Biological sex influences immune response to influenza infection in juvenile outbred mice

Abstract

Children are highly susceptible to influenza virus infection and lung inflammation. The majority of studies of influenza pneumonia in mice are in adult, inbred, same-sex mice with limited variation in disease outcomes, unlike what is observed in children. In this study, juvenile outbred mice were challenged with influenza A/California/07/2009 H1N1 virus to examine the temporal effect of infection on lung inflammation in genetically dissimilar male and female mice. In this model, influenza infection induced variable disease severity characterized by viral load, airspace leak and inflammation, and lung function. This response was not only clustered by sex at 3 and 4 days postinfection but also was associated with disease severity by day 8. Influenza virus infection induced altered production of cytokines and chemokines by sex and timing. Males displayed an early cytokine and chemokine response driven by severity, whereas females showed a late response only in severe mice. This was independent of viral RNA load. Flow cytometry was then conducted to analyze lymphocytes. In male mice, neutrophil recruitment was elevated by disease severity. In contrast, female mice showed increased T cell numbers late in infection regardless of severity. Finally, unsupervised clustering of mice by immune parameters at day 8 showed similarity by sex and severity, with more variability in male mice. These data demonstrate that the immune response to influenza virus infection is impacted by sex and disease severity in pediatric equivalent outbred mice. This complexity of immune activation is important to consider when evaluating drivers of disease pathogenesis or planning therapeutic interventions targeting immune mediators.

NEW & NOTEWORTHY

This study defines the temporal immune response to influenza A virus infection in the juvenile lung. Lung inflammation differs by biological sex and disease severity in juveniles, which is reflected in cytokine and chemokine mediators and innate and adaptive immune cells recruited to the tissue. These data inform future generations of computational models of severe disease drivers and impact the potential use of immunomodulatory therapy in children with influenza A pneumonia.

INTRODUCTION

Influenza disproportionately impacts young children, leading to increased rates of hospitalization and morbidity. Children often have higher fevers, greater lymphadenopathy, and more gastrointestinal symptoms than adults during an influenza virus infection; however, this is not due to greater viral titers (1, 2). Since differences in influenza outcomes cannot be explained by a higher viral burden in children, it is likely that immune-mediated mechanisms are responsible. It has been demonstrated that infants and children exhibit distinct immune responses to the same pathogen when compared with adults. Newborns have an immature immune system, with virtually no immune memory. At this stage, the newborn immune response is predominated by innate immune cells such as neutrophils and monocytes that have weakened effector responses compared with adults (3). Furthermore, the adaptive immune system during early life is primed for immune suppression to create self-tolerance to alloantigens (4). Together, an immature innate response and immunosuppressive adaptive response create an anti-inflammatory environment early during development.

Discrepancies between the immune responses of adults and juveniles to influenza virus infection can be effectively modeled in specific pathogen-free mice, which lack pre-existing immunity to the virus, in contrast to human adults who possess prior immune memory. Juvenile mice have delayed interferon-γ and IL-10 production during primary influenza challenge and a delayed and diminished T cell response compared with adult mice (5). Furthermore, adult mice and older juvenile mice have a Th1-skewed CD4⁺ response, but neonatal mice have no clear skewing (5). The innate immune response to influenza virus is also distinct in juvenile mice compared with adults. Juvenile mice have sustained NLRP3 activation, type I interferons, and monocyte chemoattractant protein (MCP)-1 production (6). Aberrant MCP-1 production leads to excessive recruitment of monocytes into the lung and persistent inflammation. The monocyte gene expression signatures from juvenile mice are enriched for inflammatory signatures compared with adult monocytes and contribute to lung injury (6). The mouse model has helped us understand that the magnitude and kinetics of the immune response likely contribute to worse outcomes in the context of juvenile influenza infection. However, these studies were performed on inbred mice, which do not represent the genetic diversity or the breadth of influenza virus infection severity present in the human population.

Jackson diversity outbred (JDO) mice are derived from interbreeding of the eight collaborative cross strains (7). As a result, JDO mice are genetically distinct and display a range in disease phenotypes (8). Previous studies from our group have shown that adult and juvenile JDO mice exhibit a wide range of disease severity in response to influenza virus infection (9, 10). In this study, we leveraged this outbred mouse model to examine the effect of sex and severity on influenza infection-driven lung inflammation in pediatric equivalent mice.

MATERIALS AND METHODS

Experimental Design and Influenza Virus Infection

Juvenile JDO mice were obtained from Jackson Laboratory and were between 19 and 24 days of age upon arrival. The mice were allowed a 72-h acclimation period before being intranasally inoculated with 1 × 10⁴ plaque-forming units (PFUs) of influenza A/California/07/2009 H1N1, diluted in 10 μL of phosphate-buffered saline (PBS). The mice were maintained under pathogen-free conditions at UPMC Children’s Hospital of Pittsburgh, and all animal procedures were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (Protocol No. 23073501). Mice were weighed frequently throughout infection and euthanized using pentobarbital injection followed by exsanguination. Naïve “day 0” mice were euthanized 3 days (males) or 6 days (females) after arrival. Influenza-infected mice were euthanized on days 3, 4, 5, and 8 postinfection in groups of 15 males and 15 females per time point.

FlexiVent

Pulmonary function was evaluated in anesthetized (90 mg/kg pentobarbital, administered intraperitoneally) and tracheotomized mice during mechanical ventilation at a rate of 150 breaths/min, with a tidal volume of 10 mL/kg and a positive end-expiratory pressure of 3 cmH₂O. A computer-controlled small-animal ventilator (FlexiVent, SCIREQ, Montreal, QC, Canada) was used for this assessment, following previously established protocols (11, 12). Baseline quasi-static lung compliance and hysteresis were determined by analyzing pressure-volume (PV-Ps) curves.

Cytokine and Chemokine Measurement

A sample from the lower right lung lobe was collected, weighed, and homogenized in 1 mL of sterile PBS. Lung homogenates were centrifuged to separate the pellet, and the supernatants were diluted 1:1 with PBS. These samples were then analyzed for cytokine and chemokine concentration using the BioRad Magpix multiplex platform and the BioPlex Pro Mouse Cytokine 23-plex assay (BioRad, Hercules, CA), following the manufacturer’s guidelines.

Flow Cytometry

For flow cytometry analysis, the left lung lobe was excised into 1 mL of PBS on ice, mechanically dissociated using sterile scissors, and digested at 37°C with 1 mg/mL collagenase media (DMEM, Gibco Fisher Scientific, Waltham, MA). After 1 h of digestion, the lung tissue was passed through 70-μm mesh filters to create a single-cell suspension. To remove any remaining red blood cells, ACK lysis buffer (Gibco Fisher Scientific, Waltham, MA) was applied. The cells were resuspended in 1 mL sterile PBS for counting and then plated in 96-well V-bottom plates for staining with myeloid and lymphoid antibody panels (Supplemental Table S1). Intracellular staining was carried out at room temperature using the Bioscience Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher Scientific, Waltham, MA). Chemokine receptors were stained separately at 37°C. The samples were analyzed on a Cytek Aurora flow cytometer (Cytek Biosciences, Fremont, CA), and data were processed using FlowJo software. Male samples were processed first and did not contain both CD90.1 and CD90.2 antibodies for detection of total T cells; however, female samples were stained with both CD90.1 and CD90.2 antibodies. Gating strategies utilized for cell populations are found in Supplemental Figs. S1–S3.

Bronchoalveolar Lavage Protein and Cell Counts

Bronchoalveolar lavage fluid (BAL) was collected by washing the mouse airways with 1 mL of sterile PBS. The BAL cells were pelleted and resuspended in fresh PBS for total cell counts. The supernatant was stored in −80°C until a protein assay was performed. Protein levels were quantified using the Pierce BCA protein assay, according to the manufacturer’s instructions (Thermo Fisher Scientific, Waltham, MA).

RNA Isolation and qPCR

Mouse lungs were snap-frozen in liquid nitrogen. RNA was extracted according to the protocol provided with the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany). cDNA synthesis was performed using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Quantitative PCR (qPCR) was carried out with SsoAdvanced Universal Probes Supermix (Bio-Rad, Hercules, CA) and target-specific TaqMan real-time PCR assay primer probes (Thermo Fisher Scientific, Waltham, MA). To assess viral RNA load, quantitative real-time RT-PCR was performed on lung RNA targeting the viral matrix protein (M1), as previously described (13, 14). The primers and probe used were as follows: Forward Primer: 5′-GGACTGCAGCGTAGACGCTT-3′, Reverse Primer: 5′-CATCCTGTTGTATATGAGGCCCAT-3′, Probe: 5′-/56 FAM/CTCAGTTAT/ZEN/TCTGCTGGTGCACTTGCCA/3IABkFQ/-3′.

Statistical Analysis

Measurements of 23 cytokines and 16 immune cell types were obtained from the experimental data. Mice with IDs 30, 61 (males), and 90 (female) were excluded from all analyses due to missing data points. To standardize the data, a log1p transformation (base 10) was applied, with a minimum value of 0.03 added to all measurements before transformation. This approach ensured that the standardized measurements followed a normal distribution, facilitating subsequent statistical analyses. For each sex at each experimental time point, the sample size was n = 14 for day 0 (female) and day 5 (male); for all other time points n = 15, unless otherwise specified. Regulation profiles were visualized using heatmaps generated from log1p-transformed data, normalized to the mean values of measurements at day 0 (uninfected baseline), to evaluate up- or downregulation over the course of infection. A two-way ANOVA test was conducted to compare uninfected mice (day 0) with infected mice at each time point, enabling the assessment of sex-specific and severity-specific regulation dynamics for cytokines and immune cell types. P values were adjusted using Benjamini–Hochberg method to account for false discovery rates. Adjusted P values less than 0.05 were considered statistically significant. This GitHub repository contains all of the code to reproduce the results from the manuscript: https://github.com/Salty-Ramen/Characterizing-Juvenile-JDO-Immune-Response.

Data Cohort Labels

The experimental cohort consisted of mice of two biological sexes, with 75 males and 75 females included in the study. Of these, 60 males and 60 females were infected with influenza virus. BAL cell counts; BAL fluid protein levels, lung hysteresis, and viral RNA load were measured for each mouse at every time point. Within each sex, infected mice with values higher than the median for at least two of these four metrics were classified as “Severe Phenotype,” whereas those that did not meet this criterion were classified as “Mild Phenotype.” Raw data values utilized are listed in Supplemental Spreadsheet S1, SeverityLabelsAssignment.xlsx.

Unsupervised Machine Learning

Principal component analysis (PCA) was used on log10 fold change transformed data to reduce the dimensionality of the unlabeled data set comprising all cytokine and immune cell type measurements, enabling visualization of the high-dimensional data in a two-dimensional space. To validate the clusters formed in the two-dimensional PCA space, hierarchical clustering was performed independently on the same data set in conjunction with PCA. The statistical significance of the association between the resulting clusters and the severity labels was assessed using Fisher’s exact test. All computations mentioned thus far have been executed in R (R Software, San Francisco, CA).

RESULTS

Characterizing Influenza A Infection Dynamics of Juvenile Outbred Mice

Male and female juvenile JDO mice were infected with influenza virus A/California/07/2009 H1N1 and were weighed frequently during the acute infection period. All mice gained weight over the course of the study period, with male mice gaining more weight than female mice at later time points (Fig. 1A). Lung viral RNA load was calculated by measuring viral M1 RNA via qPCR and using the ΔΔCT method to quantify relative amounts of viral replication within the lung. Viral loads were comparable between male and female mice throughout the infection time course, with the exception of day 5 postinfection, where female mice exhibited a significantly lower viral load compared with males (Fig. 1B). Interestingly, this is also the time when male and female weight gain began to diverge. BAL fluid was collected to assess airway inflammation (total BAL cell counts, Fig. 1C) and lung leak (BAL protein levels, Fig. 1D). In naïve mice, male mice displayed a statistically significant higher BAL cell count than female mice (Fig. 1C). A difference was seen again at day 8 postinfection, where the males had a higher BAL cell count than the females. BAL protein concentrations remained statistically indistinguishable between the males and females during infection. Lung hysteresis data were examined to evaluate lung function during the infection course; females showed statistically higher levels than males in naïve mice as well as infected mice at days 3 and 5 days postinfection (Fig. 1E). Although female mice exhibited a more rapid decline in viral RNA load, their body weight gain was slower, indicating a potential difference in disease tolerance and severity during influenza infection.

Figure 1.

Characterizing the infection dynamics of juvenile outbred mice when infected with influenza A virus. A: bodyweights were plotted as fractions normalized to uninfected weights, where 1 = 100% of starting weight. Wilcoxon test was performed to evaluate differences between male and female mice. B–D: viral RNA load, bronchial alveolar lavage (BAL) cell counts, and BAL protein levels were log10 transformed for plotting and statistical analyses. E: hysteresis was evaluated as a measure of lung function. Error bars illustrate the standard error on the mean value. B–E: student’s t test was performed comparing within sex at each time point. The P values obtained were corrected for false discovery rates (FDR) using Benjamini–Hochberg method.

Early Sex-Dependent Immune Responses in Juvenile JDO Mice Converge to Severity-Driven Responses at Later Time Points

Viral RNA load, BAL infiltrates, lung hysteresis measurements, and weight curves were found to be different between male and female juvenile mice infected with influenza A virus. Viral RNA load, BAL infiltrates, and lung hysteresis were combined to create a “Severity Label” to classify the infected mice as “severe” and “mild.” To investigate if the immune response of juvenile mice to influenza A virus infection could be delineated by biological sex and the assigned severity label, a principal component analysis (PCA) was conducted to project the high-dimensional data set into a two-dimensional plane. The immune response remained largely indistinguishable across all mice at 3 days postinfection (Fig. 2A). At 4 and 5 days postinfection, the immune responses of male and female mice were distinct (Fig. 2, B and C). At 8 days postinfection, the stratification is primarily led by the infection severity, regardless of the sex (Fig. 2D). Fisher’s exact test was performed to statistically support the association between hierarchical clusters and the associated labels (Supplemental Fig. S4).

Figure 2.

Principal component analysis on cytokine, chemokine, and immune cell measurements. A–D: the entire immune response dataset comprising of 23 cytokines and chemokines and 16 immune cell types were transformed into log10 fold-change values and normalized. Principal component analysis (PCA) was conducted on the combined fold-change dataset to discern broad patterns in the dataset and plotted for each time point. Hierarchical clustering was performed independently on this dataset and overlayed on the corresponding PCA plot to verify clustering based on severity labels or sex. The top 2 clusters at each time point were used to perform Fisher’s exact test to statistically support the association between hierarchical clusters and severity labels or sex.

Severe Females Show a Late, Secondary Upregulation of Immune Signaling, Whereas Males Show a Stronger Initial Response

To further characterize the immune response profile of outbred juvenile mice, a panel of 23 immune cytokines and chemokines was measured from homogenized lungs. To visualize the dynamics of these cytokines and chemokines throughout the infection, they were log10-transformed and their profiles plotted over time. Two-way ANOVA was used to establish differences between male and female mice as well as between severe- and mild-labeled mice. To examine potential interaction effects between biological sex and disease severity, we used linear models including interaction terms. In addition, hierarchical clustering was performed to categorize cytokines and chemokines based on their temporal expression patterns (Supplemental Fig. S5).

The first cluster of cytokines and chemokines identified through hierarchical clustering was characterized by sex-specific expression patterns (Fig. 3). Infected male mice exhibited significantly higher upregulation of IL-1α, TNF-α, and granulocyte-macrophage colony-stimulating factor (GM-CSF) compared with infected females at all time points postinfection. IL-2, IL-1α, TNF-α, and IFN-γ peaked at 3 days postinfection in females but declined thereafter. In contrast, these cytokines in males displayed a delayed and more sustained peak beginning at 4 days postinfection. This sustained peak in males presented differently in severe and mild mice, especially in IL-10 and IL-1β, with the mild males showing significant downregulation at 5 and 8 days postinfection. Notably, severe female mice showed a pronounced secondary surge across all cytokines in this group at 8 days postinfection.

Figure 3.

Group 1 cytokine and chemokine response of JDO mice to influenza A virus infection. A total of 23 cytokines and chemokines were measured, and the values were normalized by the lung lobe sample weight. A subset was chosen and grouped based on their dynamic profile using hierarchical clustering. Error bars illustrate the standard error on the mean value. Two-way ANOVA was performed comparing sex and severity at each time point in relation to control (uninfected mice). *Sex, Ŝeverity, and +for interaction P < 0.05. Actual P values are listed in Supplemental Spreadsheet S3—ANOVA_table.xlsx. JDO, Jackson diversity outbred.

The second cytokine and chemokine cluster primarily comprised chemokines, which were consistently elevated in severe male mice throughout the first 5 days of infection (Fig. 4). At 3 days postinfection, all groups exhibited chemokine upregulation, but levels were highest in severe mice. By 4 days postinfection, severe males maintained elevated chemokine levels, whereas female and mild male mice showed stable or declining levels. This divergence continued at 5 days postinfection, with female and mild male groups demonstrating downregulation, whereas severe males sustained high expression. At 8 days postinfection, severe males exhibited marked downregulation across this chemokine group, whereas severe females showed a distinct late-phase upregulation.

Figure 4.

Group 2 cytokine and chemokine response of JDO mice to influenza A virus infection. A total of 23 cytokines and chemokines were measured, and the values were normalized by the lung lobe sample weight. A subset was chosen and grouped based on their dynamic profile using hierarchical clustering. Error bars illustrate the standard error on the mean value. Two-way ANOVA was performed comparing Sex and Severity at each time point in relation to control (uninfected mice). *Sex, Ŝeverity, and +for interaction P < 0.05. Actual P values are listed in Supplemental Spreadsheet S3—ANOVA_table.xlsx. JDO, Jackson diversity outbred.

Taken together, these patterns suggest a dichotomy in response phenotypes: mild male and female mice exhibit early, moderate cytokine and chemokine peaks, whereas severe males demonstrate a robust, delayed response. Severe females uniquely display both a mild early peak and a secondary, pronounced upregulation at 8 days postinfection. Importantly, females have lower viral RNA load than males earlier during infection but still fail to thrive (gain weight), suggesting that severity in females may be driven more by inflammation than viral sensing or replication.

Male Juvenile JDO Mice Show a Neutrophil Response, Whereas the Females Demonstrate a Strong, Late-Surging Lymphoid Response

Lung immune cell population profiles were ascertained by performing flow cytometry on lung immune cells. The frequencies of each cell type expressed as a percentage of CD45⁺ cells were log10 transformed, and their profiles were plotted over time (Fig. 5). Females exhibited a higher frequency of alveolar macrophages than males throughout the infection time course. Alveolar macrophages in males increased at 3 days postinfection, peaking at day 4. In contrast, females had high alveolar macrophage percentages when uninfected with no increase until 4 days postinfection. At day 5, both male and female mice showed decreased alveolar macrophage frequencies. However, this decrease varied across groups; severe mice, regardless of sex, showed a sustained decrease, whereas mild mice showed an increase at 8 days postinfection. Although the mechanism of alveolar macrophage decline in mice during influenza infection has not been established, these results are consistent with the literature (15, 16). It is important to note that the influenza virus-induced decrease in alveolar macrophages was shown to be mouse strain dependent (17). Interstitial macrophages were elevated in both male and female mice by severity at days 3 and 4 postinfection. Male mice then showed a more robust decrease in interstitial macrophages compared with female mice at later time points.

Figure 5.

Myeloid immune cell response of JDO mice to influenza A virus infection. A total of 16 immune cell frequencies were measured, and the values were normalized by the lung lobe sample weight. Error bars illustrate the standard error on the mean value. Two-way ANOVA was performed comparing Sex and Severity at each time point in relation to control (uninfected mice). *Sex, Ŝeverity, and +for interaction P < 0.05. Actual P values are listed in Supplemental Spreadsheet S3—ANOVA_table.xlsx. JDO, Jackson diversity outbred.

Inflammatory monocytes (CCR2⁺ Ly6c^Hi) remained at baseline at 3 days postinfection. At 4 and 5 days postinfection, however, infected mice showed sex-specific differences, with inflammatory monocytes decreasing in the lungs of female mice while remaining close to baseline in the lungs of male mice. At 8 days postinfection, inflammatory monocytes increased further in severe mice and decreased in mild mice, with mild females showing the most pronounced reduction. The frequency of nonclassical monocytes (CCR2⁺ Ly6c^Lo) remained stable until day 4 postinfection. However, at day 5 postinfection, the frequency of nonclassical monocytes declined sharply in all groups but returned to baseline levels by day 8 in mild mice, and stayed low in the severe group, regardless of biological sex. Taken together, these data support the idea that inflammatory monocytes contribute to disease severity but suggest that either biological sex or viral replication impact temporal resolution.

Neutrophils demonstrated pronounced differences in regulation between sexes when infected with influenza A virus. In males, the frequency of neutrophils increased sharply, peaking at 4 days postinfection and decreased thereafter at 5 and 8 days postinfection. Neutrophil frequencies were consistently higher in severe males 4 and 5 days postinfection compared with mild males. In contrast, neutrophil responses were not strongly induced in female mice following infection and even decreased in frequency starting 4 days postinfection. Male mice showed a sharp increase in dendritic cells at 3 days postinfection, followed by a return to uninfected levels for the remainder of the infection time course. Dendritic cell levels in female mice remained at uninfected levels throughout the infection, except at day 5 postinfection, during which a mild decrease was observed, primarily in severe female mice.

Taken together, the myeloid response to influenza A virus in juvenile, male mice was greater than that of females. Male mice had a longer period of viral RNA load; however, so inflammatory monocytes, which were most differentially regulated 5 days postinfection, might be responding to persistent viral antigen. Neutrophil and dendritic cell responses displayed earlier differences in regulation, suggesting that biological sex may influence their response to influenza A virus. Disease severity corresponded to a heightened neutrophil response and a delayed reduction in inflammatory monocytes.

We then examined the T cell response to influenza A virus infection in our cohort (Fig. 6). Natural killer (NK) cells and CD4⁺ T cell subsets were differentially regulated in juvenile male and female mice following influenza infection. Baseline frequencies of NK and Th1 cells were higher in females and were also induced to higher levels than males postinfection. Severity did not impact these two cell types. The frequency of regulatory T cells in females following infection remained largely unchanged. In males, we saw the frequency decrease as early as 3 days postinfection in mild males, with severe males dropping more slowly until 8 days postinfection. The abundance of effector CD4⁺ T cells was higher in females at baseline, but increased in severe males 3 days postinfection. The frequency of effector CD4⁺ T cells remained elevated in severe males until 8 days postinfection, where a slight increase in mild males and a slight decrease in severe males brought the two groups to similar levels. In females, the greatest increase in effector CD4⁺ T cells was seen 8 days postinfection and was not influenced by disease severity. Finally, Th17 cells were induced in both mild and severe female mice 8 days postinfection, but decreased in both groups of male mice at the same time point. Together, females had greater induction of a phenotypically diverse CD4⁺ T cell response, with increases in effector, Th1, and Th17 CD4⁺ T cells, as well as a larger NK cell response.

Figure 6.

Lymphoid immune cell response of JDO mice to influenza A virus infection. A total of 16 immune cell frequencies were measured, and the values were normalized by the lung lobe sample weight. Error bars illustrate the standard error on the mean value. Two-way ANOVA was performed comparing Sex and Severity at each time point in relation to control (uninfected mice). *Sex, Ŝeverity, and +for interaction P < 0.05. Actual P values are listed in Supplemental Spreadsheet S3—ANOVA_table.xlsx. JDO, Jackson diversity outbred.

Effector CD8⁺ T cells were robustly induced in both males and females following infection, but were temporally distinct. Females, regardless of severity, had high levels of effector CD8⁺ T cells in the lung by 3 days postinfection, which stayed stable until 8 days postinfection, when severe females displayed a sharp increase in effector CD8⁺ T cell frequency. With males, severity regulated the frequency of effector CD8⁺ T cells earlier, with severe males having a higher frequency 3–5 days postinfection, which returned to almost baseline levels in both male groups by 8 days postinfection.

The T cell response to influenza A virus in juvenile mice revealed notable sex-specific and severity-dependent dynamics. Females mounted a broader and more sustained CD4⁺ T cell response, with elevated induction of effector, Th1, and Th17 subsets, alongside consistently higher NK cell frequencies, regardless of disease severity. In contrast, males exhibited a decline in regulatory T cells and a severity-driven increase in effector CD4⁺ T cells early during infection, which eventually converged with mild levels by day 8. Effector CD8⁺ T cell responses were robust in both sexes but followed distinct trajectories: females exhibited an early and sustained response that peaked late in severe cases, whereas males showed an earlier severity-driven induction that was resolved by day 8. These findings suggest that biological sex shapes both the magnitude and timing of antiviral T cell responses, with disease severity further modulating CD8⁺ T cell dynamics, particularly in males.

Severity, Sex, and Individual Variation Shape Immune Responses in Outbred Mice during Influenza Virus Infection

Clustering analysis showed that immune response profiles aggregated based on severity labels at day 8 postinfection. Independent hierarchical clustering was performed to validate the significance of the cluster formation, and the hierarchical clustering results reinforce the notion of severity-driven clustering. To analyze the spectrum of immune response profiles at 8 days postinfection, a heatmap was plotted with log10 fold change data (compared with naïve) of all the cytokines, chemokines, and immune cell frequency data obtained from the mice (Fig. 7). The heatmap columns were segmented based on the hierarchical clustering results. Heatmap rows were clustered by biological function. There was one severe cluster that included both males and females and was characterized by high-level upregulation of myeloid-derived proinflammatory cytokines and chemokines like IL-12, IL-1, TNFα, macrophage inflammatory protein (MIP)-1b, and IL-6. Furthermore, this cluster had increased lymphocyte-derived cytokines including IL-4, IL-5, and IL-10. Cellular regulation, on the other hand, appeared to be more sex-specific, with males and females within a severity cluster exhibiting differential immune profiles. There was also a primarily mild cluster that included both males and females. This cluster had lower relative abundances of proinflammatory cytokines and chemokines compared with the severe cluster, but no remarkable changes in cellular immune profiles, which again appeared to be more sex-specific. In addition, there were six mice in this group that showed responses distinct enough to be classified separately by the hierarchical clustering algorithm. Notably, these distinct mice were all males and were a mixture of both mild- and severe-labeled phenotypes. A second heatmap with the mice grouped by sex shows that Th17 activation is greater in males versus females, suggesting an alternative inflammatory mechanism in males post-influenza (Supplemental Heatmap S1).

Figure 7.

Day 8 immune response heatmap. Log10 fold change data of all cytokines, chemokines, and immune cells at 8 days postinfection were used to generate a heatmap. The color bar represents z-scores of the measurements normalized to the baseline. Clusters 2 (mild) and 3 (severe) were the top 2 clusters identified using Hierarchical clustering, having a significant association with the severity labels.

DISCUSSION

Our group previously quantified the immune response profiles of juvenile outbred mice at 21 days post-infection (9). This recovery time point is after viral clearance has occurred. At that time point, infected female mice exhibited significant upregulation of diverse cytokines and chemokines (including IFN-γ, TNF-α, IL-1α/β, IL-12, IL-9, and Eotaxin) and increased infiltration of CD4⁺ T cells and exudate macrophages compared with naive female controls. In contrast, infected male mice displayed less robust cytokine production at 21 days postinfection relative to naive males; IL-6 was the only cytokine significantly upregulated in this group. Furthermore, infected male mice showed limited immune cell infiltration, with interstitial macrophages being the primary exception. Although this previous study characterized the immune phenotype at 21 days postinfection during recovery, the dynamics at earlier acute infection time points, which impact acute disease severity, remained undefined. Therefore, the present study investigates these acute immune dynamics. In addition, this work aims to provide a more granular analysis of the juvenile immune response to influenza A virus infection by categorizing mice based on disease severity (“mild phenotype” and “severe phenotype”) and identifying immunological features potentially contributing to these distinct outcomes.

We investigated immune response profiles in juvenile outbred mice following infection with a pandemic H1N1 influenza strain at 3-, 4-, 5-, and 8-days post inoculation. To assess the influence of biological sex on immune dynamics, PCA was performed on the log10 fold-change data at each time point. This analysis revealed that juvenile male mice exhibited distinct immune responses compared with female mice, particularly as the infection progressed. Viral RNA load was similar between sexes early during infection, until 5 days postinfection, when males had significantly higher levels of viral RNA compared with female mice. Body weight gain also differed between the sexes later during infection; with females 5- and 8-days postinfection exhibiting a failure to thrive phenotype, whereas males continued to gain weight throughout the study. In adult outbred male mice, weight loss positively correlated with BAL cellularity, protein leak, inflammatory cytokines, and viral burden, but not lung pathology (10). In both humans and the mouse model of influenza infection, females often experience greater morbidity and mortality. In a human challenge study, females were more likely to experience symptoms (18). Since body weight loss during influenza infection is the most common observable clinical symptom measured in mice, the failure to thrive phenotype observed in females suggests juvenile female mice also experience more symptoms. Furthermore, inbred mouse studies have observed greater body mass loss and hypothermia in adult female mice in response to a moderate influenza viral dose, as was used in this study, and faster viral clearance (19–21). Together, these data support previous findings that infection dynamics differ between males and females.

To identify factors contributing to the observed PCA clustering, we examined cytokine, chemokine, and immune cell dynamics over time, using statistical tests (e.g., two-way ANOVA) for comparisons between groups. During the initial phase of infection (3–4 days postinfection), male mice displayed significantly stronger upregulation of a broad panel of cytokines and chemokines. Many cytokines indicative of myeloid-derived inflammation, like IL-1α, IL-1β, TNF-α, and GM-CSF, were elevated in male mice (regardless of severity label) compared with female mice on days 3, 4, and 5 post-infection. Some cytokines and chemokines were specifically enhanced in severe male days 4 and 5 postinfection, including IL-6, G-CSF, regulated on activation, normal T cell expressed and secreted (RANTES), MCP-1, MIP-1α/β, and keratinocyte-derived cytokine (KC). This is consistent with findings in adult male outbred mice, where elevated IL-6 and G-CSF were observed with increasing viral burden and weight loss (10). Furthermore, mild and severe female mice did not have differential expression of most cytokines and chemokines until 8 days postinfection, where severe females had a steep increase in almost all cytokines and chemokines discussed. These levels often reached or exceeded the same concentration as severe males 8 days postinfection. Taken together, male mice had a robust, antiviral cytokine and chemokine response early during infection that generally waned by 8 days postinfection, with many shared cytokines and chemokines between mild and severe males. Females, on the other hand, had a more modest cytokine and chemokine response, with paradoxical evidence suggesting earlier viral clearance. However, severe females uniquely saw a second spike in proinflammatory cytokine and chemokine production, a pattern not present in males regardless of severity label or mild females, suggesting that disease severity in females may be driven by inflammation and not a lack of viral control. Generally, male and female mice have similar levels of viral replication during influenza infection, with females experiencing greater levels of cytokinemia (22, 23). These studies often focus on a single time point following infection, however, and do not capture the temporal dynamics examined in this study. These data show that cytokine and chemokine production during influenza infection is regulated by sex, severity, and infection kinetics. Pathogenic roles for the cytokines and chemokines strictly increased in severe cases should be examined further.

Myeloid cell dynamics throughout infection revealed distinct temporal patterns between males and females. For example, CCR2⁺ Ly6c^Hi monocytes, a cell well known to play a role in lung immunopathology during infection (6, 24), peaked at 3 days postinfection in both sexes and remained elevated in mild and severe males until 8 days postinfection, but steadily declined in mild females 4–8 days postinfection, indicating CCR2⁺ Ly6c^Hi monocytes may be regulated by both sex and severity. Neutrophils appeared to be primarily regulated by sex; on the other hand, as neutrophil frequencies peaked in male mice 4 days postinfection, regardless of severity label, and steadily declined in females throughout the course of the infection. Overall, myeloid infiltration during influenza infection was stronger in males, where duration and cellular dynamics did not correspond to the severity of infection, whereas myeloid responses in females were more blunted and driven by severity. Others have identified increased myeloid-derived inflammation in females following influenza infection, but this response is primarily driven by hormonal fluctuations, which may not be relevant in our juvenile model (25, 26). In addition, several other groups have investigated the role of testosterone in the response to influenza infection in adult and aged mice. Testosterone treatment in female mice reduced influenza mortality, and aged male mice with lower testosterone had worse influenza viral clearance (27–29). Again, testosterone may not be a primary driver in juvenile mice. The impact of sex differences in the response to viral infections in adult mice has been reviewed (30).

Lymphoid cell infiltration was highly sex-specific in both naïve and infected mice. It is worth highlighting a limitation of this study and a consideration for others working with outbred mice—genetic diversity led to the expression of CD90.1 and CD90.2 alleles in different mice. In the initial study performed on male mice, this allelic diversity was not considered in the antibody panel design, so T cells from male mice were gated on expression of T cell receptor (TCR)β alone, a feature that can be downregulated in response to TCR activation, whereas the female study included both CD90.1 and CD90.2 targeting antibodies so that T cells could be gated on both TCRβ and CD90 expression. Female mice showed elevated levels of effector CD4⁺ T cells, effector CD8⁺ T cells, NK cells, Th1, and Th17 cells at days 5 and 8. Male mice, however, had small increases in Th1 and effector CD8⁺ T cells on days 3 and 4 post-infection, which returned to near baseline frequencies by day 8. Within males, severity seemed to dictate the magnitude of the response, with severe males displaying a sharper increase in the frequency of effector CD4⁺ and CD8⁺ T cells 3–5 days postinfection, but eventually converging back to baseline in both severe and mild males by 8 days postinfection. Although the limitation in panel design highlighted above may explain some discrepancy between males and females, within each sex and compared with baseline, 8 days postinfection female mice had a higher amplitude of lymphoid responses regardless of disease severity. These data support previous findings that convalescent inflammation in the juvenile, outbred female lung is sustained at least 21 dpi and is driven in part by an imbalance in Th1/regulatory T cell ratio, a phenomenon that does not present in male counterparts (9). The female lymphoid response is also broad, including Th1 and Th17 responses, mirroring a lack of T cell skewing seen in neonatal mice, which supports increased disease severity in the juvenile population caused by a robust yet inappropriate inflammatory response (5). This biphasic response showcased by females, especially severe labeled, might be explained by the initial peak in the myeloid response not being substantial enough to control the infection, leading to a secondary upregulation of a diverse array of signals reflective of a chronic inflammatory phenotype.

Mouse models of influenza infections are primarily based on inbred, genetically identical, adult mice. Although valuable for studying specific immune responses to influenza, the lack of genetic diversity in inbred mouse models compared with human populations may limit the discovery of unique immunological pathways. Outbred mouse models present an opportunity to bridge this gap, as they better reflect the genetic variation seen in humans. Furthermore, since influenza severity is exacerbated in the very young and very old, examining infection in juvenile mice provides a more relevant model for understanding disease mechanisms in at-risk human populations and addresses a gap in the current literature, which predominantly focuses on adult mice. Most of the influenza studies conducted in outbred mice have focused on genetic effects on disease outcome. Mx1 status has been associated with severity, and other quantitative trait loci studies have revealed genetic differences that contribute to antibody memory (31, 32). In addition, the outbred mouse genetic signature was shown to be similar to that in human blood (33). Furthermore, influenza virus susceptibility was shown to track with several loci on the X chromosome in genetically distinct mice, implying sex-specific differences (34). Herein, we investigated individual variation in immune responses by constructing immune response profiles at each time point. Independent PCA and hierarchical clustering performed on the log10 fold change data suggest that despite the genetic variance within the mouse population used in this experiment, overarching features in immune responses are conserved across categories such as biological sex and severity phenotype. Day 8 responses were of particular interest based on the PCA results; the immune responses showed diversity based on both the biological sex of the mice and their assigned severity label. Although the unsupervised learning methods suggest relative similarity among severe mice across sexes, there are a few mice that can be considered outliers in terms of their response profiles. For example, an outlier male mouse labeled severe exhibited high levels of IL-10, IL-4, IL-5, RANTES, and TNF-α, similar to other mice belonging to the severe cluster. However, this mouse also showed low levels of Th-17, MCP-1, and MIP-1α, a combination not observed in any other mice from this cohort. Such unique patterns may suggest the existence of more than one pathway leading to a severe phenotype. Moreover, there were also mice that exhibited an immune response profile largely similar to mice belonging to severe clusters, yet were labeled with a mild phenotype, indicating that a given immune profile can manifest as either mild or severe depending on factors such as the animal’s genetic makeup. Overall, individual variation exists alongside broad trends in the immune response to influenza infection in juvenile outbred mice, likely resulting from their genetic diversity, which leads to differences in disease presentation.

In conclusion, this study has identified differences in the inflammatory response to influenza A in outbred juvenile animals driven by biological sex and/or disease severity. Male mice displayed a greater initial inflammatory cytokine and chemokine response than females, which was further increased in severe disease. Female mice had a delayed increase in these cytokines and chemokines only in severe mice. Male mice had a more robust innate immune response in terms of neutrophil recruitment, again higher in severe animals, consistent with the chemokine response observed. In contrast, females had elevated T cell recruitment versus males regardless of severity. These findings have significant implications for emerging immunomodulatory therapies for respiratory viral infections. The data imply that single biologics may have highly divergent efficacy in a temporal and sex-dependent manner. Understanding temporal immune responses in pediatric equivalent mice will allow for the development of computational models to define dominant mechanism of disease pathology.

Supplementary Material

Supplemental Figs. S1–S5: https://doi.org/10.6084/m9.figshare.32056029.

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.32056110.

Supplemental Spreadsheets S1: https://doi.org/10.6084/m9.figshare.32056119.

Supplemental Spreadsheets S2: https://doi.org/10.6084/m9.figshare.32056149.

Supplemental Spreadsheets S3: https://doi.org/10.6084/m9.figshare.32056161.

Supplemental Heatmap S1: https://doi.org/10.6084/m9.figshare.32128801.

DATA AVAILABILITY

The source data are available in Supplemental Spreadsheet S1–S3 and to verified researchers upon request by contacting the corresponding author.