Her2-positive metastatic breast cancer — Cuellar was interested in testing whether the same approach could be used for siRNA delivery.
Genentech’s antibody conjugate technology is unique, she noted, because drugs can be loaded onto specific positions on the antibody, namely heavy chains or light chains, which helps in create homogenous drug product — a critical aspect of developing clinic-ready drugs.
For example, in 2005 researchers from Harvard Medical School reported on the successful delivery of siRNAs in vivo using a protamine-antibody fusion protein. While those researchers were able to deliver the siRNAs specifically into cancer cells of interest, in Genentech’s hands the system tended to form “large aggregates,” Cuellar said at the GTC meeting.
As a result, company scientists could not obtain “pure drug” and had difficulty achieving efficient gene knockdown.
Having a wide portfolio of antibodies to work with, Cuellar and others at Genentech then explored alternative antibodies to deliver the siRNAs they had obtained through the company’s collaboration with Thermo Fisher Scientific, she said. Among these were antibodies that deliver to “a variety of antigens,” including ones that internalize slowly via the endocytic recycling machinery or rapidly into lysosomes.
The Genentech team also examined the use of cleavable and non-cleavable linkers to attach their siRNAs to the antibodies, thinking that “maybe the siRNA would need to be detached from the antibody … before it entered RISC,” she said during her presentation.
In order to avoid confounding effects that could result from targeting genes associated with cell viability, Cuellar and her colleagues focused on non-essential genes, namely PPIB, a housekeeping gene that encodes cyclophilin B.
They also ran their in vitro experiments in cells that either naturally expressed the antigens for the antibody carriers being studied, and in ones engineered to do so in order to see whether their antibody-RNAi conjugates, or ARCs, could target cells expressing low, medium, or high levels of the antigen.
Using antibodies that target lysosomal-trafficking antigens, the Genentech researchers first looked to see whether the attachment of an siRNA onto an antibody would impact its gene-silencing activity.
When compared to free siRNAs and controls, ARCs against PPIB silenced their target “as well, if not better” than the others when transfected into cells, “suggesting that attachment to the antibody does not impair function of the siRNAs,” Cuellar noted.
Additional analysis confirmed that the knockdown effect was sequence-dependent and not caused by free siRNAs that may have inadvertently been present alongside the ARCs.
The researchers then examined whether the conjugation of siRNAs to the antibodies would impact antibody binding to cells, and found that ones incorporating both cleavable and non-cleavable linkers performed “almost as well” as naked antibodies.
Next, the team tested whether the ARCs were capable of delivering their RNAi payload, transfecting them into cells, as well as adding them onto cells to permit receptor-mediated endocytosis.
For several of the ARCs tested, the researchers saw “pretty good” silencing, upwards of 95 percent in some cases, Cuellar said. Importantly, silencing was achieved with ARCs featuring cleavable or non-cleavable linkers, although those with non-cleavable linkers supported “a greater degree” of silencing at the protein level.
The gene inhibition was found to be dependent on siRNA sequence, as well as the covalent conjugation to the antibody carrier and the expression of the relevant antigen, she added. “If the antigen isn’t expressed on the cells, you get no silencing.”
Concerned that the effects observed were the result of a non-RNAi effect, the scientists conducted 5’ RACE analysis, confirming that the silencing was, indeed, the result of bona fide RNAi.
When the Genentech team tested recycler antibodies carrying siRNAs, they were also able to achieve target knockdown via RNAi, again with ARCs with both cleavable and non-cleavable linkers.
Notably, not every ARC tested resulted in gene silencing — an outcome Cuellar said is still a mystery since the siRNAs internalized into the cells. “Tons of siRNAs are getting into the cell, but not … RISC,” she said, which shows that “we still don’t understand which pathway can lead to productive silencing.”
The Genentech investigators then tested the ARCs that were most effective in vitro for their ability to deliver siRNAs to tumor cells in rodent xenograft models.
They found that they could achieve “moderate” delivery specifically into tumors, and not in any other tissues including the liver, Cuellar said.
“Most of the delivery we see is near blood vessels,” she added, “but we were excited to see that we could have targeted delivery to tissues outside of the liver and outside of blood.” Additional testing confirmed that the ARCs could trigger target gene silencing in the tumor cells, although this, too, was moderate.
Despite the findings, the Genentech researchers are still unclear on how the antibody conjugates deliver siRNAs into cells and RISC, Cuellar said. “We know they are getting into the endocytic machinery, but it is unclear … how they are getting out into the cytoplasm.”
Speculating that certain endocytic pathway components may be at work, the scientists performed experiments where they knocked down different endocytic pathway genes in cells, waited a few days, then added ARCs and looked for either enhancement or abrogation of target silencing.
One gene that caught their attention was HPS4, a regulator of lysosome biogenesis that, when knocked down, enhanced RNAi silencing. Cuellar noted that this finding jibed with a previous report that knockdown of the gene enhanced the activity of transfected siRNAs in human cells.
Exactly how HSP4 functions as a negative regulator of ARC silencing is not known, but Cuellar said that its inhibition could prove to be a way to boost the RNAi activity of functional ARCs, although its suppression does not appear to confer gene-silencing abilities to non-functional conjugates.