RIPK3 interacts with MAVS to regulate type I IFN-mediated immunity to Influenza A virus infection

The type I interferon pathway plays a critical role in both host defense and tolerance against viral infection and thus requires refined regulatory mechanisms. RIPK3-mediated necroptosis has been shown to be involved in anti-viral immunity. However, the exact role of RIPK3 in immunity to Influenza A Virus (IAV) is poorly understood. In line with others, we, herein, show that Ripk3-/- mice are highly susceptible to IAV infection, exhibiting elevated pulmonary viral load and heightened morbidity and mortality. Unexpectedly, this susceptibility was linked to an inability of RIKP3-deficient macrophages (Mφ) to produce type I IFN in the lungs of infected mice. In Mφ infected with IAV in vitro, we found that RIPK3 regulates type I IFN both transcriptionally, by interacting with MAVS and limiting RIPK1 interaction with MAVS, and post-transcriptionally, by activating protein kinase R (PKR)—a critical regulator of IFN-β mRNA stability. Collectively, our findings indicate a novel role for RIPK3 in regulating Mφ-mediated type I IFN anti-viral immunity, independent of its conventional role in necroptosis.


Author summary
Influenza A virus (IAV) is a pulmonary pathogen that presents a significant threat to human health through seasonal epidemics and occasional, highly lethal pandemics. Type I IFN is an essential component of anti-viral immunity to influenza infection and pulmonary macrophages are the major source of type I IFN. Recently, we have shown that programmed macrophage cell death (apoptosis) plays a key role in immunity to influenza infection. Interestingly, another cell death program, termed necroptosis, has been also implicated in control of viral infection via receptor-interacting serine/threonine protein kinase 3 (RIPK3). In the present study, we define a novel role of RIPK3 in regulating the type I IFN pathway to protect against lethal IAV infection, which is independent of necroptosis. RIPK3 regulates type I IFN signaling at both the transcriptional and post-

Introduction
Pulmonary macrophages (Mφ) reside in the unique extraepithelial environment of the lower airways and are the main source of one of the key components of host anti-viral immunity: type I IFN (primarily IFN-α and β) [1][2][3]. However, many highly pathogenic viruses, including IAV, have evolved to reach the lower respiratory tract and effectively sidestep the type I IFN pathway in Mφ. Initial recognition of IAV-ssRNA occurs by the cytosolic RNA helicase retinoic acid-inducible gene I (RIG-I) that interacts with the mitochondrial anti-viral-signaling protein (MAVS) to activate the interferon regulatory factor 3 (IRF3)-mediated type I IFN pathway, upstream of the TANK-binding kinase 1 (TBK1) [4]. Subsequent binding of type I IFNs to their heterodimeric receptor (IFNαR) leads to activation of the JAK/STAT pathway and the transcription of IFN-inducible genes (ISGs), such as the double-stranded RNA-dependent protein kinase R (PKR), which is critical in controlling viral replication, by regulating proteins involved in inhibiting both host and viral translation [5] as well as IFN-β mRNA integrity [6]. Resident alveolar Mφ are the first immune cells to encounter IAV in the airways and orchestrate the immune response [7]. While both the frequency and number of resident alveolar Mφ are constant shortly after infection [8], the frequency and total cell number of bone marrow derived-monocytes, recruited in a CCR2-dependent manner, are significantly increased and represent the major source of Mφ in the lungs during IAV infection [9,10]. We, and others, have shown that the induction of type I IFN by pulmonary Mφ is indispensable during IAV infection [1,2,10]. Thus, it is not surprising that IAV has evolved several strategies to inhibit the type I IFN axis, including encoding the virulence factor PB1-F2, which specifically targets mitochondria to induce early apoptosis in Mφ [11,12] to limit the production of type I IFN [13]. Interestingly, the receptor interacting serine/threonine protein kinase (RIPK) family members (RIPK1 and RIPK3) regulate necroptosis (a form of programmed necrosis) and play a critical role in immunity to viral infections. For example, RIPK3-mediated necroptosis was shown to be important in the host defense against vaccinia virus [14], murine cytomegalovirus (MCMV) [15], as well as IAV [16]. Additionally, a cell death-independent role for RIPK1 and RIPK3 in inflammation has also been described in myeloid cells. In models of LPSinduced inflammation using bone marrow-derived Mφ (BMD-Mφ) [17], or bone marrowderived dendritic cells (DC) [18], RIPK3-deficient cells failed to release pro-inflammatory cytokines [19,20]. Consistent with these studies, it has been also demonstrated that RIPK1 regulates the production of potent inflammatory cytokines, including TNF-α [21]. Importantly, it has recently been shown that RIPK3 confers enhanced viral clearance and protection to IAV by modulating apoptotic and necroptotic cell death in infected lung structural cells [16], while its expression moderately affects the pro-inflammatory and anti-viral signature of fibroblasts [22]. However, the function of RIPK3 in lung immune cells, which contribute significantly to immunity to IAV infection has not been well understood.
In this report, we sought to further delineate the role of RIPK3 in immunity to pulmonary IAV infection. Herein, we define RIPK3 as an essential component of host defense against IAV infection. Surprisingly, RIPK3-deficient mice were extremely susceptible to IAV infection due to a significant reduction in type I IFN. Pulmonary Mφ from RIPK3-deficent mice failed to mount an effective type I IFN response to IAV. Mechanistically, we demonstrated that RIPK3 was upregulated in IAV-infected Mφ and its induction was required for optimal production of type I IFN at two steps: via interaction with MAVS to regulate IFN-β transcription and via activation of PKR to stabilize IFN-β mRNA. Notably, the loss-of-function in RIPK3 has no effect on cell death responses to IAV-infected Mφ, both in vitro and in vivo, indicating a new cell-death independent function for RIPK3 in innate anti-viral responses.

RIPK3-deficient mice are highly susceptible to IAV infection and exhibit a heightened pulmonary viral load
To examine a potential role for RIPK3 in immunity to IAV infection, wild-type (WT) and Ripk3 -/mice were infected with a low dose (50 pfu) of IAV. Ripk3 -/mice exhibited significant morbidity as shown by increased weight loss (Fig 1A) as well as mortality ( Fig 1B) compared to WT mice. Similar data were obtained using a higher dose of (90 pfu %LD 50 ) of IAV infection (S1A and S1B Fig). This increase of mortality was associated with a significantly increased pulmonary viral load ( Fig 1C) and decreased levels of active type I IFN in both the airways ( Fig  1D) and the lungs (Fig 1E and S1C Fig). Corresponding to the increased pulmonary viral load ( Fig 1C), Ripk3 -/lungs had a higher frequency of viral nucleoprotein (NP) + cells in both epithelial cells (Non-leukocytes, CD45 -NP + ) as well as leukocytes (CD45 + NP + ) ( Fig 1F). Interestingly, the percentage of NP + pulmonary Mφ (CD45 + F4/80 + CD19 -) was higher in the lung (S1D Fig) and BAL (Fig 1G) of Ripk3 -/mice, indicating RIPK3-deficient Mφ are more susceptible to IAV infection in vivo. The initial control of virus propagation is a major determinant of an adequate host immune response, allowing elimination of the pathogen with minimal immunopathology. In line with this, the significant increase in viral load in Ripk3 -/mice correlated with markedly enhanced inflammation (Fig 1H, S2A and S2B Fig) and immunopathology ( Fig 1I and S1E Fig) as well as reduced pulmonary function ( Fig 1J). Collectively, these data indicate that RIPK3 plays an indispensable role in immunity to IAV infection by regulating host pulmonary anti-viral responses and reducing pulmonary immunopathology.

RIPK3 is required for optimal induction of type I IFN in Mφ infected with IAV
During the steady state, the pulmonary compartment is primarily comprised of resident alveolar Mφ (AMφ). However, after pulmonary infection the recruitment of monocyte/Mφ from the bone marrow is critical for host defense to infection [23]. Since Mφ are the primary source of type I IFN in response to pulmonary viral infections [1][2][3]10] and Ripk3 -/mice elicited attenuated type I IFN responses to IAV, we next determined whether Ripk3 -/-Mφ are impaired in their ability to produce type I IFN in vitro. Consistent with the significant reduction of type I IFN and increased viral load in the lungs of RIPK3-deficient mice, a significant reduction of total active type I IFN and IFN-β was observed in IAV-infected Ripk3 -/-BMD-Mφ (Fig 2A and  In addition, RIPK3-deficient BMD-Mφ were more permissive to IAV infection, as evaluated by qPCR for IAV NS1 transcripts (Fig 2C), flow cytometry for NP protein (Fig 2D), or standard plaque assay (S3C Fig). Similarly, WT BMD-Mφ treated with the selective inhibitor of RIPK3 activity (GSK '843) also exhibited less active type I IFN upon IAV infection ( Fig 2E) with a higher viral load ( Fig 2F). Finally, we sought to extend the role of RIPK3 in immunity to IAV in human monocyte-derived Mφ. Using monocyte-derived Mφ generated from peripheral blood mononuclear cells (PBMC) obtained from healthy donors and infected with a human strain of IAV (H3N2), RIPK3-inhibited PBMC released significantly less type I IFN ( Fig 2G) and exhibited an elevated viral load ( Fig 2H). These data collectively indicate that RIPK3 regulates the induction of type I IFN in murine and human Mφ infected with IAV.
As necroptosis has been shown as a mechanism involved in cytokine release (e.g. IL-1) [19], and RIPK3 is a key player in this cell death pathway, we initially hypothesized that RIPK3-mediated necroptosis is required for the secretion of type I IFN from IAV-infected BMD-Mφ. Consistent with other experimental models using LPS [24], which show that the addition of the pan-caspase inhibitor (zVAD-FMK) is required for the induction of necroptosis (S3D Fig), we also found that inhibition of caspases via zVAD increased necroptosis in IAV-infected BMD-Mφ ( Fig 2I). In these experimental models, the induction of necroptosis was RIPK1and RIPK3-dependent, as necroptosis was completely abrogated by addition of necrostatin-1 (Nec-1) or the loss-of-function of RIPK3 (Fig 2I and S3D Fig). However, in the absence of zVAD, using the LDH assay (Fig 2J) Fig 3A). Additionally, we found that prior to IAV infection, RIPK3 is primarily localized in the cytoplasm (Fig 3B and 3C). However, following IAV infection of WT BMD-Mφ, the levels of cytoplasmic RIPK3 markedly decreased, while there was an increase in RIPK3 translocation to the mitochondria (Fig 3B and 3C). Interestingly, we also found that RIPK3 interacted with MAVS ( Fig 3D and S4B Fig) Fig 3F). Furthermore, this enhanced interaction between RIPK1/MAVS led to a significantly increased level of phosphorylation of the downstream mediators of MAVS signalling, TBK1 and the transcription factor IRF3 (Fig 3G and  3H). A similar trend was also observed in stimulation of Ripk3 -/-Mφ with the RIG-I ligand, 5'triphosphate (ppp) dsRNA, which led to increased levels of interaction between RIPK1 and MAVS, increased phosphorylation of IRF3, as well as IFN-β transcripts (S4D- S4F Fig). To test whether RIPK1 directly mediates TBK1 activation, we next inhibited RIPK1 using Nec-1 and prepared prior to and 6 days after IAV infection. At low power, inflammation is absent in both Wild Type and Ripk3 -/-(day 0). At high power, the inflammatory infiltrate is composed of lymphocytes, histiocytes and neutrophils within the alveolar space (solid arrow) and bronchiolar lumen (dotted arrow), shown at 6 days postinfection. Scale bar represents 1mm (low magnification) and 50μm (higher magnification). Using flexivent, total respiratory resistance (J) of uninfected or IAV-infected mice was measured following methacholine challenge at day 6 post-infection. Data are represented as mean ± SEM. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 between genotypes as indicated, in J, † indicate significant differences over baseline parameter readings of the same genotype. Except in A and B (as indicated), n = 4-8 animals per group per time point.  demonstrated that TBK1 activation was completely abolished in IAV-infected Ripk3 -/-BMD-Mφ ( Fig 3H and S4G Fig). Moreover, inhibition of RIPK1 by Nec-1 significantly decreased the levels of IFN-β mRNA after IAV infection ( Fig 3I). These data collectively indicate that in the absence of RIPK3, there is increased activation of the RIPK1-dependent RIG-I/ MAVS signaling pathway. Additionally, these findings support that RIPK3 plays as a negative regulator of RIPK1-mediated type I IFN signaling pathways during IAV infection.

RIPK3 activates the PKR pathway in IAV-infected BMD-Mφ, increasing the integrity of IFN-β transcripts and promoting protection
Given the increased activation of TBK1/IRF3 signaling in Ripk3 -/-BMD-Mφ, we next investigated whether this leads to an upregulation at the transcriptional level of IFN-β. In line with the increased TBK1/IRF3 signaling, IFN-β transcripts were elevated in IAV-infected RIPK3-deficient BMD-Mφ ( Fig 4A). Similarly, the levels of IFN-β transcripts were also significantly elevated in the lungs of Ripk3 -/mice after 3 and 6 days of IAV infection ( Fig 4B). These results were surprising as the levels of IFN-β protein were significantly reduced in both IAV-infected Ripk3 -/-BMD-Mφ (Fig 2A and 2B) and lungs ( Fig 1D and 1E and S1C Fig). To address the disparity between the transcription and translation of IFN-β in RIPK3-deficient BMD-Mφ and lungs, we next investigated the mechanism involved in IFN-β post-transcriptional regulation. Protein kinase R (PKR) is an important player in the host response to viral infections mainly via phosphorylating the α subunit of the translation initiation factor eIF2 (eIF2α) that inhibits both host and viral mRNA translation, thus suppressing viral propagation [5,25]. Additionally, several studies also demonstrate that PKR plays a key role in augmenting type I IFN responses to viral infection [6,26]. The mechanism of PKR function is either at a translational level through activation of eIF2α kinase [27] or at a post-transcriptional level by preserving the integrity of IFN-β mRNA, via the maintenance of its poly(A)-tail [6]. We found that during IAV infection there was no difference in phosphorylation of eIF2α in Ripk3 -/or WT BMD-Mφ ( Fig 4D). However, after IAV infection, activation of PKR was markedly reduced in Ripk3 -/-BMD-Mφ, compared to the WT as assessed by confocal microscopy (Fig 4C) or western blot (Fig 4D), and no significant effects were observed in the total PKR protein (S5A Fig).
Correlating to these in vitro findings, we also found a significant reduction of PKR phosphorylation in the lungs of IAV-infected Ripk3 -/mice, compared to infected WT mice ( Fig 4E). Furthermore, the reduction of PKR activation in IAV-infected Ripk3 -/-BMD-Mφ correlated with diminished IFN-β mRNA stability, as evaluated by the levels of IFN-β mRNA after synthesis of cDNA with oligo(dT) primers, compared to hexamer primers ( Fig 4F). Comparable to another viral model [6], this effect was specific to IFN-β mRNA, as the levels of GAPDH (S5B Fig  IAV, PKR activation is profoundly compromised, leading to reduced IFN-β mRNA stability and thus IFN-β production. Finally, to directly address whether the reduced anti-viral function of Ripk3 -/-BMD-Mφ in vitro translates to an impaired control of IAV replication in vivo, we adoptively transferred (intratracheally) BMD-Mφ from either Ripk3 -/or WT mice into Rag1-deficient mice (lacking B and T cells), which were then infected intranasally with IAV (Fig 4I). At day 3 post IAVinfection, the Rag1 -/mice that received Ripk3 -/-BMD-Mφ showed a significantly increased pulmonary viral titre in comparison to the Rag1 -/mice that received WT BMD-Mφ ( Fig 4J). Finally, to provide the direct link between reduction of type I IFN in RIPK3-deficient mice and susceptibility to IAV infection, we reconstituted IFN-β in the lungs of Ripk3 -/mice and showed that there was a significant reduction in pulmonary viral load, which was comparable to the viral load in infected WT mice (Fig 4K). Taken together, our data provide the first evidence that RIPK3 intrinsically regulates anti-viral immunity in Mφ, independent of its conventional role in necroptosis, by driving PKR activation and the IFN-β anti-viral effector program.

Discussion
In the present study, we define a novel and critical role of RIPK3 in host defense against IAV. Our findings provide strong evidence that in IAV-infected Mφ, RIPK3 regulates type I IFN production at both the transcriptional level, via interaction with the RIG-I/MAVS signaling pathway, as well as the post-transcriptional level, via activation of PKR ( Fig 5).
RIPK3 was initially identified as a master regulator of necroptosis [28]. Genetic studies have undoubtedly shown the critical physiological role of RIPK1/3 dependent necroptosis in embryonic development [29,30]. Moreover, accumulating evidence indicates that RIPK1/3 are also key players in host defense. Several viral infections have been shown to initiate RIPK1/ 3 mediated necroptosis, which contributes to host immunity against the infection [14,15,31]. While a previous study suggested that the protection and survival of RIPK3-deficient mice is comparable to WT mice following IAV infection [32], here we have demonstrated that RIPK3deficient mice are remarkably susceptible to pulmonary IAV infection, which is also in line with a recent study by Balachandran's group [16]. The exact nature of this difference is unknown, but we speculate that the strain of IAV, as well as the low dose of IAV (~0.4 LD 50 ), which was weight-adjusted, may potentially explain these differences. Additionally, following IAV infection, it was shown that RIPK3 was critical in the production of IL-1β by Mφ via the NLRP3 inflammasome [33]. Although, the in vivo consequences of this deficiency were not investigated in that study, activation of the NLRP3 inflammasome was previously shown to be crucial in immunity to IAV infection [34].

Fig 5. RIPK3 enhances innate anti-viral immunity against Influenza A virus.
Pulmonary infection by IAV triggers the recruitment of monocytes from the bone marrow that differentiate into macrophages. IAV encounters and infects those macrophages, where viral RNA activates the RIG-I/MAVS pathway, leading to production of the key anti-viral cytokine IFN-β. IAV-induced RIPK3 interaction with MAVS at the mitochondria and may represent an immune evasion strategy to decrease IFN-β production. In the absence of RIPK3, there is increased RIPK1/ MAVS interactions, which enhance downstream signaling, resulting in higher TBK1/IRF3 activation and IFN-β mRNA levels. However, this mechanism is counteracted by the RIPK3-mediated activation of PKR. PKR stabilizes IFN-β mRNA through the poly(A) tail, leading to increased IFN-β protein production and, ultimately, host protection.
https://doi.org/10.1371/journal.ppat.1006326.g005 RIPK3 deficient mice are fully resistant to murine cytomegalovirus (MCMV) [15], murine hepatitis virus [35] and lymphocytic choriomeningitis virus [36], but they are particularly susceptible to vaccinia virus [14]. The susceptibility of RIPK3-deficient mice to vaccinia has been directly linked to necroptosis in a RIPK1-dependent manner [14], while their resistance to MCMV-despite their inability to induce necroptosis-was RIPK1-independent [15]. These differences might be the reflection of a dual regulatory role of RIPK1 in the transcription of cytokines, as well as cell death. For instance, RIPK1 was shown to be essential in inducing inflammatory cytokines (IL-6, IL-1β, TNF-α) in response to bacterial [37] and viral infections [38]. This pro-inflammatory role may be explained by its ability to trigger NF-κB activation via a TLR3/TRIF-dependent pathway [39,40]. In the case of type I IFN, during dsRNA responses [38] and following VSV or Sendai virus infection [41], RIPK1 has been shown to be involved in upregulation of type I IFN signaling and can interact with the RIG-I/MDA5/ MAVS complex [4,42,43]. Similar to these studies, our data indicate that during IAV infection, in the absence of RIPK3, there is a markedly increased interaction between RIPK1 and MAVS compared to WT that leads to enhanced activation of TBK1/IRF3 and transcription of IFN-β mRNA in Mφ. Interestingly, others have previously reported that RIPK3 negatively regulates the TRIF-RIPK1-induced NF-κB pathway [39]. Our data suggest a potentially similar mechanism in which the interaction of RIPK3 with MAVS limits its interaction with RIPK1 to dampen TBK1/IRF3 activation. Whether RIPK3 directly inhibits RIPK1 recruitment to RIG-I/ MAVS, or recruitment of other partners in the complex, requires further investigation. Taken together, our data support the function of RIPK1 as an activator of host immunity, while RIPK3 serves to limit RIPK1 activity, regulating the inflammatory response at the signaling level during the "tug-of-war" between host defense and tolerance.
Several recent reports [17,18,33,44] describe the functional role of RIPK3 in regulating pro-inflammatory cytokines, independent of necroptosis. Herein, we showed that RIPK3 controls IFN-β production at the mRNA level, by regulating the stability of IFN-β transcripts through the activation of PKR. It is well established that PKR-deficient cells are impaired in type I IFN production following viral infections. Interestingly, a study by Schulz and colleagues revealed that PKR activation is indispensable in the production of IFN-β to MDA5-mediated viruses (e.g. ECMV) but dispensable for RIG-I-mediated viruses (e.g. Sendai virus, ΔNS1-IAV) in infected DC. However, another study indicated that the optimal production of type I IFN in pulmonary macrophages infected with IAV was dependent on activation of PKR [45]. Similarly, we also found that PKR plays an indispensable role in the stability of IFN-β mRNA by maintaining the poly(A) tail in IAV-infected BMD-Mφ, rather than controlling IFN-β production through eIF2α kinase activation. How RIPK3 regulates PKR activation and which molecular mechanisms are involved upstream of RIPK3 in this process need further investigation.
Interestingly, the DNA-dependent activator of IFN regulatory factors (DAI) is a recently characterized sensor of IAV that modulates both cell death responses [46] and inflammation [47]. In the study by Thapa et al., DAI was shown to promote apoptosis and necroptosis upstream of RIPK3 in IAV-infected fibroblasts, supporting DAI as an activator of RIPK3. Moreover, DAI was first identified as an activator of TBK1/IRF3 in the type I IFN response to herpes simplex virus 1 [48]. These studies are intriguing and may reveal the differential role of RIPK3 in regulating type I IFN production via sensing IAV genomic RNA and the RIG-I/ MAVS axis versus necroptosis via DAI/MLKL axis. Potentially, the differential expression of RIG-I versus DAI in different cell types can dictate the functional role of RIPK3 in immune cells (e.g. macrophages) versus structural cells (e.g. fibroblasts or epithelial cells). Furthermore, the replicative capacity of IAV and the levels of cytosolic IAV genomic RNA in each cell type may preferentially activate one axis versus the other. Certainly, this is a very exciting area of research and further investigation is required to determine the mechanisms regulating innate immunity to IAV infection.
Moreover, although in the current study we demonstrate that the function of RIPK3 is dispensable in Mφ death modality during IAV infection, the contribution of RIPK3 in pathogenesis of IAV in vivo is certainly more complex and we cannot exclude its potential role as a death kinase in other immune cells or structural cells. In this context, a recent publication highlighted the critical role of RIPK3-mediated necroptosis in promoting immunity to IAV in fibroblasts [16]. Interestingly, they also showed that the levels of type I IFN, although modest, were significantly reduced in RIPK3-deficient fibroblasts after infection with IAV [22]. However, fibroblasts contribute substantially less to type I IFN production than macrophages, which have been demonstrated to be the main producer of type I IFN during IAV infection [10]. Both our study, as well as Balachandran's studies indicate that RIPK3 has evolved to promote viral clearance through distinct mechanisms in immune (macrophages) and lung structural cells (fibroblasts). Furthermore, the function of RIPK3 also appears to differ among immune cells. For instance, RIPK3-deficient DC were impaired in the production of proinflammatory cytokines following LPS stimulation, while Mφ were not [18]. Our data also support this notion since the production of type I IFN was only impaired in Mφ but not DC. Thus, the mechanisms underlying the differential activation of RIPK3 is certainly cell, as well as pathogen specific.
Pulmonary Mφ convert into highly active cells following detection of IAV viral particles by PRRs and become the major source of type I IFN [10]. As the initial control of virus propagation through type I IFN is a major determinant of an adequate host immune response that eliminates the pathogen with minimal immunopathology, IAV has evolved multiple strategies to subdue Mφ type I IFN pathways. We have recently demonstrated that the mitochondrial PRR belonging to the NOD-like family (NLRX1) plays a critical role in Mφ by maintaining mitochondrial fitness and preventing IAV-induced cell death to maximize type I IFN production [13]. IAV's strategies for suppressing type I IFN in Mφ is not limited only to PRRs, as they also target Mφ eicosanoid pathways [10]. In line with this, one may envision a scenario where IAV evolved a strategy to upregulate RIPK3 to dampen RIPK1/MAVS-mediated TBK1/IRF3/ type I IFN production to facilitate its replication. However, the host may have counter-evolved to promote a secondary role of RIPK3 in activating PKR and eliciting type I IFN responses through mRNA stability. Additional studies are required to address this unique function of RIPK3 in regulating type I IFN responses, following infection with other strains of IAV or pathogens.
In summary, we provide in vivo evidence showing that the lack of RIPK3 limits the production of type I IFN, which results in enhanced IAV propagation, early excessive host inflammatory responses that contribute to pulmonary tissue and vasculature damage/dysfunction and, ultimately, enhanced mortality. The integrity of type I IFN pathways is essential in anti-viral immunity and identification of molecular mechanisms that are involved in maintaining this will undoubtedly provide new opportunities for targeted therapy of highly pathogenic strains of IAV.

Materials & methods Mice
Six-to ten-week-old C57BL/6 mice were purchased from Jackson Laboratories. Ripk3 -/-, kindly provided from Vishva Dixit (Genentech, San Francisco), and Mavs -/-, a kind gift from Dr. Salman Qureshi (McGill University), were bred at McGill University. Experiments were performed using age-and sex-matched mice.

Isolation and culture of primary macrophages and cell lines
Murine Bone Marrow-Derived Macrophages (BMD-Mφ) were isolated following aseptic flushing of tibiae and femurs of eight-to ten-week-old mice. Macrophages were differentiated from bone marrow precursors for 7 days in RPMI-1640 supplemented with 30% (vol/vol) L929 cell-[American Type Culture Collection (ATCC)] conditioned medium, 10% (vol/vol) FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 1% essential and nonessential amino acids, 10mM HEPES and 100 U/mL penicillin/streptomycin. To generate BMDC, bone marrow was cultured in RPMI-1640 supplemented with 10% (vol/vol) FBS, 2 mM L-glutamine, 1 mM sodium pyruvate, 1% nonessential amino acids, 100 U/mL penicillin/streptomycin and 0.35% Β-mercaptoethanol, containing 20 ng/ml of GM-CSF as described previously [49]. Alveolar macrophages (AMφ) were collected by bronchoalveolar lavage of naïve mice using cold, sterile PBS. AMφ were cultured in RPMI-1640 supplemented with 10% (vol/vol) FBS, 2 mM L-glutamine, 10mM HEPES and 100 U/mL penicillin/streptomycin. After 1h adhesion, AMφ were washed with PBS and placed in fresh media. Madin-Darby Canine Kidney cells (MDCK) were obtained from ATCC and maintained in Dulbecco's Modified Eagle Medium (DMEM) enriched with 10% (vol/vol) FBS, 2mM L-glutamine, and 100 U/mL of penicillin/streptomycin. To generate human monocyte-derived macrophages, peripheral blood mononuclear cells (PBMCs) were isolated from healthy donors using Ficoll-Paque PLUS (GE Healthcare, Burlington, ON, Canada), according to the manufacturer's protocol. PBMCs were then cultured in RPMI with 2% human serum with 20ng/mL of human M-CSF. Monocytes were differentiated for 7 days with fresh media added every second day. All reagents and supplements pertaining to cell culture were purchased from GIBCO. Cells were seeded at a density 0.5-1.5x10 6 cells/well of a 6-well plate.

Viruses and infections
All in vitro and in vivo infections were performed using influenza A/Puerto Rico/8/34 (H1N1) virus (IAV), kindly provided by Dr. Jonathan A. McCullers (St. Jude Children Research Hospital), except for in vitro infections of human cells that were performed using the clinical strain H3N2 A/Hong-Kong/1/68. Mice were challenged intranasally (in 25μL PBS) with IAV at a sublethal dose of 50 pfu or a lethal dose of 90 pfu. In vitro, BMD-Mφ were seeded in tissue culture plates the day before infection, unless indicated otherwise, and infections were performed in fresh medium lacking L929 cell-conditioned medium with 1, or 5 multiplicities of infection (MOI) of virus. Viruses were propagated and isolated from MDCK cells and titrated using standard plaque assay in MDCK cells [50].

Histopathological analysis
Lungs were inflated and fixed for at least 24 hours with 10% formalin, and then embedded in paraffin. 5 μm sections were cut and stained with hematoxylin-eosin. Slides were scanned at a resolution of 200X magnification (Nanozoomer scanner, Hammamatsu, Japan) and pictures were taken using NDPI viewer (Hammamatsu, Japan).

Analysis of pulmonary function
Airway responses to methacholine were evaluated using a small animal ventilator (flexiVent apparatus and flexiVent 5.1 software) as previously described [51].

RNA isolation and RT-qPCR
RNA from BAL of IAV-infected mice or from BMD-Mφ was extracted using Qiazol reagent (Qiagen) according to manufacturer's instructions. Five hundred ng of RNA was reverse transcribed using the Quantitect Reverse Transcription kit (Qiagen), as directed by the manufacturer. cDNA was generated by qPCR using EvaGreen SYBR Green (Biorad) and the following primers: GAPDH-forward:

Western blot
Cells obtained from BAL of WT or Ripk3 -/mice at day 3 post-infection or BMD-Mφ were lysed in lysis buffer (1% Triton X-100, 150mM NaCl, 20mM Hepes pH7.5, 10% glycerol, 1mM EDTA, supplemented with anti-protease and anti-phosphatase cocktails, Roche) and protein concentration was determined using BCA assay (Pierce). 20 μg of protein were resolved by SDS-PAGE and transferred onto PVDF membranes (Biorad). Membranes were blocked and incubated overnight at 4˚C with gentle agitation with primary antibodies. The following primary antibodies were used: anti-RIPK3 ( . Primary antibodies were followed by HRP-conjugated secondary antibodies and signal was detected using Clarity ECL kit (Biorad) and acquired on Chemidoc MP System (Biorad). Densitometry analyses were performed using ImageJ software (NIH).

Mitochondria isolation
Mitochondrial and cytosolic fractions from IAV-infected or uninfected WT BMD-Mφ were extracted using Qproteome Mitochondria Isolation Kit (Qiagen) following manufacturer's intructions. Purified mitochondria were lysed with RIPA buffer and further analysed by western blot using antibodies against RIPK3, CYPD, MAVS and actin.
Coverslips were mounted (ProLong Gold Anti Fade, Invitrogen) onto microscope slides. Images were acquired using a Zeiss LSM 700 laser-scanning confocal microscope.

Adoptive transfer model of infection
BMD-Mφ from Wild Type and Ripk3 -/mice were generated as described previously. On day 7 of differentiation, BMD-Mφ were harvested and resuspended at a density of 1 x 10 6 cells per 50μL. BMD-Mφ were then transferred by the intratracheal route into naïve Rag1 -/mice. After 2 hours, Rag1 -/were intranasally infected with 500 PFU of IAV. Lungs were harvested and processed as previously described for viral load analysis.

Interferon-β treatment
Recombinant murine interferon-β was purchased from R&D Systems (#8234-MB-010). Mice were intranasally infected with 50 pfu of IAV. On day 2 post-infection, mice were given intranasally either PBS or IFN-β (2000U). Mice were euthanized on day 3 post-infection and lungs were harvested and processed to determine pulmonary viral load as previously described.