Inflammatory Monocytes Drive Influenza A Virus–Mediated Lung Injury in Juvenile Mice

Influenza A virus (IAV) is a highly contagious RNA virus that can infect the respiratory tracts of both humans and animals. Each year, IAV infection affects up to 40% of children in the United States and is responsible for the hospitalization of ∼1/1000 children <5 y of age (1, 2). These children often develop lower respiratory tract infections, which may then progress to severe, life-threatening disease. In the 2009 H1N1 pandemic, the pediatric population accounted for 33% of all severe respiratory tract infections (3). It is notable that in the United States, the number of IAV-related deaths in previously healthy children is similar to the number of deaths in children with high-risk conditions, including lung disease and neurologic disorders. In contrast, the majority of adults who develop severe IAV infections have predisposing conditions that place them at increased risk of death (1, 4, 5). Although it is tempting to ascribe the increased susceptibility to severe IAV to the size of the pediatric airways and/or to differences in viral burden, there is little evidence that these are the factors responsible for the morbidity and mortality in this age group (6). In fact, several studies have shown that viral burden is similar between children and adults infected with IAV and that viral titer does not correlate with severity of illness (7–10). Instead, the host inflammatory response to IAV has been implicated in IAV-induced lung injury (9, 11–14). However, the mechanisms by which age-dependent differences in the innate immune response to IAV may influence the development of lung injury in infected children remain unknown (2, 6).

The innate immune response to IAV begins in the respiratory epithelium, which is the primary target of IAV. Detection of IAV in these cells by pattern recognition receptors leads to the production of antiviral IFNs and inflammatory cytokines. Type I IFNs, including IFN-α and -β, lead to the transcription of IFN-stimulated genes that aim to eliminate the virus and prevent its spread by promoting an antiviral state in nearby cells. Type I IFNs must be carefully regulated to maximize viral clearance while inflicting minimal damage to host cells. Excessive IFN-α/β has been implicated in lung injury and death in severe IAV infection (14, 15). In addition to halting normal transcription, which can damage the host cell, type I IFNs amplify the production of MCP-1, the primary chemokine responsible for recruiting inflammatory monocytes to the lungs during infection (16–18). These monocytes have been implicated in IAV-induced lung injury (13, 19). Importantly, elevated MCP-1 levels have been associated with severity of illness in pediatric IAV infection (20).

Together with alveolar macrophages, monocytes recruited by MCP-1 coordinate the inflammatory response to IAV, in part through activation of the NOD-like receptor family, pyrin domain–containing 3 (NLRP3) inflammasome (21, 22). This inflammasome consists of a pattern recognition receptor (NLRP3), an adaptor protein (apoptosis-associated speck-like protein [ASC]), and an effector protein (caspase-1). Activation of the NLRP3 inflammasome leads to processing of the potent proinflammatory cytokines of the IL-1 family, including IL-1β and IL-18, into their mature and active forms. Pro–IL-18 is constitutively expressed, but a priming signal is required for the production of pro–IL-1β (23). In this first step, detection of IAV by extracellular or cytosolic pattern recognition receptors activates the transcription factor NF-κB and leads to the production of pro–1L-1β. A second signal is then required for inflammasome assembly and activation. Both of the IAV proteins, M2 and PB1-F2, may serve as this activation signal (24, 25). Following this second signal, the intermediate filament, vimentin, acts as a scaffold upon which the components of the NLRP3 inflammasome converge (26). This convergence leads to activation of caspase-1 and cleavage of the proforms of IL-1β and IL-18 into their mature forms. Early NLRP3 inflammasome activity is protective in lethal models of IAV infection in adult mice (21, 22, 27). In contrast, excessive NLRP3 inflammasome activation may exacerbate lung injury and contribute to mortality during IAV infection (12, 14, 25). Therefore, similar to type I IFN signaling, NLRP3 inflammasome activation must be tightly regulated to achieve pathogen clearance while minimizing collateral damage to host tissues. Notably, type I IFNs have been shown to increase NLRP3 inflammasome activation (28).

Increased mortality and pulmonary pathology have been reported in juvenile mice infected with IAV compared with adults (15, 29), but the mechanisms underlying this susceptibility to acute lung injury are unknown. To study the differences between juvenile and adult responses to IAV, we developed a juvenile mouse model of IAV lower respiratory tract infection (12). We report here that IAV-induced lung injury in juvenile mice is characterized by excessive production of type I IFNs, increased MCP-1 secretion, prolonged inflammatory monocyte and monocyte-derived cell (Ly6C⁺CD64^+/−) recruitment, and sustained NLRP3 inflammasome activation as compared with IAV-induced lung injury in adult mice. RNA sequencing was performed to assess the transcriptional signatures of recruited monocytes to the juvenile and adult IAV-infected lungs. Genes associated with a proinflammatory signature were upregulated in the juvenile monocytes compared with adult monocytes. Importantly, using anti-CCR2 Abs, we prevented the recruitment of Ly6C⁺CD64^+/− monocytes to the lungs of IAV-infected juvenile mice, which resulted in decreased type I IFNs, decreased NLRP3 inflammasome activation, and increased survival of juvenile mice. Collectively, these findings demonstrate a key role for Ly6C⁺CD64^+/− monocytes and monocyte-derived macrophages in the pathogenesis of IAV-mediated lung injury in juvenile mice that appears to distinguish juvenile IAV infection from adult IAV infection.

All procedures complied with United States federal guidelines and were approved by the Institutional Animal Care and Use Committee at Northwestern University (protocol IS00001729).

129S wild-type mice were provided by Jackson Laboratories and bred in house.

Influenza virus strain A/WSN/1933 (WSN) was grown for 48 h at 37.5°C and 50% humidity in the allantoic cavities of 10–11-d-old fertile chicken eggs (Sunnyside Hatchery). Viral titers were measured by plaque assay in Madin-Darby canine kidney epithelial cells (American Type Culture Collection). Virus aliquots were stored in liquid nitrogen and freeze/thaw cycles were avoided.

Juvenile (4-wk-old) and adult (8–10-wk-old) mice were anesthetized with isoflurane and infected intratracheally (i.t.) with WSN (25 PFU) or an equal volume of PBS.

Mice were euthanized and lungs were perfused via the right ventricle with 10 ml HBSS with Ca²⁺ and Mg²⁺. A 20-gauge angiocatheter was sutured into the trachea, heart and lungs were removed en bloc, and then lungs were inflated with up to 0.8 ml of 4% paraformaldehyde (PFA) at a pressure not exceeding 16 cm H₂O. The heart and lungs were fixed in 4% PFA overnight at 4°C, processed, embedded in paraffin, sectioned (4 μm thickness), and stained with H&E by the Mouse Histology and Phenotyping Laboratory at Northwestern University (Chicago, IL). Images were acquired with TissueGnostics automated slide imaging system (TissueGnostics, Vienna, Austria) at the Northwestern University Center for Advanced Microscopy (Chicago, IL).

Mice were euthanized, and lungs were surgically removed en bloc. Lungs were weighed in a tared container. The lungs were then dried at 70°C in a SpeedVac SC100 evaporator (Thermo Scientific, Waltham, MA) until a constant weight was obtained, and the wet-to-dry weight ratio was calculated.

A 20-gauge angiocatheter was ligated into the trachea and the lungs were lavaged twice with sterile PBS (0.04 μl/kg of body weight). The bronchoalveolar lavage fluid (BALF) was centrifuged at 1000 × g for 10 min. The pellet was resuspended and the cells were counted using the Invitrogen Countess Automated Cell Counter (Invitrogen, Grand Island, NY). Protein levels in the supernatant were measured by Bradford assay (Bio-Rad) and cytokine levels were measured using ELISA. IL-18 was measured using the mouse IL-18 ELISA Kit (MBL International, Woburn, MA) according to the manufacturer’s instructions. Caspase-1 was measured using the mouse caspase-1 ELISA Kit (Adipogen, San Diego, CA). TNF-α and IL-6 were measured using the mouse TNF-α and IL-6 Ready-SET-Go! ELISA Kits (eBioscience, San Diego, CA). IFN-α was measured using the mouse IFN-α ELISA Kit (PBL Assay Science, Piscataway, NJ).

Mice were anesthetized with ketamine (45 mg/100 g of body weight) and a tracheotomy was performed. Pancuronium was administered i.p. (0.08 μg/kg of body weight). Each mouse was mechanically ventilated on a computer-controlled piston ventilator (flexiVent; SCIREQ, Montreal, Quebec, Canada) with a tidal volume of 10 ml/kg at a frequency of 150 breaths/min against 2–3 cm H₂O positive end-expiratory pressure. Pressure and flow data were collected during a series of standardized volume perturbation maneuvers. These data were used to calculate both total lung resistance and elastance using the single-compartment model. To calculate specific compliance, the static compliance for each mouse was normalized to its inspiratory capacity, which was measured by the flexiVent.

For plaque assay, lungs were homogenized in PBS (20 μl/mg lung). For Western blot, lungs were homogenized in RIPA buffer with protease inhibitor (20 mM Tris-HCl, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, Roche cOmplete ULTRA Tablet). Homogenized lungs were centrifuged at 1000 × g. The supernatant was frozen at 80°C.

Confluent monolayers of Madin-Darby canine kidney epithelial cells were infected with stock virus or lung homogenate serially diluted in 1% BSA DMEM for 2 h at 37°C. Plates were washed with PBS and an overlay of 50% 2× replacement media (2× DMEM, 0.12M NaHCO₃, 2% penicillin-streptomycin, and 1% HEPES), 50% Avicel (2.35%), and N-acetyl trypsin (1.5 μg/ml) remained on the cells for 72 h at 37°C. Plates were washed with PBS and fixed with 0.2% PFA before staining the monolayers with naphthalene blue-black stain.

Lungs were prepared in the same manner as for histology (see above). Sections were deparaffinized, rehydrated, and subjected to heat-induced Ag retrieval in a 10 mM solution of sodium citrate with 0.05% Tween 20 for 20 min, followed by a 20 min cooldown to room temperature. Lung specimens were then immunostained for IAV nonstructural protein 1 (NS1) using a polyclonal rabbit Ab (cat. no. GTX125990; GeneTex, Irvine, CA) at 1:500 dilution with a Vectastain ABC kit (cat. no. PK-4001), an avidin/biotin blocking kit (cat. no. SP-2001), and a 3,3'-diaminobenzidine peroxidase substrate kit (cat. no. SK-4100), all from Vector Laboratories (Burlingame, CA) and used according to the manufacturer’s protocols. Images were acquired with a Nikon Eclipse E800 microscope with Nikon DS-Fi2-U3 camera using NIS-Elements BR software (Nikon Instruments, Melville, NY).

The presence of indicated proteins in lung homogenates from day 5 postinfection (p.i.) was assessed by Western blotting using the following Abs: anti-NLRP3 (Adipogen), anti–caspase-1 (14F468) (sc-56036; Santa Cruz), anti-ASC (Adipogen), anti–IL-18 (5180R-10; Biovision), and anti-actin (sc-1615; Santa Cruz). Signals were detected following incubation with IRDye secondary Abs (1:10,000; LI-COR Biosciences) for 2 h at room temperature using the LI-COR Odyssey Fc Imaging System (Table I).

Approximately 150 μl of blood was collected from the submandibular vein in an EDTA-coated tube (BD, Franklin Lakes, NJ) for flow cytometry analysis or in a Microtainer tube with serum separator (BD) for ELISA analysis. MCP-1 was measured in the serum using the Mouse/Rat MCP-1 Quantikine ELISA Kit (R&D Systems).

Mice were euthanized and lungs were perfused via the right ventricle with 10 ml HBSS with Ca²⁺ and Mg²⁺. The lung lobes were removed and inflated with enzyme solution (5 ml of 0.2 mg/ml DNase I and 2 mg/ml collagenase D in HBSS with Ca²⁺ and Mg²⁺) using a 30G needle. The tissue was minced and then processed in gentleMACS dissociator (Miltenyi) according to the manufacturer’s instructions. Processed lungs were passed through a 40-μm cell strainer and RBCs were lysed with BD Pharm Lyse (BD Biosciences, San Jose, CA). Remaining cells were counted with a Countess cell counter (Invitrogen, Grand Island, NY). CD45 microbeads were added and cells were eluted according to Miltenyi manufacturer’s instructions. Cells were stained with viability dye Aqua (Invitrogen) and unspecific Ab binding was inhibited by adding Fc Block (553142, clone 2.4G2; BD Pharmingen). Blocked cells were stained with a mixture of fluorochrome-conjugated Abs (see Tables II and III for lists of fluorochromes, Abs, manufacturers, and clones, as applicable). Data were acquired on a BD LSR II flow cytometer using BD FACSDiva software (BD Biosciences), and data analyses were performed with FlowJo software (Tree Star, Ashland, OR). Cell populations were identified using sequential gating strategy, and the percentage of cells in the live/singlets gate was multiplied by the number of live cells to obtain an absolute live cell count. Cell number was normalized to dry lung weight. The expression of activation markers is presented as median fluorescence intensity (MFI).

Serum was stained with a mixture of fluorochrome-conjugated Abs (see Table IV for lists of fluorochromes, Abs, manufacturers, and clones). Data were acquired on a BD LSR II flow cytometer using BD FACSDiva software (BD Biosciences) and data analyses were performed with FlowJo software (Tree Star). Cell populations were identified using sequential gating strategy.

On day 2 p.i, IAV-infected juvenile mice (4-wk-old) were anesthetized with isoflurane, and 50 μl of anti-CCR2 hybridoma supernatant (clone: MC21, 1.83 μg/μl) was administered retro-orbitally (30).

Following an i.t. infection with IAV, lungs were harvested, enzymatically digested, and enriched for CD11c. Using a sequential gating strategy and FACS, recruited monocytes were isolated based on their surface marker expression of CD45⁺, Ly6C⁺, CD64⁺, Siglec F⁻, and CD11b⁺. Sorted cells were lysed in RLT lysis buffer (Qiagen) and samples were stored at −80°C until all time points were collected. Total RNA was isolated using the RNeasy plus mini kit (Qiagen) and RNA quality and quantity was assessed using a TapeStation 4200 (Agilent). mRNA was isolated from all samples using poly(A) enrichment (NEBNext). RNA-seq libraries were prepared using NEBNext RNA Ultra chemistry (New England Biolabs) and sequenced on an Illumina NextSeq 500 using a 75-cycle single-end high-output sequencing kit. Resulting libraries had an average read depth of 8 million reads, with a range of 4–12 million reads. Reads were demultiplexed (bcl2fastq) and fastq files were aligned to the mm10 mouse genome using TopHat2; counts were obtained using Htseq. Data were normalized and filtered for low counts using edgeR package. Differential expression analysis was performed using edgeR and differentially expressed genes were selected using a 1.0 fold change cutoff for pairwise comparisons and volcano plots. Following ANOVA and Benjamini–Hochberg adjustment, k-means clustering was run on remaining genes (false discovery rate [FDR] <0.05) using Gene-E (https://software.broadinstitute.org/GENE-E). Enrichment analysis was performed using GOrilla (31). Gene expression data are available from Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra), accession numbers SRR6369110, SRR6369111, SRR6369109, SRR6369121, SRR6369135, SRR6369134, SRR6369120, SRR6369289, SRR6369288, SRR6369345, SRR6369348, SRR6369347, SRR6369350, and SRR6369349. https://trace.ncbi.nlm.nih.gov/Traces/study/?acc=SRP126565

Data are expressed as mean ± SD. Differences between two groups were assessed using a Student t test. Differences between three or more groups were assessed using two-way ANOVA with the Sidak multiple comparisons test. The p values < 0.05 were considered to be significant. The log-rank test was used in the analysis of the Kaplan–Meier curve. All analyses were performed using GraphPad Prism software version 6.0 for Windows (GraphPad Software, San Diego, CA).

To assess whether naive juvenile mice have increased susceptibility to IAV infection as compared with naive adult mice, juvenile (4-wk-old) and adult (10-wk-old) mice were infected i.t. with IAV (25 PFU WSN). IAV-infected juvenile mice displayed a ∼2-fold increase in mortality compared with IAV-infected adult mice. Juvenile mice had a median survival of 7.5 d p.i., whereas adult mice had a median survival of 13 d p.i. All of the juvenile mice died by 9 d p.i., whereas >50% of the adult mice survived the infection (Fig. 1A). Weight loss, lethargy, coat ruffling, and febrile shaking were observed in both juvenile and adult mice, but were transient in the adult survivors (Fig. 1B). A common misbelief is that juvenile mice would have increased mortality simply because they are smaller than adult mice, so we examined the survival of juvenile and adult mice infected with an equivalent amount of IAV per gram of body weight (PFU/g). When mice were given 1 PFU/g, we again found excess mortality in the juvenile mice. Even when adult mice were infected with 2 PFU/g and compared with juvenile mice infected with 1 PFU/g, juvenile mice continued to demonstrate worse survival (Supplemental Fig. 1A). Increased mortality was consistently seen in juvenile mice compared with adult mice regardless of inoculation dose of IAV (Supplemental Fig. 1B–E). At each infectious dose of IAV administered, fewer juvenile mice survived the infection than adult mice.

To determine whether a failure to control IAV infection was responsible for the increased mortality observed in juvenile mice compared with adult mice, we examined viral titers in the lungs of juvenile and adult mice over the course of IAV infection. Plaque assays performed on the lung homogenates of juvenile and adult mice 3, 5, and 7 d p.i. with IAV failed to show a difference in viral titers between the two age groups (Fig. 1C). Immunohistochemistry for IAV NS1 in the lungs of juvenile and adult mice 5 d p.i. supported this finding of equivalent viral burden in both age groups (Fig. 1D).

IAV infection can result in acute lung injury, with increased alveolar epithelial permeability, leakage of protein containing fluid into the alveolar space, and infiltration of immune cells into the lungs, which ultimately lead to impairment in gas exchange and respiratory failure (2, 32). Histologic examination of the lungs of IAV-infected juvenile and adult mice 7 d p.i. revealed more alveolar damage and infiltrating immune cells in the lungs of juvenile mice as compared with age-matched, saline control–treated juvenile mice and IAV-infected adult mice (Fig. 2A, 2B). In accordance with the histological demonstration of lung injury, IAV-infected juvenile mice displayed an increase in the lung wet-to-dry weight ratio as well as a 5-fold increase in the total protein content in BALF that was not observed in IAV-infected adult mice (Fig. 2C, 2D). BALF from IAV-infected juvenile mice also contained more leukocytes than BALF from IAV-infected adult mice (Fig. 2E).

Consistent with the more significant lung injury seen in juvenile mice, lung mechanics were altered in juvenile mice infected with IAV, but not in adult mice. Pulmonary specific compliance was decreased in IAV-infected juvenile mice compared with saline-treated controls 7 d p.i, but was similar between saline-treated and IAV-infected adult mice (Fig. 2F). As expected, baseline resistance was increased in uninfected juvenile mice compared with adult mice, because of the smaller size of the juvenile airways. Infection with IAV did not change resistance in adult mice, but resulted in a considerable increase in resistance in juvenile mice (Fig. 2G). These findings of decreased compliance and increased resistance in juvenile mice 7 d p.i. with IAV are consistent with the increased incidence of acute lung injury, respiratory failure, and death seen in juvenile mice compared with adult mice. Combined, these data indicate that healthy juvenile mice develop a severe respiratory tract infection and acute lung injury following IAV infection that resulted in increased recruitment of inflammatory cells and increased mortality, which was not observed in healthy adult mice infected with IAV.

The innate immune response to IAV comprises an antiviral response, orchestrated by type I IFNs, and an inflammatory response, coordinated by NLRP3 inflammasome activation. The antiviral activity of the type I IFNs is primarily responsible for control of viral replication and promotion of viral clearance. But type I IFNs can also enhance the inflammatory response by increasing the production of MCP-1, the chemokine that recruits inflammatory monocytes to the lungs during infection (16–18). Measurement of type I IFNs in the BALF of IAV-infected juvenile and adult mice revealed that juvenile mice had twice the IFN-β levels as adult mice on day 5 p.i. (219 ± 42 pg/ml versus 114 ± 15 pg/ml, respectively) (Fig. 3A). Additionally, IAV-infected juvenile mice had increased levels of MCP-1 in both serum and BALF on day 7 p.i., which was not observed in IAV-infected adult mice (Fig. 3B, 3C). Importantly, these increases in type I IFN and MCP-1 were present in BALF of IAV-infected juvenile mice despite similar viral titers in juvenile and adult mice (see Fig. 1C). Therefore, the increased mortality seen in juvenile mice compared with adult mice during IAV infection was not explained by impaired viral clearance in juvenile mice, but was associated with increased type I IFNs in BALF and increased MCP-1 in BALF and serum.

Recognition of IAV infection by extracellular or cytosolic pattern recognition receptors leads to activation of NF-κB, production of proinflammatory cytokines, and activation of the NLRP3 inflammasome (11). TNF-α and IL-6 are among the proinflammatory cytokines secreted in response to IAV infection. Increased levels of TNF-α and IL-6 correlate with severity of symptoms and degree of lung injury following infection with IAV (9, 20, 33, 34). In adult mice, TNF-α levels were elevated 3 d p.i. (42 ± 37 pg/ml), peaked at 5 d p.i. (171 ± 25 pg/ml), and returned toward baseline at 7 d p.i. (57 ± 32 pg/ml). In contrast, TNF-α levels were elevated 3 d p.i. (204 ± 114 pg/ml) and remained elevated at 5 and 7 d p.i (175 ± 34 and 187 ± 40 pg/ml) in juvenile mice (Fig. 4A). Similarly, IL-6 levels continually increased in juvenile IAV-infected mice over the course of infection, but not in adult mice (Fig. 4B, Supplemental Fig. 2). Combined, these findings suggest that juvenile mice infected with IAV exhibit a robust activation of NF-κB–dependent proinflammtory cytokine production.

We next evaluated activation of the NLRP3 inflammasome in juvenile and adult IAV infection. We began by performing Western blot analysis of lung homogenates to compare the inflammasome components, NLRP3, ASC, and procaspase-1, in juvenile and adult mice. At baseline, juvenile mice had increased levels of NLRP3 protein expression in lung homogenates compared with adult mice. Following infection with influenza, both juvenile and adult mice had a 2.5-fold increase in NLRP3 protein levels (Fig. 4C). ASC levels were similar in the two age groups prior to IAV infection; however, ASC levels increased 3-fold in juvenile mice whereas they remained unchanged in adult mice infected with IAV (Fig. 4D). Procaspase-1 protein levels were similar between juvenile and adult mice at baseline and unchanged in response to IAV infection (Fig. 4E).

Once activated, the components of the NLRP3 inflammasome assemble into a multiprotein complex, which induces self-cleavage of caspase-1, that then mediates the cleavage of pro–IL-1β and pro–IL-18 into biologically active, mature forms. We evaluated NLRP3 inflammasome activation in juvenile and adult mice infected with IAV by measuring mature caspase-1 and mature IL-18 in the BALF. We found increasing levels of both mature caspase-1 and mature IL-18 in juvenile mice over the course of infection. IL-18 levels increased from undetectable to 541 ± 169 pg/ml between 0 and 7 d p.i. in the juvenile mice, whereas IL-18 levels were only 74 ± 49 pg/ml 7 d p.i. in adult mice. Similar results were observed for mature caspase-1 levels (Fig. 4F, 4G). This was despite falling viral titers over the same time period (see Fig. 1E). In accordance with this, measurement of mature caspase-1 and mature IL-18 in the lung homogenates of juvenile and adult mice by Western blot on day 7 p.i. showed increased mature caspase-1 and a trend toward increased mature IL-18 in juvenile mice in response to IAV (Fig. 4H, 4I). Taken together, the increase in NLRP3 and ASC in juvenile mice demonstrate increased NLRP3 inflammasome components in juvenile IAV infection. In addition, the increase in mature caspase-1 and mature IL-18 in IAV-infected juvenile mice suggest increased activation of the NLRP3 inflammasome in this age group.

In order to better understand which inflammatory cells may be contributing to the increase in inflammatory cytokines in IAV-infected juvenile mice, we performed a systematic flow cytometric analysis of whole-lung lysates (12, 35). The pan-hematopoietic marker CD45 was used first to identify immune cells in the lungs. Additional surface markers were then used to distinguish between neutrophils (Ly6G⁺, CD11b⁺, CD24⁺), eosinophils (CD11b⁺, CD24⁺, Siglec F⁺, CD11c⁻, CD64⁻), dendritic cells (CD11b^+/−, CD11c⁺, CD24⁺, MHC II⁺, CD64⁻), alveolar macrophages (CD64⁺, CD11c⁺, Siglec F⁺), steady state interstitial peribronchial macrophages (CD11b⁺, CD11c⁺, CD64⁺, Siglec F⁻, MHC II⁺, CD24⁻, Ly6C⁻) (36), and classical monocytes/monocyte-derived cells (CD11b⁺, Ly6C⁺, CD64^+/−) using 10-color flow cytometry (Supplemental Fig. 3A, data not shown) (28, 35). Monocyte-derived cells (MoDC) are virtually absent in the untreated, control animals (35), therefore we combined classical monocytes (Ly6C⁺CD64⁻) and MoDC (Ly6C⁺CD64⁺) during the analysis to enable comparison over baseline and referred to these as “recruited (Ly6C⁺CD64^+/−) monocytes/MoDC” (Supplemental Fig. 3B).

Saline-treated juvenile and adult mice had similar numbers of neutrophils and eosinophils in their lungs, but juvenile mice had significantly more Ly6C⁺CD64^+/− monocytes/MoDC and alveolar macrophages at baseline (Fig. 5A). In both age groups, alveolar macrophages decreased over the course of infection. In contrast, recruited Ly6C⁺CD64^+/− monocytes/MoDC steadily increased from day 5 to day 7 p.i. in both juveniles and adults. Notably, twice as many Ly6C⁺CD64^+/− monocytes/MoDC were recruited to juvenile lungs than adult lungs. These Ly6C⁺CD64^+/− monocytes/MoDC represented 45% of the inflammatory cells recovered from the lungs of IAV-infected juvenile mice on day 7 p.i. (data not shown). These results are consistent with our observation of increased MCP-1 in serum and BALF from IAV-infected juvenile mice (see Fig. 3B, 3C).

To determine whether recruited Ly6C⁺CD64^+/− monocytes/MoDC expressed the NLRP3 inflammasome, we performed intracellular staining for NLRP3 and caspase-1 on cells isolated from juvenile and adult mice infected with IAV by flow cytometry. Seven days p.i., individual recruited Ly6C⁺CD64^+/− monocytes/MoDC from juvenile mice had a 54% increase in MFI for NLRP3 and a 39% increase in caspase-1, as compared with individual recruited monocytes/MoDC from adult mice, which had a 30% and 37% increase in MFI for NLRP3 and caspase-1, respectively. (Fig. 5B–E).

It was unclear if the exacerbated lung injury seen in IAV-infected juvenile mice was related to the increased number of monocytes recruited to juvenile lungs, or if there were properties intrinsic to recruited juvenile monocytes contributing to IAV-mediated lung injury in juvenile mice. To further explore the possibility that these recruited monocytes were uniquely inflammatory in juvenile mice, we performed cell sorting, RNA isolation, and mRNA enrichment to obtain the expression profiles of MoDC isolated from juvenile and adult lungs. Comparative analysis of the juvenile and adult data sets by pairwise comparison revealed differences between the two groups in their response to IAV (Fig. 6A). The largest differences occurred on day 3 p.i., when 692 genes were downregulated and 149 genes were upregulated in juvenile versus adult mice. To further characterize these differences we performed k-means clustering on 5776 genes that were differentially expressed between all groups (FDR <0.05) (Fig. 6B). We went through several clustering iterations and selected k = 5. As expected, the inflammatory response was associated with downregulation of a large number of genes related to steady-state metabolic processes (cluster V), and clusters I–IV were related to the development of inflammatory responses (Fig. 6B, 6C). However, clusters I–IV revealed differential activation patterns between juvenile and adult MoDC. MoDC from adult mice exhibited early upregulation of genes involved in Ag processing (MHC II genes and Cd74) as well as molecular chaperones involved in adaptive unfolded protein response and response to stress (Fig. 6B, 6C, cluster I). In contrast, juvenile mice demonstrated early activation of the genes involved in negative regulation of Ag receptor–mediated signaling pathway (cluster II). Clusters III and IV were enriched with the genes involved in immune response to virus and inflammation and were shared between juvenile and adult mice. As anticipated, evaluation of the NLRP3 inflammasome genes, Nlrp3, Casp1, and Il18 did not reveal a difference at the transcriptional level between juveniles and adults (Fig. 6D). However, increased expression of genes that potentiate inflammasome signaling was seen in juveniles compared with adults, including Il18rap (IL-18 accessory protein) (37), Cmpk2 (UMP-CMP kinase 2, mitochondrial) (38) and Myd88 (primary response protein MyD88) (37). Next, we investigated whether juvenile and adult MoDC demonstrated more proinflammatory or reparative properties (36, 39, 40). Genes associated with a more proinflammatory signature (Cxcl10, Mmp8, Col3a1) (41, 42), were found to be upregulated in the juvenile compared with the adult MoDC following IAV infection (Fig. 7A). Additionally, many reparative genes (Il10, Retnla, Vsig4) were more downregulated in juvenile MoDC. Expression patterns for several genes of interest uncovered differences between juvenile and adult responses to IAV (Fig. 7B). Combined, these data suggest that the MoDC found in the juvenile murine lung adopt a more proinflammatory transcriptome over the course of IAV infection.

We hypothesized that recruitment of monocytes with a proinflammatory expression profile was contributing to the increased lung injury seen in juvenile IAV infection. Classical Ly6C⁺ monocytes are characterized by the high expression of the receptor for MCP-1: CCR2 (43). Therefore, we depleted Ly6C⁺ monocytes in juvenile mice with anti-CCR2 Ab on day 2 p.i. Treatment with anti-CCR2 Ab depleted circulating Ly6C⁺ monocytes (Fig. 8A) and prevented monocyte recruitment to the lungs of IAV-infected juvenile mice as demonstrated by flow cytometry (Fig. 8B). As previously observed, juvenile mice infected with IAV had significant weight loss over the course of infection and by 9 d p.i., 100% of the mice had died. Treatment with the isotype control Ab (rat IgG2b) did not protect juvenile mice from IAV-induced mortality (Supplemental Fig. 4). In contrast, juvenile mice treated with anti-CCR2 had marginal weight loss following IAV infection and only 20% mortality at 9 d p.i. (Fig. 8C, 8D). Prevention of monocyte recruitment did not impact viral clearance, as viral titers were similar in the homogenized lungs of untreated and anti-CCR2–treated juvenile mice (Fig. 8E). However, examination of BALF by ELISA for type I IFN showed markedly decreased type I IFN production in IAV-infected juvenile mice treated with anti-CCR2 (Fig. 8F). In addition, decreased inflammation and lung injury were seen in IAV-infected mice treated with anti-CCR2 as measured by protein leakage and cellular infiltration into BALF (Fig. 8G, 8H). Consistent with this, decreased caspase-1 and IL-18 levels were found in BALF of anti-CCR2–treated IAV-infected juvenile mice on day 7 p.i. (Figs. 8I, 8J), similar to those observed in adult IAV-infected mice at this time point (see Fig. 4F, 4G). Therefore, prevention of monocyte recruitment to the lungs during juvenile IAV infection is associated with decreased lung injury and decreased NLRP3 inflammasome activation, but does not impair viral clearance. A schematic representation of our findings is depicted in Fig. 9.

Innate immune responses to a variety of stimuli have been shown to differ between children and adults. It has been proposed that these differences impair antiviral responses in children and leave them particularly susceptible to severe viral infections (44–50). However, very little is known about how innate immune responses, including type I IFN production and NLRP3 inflammasome activation, might differ between children and adults infected with IAV. Recent data from natural IAV infection in children suggest an excessive, rather than an impaired, inflammatory response to IAV (9). Consistent with our findings in mice, viral titers were equal between children and adults, and did not correlate with severity of illness. This challenges the long-standing dogma that life-threatening IAV infection in children is due to an inability to control the infection. It may also explain the limited utility of antiviral medications in this age group (51).

IAV infection is much more likely to become life-threatening in healthy children than in healthy adults, but the mechanisms underlying this increased risk of IAV-induced lung injury and death in children are unknown (6). To investigate the increased burden of severe IAV infection in children, we compared IAV infection in 4-wk-old juvenile mice and in 10-wk-old adult mice. Combined, our data demonstrate that juvenile mice have an exaggerated inflammatory response to IAV infection that persists beyond viral clearance and is associated with increased lung injury and death.

We chose 4-wk-old mice for our model of pediatric influenza infection because lung development in this age group of mice roughly corresponds to the lung development of children between 6 mo and 5 y of age (52, 53), and these are among the children who are at highest risk of severe influenza infection (6). We corrected for lung size when analyzing the immune cell population in the lungs of juvenile and adult mice, as larger lungs would be expected to support higher numbers of these cells regardless of infection status and without necessarily impacting or reflecting disease progression (see Fig. 5A). When multiple inoculation doses were tested, severity of illness was consistently worse in juvenile mice, indicating that juvenile mice are at higher risk of severe IAV infection independent of infectious dose. This was true even when mice were dosed per gram of body weight (see Supplemental Fig. 1A). Notably, 4-wk-old mice are only ∼20% smaller than 10-wk-old adult mice, so size could not explain the increased mortality in this age group. In addition, plaque assay showed equal viral titer per volume of homogenized lung tissue in juveniles and adults and equal rates of clearance (see Fig. 1D). Perhaps most importantly, juvenile mice showed increasing Ly6C⁺CD64^+/− monocyte/MoDC recruitment (see Fig. 5A) and rising NLRP3 inflammasome activation (see Fig. 4F, 4G) even as viral titers were decreasing. Therefore, amount of virus alone could not explain the increased degree of injury in juvenile lungs.

The chemokine responsible for recruiting monocytes to the lungs, MCP-1, is one of the proinflammatory proteins that has been associated with severity of IAV infection in children (20). In support of this, we report increased MCP-1 in the serum and BALF of IAV-infected juvenile mice compared with IAV-infected adult mice, which correlates with the degree of lung injury in juvenile mice. Importantly, type I IFNs have been shown to amplify the release of MCP-1 during viral infection (17). We demonstrate excessive production of type I IFNs in IAV-infected juvenile mice, which may explain the elevation of MCP-1 in these animals (see Fig. 3). This increase in MCP-1 accounts for the increased presence of Ly6C⁺CD64^+/− monocytes/MoDC in the lungs of IAV-infected juvenile mice compared with IAV-infected adult mice (see Fig. 5A). Notably, these recruited Ly6C⁺CD64^+/− monocytes/MoDC are a primary source of NLRP3 inflammasome activity and the proinflammatory cytokines associated with IAV-related lung injury in our model of juvenile IAV infection (see Fig. 8I, 8J).

Deficiency of the NLRP3 inflammasome is harmful in adult models of IAV infection (21, 22, 54). However, when we depleted monocytes using anti-CCR2 Ab 2 d following infection, effectively inhibiting NLRP3 inflammasome activation late in juvenile IAV infection (see Fig. 8I, 8J), we found improved survival. This is consistent with recent reports that early inhibition of NLRP3 inflammasome activity in IAV infection is detrimental, whereas late inhibition is protective (12, 55). Our data extends this finding by implicating recruited monocytes as the source of late and harmful NLRP3 inflammasome activity.

Prevention of Ly6C⁺CD64^+/− monocyte recruitment in juvenile mice was also associated with decreased type I IFN in the later stages of infection, but did not impair viral clearance (see Fig. 8E, 8F). We reason that preserved type I IFN production early in infection enabled adequate viral clearance in these mice, whereas inhibition of late IFN secretion (day 5 p.i.) with Ly6C⁺ monocyte depletion prevented unnecessary perpetuation of the inflammatory response. Consistent with this, we report decreased lung injury in juvenile mice treated with anti-CCR2 (see Fig. 8G, 8H). In addition, we show that recruited Ly6C⁺CD64^+/− monocytes/MoDC are a primary source of NLRP3 inflammasome activity, as caspase-1 and IL-18 were decreased in the absence of recruited Ly6C⁺CD64^+/− monocytes/MoDC (see Fig. 8I, 8J). However, CCR2 is expressed by a small percentage of T cells at baseline, and its expression on T cells is increased in inflamed tissues, so it is possible that the effects of anti-CCR2 treatment were not limited to monocytes. We did not examine the impact of anti-CCR2 on T cell migration to the lungs during IAV infection, but in other models of inflammation, T cell recruitment was not altered by anti-CCR2 (56). In addition, the timing of anti-CCR2 administration at 2 d p.i. makes it less likely that the protection seen with this therapy was due to T cell responses alone, as these cells are expected to be prominent later in infection. Of note, the stimulus for late type I IFN production despite adequate viral clearance in juvenile IAV infection is unknown at this time. Furthermore, the mechanism by which inhibition of monocyte recruitment to the lungs prevents this late IFN response deserves further study.

To our knowledge, our examination of the gene expression profiles of monocytes recruited to the lungs during IAV infection is the first to suggest that juvenile recruited monocytes acquire an inflammatory phenotype that is distinct from adult recruited monocytes (see Fig. 7). In this study, we have highlighted the upregulation of some inflammatory genes and the downregulation of reparative genes that may be important in the juvenile response to IAV. Although we found increased amounts of mature caspase-1 and IL-18 proteins in juvenile lung homogenates and BALF, we were not surprised that these differences were not also found at the transcriptional level. These proteins require the NLRP3 inflammasome to be activated prior to conversion into their mature and active forms, so their gene expression would not be expected to predict their protein levels. We did, however, see elevation of some genes potentially involved in NLRP3 inflammasome activation (CMPK2), IL-18 receptor activity (IL-18rap), and downstream IL-18 signaling (MyD88) (38, 57). Other genes we found interesting included MMP-8, which has been associated with death from sepsis in children (58), and was markedly elevated in our juvenile recruited monocytes. We also found it interesting that members of the MHC II, H2-Ab1 and H2-Aa, showed a trend toward elevation in adult MoDC over juvenile MoDC. As these genes are involved in Ag presentation and the switch from innate to adaptive immunity (59), their deficit in juvenile MoDC could prolong the innate inflammatory response. We also identified IFN-stimulated genes that may differ between juveniles and adults (Mx1, Rsad2, and Ch25h). Future experiments will explore the contribution of the proinflammatory and reparative genes that comprised the transcriptomic fingerprint of juvenile recruited monocytes and may contribute to their injurious impact on juvenile lungs during IAV infection.

In this report, we establish that juvenile mice develop much more severe lung injury during IAV infection than adult mice, even though they are able to clear IAV from their lungs at the same rate as adult mice. To our knowledge, this study demonstrates for the first time that juvenile mice produce greater amounts of type I IFNs, secrete more MCP-1, recruit more monocytes to their lungs, and have increased activation of the NLRP3 inflammasome in response to IAV. Inhibition of monocyte recruitment to the lungs in juvenile mice ameliorated lung injury and inflammation, and implicated this cell type in type I IFN production, NLRP3 inflammasome activation, and alveolar epithelial barrier dysfunction. Future investigations into therapies directed at these pathways in juveniles are warranted.

We thank Jennifer Davis for acquiring and processing the microscopic images in Figs. 1 and 2. The illustration in Fig. 9 was created by Jacqueline Schaffer. Histology services were provided by the Northwestern University Research Histology and Phenotyping Laboratory. Imaging was performed at the Northwestern University Center for Advanced Microscopy. Flow cytometry was performed in the Northwestern University Flow Cytometry Core Facility.

The authors have no financial conflicts of interest.