Juvenile, but Not Adult, Mice Display Increased Myeloid Recruitment and Extracellular Matrix Remodeling during Respiratory Syncytial Virus Infection

Early life respiratory virus infection, including respiratory syncytial virus (RSV), has been linked to the subsequent onset of asthma and with decreasing air quality worldwide, and chronic pulmonary disorders are becoming an ever-increasing healthcare burden (1–5). RSV infects 37 million children annually (3, 4). As no vaccine is currently available, a better understanding of the initial immunological progression of this pathogen is critical to developing more specific targeted therapies (3–5). Much of the work defining the immune response to RSV infection has been performed in mice typically 6–8 wk old, with large doses of RSV (in the range of 1 × 10⁶–1 × 10⁷ PFUs) in BALB/C mice (6–10). These studies may, therefore, be missing the responses to early life RSV infection characteristic of a 1- to 3-y-old human, which would be better modeled in younger mice while observing the acute inflammatory contributions of type 1 immunity against the background of age-related lung development (11–14). Additionally, in vitro systems have been used to study the response of leukocytes and epithelial cells to RSV infection (15, 16). These studies showed that the CCL2, CCL3, and CCL4 (also known as MCP-1, MIP-1α, and MIP-1β, respectively) are drivers of RSV-related immunity, initially released by alveolar macrophages (AM) and the infected epithelium to recruit leukocytes, which then release additional cytokines and chemokines, and also produce factors that lead to extracellular matrix (ECM) modification (10, 15–17). Myeloid recruitment and subsequent cellular activity is at the heart of the lung pathology seen in influenza infection models, leading us to ask the following question: could RSV pathology also be driven by overzealous myeloid recruitment (17, 18)?

There are few data on age-related differences in response to RSV infection, with studies largely focused on responses in adult mice that had been infected as neonates or juveniles (8, 19). We directly tested whether young and old mice respond differently to infection, hypothesizing that younger mice would demonstrate age-related immunological difference(s) associated with RSV infection. We conducted a study to examine the initial myeloid recruitment into the lungs of RSV-infected 3-wk-old (hereafter referred to a juvenile) and 8-wk-old (hereafter referred to as adult) C57BL/6 (B6) mice to elucidate age-related differences. When compared with adult mice, we found that RSV-infected juvenile mice have sustained myeloid recruitment (including polymorphonuclear neutrophils [PMN], eosinophils [EO], and monocytes [MO]) in the infected lung, as well as resilience of the AM compartment, and maintenance of Ccl2, Ccl3, and Ccl4 RNA expression, the relevant chemokines for myeloid recruitment (17, 18). Juvenile mice demonstrated enhanced αSma expression, which is indicative of myofibroblast activity, and displayed increased hyaluronan (HA) deposition in the lung parenchyma, which has been attributed to asthma progression (2, 6, 20). To translate these murine findings toward human relevance, we used an RSV-infected ex vivo human cell culture system including human bronchial epithelial cells (BECs), human lung fibroblasts (HLFs), and the U937 human MO cell line. This human cell culture system supported these mouse data also revealed elevated expression of matrix metalloproteinase-7 (MMP7) and MMP9, both of which have been implicated in ECM pathology (21, 22). Taken together, these data support our hypothesis of age-related differences in the immune response to RSV infection and that, like influenza, MO can be a key driver in lung pathology, possibly contributing to the airway-remodeling characteristic of asthma.

Male and female 3- and 8-wk-old C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME) and housed at the Benaroya Research Institute’s (BRI’s) Biosafety Level-2 mouse facility, where they were given an intranasal inoculation of 40 μl of PBS or 40 μl of RSV (containing 1.5 × 10⁵ PFUs of Line-19 RSV) then euthanized with 1.5 ml of avertin at 18 or 72 h. All animal procedures were approved by the BRI’s Institutional Animal Care and Use Committee.

Bronchiolar lavage (BAL) was performed by collecting three separate samples of 0.6 ml from the juvenile mice and 0.8 ml from the adult mice, and the lavages were centrifuged at 1500 rpm for 5 min to separate the cells. After collection, the cells were treated in ammonium-chloride-potassium lysis buffer for 10 min at 4°C to remove the RBCs, then washed in PBS.

The right lung was perfused with 3 ml of PBS and removed for digestion with collagenase mix (isolated from Clostridium histolyticum; Sigma-Aldrich (St. Louis, MO)); the lung tissue was minced, then transferred to a 12-well plate containing 2 ml of the collagenase mix per sample, incubated at 37°C for 45–60 min, regurgitated through a 16-gauge needle five times to further break down the tissue, then pressed through a 40-μm filter, which was washed with PBS. The tissue slurry was centrifuged at 1500 rpm for 5 min, the supernatant was decanted, and the cells were collected, ammonium-chloride-potassium treated for 10 min at 4°C, and then washed in PBS.

The cells yielded from the BAL and lung were stained using fluorescent-labeled Abs at 4°C in the dark for 20 min, washed with PBS, and then examined on the BD Biosciences (San Jose, CA) LSR II platform in the BRI flow cytometry core.

Leukocytes were stained for CD11b, CD11c, CD45.2, CD64, CD103, Ly6C, Ly6G, MHC class II, SiglecF (all BioLegend), CD253 (TRAIL), and viability (eBioscience). The eBioscience (Thermo Fisher Scientific, Foster City, CA) Annexin V Apoptosis Detection Kit was used in accordance with the manufacturer’s instructions to determine which of the cells that absorbed the viability dye had undergone a programmed cell death process. See Supplemental Fig. 1 for the complete staining and pictorial gating scheme.

The gene expression of Ccl2, Ccl3, Ccl4, and αSma was determined from the RNA isolation of the upper-left lung lobe of each infected mouse. After PBS perfusion, the lung lobe was snap frozen on dry ice and RNA was later isolated using the Nucleospin RNA Isolation Kit (Takara Bio, Mountain View, CA), converted to cDNA, then quantified with TaqMan Primer/Probe sets from Applied Biosystems (a subsidiary of Thermo Fisher Scientific) to determine script copy quantities on the Thermo Fisher Scientific (Waltham, MA) QuantStudio-5 real-time quantitative PCR (qPCR) platform.

The lower-left lung was inflated with a 50/50 mixture of OCT-PBS and placed in 10% formalin for histological preparations by the BRI Histology Core, using a biotinylated HA-binding protein (prepared in-house), then counter-stained with streptavidin HRP label (Biocare Medical, Pacheco, CA). These preparations underwent HA area quantitation, as previously described, using the ImageJ software package (National Institutes of Health, Bethesda, MD) (20).

With approval from the Seattle Children’s Hospital Institutional Review Board and following informed consent (written consent was obtained from parents of subjects and assent was obtained from children equal to or older than age 10 y), primary human BECs were isolated from healthy, nonasthmatic pediatric donors (age 6–18 y) undergoing elective surgical procedures requiring endotracheal intubation at Seattle Children’s Hospital as previously described (20). BECs were expanded in submerged culture and then passaged into transwells and differentiated at an air-liquid interface for 3 wk (passage ≤3). Cocultures were established using commercially available pediatric HLFs (Lonza, Walkersville, MD), as previously described (23, 24). BECs were infected with RSV L19 at a multiplicity of infection of 1, as described previously (25). U937 cells, a human MO cell line, (American Type Culture Collection, Manassas, VA) were placed in the basal chamber of the transwell coculture system 24 h after BEC RSV infection and then harvested at the designated time-points. Duplicate wells were run for each isolate with RNA from the U937 cells collected using the Qiagen RNA Isolation Kit (Germantown, MD) with the RNA for each duplicate well consolidated before cDNA generation. Molecular targets for: CCL2, CCL3, CCL4, MMP7, MMP9, TSG6, HAS3, and HYAL1 were analyzed as described in the “Whole–lung RNA isolation and quantitative PCR analysis” section above.

All statistical analysis was performed using Prism (GraphPad Software) version 8.3.1; all results depict unpaired t tests (mouse-based experiments) or paired t tests (human-based cell cultures), with statistical significance reported as follows: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

We determined the profile of myeloid cell recruitment into the airways and lung tissue following initial RSV infection of male and female juvenile and adult mice. Both age groups displayed an acute increase in total leukocytes in the BAL fluid (BALF) 18 h postinfection; however, despite the significantly greater initial cellular presence in juveniles, all groups displayed similar cellularity by 72 h (Fig. 1A). Neutrophilia, a classic sign of infection, was observed in the BALF of both age groups at 18 h postinfection (Fig. 1B); however, there was a significantly greater presence in the juvenile lungs, that resolved by 72 h (data not shown). Only the juvenile mice demonstrated EO recruitment at 18 h (Supplemental Fig. 2), which, like the PMNs, also resolved by 72 h (data not shown). Both age groups also recruited MO into the airway by 18 h postinfection (Fig. 1C) that greatly diminished, but remained elevated, at 72 h postinfection (Supplemental Fig. 2). The adult mice demonstrated a greater overall leukocyte recruitment into the lung tissue itself at 18 h that resolved by 72 h, whereas the juvenile leukocyte recruitment totals remained constant in the RSV-infected tissue 72 h, showing a persistent leukocyte presence in the infected lung (Fig. 2A). PMN counts were similarly elevated in both infected age groups at 18 h (Supplemental Fig. 2); however, neutrophilia endured in the juvenile lung at 72 h (Fig. 2B). There was not a significant EO recruitment among either infected age group at 18 h (data not shown), but in line with the other granulocyte PMNs, the juvenile mice demonstrate a significant presence at 72 h (Supplemental Fig. 2). MO recruitment into the lung tissue resembled PMNs in an age-related manner, with increases in both age groups at 18 h (Supplemental Fig. 2), and sustained numbers in the juvenile, but not adult, mice at 72 h (Fig. 2C). Increased survival could contribute to this continued myeloid presence; however, we propose that greater recruitment is the dominant factor (26). RSV-infected juvenile mice demonstrate a continued myeloid recruitment at 72 h postinfection where the adult mice have a reduced presence similar to their age-related PBS controls. It is worth noting that there was no observed difference in the myeloid recruitment between the male and female sexes of the juvenile mice, although males have been shown to have relatively worse complications due to RSV infection (data not shown) (3, 10). This line of experimentation was attempted at the 6-d timepoint to determine if the observed pattern of myeloid recruitment continued and by that mark all infected mice had leukocyte levels equivalent to their age-related PBS-treated controls (data not shown).

Myeloid recruitment has been linked to lung pathology in the face of viral infection (17, 18). Given the sustained myeloid cell infiltration in the lungs of RSV-infected juvenile mice, we next examined whether the normal RSV immune evasion factors were involved (27, 28). We examined the RNA expression of the antiviral genes Mx2 (which is downstream of IFNAR signaling), Tlr3 (whose transcript generation is a product of the IFNα initially generated by retinoic acid-inducible gene I [RIG-I] activity), and Ifnl (which is primarily generated by the infected epithelium) with no deficiency seen in the expression of each of the three for both infected age groups (Supplemental Fig. 3) (29, 30). AMs are the first leukocytes to encounter virus in the pulmonary airway whereupon they initiate the immune response, and there was no difference observed between the two age groups in the BALF at 72 h (Supplemental Fig. 2). The AM presence was also consistent between the two age groups in the lung tissue at 18 h (Supplemental Fig. 3) with a significant decrease then seen at 72 h in the lung tissue of the adult mice (Fig. 3A). Annexin V staining was incorporated to determine if the observed differences between the AM populations in the juveniles and adults were due to a deficiency in apoptotic activity. The RSV-infected juvenile mice demonstrated a significant difference compared with their age-related PBS controls for the apoptotic cell counts; however, apoptotic cells were significantly less than seen in the RSV-infected adult mice (Fig. 3B). PMN-generated TRAIL has been indicated as the driver of AM apoptosis in the infected lung landscape; CD253 (TRAIL) staining was incorporated to determine if the juvenile mice were deficient in TRAIL expression, with no discernible differences between the age groups noted in BALF at 18 h (Supplemental Fig. 3), suggesting that any juvenile AM resilience is possibly due to RSV-driven apoptotic evasion and not a cellular deficiency (27, 31). As the juvenile mice demonstrated a continued overall leukocyte recruitment into the lung tissue at 72 h, we examined the expression of chemokines that recruit these cells. qPCR was performed on the RNA extracted from whole-lung isolates of each age group, in which we found comparable mRNA expression for the classic myeloid chemokines CCL2, CCL3, and CCL4 in both age groups at 18 h (data not shown). Only the juvenile mice displayed a dramatic increase in transcript expression at 72 h for these chemokines (Fig. 4A–C, respectively). These data suggest that this continued CCL expression contributed to the continued myeloid recruitment in the younger mice at 72 h postinfection, possibly fed by the continued AM and MO presence (17, 18). We also examined the apoptotic activity of the juvenile MO population, but found no demonstrated deficiency in that function for that population (Supplemental Fig. 3).

One aspect of the inflammatory response to respiratory virus infection is a remodeling of the ECM. Myeloid recruitment contributes to lung pathology through the direct modification of the ECM, but also through an inability to properly scavenge debris naturally generated by pathogen clearance. MO in the lungs of RSV-infected juvenile mice showed reduced expression of CD64 at 18 h postinfection (Fig. 5A) with sustained reduction at 72 h postinfection (Fig. 5B). CD64 is the Fc receptor that recognizes the liver-derived pentraxin serum amyloid P (SAP), which coats cellular debris and free HA to be phagocytized. In the absence of CD64, reduced debris clean-up occurs, leading to increased HA accumulation seen in the juvenile B6 mice (32–34). Infected lungs from juvenile mice also show a significant increase in αSma expression at 72 h, indicating fibroblast to myofibroblast transition (Fig. 5C). This increased myofibroblast activation can further increase deposition of HA and other ECM components, such as versican, which has been implicated in asthma (35, 36). These data lead us to examine the lung histology to glean any pertinent observations these images presented to us. The PBS-treated adult B6 mice displayed more HA staining when compared with the juvenile mice (Fig. 6A, 6B); however, the RSV-treated juvenile mice showed increased HA deposition in the parenchyma (Fig. 6C) when compared with the adult mice (Fig. 6D). Increased parenchymal HA could further contribute to continuing unproductive inflammation with lower m.w. HA fragments stimulating dendritic cells, further driving myeloid recruitment, effecting more ECM modification, thus feeding a cycle that could perpetuate the remodeling that primes the lung landscape for hyperresponsiveness (37–40). To analyze differences in parenchymal HA content between the juvenile and adult mice, the ImageJ software package was used to compare the pixel area among the age groups; there was a statistically significant increase in parenchymal staining in the juvenile mice with no difference observed when the airway staining was compared (Supplemental Fig. 4).

Because of MO plasticity and elevated recruitment in the juvenile mice in our mouse model, we further investigated their performance during RSV infection in an ex vivo human airway epithelial cell model system cocultured with a human MO cell line (U937). The airway epithelial cultures were infected for 24 or 72 h with RSV, the U937 cells were harvested at each timepoint, and the expression of chemokine- and ECM-remodeling genes was assessed. U937 cells displayed elevated expression of CCL2, CCL3, and CCL4 (Fig. 7A–C, respectively), which contribute to myeloid recruitment into the infected space, consistent with what was seen in the lungs of the infected juvenile mice. U937 cells also showed virus-driven expression of MMP7 and MMP9 (Fig. 8A, 8B, respectively), with MMP7 expression increasing at 72 h and MMP9 production remaining constant. Both of these MMPs have been implicated in ECM pathology and increased myeloid recruitment (21, 22).

In addition to making RNA transcripts for myeloid recruitment factors and direct ECM modifiers, human U937 MO generate transcripts for matrix components that directly impact inflammation and the status of the ECM. Expression of TNF-stimulated gene-6 (TSG6) was elevated at 24 h from U937 cells (Fig. 8C). TSG-6 is an HA-binding protein that can sequester chemokines, cross-link the increased HA content, and initiate PMN apoptosis with production of TSG-6 returning to the levels expressed by the PBS controls at 72 h, which could contribute to continuing inflammation (41). The MO initially generate transcripts for HA synthase 3 (HAS3) (Fig. 8D), which is the HAS capable of generating a lower m.w. HA product, which, in itself, could also antagonize myeloid cells, further contributing to an inflammatory state (37–40). To address the increase in HA content, the MO generate hyaluronidase 1 (HYAL1) (Fig. 8E), which is responsible for the internal metabolism of phagocytized HA (39). Although many studies implicate RSV as driving a type 2 immune response, a recent study revealed a lack of type 2 helper T cells in a human-based RSV infection trial (13). The U937 MO in the human cell cultures did generate TNFSF15 (Supplemental Fig. 4). TNFSF15 drives innate lymphoid cell 2 (ILC2) generation of IL-5 and IL-13, which could initiate the EO recruitment not seen in the mouse lung tissue at 18 h but is present in the juvenile lung by 72 h (Supplemental Fig. 2) (42).

We have demonstrated that juvenile mice exhibit a distinct myeloid recruitment pattern in response to RSV infection as compared with adult mice, including exacerbating factors that support this recruitment, with a resultant impact on lung ECM, which could contribute to remodeling. In an ex vivo human airway epithelial cell–MO coculture model system, we then confirmed the presence of the chemokines responsible for myeloid recruitment and certain ECM impactors that may drive airway remodeling. To our knowledge, we are the first group to compare the age-related immune response of mice using an environmentally relevant dose of RSV, then use those initial findings as an examination template to interrogate a human cell–based ex vivo lung model (using airway epithelial cells from children), showing that increased myeloid recruitment potentiates ECM remodeling, supporting the concept that RSV infection may contribute to the development of asthma.

RSV causes 59,000 in-hospital deaths annually across the globe in children under the age of five with increasing morbidity seen in adult populations, especially the elderly and immunocompromised (4, 14, 43). Early life RSV infection has been linked to the progression to and onset of asthma. In a Swedish patient record-based study, Sigurs et al. (4) reported that 47% of the hospital admissions due to RSV-related bronchiolitis were associated with a later-life diagnosis of asthma. Although many RSV-related studies use higher doses of virus generating a type 2-like immune response, a recent study demonstrated that the typical RSV-infected child exudes an infectious cloud that extends 1 m around the person and contains 1–2 × 10⁵ PFUs of virus (43). Therefore, our age-appropriate mice using a lower, more environmentally relevant dose of RSV provide a measuring stick against which to compare the juvenile immune response with that of the adults (in 8-wk-old B6 mice) to elucidate why a virus that is so ubiquitous has such an impact in early childhood. Recent studies have also shown that the architecture and composition of a 3-wk-old mouse lung typically resembles a 2-y-old human lung and an 8-wk-old mouse lung typically resembles that of an adult human lung (11–13).

The sustained myeloid response and CCL gene expression observed in the juvenile mouse lung at 72 h should be taken into account during subsequent RSV vaccine trials, considering enduring neutrophilia was implicated in the vaccine-enhanced disease associated with the 1969 RSV vaccine trial, whereas MO are noted to exacerbate lung pathology during viral infection (5, 6, 17). Both infected age groups had an initial myeloid recruitment into the airways; however, by 72 h, that recruitment had resolved, as it appears the complications caused by RSV infection transition to the lung tissue itself. The enduring CCL production in the infected juvenile mice could be attributed to the persistent AM and increased MO presence compared with the adults (27). The immune evasion of RSV has been linked to the blockage of initial RIG-I activity (disrupting IFN-α production) and subsequent IFNAR signaling, although our B6 mice did not appear to have such difficulty with the upregulation of Tlr3 and Mx2, respectively, in addition to the upregulation of Ifnl. There was an apparent lack of AM apoptosis in the juvenile mice that could be caused by the cell populations RSV infected, making AMs a possible viral reservoir to infect the parenchymal tissue, further driving pathology (20, 27, 29, 30).

Free, low m.w. HA has consistently been linked to dendritic cell activation, inflammation, and further myeloid recruitment (37, 38). CD64 activity through SAP is the natural process by which free HA and other cellular debris are eliminated that would otherwise be contributing to inflammation causing further lung pathologic conditions. This raises the question of whether the apparent lack of MO function through CD64 in the younger mice is merely a result of naive cellular activity, the association with an incomplete ECM in the younger organisms (as previous studies have shown), or some combination of both (36, 38, 44, 45)? HA equilibrium is essential for a return to hemostasis, immune regulation, and the continuance of proper lung function, with the absence of proper HA mechanics driving airway pathology (46, 47).

Translating our mouse findings into a human primary airway epithelial cell–based infection model, we discovered that MO exposed to an infected epithelium generate increased RNA transcript levels of CCL2, CCL3, and CCL4 along with previously characterized ECM modifiers MMP7 and MMP9. The CCLs would contribute to the enduring myeloid recruitment seen in the mice, whereas the MMPs would contribute to matrix pathology while further allowing PMN infiltration (21, 22). The human U937 MO cocultured with airway epithelial cells initially generate transcripts for TSG6, whose incorporation into the ECM would initiate PMN apoptosis, sequester chemokines (such as CXCL2) for enzymatic degradation, and cross-link HA to establish a more proinflammatory ECM in the lung. However, production of TSG6 returns to the levels expressed by the PBS controls at 72 h. Does this allow for the space along the HA chains to then be occupied by the more inflammatory chondroitin sulfate versican, which could contribute to continuing inflammation (35, 36, 48–50)? The MO also have the potential themselves to generate low m.w. HA, although they further identify the need to digest the free HA that would contribute to inflammation through the generation of HYAL1 transcripts. EO recruitment resembles that of other myeloid cells in the BALF; however, after limited recruitment in the lung at 18 h, there is a significant recruitment at 72 h, consistent with an enhanced type 2 response that can be attributed to MO-derived TNFSF15 driving the ILC2 production of IL-5 and IL-13 and possibly skewing the later immune response that has been observed in some RSV infection studies (7, 9, 42).

The study of RSV infection progression in humans is hampered by many limitations. It is, of course, difficult and unethical to study the progression of RSV infection directly in lung tissue from human children, and the children who do typically present to the emergency department with the symptoms of RSV bronchiolitis are already far enough into an infection course to preclude characterization of the time course of pulmonary myeloid recruitment. In this study, we attempted to profile the immune response of RSV-infected juvenile and adult mice. However, findings from murine studies may not accurately model human disease in which human pathogens such as RSV are concerned. Therefore, we also attempted to model lung epithelial responses to RSV infection and their effects on MOs and ECM in a human cell culture system. A strength of the current study is that we replicated our initial mouse findings in an ex vivo human culture system, demonstrating the potential for RSV infection to drive ECM modification. A limitation of our work is that outcome measures are primarily based on gene expression and not protein data. One important limitation of our composite ex vivo human cell culture system is that it is not possible to separate contributions from individual cell types to secreted cytokine pools assayed in the cell culture media, as all cells share the same media. Thus, gene expression analysis is the best surrogate for the contributions of individual cell types in this system. Another important limitation of this work is that we are unable to isolate our findings to a specific pathway, given the complex interactions during RSV infection. These signaling pathways are an important area of ongoing research in our laboratory in which future studies will focus on AM activity (another possible source of the Ccl2, Ccl3, and Ccl4 seen in the whole mouse lung isolates) and any epithelial-derived signals that may steer the immune reaction in our human primary cell cultures (15, 16, 27).

Overall, the increased myeloid recruitment, ECM disruption, and human cell line results suggest that RSV does indeed have the ability to potentiate the lung architecture remodeling observed in asthma. These findings warrant further investigation, as all factors combined are likely to prime the lung landscape toward the settings for airway dysfunction in the face of an environmental challenges. We are currently conducting further studies using GFP-labeled RSV to determine in which leukocytes the virus resides at 72 h, along with completing preliminary studies that point to necroptosis as the possible driver of sustained myeloid recruitment.

Taken as a whole, these data provide a potential mechanistic explanation for the differences seen in the response to RSV infection between children and adults. Further studies are required to link the changes in ECM deposition and remodeling to subsequent lung function deficits seen in some RSV-infected infants.

We thank Dr. Sabine Spath for consultation on flow cytometry staining panels and experimental design. We thank Dr. Kate Pothoven for consultation on viral immunology in mouse compared with human experimental systems. We thank Dr. Adam Wojno and the BRI’s Flow Cytometry Corps; Dr. Andrew Burich and BRI’s vivarium facility; and Dr. Pamela Y. Johnson and the BRI Histology Core for facilitating these experiments. We thank Dr. Virginia M. Green for editing the manuscript.

The authors have no financial conflicts of interest.