Defective Viral Genomes as Drivers of Cell Fate During the Virus-Host Interaction
Despite extensive progress in the identification of host immune pathways and viral molecular motifs involved in recognition and clearance of viruses, a critical gap of knowledge remains: How is viral control achieved during infections with viruses that effectively counteract initial recognition by the host immune system? Since most clinically relevant viruses evade the immune response long enough to replicate and spread, we propose that virus clearance does not depend on the detection of replicating viral genomes, as predicted based on current paradigms. Rather, it depends on sensing viral products generated once viral replication has been established and jeopardizes host survival (Figure 1). Full understanding of the requirements for immune-mediated viral clearance requires unraveling these largely uncharacterized interactions among diverse viral replication products and the host.
Figure 1. Current and proposed models for detection of viruses that evade immune recognition. Current models predict that antiviral immunity (represented as interferon (IFN) a/b) begins when specialized cellular proteins (RIG-I/MDA5 or others) bind danger signals that originate from the viral genome. This initial response is maximized via positive feedback by IFNs. However, many viruses block the production and signaling of IFNs through virus encoded proteins (Viral Antagonists). Our model proposes that distinct viral products produced during viral replication at high titers provide potent danger signals that stimulate viral sensors, thereby overcoming viral antagonism, and is independent of type I IFN feedback. Supporting evidence is found in references1-5.
Virus replication generates multiple products in addition to full-length standard viruses (Figure 2). These products include a variety of slightly mutated versions of the virus (quasispecies) that enhance its fitness, and a much less understood set of severely truncated genomic products collectively named defective viral genomes (DVGs) for their inability to replicate without a helper virus. Although most viruses generate a discrete population of DVGs, they have been considered irrelevant byproducts of virus replication and their biological role largely overlooked. However, in published and unpublished data, we have shown that DVGs serve as de facto danger signals that trigger immunity in infections with a number of paramyxoviruses in mice and humans and only cells with high numbers of DVGs activate antiviral responses (Figure 3). These data position DVGs as pivotal determinants of infection outcome.
Figure 2. Viral replication products. At least three major viral products are produced during virus replication. Full-length viruses, viral quasispecies that contain small mutations that allow for the selection of the fittest, and defective viral particles containing severely truncated genomes (DVGs). DVGs are unable to replicate in the absence of helper virus, but retain packaging signals. The biological role of DVGs is largely unknown.
Figure 3. DVGs stimulate the antiviral response during infection. RNA FISH-IFA staining of DVG (green), Sendai virus full-length genome (gSeV) and viral mRNA (orange and red), nuclei (DAPI; blue) (top), and IRF3 (purple)(bottom) in cells infected with SeV. Intranuclear localization of IR3 indicates activation of antiviral responses. This is only evident in cells enriched in DVGs (white arrow) (100 X).
One aspect of our research focuses on establishing the impact of DVGs on paramyxovirus maintenance and host survival. This work will elucidate the cellular and molecular basis of host immune responses triggered by DVGs and allow us to harness them as markers of disease prognosis and vaccine adjuvants. We can frame this aspect of our work into four main areas:
What is the molecular basis for the recognition of DVGs as danger signals that trigger immunity?
It has been proposed that either high viral loads are needed to achieve the minimal threshold for triggering of host immune responses, or that cell damage or death are pre-requisites for immune stimulation. However, our data strongly support a paradigm-shifting model in which DVGs generated after the virus reaches high titers are the primary inducers of host immune responses, even in the presence of viral-encoded antagonists (Figure 1). In agreement with our model, the absence of DVGs greatly compromises antiviral responses to Sendai virus, respiratory syncytial virus, and influenza virus even in infections with large doses of virus. A critical question arises: what cellular and molecular events allow for recognition and response to DVGs in the presence of strong viral antagonists of immune detection? Answering this question will reveal the cellular and molecular basis for maximal host responses to virus infection and highlight potentially novel points of intervention with virus replication.
A trivial explanation for the immunostimulatory activity of DVGs is that they accumulate to higher levels than full-length genomes in infected cells providing more viral danger signals. However, viral danger signals levels per se cannot explain why infections with high doses of standard virus fail to trigger immune responses. We are testing three potential mechanisms for DVG recognition that are not mutually exclusive: 1. Unique molecular motifs facilitate the binding of viral RNA by immune sensors, 2. Viral detection requires localization to specific intracellular compartments devoid of viral antagonists, 3. Unknown cellular pathways are engaged to modulate viral sensing positively.
Preliminary evidence suggests that a combination of these mechanisms is required for maximal viral sensing. We discovered a unique molecular motif present in DVGs, but not in the full-length viral RNA, that complements other viral associated molecular patterns and is required for maximal activation of host antiviral responses (Figure 4). In addition, single cell level analysis revealed a distinct intracellular distribution of full-length genomes and DVG throughout the infection suggesting a critical intracellular spatial component of viral sensing that is largely not understood. We are interested in characterizing the cellular processes leading to this distribution and how it relates to virus detection by the host.
Figure 4. A unique stem loop motif confers strong stimulatory activity to a SeV DVG. (A) Schematics of the full Sendai virus genome and its major stimulatory DVG. Predicted essential motifs for immunostimulation include 5’-triphosphates and 5’-3’ complementary structures found in DVGs but not in standard genomes (grey boxes). Blue and red squares indicate regions of interest. (B) Structural prediction for DVG lacking the 3’ complementary region denoting candidate sequences marked in A. (C) IFNB1 expression upon transfection of in vitro transcribed RNA lacking motifs 5-51 and 70-114. These mutants preserved their overall predicted structure. Motif 70-114 is essential for the stimulatory activity of this DVG independently of the 3’ complementary end and can be transferred to inert RNA conferring immunostimulatory potential.
Can we harness DVGs as biomarkers of infection outcome?
We found that roughly half of the nasopharyngeal samples (20/41) from children infected with respiratory syncytial virus tested positive for DVGs, and there was a strong correlation between the presence of DVGs and evidence of active antiviral immunity (Figure 5). Based on this, we propose that DVGs impact viral disease outcome in humans and that their presence in acute respiratory infections predicts viral clearance and full disease recovery, while their absence suggests poor engagement of host immune responses and the need for more aggressive treatments. Current projects focus on addressing the follow questions: 1. Do DVGs determine the quality of immune responses in humans? 2. How does the host response relate to infection outcome? 3. Is the response to DVGs affected by co-morbidities, co-infections, or other host conditions? Our goal is to establish a framework for testing DVGs as biomarkers of infection outcome and to ultimately use this information to optimize management of viral diseases.
Figure 5. DVGs are detected in patients infected with respiratory syncytial virus. (A) PCR assay specifically detects DVGs in 50% of nasopharyngeal aspirates from infected patients (20/41). One negative and one positive patient are shown. Patients analyzed in this cohort contained equivalent levels of standard RSV (gRSV). (B) Presence of DVGs correlates with expression of antiviral genes detected in human samples.