Type I Interferon Signaling Regulates Ly6Chi Monocytes and Neutrophils during Acute Viral Pneumonia in Mice

Type I interferon (IFN-I) plays a critical role in the homeostasis of hematopoietic stem cells and influences neutrophil influx to the site of inflammation. IFN-I receptor knockout (Ifnar1 −/−) mice develop significant defects in the infiltration of Ly6Chi monocytes in the lung after influenza infection (A/PR/8/34, H1N1). Ly6Chi monocytes of wild-type (WT) mice are the main producers of MCP-1 while the alternatively generated Ly6Cint monocytes of Ifnar1 −/− mice mainly produce KC for neutrophil influx. As a consequence, Ifnar1 −/− mice recruit more neutrophils after influenza infection than do WT mice. Treatment of IFNAR1 blocking antibody on the WT bone marrow (BM) cells in vitro failed to differentiate into Ly6Chi monocytes. By using BM chimeric mice (WT BM into Ifnar1 −/− and vice versa), we confirmed that IFN-I signaling in hematopoietic cells is required for the generation of Ly6Chi monocytes. Of note, WT BM reconstituted Ifnar1 −/− chimeric mice with increased numbers of Ly6Chi monocytes survived longer than influenza-infected Ifnar1 −/− mice. In contrast, WT mice that received Ifnar1 −/− BM cells with alternative Ly6Cint monocytes and increased numbers of neutrophils exhibited higher mortality rates than WT mice given WT BM cells. Collectively, these data suggest that IFN-I contributes to resistance of influenza infection by control of monocytes and neutrophils in the lung.


Introduction
Type I interferons (IFN-I) are produced by different cell types including alveolar macrophages (AM), plasmacytoid dendritic cells, and epithelial cells following virus infection in the lung [1,2].These IFN-I cytokines engage a unique heterodimeric IFN-a receptor (IFNAR) to induce various antiviral effectors [3].IFN-Irelated antiviral effectors, including protein kinase PKR [4], 29-59oligo A synthetase [5], and Mx-GTPase [6], control influenza virus infection by their own or in cooperation with various other signaling pathways [7,8].Even though influenza NS1 protein provides antagonistic properties against IFN-I-inducible antiviral proteins, which help virus to circumvent host barriers [9], increased susceptibility in Ifnar1 2/2 mice indicates that IFN-I signaling still plays a significant role in protecting the host after influenza infection in vivo [10,11].
Monocytes emigrate from bone marrow (BM) thorough CCR2 receptor-mediated signaling and then monocyte-derived cells mediate inflammatory responses against influenza infection [12,13].Because of the important function of these monocytes, IFN-I signaling-mediated monocyte differentiation should be examined to better understand the regulation of leukocyte differentiation at large.In accordance, one recent study showed that IFN-I signaling triggers hematopoietic stem cell (HSC) proliferation [14].Similarly, mice lacking IFN regulatory factor-2, a suppressor of IFN-I signaling, fail to maintain quiescent HSC [15].These studies of the effect of IFN-I on regulation of cell homeostasis may explain different cell constitutions of Ifnar1 2/2 mice.However, specific cell populations directly affected by IFN-I during hematopoiesis and their contributions toward unique cell composition in peripheral tissues are not yet described.Thus, the role of IFN-I on the regulation of overall monocyte differentiation and infiltration into inflamed tissue needs to be analyzed in depth.
Although neutrophils are universally accepted as important in bacterial infection resistance [16], their role in viral infection remains controversial.Tate and colleagues reported that mice undergo more pronounced disease when neutrophils are absent [17,18].However, depletion of neutrophils not only results in increased viral burden but also in decreased lung inflammation, indicating that neutrophils contribute to control virus dissemination but may augment overall pathogenesis [19].Clinically, excessive neutrophil recruitment, especially after highly pathogenic avian H5N1 or 1918 pandemic influenza infection, seems to play a detrimental role in acute lung injury [20,21].Thus, neutrophil infiltration seems to be closely related to tissue damage following infection as well as to inflammation, and the feed-back regulation of neutrophil recruitment to the site of infection needs to be tightly regulated.
In the current study, we adopted the influenza infection model and found that Ifnar1 2/2 mice undergo more acute and severe inflammation than B6 wild-type (WT) mice.IFN-I was directly involved in Ly6C hi monocyte differentiation from its precursor and these Ly6C hi monocytes exclusively provided MCP-1 in the lung after influenza infection.Further, Ifnar1 2/2 mice with defects in monocyte maturation produced excess KC chemokine and developed high mortality and severe neutrophilia when compared with WT mice.Our results suggest that IFN-I is required to resist influenza infection by orchestrating the leukocyte population in the lung and chemokines produced by those cells.

Results
Ifnar1 2/2 mice develop more acute and severe lung inflammation than WT mice Because many previous studies indicate a crucial role for IFN-I in host defense against influenza infection, we looked for crucial regulatory factors that are mainly regulated by IFN-I after influenza infection using Ifnar1 2/2 mice of B6 background.First we challenged Ifnar1 2/2 and WT mice with a lethal dose (1610 5 pfu) of PR8 virus.The Ifnar1 2/2 mice started to die 5 days post infection (dpi) and all were dead within 8 dpi while approximately 50% of WT mice survived (Figure 1A).When we decreased the challenge dose of PR8 virus (2610 4 pfu), virus infection killed Ifnar1 2/2 mice from 5 dpi and no mice survived at 13 dpi while 80% of WT B6 mice survived (Figure 1A).WT mice started to regain body weight 8 or 9 days after infection while Ifnar1 2/2 mice continued to lose weight until they died (Figure 1A).Since the contribution of IFN-I in viral clearance is controversial [22,23], we next addressed viral titer in the lung at 2 and 5 dpi with influenza PR8 virus (1610 5 pfu).Intriguingly, Ifnar1 2/2 mice showed higher viral titer in the lung than WT mice at 2 dpi but not at 5 dpi (Figure 1B).However, total protein levels were significantly higher in the bronchoalveolar lavage fluid (BALF) of Ifnar1 2/2 mice at 5 dpi than in WT mice (Figure 1C).Of note, significant levels of IL-6, TNF-a and IP-10 were determined at 3 dpi in the BALF of Ifnar1 2/2 mice and high levels IFN-c and IL-6 at 5 dpi when compared with levels in WT mice (Figure 1D).Lung histopathology of WT and Ifnar1 2/2 mice after H&E staining revealed more edema, alveolar hemorrhage, alveolar wall thickness, and neutrophil infiltration in Ifnar1 2/2 mice than in WT mice at 5 dpi (Figure 1E and 1F).Staining specifically for myeloperoxidase (MPO), most abundantly present in the granules of neutrophils, confirmed increased numbers of MPO + neutrophils in the lung of Ifnar1 2/2 mice after influenza infection (Figure 1F).

IFN-I signaling is involved in leukocyte infiltration in the lung after influenza infection
As Ifnar1 2/2 mice of B6 background exhibited severe pathology in terms of hyper secretion of pro-inflammatory cytokines and neutrophilia in lung after infection with influenza virus, we examined profiles of infiltrated cell populations in BALF in a time-dependent manner.From 1 dpi with influenza virus, predominant numbers of neutrophils (Ly6C int Ly6G + ) were infiltrated into the lung of WT and Ifnar1 2/2 mice (Figure 2A).Of note, the proportion (Figure 2A) and absolute numbers (Figure 2B) of infiltrated neutrophils in BALF were much higher in Ifnar1 2/2 mice than in WT mice.Furthermore, the numbers of neutrophils in WT mice peaked at 3 dpi and decreased at 5 dpi while neutrophils in Ifnar1 2/2 mice increased until 5 dpi, when mice began to die (Figure 2B).Meanwhile, monocytes (Ly6C + Ly6G 2 ) were gradually infiltrated into the lung of WT mice in a time-dependent manner after influenza infection (Figure 2A and 2B).However, fewer monocytes were recruited into the lung of Ifnar1 2/2 mice than in WT mice (Figure 2A and 2B) and they were Ly6C int rather than Ly6C hi (Figure 2C).Since Ly6C hi monocytes have similar phenotype to myeloid-derived suppressor cells, which expand during cancer, inflammation, and infection [24], we tested their ability to suppress CD4 + T cells.However, we did not find any evidence of CD4 + T cell suppression by Ly6C hi monocytes (Figure S1A).We next examined expression levels of co-stimulatory molecules, such as CD40, CD80, CD86, and MHCII, on surfaces of the respective cell populations, but overall these markers were not significantly different between cells isolated from WT and Ifnar1 2/2 mice (Figure S1B).We further analyzed surface expression of various markers on monocytes of infected WT and Ifnar1 2/2 mice (Figure S1C).Monocytes in the lung of infected WT and Ifnar1 2/2 mice expressed macrophage-related markers and they were negative for markers specific for dendritic cells, lymphocytes, and natural killer cells [25].
Because MyD88 signaling cooperates with IFN-I in Ly6C hi monocyte recruitment in a Listeria monocytogenes infection model [26], we tested whether Toll-like receptor (TLR) or RIG-I-like receptor (RLR) signaling is involved in Ly6C hi monocyte regulation in our model.However, mice have defects in TLR (Myd88 2/2 Trif 2/2 ) or RLR (Ips-1 2/2 ) signaling normally generated Ly6C hi monocytes (Figure S2A).Although IPS-1 is involved in IFN-I expression, deletion of IPS-1 can be compensated by MyD88 signaling after influenza infection [27].Indeed, Ips-1 2/2 mice produced IFN-a comparable to Ips-1 +/+ mice (Figure S2B).Overall, direct engagement of IFN-I signaling through IFNAR but not TLR or RLR signaling seems to play a crucial role in Ly6C hi monocyte infiltration into the lung for host defense after influenza infection.

Author Summary
Type I interferon (IFN-I) was originally reported as a molecule that interferes with influenza virus replication.Various IFN-I inducible antiviral proteins contribute to dampening virus replication and dissemination.Thus, loss of IFN-I signaling attenuates antiviral response and aggravates disease.Recent studies suggest the possible role of IFN-I in hematopoiesis, which subsequently might have an effect on the immune cell response at the site of infection.Indeed, IFN-I signaling-defective mice have been shown to develop aberrant cell populations.The aim of this current study was to clarify the mechanisms of IFN-I signaling in the regulation of monocytes and neutrophils.We show that IFN-I is directly involved in monocyte differentiation and that loss of IFN-I signaling allows mice to generate monocytes whose gene profile is significantly different.We found that monocytes are an important source of chemokines for further monocyte recruitment, but IFN-I-defective monocytes produce chemokines for neutrophil recruitment.As a result, mice lacking IFN-I signaling recruit more neutrophils and a reduced number of alternatively generated monocytes.Thus, our findings indicate that authentic monocyte differentiation, which requires IFN-I signaling, is critical in controlling neutrophils and protecting mice against influenza virus infection.
Ly6C pos monocytes from influenza-infected lung of WT and Ifnar1 2/2 mice have different characteristic features To assess the innate immune cells of WT and Ifnar1 2/2 mice in more detail, we next examined their morphologies (Figure 2D).Both AM and neutrophils in the lung seemed identical in WT and Ifnar1 2/2 mice except that some neutrophils from Ifnar1 2/2 mice exhibited larger size and had a more diffused nucleus, but monocytes were clearly different in the lung of WT and Ifnar1 2/2 mice after influenza infection.Interestingly, Ly6C hi monocytes morphologically resemble the foamy macrophages previously found in the lung of Mycobacterium bovis bacillus Calmette-Gue ´rininfected mice [28].To confirm this, we stained Ly6C pos monocytes with Nile red, which mainly stains lipid body, and found that only Ly6C pos monocytes isolated from the lung of WT mice were positively stained (Figure 2E).Next we generated BM chimeric mice (WT BM into Ifnar1 2/2 mice and vice versa) to confirm whether the defect in IFN-I signaling in the hematopoietic cell lineage can trigger the alteration of monocyte phenotypes in the lung of influenza-infected mice.As a result, Ly6C hi monocytes were generated in Ifnar1 2/2 recipient mice reconsti-tuted with WT BM cells but were not detected in WT recipient mice that received Ifnar1 2/2 BM cells (Figure 2F).These suggest that IFN-I signaling in the hematopoietic cell lineage plays an indispensable role for differentiation of Ly6C hi monocytes against influenza infection.Monocytes from WT mice produce MCP-1 while those from Ifnar1 2/2 mice produce KC against influenza infection We next measured the chemokines responsible for leukocyte migration against influenza infection in the BALF at different time points.Of note, expression levels of MCP-1 and KC, the main chemokines for CCR2-and CXCR2-dependent cell recruitment, respectively [29,30], were very different in the BALF of WT and Ifnar1 2/2 mice (Figure 3A).The Ifnar1 2/2 mice had significantly lower levels of MCP-1 than the WT mice but had predominant levels of KC at 3 and 5 dpi (Figure 3A).In addition, MIP-2, another well-known molecule for neutrophil recruitment [31], was significantly higher in Ifnar1 2/2 mice at 3 dpi (Figure 3A).Both Ly6C hi and Ly6C int monocytes obtained from the lung of WT and Ifnar1 2/2 mice expressed CCR2 but not CXCR2 (Figure 3B), suggesting the critical role of MCP-1 for monocyte recruitment into influenza-infected lung.
To directly compare chemokine expression by individual cell populations, we sorted Ly6C pos monocytes from influenza-infected lung of WT and Ifnar1 2/2 mice and then cultured them in vitro.Although both monocytes and AMs isolated from WT mice produced MCP-1 (Figure 3C), recovered Ly6C hi monocytes were quantitatively overwhelming over AMs (2.6610 5 vs. 3.1610 4 in BALF per mouse at 5 dpi) (Figure 2B).These findings suggest that Ly6C hi monocytes are the main producer of MCP-1 among leukocytes in WT mice after influenza infection.In contrast, in Ifnar1 2/2 mice, KC was highly produced by both Ly6C int monocytes and AMs, but Ly6C int monocytes secreted significantly less MCP-1 than WT Ly6C hi monocytes (Figure 3C).It is noteworthy that types of chemokines expressed in monocytes from WT and Ifnar1 2/2 mice were in complete contrast; gene expression profiles analyzed by gene chip experiments support these results (Figure 3D).Collectively, these data suggest that IFN-I is crucial for determining monocyte characteristics that dramatically influence chemokine production.Monocytes from WT and Ifnar1 2/2 mice show different gene profiles For macroscopic comparison between monocytes recruited in the lung of influenza-infected WT and Ifnar1 2/2 mice, we performed gene chip analysis (Figure 4).As expected, IFN-Iregulated genes (e.g., Mx, Oas, Irf, etc.) were significantly decreased in monocytes isolated from Ifnar1 2/2 mice.Of the genes elevated in Ly6C int monocytes isolated from Ifnar1 2/2 mice, S100a8/ S100a9 and Trem1 are associated with inflammatory responses in various diseases [32,33].In contrast, negative regulators of inflammation, Trim21 and Trim30 [34,35], were up-regulated in Ly6C hi monocytes isolated from WT mice when compared to Ly6C int monocytes from Ifnar1 2/2 mice.These inflammationbiased gene expressions by Ly6C int monocytes may explain the higher of Ifnar1 2/2 mice to influenza infection.Ly6C hi monocytes from WT mice were also superior in expressing genes involved in lipid metabolism (e.g., Apoe, Apoc2, etc.).Interestingly, the 1918 pandemic influenza virus was found to block lipid metabolism as part of its evasion strategy against antiviral responses [36].Furthermore, influenza infection causes prominent inflammation in Apoe 2/2 mice [37], indicating that a defect in lipid metabolism in Ifnar1 2/2 mice might contribute to worsen inflammation.Collectively, we confirmed that monocytes from WT and Ifnar1 2/2 mice have significantly different characteristics and that lack of IFN-I signaling changes gene expression bias to augment inflammation.

IFN-I signaling is required for Ly6C hi monocyte generation
Since the different gene expression patterns in Ly6C pos monocytes of WT and Ifnar1 2/2 mice after influenza infection can be ascribable to altered monocyte differentiation from their precursors, we next analyzed BM where hematopoiesis occurs and provides common monocyte precursors.The proportion of Ly6C hi monocytes was significantly lower (6.961.9 vs. 1.060.3%) in the BM of Ifnar1 2/2 mice than in WT mice, while the proportion of Ly6C int monocytes was comparable to WT mice (17.260.6 vs. 16.361.5%)at 5 dpi (Figure 5A).However, there were no significant differences in cell morphology between WT and Ifnar1 2/2 mice (data not shown), and Ly6C pos monocytes of WT mice did not show lipid bodies unlike Ly6C hi monocytes in the lung post influenza infection (Figure 5B).Because we found higher levels of Ly6C expression in BM monocytes from WT than in Ifnar1 2/2 mice at 5 dpi (Figure 5C), we further assessed whether IFN-I signaling can directly affect differentiation of naı ¨ve BM cells in vitro.When BM cells were stimulated with WT BALF collected at 5 dpi or directly infected with influenza virus, WT BM cells were able to differentiate into Ly6C hi monocytes but Ifnar1 2/2 BM could not (Figure 5D).To confirm that this maturation defect of Ly6C hi monocytes from Ifnar1 2/2 BM is due to lack of IFN-I signaling, we co-cultured PR8-infected WT BM cells with or without anti-IFNAR1 blocking antibody.When treated with anti-IFNAR1 antibody, WT BM cells failed to differentiate into Ly6C hi monocytes (Figure 5E).Importantly, these Ly6C hi monocytes derived from WT BM stimulated by influenza virus dominantly produced MCP-1 when compared to Ly6C int monocytes derived from WT or Ifnar1 2/2 BM, while Ifnar1 2/2 Ly6C int monocytes produced robust KC instead (Figure 5F).From this observation, we suggest that absence of Ly6C hi cells in lung of Ifnar1 2/2 mice results from failure in differentiation of their precursors due to lack of IFN-I signaling.

IFN-I signaling on hematopoietic cells is required to influenza infection
To clarify the origin and roles of monocytes in depth in the absence of IFN-I signaling, we used BM chimeric mice.Ifnar1 2/2 recipient mice reconstituted with WT BM cells (WRK) have significantly more Ly6C hi monocytes than KRK (Ifnar1 2/2 donor and Ifnar1 2/2 recipient) mice (Figure 6A).Concomitantly, the MCP-1 level in the BALF was correlated with Ly6C hi monocytes (i.e., WRW and WRK) after influenza infection (Figure 6B).The fact that KRW mice elicited no higher level of MCP-1 than found in the BALF of KRK mice after influenza infection suggests that radio-resistant WT parenchymal cells do not play a major role in MCP-1 production (Figure 6B).In addition, Ly6C hi monocytes derived from WRW and WRK mice produce MCP-1 efficiently following in vitro culture (Figure 6C), clearly indicating that Ly6C hi monocytes derived from WT BM mice exclusively produce MCP-1 in response to influenza infection.To the contrary, although we saw a distinct correlation between KC production with IFN-I deficiency in monocytes (Figure 6C), WRK chimera mice still had significantly augmented KC in BALF even with high levels of MCP-1 and Ly6C hi monocytes (Figure 6B).Thus it seems likely that the complete regulation of KC also needs the IFN-Idependent signaling pathway in cells other than monocytes or there may be another regulator for KC in the absence of IFN-I signaling.
Finally, we sought to find the relationships between the defect in Ly6C hi monocyte generation and the susceptibility of chimeric mice against influenza infection.The loss of intrinsic Ly6C hi monocytes in KRW mice resulted in the increased susceptibility of mice against influenza infection as compared with WRW mice (Figure 6D).In addition, the WRK chimera mice showed less susceptibility to influenza infection than did KRK mice with the restoration of Ly6C hi monocytes (Figure 6D).To see whether neutrophils can augment inflammation and tissue damage, we isolated neutrophils from infected Ifnar1 2/2 mice and transferred them to naı ¨ve WT mice (Figure S3A).Then effect of neutrophil transfer on the lung inflammation was observed without additional infection or any treatment.Histologically, recipient WT mice had destruction in epithelial layers and showed inflammatory lesions while PBS-treated mice did not at 2 days after transfer (Figure S3B).Inflammatory cytokines in the BALF were also induced after neutrophil transfer, indicating activated neutrophil itself can cause inflammation (Figure S3C).Thus, we suggest that excess neutrophils can worsen disease even if there are good reasons for recruitment after virus infection.Even though parenchymal cells appear to maintain host defense, our data clearly show the importance of IFN-I signaling in hematopoietic cells in protection against influenza infection.

Discussion
Despite numerous previous studies focused on the direct antiviral nature of IFN-I, the data we present here suggest that IFN-I-dependent generation of Ly6C hi monocytes after influenza infection might be critical for the attenuation of neutrophil infiltration and hence prevent severe tissue damage caused by uncontrolled inflammation (Figure 7).Ly6C hi monocytes in the lung were previously reported as TNF/iNOS-producing dendritic cells [38] or inflammatory monocytes [12,39].In our observations, however, Ly6C hi monocytes were closest in morphology to the highly vacuolated foamy macrophages, which were found in granulomas in the lung of Mycobacterium bovis-infected mice [28].
Previously, it was proposed that there is no up-regulation of MCP-1 in the absence of IFN-I signaling; as a consequence, without up-regulation there will be fewer Ly6C hi monocytes in various tissues [26,40,41].We show here that only mice reconstituted with WT BM cells (WRW and WRK), but not with Ifnar1 2/2 BM cells (KRK and KRW), generated Ly6C hi monocytes and produced high amounts of MCP-1 in the BALF against influenza infection.Further, BM cells from Ifnar1 2/2 mice were unable to differentiate into MCP-1-producing Ly6C hi monocytes by in vitro stimulation, indicating that IFN-I signaling on hematopoietic cells is required for differentiating Ly6C hi monocytes.Our findings are the first to delineate a notion that MCP-1 is mainly produced by Ly6C hi monocytes, which are regulated by IFN-I against influenza infection.Additionally, it seems plausible that AMs and other cells (e.g., pulmonary MCP-1 is a chemokine that has great importance in CCR2mediated monocyte recruitment after lung inflammation [42].Although other chemokines such as MCP-2, MCP-3, and MCP-4 also contribute to attract CCR2-expressing monocytes after influenza infection, overexpression of MCP-1 results in elevated monocyte recruitment [12].In this regard, MCP-1-deficient mice show significantly reduced macrophage infiltration [43].Moreover, MCP-1 is capable of activating AM [44] and of blocking MCP-1-augmented damage on epithelial cells after influenza infection [45].Thus, Ly6C hi monocytes might play an important role for host defense against influenza infection by mediating MCP-1. Unlike MCP-1 whose level was dramatically decreased in Ifnar1 2/2 mice, KC was significantly increased after influenza infection.We observed that WRK and KRW chimeric mice produced intermediate KC when compared to WRW and KRK mice, suggesting that KC can be produced by both radio-sensitive and -resistant cells.AMs, which produced KC in vitro, seemed especially important producers as did radio-resistant cells (e.g., epithelial cells) as proposed by others [46].Since KC with MCP-1 is reversely correlated in BALF and culture supernatant with monocytes from the lung of WT and Ifnar1 2/2 mice, further studies are needed to elucidate a putative role of IFN-I-dependent Ly6C hi monocytes on neutrophil regulation as well as possible feedback regulation of KC by MCP-1 production.
Previously it was proposed that Ly6C int monocytes were converted from Ly6C hi monocytes after migration to the site of inflammation [41].Those activated forms of Ly6C int monocytes were characterized by high CX 3 CR1 but no CCR2 expression [47].In our study, however, Ly6C int monocytes generated in influenza-infected Ifnar1 2/2 mice expressed high levels of CCR2.Thus, it seems likely that Ly6C int monocytes in Ifnar1 2/2 mice could be an alternative to Ly6C hi monocytes generated under IFN-I-deficient conditions rather than a different cell subset of Ly6C expressing monocytes.
It has been reported that the spleen also stores monocytes and these cells deploy to inflammatory sites to regulate inflammation [25].To address this issue, we compared characteristics of monocytes in the spleen before and after influenza infection (Figure S4A).Interestingly, Ly6C expression level in the monocytes was much lower in the spleen of Ifnar1 2/2 mice than in WT mice in the steady-state condition (Figure S4B).The number of Ly6C hi monocytes stored in the spleen was also decreased in Ifnar1 2/2 mice compared to WT mice, whereas the number of neutrophils was comparable before influenza infection (Figure S4C).We speculate that the reduced numbers of WT and Ifnar1 2/2 monocytes after influenza infection in the spleen may indicate the spleen is a reservoir that provides monocytes to the lung.
After finding that IFN-I signaling is involved in Ly6C hi generation, we next sought to clarify the relationships between the defect in Ly6C hi monocytes and susceptibility to influenza infection.Viral titer in lethally infected lung culminated around 2 dpi and continuously decreased.Although Ifnar1 2/2 mice were intact in viral clearance and showed similar viral burden at 5 dpi compared to WT mice, they developed severe inflammation and consequently higher susceptibility.In previous studies, attenuated inflammation provided resistance to influenza infection even with increased viral burden [12,48].These findings indicate that virusinduced inflammation could be more critical than viral burden itself in the course of influenza pathology.Thus, we suggest that severe and acute lung inflammation in Ifnar1 2/2 mice, especially with uncontrolled accumulation of neutrophils due to massive KC production, contributes to increased susceptibility of those mice to influenza infection.
Accumulation of neutrophils is one of the most important events during acute respiratory distress syndrome [49].Neutrophils are generally thought to aggravate lung injury after influenza infection [50], especially in severe infections such as those we assessed in our study.To address the contribution of neutrophils to virus-induced airway hyperresponsiveness as proposed by others [51,52], we used CXCR2 blocking Ab and CXCR2 antagonist SB225002 [29,53].These materials partially dampened neutrophil responses when moderate or low doses of influenza were given, but they could not efficiently block massive influxes of neutrophils after lethal influenza infection in Ifnar1 2/2 mice (data not shown).However, we found that neutrophils can augment inflammation and tissue damage (Figure S3); also loss of Ly6C hi monocytes in KRW mice augmented neutrophil infiltration compared to WRW mice, indicating the balance between neutrophils and Ly6C hi monocytes are reversely correlated.Moreover, when chimeric mice were lethally challenged, susceptibility of these mice was directly proportional to neutrophil numbers.Our data suggest that uncontrolled neutrophils may aggravate the outcome of excess inflammation against virus infection.Even though neutrophils are thought to augment inflammation and make disease worse, we still must consider their protective role.We showed Ifnar1 2/2 mice had higher peak virus titer but were able to successfully control virus replication.Regarding previous report that neutrophils can limit virus replication [19], it seems plausible that excessively recruited neutrophils may contribute to observed virus elimination in Ifnar1 2/2 mice.However, several lines of evidence lead us to speculate that destructive trait of neutrophils dominated over their positive role during severe and acute viral pneumonia.IFN-I is used to treat several diseases, including hepatitis B virus infection [54], chronic hepatitis C virus infection [55], and multiple sclerosis [56].Since IFN-I, which is produced by virus infection, can migrate and affect BM to switch on the production of functional monocytes [57], therapeutically administrated IFN-I can communicate with BM leukocytes.This possibility suggests that clinical uses of IFN-I should be investigated in terms of modified patient leukocyte profiles, especially in those receiving prolonged IFN-I therapy.
Influenza virus NS1 protein antagonizes IFN-I responses, and influenza virus lacking the NS1 gene replicates inefficiently in tissue culture and normal egg culture conditions and shows attenuated phenotype in WT mice; however, it replicates far more efficiently in IFN-deficient Vero cells and pathogenic in Stat1 2/2 mice [58].In the clinical context, it is important to note that the NS1 protein of highly pathogenic viruses, such as H5N1 avian influenza and the 1918 pandemic influenza virus, has stronger suppressive effects on IFN-I [59,60].These viruses also are involved in more acute recruitment of neutrophils, severe lung injury, and aggressive inflammatory cytokine production (so-called cytokine storm) as found in Ifnar1 2/2 mice [20,61,62].Thus our results in Ifnar1 2/2 mice merit further study to help understand the pathogenesis of highly pathogenic influenza virus.
Our findings show the specific function of IFN-I on Ly6C hi monocyte differentiation and address the impact of this event in the lung after influenza infection.Further studies on regulation of neutrophils and Ly6C hi monocytes by potential pandemic virus may provide insight that will prove useful for development of novel therapeutic targets.

Ethics statement
All animal experiments were approved by the Institutional Animal Care and Use Committee of the International Vaccine Institute (Approval No: PN 1003), and all experiments were carried out in strict accordance with the Guide for the Care and Use of Laboratory Animals, Institute of Laboratory Animal Resources Commission on Life Sciences National Research Council, USA.All experiments were performed under anesthesia with a mixture of ketamine (100 mg/kg) and xylazine (20 mg/kg), and all efforts were made to minimize suffering.

Sample preparation
To obtain BALF, tracheas were cannulated after exsanguination and lungs were washed with 1 ml of PBS.BALF samples were centrifuged (8006g, 5 min) to isolate cells and supernatants were centrifuged again (13,0006g, 1 min) to completely remove remaining cells.BM cells obtained from femurs and tibias and red blood cells were removed before analysis.In some experiments, cells were cultured in vitro (1610 5 cells/well) for 4 h in RPMI (Gibco, Auckland, New Zealand) supplemented with 10% FBS (Gibco) to measure chemokines.Cultured cells were removed from supernatants by centrifugation (2,3006g, 3 min) and supernatants were used for further analysis.

Virus plaque titration
Total lung was removed and homogenized to prepare lung extracts in 1 ml of PBS (pH 7.4).Confluent Madin-Darby canine kidney (MDCK) cells were washed with MEM (Gibco) once and treated with virus for 30 min at room temperature.After a wash with MEM, the plate was overlaid with MEM containing 1% lowmelting-point agarose and 10 mg/ml of trypsin and incubated at 37uC for 3 days.

Measurement of total protein in BALF
We measured total protein in BALF samples by BCA Protein Assay Kit (Pierce, Rockford, IL) according to the manufacturer's instructions.

Cytokine and chemokine detection
The levels of MCP-1, IL-6, TNF-a, and IFN-c were measured by Mouse Inflammatory Cytometric Bead Array Kit (BD Biosciences, San Jose, CA).The levels of KC, MIP-2, and IP-10 were measured by DuoSet Mouse ELISA Kit (R&D Systems, Minneapolis, MN) according to the manufacturer's instructions.

Histology
Lungs were removed from naı ¨ve or infected B6 and Ifnar1 2/2 mice and washed using PBS before being fixed with 4% formaldehyde for 1 h at 4uC.The tissues were embedded in paraffin and stained with H&E.To detect MPO expression, tissues were dehydrated in sucrose solutions (10, 20, and 30%) after fixation and embedded in OCT compound (Sakura Finetec, Tokyo, Japan).Cryo sections (5 mm) were fixed in ice-cold acetone and blocked with FcRII/III mAb (2.4G2; BD Pharmingen, San Jose, CA) in PBS.Then, tissues were stained with FITCconjugated anti-MPO (2D4; Abcam, Cambridge, MA) for confocal microscopy.Histopathological score was assessed by a pathologist using a blind test.As previously described [63], we used a scoring system of 20 points to evaluate the level of lung tissue destruction, epithelial cell layer damage, polymorphonuclear cell infiltration into the site, and alveolitis.

Cytospin and Nile red staining
To cytospin cells on Cytoslide (Thermo Scientific, Asheville, NC), sorted cells were centrifuged at 1,000 rpm for 10 min using CytoSpin 4 Cytocentrifuge (Thermo Scientific).Then cells were fixed and stained with H&E.For Nile red staining, stock solution (Sigma, St. Louis, MO; 0.1 mg/ml in acetone) was diluted 1:5,000 in PBS and cells were stained for 30 min at 37uC.Samples were washed twice with Ca 2+ /Mg 2+ -free HBSS and cytospun.Then fixed specimens (3.7% formaldehyde) were stained with DAPI and washed twice before mounting.

BM cell in vitro stimulation
Cells were prepared from BM of naı ¨ve WT and Ifnar1 2/2 mice.After being washed twice with RPMI, cells (1610 7 ) were either infected with PR8 (5610 6 pfu/3 ml) virus or mock infected for 30 min at room temperature.Cells were washed in RPMI twice and cultured for 5 days in a 1:1 mixture of PBS and RPMI containing 10% FBS in culture dishes (Nunc, Roskilde, Denmark).In some groups, we used BALF from infected WT mice for stimulation.To inhibit IFN-I signaling, we used anti-IFNAR1 blocking antibody (100 ng/ml; MAR1-5A3; BioLegend, San Diego, CA).

Microarray analysis
Monocytes were sorted from lung of WT and Ifnar1 2/2 mice at 5 dpi.RNA from each cell subset was extracted by RNA Isolation Kit (Qiagen, Valencia, CA).cDNA microarray analysis was performed using a MouseRef-8 v2 Expression Beadchip Kit (Illumina, Inc., San Diego, CA).

Statistics
We used a paired two-sample t-test for analysis, except for survival data for which we used Kaplan-Meier analysis.*p,0.05,**p,0.01,and ***p,0.001were considered significant.

Figure 1 .
Figure 1.Ifnar1 2/2 mice are more susceptible to influenza infection than wild-type (WT) mice.(A) Survival and weight loss of WT and Ifnar1 2/2 mice after intranasal (i.n.) infection with 1610 5 pfu or 2610 4 pfu PR8 virus.(B) Viral load was measured by plaque assay in the lung of mice infected with 1610 5 pfu PR8 virus at 2 (n = 6) and 5 (n = 11) days post infection (dpi).Total protein (C) and cytokine (D) levels were determined in the BALF of PR8-infected mice at 3 and 5 dpi.(C, D) Values are mean + SD and results are representative of three independent experiments with n.5 for each group.Histopathological scores (E) and representative photos (F) of lungs from WT and Ifnar1 2/2 mice (3 mice/group) after infection with 1610 5 pfu PR8 virus and staining with H&E or MPO.Bars: 100 mm or 25 mm (enlarged pictures).doi:10.1371/journal.ppat.1001304.g001

Figure 2 .
Figure 2. Ifnar1 2/2 mice develop different cell population profiles in the lung after influenza infection.Mice were infected i.n. with 1610 5 pfu PR8 virus.(A) Cells were obtained from BALF at 1, 3, and 5 days post infection (dpi) and cellular compositions were determined by flow cytometry.AM, alveolar macrophages; N, neutrophils; DP, double-positive cells; Mo, monocytes.Numbers indicate cell population percentages within gates.(B) Absolute cell numbers of each subset in the BALF of mice pre and post infection.(C) CD11c and Ly6C expression levels in the monocytes isolated from the BALF of WT and Ifnar1 2/2 mice at 5 dpi.MFI; mean fluorescence intensity.Indicated cell subsets were isolated from the lung of WT and Ifnar1 2/2 mice at 5 dpi and stained with H&E (D) or Nile red (E) after cytospin.Original magnification 640.(F) Recipient WT and Ifnar1 2/2 mice were lethally irradiated and reconstituted with bone marrow (BM) cells from Ifnar1 2/2 and WT mice, respectively.Cell populations in the BALF isolated from recipient WT and Ifnar1 2/2 mice at 3 dpi were analyzed.Values are mean + SD and results are representative of three independent experiments with n.5 for each group.doi:10.1371/journal.ppat.1001304.g002

Figure 3 .
Figure 3. Ifnar1 2/2 mice produce significantly lower MCP-1 and higher KC than WT mice.(A) Levels of MCP-1, KC, and MIP-2 were determined in the BALF of PR8 virus (1610 5 pfu)-infected WT and Ifnar1 2/2 mice at 1, 3, and 5 dpi.Values are mean + SD and results are representative of three independent experiments with n.4 for each group.(B) Expression patterns of chemokine receptors, CXCR2 and CCR2, on monocytes (Mo) and neutrophils (N) were analyzed by flow cytometry.Histograms are representative of two independent experiments and numbers indicate MFI.(C) Monocytes (Mo) and alveolar macrophages (AM) were isolated using FACS Aria and plated onto 96-well plates.After culture for 4 h, MCP-1 and KC levels in the supernatants were analyzed.Data are representative of three independent experiments and values are mean + SD of three independently cultured wells.(D) mRNA expression levels of MCP-1 and KC were analyzed by gene chip data using monocytes from WT and Ifnar1 2/2 mice 5 dpi.The expression levels were normalized by GAPDH expression levels.Data are mean values of two independent experiments.doi:10.1371/journal.ppat.1001304.g003

Figure 4 .
Figure 4. Monocytes of Ifnar1 2/2 mice express different patterns of genes.WT and Ifnar1 2/2 mice were infected i.n. with 1610 5 pfu PR8 and monocytes were sorted.Total RNA was extracted and gene expression profiles were analyzed on gene chip.Data are representative of two independent experiments.doi:10.1371/journal.ppat.1001304.g004

Figure 5 .
Figure 5. Ifnar1 2/2 bone marrow (BM) cells are unable to differentiate into MCP-1-producing Ly6C hi monocytes.(A) BM cells were isolated from WT and Ifnar1 2/2 mice pre-and post-influenza infection and analyzed by flow cytometry.(B) Monocytes (Mo) and neutrophils (N) were isolated, cytospun, and stained by H&E (640).(C) MFI of Ly6C expression of monocytes in the BM of WT and Ifnar1 2/2 mice following influenza infection (1610 5 pfu).Mean values are in parenthesis.(D) BM cells from naı ¨ve WT and Ifnar1 2/2 mice were prepared and stimulated in vitro with BALF from PR8 virus (1610 5 pfu)-infected WT mice or with PR8 virus (0.5 multiplicity of infection) alone.After 5 days of culture, subsets of Ly6C hi and Ly6C int were counted and the percentage of Ly6C hi among Ly6C pos was determined.(E) After BM cells were prepared from naı ¨ve WT mice, cells were infected with PR8 virus and cultured for 5 days with/without anti-IFNAR1 blocking Ab (100 ng/ml).(F) BM cells infected with PR8 and cultured in vitro for 5 days.Cultured cells were further sorted into Ly6C neg , Ly6C int , and Ly6C hi .Each cell subset was cultured in vitro for 4 h and chemokine levels in the supernatant were analyzed.Values are mean + SD and data are representative of at least two independent experiments.doi:10.1371/journal.ppat.1001304.g005

Figure 6 .
Figure 6.IFN-I signaling on hematopoietic cell lineage helps protect mice from lethal influenza infection.Bone marrow (BM) cells from naı ¨ve WT (W) or Ifnar1 2/2 (K) mice were grafted to lethally irradiated WT or Ifnar1 2/2 recipient mice.(A-C) Sets of BM grafted mice were challenged with influenza PR8 virus (1610 5 pfu) and analyzed at 3 dpi.(A) Number of monocytes (Mo) and neutrophils (N) were determined.(B) Expression levels of MCP-1 and KC were measured in BALF.(C) Monocytes in the lung were pooled and cultured for 4 h to measure chemokine production in vitro.Values are mean + SD of triplicate experiments with n.4 for each group.(D) Mice were challenged with lethal dose of influenza PR8 (1610 5 pfu) and survival was monitored for 2 weeks.WRW (n = 17), KRW (n = 16), WRK (n = 16), and KRK (n = 11).Data were pooled from three independent experiments.doi:10.1371/journal.ppat.1001304.g006

Figure 7 .
Figure 7. Suggested model for the role of IFN-I on monocyte and neutrophil regulation.After influenza infection, hematopoietic stem cell differentiation might be triggered by IFN-I and then MCP-1producing Ly6C hi monocytes are recruited to the lung.In the absence of IFN-I signaling, Ly6C int monocytes are alternatively generated and produce KC, which might in turn aggravates lung pathology by attracting excess neutrophils to lung.doi:10.1371/journal.ppat.1001304.g007