Maximal interferon induction by influenza lacking NS1 is infrequent owing to requirements for replication and export

Abstract

Influenza A virus exhibits high rates of replicative failure due to a variety of genetic defects. Most influenza virions cannot, when acting as individual particles, complete the entire viral life cycle. Nevertheless influenza is incredibly successful in the suppression of innate immune detection and the production of interferons, remaining undetected in >99% of cells in tissue-culture models of infection. Notably, the same variation that leads to replication failure can, by chance, inactivate the major innate immune antagonist in influenza A virus, NS1. What explains the observed rarity of interferon production in spite of the frequent loss of this, critical, antagonist? By studying how genetic and phenotypic variation in a viral population lacking NS1 correlates with interferon production, we have built a model of the “worst-case” failure from an improved understanding of the steps at which NS1 acts in the viral life cycle to prevent the triggering of an innate immune response. In doing so, we find that NS1 prevents the detection of de novo innate immune ligands, defective viral genomes, and viral export from the nucleus, although only generation of de novo ligands appears absolutely required for enhanced detection of virus in the absence of NS1. Due to this, the highest frequency of interferon production we observe (97% of infected cells) requires a high level of replication in the presence of defective viral genomes with NS1 bearing an inactivating mutation that does not impact its partner encoded on the same segment, NEP. This is incredibly unlikely to occur given the standard variation found within a viral population, and would generally require direct, artificial, intervention to achieve at an appreciable rate. Thus from our study, we procure at least a partial explanation for the seeming contradiction between high rates of replicative failure and the rarity of the interferon response to influenza infection.

Modulation of Influenza A virus NS1 expression reveals prioritization of host response antagonism at single-cell resolution

“…Finally, our single cell analysis further highlighted that the observed difference in host response sensitivity is a combined effect of both variation in the ‘frequency” of NS1 (percentage of infected cells that are NS1 positive) and the overall NS1 “abundance” (the level of NS1 within individual cells). In the case of IFN and pro-inflammatory cytokine inductions, both the highest amount of viral transcription and the lowest NS1 level are required, consistent with recent findings by Vicary et al (2023) . During wild type IAV infection, this is an extreme rare event ( Supplementary Figure S5C ).…”

“…However, they found silencing NS1 does not affect host antiviral response, and the authors attributed this to the limited immune capacity in immortalized systems. Follow-up single cell transcriptomic studies highlighted that viral NS1 expression levels in-part anti-correlates with host immune gene expression but did not define which host pathways are more impacted by varying NS1 levels ( Russell et al, 2018 ; Ramos et al, 2019 ; Sun et al, 2020 ; Vicary et al, 2023 ).…”

The use of single-cell RNA-seq to study heterogeneity at varying levels of virus–host interactions

“…The key limitation requiring fragmentation and end-choice is due to both the read-length limitation of Illumina sequencing, as well as length requirements for efficient bridge amplification in an Illumina flow cell [ 39 ]. From this partial genetic information, several studies nevertheless were able to identify architectural changes to viral genomes, for instance, large internal deletions in influenza A virus [ 31 , 33 , 34 ]. These deletions form the basis of defective viral genomes (DVGs, also called nonstandard viral genomes or nsVGs), which are associated with a general loss of viral fitness as well as the induction of an innate immune response [ 40 , 41 ].…”

“…For viral reads, this will, in turn, depend on the level of viral transcripts in an infected cell. Thus, successful genotyping, including architectural variation, is impacted by the level of viral transcripts [ 34 ]. This can lead to certain incorrect conclusions; for instance, if the probability that a large deletion in the viral genome is detected is proportional to the level of coverage, then that deletion will be found to be correlated with high levels of viral transcription regardless of its actual biological impact.…”

“…This, in turn, can be correlated with viral infection, levels of viral transcripts, and, as discussed in the prior section, viral genetic variation to determine what features may lead to the induction of a response. Thus far, it has been illuminating that while there is a considerable contribution of viral variation to the induction of an innate immune response, even stimulatory variants within a population tend to induce a stochastic, if more common, production of interferons, indicating a role for chance events or preexisting host cell heterogeneity [ 34 , 65 ].…”