The Host Response to Influenza A Virus Interferes with SARS-CoV-2 Replication during Coinfection
In golden hamsters, influenza A virus infection reduces SARS-CoV-2 replication during coinfection, after preinfection, and even one to two weeks after influenza has been cleared, while SARS-CoV-2 leaves influenza replication in vivo unchanged.
- Senior authors
- Benjamin R. tenOever
- Correspondence
- Benjamin R. tenOever
Research areas & themes
Citation
Oishi K, Horiuchi S, Minkoff JM, tenOever BR. The host response to influenza A virus interferes with SARS-CoV-2 replication during coinfection. Journal of Virology. 2022. Volume 96, issue 15, article e00765-22.
DOI 10.1128/jvi.00765-22. PMID 35862681. PMCID PMC9364782.
One-sentence contribution
In golden hamsters, influenza A virus infection reduces SARS-CoV-2 replication during coinfection, after preinfection, and even one to two weeks after influenza has been cleared, while SARS-CoV-2 leaves influenza replication in vivo unchanged.
Executive summary
SARS-CoV-2 and influenza A virus both spread by the airborne route, both infect the human airway, and both were documented as coinfections early in the COVID-19 pandemic. Whether coinfection makes disease worse was unsettled, with published animal studies reaching different conclusions depending on model, strain and sex. The authors asked how the two viruses interact when a single host carries both. In Vero E6 cells, which lack the type I and type III interferon genes, SARS-CoV-2 was modestly delayed by influenza at twenty-four hours and influenza was unaffected. In interferon-competent A549 cells expressing ACE2, SARS-CoV-2 was again delayed while influenza titres fell substantially at every time point. In golden hamsters infected intranasally with either virus alone or with both simultaneously, influenza replication was unchanged by SARS-CoV-2 at every time point, while SARS-CoV-2 titres were lower at three days and cleared from the lung by five days in coinfected animals rather than persisting to five and seven days as in single infection. Weight and lung histology tracked with SARS-CoV-2 rather than with coinfection as such. Preinfection with influenza three days ahead reduced subsequent SARS-CoV-2 replication, whereas preinfection with SARS-CoV-2 did not change influenza. The suppression persisted when SARS-CoV-2 was given seven or fourteen days after influenza, at which point only twenty-five genes remained differentially expressed although ISG15 and IRF7 remained elevated. The authors read this as immune priming rather than direct competition.
Scientific context
Coinfection with SARS-CoV-2 and other respiratory viruses was reported repeatedly early in the pandemic, with a meta-analysis of the first four months finding viral coinfection in about three percent of hospitalised COVID-19 patients and influenza A virus among the most common partners. More than half of the roughly thirty SARS-CoV-2 proteins had been reported by others to interfere with interferon induction or signalling, which raised the concern that a SARS-CoV-2-infected airway might be unusually permissive to a second pathogen. The two viruses use different entry receptors, ACE2 for SARS-CoV-2 spike and sialic acid alpha 2,6-galactose for human influenza haemagglutinin, but both are reported to infect alveolar type II cells. Prior experimental work by other groups had produced divergent results. In K18-hACE2 mice, coinfection increased disease severity and both preinfection and coinfection with influenza raised SARS-CoV-2 loads with delayed clearance. In golden hamsters, one study found increased severity with lower SARS-CoV-2 lung loads but prolonged shedding, and another using an H3N2 strain in female hamsters found robust SARS-CoV-2 replication alongside low influenza titres, with Mx1 implicated in inhibition of influenza. The present study enters this disagreement with a single model, matched strains, and both simultaneous and sequential designs.
Central question
When one host carries both SARS-CoV-2 and influenza A virus, does either virus interfere with the replication of the other, and does the direction and timing of that interference depend on the host interferon response rather than on direct competition for cells?
Experimental strategy
The design isolates the contribution of the interferon system twice over, once in cell culture and once in time. In culture, the same coinfection is run in Vero E6 cells, which lack the type I and type III interferon genes, and in A549 cells engineered to express ACE2, which are interferon competent and support both viruses, so a phenotype present only in the latter can be attributed to the host response rather than to receptor competition or cell death. In vivo, the golden hamster is used because it supports robust replication of both viruses in the respiratory tract and reproduces pathological and immunological features of human disease. Three temporal arrangements are then compared. Simultaneous coinfection asks whether the two viruses interfere while both are replicating. Preinfection at three days in each direction asks whether an established infection changes what follows. Challenge at seven and fourteen days after influenza, when no infectious influenza is recoverable, separates any effect of the ongoing infection from a residual change in the airway. Readouts combine plaque titration of each virus on its own indicator cell line, body weight, lung histology, messenger RNA sequencing of lung at five days after coinfection and at seven days after influenza alone, and quantitative RT-PCR for ISG15 and IRF7 as markers of a persisting interferon-stimulated state.
Key findings
-
In Vero E6 cells at low multiplicity, SARS-CoV-2 titres were lower in the presence of influenza at twenty-four hours but comparable at forty-eight and seventy-two hours, while influenza titres were unchanged at all time points (Figures 1A and 1B).
-
In A549-ACE2 cells, SARS-CoV-2 again showed reduced titre at twenty-four hours and recovery by forty-eight hours, while influenza titres were substantially lower in the presence of SARS-CoV-2 at every time point (Figures 1C and 1D). Because the two cell lines differ in interferon competence, the authors read the influenza phenotype as host-response dependent.
-
In simultaneously coinfected hamsters, SARS-CoV-2 lung titres matched single infection at one day, were lower at three days, and were cleared by five days, whereas infectious SARS-CoV-2 was recovered from all singly infected animals at five days and from a quarter at seven days (Figure 2B). Influenza titres were unaffected by coinfection at every time point (Figure 2C).
-
Weight change and lung histology followed SARS-CoV-2 rather than coinfection. Coinfected animals and those given SARS-CoV-2 alone showed a delayed lack of weight gain relative to control and influenza-only animals, and both showed extensive mononuclear infiltration on haematoxylin and eosin staining at three days (Figures 2D and 2E).
-
Messenger RNA sequencing of lung at five days showed that coinfection produced a more pronounced inflammatory response than influenza alone, and that the genes induced by SARS-CoV-2 or by coinfection included those induced by influenza plus additional ones, with the SARS-CoV-2-associated set enriched for NF-kappaB-dependent chemokines including CCL2, CCL5, CCL7 and CXCL10 (Figures 2F and 2G, Table 1).
-
Preinfection with influenza three days before SARS-CoV-2 reduced SARS-CoV-2 titres at every time point tested, while influenza titres were unchanged by SARS-CoV-2 given during active influenza replication (Figures 3B and 3C).
-
Preinfection with SARS-CoV-2 three days before influenza did not affect influenza replication, and SARS-CoV-2 titres were also unchanged by the later influenza challenge (Figures 3E and 3F). No delay in clearance of either virus was observed in either sequential arrangement.
-
Challenging influenza-recovered hamsters with SARS-CoV-2 seven days after influenza, when infectious influenza was no longer detectable, gave lower SARS-CoV-2 titres at one and three days, with titres converging at five days (Figures 4A and 4B). At fourteen days after influenza, SARS-CoV-2 titres were still lower at one day (Figures 4C and 4D).
-
Sequencing of lung at seven days after influenza alone found only twenty-five differentially expressed genes relative to control, which the authors take as indicating a largely restored baseline. Among the modestly enriched transcripts they note markers the paper associates with immune priming and with increased macrophage presence, including C1QA, C1QB and C1QC (Table 2).
-
By quantitative RT-PCR, ISG15 and IRF7 were elevated in lung at seven days after influenza, with ISG15 still elevated at fourteen days (Figures 4E and 4F). The authors read this as a residual interferon-stimulated state, or alternatively as increased resident immune cells in the airway, and explicitly present these as two possibilities rather than one conclusion.
Mechanistic model
The study does not establish a definitive mechanism, and the authors state directly that it remains unclear whether the inhibition of SARS-CoV-2 reflects direct effects within coinfected cells or bystander effects on neighbouring cells, and that resolving this would require single-cell approaches not attempted here. What the data constrain is the direction, the durability, and the plausible source of the effect. Because suppression of early SARS-CoV-2 replication persists a week or two after influenza has been cleared, when almost no transcriptional difference remains detectable in bulk lung, direct competition for target cells is not a sufficient explanation, and the authors describe the phenotype as likely a product of immune priming with type I and type III interferon. They note two candidate substrates for that priming, namely residual low-level interferon-stimulated gene expression, supported by the elevated ISG15 and IRF7, and an increased population of resident immune cells, supported by the complement component enrichment. They do not distinguish these. They also invoke kinetics, citing earlier work from the group in the same model showing that interferon-stimulated gene and chemokine peaks occur around three days for influenza and later and more persistently, from five to seven days, for SARS-CoV-2, and propose that this offset contributes to asymmetric interference. The failure of SARS-CoV-2 to suppress influenza in hamsters despite Mx1 induction is interpreted as influenza being able to replicate in an interferon-primed lung given enough susceptible cells, and the authors note that their own earlier work found Mx1 induction by SARS-CoV-2 to be delayed in the airway and more evident in tissues that are not productively infected. The in vitro suppression of influenza by SARS-CoV-2 in interferon-competent cells but not in hamsters is attributed to the lung offering an expansive and heterogeneous cellular landscape not reproduced in culture, which is interpretation rather than demonstration.
Conceptual or technical advance
The work supplies a controlled comparison in one model of all three timings that matter epidemiologically, namely simultaneous infection and each virus preceding the other, and adds the recovered-host arm that separates an active infection from its aftermath. That last arm is what makes the argument, because it shows an effect on SARS-CoV-2 replication persisting into a window where the airway looks transcriptionally almost normal. Practically, the finding that coinfected animals show a host response comparable to SARS-CoV-2 alone, with no prolonged shedding and no delayed clearance, gives a reason to doubt that cocirculation of the two viruses would by itself increase disease severity, a conclusion the authors state while also noting that whether the interference affects severity remains unclear. The study also documents an asymmetry that makes the interferon-competence of the experimental system a variable rather than a detail, since the suppression of influenza by SARS-CoV-2 appears in interferon-competent cells, not in interferon-deficient cells, and not in the animal.
Relationship to the broader research program
The paper sits within the laboratory's use of the golden hamster as a model for comparing host responses to respiratory viruses, and it relies directly on earlier work from the group in the same model with the same virus strains for the kinetics of interferon-stimulated gene and chemokine induction. It also builds on the group's earlier finding that SARS-CoV-2 induces a more pronounced inflammatory signature than influenza in hamsters, an observation reported as consistent with patient data and with the ferret model. Marked as category 3 synthesis, a recurring thread across the corpus is the treatment of the host response, rather than the virus, as the variable that determines outcome, whether that is defective genome production driving interferon in the nucleoprotein work or interferon priming setting the ceiling on a second infection here. A second thread is the pairing of interferon-deficient and interferon-competent systems as a routine control, used in this paper as Vero E6 against A549-ACE2.
Related publications
- Horiuchi et al., predecessor. Same laboratory, cited for the hamster kinetics of interferon-stimulated gene and chemokine induction in response to the same SARS-CoV-2 and influenza strains, and shares an author with the present paper.
- Blanco-Melo et al. 2020 and related work on the SARS-CoV-2 host response, predecessor. Same laboratory, cited for the finding that SARS-CoV-2 induces a stronger inflammatory signature than influenza and for Mx1 induction patterns across tissues.
- Published golden hamster and K18-hACE2 mouse coinfection studies from other groups, companion. Cited and contrasted throughout the discussion as producing partly conflicting results attributed to model, strain and sex, and explicitly attributed by the authors to those other laboratories.
Limitations and boundaries
The animal work uses male Syrian golden hamsters only, at six to eight weeks of age, which matters because the authors themselves note that another group's divergent result came from female hamsters and that sex, age and strain may influence the outcome. One SARS-CoV-2 isolate and one H1N1 strain were used, at fixed doses delivered intranasally, so dose and strain dependence is untested. Group sizes are small, with three or four animals per time point for titration and eight per condition for weight. Titres are from lung homogenate, with no assessment of the upper respiratory tract, no transmission or shedding measurements, and no barcoded viruses to resolve whether interference occurs within the same cells, an experiment the authors name as needed. The mechanistic claim rests on an association between residual ISG15 and IRF7 expression and reduced early SARS-CoV-2 replication, with no interferon blockade, receptor knockout, or cell depletion to test causation, and the authors offer immune priming and increased resident immune cells as alternatives they cannot separate. Cell marker antibodies for hamster are limited, which the authors give as the reason cellular resolution was not attempted. Transcriptomics is bulk lung at two time points. The in vitro and in vivo results disagree on whether SARS-CoV-2 suppresses influenza, and the explanation offered for that disagreement is interpretive. Finally, the conclusion that cocirculation is unlikely to increase disease severity is drawn from an animal model and does not rest on clinical data.
Audience summaries
25 words
In hamsters, influenza infection holds back SARS-CoV-2 replication during coinfection and for up to two weeks afterwards, while SARS-CoV-2 does nothing to influenza in the animal.
75 words
Golden hamsters were given SARS-CoV-2, influenza A virus, or both, either together or three, seven or fourteen days apart. Influenza consistently reduced SARS-CoV-2 replication and hastened its clearance, including when given a week or two before, when influenza itself was gone but interferon-stimulated genes remained elevated. SARS-CoV-2 never changed influenza titres in the animal, although it did suppress influenza in interferon-competent cells. Coinfected animals looked like SARS-CoV-2 infection alone by weight and histology.
150 words
Oishi and colleagues compared SARS-CoV-2 and influenza A virus H1N1 coinfection in interferon-deficient Vero E6 cells, in interferon-competent A549-ACE2 cells, and in golden hamsters. In culture, influenza delayed SARS-CoV-2 in both cell lines, while SARS-CoV-2 suppressed influenza only in the interferon-competent line. In hamsters, simultaneous coinfection lowered SARS-CoV-2 titres at three days and cleared the virus by five days without altering influenza at any point. Influenza given three days earlier reduced subsequent SARS-CoV-2 replication, and the reverse arrangement had no effect on influenza. The suppression persisted when SARS-CoV-2 was given seven or fourteen days after influenza, a point at which only twenty-five lung genes were differentially expressed although ISG15 and IRF7 remained elevated. The authors attribute the effect to interferon-mediated immune priming rather than direct competition, state that they cannot distinguish direct from bystander effects, and conclude that cocirculation is unlikely to raise disease severity.