A Versatile RNA Vector for Delivery of Coding and Noncoding RNAs

lab-led

Replication-incompetent influenza-based vectors carrying no DNA intermediate deliver coding messages and functional small RNAs together to primary human and mouse cells and to mouse lung, with output and cytotoxicity tuned by microRNA target sites in the nucleoprotein segment.

Sonja Schmid; Lum C. Zony; Benjamin R. tenOever

2014 · Journal of Virology · methods/resource

Senior authors

Benjamin R. tenOever

Correspondence

Benjamin R. tenOever

Research areas & themes

Citation

Schmid S, Zony LC, tenOever BR. A Versatile RNA Vector for Delivery of Coding and Noncoding RNAs. Journal of Virology. 2014. 88(4), 2333-2336.

DOI 10.1128/jvi.03267-13. PMID 24307584. PMCID PMC3911536.

One-sentence contribution

Replication-incompetent influenza-based vectors carrying no DNA intermediate deliver coding messages and functional small RNAs together to primary human and mouse cells and to mouse lung, with output and cytotoxicity tuned by microRNA target sites in the nucleoprotein segment.

Executive summary

Delivering foreign RNA to a chosen tissue remains the practical bottleneck between RNA interference as a laboratory technique and RNA interference as a therapy, and the vectors in common use either integrate into the host genome or saturate the endogenous silencing machinery. This short report develops an alternative built on influenza A virus. Starting from a published hemagglutinin-deleted virus-like vector system propagated on complementing cells, the authors replaced the segment 4 open reading frame with a primary microRNA transcript and showed that the resulting vector expressed miR-124 in normal human dermal fibroblasts at levels roughly three times those from a replication-competent virus expressing the same microRNA from segment 8. The vector carried a green fluorescent protein message on one segment and a microRNA on another at the same time, and it processed all five hairpins of the miR-302/367 cluster from a single segment. Delivered miR-124 was functional, reducing its endogenous target polypyrimidine tract binding protein by 75 percent at 2 days. To control dose and toxicity the authors used a nucleoprotein segment bearing target sites for the ubiquitous miR-93, which lowered artificial microRNA output about fivefold while preserving target knockdown and removed the cytotoxicity seen with the wild-type nucleoprotein vector. Knockdown also worked in bone marrow derived macrophages, and intranasal delivery gave detectable miR-124 in whole lung. In vivo silencing was not tested.

Scientific context

Two facts frame the work. First, no RNA virus lacking a DNA intermediate had been found to encode a canonical microRNA, yet several groups including this laboratory had already shown that RNA viruses can be engineered to produce functional small RNAs. Second, artificial microRNAs with perfect complementarity silence far more effectively than endogenous microRNAs with partial complementarity, which makes them attractive as therapeutics, and yet in vivo delivery remains the limiting step. Existing options carry known liabilities. Vectors derived from DNA viruses and lentiviruses integrate and can perturb endogenous gene expression, and sustained high-level expression of small RNAs from integrating vectors had been reported by the Kay group to saturate the endogenous silencing pathway with lethal consequences in mice. Influenza-based vectors with limited replicative capacity had already reached regulatory approval in the vaccine setting, which the authors cite as evidence for the platform's safety profile. The specific starting material was a hemagglutinin-deleted influenza virus-like vector system and a segment 4 construct with green fluorescent protein flanked by hemagglutinin packaging sequences, both described previously by other groups and by this laboratory.

Central question

Can a replication-incompetent, influenza-derived RNA vector that never passes through a DNA intermediate deliver functional small RNAs, alone or together with a protein-coding message, to primary cells and to the respiratory tract, and can its output and its toxicity be tuned rather than merely accepted?

Experimental strategy

The design treats the influenza genome as a modular cassette system. Because the virus has eight segments, a coding message and a noncoding RNA can be assigned to different segments and tested for independent expression, and because packaging signals rather than open reading frames define a usable segment, an open reading frame can be swapped for a primary microRNA transcript. Replication incompetence comes from deleting hemagglutinin and supplying it in trans from a complementing MDCK line, which confines propagation to the producer cells. To make dose adjustable the authors reach for the reverse of a delivery function, using a previously characterized nucleoprotein segment carrying perfect target sites for miR-93, a microRNA expressed in essentially all mammalian cells, so that the host silencing machinery itself limits vector activity. Since that segment attenuates nucleoprotein, a second complementing line supplying both hemagglutinin and nucleoprotein was built to produce it. Readouts are deliberately layered. Small RNA northern blot shows that the RNA is made, western blot and quantitative PCR of an endogenous or transgenic target show that it works, a viability assay shows what the vector costs the cell, and intranasal delivery asks whether any of it survives contact with an animal.

Key findings

1. A virus-like vector expressing the primary miR-124-2 transcript from segment 4 produced about threefold more miR-124 in normal human dermal fibroblasts than a replication-competent influenza expressing the same microRNA from segment 8 (Fig. 1B). The authors report this comparison as unexpected.
2. A single vector carrying green fluorescent protein on segment 4 and miR-124 on segment 8 gave both high green fluorescent protein expression and high miR-124 levels in primary human fibroblasts at 1 day (Fig. 1C and 1D), establishing simultaneous delivery of coding and noncoding RNA.
3. The five-hairpin miR-302/367 cluster cloned into segment 4 yielded detectable miR-302 and miR-367 (Fig. 1E). Because these are the 5' and 3' ends of the transcript, the authors interpret detection of both as evidence that all five hairpins were processed, an inference the northern blot cannot confirm directly since it does not resolve miR-302a through miR-302d.
4. Vector-delivered miR-124 was functional. Polypyrimidine tract binding protein, an established endogenous miR-124 target, fell by 75 percent at 2 days and 60 percent at 3 days after treatment (Fig. 2A), with miR-124 detectable at 1, 2, and 3 days (Fig. 2B).
5. A nucleoprotein segment targeted by miR-93 reduced artificial microRNA against green fluorescent protein about fivefold relative to the wild-type nucleoprotein vector, bringing it to levels comparable with endogenous microRNAs (Fig. 3B), and yet both vectors knocked down green fluorescent protein transcripts comparably at 1 day in primary lung cultures from green fluorescent protein transgenic mice (Fig. 3C).
6. Cytotoxicity separated the two vectors. The wild-type nucleoprotein vector left roughly 50 percent cell survival at 4 days, while the miR-93-targeted vector was not significantly different from mock treatment (Fig. 3D). Both gave 42 percent knockdown of green fluorescent protein protein at 4 days (Fig. 3E), and nucleoprotein expression was lower with the targeted vector as expected.
7. In bone marrow derived macrophages, the vector expressed miR-124 at 1 day (Fig. 4A), reduced green fluorescent protein transcripts by approximately 40 percent at 1 day (Fig. 4B), and reduced green fluorescent protein protein in nucleoprotein positive cells at 3 days by flow cytometry (Fig. 4C).
8. Intranasal treatment of C57BL/6 mice with the miR-124 vector gave detectable miR-124 in whole lung at 1 day (Fig. 4D). The authors state explicitly that demonstrating in vivo silencing will require future studies, so delivery in the animal is shown and silencing in the animal is not.
9. The authors make a priority claim in the text, stating that this is the first demonstration that replication-incompetent virus-like vectors lacking a DNA intermediate can be used to deliver small RNAs. That claim is the paper's own and is recorded here as such.

Mechanistic model

This is a vector engineering report and it does not set out to establish a biological mechanism. The mechanisms it relies on are already established elsewhere and are used here as tools, namely Drosha and Dicer processing of a primary microRNA transcript delivered as part of a viral segment, loading of the mature small RNA into the silencing complex, and cleavage of a perfectly complementary target, alongside the reciprocal use of perfect host microRNA target sites in an essential viral segment to have the host silencing machinery destroy the vector's own message. What the data constrain is the input and output relationship. Small RNA abundance, target knockdown, and cytotoxicity can be moved by choice of segment and by the presence of miR-93 target sites in the nucleoprotein segment. What the data do not constrain is why a fivefold reduction in artificial microRNA levels leaves knockdown essentially unchanged, and the authors do not propose a mechanism for that dissociation. The relationship between nucleoprotein expression and the observed cytotoxicity is correlative here, since nucleoprotein level, vector activity, and toxicity all move together and were not separated.

Conceptual or technical advance

The report converts an attenuated influenza system into a tunable RNA delivery platform with three properties that are hard to obtain together. The vector carries no DNA intermediate and does not integrate, it can carry a protein-coding message and one or more small RNAs at once, and the amount of small RNA it produces is set by an engineered feature rather than by the dose given, which in these experiments decoupled useful silencing from measurable cytotoxicity. The introduction of a hemagglutinin and nucleoprotein complementing cell line is the enabling technical step, since it allows production of vectors whose essential nucleoprotein is itself silenced in the target cell. The authors note that a tissue-restricted microRNA could be substituted for the ubiquitous miR-93 to add tissue specificity, and that loop architecture can be used to tune processing efficiency, both of which are proposed extensions rather than results.

Relationship to the broader research program

This paper sits at the junction of two lines that run through the corpus. One is the engineering of RNA viruses to make small RNAs, represented in the reference list by earlier laboratory work on engineered RNA viral synthesis of microRNAs, on noncanonical cytoplasmic processing of viral microRNAs, and on in vivo delivery of cytoplasmic RNA virus-derived microRNAs. The other is the use of host microRNA target sites to restrict viral gene expression, represented by the laboratory's miR-93 attenuation of influenza and by the hematopoietic restriction of influenza through miR-142, both cited here. The novelty relative to those is the combination, since the same microRNA target site logic that was previously used to interrogate biology is used here to make a therapeutic-style vector safe. Placing this vector work beside later corpus entries that deliver or restrict RNA cargo in animals would be category 3 synthesis and belongs to central assembly.

Related publications

• Varble et al. 2010, engineered RNA viral synthesis of microRNAs. Methodological foundation, the source of the segment 8 microRNA-expressing influenza used here as the comparison virus.
• Langlois et al. 2012, in vivo delivery of cytoplasmic RNA virus-derived miRNAs. Predecessor from the same laboratory, cited for the estimate of roughly fifty thousand microRNA copies per cell from influenza-based expression systems.
• Perez et al. 2009, microRNA-mediated species-specific attenuation of influenza A virus. Methodological foundation, the source of the miR-93-targeted nucleoprotein segment used to tune vector activity.
• Langlois et al. 2012, hematopoietic-specific targeting of influenza A virus. Companion, cited as prior use of host microRNA target sites to restrict influenza to selected cell types.
• Chua et al. 2013, influenza A virus utilizes suboptimal splicing to coordinate the timing of infection. Methodological foundation, cited for the virus-like vector production system used here.
• Shapiro et al. 2010, noncanonical cytoplasmic processing of viral microRNAs. Predecessor from the same laboratory on small RNA processing from RNA viral contexts.
• Varble et al. 2013, an in vivo RNAi screening approach to identify host determinants of virus replication. Conceptual extension, cited as the source of the artificial microRNA against green fluorescent protein used here.
• tenOever 2013, RNA viruses and the host microRNA machinery. Review or synthesis from the same laboratory, cited for the prospect of using cell or tissue-specific microRNAs.

Limitations and boundaries

The animal data show delivery only. Detection of miR-124 in whole lung after intranasal treatment does not demonstrate silencing of any target in vivo, and the authors say so. All silencing results come from cell culture, over a window of one to four days after treatment, and the report contains no data on durability beyond that, on repeat dosing, on immune responses to the vector, or on biodistribution. Two of the three knockdown readouts use a transgenic green fluorescent protein reporter rather than an endogenous gene, and the one endogenous target tested is polypyrimidine tract binding protein with a single microRNA, miR-124. The processing claim for the miR-302/367 cluster rests on detecting the terminal members, since the northern blot cannot distinguish the four miR-302 species. Cell survival was measured with a single luminescent viability assay at one time point and does not address sublethal effects or saturation of the endogenous silencing machinery, which is precisely the failure mode the authors invoke against integrating vectors. Finally, the system is built on one mouse-adapted influenza strain and one complementing cell background, and the safety inference drawn from approved live attenuated influenza vaccines is an argument by analogy rather than a result from this study.

Audience summaries

25 words

An influenza-derived vector that cannot replicate delivers protein-coding messages and silencing RNAs together into primary cells and mouse lung, with output tuned to avoid killing cells.

75 words

Turning RNA interference into medicine depends on getting the RNA where it is needed. This work adapts a crippled influenza virus into a delivery vehicle that never becomes DNA and cannot integrate into the genome. It carried a fluorescent reporter and a silencing RNA at once, knocked down target genes in human and mouse primary cells, and reached mouse lung after nasal delivery. Building in a self-limiting switch removed the vector's toxicity without losing silencing.

150 words

This short report develops replication-incompetent influenza-based virus-like vectors as a delivery system for small RNAs. Replacing the segment 4 open reading frame with a primary microRNA transcript produced roughly threefold more miR-124 in primary human fibroblasts than a replication-competent virus expressing it from segment 8, and a single vector delivered a fluorescent protein message and a microRNA at the same time, or processed a five-hairpin microRNA cluster. Delivered miR-124 lowered its endogenous target polypyrimidine tract binding protein by 75 percent. To control dose, the authors used a nucleoprotein segment bearing target sites for the ubiquitous miR-93, which cut small RNA output about fivefold, preserved target knockdown, and abolished the cytotoxicity of the untargeted vector, and they built a hemagglutinin and nucleoprotein complementing cell line to produce it. Knockdown also worked in bone marrow derived macrophages. Intranasal delivery placed miR-124 in mouse lung, though in vivo silencing was not assessed.

Discoveries supported by this paper

Documented publication relationships

• Engineered RNA viral synthesis of microRNAs — methodological foundation. Direction: this paper → linked paper. stated in the Related publications section of 2014-schmid-a-versatile-rna-vector-for-deliver
• Influenza A Virus Utilizes Suboptimal Splicing to Coordinate the Timing of Infection — methodological foundation. Direction: this paper → linked paper. stated in the Related publications section of 2014-schmid-a-versatile-rna-vector-for-deliver
• Noncanonical cytoplasmic processing of viral microRNAs — predecessor. Direction: this paper → linked paper. stated in the Related publications section of 2014-schmid-a-versatile-rna-vector-for-deliver

Pathogens

Technologies