A critical appraisal of the evidence for antiviral RNA interference in mammals, accepting that mouse embryonic stem cells generate virus-derived small interfering RNAs while finding the somatic cell case unproven, because the viral proteins invoked also antagonize interferon.
Cullen BR, Cherry S, tenOever BR. Is RNA Interference a Physiologically Relevant Innate Antiviral Immune Response in Mammals? Cell Host & Microbe. 2013. 14(4), 374-378.
DOI 10.1016/j.chom.2013.09.011. PMID 24139396. No PMCID recorded in the inventory.
Article type is Minireview, labeled as such in the running head of the article itself. It presents no new experimental data, and every experimental result discussed in this record was produced by the laboratory named with it, not by the authors of this piece.
A critical appraisal of the evidence for antiviral RNA interference in mammals, accepting that mouse embryonic stem cells generate virus-derived small interfering RNAs while finding the somatic cell case unproven, because the viral proteins invoked also antagonize interferon.
RNA interference is an established antiviral mechanism in plants, nematodes, and arthropods, and whether it serves the same role in mammals had appeared settled in the negative until two 2013 Science papers reopened it. This Minireview evaluates those two papers, from the Voinnet group (Maillard and colleagues) and the Ding group (Li and colleagues), together with a study from the Sullivan group (Seo and colleagues) published in the same issue of Cell Host and Microbe. The authors accept that the evidence for small interfering RNA production in mouse embryonic stem cells is strong, since virus-derived small RNAs of the expected size and phasing were detected after encephalomyocarditis virus infection, were lost in Dicer negative cells, and a nodamura virus lacking its B2 suppressor was partly rescued in cells lacking all four Argonaute proteins. They are considerably more cautious about somatic cells and about the in vivo data, because the central inference there depends on treating the nodavirus B2 protein and the Ebola virus VP35 protein as selective inhibitors of small interfering RNA production, when both are double-stranded RNA binding proteins that also antagonize the interferon response, and the point mutation used to inactivate B2 suppressor function also abolishes double-stranded RNA binding. The authors set out the experiment they consider decisive, which is genetic rescue of a suppressor-deficient virus in RNA interference-compromised somatic cells combined with direct evidence of viral messenger RNA silencing, and note that it has not been done.
The article assembles the state of the field as of 2013. Dicer cleavage of long double-stranded RNA into small interfering RNAs, loading of one strand into an Argonaute-containing silencing complex, and the resulting cleavage of complementary RNA are established in invertebrates and plants, where viral suppressors of RNA silencing have evolved repeatedly, the nodavirus B2 protein being among the best characterized by the Ding and Schneemann groups. In mammals the picture had been different. Deep sequencing across many virus and host combinations, notably the survey by Parameswaran and colleagues, either failed to find virus-derived small interfering RNAs or found them at very low levels, and long double-stranded RNA transfected into mammalian somatic cells does not yield functional small interfering RNAs but instead triggers the interferon response, which is absent from most organisms that use RNA interference antivirally. Against this, work from the Filipowicz, Hannon and Svoboda groups had shown that germ cells, embryonic stem cells, and some embryonal carcinoma lines do process long double-stranded RNA productively, and interferon inducibility is reported to be absent in those same cell types, which the authors note as a suggestive complementarity rather than a demonstrated relationship.
Does RNA interference constitute a physiologically relevant antiviral immune response in mammals, and specifically, do the recently reported virus-derived small interfering RNAs in embryonic stem cells, in cultured somatic cells, and in infected mice actually direct the silencing of viral RNA and restrict virus replication?
The strategy of the piece is evidentiary rather than experimental. The authors take a standard for what would establish antiviral RNA interference, quoted from Maillard and colleagues, namely genetic rescue of a suppressor-deficient virus in host cells whose RNA interference machinery is compromised, and they add a second requirement of their own, direct demonstration that the detected small RNAs silence viral messenger RNAs. They then apply that standard in turn to the embryonic stem cell data, the cultured somatic cell data, and the in vivo data, and in each case identify which of the two requirements is met and which alternative explanation remains open. A recurring analytical move is to ask whether a phenotype attributed to loss of suppressor activity against RNA interference could equally be explained by loss of double-stranded RNA binding and therefore loss of interferon antagonism.
The numbered entries below are the findings of the cited laboratories as the Minireview reports them, followed by the assessment the authors of the Minireview offer.
The article does not establish a mechanism, and its purpose is the opposite, namely to specify which mechanistic claims the available data can and cannot support. The authors accept a Dicer-dependent, Argonaute-dependent pathway that generates virus-derived small interfering RNAs in mouse embryonic stem cells and is lost on differentiation, with the molecular basis of that loss unknown. For somatic cells they leave the question open and identify the specific ambiguity that prevents resolution, which is that the nodavirus B2 protein and Ebola virus VP35 bind double-stranded RNA and therefore have at least two separable potential functions, blocking Dicer access and blocking pattern recognition receptor-driven interferon induction, and that the mutants used do not separate them. The figure accompanying the article presents interferon, microRNA, and RNA interference as three systems used to different degrees across taxa, with somatic mammalian cells shown with uncertain RNA interference usage, and this is offered as a framework rather than as a result. A possible reciprocal relationship between interferon competence and RNA interference competence is raised as a suggestion and is not tested anywhere in the work discussed.
The lasting contribution is a set of evidentiary criteria. The article states plainly what an experiment must show before a virus-derived small RNA can be called an antiviral effector, namely genetic rescue of a suppressor-deficient virus in cells whose RNA interference machinery is genetically compromised, combined with direct evidence that the small RNAs silence viral RNA, and it names the specific control that separates suppression of silencing from interferon antagonism when double-stranded RNA binding proteins are involved. It also identifies cell state rather than viral countermeasure as the more likely explanation for why embryonic stem cells and somatic cells differ, on the grounds that a strictly viral explanation cannot account for encephalomyocarditis virus producing abundant small interfering RNAs in stem cells and none in somatic cells. Those criteria make the disagreement testable rather than rhetorical.
The question of whether mammalian cells mount antiviral RNA interference recurs throughout this corpus, and this Minireview marks the position taken in 2013, which is skeptical for somatic cells and accepting for embryonic stem cells. The tenOever laboratory contribution to the argument is the demonstration by Perez and colleagues that influenza NS1 does not suppress RNA interference, together with the laboratory's use of the small RNA machinery as an engineering tool, cited here as the ability to build viruses that express small RNAs against host messenger RNAs. Both of those threads run forward into later corpus entries on microRNA-targeted viruses and on RNase III biology. Setting this article beside later primary work from the laboratory on RNA interference and on RNase III nucleases as antiviral effectors would be category 3 synthesis, and it should be assembled centrally rather than asserted from this article, which by its nature reports the state of an argument at one moment and is not evidence about work published afterwards.
The article contains no new data and its conclusions are only as good as the three studies it evaluates, all of which rest heavily on one virus, nodamura virus, and one suppressor protein, B2, in mouse and hamster systems. The authors are explicit that the functional relevance of viral small interfering RNAs in mammalian somatic cells remains an open question, that the in vivo attenuation data do not establish that small interfering RNAs are antiviral, and that a detailed identification of which cell types in an animal produce viral small RNAs is still needed. Nothing in the article addresses human viral pathogens directly, and the authors name that as the critical unaddressed question. As a Minireview it also reflects the judgment of its three authors at the time of writing and should not be read as a consensus position of the field or as evidence about experiments published later.
A 2013 Minireview weighing new claims that RNA interference fights viruses in mammals, accepting the stem cell evidence and finding the somatic cell case unproven.
RNA interference protects plants and insects from viruses, but whether mammals use it the same way has been disputed. Reviewing three 2013 studies, these authors accept that mouse embryonic stem cells generate virus-derived small silencing RNAs, and argue that the parallel claim for ordinary body cells is not yet supported, because the viral proteins used to make the case also block interferon, so the two explanations have not been separated experimentally.
Two 2013 Science papers reopened the question of whether RNA interference is an antiviral immune pathway in mammals, and this Minireview assesses them alongside a study in the same Cell Host and Microbe issue. The authors find the embryonic stem cell evidence convincing at the level of small RNA production, since encephalomyocarditis virus yields abundant Dicer-dependent small interfering RNAs in these cells and a nodavirus lacking its B2 protein is partly rescued when all four Argonaute proteins are removed. They are unpersuaded for somatic cells and for infected mice, because the inference there depends on B2 and on Ebola VP35 acting selectively against silencing, when both bind double-stranded RNA and also antagonize interferon, and the mutations used remove both activities together. They set out the experiments that would settle it, including challenge of interferon-deficient mice and direct demonstration that the small RNAs silence viral messenger RNA.
lab-led for Backes 2014, Benitez 2015 on engineered RNA interference, Aguado 2015 and the tenOever 2013 review, co-led with the Cherry laboratory for Backes 2012, and collaborative for Cullen 2013, a Minireview led by Bryan Cullen with Sara Cherry and tenOever