Increased lethality in influenza and SARS-CoV-2 coinfection is prevented by influenza immunity but not SARS-CoV-2 immunity

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the cause of the ongoing coronavirus disease 2019 (COVID-19) pandemic. The continued spread of SARS-CoV-2 increases the probability of influenza/SARS-CoV-2 coinfection, which may result in severe disease. In this study, we examine the disease outcome of influenza A virus (IAV) and SARS-CoV-2 coinfection in K18-hACE2 mice. Our data indicate enhance susceptibility of IAV-infected mice to developing severe disease upon coinfection with SARS-CoV-2 two days later. In contrast to nonfatal influenza and lower mortality rates due to SARS-CoV-2 alone, this coinfection results in severe morbidity and nearly complete mortality. Coinfection is associated with elevated influenza viral loads in respiratory organs. Remarkably, prior immunity to influenza, but not to SARS-CoV-2, prevents severe disease and mortality. This protection is antibody-dependent. These data experimentally support the necessity of seasonal influenza vaccination for reducing the risk of severe influenza/COVID-19 comorbidity during the COVID-19 pandemic.

T he pandemic of coronavirus disease 2019 (COVID- 19), which is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has posed serious threats to global health and economy. The COVID-19 pandemic presents a broad spectrum of severities, ranging from asymptomatic presentation to severe pneumonia. Most of the mildly to critically affected patients show respiratory complications, including moderate to severe pneumonia, which can further progress to acute respiratory distress syndrome, sepsis, and multiple organ dysfunction in severely ill patients 1 . Most of these clinical symptoms are associated with the respiratory system, specifically the lungs, resulting in the depleted lung functionality 2 . Several risk factors for severe COVID-19 disease, such as the age, sex, and obesity have been identified 3 . However, more research is needed to better understand the involvement of these and other risk factors in severe illness or complications. Whether coinfection with other pathogens may impact disease severity is yet to be elucidated. Previous studies have reported coinfections between SARS-CoV-2 and common respiratory viruses, including rhinovirus, influenza virus, metapneumovirus, parainfluenza virus, and respiratory syncytial virus 4 . A meta-analysis of 30 studies including 3834 patients with COVID-19 revealed that overall, 7% of hospitalized patients with COVID-19 had a bacterial coinfection, while the pooled proportion of viral coinfections was 3%, with respiratory syncytial virus and influenza A virus (IAV) being the most common 5 . However, the prevalence and disease outcomes of viral coinfections in the SARS-CoV-2-positive population remain controversial [6][7][8][9] .
IAV is a leading cause of respiratory infection, resulting in respiratory disease 10 . Complications involving secondary infections with other pathogens, mostly bacteria, significantly exacerbate the risk of severe disease due to IAV and other viral respiratory pathogens 11-14 . However, whether coinfection with IAV and SARS-CoV-2 presents more severe disease than a single infection is not clear.
In this study, we delineate the interplay between IAV and SARS-CoV-2 infections. For this purpose, we employ transgenic mice expressing human angiotensin-converting enzyme 2 (hACE2) under control of the human cytokeratin 18 promoter (K18-hACE2 mice) to establish a SARS-CoV-2-susceptible murine model 15 . We show that IAV-infected mice have enhanced susceptibility to developing severe disease upon coinfection with SARS-CoV-2 two days post influenza infection (dpIi). In contrast to nonfatal influenza disease and lower rates of mortality due to SARS-CoV-2 alone, this coinfection is associated with elevated influenza viral loads in respiratory organs, while SARS-CoV-2 viral load was reduced following coinfection. Remarkably, prior immunity to influenza, but not to SARS-CoV-2, prevents severe disease and mortality. Finally, we show that this protection is antibody dependent.

Results and discussion
Severe mortality in K18-hACE2 mice coinfected with SARS-CoV-2 and IAV. To delineate the interplay between IAV and SARS-CoV-2 infections, we used K18-hACE2 mice to establish a SARS-CoV-2-susceptible murine model 15 . Non/partially lethal viral doses were chosen for each virus to allow us to distinguish more severe outcomes of coinfection. The doses were determined based on challenge experiments in which K18-hACE2 mice were infected with different doses of mouse-adapted IAV (A/Puerto Rico/8/1934 H1N1 (PR8)) or SARS-CoV-2 ( Supplementary  Fig. 1). To mimic coinfection, the mice were infected with IAV and subsequently infected with SARS-CoV-2. While the terms 'coinfection' and 'superinfection' are often used interchangeably, the use of 'coinfection' here refers to a sequential infection with two viruses within a very short time, with the second infection occurring prior to the elimination of the first virus.
First, the outcome of SARS-CoV-2 infection was tested two dpIi during the presymptomatic stage of influenza (Supplementary Fig. 1a). At this stage, mice do not present any disease manifestations, but the viral titer of IAV in the lungs is high 16 . The IAV-infected mice started losing weight at 5 dpIi and exhibited maximal morbidity at 9-10 dpIi (75% of their initial body weight) (Fig. 1a). At 11 dpIi, all the mice began to gain weight and returned to their initial body weight by 18 dpIi. Remarkably, the mice infected with SARS-CoV-2 at two dpIi exhibited earlier and increased weight loss compared to the mice with IAV infection alone. Moreover, in contrast to the survival and only 38% mortality of the only IAV-and only SARS-CoV-2infected mice, respectively, all coinfected mice died by 5-7 days post SARS-CoV-2 infection (dpSi) (Fig. 1b, **** p < 0.0001).
Subsequently, we examined coinfection with SARS-CoV-2 at five dpIi during the early symptomatic stage. The IAV-infected and coinfected mice started to lose weight at 6-7 dpIi (Fig. 1c). However, while the IAV-infected mice reached their maximal weight loss at 8 dpIi (79% of their initial weight), the coinfected mice continued to lose weight until 10 dpIi (73% of their initial weight). Additionally, the recovery period of the coinfected mice was prolonged compared to that of the IAV-infected mice. While the IAV-infected mice reached 97% of their initial weight at 11 dpIi, the coinfected mice reached 94% of their initial weight by 17 dpIi and returned to their baseline weight only 22 dpIi (Fig. 1c). Moreover, coinfection with SARS-CoV-2 at 5 dpIi resulted in 70% mortality compared to the 43% mortality observed among the mice infected with SARS-CoV-2 alone (p = 0.08) (Fig. 1d). Finally, we assessed coinfection when SARS-CoV-2 was administered in the late-symptomatic stage of influenza disease during which maximal morbidity was detected (8 dpIi, Fig. 1e, f). Interestingly, SARS-CoV-2 infection at 8 dpIi had no effect on the body weight or survival rate of the mice. These results suggest that infection with SARS-CoV-2 early upon infection with IAV results in a more severe disease, while a later SARS-CoV-2 infection has no severe effect.
In the human population, coinfection is most likely to occur during the asymptomatic period. Additionally, during the asymptomatic stage of influenza disease, as represented by the experiments with SARS-CoV-2 infection at 2 dpIi (Fig. 1a, b), coinfection resulted in the most severe and lethal disease. Therefore, we chose to focus on this stage, and hereafter, coinfection refers to infection with IAV followed by infection with SARS-CoV-2 two days later.
Increased IAV and decreased SARS-CoV-2 viral loads in coinfected mice. To correlate the severe outcome observed in the coinfected mice and the viral load, the IAV and SARS-CoV-2 viral loads were quantified in the lungs, nasal turbinates (N.T.), and brains ( Fig. 2). A significant increase in the IAV RNA levels was observed in the lungs of the coinfected mice at 4 dpIi, but not at 6 dpIi, compared to those in the mice with IAV infection (4.6-fold increase). In the N.T., a significant increase in the IAV viral titer was observed at both 4 and 6 dpIi (11-fold and 2.3-fold increases, respectively) in the coinfected mice compared to that in IAV-infected mice (Fig. 2b, c), which coincided with the exacerbated disease. In contrast, the SARS-CoV-2 load was reduced in the coinfected mice compared to that in the SARS-CoV-2-infected mice in both the lungs and the N.T. (albeit nonsignificantly) at 4 and 6 dpIi (Fig. 2d, e). These findings are consistent with previous evidence of pathogenic competition between respiratory viruses, such as influenza and seasonal coronaviruses 17,18 . A possible explanation for the reduced SARS-CoV-2 viral load might be the induction of the innate immune response activated by the IAV infection prior to the SARS-CoV-2 infection, inhibiting the establishment of infection and replication [19][20][21] . In addition to innate immunity mechanisms underlying the reduced SARS-CoV-2 viral load, IAV infection may interfere with SARS-CoV-2 infection in vivo. Our data contradict those of a recent study showing that coinfection with IAV enhances SARS-CoV-2 infectivity 22 . In this previous study, no significant differences were found in disease manifestations between coinfected K18-hACE2 mice and SARS-CoV-2-monoinfected mice 22 . However, the SARS-CoV-2 infection dose in our study is 4-log lower than that in the Bai et al.'s study, and no data comparing coinfected mice and IAV-monoinfected mice were presented by Bai et al., making comparison between the two studies difficult.
Notably, although the SARS-CoV-2 viral load was reduced in the coinfected mice, the remaining levels were sufficient to trigger the lethal outcome of the coinfection. Histopathological analysis of coinfected K18-hACE2 mice. Lung pathology was observed in all infected groups and was consistent with the IAV viral load, with the coinfected mice presenting the most severe pathology (Fig. 2g); the lungs of the IAV-infected mice appeared moderately to strongly affected, showing necrosis of the bronchiolar epithelium with massive peribronchiolar infiltration of neutrophils and lymphocytes (Fig. 2g). The SARS-CoV-2-infected lungs showed milder histological changes characterized by minor-to-mild interstitial infiltration of lymphocytes and some neutrophils, with no obvious necrosis (Fig. 2g). The lungs of the coinfected mice showed a histopathological pattern similar to that observed in the IAV group but were more severe, including involvement of the lung parenchyma. Small hemorrhages were observed in the interstitium of some samples, and thickening of the pleura was identified in 3 of 5 animals (Fig. 2g). Taken together, these results ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26113-1 suggest that the increased morbidity and mortality detected in the coinfected mice are associated with elevated levels of IAV and exacerbated pathology in the respiratory system.
K18-hACE2 mice are a valuable model of SARS-CoV-2 infection and have been studied extensively 23,24 . However, in K18-hACE2 mice, SARS-CoV-2 also enters the brain, which might affect the disease severity 15 . To examine whether the increased morbidity was brain related, SARS-CoV-2 viral load was measured in brain tissues, and was found in only 3 of 10 SARS-CoV-2-infected mice ( Fig. 2f; at 4 dpSi). Importantly, no virus was detected in the brains of the coinfected mice, and no brain pathology was apparent at 4 dpSi ( Supplementary Fig. 2).
Coinfection in AdV-hACE2 mice model. Subsequently, to exclude the possibility that the severe outcome of the coinfected mice is model specific and brain related, we examined an additional independent COVID-19 mouse model, i.e., hACE2expressing human Ad5 (AdV-hACE2) mice. In the AdV-hACE2 model, C57BL/6J mice were inoculated via the intranasal (i.n.) route with replication-defective adenovirus encoding the hACE2 receptor, resulting in lung-specific pathology 25 . Similar to the K18-hACE2 mice, the AdV-hACE2 mice were significantly more sensitive to coinfection with SARS-CoV-2 at 2 dpIi (Supplementary Fig. 3). While all monoinfected mice survived the infection, only 9% of the coinfected mice survived (Supplementary Fig. 3). These results support that coinfection lethality is not model specific.
Upregulation of inflammation-related genes in the lungs of coinfected mice. To elaborate upon the host response to each monoinfection versus coinfection, we assessed the expression of immune-related genes in the lungs at 4 dpIi and 2 dpSi.
Overall, in the lungs of the SARS-CoV-2-infected mice, no alterations in the mRNA levels of the tested genes were observed compared to those in the lungs of the uninfected mice (Fig. 3) most likely due to the low infection dose (10 PFU/mouse) and short time post infection (2 days). After the infection with IAV, all tested genes were overexpressed (Fig. 3). Remarkably, the IAV and SARS-CoV-2 coinfection resulted in a significantly larger elevation of gene expression than exhibited upon IAV infection alone, indicating robust induction of the immune system that may have led to exacerbated disease.
Preexisting immunity to SARS-CoV-2 does not rescue coinfected mice. Then, we examined whether preexisting immunity to SARS-CoV-2 prevents the severe manifestations in the coinfected mice. Efficient immunity to SARS-CoV-2 was achieved by intramuscular (i.m.) immunization of SARS-CoV-2 (Supplementary Fig. 4a, b). This route of immunization induced both humoral and cellular responses ( Supplementary Fig. 4a, b, respectively) against SARS-CoV-2 and was sufficient to protect the mice from a SARS-CoV-2 challenge (Fig. 4a-c). However, while preexisting immunity to SARS-CoV-2 completely prevented the mortality caused by SARS-CoV-2 infection, accounting for 50%, it had no effect on the morbidity and full mortality caused by the IAV and SARS-CoV-2 coinfection (Fig. 4b, c). To rule out the possibility that the inability of SARS-CoV-2 immunization to rescue the coinfected mice was a result of the i.m. immunization route, we examined another SARS-CoV-2 route of immunization. The acquisition of preexisting immunity to SARS-CoV-2 by the i.n. route ( Supplementary Fig. 4c, d) did not protect the mice from coinfection ( Supplementary Fig. 5), which is similar to the results of infection via the i.m. route, even though both routes of infection induced humoral and cellular responses to SARS-CoV-2 ( Supplementary Fig. 4).
To determine whether preexisting immunity to IAV can prevent coinfection manifestations, mice were vaccinated i.m. with IAV 30 days prior to the viral infection with IAV and/or SARS-CoV-2. This preexposure to IAV induced both humoral and cellular responses ( Supplementary Fig. 6a, b, respectively) against IAV and alleviated the morbidity observed upon infection with IAV (Fig. 4f). Preexposure to IAV had no effect on the survival rate upon SARS-CoV-2 infection (Fig. 4e, p = 0.64). Remarkably, immunity to IAV prevented the severe manifestations and mortality caused by the IAV and SARS-CoV-2 coinfection. Neither weight loss nor increased mortality was detected in the coinfected mice vaccinated against IAV compared to those in the coinfected mice without preexisting immunity (Fig. 4e, f).
These data further support the notion that the severe manifestations of coinfection are not the result of a more severe SARS-CoV-2 disease but rather of IAV.
Prior immunity to IAV rescues coinfected mice and is antibody dependent. Next, we examined whether the protection of coinfected mice by prior immunization is due to cellular or humoral immunity. First, T-cell depletion was conducted (Supplementary Fig. 7). Mice were immunized by the i.m. route to IAV to induce immunity; after 30 days, CD4 + or CD8 + T cells were depleted, the mice were coinfected with IAV and SARS-CoV-2, and the ability of preexisting immunity to rescue the mice in the presence or absence of T cells was tested (Fig. 4g, h). The depletion of CD4 + or CD8 + T cells had no effect on the survival rate of the preimmune coinfected mice. This result suggests that the protection provided by preimmunity to IAV is T cell independent. Next, to delineate the role of anti-IAV Fig. 2 Increased IAV and decreased SARS-CoV-2 loads in coinfected mice. Schematic timeline (a). K18-hACE2 mice were sacrificed at 4 or 6 dpIi and 2 or 4 dpSi. The IAV RNA load was measured in the lungs (b) and nasal turbinates (N.T.) (c), and the PFU equivalent per organ (pfuE/organ) was calculated. The SARS-CoV-2 loads in the lungs (d), N.T. (e), and brains (f) were determined by a PFU assay. Each symbol represents one mouse (10 mice per group, except 4 mice for brain testing at 4 dpIi). The lines represent the means. The error bars represent the SEs. **p = 0.0062 (b); ***p = 0.0008, *p = 0.0411 (c); *p = 0.0212, **p = 0.0056 (d); ns, not significant; two-tailed unpaired Student's t-test. Hematoxylin and eosin (H&E) staining (g) of lung sections from naive K18-hACE2 mice or after i.n. administration of SARS-CoV-2 (4 dpSi), IAV (6 dpIi  antibodies in the protection of coinfected mice, we performed passive transfer of anti-IAV sera to coinfected mice. Anti-IAV antibodies rescued the coinfected mice (Fig. 4i), indicating that the protection was indeed antibody dependent.
Altogether, our data show that IAV-SARS-CoV-2 coinfection results in severe and lethal disease in mice. The severe manifestations were associated with an elevated IAV viral load, robust induction of innate immunity, and exacerbated pathology in respiratory organs. Importantly, prior humoral immunity to influenza, but not to SARS-CoV-2, prevented disease and mortality, suggesting that influenza intervention by prior vaccination may prove valuable in reducing the risk of severe influenza/SARS-CoV-2 comorbidity. For the infection with Ad-GFP-hACE2 25 , C57BL/6J mice were injected intraperitoneally (i.p.) with 2 mg of an anti-IFNAR (interferon-α/β receptor) monoclonal antibody (mAb) (LEINCO Technologies, Inc., St. Louis, MO, USA, code I-401-100) in 0.5 ml. After 24 h, 20 µl of Ad-GFP-hACE2 were administered to the anesthetized mice. After 3 days, the mice were treated i.n. with 80 PFU of PR8, and after 2 days, the mice were treated i.n. with 10 5 PFU of SARS-CoV-2 (all i.n. administrations were performed under anesthesia).
Determination of the viral load in organs. The viral loads were determined at 2 and 4 dpSi or 4 and 6 dpIi. In the coinfected group, the viral load was determined at 4 and 6 dpIi (2 and 4 dpSi). Each group included 10 mice, except for the experiments examining brains at 2 dpSi, which included a group of 4 K18-hACE2 animals. The lungs, N.T., and brains were harvested and stored at −80°C until further processing. The organs were processed for titration in 1.5 ml of ice-cold PBS. Tissues were homogenized (ULTRA-TURAX® IKA R104) for 30 s in 1.5 ml of ice-cold PBS, followed centrifugation (270g, 10 min, 4°C) and collection of the supernatants for virus titration. Processed tissue homogenates were divided for immediate RNA extraction for the IAV RNA determination and gene expression analysis or kept at −80°C until further processing for the viral titration (used for SARS-CoV-2 PFU assay).
The SARS-CoV-2 viral load was determined using a PFU assay 27 . Serial dilutions of extracted organ homogenates from mice infected with SARS-CoV-2 or coinfected with IAV and SARS-CoV-2 were prepared in infection medium (MEM containing 2% FBS) and used to infect Vero E6 monolayers in duplicate (200 µl/well). The plates were incubated for 1 h at 37°C to allow viral adsorption. Then, 2 ml overlay (MEM containing 2% FBS and 0.4% tragacanth; Merck, Israel) were added to each well, and the plates were incubated at 37°C in a 5% CO 2 atmosphere for 48 h. The medium was then aspirated, and the cells were fixed and stained with 1 ml/well crystal violet solution (Biological Industries, Israel). The number of plaques per well was determined, and the SARS-CoV-2 PFU titer was calculated.
The IAV RNA level was determined using real-time RT-PCR (see below). Total RNA extracted from the lungs of mice infected with IAV or SARS-CoV-2 or coinfected at 2 dpSi and 4 dpIi was used to measure the differential expression of genes by qRT-PCR using corresponding specific primers printed on 96-well plates (Custom TaqMan Array Plates, Applied Biosystems TM ) as previously described 28 . Briefly, 1 µg of cDNA was synthesized from RNA using a Verso cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. The samples were subjected to qPCR with TaqMan® Fast Advanced Master Mix (7500 Real-Time PCR System, Applied Biosystems, Thermo Fisher Scientific). The housekeeping gene GAPDH was used to normalize the fold change of each gene compared to the mock-infected control at the same time point, which was calculated as ΔΔCT.
Histopathology. For the hematoxylin and eosin (H&E) general histopathology evaluation, 6 dpIi (4 dpSi) mice (n = 5 per group) were anesthetized and then perfused transcardially with PBS, followed by 4% paraformaldehyde. The lungs and brains were isolated and fixed in 4% paraformaldehyde at room temperature for 2 weeks. The fixed tissues were sent to 'Patho-Logica' (Rehovot, Israel) for processing and analysis. Serial coronal 4 µm thick sections were obtained, and selected sections were stained with H&E for light microscopy examination. Pictures were taken using an Olympus microscope (BX60, serial No. 7D04032) equipped with a microscope camera (Olympus DP73, serial No. OH05504) at objective magnifications of X4 and X10.
Plaque reduction neutralization test. The SARS-CoV-2-neutralizing antibody levels were determined using a plaque reduction neutralization test (PRNT) 27 . Briefly, all sera were heat inactivated at 60°C for 30 min, then serially diluted twofold in 400 µl of infection medium, mixed with 400 µl of 300 PFU/ml SARS-CoV-2, and incubated at 37°C in an atmosphere of 5% CO 2 for 1 h. Then, 200 μl of each serum/virus mixture was added in duplicate to Vero E6 cell monolayers, and the cells were incubated for 1 h at 37°C. A virus mixture without serum was used as a control. Two milliliters overlay were added to each well, and the plates were incubated at 37°C in a 5% CO 2 atmosphere for 48 h. The medium was then aspirated, and the cells were fixed and stained with 1 ml/well crystal violet solution. The number of plaques per well was determined, and the serum dilution that neutralized 50% of the virions (NT 50 ) was calculated using Prism software (GraphPad Software Inc.).
filtered, separated on Ficoll-Paque (GE), and washed with medium. Then, 4 × 10 5 cells from each sample were plated into 96-well ELISpot plates in duplicate in the presence of peptides representing H-2Kb immunodominant epitopes at a final concentration of 2 μg/ml and incubated at 37°C for 24 h. The naive samples consisted of pools from 2 mice. The frequency of IFNγ-secreting cells was determined using a Mouse IFN-γ Single-Color ELISpot kit (Cellular Technology Limited, Germany) according to the manufacturer's instructions. The frequency of cytokine-secreting cells was quantified with an ImmunoSpot S6 Ultimate reader and analyzed with ImmunoSpot software (Cellular Technology Limited, Germany). The following Db-restricted antigens were used for the stimulation: influenza PR8-derived peptide: NP366-374 ASNENMETM SARS-CoV-2-derived peptide: Spike 539-546 VNFNFNGL For all assays, antigen-free cells supplemented with medium were used as a negative control.
Passive immunization. Six to eight week-old female K18-hACE2 mice were immunized against IAV by i.m. injection of PR8 (10 6 PFU/mouse). Four weeks later, sera were collected and pooled, and the levels of anti-PR8 IgG antibodies in the anti-IAV sera were determined. The sera were administered (i.p. 200 µl/mouse) 1 day prior to infection with IAV. Mice were coinfected as described above.