Oncogene c-Myc Regulates Upper Respiratory Epithelial Repair for Host Immunity During Acute Influenza Infection

Abstract

The upper respiratory epithelium protects the airway from microbial exposure, environmental irritants, and viral pathogens such as influenza A virus (IAV). Situated at the airway–gastrointestinal junction, the larynx must integrate this barrier function with fine neuromuscular control under constant immune stimulation. Clinical syndromes such as post-viral vagal neuropathy—marked by chronic cough, throat clearing, dysphonia, and vocal fatigue—highlight critical gaps in our understanding of how the larynx and upper airway remodel after viral injury. To address this, we employed mouse models integrating genetics, imaging, and single-cell transcriptomics to dissect IAV-induced injury and repair in the upper respiratory tract. IAV infection displayed regionally restricted tropism toward ciliated, secretory, neuroendocrine, and basal epithelial cells, eliciting rapid clearance, acute neutrophil invasion, and a sustained intraepithelial population of cytotoxic CD8⁺ natural killer T (NKT)-like cells. Single-cell analysis revealed injury-induced Myc expression in KRT5⁺ basal progenitors, which activated a proliferative repair program and an epithelial–immune communication circuit centered on a MYC-dependent CXCL10–CXCR3 axis. Basal-specific Myc deletion preserved viral replication and clearance but blunted injury-induced proliferation, skewed lineage differentiation, and impaired CD8⁺ NKT cell recruitment. These findings define a coordinated epithelial–immune program governing post-IAV mucosal restoration and offer mechanistic insight into persistent upper airway sequelae.

Introduction

The upper respiratory epithelium serves as a frontline barrier defense, shielding the airways from a constant barrage of symbiotic microbes, environmental irritants, and a range of invading pathogens.^1–4 Mucosal immunity is critical to maintain organ homeostasis following epithelial insult.^5,6 Among common respiratory pathogens, influenza A virus (IAV) remains a major global threat, causing substantial morbidity, acute respiratory distress, and death each year.⁷ While epithelial cells exhibit significant regenerative capacity, it is not well-described how the upper respiratory epithelium remodels in post-viral settings.

The larynx represents a unique and understudied component of this system. As a neurologically complex sensorimotor organ positioned at the intersection of the respiratory and gastrointestinal tracts, the larynx must integrate epithelial defense with fine-tuned neuromuscular control amid constant immunologic challenges. Work in papillomaviral disease has highlighted this complexity: in severely immunocompromised mice, laryngeal infection with MmuPV1 leads to severe dysplasia and invasive cancer persisting for months,⁸ whereas immunocompetent mice demonstrate protection against recurrent or vertical infection,⁹ underscoring the importance of epithelial–immune crosstalk in shaping local antiviral responses. Yet, beyond these models, the impact of other clinically relevant viruses—such as influenza—on laryngeal and upper airway biology remain surprisingly unexplored, with limited data on multi-organ, systems-level analysis.

Recent single-cell studies have revealed the upper respiratory epithelium to be highly compartmentalized, with distinct proximal–distal and basal–luminal transcriptional programs across the pharyngolaryngeal and tracheobronchial regions.¹⁰ Many respiratory viruses enter through, and initially engage with, sentinel epithelial cells at these mucosal surfaces.⁶ Early repair responses trigger inflammatory cascades that are central to local immune regulation and, when dysregulated, contribute to chronic disease.^1,6 Acute upper respiratory infections are among the most common illnesses globally;¹¹ although typically self-limiting, a subset of patients experience persistent sensory and motor dysfunction long after acute illness, suggesting systems-level changes. Post-viral vagal neuropathy exemplifies this phenomenon, often manifesting as chronic cough, throat clearing, dysphonia, vocal fold paresis, and vocal fatigue.^12–17 Despite strong clinical associations linking these symptoms to antecedent viral illness, the mechanisms driving post-viral laryngeal dysfunction remain undefined.

To address this gap, we sought to define the upstream cellular and molecular events that shape epithelial repair and immune activation following IAV-induced injury in the mouse upper respiratory tract. Using an integrative approach combining mouse genetics, cell-based assays, and single-cell transcriptomics, we uncovered coordinated epithelial and immune responses that extends well beyond the acute phase of infection. We show that IAV infection targeting discrete epithelial lineages (ciliated, secretory, neuroendocrine, and basal) elicits a rapid Ly6G⁺ neutrophilic response and establishes a persistent intraepithelial population of cytotoxic CD8⁺ NKT-like cells marked by a durable effector transcriptome. We further identify ectopic Myc induction within injured subglottic and tracheal pseudostratified epithelium and demonstrate that KRT5⁺ basal progenitors deploy a Myc-dependent proliferative and chemokine program that shapes downstream immune behavior. Basal-specific loss of Myc blunts epithelial expansion and disrupts lineage differentiation, while attenuating CXCL10-dependent recruitment of CXCR3+ CD8⁺ NKT cells. Together, these findings delineate a systems-level epithelial–immune circuit governing post-viral remodeling in the upper respiratory tract and offer mechanistic insight into clinically relevant sequelae such as post-viral laryngeal sensorimotor dysfunction.

Results

Spatial and temporal dynamics of IAV infection reveal regionally restricted epithelial tropism and rapid viral clearance.

To delineate the spatiotemporal progression of IAV infection along the upper respiratory tract, we generated large-scale coronal sections encompassing the pharyngolaryngeal-to-tracheobronchial axis from wild-type (WT) mice at 1, 3, 7, and 8 days post-infection (DP-IAV) (Fig. 1A,B). Multiplex RNA in situ hybridization with probes targeting IAV transcripts and the interferon-γ-inducible chemokine Cxcl10—known to orchestrate antiviral immunity¹⁸—revealed initial viral localization to the airway epithelium at 1DP-IAV, with peak epithelial coverage at 3DP-IAV (Fig. 1C,D). Viral transcripts persisted through 7DP-IAV but were eradicated by 8DP-IAV, aligning with established timelines of innate immune-mediated clearance.¹⁹ Cxcl10 expression mirrored viral burden, with negligible expression at 8DP-IAV, thereby underscoring its role in recruiting effector cells to sites of active replication (Fig. 1C). Consistent with prior reports,²⁰ IAV induced transient morbidity with decrease in body weight of ~20%, reaching its lowest point at 10DP-IAV before returning to near baseline levels by 21DP-IAV (Fig. 1E).

Fig 1. IAV infection induces transient epithelial injury with broad viral tropism.

(A,B) Schematics of upper airway coronal sectioning and IAV infection timeline. (C) H&E and RNAscope detection of IAV-mRNA (white) and Cxcl10 (red) at 1, 3, 7, and 8DP-IAV. (D) Quantification of epithelial IAV-mRNA coverage (one-way ANOVA with Dunnett’s post-hoc test; mean ± SEM; n = 3/group). (E) Body weight change following infection (Welch’s t-test; mean ± SEM; n = 8/group). (F) Timeline schematic corresponding to panels G-G′. (G,G′) H&E and immunofluorescence for KRT5 (green) and KRT17 (magenta) showing basal cell hyperplasia at 3–14DP-IAV with resolution by 21DP-IAV; dashed boxes indicate higher-magnification insets. (H) Quantification of KRT5/KRT17 epithelial coverage across airway compartments (two-way ANOVA with Fisher’s LSD post-hoc test; mean ± SEM). (I,J) Schematics highlighting epithelial diversity and airway regions enriched for IAV-mRNA. (K-P) RNAscope co-localization of IAV-mRNA with epithelial lineage markers (FOXJ1, SCGB1A1, ASCL1, KRT5, KRT13) at 3 and 7DP-IAV. DAPI, blue. Scale bars: 500 μm (C,G), 100 μm (others). *P < 0.05, **P < 0.005, ***P < 0.0005; ns, not significant.

Building on this temporal framework, we interrogated regional vulnerabilities by reorienting sections along dorsal-ventral and medial-lateral axes at peak infection (3DP-IAV). Pseudostratified epithelia of the subglottis, trachea, and extrapulmonary bronchi exhibited the highest susceptibility to IAV-induced cytopathy, whereas stratified, squamous regions of the glottis and supraglottis harbored fewer infected cells (Fig. 1C; Fig. S1A). This topographic bias suggested that epithelial architecture and cell-type composition dictate viral entry and propagation.

To link infection dynamics to regenerative potential, we traced basal progenitor responses using immunofluorescence for keratins KRT5 and KRT17—hallmarks of proliferative basal cells—across acute (3–7DP), subacute (14DP), and resolution (21DP) phases (Fig. 1F). Hyperplasia of KRT5⁺KRT17⁺ cells emerged by 3DP-IAV, crested at 7–14DP-IAV, and normalized by 21DP-IAV (Fig. 1G–G′). Compartment-specific quantification (supraglottic surface epithelium [SE], subglottic SE, tracheal SE, and submucosal glands [SMG]) confirmed that expansion was confined to pseudostratified zones of prior viral enrichment, with maximal accrual in subglottic and tracheal epithelia (Fig. 1H; Fig. S1B). Thus, basal cell activation represents a targeted reparative program, spatially attuned to sites of epithelial compromise.

We next resolved the cellular substrates of infection through colocalization of IAV transcripts with lineage-defining markers (Fig. 1I–P). In the trachea, viral RNA predominantly colocalized with FOXJ1⁺ ciliated and SCGB1A1⁺ club cells (Fig. 1K), while in the laryngeal epiglottis, SCGB1A1⁺ secretory cells were preferentially targeted (Fig. 1L). Lineage tracing in Ascl1^CreER;tdTomato mice further implicated subglottic and tracheal neuroendocrine cells as viral reservoirs (Fig. 1M). Additionally, IAV enrichment extended to KRT13⁻ suprabasal layers of stratified squamous hillocks and arytenoid mucosa (Fig. 1N,O), as well as dispersed KRT5⁺Ki67⁺ basal cells (Fig. 1P; Fig. S1C). Collectively, these data delineate a broad yet selective epithelial tropism, wherein IAV exploits regionally specialized sentinel cells to establish foothold, culminating in viral detection and progenitor-driven remodeling.

Neutrophil epithelial invasion and persistent CD8⁺ NKT-like sentinels drive post-IAV airway immunity and repair.

Given that viral clearance hinges on coordinated innate and adaptive immunity to preserve epithelial barrier function,⁵ we systematically profiled immune infiltration to dissect its contributions to resolution and homeostasis. Prior studies have implicated neutrophils in early antiviral containment and CD8⁺ T cells in late-stage cytotoxicity;^18,21 thus, we prioritized these compartments using Ly6G (neutrophil) and CD3 (pan-T cell) immunostaining in WT tissues at acute (3DP-IAV, 7DP-IAV) and chronic (21DP-IAV) stages (Fig. 2A–F). In uninfected controls, Ly6G⁺ neutrophils resided in the subepithelial compartment. At 3DP-IAV, however, these cells invaded the pseudostratified epithelium, colocalized with IAV enriched cells, aggregating into luminal-extruding spheres suggestive of NETosis-mediated viral entrapment (Fig. 2D,D′; Fig. S1D). This breach resolved by 7DP-IAV (Fig. 2D′), indicating a temporally restricted role in acute containment.

Fig 2. Acute IAV infection recruits neutrophils and establishes persistent intraepithelial CD8⁺ NKT cells.

(A,B) Schematics of IAV infection timeline and airway sectioning. (C) H&E staining of sa-line-treated controls at 3 DP. (D,D′) Immunofluorescence for Ly6G⁺ neutrophils at 3 and 7 DP-IAV demonstrating epithelial infiltration; dashed boxes indicate insets. (E) Immunofluorescence showing CD3⁺ intraepithelial T cells in subglottic and tracheal epithelium at 7 DP-IAV that persist at 21 DP-IAV. (F) Quantification of CD3⁺ epithelial coverage (one-way ANOVA with Dunnett’s post-hoc test; mean ± SEM). (G) Bulk RNA-seq showing increased expression of cytotoxic T-cell genes (Cd8a, Cd8b1, Ccl5, Gzma, Gzmb) at 21 DP-IAV. (H) RNAscope co-localization of Gzmb and Ccl5 transcripts at 21 DP-IAV. (I) Top enriched GO biological processes from bulk RNA-seq at 21 DP-IAV. (J) Schematic of scRNA-seq workflow. (K) UMAP of immune populations at 7 DP-IAV (n = 8 biological replicates). (L) Dot plot of top cluster markers. (M) Immune cluster composition comparing saline and IAV. (N,O) Vio-lin plots showing enrichment of cytotoxic and tissue-resident gene signatures in CD8⁺ NKT cells. DAPI, blue. Scale bars: 500 μm (C,D), 100 μm (others). *P < 0.05, **P < 0.005, ***P < 0.0005; ns, not significant.

In contrast, intraepithelial CD3⁺ T cells accrued progressively, peaking at 7DP-IAV and remaining significantly expanded at 21DP-IAV over saline controls (Fig. 2E,F). To molecularly characterize this resident population, bulk RNA sequencing of 21DP-IAV laryngeal and tracheal tissues identified a cytotoxic signature dominated by Cd8a, Cd8b1, Gzma, Gzmb, and Ccl5 transcripts (Fig. 2G). In situ validation confirmed colocalization of Gzmb+Ccl5+ cells within the epithelium at 21DP-IAV (Fig. 2H). Gene Ontology (GO) analysis of differentially upregulated genes was enriched for T cell-mediated immunity (Cd8a, Prf1), pyroptosis (Gzma, Gzmb), epithelial differentiation (Sult1b1, Dhrs9, Cdk1), metabolic regulation (Ccl5, Cdk1), and apoptosis (Gzma, Prf1, Gzmb) (Fig. 2I), implying multifaceted contributions to both viral lysis and tissue restitution.

To integrate these immune dynamics with epithelial repair and viral hotspots, we conducted single-cell RNA sequencing (scRNA-seq) on en bloc-dissociated tissues along the pharyngolaryngeal-to-tracheobronchial axis from 7DP-IAV—the timepoint between peak inflammatory repair and viral clearance (Fig. 2J). Unbiased clustering resolved 13 immune subsets (Fig. 2K,L; Fig. S1E–G), with compositional shifts post-IAV dominated by T/NK lineage expansion (Fig. 2M). Notably, we identified a transcriptionally distinct CD8⁺ NKT-like cell cluster, co-expressing canonical T cell (Cd3d, Cd8b1), NK (Nkg7), and Th1/iNKT markers (Klrk1, Tbx21, Ifng), alongside cytotoxic effectors (Gzma, Gzmb, Ccl5) (Fig. 2L,N; Fig. S1H). Consistent with a tissue-resident phenotype, these cells upregulated canonical residency markers (Cd69, Itga1) and mirrored transcriptional programs observed in our bulk-seq dataset (Fig. 2N,O), suggesting that CD8⁺ NKT cells may function as durable sentinels for immune surveillance.

scRNA-seq uncovers MYC-driven proliferative basal progenitors in post-IAV epithelial repair.

To dissect the epithelial regenerative repertoire during the transitional phase of IAV clearance, we interrogated EpCAM⁺-enriched fractions from our 7DP scRNA-seq atlas (Fig. 3A). Dimensionality reduction and unsupervised clustering delineated 20 distinct transcriptional states spanning the upper respiratory epithelium (Fig. 3A–D; Fig. S1I), which segregated into laryngeal, tracheal, and SMG domains in accordance with anatomical and prior transcriptomic delineations.¹⁰ A salient feature was the emergence of proliferative epithelial clusters, disproportionately represented in our post-IAV condition (Fig. 3B,C,E), implying a stimulus-dependent expansion or transdifferentiation program in response to viral perturbation.

Fig 3. scRNA-seq identifies Myc-driven basal progenitor programs within IAV-enriched epithelial populations.

(A) Schematic of scRNA-seq experimental design. (B) Integrated UMAP of upper airway epithelial populations highlighting regional and cell-type diversity. (C) UMAP colored by condition (saline vs IAV) identifying IAV-enriched epithelial clusters. (D) Dot plot of top markers defining epithelial clusters. (E) Cluster composition comparing saline and IAV conditions. (F) In silico viral transcript mapping demonstrating preferential IAV enrichment in epithelial lineages at 7DP-IAV. (G–I) UMAPs highlighting IAV-enriched epithelial subsets and associated cell-cycle states (G1, S, G2M). (J) Dot plot of markers defining IAV-enriched clusters. (K) Integrated regulatory GSEA identifying activation of MYC target pathways in IAV-enriched basal progenitors. WT, wild-type; DP, days post.

To pinpoint viral niches at cellular resolution, we augmented the reference transcriptome with the complete IAV genome sequence (TableS1), enabling imputation of viral transcripts across clusters. This revealed a stark epithelial tropism, with highest IAV-mRNA enrichment in secretory, ciliated, and—critically—the novel IAV-enriched (e.g. responsive) subsets (Fig. 3F; TableS1), thereby validating our spatial mapping. Sporadic detection in CD8⁺ NKT cells (<15%) and vascular smooth muscle (<5%) hinted at intercellular viral transfer or bystander activation (Fig. 3F).

To confirm that the synthetic ‘IAV-gene’ did not confound dimensionality reduction or clustering in our scRNA-seq analysis, we reprocessed the dataset after its exclusion, maintaining identical parameters and fixed seeds. Cell-type labels were aligned by barcode to permit one-to-one comparison. Clustering outcomes were identical, with an adjusted Rand index (ARI) and normalized mutual information (NMI) of 1.00, and a perfectly diagonal confusion matrix (TableS1). PCA embeddings (PCs 1–30) showed near-perfect congruence via Procrustes analysis (Gower’s m²=1.26×10⁻⁴, r=0.99994; TableS1), with gene loading vectors exhibiting high concordance (median Pearson r=0.99936 across PCs). Local topology, assessed by k-nearest neighbor (kNN) Jaccard overlap, remained unchanged (median=1.00; 10th/50th/90th percentiles: 1.00/1.00/1.00). UMAP visualizations confirmed preservation of manifold structure and cell-type organization (Fig. S2A). These metrics collectively affirm that the synthetic ‘IAV-gene’ exerted no discernible influence on the transcriptomic landscape or clustering solution.

The temporal coincidence of our IAV-enriched subpopulation with peak basal hyperplasia prompted us to posit a progenitor-like identity underpinning repair. Subclustering unmasked proliferative heterogeneity, with G2/M− and S-phase signatures evoking cell-cycle traversal (Fig. 3G–I; TableS1). Preeminent among these was a basal-enriched contingent, defined by Trp63, Krt5, Krt14, Krt17, and Ngfr, and biased toward S-phase progression—hallmarks of replicative competence in response to genotoxic stress (Fig. 3G–J). Pathway analysis through irGSEA unveiled marked upregulation of MYC target gene sets in the IAV-enriched basal subpopulation (Fig. 3K), inspiring targeted follow-up experiments.

Basal-specific Myc ablation impairs proliferative repair and lineage commitment following IAV injury.

The proto-oncogene transcription factor c-Myc (hereafter Myc) constitutes a central regulator of immune cell proliferation, metabolic reprogramming, and differentiation, coordinating rapid cellular responses during activation and inflammatory states.^22–24 Interrogation of our scRNA-seq dataset revealed selective enrichment of Myc—but not L-Myc or N-Myc—within the IAV-enriched population, with the highest expression in the basal subpopulation, where 72% of basal cells expressed Myc (Fig. 4A,B; Fig. S2B,C). Pseudotemporal ordering further supported a stem-like trajectory, bifurcating toward goblet, luminal-hillock, differentiating club, mature club, and stress-activated epithelial fates (Fig. 4C). Spatial validation via in situ hybridization confirmed widespread ectopic Myc expression throughout the pseudostratified epithelium spanning the subglottis, trachea, and extrapulmonary bronchi during the acute post-IAV phase (Fig. 4D). Collectively, these findings designate a MYC-driven basal cell program as a central mediator of acute epithelial regeneration following IAV injury.

Fig 4. MYC promotes basal progenitor expansion and lineage remodeling during post-IAV epithelial repair.

(A,B) Violin and feature plots showing Myc expression across IAV-enriched epithelial populations at 7DP-IAV. (C) Monocle3 pseudotime trajectory analysis of IAV-enriched epithelial subsets. (D) Immunofluorescence at 3DP-IAV demonstrating ectopic Myc expression in damaged pseudostratified epithelium. (E) Experimental schematic for basal cell-specific Myc deletion. (F) Immunofluorescence confirming Myc loss following tamoxifen induction. (G,H) Quantification of epithelial Myc-mRNA and IAV-mRNA coverage (Student’s t-test; mean ± SEM). (I) EdU and TUNEL staining assessing proliferation and apoptosis at 7DP-IAV. (J,K) Quantification of proliferative and apoptotic indices (one-way ANOVA with Tukey’s post-hoc test; mean ± SEM). (L) Immunofluorescence of lineage marker–positive cells in tracheal epithelium at 21DP-IAV. DAPI, blue. Scale bars: 500 μm (D), 100 μm (others). *P < 0.05, **P < 0.005, ***P < 0.0005; ns, not significant.

To assess the functional requirement of Myc in epithelial repair, we generated tamoxifen-inducible basal cell–specific knockout mice (Krt5^creER;Myc^flox/flox, hereafter Krt5^creER;Myc). Experimental animals received three consecutive tamoxifen doses five days prior to IAV inoculation, with tissue harvest performed at 7 and 21DP-IAV (Fig. 4E). Immunofluorescence microscopy with quantification of Myc-mRNA coverage confirmed efficient epithelial Myc deletion (Fig. 4F,G). Phenotypically, Krt5^creER;Myc mutants mirrored heterozygous controls in acute morbidity, incurring ~20% weight loss at 7DP-IAV (Fig. S2D).

To determine whether Myc deletion impacted viral replication and/or clearance, we assessed IAV transcript localization during the reparative phase. At 7DP-IAV, both control and Krt5^creER;Myc mutants exhibited similar levels of epithelial IAV-mRNA enrichment, and by 10DP-IAV, viral transcripts were undetectable in both groups, indicating normal viral replication and clearance (Fig. 4H; Fig. S2E).

Given the role of Myc in cell cycle regulation,^23,25 we next evaluated epithelial proliferation and apoptosis using EdU incorporation and TUNEL staining at 7DP-IAV. Homeostatic epithelia showed no differences in EdU or TUNEL positivity between control and mutant mice (Fig. 4I–K). However, following IAV injury, Krt5^creER;Myc mutants exhibited a marked reduction in epithelial cell proliferation compared to infected controls, while apoptosis levels remained unchanged (Fig. 4I–K). These findings indicate that Myc drives mitogenic surge selectively upon injury, decoupling it from homeostatic turnover, apoptosis, or antiviral orchestration.

Anticipating chronic sequelae from blunted epithelial expansion, we scrutinized 21DP-IAV histology, yet discerned similar cell abundance and no overt dysmorphia in mutants versus controls, irrespective of IAV status (Fig. S3A,B). Given that all upper airway epithelial lineages originate from KRT5⁺ progenitors (Fig. S2F), we next examined lineage-defined cell types in Krt5^creER;Myc mutants compared to infected controls to assess sustained architectural repair at 21DP-IAV. In mutant epithelium, we observed trends toward reduced KRT14⁺ basal cells, modest expansion of KRT13⁺ hillock cells, and a decrease in acetylated-tubulin⁺ ciliated cells, although none of these changes reached statistical significance (Fig. 4L; Fig. S3B). Notably, we identified a significant reduction in SCGB1A1⁺ club cells in injured mutants, accompanied by disrupted ciliated and club cell distribution at 21DP-IAV (Fig. 4L; Fig. S3B). Collectively, our data indicate that Myc acts as a critical injury-induced regulator of basal progenitor function, coordinating both epithelial expansion and lineage specification. Loss of Myc blunts proliferation and disrupts differentiation trajectories, resulting in mispatterned epithelial remodeling at 21DP-IAV.

Epithelial-Immune Crosstalk Establishes a MYC-Dependent Cxcl10–Cxcr3 Axis Driving CD8⁺ NKT-like Cell Recruitment Post-IAV.

Epithelial cells, as primary targets of IAV-mediated cytopathy, orchestrate immune recruitment via secreted mediators to expedite viral clearance and mucosal restoration.^5,6 To test the hypothesis that Myc expression in basal cells modulates immune recruitment, we examined differentially expressed genes in our scRNA-seq dataset, focusing on IAV-enriched epithelial cells. Cxcl10 emerged as the singular chemokine significantly upregulated, with Myc+ co-expression confined to the injury-responsive basal population (Fig. 5A; Fig. S3C), positioning it as a candidate for coordinating immune influx.

Fig 5. MYC-dependent epithelial CXCL10 signaling recruits CD8⁺ NKT cells during post-IAV repair.

(A) UMAP of IAV-enriched epithelial populations showing co-expression of Myc and Cxcl10 at 7DP-IAV. (B-D) CellChat analysis identifying CXCL10–CXCR3 as the dominant IAV-enriched-to-CD8⁺ NKT signaling axis. (E) Schematic of tamoxifen induction and IAV infection timeline in Krt5^CreER;Myc mutant mice. (F,G) Whole-mount RNAscope imaging and quantification showing reduced epithelial Cxcl10 expression in Krt5^CreER;Myc mutants at 7DP-IAV. (H) Immunofluorescence showing reduced CD45⁺ immune infiltration and epithelial Gzmb/Ccl5 expression in mutants at 7 and 21DP-IAV. (I) Quantification of intraepithelial CD3⁺ cells. (J) Quantification of epithelial Gzmb expression at 21DP-IAV. DAPI, blue. Scale bars: 500 μm (F), 100 μm (others). *P < 0.05, **P < 0.005, ***P < 0.0005; ns, not significant.

To map epithelial-immune communication, we applied CellChat to infer ligand-receptor interactions.²⁶ Cxcl10 transcripts were detected across epithelial (serous-ductal, IAV-enriched, basal-myoepithelial) and immune (TR-macrophage, neutrophil) populations, yet its receptor, Cxcr3, was exclusively expressed in our CD8⁺ NKT cells (Fig. 5B–D). Probabilistic signaling analysis underscored the Cxcl10–Cxcr3 axis as the predominant conduit linking IAV-enriched epithelia to CD8⁺ NKT cells (Fig. 5C,D), implicating this pathway in precise immune recruitment during repair. In contrast, interrogation of the Cxcl2–Cxcr2 axis—previously tied to neutrophil dynamics²¹—revealed no epithelial Cxcl2 expression at 7DP-IAV, excluding its role in epithelial-driven responses at this stage (Fig. 5C).

To interrogate the functional control exerted by Myc over this epithelial–immune chemokine axis, we profiled immune recruitment dynamics in Krt5^creER;Myc mutants versus heterozygous controls at acute 7DP-IAV and convalescent 21DP-IAV phases (Fig. 5E). In situ hybridization disclosed a profound depletion of Cxcl10-mRNA across the pseudostratified airway mucosa in mutants relative to controls specifically at the acute 7DP-IAV stage (Fig. 5F–H), underscoring the requisite role of Myc in injury-induced chemokine induction. This blunted CXCL10 gradient translated to impaired mucosal immunity whereby mutants displayed significantly reduced CD45+ leukocyte and CD3+ T-cell infiltration at 7DP-IAV (Fig. 5H,I; Fig. S3D), with trends toward persistence at 21DP-IAV that fell short of statistical significance. Although overall CD3+ accrual normalized by 21DP-IAV, epithelial expression of hallmark CD8+ NKT effector genes—Gzmb and Ccl5—was durably subdued in mutants (Fig. 5H,J). Notably, Gzmb transcripts exhibited sustained, statistically significant attenuation at 21DP-IAV (Fig. 5H,J), implying that Myc orchestrates not only initial recruitment but also the long-term priming and cytotoxic arming of intraepithelial effectors. Collectively, these findings establish that Myc orchestrates a Cxcl10–Cxcr3–dependent epithelial–immune signaling circuit, enabling selective recruitment and functional maturation of CD8⁺ NKT cells for coordinated mucosal repair following IAV injury.

Discussion

Viral infections in the airway play a pivotal role in modulating host immune responses, influencing both the resolution of infection and the preservation of epithelial integrity essential for airway health and function. Here, we define a MYC-dependent epithelial–immune circuit that orchestrates repair of the upper respiratory tract after IAV injury. Using spatial mapping, scRNA-seq, and conditional genetics, we show that IAV targets regionally specialized epithelial lineages and elicits a MYC-driven basal progenitor response that is coupled to recruitment of cytotoxic CD8⁺ NKT effector cells via a CXCL10–CXCR3 axis. Basal cell–specific Myc deletion uncouples these processes such that viral replication/clearance and gross architecture are preserved, but injury-induced proliferation, lineage allocation, and cytotoxic T cell recruitment are attenuated, resulting in subtle yet mispatterned remodeling of secretory and ciliated compartments.

Our findings extend a growing literature positioning Myc as a central regulator of epithelial regeneration beyond its canonical oncogenic role.^24,25,27,28 Myc has been implicated in coordinating proliferation and metabolic rewiring in diverse regenerative contexts, including intestinal, hepatic, and skeletal tissues.^29–31 In the lung, c-Myc promotes activation of the peribronchiolar smooth muscle cell niche and Fgf10 expression after naphthalene injury, and its deletion in this compartment attenuates epithelial repair.³² Work in corneal epithelium further demonstrates c-Myc maintains epithelial architecture at homeostasis and governs the balance among stratification, differentiation, and exfoliation during wound repair, in part through modulation of p63-dependent cell-fate programs.²⁵ Our data add a complementary layer by showing that, in the upper respiratory tract, Myc is selectively induced in KRT5⁺ basal progenitors of pseudostratified epithelia after viral injury and is required for a full proliferative burst and appropriate differentiation into club and ciliated lineages. Across epithelial organs, c-Myc thus emerges as a conserved coordinator of injury-induced regeneration whose precise regulation is essential for maintaining tissue homeostasis. Whether this program is conserved across non-viral airway challenges such as chemical/environmental irritants and bacterial infection remains to be determined. In homeostatic conditions, basal cells of the stratified squamous laryngeal epithelium displayed robust Myc expression; nonetheless, this region experienced negligible viral insult, precluding any detectable MYC-mediated repair program. Accordingly, we propose that similar MYC-regulated programs exist in this niche but are engaged only after substantial epithelial compromise.

A key feature of our model is the integration of the MYC-driven basal progenitor program with localized cytotoxic immunity. We identify a persistent intraepithelial CD3⁺ T cell population at 21DP-IAV, exhibiting a tissue-resident, CD8⁺ NKT cytotoxic transcriptional signature. scRNA-seq and CellChat analysis revealed that Myc⁺ basal and IAV-enriched epithelial cells are major sources of the chemokine Cxcl10, whereas its receptor Cxcr3 is largely confined to these CD8⁺ NKT cells, positioning this axis as a dominant conduit of epithelial–immune crosstalk. The CXCL10–CXCR3 axis is well recognized for directing effector T-cell trafficking to injured tissues.^33–38 In influenza models, CXCR3⁺ CD8⁺ T cells and lung tissue-resident memory (T_rm) cells provide rapid heterosubtypic protection upon re-exposure, yet subsets of CXCR3⁺ or T_rm CD8⁺ T cells can also exacerbate chronic lung pathology without significantly altering viral clearance, indicating that functional outcomes are context dependent.³⁵ Consistent with this complexity, Myc-deficient airways show reduced epithelial Cxcl10 expression and diminished early CD3⁺ recruitment, indicating that MYC tunes both the magnitude and positioning of cytotoxic cells during repair.

Several caveats warrant careful consideration when interpreting our results. First, our identification of a CXCR3⁺ CD8⁺ NKT-like population does not resolve its full heterogeneity. CXCR3 marks diverse effector subsets—including Th1-biased CD4⁺ T cells, canonical CD8⁺ T_rm, type I/II NKT cells, MAIT cells, γδ T cells, and FOXP3⁺ regulatory T cells—each of which may respond differently to epithelial Cxcl10 and exert distinct tissue-level effects.^37–40 Second, the functional implications of sustaining these cytotoxic cells within the upper airway epithelium remain uncertain. Their transcriptional profile, defined by residency markers (Itga1/CD49a) and effector genes (Gzma, Gzmb, Ccl5), suggests they are primed memory-like cells capable of rapid antiviral defense. Indeed, emerging evidence positions CD49a as a compartmental discriminator for CD8⁺ T_rm cells in barrier tissues like skin, where it underpins rapid effector deployment (Gzma, Gzmb, Ccl5, Nkg7, Cxcr3) upon restimulation.³⁹ By extension, our cells may afford protective immunosurveillance against reinfection, aligning with canonical T_rm benefits in respiratory models.^39,40 However, their persistence may also potentiate maladaptive inflammation during recurrent insults. Distinguishing protective versus pathogenic roles will require studies incorporating lineage tracing, targeted depletion, cytokine blockade, and secondary-challenge paradigms.

These interpretive nuances extend to broader methodological constraints that shape our conclusions. For instance, we examined a single IAV strain and dose in young adult mice, leaving open questions about how aging, viral variants, or comorbidities (e.g., asthma or reflux) might reshape the MYC–CXCL10–CD8⁺ NKT pathway. Our conditional Myc knockout targeted KRT5⁺ basal cells broadly, without separating rare stem-like pools (e.g., KRT14⁺) from more differentiated ones, nor does it address potential compensation by other Myc paralogs over longer time scales. Moreover, although our analyses capture structural and molecular outcomes, they do not assess physiological consequences such as mucociliary clearance, airway biomechanics, or vagal reflex function.

Despite these limitations, our findings offer compelling clinical insights, revealing how a single viral hit can leave a lasting mark on airway tissue structure and local immune defenses. This mechanism is especially relevant to post-viral vagal neuropathy, chronic cough, voice loss (dysphonia), and related unexplained problems with laryngeal sensation and movement. In such conditions, an exaggerated or poorly resolved T_rm-like response in the laryngeal mucosa may sustain low-grade inflammation, skew epithelial differentiation, and perturb sensory circuits, thereby prolonging symptoms despite virologic recovery. Fundamentally, our results frame airway basal cells as key hubs linking repair to immune tuning, wherein MYC-driven proliferation and chemotactic cues (such as CXCL10-mediated immune influx) orchestrate balance between adaptive memory and maladaptive persistence.

To extend these findings, future studies should determine how repeated or mixed viral exposures recalibrate this epithelial–immune circuit and clarify when MYC-driven repair and CXCL10–CXCR3–mediated CD8⁺ NKT recruitment promote resilience versus precipitate maladaptive remodeling. Longitudinal approaches that integrate epithelial and immune profiling with direct assessments of neuroimmune function—such as vagal afferent activity, cough and laryngeal reflex sensitivity, and sensory nerve morphology—will be essential for establishing whether sustained cytotoxic surveillance disrupts sensory pathways in a manner that parallels human post-viral disease. In summary, our study spotlights basal Myc as a trigger for injury-timed control of epithelial growth, cell fate decisions, and chemokine-summoned influx of cytotoxic CD8⁺ NKT cells across the upper respiratory tract. This coordinated program likely contributes to efficient viral control and the establishment of local memory, but may also set the stage for persistent epithelial remodeling and neuroimmune dysregulation when dysbalanced.

Methods

Animals

All experimental procedures were performed in the American Association for Accreditation of Laboratory Animal Care (AAALAC)-certified laboratory animal facility at the University of California, San Diego (UCSD). All animal husbandry and experiments were conducted under approved Institutional Animal Care and Use Committee (IACUC) guidelines. Wild-type C57BL/6 (JAX 000664), Krt5^CreERT2 (JAX 029155), Ascl1^CreERT2 (JAX 012882), and Rosa^lxl-tdTomato (Ai14, JAX 007914) lines were purchased from the Jackson lab. Mycflox (MMRRC 036778) line was gifted from Rob Signer at UCSD. All the cre lines we used in this study were kept in C57BL/6 background. All cre driver lines in heterozygous form are viable and fertile, and no abnormal phenotypes were detected. Both male and female mice were used in the experiment. Adult mice were 8–12 weeks of age for all experiments. The experiments were not randomized, and the investigators were not blinded to allocation during experiments and outcome assessment.

Influenza A virus (IAV) infection

Influenza strain A/H1N1/PR/8 strain was obtained from ATCC (VR-95PQ). Briefly, mice were anesthetized with 3.5% isoflurane for 15 min until bradypnea was observed. Virus dissolved in 35 μl of PBS was pipetted onto the nostrils of anesthetized mice, whereupon they aspirated the fluid directly into their airways. Control and experimental groups were infected simultaneously in the same cohort by the same investigator so that direct comparison between groups was justified and appropriate.

Tissue collection and immunofluorescence

Mice were euthanized by CO2 inhalation followed by transcardial perfusion with PBS to remove circulating blood. The larynx-trachea was isolated and fixed overnight in 1%PFA or immediately embedded in OCT, flash frozen using 2-Methylbutane and liquid nitrogen and stored at −80°C. Using a cryostat, the larynx-trachea was then sectioned (12 μm) and stored at −20°C. Unfixed/lightly fixed tissue was then washed in PBS for 5mins to remove OCT, heated to boiling in 10 mM citrate buffer (pH 9) for antigen retrieval and treated with 0.5% Triton X-100 in PBS for 15mins. All sections were processed for immunostaining following a standard protocol.⁴¹ Ascl1 and Krt5 recombination was induced via Tamoxifen IP administration with dose of 100mg/kg for either two or three consecutive days. All primary and secondary antibodies used are listed in TableS2. Primary antibodies were applied overnight at 4°C, while secondary antibodies were applied for 1hr at room temperature. Sections were incubated with DAPI (1:1000 ratio) for 10 min at RT. Slides were mounted and coverslipped with Prolong Diamond mounting media (Fisher P36970), cured flat at room temperature in the dark for 24 h, and stored at 4°C. Each experiment was replicated at least twice for all timepoints and targets assessed.

Tissue processing and cell sorting

Whole upper airway (larynx-trachea) was mechanically dissociated by mincing tissue with razor blades in solution containing 5ml of RPMI1640 (Thermo Scientific) with 10% fetal bovine serum, 1mM HEPES (Life Technology), 1mM MgCl2 (Life Technology), 1mM CaCl₂ (Sigma), 0.5mgml⁻¹ collagenase D and type I/Dispase (Roche), and 0.25mg DNaseI (Roche). Minced tissue was then digested by shaking at around 150rpm for 30min at 37°C. Following incubation, upper airway pieces were mechanically dissociated further by straining through a MACS 70 μm filter. Red blood cells were removed by the addition of 1ml of RBC lysis buffer (BioLegend) to each tube and incubation at room temperature for 1min. Single-cell suspensions were pelleted (1,500rpm, 4°C, 5min) and stained with Fc blocking antibody (5mgml⁻¹, BD). For FACS, the following antibodies were used: (Immune) 1:1000 BV510-conjugated anti-CD45 (BioLegend, no. 103138), (Epithelial) 1:1000 APC-FITC-conjugated anti-Epcam (BioLegend, no. 118214), and (Endothelial) 1:1000 PE-conjugated anti-CD31 (BioLegend, no. 102508). Cells were then stained using live/dead dye (CD11b: Brilliant Violet 450, BioLegend, no. 75–0112-U025) before being resuspended in 2% FBS + 1:2000 DAPI. All FACS sorting for scRNA-seq library preparation was done on a BD FACSAria Fusion Sorter (BD Biosciences) analyzer with three lasers (405, 488 and 640nm) at the Flow Cytometry Core at VA San Diego Health Care System.

Bulk RNA-seq and data analysis

Total RNA from 10wk old adult WT upper airway (larynx/trachea) tissue was extracted as described above. cDNA libraries were constructed using Illumina TruSeq RNA Library Prep Kit V2 (Illumina) and sequenced on the HiSeq4000 platform (Illumina) at the Institute for Genomic Medicine (IGM) at UCSD. FASTQ files were aligned to the mouse reference genome (mm10) by using Bowtie2⁴² with default settings. Differential gene expression analysis was performed using Cufflinks.⁴³ P-values were adjusted by the Benjamini-Hochberg method to control False Discovery Rate (FDR) at 0.05. Heatmaps and volcano plots were generated by ggplot2 (version 3.3.2) using RStudio. (v. 2024.04.0 Build 735) running R 4.3.3 (R Core Team). Once exclusively differentially expressed genes were identified, we performed tests of enrichment using Gene Ontology (GO) annotations utilizing Enrichr (v2024)⁴⁴ as previously described.⁴⁵

10x Chromium scRNAseq and data analysis

Single-cell RNA sequencing (scRNA-seq) was performed using the 10× Genomics Chromium platform as previously described.10 Briefly, dissociated cell suspensions were loaded onto the Chromium Controller to generate single-cell gel bead emulsions, followed by cDNA synthesis and library preparation using the Chromium Single Cell 3′v3 kit (10X Genomics, Pleasanton, CA). Libraries were sequenced on an Illumina NovaSeq 6000, and raw reads were processed with Cell Ranger (v3.0.2, 10× Genomics) using the mouse reference genome (GRCm38) to generate single-cell gene barcode matrices. Influenza A Virus complete gene sequence was imputed into the reference genome for in silico transcript identification (TableS1). Downstream analyses—including quality control, normalization, clustering, and differential expression—were performed in Seurat (v4.0.5)^46,47 following the previously mentioned computational workflow.¹⁰ Unless otherwise noted, identical parameters and filtering thresholds were applied. To account for technical and biological covariates, cell cycle phase scores (S and G2/M), mitochondrial transcript percentage, and gene feature counts (nFeature_RNA) were incorporated into the dataset and regressed during data scaling to minimize their confounding effects on clustering.

To infer and quantify cell–cell signaling, we applied the CellChat R package (v1.x) to the processed single-cell transcriptomic data. Ligand–receptor interactions and their associated cofactors were evaluated using the curated CellChatDB database (mouse version, ~2,021 validated interactions).²⁶ Communication networks were represented as directed, weighted graphs, and visualized using hierarchical plots, circle plots, bubble plots, and chord diagrams. Centrality metrics (e.g. indegree, out-degree, betweenness) were computed to determine dominant signaling sources, targets, and mediators. Pattern recognition and manifold learning approaches were applied to identify global communication patterns and context-specific signaling differences across conditions.

To assess pathway-level activity across our IAV-enriched epithelial subpopulations, we applied the irGSEA R package (v2.1.5) for rank-based, single-cell gene set enrichment analysis.⁴⁸ Gene sets were derived from MSigDB (Hallmark and KEGG collections) (version 7.4.1). Normalized expression matrices from Seurat objects were imported into irGSEA, and enrichment scores were calculated using the AUCell (version 1.14.0), UCell (version 1.1.0), singscore (version 1.12.0), GSVA (version 1.40.1) and Viper (version 1.32.0) algorithms implemented within the package. Scores were computed on a per-cell basis and summarized by cell type and condition. Differential pathway enrichment was visualized as violin and dot plots to highlight compartment-specific activation patterns. All parameters followed default settings in irGSEA unless otherwise specified.

Cellular lineage and differentiation trajectories were reconstructed using Monocle3 (1.4.26) to infer dynamic transcriptional transitions within epithelial IAV-enriched subpopulations.49 Briefly, Seurat-processed data were converted to a cell_data_set object using the as.cell_data_set() function, retaining normalized expression values and annotated metadata. Dimensionality reduction was performed using UMAP and PCA embeddings previously generated in Seurat. Cells were ordered in pseudotime using learn_graph() and order_cells(), with root nodes defined by basal populations. Pseudotime-associated genes were identified using graph_test(), and trajectory-dependent gene expression patterns were visualized using plot_cells() functions. Default Monocle3 parameters were used unless otherwise noted.

RNAscope in situ hybridization

All staining procedures were performed using the RNAscope Fluorescent Multiplex Kit V2 (Advanced Cell Diagnostics, no. 323100) following the manufacturer’s instructions. The following probes from Advanced Cell Diagnostics were used: Mm-IAV (no. 521181), Mm-Myc (no. 413451), Mm-Cxcl10 (no. 408921), Mm-tdTomato (no. 317041), Mm-Gzmb (no. 490191), Mm-Ccl5 (no. 469601).

Cell proliferation and apoptosis

Cell proliferation was assessed using the Click-iT^™ EdU Cell Proliferation Kit (C10337, Invitrogen), which detects DNA synthesis through copper-catalyzed azide–alkyne cycloaddition between incorporated 5-ethynyl-2′-deoxyuridine (EdU) and a fluorescent azide probe. Adult mice received 1 mL of a 400 μM EdU solution (Invitrogen, diluted in PBS) by intraperitoneal injection, and upper airway tissues were collected 1 hour later. Samples were fixed in 1% paraformaldehyde (PFA) overnight at 4°C, washed in PBS for 30 minutes, and embedded in OCT before cryosectioning and EdU detection according to the manufacturer’s protocol. Apoptosis was measured using the Click-iT^™ Plus TUNEL Assay (C10245, Invitrogen), which labels DNA strand breaks via terminal deoxynucleotidyl transferase (TdT)–mediated incorporation of alkyne-modified nucleotides, followed by fluorescent azide detection using the same click-chemistry–based reaction. Tissue preparation, fixation, and embedding were identical to those described for EdU staining, and TUNEL labeling was performed following the manufacturer’s instructions.

Lineage tracing

Tamoxifen dissolved in corn oil was administered via intraperitoneal injection at 25 mg/kg once daily for three consecutive days.

Cell count quantification

All cell counts and area measurements were quantified at 20× magnification from 3–4 whole-mount coronal images spanning the pharyngolaryngeal-to-tracheobronchial axis (Fig. S1B). Images were analyzed using an automated pipeline in QuPath (v0.6.0). Briefly, the epithelium was manually annotated as a region of interest (ROI) and subdivided into regional compartments, including tracheal surface epithelium (SE), subglottic SE, supraglottic SE, and submucosal glands (SMG). For marker-positive cell counts or total feature area (μm²), the QuPath create_threshold function was applied within each annotated epithelial ROI to accurately label the target signal. The cell_detection function was then used within the same ROI to identify DAPI⁺ nuclei and determine total cell numbers. Threshold-defined feature counts were normalized to DAPI⁺ counts to calculate the percentage of positive cells; alternatively, threshold-defined feature area was recorded within each ROI and reported as total area (μm²) or expressed in arbitrary units (AU) as indicated.

Statistics

All analyses were performed using GraphPad Prism 10. Data were first assessed for normality (Shapiro–Wilk) and homogeneity of variance (F test) to determine whether parametric or non-parametric tests were appropriate. For normally distributed data with equal variance, statistical comparisons used t-tests (two groups) or one-way/two-way ANOVA (multiple groups or multiple variables), followed by the appropriate post-hoc corrections (Dunnett’s, Tukey’s, or Fisher’s LSD) as indicated in figure legends. When normality or variance assumptions were not met, non-parametric equivalents were applied. Results are reported as mean ± SE, and significance was defined as p ≤ 0.05.

Data and code availability

Bulk and single cell RNA-seq data have been deposited at the Gene Expression Omnibus (GEO) and are publicly available as of the date of publication. All data generated supporting the findings of this study are available in the manuscript. Further information is available from the lead author upon reasonable request. This paper does not report original code.