Rapid vector-delivered destruction of the entire cellular microRNA population shows that post-transcriptional silencing makes no measurable contribution to the acute transcriptional response to double-stranded RNA or to type I interferon, while sustained loss of microRNAs derepresses a broad set of chemokines and proinflammatory cytokines in fibroblasts and in mouse lung.
Aguado LC, Schmid S, Sachs D, Shim JV, Lim JK, tenOever BR. microRNA Function Is Limited to Cytokine Control in the Acute Response to Virus Infection. Cell Host & Microbe 2015, volume 18, issue 6, pages 714-722.
DOI 10.1016/j.chom.2015.11.003. PMID 26651947. PMCID PMC4683400.
Rapid vector-delivered destruction of the entire cellular microRNA population shows that post-transcriptional silencing makes no measurable contribution to the acute transcriptional response to double-stranded RNA or to type I interferon, while sustained loss of microRNAs derepresses a broad set of chemokines and proinflammatory cytokines in fibroblasts and in mouse lung.
Whether microRNAs participate in mammalian antiviral defence has been contested. Some reports assign microRNAs a substantial share of interferon's antiviral activity, others find no effect on virus replication, and separate work shows that the silencing complex is inactivated by cellular stress and by infection. Testing the question requires removing microRNAs quickly, in cells that are not transformed and not already responding to a replicating virus. The authors built that tool by placing the vaccinia virus poly(A) polymerase subunit VP55, which tails and degrades Argonaute-associated small RNAs, into a replication-incompetent adenovirus vector. The vector removes the abundant microRNAs within about a day, does not itself induce interferon-stimulated genes, and shows essentially no off-target transcriptional activity in cells already lacking Dicer. Applied to primary human foreskin fibroblasts, microRNA removal changed only twelve of the roughly 1,500 genes induced by transfected double-stranded RNA, and only twelve of the genes induced by six hours of interferon beta. Extending microRNA loss to nine days, by contrast, altered more than 1,700 transcripts, with central components of the antiviral machinery mostly unchanged but with a broad set of chemokines and cytokines strongly derepressed. Delivered intranasally to mice, the vector raised six of eighteen measured cytokines in whole lung within 48 hours and changed more than 2,400 transcripts, with the strongly induced set enriched for immune activation and adhesion categories. The authors conclude that microRNA control in the antiviral setting operates on cytokine output rather than on the intrinsic response itself.
Mammals defend against viruses chiefly with a protein-based stratified system of intrinsic effectors, innate cellular responses and adaptive immunity, in contrast to organisms that rely on pathogen-derived small RNAs. Whether the microRNA arm of the small RNA machinery nonetheless contributes has been disputed. Prior reports had claimed microRNAs account for as much as half of type I interferon's antiviral activity, while other work found many human viruses refractory to inhibition by endogenous microRNAs. Meanwhile, two lines of evidence suggested that the silencing complex is not available during infection at all, since it is inactivated by cellular stress and is ribosylated beyond roughly eight hours of infection. Earlier work from this laboratory had shown that poxviruses tail and degrade microRNAs through VP55 and that delivering VP55 from a replicating vesicular stomatitis virus attenuated that virus, but the inflammatory environment created by the replicating vector confounded any attempt to separate microRNA function on cytokines from secondary interferon-stimulated gene induction. The gap the paper addresses is therefore a clean, fast, inert removal of microRNAs in primary cells.
Do endogenous microRNAs shape the mammalian cellular response to virus infection, and if their influence is limited, what part of that response do they actually control?
The strategy is subtractive and depends entirely on the quality of the subtraction. VP55 removes microRNAs post-transcriptionally rather than blocking their synthesis, so depletion is fast and does not require waiting out the turnover of existing pools. Putting VP55 in a replication-incompetent adenovirus with a modified fibre for use in primary cells gives delivery into terminally differentiated cells without provoking an antiviral response, which the authors verify by transcriptome comparison of vector-treated and untreated fibroblasts. Specificity for small RNAs rather than messenger RNAs is tested in Dicer-deficient cells, where any transcriptome change caused by VP55 must be microRNA-independent. Stimuli are chosen to isolate different arms. Transfected double-stranded RNA drives the intrinsic response at a point that mimics peak viral replication but is cytotoxic, so interferon beta treatment is used in parallel because it produces a comparable transcriptional response without cell death and therefore allows a longer window. Because microRNA effects are individually small, the design then varies time rather than stimulus, comparing depletion at one day with depletion at nine days to ask whether the acute-phase null result reflects an absence of targets or simply insufficient time. Finally, the vector is given intranasally to mice so that cytokine protein, not just transcript, can be measured in tissue.
The paper does not establish a mechanism for how microRNA loss is coupled to cytokine output during infection, and the central proposal in the discussion is explicitly framed as an idea the data support rather than demonstrate. What the data constrain is the target set and the timescale. Post-transcriptional silencing exerts a constitutive, individually modest repression on a diverse group of cytokines and chemokines, demonstrated here by derepression on microRNA removal and, for IL6 and IRF1, by rescue with a protected mimic and by reporter assays with mutated target sites. That repression requires days rather than hours to change transcript levels appreciably, so it cannot shape the acute transcriptional response to double-stranded RNA or interferon, which unfolds over hours. From this the authors propose that the known inactivation of the silencing complex during infection, reported by others to occur beyond about eight hours, is itself a mechanism by which the host derepresses inflammatory transcripts without requiring new transcription, and they extend this to suggest that the influenza A virus NS1 protein, reported by others to block that inactivation, would thereby dampen cytokine output. Neither the causal link from silencing complex inactivation to cytokine burst during a real infection nor the NS1 consequence is tested in this paper. The split between transcript and protein responses in lung is described but its basis is not resolved.
The vector is a general tool. Rapid, near-complete and uniform removal of the cellular microRNA population in terminally differentiated primary cells, achieved post-transcriptionally and without provoking an antiviral response, makes it possible to define stimulus-specific microRNA target sets by subtraction, and it can be applied in vivo. Conceptually, the work separates two questions that had been conflated, whether microRNAs act during the antiviral response and whether they act on the antiviral response, and answers them differently. It also supplies a timescale argument that reconciles the conflicting literature, since assays run over hours would find nothing even where genuine targets exist. The reframing of microRNA function as a link between the intrinsic response and the recruitment and activation of immune cells makes the cytokine target set, rather than interferon-stimulated genes, the place to look for microRNA effects on infection outcome.
The work continues a line in the laboratory asking whether small RNA silencing is a physiologically relevant antiviral mechanism in mammals. Its own citations include earlier reports from the group on poxvirus degradation of host microRNAs through VP55 and terminal RNA methylation as a protective feature, on the mammalian response to virus infection being independent of small RNA silencing when VP55 is delivered from a replicating vesicular stomatitis virus, on noncanonical cytoplasmic processing of viral microRNAs, and on RNA viruses and the host microRNA machinery. The present paper is presented by the authors as an improvement on the earlier vesicular stomatitis virus delivery, because the inert adenovirus vector removes the confound of a replicating virus. Placing this against the laboratory's later work on antiviral RNA interference would be category 3 synthesis and is not attempted here.
The conclusions rest on removing microRNAs and observing what changes, which detects only repression that is active at baseline and would miss a microRNA whose function requires induction that the vector timeline does not allow. VP55 acts on poly(A) substrates and does affect some non-microRNA transcripts, notably histone cluster messages, so the Dicer-deficient control bounds but does not eliminate direct effects. The acute experiments use transfected double-stranded RNA and recombinant interferon beta as surrogates for infection rather than a replicating virus, which was a deliberate choice to avoid confounds but means no virus replication phenotype is measured here. Cell systems are human foreskin fibroblasts, immortalised fibroblasts and 293T lines, so the findings do not on their own extend to immune cell types, epithelium or other tissues. The nine-day depletion regime, while informative about kinetics, is not a physiological state for an infected cell and the authors accordingly frame the relevance as chronic infection or persistence rather than acute disease. The in vivo work is a single 48-hour timepoint in mouse lung after vector delivery with no pathogen present, with five animals per group and eighteen cytokines measured, and it therefore shows derepression rather than a consequence for infection outcome. The mechanistic claims about silencing complex inactivation during infection and about NS1 are taken from other laboratories' work and are not tested here.
Stripping cells of all microRNAs leaves the immediate antiviral response unchanged but releases a broad set of cytokines, placing microRNA control on inflammation rather than on defence itself.
MicroRNAs tune protein levels, and their role in antiviral immunity has been argued both ways. Using a harmless adenovirus vector carrying a poxvirus enzyme that destroys microRNAs within a day, the authors removed them from primary human fibroblasts and from mouse lung. The response to double-stranded RNA and to interferon was essentially unaffected. Only after days of depletion did large changes appear, concentrated in chemokines and cytokines that recruit and activate immune cells.
A replication-incompetent adenovirus delivering the vaccinia poly(A) polymerase subunit VP55 removes roughly 90 percent of abundant microRNAs from primary cells within a day without inducing interferon-stimulated genes and with minimal direct action on messenger RNAs, as judged in Dicer-deficient cells. In primary fibroblasts, microRNA loss altered 12 of 1,548 genes induced by transfected double-stranded RNA and 12 of 179 induced by interferon beta, with IRF1 among them and a miR-23 site in its untranslated region confirmed by reporter. Nine days of depletion, by contrast, changed over 1,700 transcripts while leaving most core antiviral components untouched, and the responsive set was dominated by chemokines and cytokines including IL6, whose repression was restored by a VP55-resistant let-7 mimic. Intranasal vector delivery raised six lung cytokines at the protein level in mice, in some cases without a matching transcript change. The authors frame microRNA control as acting on cytokine output over extended timescales.
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