{"paper_id":"06b28eb1-e078-47bb-99dc-1f97220a330f","body_text":"A Peripheral Neuron-to-Microglia Signaling Axis Connecting Transient Viral Infection to Persistent Neuroinflammatory States | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Biological Sciences - Article A Peripheral Neuron-to-Microglia Signaling Axis Connecting Transient Viral Infection to Persistent Neuroinflammatory States Benjamin tenOever, Justin Frere, Jonathan Dever, Leon Hsieh, Caroline Ubbud, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7179724/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Acute viral infections can cause lasting symptoms in anatomically distant, uninfected tissues, a phenomenon that challenges traditional notions of viral pathogenesis. A leading example is long COVID, a condition in which neurological and other clinical symptoms can materialize long after viral clearance 1 . In investigating the mechanisms underlying this phenomenon, we found that despite evading direct infection, olfactory sensory neurons (OSNs) demonstrate a progressive slow decline following SARS-CoV-2 infection, triggering a prolonged neuroinflammatory response that persists for weeks to months post viral resolution. Using both small animal models and human clinical samples, we demonstrate that the virus selectively infects sustentacular cells in the olfactory epithelium (OE), leading to structural disruption and secondary OSN loss. Axonal debris from degenerating OSNs accumulates in the olfactory bulb (OB), where it triggers sustained activation of resident microglia cells and persistent inflammatory signaling. These immune responses are spatially restricted to OB regions innervated by the damaged neurons and are marked by transcriptional programs involved in phagocytosis, synaptic remodeling, and debris clearance. Together, these findings delineate a conserved neuron-to-glia injury axis in which peripheral neuronal damage initiates a protracted cascade that, despite the absence of direct central nervous system infection, culminates in delayed and persistent neuroinflammation. This mechanism offers a unifying framework for how transient respiratory infections can lead to persistent neurological sequelae, including those seen in long COVID. Biological sciences/Microbiology/Virology/SARS-CoV-2 Biological sciences/Immunology/Neuroimmunology Figures Figure 1 Figure 2 Figure 3 Figure 4 Full Text Acute viral infections are increasingly recognized for their capacity to induce sustained symptoms that persist well beyond the period of viral clearance 2 . Among the most prominent examples is SARS-CoV-2, the causative agent of COVID-19, which has drawn attention to the poorly understood relationship between acute infection and long-term clinical sequelae. Despite being largely restricted to the respiratory tract, SARS-CoV-2 has consistently demonstrated the ability to affect distal tissues, both during active infection and after viral resolution 3 . Indeed, while prevalence estimates vary, the World Health Organization reports that approximately 10–20% of individuals experience symptoms beyond three months after resolution of SARS-CoV-2 infection 4 . Together these post-viral complications have collectively come to be referred to as post-acute sequelae of SARS-CoV-2 infection (PASC), or more colloquially as long COVID. Notably, neuropsychiatric symptoms rank among the most frequently reported features of long COVID 5-7 . Meta-analyses indicate that more than 20% of post-COVID patients report ongoing issues such as sleep disturbance, cognitive impairment, fatigue, anxiety, depression, pain, and prolonged anosmia, or loss of smell 8 . The underlying causes of these persistent symptoms remain elusive, in part because multiple studies have shown that SARS-CoV-2 is rarely seen to infect neural tissues in vivo 9 . Both clinical and experimental studies into alternative methods of virus-induced damage have implicated a range of potential contributory processes, including thromboembolic events 6,10 , complement dysregulation 10-12 , altered serotonergic signaling 13 , reactivation of latent herpesviruses 14,15 , sustained inflammatory activity 1,16 , immune dysregulation 17,18 , residual viral reservoirs 19 , and lasting damage incurred during acute infection 20,21 ; however, despite this evidence, it still remains unclear how a transient viral infection may be able to mediate these processes at times following viral clearance and throughout tissues spared from direct infection. Interestingly, recent evidence has further implicated olfactory disruption as a potential driver of long COVID neurocognitive symptomology 1,22 . Olfactory dysfunction is a common feature of SARS-CoV-2 infection, with anosmia affecting approximately 40% of unvaccinated and infection-naïve individuals with COVID-19 23 . Anosmia is thought to be due to robust SARS-CoV-2 infection of the OE in the upper respiratory tract during early infection. The OE is home to OSNs that express olfactory receptors on their highly ciliated apical surfaces and mediate primary olfaction by binding odorant molecules in the nasal lumen 24 . OSNs, which are largely supported in the OE by sustentacular cells, transmit these odorant signals through long axonal projections which extend from their cell bodies in the OE through the cribriform plate to the glomerular layer (GL) of the OB, where they synapse with mitral cells and other second-order neurons ( Fig. 1a ) 24 . Although the sustentacular cells of the OE are highly susceptible to SARS-CoV-2, multiple human and animal studies have shown that OSNs are largely resistant to infection due to minimal expression of the viral entry receptor angiotensin-converting enzyme 2 (ACE2) 25-27 . Anosmia is therefore believed to result from epithelial disruption and local inflammation, rather than direct neuronal infection, a model supported by the transient nature of smell loss in most patients 28 .Despite some evidence suggesting limited infection of neuronal cells, recent studies have shown that COVID-associated olfactory disturbance is associated with significant alteration of neuronal biology in the central nervous system, and these changes can be observed to persist to time points well beyond viral clearance 1,22 . Again, however, while these pathologic changes have been noted, the particular mechanism by which a largely non-neurotropic SARS-CoV-2 virus can drive these phenotypes remains unclear. Given the significant viral load that accumulates in the OE and the functional link that infection of this tissue appears to have with central nervous system dysfunction, we here sought to investigate whether a mechanistic link may exist to connect these phenomena. To investigate how SARS-CoV-2 infection of the OE might initiate biological changes leading to persistent neurological symptoms, we utilized the golden hamster model, widely regarded as the most physiologically relevant small animal system for COVID-19 29 . To this end, we characterized the temporal dynamics and tissue specificity of viral replication following intranasal challenge of SARS-CoV-2 over a 30-day period. Viral RNA levels in the OE increased sharply within 24 hours post-infection and remained elevated for ten days before returning to baseline ( Fig. 1b ). In parallel, the lower airways of the same animals exhibited similarly high viral RNA levels early after infection, but these declined more rapidly, dropping by three orders of magnitude by day 7 and becoming undetectable by day 10. ( Fig. 1c ). Infection of the OE could also be confirmed in an independent cohort of animals using immunohistochemistry (IHC) for SARS-CoV-2 nucleocapsid (N) and spike (S) proteins ( Extended Data Fig. 1a ). Notably, viral proteins were cleared from the OE more rapidly than RNA, with near-complete resolution by day 5, suggesting either higher sensitivity of RNA-based detection or delayed clearance of viral RNA following protein clearance ( Fig. 1b and Extended Data Fig. 1a ). Co-staining with the olfactory marker protein (OMP), an OSN marker, revealed minimal colocalization with viral proteins, corroborating past studies suggesting that SARS-CoV-2 infection in the OE is largely restricted to non-neuronal sustentacular cells ( Fig. 1d ) 25 . To assess the host response to SARS-CoV-2 infection in the OE, we next examined the temporal dynamics of antiviral and immune signaling. RT-qPCR analyses revealed robust induction of immune-related genes, including Isg15 , a canonical interferon-stimulated gene (ISG); Ccl5 , a chemokine involved in immune cell recruitment; Iba1 , a marker for microglia and macrophages; and Cd3e , a T cell-specific marker ( Fig. 1e-h ). All four genes were strongly upregulated during the early phase of infection, with peak expression occurring between 1 and 5 days post-infection (dpi), coinciding with maximal viral replication. These trends were also observed in the lung, although immune gene expression in the OE was more pronounced and slower to resolve ( Extended Data Fig. 1b-e ). Protein-level validation confirmed these findings, with IHC detecting increased expression of the interferon-inducible MxAprotein and CD3 T cell marker in the OE ( Fig. 1i and Extended Data Fig. 1f ). To assess whether these olfactory immune responses were conserved across SARS-CoV-2 variants or when viral biology was muted, we performed transcriptional profiling of OE tissue from hamsters infected with the original Washington strain (WA1), a cleavage site-deficient mutant (WA1 ΔPRRA), or the BA.5 Omicron variant at peak infection (4 dpi). Despite differences in viral genotype and pathogenesis, all three strains elicited nearly identical type I interferon (IFN-I) responses ( Fig. 1j ), with comparable patterns of immune gene expression and pathway enrichment, including markers of immune cell infiltration ( Extended Data Fig. 1g-h ). These data suggest that SARS-CoV-2-induced OE inflammation represents a conserved host response that is not variant or viral fitness dependent. Given the extent of SARS-CoV-2 infection in the OE, we next assessed its impact on OSNs and overall epithelial architecture. Immunostaining for OMP on fixed cross-sections of the OE revealed marked structural disruption during peak infection ( Fig 2a ). OSNs exhibited a progressive loss of their characteristic ciliated morphology, accompanied by detachment from the epithelial surface tracking with viral infection and immune response. Notably, OMP⁺ cell bodies were frequently observed within the nasal lumen, consistent with the accumulation of sloughed neuronal debris and widespread epithelial disorganization. Quantification of changes to OE structure demonstrated a significant thinning of the epithelial layer during this period, with many areas of OE showing near complete OSN ablation between 5-10dpi (Fig. 2b and Extended Data Fig. 2a) . Following viral clearance, evidence of epithelial repair became apparent in stained OE sections. Quantification of OE thickness and OMP⁺ OSN cell counts revealed a gradual recovery that eventually approached baseline levels, but only by 30dpi ( Fig. 2a-b ). This regenerative process was mirrored by an increase in Ki67 expression, a marker of cellular proliferation, localized to the basal layer of the OE, where progenitor and stem cell populations reside ( Fig. 2c ) 30 . Notably, Ki67 expression returned to baseline by 14 dpi, while the recovery of mature OSNs lagged behind, indicating a temporal gap between stem cell activation and neuronal differentiation, consistent with previous observations 30 . This delayed repopulation of the olfactory neuroepithelium may help explain why some individuals experience persistent anosmia following SARS-CoV-2 infection 31 . To further investigate the functional consequences of SARS-CoV-2 infection on OSNs and their progenitors, we compared mock to SARS-CoV-2 infected OE by single-cell RNA sequencing (scRNA-seq). This approach resolved distinct OSN populations and their lineage-associated precursors, with trajectory analysis confirming expected differentiation hierarchies ( Fig. 2d and Extended Data Fig. 2b-c ). Across cell types, infection elicited a broadly uniform transcriptional response, marked by a peak in ISG and chemokine expression at 3dpi ( Fig. 2e) . These immune signatures inversely correlated with the expression of genes involved in axoneme assembly, the structural foundation of motile cilia, suggesting that inflammation may suppress ciliogenic programs during peak infection ( Fig. 2f and Extended Data Fig. 2d ). Given that odorant receptors are localized to the cilia of OSNs 32 , we assessed their expression levels and observed significant downregulation at 1, 3, and 5 days post-infection in both scRNA-seq and bulk RNA-seq datasets ( Extended Data Fig. 2e–g ). Notably, while transcriptional repression of odorant receptors was evident following infection with both the original SARS-CoV-2 strain (WA1) and the Omicron variant, the extent of downregulation was consistently reduced in Omicron-infected samples, consistent with reports indicating that this variant induces smell loss at a reduced rate compared to other variants 33 , 34 . While these findings provide a molecular explanation for infection-induced anosmia and align with previous studies 25 , they also reveal that these epithelial abnormalities largely resolve by 14dpi. This transient nature of OE disruption therefore fails to account for the persistent neuroinflammation observed both in this same small animal model as well as some individuals long after viral clearance, which occurs within 7dpi 1,35,36 . To explore this disconnect, we assessed the impact of SARS-CoV-2 infection on neuronal biology in the OB, the first central nervous system structure to receive input from OSNs and a potential site for longer-term immune engagement. To this end, we first examined the expression of Fos, a canonical immediate early gene (IEG) and established marker of neuronal activity 37 . IEGs are rapidly and transiently induced in response to cellular stimuli and play essential roles in synaptic plasticity, learning, and memory 37 . RT-qPCR analysis revealed a marked reduction in Fos transcript levels in the OB coinciding with the onset of OE damage and OSN loss. Notably, this suppression persisted throughout the 30-day post-infection period, indicating a sustained depression of OB neuronal activity following SARS-CoV-2 exposure ( Extended Data Fig. 3a ). To further characterize this response, we performed bulk RNA sequencing of the OB across the same time course. Differential expression analysis confirmed that, like Fos , other odorant-evoked neuronal IEGs that have been reported, such as: Arc, Egr1, Egr3, Fosb, Npas4 , and Nr4a1 38 , were also significantly downregulated ( Fig. 3a and Extended Data Fig. 3b ). Together, these findings suggest that SARS-CoV-2 infection induces broad and long-lasting suppression of neuronal activity in the OB, potentially disrupting neurobiological functions well beyond the period of viral clearance. To better define the molecular underpinnings of sustained OB dysfunction, we performed gene set enrichment analysis (GSEA) on our longitudinal OB RNA-sequencing data. This analysis revealed persistent upregulation of IFN-I and chemokine signaling pathways, with a notable late-phase intensification emerging at 14dpi and persisting through at least 30dpi ( Fig. 3b-c ). These prolonged immune signatures in the OB stood in sharp contrast to the transient and steadily resolving responses observed in the OE ( Fig. 2e, Extended Data Fig. 2d ), indicating a decoupling of peripheral viral clearance from central immune activity. These data were further validated by RT-qPCR within an additional cohort of infected animals using Isg15 and Ccl5 as proxies for the innate antiviral response ( Extended Data Fig. 3c-d ). To determine whether this prolonged transcriptional activity reflected ongoing viral presence, we next examined OB tissues for evidence of SARS-CoV-2. RT-qPCR targeting subgenomic nucleocapsid (sgN) RNA was negative except for trace positivity at 4dpi, the peak of systemic viral burden and well before the late-phase intensification of OB immune activation; further IHC for S and N protein revealed an absence of viral protein in the OB throughout the observed period ( Fig. 3d and Extended Data Fig. 3e ). These findings align with previous reports indicating that direct infection of the central nervous system by SARS-CoV-2 is minimal to absent and point instead to a non-infectious trigger of OB inflammation 1,35,39 . Given the absence of direct OB infection, we next sought to identify the cellular mediators of this delayed inflammatory response. We interrogated our GSEA datasets for signatures of immune cell infiltration and observed significant enrichment for transcriptional profiles of microglia, the resident macrophage population of the brain ( Fig. 4a ). Notably, microglial signatures peaked at 4 dpi, declined by 5 dpi, and then rose again to a second peak at 14 dpi—following viral clearance from the olfactory epithelium—and remained elevated through 31 dpi, closely paralleling the temporal dynamics of interferon and chemokine pathway activation ( Fig. 1b, 3b-c and Extended Data Fig. 1a) . RT-qPCR and IHC validation further confirmed sustained upregulation of the macrophage/microglia marker gene Iba1 beyond 30dpi in an independent cohort ( Extended Data Fig. 4a-b ). To further characterize IFN-I signaling in the OB, tissue sections from SARS-CoV-2-infected hamsters were stained for MxA, a canonical ISG 40 , showing robust MxA induction throughout the 30dpi period ( Fig. 4b) . Importantly, these responses were most prominent at the outermost layers of the OB, particularly within the olfactory nerve layer (ONL) and glomerular layer (GL), regions enriched for OMP + axonal projections from OSNs in the OE ( Fig. 4b ) 41 . To refine these observations, OB sections were stratified into OMP + (ONL and GL) and OMP – (external plexiform layer (EPL), mitral cell layer (MCL), internal plexiform layer (IPL), and granule cell layer (GCL)), regions, and levels of Iba1 and MxA expression were independently quantified in each compartment ( Extended Data Fig. 4c ). These analyses revealed significantly elevated expression of both gene products through 14dpi, with the strongest signals consistently localized to the OMP + regions ( Fig. 4c-f, Extended Data Fig. 4d-g ). Notably, fluorescent co-localization of MxA and IBA1 demonstrated that microglia were the predominant source of MxA expression at later time points, and that this activation was sustained in microglia well beyond viral clearance ( Fig. 4g ). Moreover, while MxA + microglia were observed throughout the OB, their staining intensity was markedly higher in the OMP+ region, suggesting that proximity to OSN axons may shape long-lasting microglial activation. Having implicated microglia as the probable drivers of sustained immune activity in the OB, we next aimed to more fully define the nature of their activation. To this end, we analyzed OB RNA-Seq time course data to profile transcriptional signatures associated with core microglial functions. These analyses revealed robust and persistent upregulation of genes involved in phagocytosis and antigen presentation (e.g., Aif1, Cd68, Cd14, Fcgr1a, Fcgr2b, Itgam, Itgb2, Ptprc, Cd74, Cd209a, B2m ), innate immune sensing and inflammatory signaling (e.g., Tlr2, Tlr4, Siglec1, Syk, Il1b, Tnf, Il18, Cx3cr1 ), immune cell adhesion and trafficking (e.g., Itgb1, Lyx ), and apoptotic cell clearance ( Trem2, Cd36, Mrc1, Cd163 ), reflecting a broad immune activation state associated with post-injury homeostasis ( Fig. 4h-j ). In addition to the above microglia signatures, arguably the most striking feature was the sustained upregulation of genes associated with the classical complement cascade, a pathway increasingly recognized for its role in activity-dependent synaptic pruning 42 . Notably, we observed persistent elevation of C1qa, C1qb, C1qc , and downstream components such as C3, C4a, C3ar1 , and C5ar1 , all of which have been implicated in tagging synapses for elimination during development and in response to injury ( Fig. 1h; Extended Data Fig. 4h ). This transcriptional profile suggests that microglia in the OB remain engaged in complement-mediated remodeling well beyond the resolution of acute infection, potentially contributing to circuit refinement or, conversely, maladaptive synaptic loss. Given this, we next sought to identify a potential trigger for these activities. Here, the localization of microglial recruitment and immune signaling to the outer OB provided an important clue that OSN projections may be serving as this trigger. Following initial SARS-CoV-2 infection in the OE, OSN cell bodies were observed to be damaged and cleared from the OE. Despite this clearance, OSN axonal projections were still notably visible feeding into the outer OB, suggesting a potential delayed clearance of these OSN components ( Fig. 4k ). Quantification of OMP staining intensity in the outer OB confirmed this delayed clearance, as OMP signal was largely sustained in the OB until ~14dpi, at which point marker intensity decreased significantly. This significant decrease was sustained out to 30dpi ( Fig. 4l ) and notably aligned with the upregulation of inflammatory genes and genes important to microglial debris clearance ( Fig. 4h-j ). Further evidence of this delayed clearance of OSN components in the OB could be seen in comparison of bulk sequencing data from the OB and OE. These data showed that at early time points following SARS-CoV-2 infection (4dpi), marker genes of OSN cells decreased at significantly greater rates in the OE than the OB, suggesting delayed degradation of OB OSN components ( Fig. 4m ). Analysis of full time-course sequencing data showed that maximum negative enrichment of OSN marker genes in the OB was delayed until at least 8dpi (Extended Data Fig. 4i ). Together, these data suggest that OSN axonal projections remain fixed in tissue despite loss of OSN cell bodies due to SARS-CoV-2 infection; microglia may thus be recruited and activated to clear this neuronal debris. Although molecular profiling of the human OB following COVID-19 remains challenging, primarily due to the difficulty of obtaining this tissue post-mortem, several datasets are available. To explore whether the immune signatures observed in our animal models are recapitulated in humans, we conducted a meta-analysis of publicly available mass spectrometry data from post-COVID donor OBs. This analysis revealed consistent upregulation of IFN-I signaling and antigen presentation pathways, closely mirroring the transcriptional patterns observed in our preclinical studies ( Extended Data Fig. 4j ). To further support these findings, we examined RNA-sequencing data from OB tissue collected from two additional post-COVID donors, which likewise showed elevated expression of genes associated with microglial activation and phagocytic activity, including TREM2, FCGR1A, CD209, CD68, ITGAM, FCGR2B, and TLR4 ( Extended Data Fig. 4k ). While the limited number of human samples tempers definitive interpretation, these converging signals suggest that the same immune processes observed in model systems may also be operative in the human brain following SARS-CoV-2 infection, offering a potential mechanistic link to the lingering neurological symptoms reported in long COVID. Although SARS-CoV-2 primarily targets the respiratory tract, clinical and imaging studies have shown that even mild infections can produce persistent structural and functional brain changes detectable long after viral clearance 39 . These findings have led to the hypothesis that a persistent viral reservoir underlies post-acute symptoms, but evidence of a sustained and functionally replicative viral reservoir in the brain is currently lacking 43 . Challenging this concept, here, we propose an alternative mechanism by which a transient respiratory infection leads to lasting neurological symptoms. We show that SARS-CoV-2-induced damage to the OE causes bystander loss of OSNs. Due to the unique anatomy of OSNs, whose axons traverse the cribriform plate to the OB, this peripheral injury extends centrally. As a product of their length, degenerating OSNs take time to be cleared, triggering sustained microglial activation and persistent immune signaling. This phenomenon has notably also been observed in other rodents and zebrafish, suggesting induction of a conserved damage response to OSN injury 44-46 . Together, these findings outline a conserved mechanism by which a self-limited peripheral infection can drive prolonged, localized neuroinflammation. By uncovering a damage-mediated glial response in the absence of direct viral invasion and temporally removed from acute peripheral infection, this work provides insight into the pathophysiology of long COVID and broader principles by which acute peripheral insults can durably alter brain function. This work therefore highlights new therapeutic opportunities for post-viral and neuroinflammatory syndromes. Declarations Acknowledgments We would like to thank the Zegar Family Foundation and the Conestoga Road Foundation for enabling these studies. Author contributions The study was designed by J.T.F. and B.T. The experiments were performed J.T.F, J.D., L.H., C.A., S.U., R.A.S., T.T.W, and C.C.C. Data analysis was performed by J.T.F. Preparation and writing of the manuscript was performed by J.T.F, J.M., and B.T. Competing interests The authors declare that they have no competing interests. Materials and Methods Viruses and cells SARS-CoV-2 (USA-WA1/2020) (NR-52281) (Biodefense and Emerging Infections Research Resources Repository, BEI Resources) was propagated in Vero-E6 cells (American Type Culture Collection) that were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 2% fetal bovine serum (FBS) (Millipore Sigma), 1 mM Hepes (Lonza Bioscience), and 1% penicillin/streptomycin (Thermo Fisher Scientific). Propagation culture supernatants were filtered using Amicon Ultra-15 centrifugal filter units (Sigma-Aldrich). Viral stocks were sequenced to ensure that the furin cleavage site was maintained and then stored at -80°C until use. Following single use, stocks were discarded to prevent freeze-thaw cycling of viral particles. Infectious viral titers were determined via plaque assay in Vero-E6 cells with an overlay of DMEM supplemented with 2% FBS, 1 mM HEPES, and 0.7% Oxoid agar (Thermo Fisher Scientific). Plaque assays were fixed with 4% paraformaldehyde (PFA) (Thermo Fisher Scientific) prior to staining with crystal violet solution (Sigma-Aldrich). All infections in these studies were performed using passage three or four SARS-CoV-2 virus. Hamster Studies Male Syrian golden hamsters ( Mesocricetus auratus ) aged six to seven weeks were obtained from Charles River Laboratories. Hamsters were allowed to acclimate to U.S. Centers for Disease Control and Prevention (CDC)- and U.S. Department of Agriculture (USDA)-approved biosafety level 3 (BSL-3) facilities at either Icahn School of Medicine at Mount Sinai (ISMMS) or New York University Langone Health (NYULH) for at least 7 days prior to inclusion in experiments. Hamsters were randomly assigned into treatment groups and treated intranasally with 100uL of either phosphate-buffered saline (PBS) (Gibco) or a viral stock containing 1000 plaque-forming units (PFU) of SARS-CoV-2 (diluted as needed in PBS) under ketamine/xylazine anesthesia. Hamsters were euthanized at various time points following infection via intraperitoneal injection of pentobarbital and cardiac perfusion with either 60 mL of cold PBS (for experiments where tissues were taken for RNA-based assays) or 30 mL of cold PBS followed by 30 mL of cold 4% PFA (for experiments where tissues were taken for exclusively histology-based assays). Tissues set to be analyzed via RNA-based assays were harvested into TRIzol (Thermo Fisher Scientific) and homogenized in Lysing Matrix A homogenization tubes (MP Biomedicals) using a Fast-Prep-24 5G bead grinder and lysis system (MP Biomedicals). Samples were homogenized for 40 seconds at 6 m/s for two cycles. Tissues harvested for histological assays were collected into 4% PFA and allowed to fix for at least 72 hours before being washed three times in PBS to prepare for downstream processing. All animal experiments were performed according to protocols approved by the IACUCs and Institutional Biosafety Committees at ISMMS and NYULH. All infections and handling of infectious material took place in BSL-3 facilities. Quantitative Reverse Transcription Polymerase Chain Reaction (RT-qPCR) RNA was isolated from tissues homogenized in TRIzol via standard phenol-chloroform extraction. One microgram of extracted total RNA was reverse transcribed using the SuperScript III reverse transcriptase system and oligo(dT) primers (Thermo Fisher Scientific). The resulting cDNA was assayed via qRT-PCR using the KAPA SYBR Fast Master Mix (KAPA Biosystems) and a Thermo Fisher Quantstudio 7 Instrument (Thermo Fisher Scientific). Delta-delta cycle threshold or inverted raw Ct values were used to compare transcriptional changes between SARS-CoV-2- and mock-treated hamsters as detailed in the primary text. Graphical representations of these data were generated using GraphPad Prism (version 10) (GraphPad Software). Decalcification of Snout Tissues and Histological Processing Prior to paraffin embedding, formalin-fixed head and snout tissues were processed as detailed above (see “Hamster Studies”) and were washed in PBS three times. These tissues were then dissected and trimmed to leave only connected bony snout and anterior cranial cavity (containing the olfactory bulb). These snout and anterior cranial cavity tissues were then placed into at least five-fold volumetric excess of EDTA Bone Decalcifier Buffer (pH 7.4) (alternatively labeled “Buffered Versenate, pH 7.4”) (StatLab) and placed on a rocker (10 rpm) at room temperature for 10 days. EDTA Bone Decalcifier Buffer was replaced with fresh buffer every two days during this incubation. Following the tenth day of incubation, snout tissues were rinsed three times in PBS and then embedded in paraffin. Paraffin blocks were cut into 5uM sections using a microtome and mounted onto charged glass slides for downstream processing. Immunofluorescent Staining and Imaging Slide-mounted sections were processed for immunofluorescent staining as described previously 1,20 . In brief, sections were deparaffinized via submersion in xylene and rehydrated in gradated ethanol solutions. Antigen retrieval was performed by submerging slides in IHC-Tek Epitope Retrieval Solution (IHC-Tek cat. IW-1100) and steaming for 45 minutes in the IHC-Tek Epitope Retrieval Steamer (IHC-Tek cat. IW-1102). Tissue was blocked with a solution of tris-buffered saline (TBS) (Fisher Scientific) supplemented with 10% goat serum (Millipore Sigma) and 1% bovine serum albumin (Thermo Fisher Scientific). Primary antibodies were diluted in TBS supplemented with 1% bovine serum albumin and added to slides. Slides were incubated overnight at 4° C and then washed with a solution of TBS supplemented with 0.025% Triton X-100 (Thermo Fisher Scientific). Fluorophore-conjugated secondary antibodies (goat anti-rabbit IgG AlexaFluor 568 and goat anti-rabbit IgG AlexaFluor 647 [Thermo Fisher Scientific, Cat# A-11011 and A-21245, respectively]) were diluted to a concentration of 1:1000 in TBS with 1% bovine serum albumin and added to slides as appropriate given primary antibody staining. Slides were incubated at room temperature in the dark for 1 hour and were then washed in TBS supplemented with 0.025% Triton X-100. Section nuclei were then stained with 4′,6-diamidino-2-phenylindole (DAPI) and rinsed in PBS prior to coverslipping with ProLong Diamond Antifade Mountant (Thermo Fisher Scientific). Slides were imaged using a Zeiss Axio Observer epifluorescence and crossed-polarization inverted microscope platform with initial processing and stitching using Zeiss ZEN (Blue Edition) software (Carl Zeiss Microscopy). Bulk RNA Sequencing (FDA, retrieve our own data) RNA-sequencing FASTQ files analyzed in these experiments were retrieved from NCBI GEO datasets, GSE161200 and GSE203001. Each dataset contained longitudinally sampled olfactory tissues from SARS-CoV-2- and control-treated animals. Following data retrieval, RNA-sequencing reads were aligned to the Mesocricetus auratus transcriptome (assembly BCM_Maur_2.0) using Salmon 1.10 software. Differential expression analysis was then conducted on each experiment independently using DESeq2 to compare SARS-CoV-2-infected tissues to control tissues. All samples were compared to internal experimental controls, and all log2(Fold Change) results comparing infected to uninfected subjects were obtained from experimentally matched samples. All genes with P adj < 0.1 were considered “differentially expressed genes” (DEGs). Graphical representations of data were generated using the R programing language and the ggplot2 and heatmap packages. Single-Cell RNA-Sequencing Analysis Raw FASTQ files generated from single-cell RNA sequencing of longitudinal SARS-CoV-2-infected and control hamster olfactory epithelium (OE) tissues were retrieved from the 4DN data portal 25 . These files were processed and aligned to the Mesocricetus auratus genome (assembly and annotation BCM_Maur_2.0) using Cell Ranger (10x Genomics). The resulting files were processed in Seurat 5.1.0 using the R programming language (The R Foundation). Cells with mitochondrial gene expression surpassing 5% of the total transcriptome and with greater than 6,000 or less than 500 uniquely expressed genes were excluded from analysis in Seurat. Samples were normalized using the SCTransform method and subsequently integrated. Cellular clusters were defined with 30 dimensions using a resolution of 0.5. Conserved cluster-specific markers were used to determine cluster identity. Differential expression analysis was then conducted using the FindMarkers function in Seurat to compare cell-type-specific transcriptional changes in all genes between infected and uninfected conditions. Genes in each respective differential expression analysis were ranked by Log2(Fold Change) and analyzed with preranked Gene Set Enrichment Analysis (GSEA) conducted by the fgsea R package. Gene sets assessed included the pre-curated C5 (ontology) and C2 (curated) gene set collections maintained by the Molecular Signatures Database (MSigDB) (UC San Diego, Broad Institute). Additional analyses included cellular trajectory (pseudotime) analysis, which was conducted on identified neuronal and precursor populations using the Monocle 3 package in R, and statistical analysis of individual or clustered gene expression, which was conducted with the ggpubr R package. Visualizations of single-cell RNA-sequencing data were created using the R programming language and the ggplot2, Seurat, Monocle3, and fgsea packages. Quantification of Microscopy Data Following stitching and initial processing of microscopy images in the Zeiss ZEN (Blue Edition) software, imaging files were imported into QuPath (0.5.1) for further processing and quantification. All images that were part of a staining set were combined into a QuPath project and processed analogously to allow consistency in visualization. For quantification of OE thickness, distance was measured from the luminal surface of the epithelium to the lamina propria on slide sections stained for olfactory marker protein (OMP), a marker for olfactory sensory neurons (OSNs). Measurements were taken at 100mm intervals on the epithelial layer overlaying the septum or medial-most wall of each nostril, and these values were averaged for each subject to give a mean value of OE thickness. For measurement of all cells and of OSNs per specific length of OE, the epithelial layer of the septum or medial-most wall of each nostril was encircled using the QuPath “Closed Polygon Annotation” tool, capturing all cells between the luminal surface of the OE and the lamina propria. Nuclei and OMP positivity within this capture area were algorithmically quantified using the QuPath “Positive Cell Detection” tool. For OE quantification images, nuclei were labeled with DAPI, and OMP was labeled with Cy5. “Positive Cell Detection” tool settings reflected this labeling and cells were designated as OSN positive if the mean Cy5 intensity in the nucleus plus a 2um expansion zone surpassed a threshold of 1,000 Cy5 brightness intensity units. This threshold was verified visually by the experimenter to appropriately capture cells that were OMP positive (OMP+) and exclude cells that were OMP negative (OMP-). The luminal surface of the encircled olfactory epithelial layer was also traced to measure the span of OE captured. Reported values (cells per length of OE; OSNs per length of OE; fraction of OMP+ OE cells) were calculated from these measurements. Analogous protocols were followed for analysis and quantification of Ki67 (rather than OMP) positivity. For quantification of areas of “bare” OE, the entire length of the luminal surface of all visible nostril epithelium was traced and measured in OMP-labeled sections. These same surfaces were then subsequently re-traced, but only in areas where OMP+ cells could be visually observed in the epithelium underlying the traced surface. “Bare” lengths of membrane were calculated by subtracting the OMP+ luminal surface length from the total luminal surface length. Quantification of OMP (Cy5), MxA (Alexafluor 568), and IBA1 (Alexafluor 488) in images of labeled olfactory bulb tissues was performed by manually outlining the olfactory bulbs in the images using the QuPath “Closed Polygon Annotation” tool. These outlines were refined through the use of the QuPath “Pixel Classifier” tool, which was set to only retain areas where tissue was present within the encircled region. These areas were further segmented into an OMP+ “outer region” (containing outer areas of the olfactory bulb including the glomerular layer and olfactory nerve layer) and an OMP- “inner region” (containing inner areas of the olfactory bulb including the external plexiform layer, mitral cell layer, internal plexiform layer, and granule cell layer) by using the QuPath “Pixel Classifier” tool. Areas above and below a threshold of 400 Cy5 fluorescence intensity units (here marking OMP staining) were annotated as OMP+ and OMP-, respectively. A schematic displaying this process can be seen in Extended Data Fig. 4a. Mean fluorescence intensities for each channel and tissue area were calculated within OMP- and OMP+ regions, and the QuPath “Positive Cell Detection” and “Pixel Classifier” tools were used as described above to quantify cells positive for IBA1 and MxA in each region. The area of these annotations, as well as the mean fluorescence intensity of all markers within these annotations, were also quantified to enable multivariable measurements such as MxA staining intensity within the confines of IBA1+ regions. All image quantification measurements were calculated from these raw values. Each olfactory bulb was quantified independently and reported as an independent data point. Additional Declarations There is NO Competing Interest. Supplementary Files FrereetalExtDataFig14.pdf Extended Data Figure 1 | Antiviral and immune gene signatures in the olfactory epithelium are variant-independent. a. Immunofluorescence staining of olfactory epithelium (OE) slices from mock- and SARS-CoV-2–infected hamsters at the indicated days post-infection (dpi), showing detection of viral spike (S; red) and nucleocapsid (N; green) proteins. DAPI (blue) marks nuclei. Insets show digitally magnified views of boxed areas. b–e. RT-qPCR analyses of antiviral response genes expressed over time in the lungs of SARS-CoV-2-infected hamsters, including Isg15 (b), Ccl5 (c), Iba1 (d), and Cd3e (e). Data represent fold change versus mock at each time point. Means and SEMs from 3 independent biological replicates per group are shown. f. Immunofluorescence staining for SARS-CoV-2 S protein (red) and CD3+ T cells (magenta) in the OE at 5 dpi versus mock. Nuclei are stained with DAPI (blue). g-h. Gene set enrichment analysis (GSEA) of positively enriched antiviral and immune pathways (g) and cell-types (h) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant. Pathways (labeled A-E or A-D) are defined at the bottom of the figure. Extended Data Figure 2 | Neuronal loss following SARS-CoV-2 infection disrupts homeostatic gene expression and olfactory signaling in the olfactory epithelium. a. Quantification of “bare” olfactory epithelium (OE), defined as the fraction of OE devoid of olfactory sensory neurons (OSNs; OMP+) at the indicated days post-infection (dpi). b. Schematic of olfactory neurogenesis and cell type hierarchy, illustrating differentiation from horizontal basal cells to mature OSNs. c. Pseudotime projection of single-cell transcriptomic data showing a differentiation trajectory from basal cells to mature neurons. Cellular trajectory analysis was performed on neuronal and precursor populations using Monocle 3 in R. Visualizations were generated using ggplot2, Seurat, Monocle3, and fgsea. d. Gene set enrichment analysis (GSEA) showing changes in response to cytokines at the indicated time points. Significance is shown as –log10(padj-value) × sign(NES); dashed lines represent false discovery rate thresholds of padj=0.05. e. Violin plot showing olfactory receptor (OR) gene expression in mature OSNs at 1, 3, and 10 dpi compared to mock. f. Heat map showing differential OR gene expression (average log₂ fold change) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant, compared to mock. g. GSEA of negatively enriched pathways (left) and cell types (right) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant. Pathways (labeled A-F) are defined at the bottom of the figure. Extended Data Figure 3 | SARS-CoV-2 suppresses neuronal activity and elicits sustained immune signatures in the olfactory bulb in the absence of local infection. a. RT-qPCR analysis of Fos expression in the olfactory bulb (OB) from mock- and SARS-CoV-2–infected hamsters at the indicated time points. b. Gene set enrichment analysis (GSEA) quantifying immediate early gene (IEG) programs in OB tissues from hamsters infected at the indicated doses (PFU; plaque-forming units) and days post-infection (dpi). Enrichment scores are shown with dot size proportional to statistical significance (padj; adjusted p-values). c–d. RT-qPCR analysis of Isg15 (c) and Ccl5 (d) expression in OB from mock- and SARS-CoV-2–infected hamsters at the indicated time points. e. SARS-CoV-2 subgenomic N (SgN) RNA levels in the OB of mock- and SARS-CoV-2-infected animals, measured by RT-qPCR and expressed as the inverse of the cycle threshold (1/(Ct value)). Extended Data Figure 4 | Localized interferon signaling and microglial activation within the olfactory bulb are spatially enriched in OSN projection zones. a. RT-qPCR analysis of Iba1 in olfactory bulb (OB) tissue from mock- and SARS-CoV-2-infected hamsters over time post-infection. b. Immunofluorescent images of OBs stained for IBA1 (green) and DAPI (blue) at the indicated days post-infection (dpi). Lower panels show digitally magnified views of the boxed regions. c. Schematic illustrating stratification of the OB based on olfactory marker protein (OMP) expression into outer (OMP+) and inner (OMP-) regions. Outer region: olfactory nerve layer (ONL), glomerular layer (GL). Inner region: granule cell layer (GCL), internal plexiform layer (IPL), mitral cell layer (MCL), external plexiform layer (EPL). d. Mean fluorescence intensity of MxA in OMP- (inner) OB regions in mock- and SARS-CoV-2 infected hamsters at the indicated dpi. e-f. Percentage of IBA1+ cells within OMP+ outer regions (e) and OMP- inner regions (f) of the OB in mock- and SARS-CoV-2-infected hamsters across time points. g. Fraction of IBA1+ tissue within OMP- inner OB regions in mock- and SARS-CoV-2-infected hamsters at the indicated dpi. h. Gene set enrichment analysis (GSEA) of synaptic pruning programs in OBs from infected animals across all time points and infection doses. i. Analysis of full time-course sequencing data showing changes in significance of GSEA for olfactory sensory neuron (OSN) marker genes at the indicated time points. Significance is shown as –log10(p-adjusted value) × sign(NES); dashed lines represent false discovery rate thresholds. j. Heat map displaying differentially abundant type I interferon signaling and antigen presentation proteins in the OB at early and late time points post-infection as measured by mass spectrometry. Scale representative of log2(fold change) in abundance. k. Heat map displaying fold change of genes associated with microglial activation and phagocytic activity in human post-COVID OB tissue compared to healthy donor OB tissue as determined via RNA-sequencing. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {\"props\":{\"pageProps\":{\"initialData\":{\"identity\":\"rs-7179724\",\"acceptedTermsAndConditions\":true,\"allowDirectSubmit\":true,\"archivedVersions\":[],\"articleType\":\"Biological Sciences - Article\",\"associatedPublications\":[],\"authors\":[{\"id\":493291702,\"identity\":\"d9848507-7f8f-49fa-b309-efe5aca3a48d\",\"order_by\":0,\"name\":\"Benjamin tenOever\",\"email\":\"data:image/png;base64,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\",\"orcid\":\"https://orcid.org/0000-0003-0324-3078\",\"institution\":\"New York University\",\"correspondingAuthor\":true,\"prefix\":\"\",\"firstName\":\"Benjamin\",\"middleName\":\"\",\"lastName\":\"tenOever\",\"suffix\":\"\"},{\"id\":493291703,\"identity\":\"6e299a28-cea3-40ef-a3f2-8eb7dc4800da\",\"order_by\":1,\"name\":\"Justin Frere\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Icahn School of Medicine at Mount Sinai\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Justin\",\"middleName\":\"\",\"lastName\":\"Frere\",\"suffix\":\"\"},{\"id\":493291704,\"identity\":\"f2f98eb6-bd01-44b0-880c-9b89ce1b31c1\",\"order_by\":2,\"name\":\"Jonathan Dever\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"New York University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Jonathan\",\"middleName\":\"\",\"lastName\":\"Dever\",\"suffix\":\"\"},{\"id\":493291705,\"identity\":\"ffca18ff-de00-4cc7-b962-43465f421471\",\"order_by\":3,\"name\":\"Leon Hsieh\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"New York University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Leon\",\"middleName\":\"\",\"lastName\":\"Hsieh\",\"suffix\":\"\"},{\"id\":493291706,\"identity\":\"b5320ede-c977-48e1-997e-0e725a775260\",\"order_by\":4,\"name\":\"Caroline Ubbud\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"New York University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Caroline\",\"middleName\":\"\",\"lastName\":\"Ubbud\",\"suffix\":\"\"},{\"id\":493291707,\"identity\":\"8f0a9010-d5ed-47f1-a8b9-2c4de1463bcb\",\"order_by\":5,\"name\":\"Skyler Uhl\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"New York University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Skyler\",\"middleName\":\"\",\"lastName\":\"Uhl\",\"suffix\":\"\"},{\"id\":493291708,\"identity\":\"bfde467d-a5af-499d-804e-e6df034cda82\",\"order_by\":6,\"name\":\"Randal Serafini\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"Icahn School of Medicine at Mount Sinai\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Randal\",\"middleName\":\"\",\"lastName\":\"Serafini\",\"suffix\":\"\"},{\"id\":493291709,\"identity\":\"2dcc4488-0ac5-405a-8a4f-35579147b010\",\"order_by\":7,\"name\":\"Judith Minkoff\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"New York University\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Judith\",\"middleName\":\"\",\"lastName\":\"Minkoff\",\"suffix\":\"\"},{\"id\":493291710,\"identity\":\"d5570e09-72b0-4f56-bc9f-ce80dd67ccc9\",\"order_by\":8,\"name\":\"Tony Wang\",\"email\":\"\",\"orcid\":\"https://orcid.org/0000-0001-8983-1721\",\"institution\":\"FDA\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Tony\",\"middleName\":\"\",\"lastName\":\"Wang\",\"suffix\":\"\"},{\"id\":493291711,\"identity\":\"39d4fea4-93b5-4f9c-a33c-d41d211ff680\",\"order_by\":9,\"name\":\"Charles Chung\",\"email\":\"\",\"orcid\":\"\",\"institution\":\"FDA\",\"correspondingAuthor\":false,\"prefix\":\"\",\"firstName\":\"Charles\",\"middleName\":\"\",\"lastName\":\"Chung\",\"suffix\":\"\"}],\"badges\":[],\"createdAt\":\"2025-07-21 16:56:02\",\"currentVersionCode\":1,\"declarations\":\"\",\"doi\":\"10.21203/rs.3.rs-7179724/v1\",\"doiUrl\":\"https://doi.org/10.21203/rs.3.rs-7179724/v1\",\"draftVersion\":[],\"editorialEvents\":[],\"editorialNote\":\"\",\"failedWorkflow\":false,\"files\":[{\"id\":87980240,\"identity\":\"93e33b3e-89c2-4696-8866-ed0a1c58a693\",\"added_by\":\"auto\",\"created_at\":\"2025-07-31 05:49:14\",\"extension\":\"png\",\"order_by\":1,\"title\":\"Figure 1\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":1557670,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eSARS-CoV-2 infection induces robust antiviral and immune signaling in the olfactory epithelium.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e Schematic (left) and immunofluorescent staining (right) of a coronal section through the hamster olfactory system, highlighting airflow through the nasal cavity and the anatomical relationship between the olfactory epithelium (OE) and olfactory bulb (OB). Olfactory sensory neurons (OSNs) are stained for Cy5-labeled olfactory marker protein (OMP, magenta); nuclei are stained with DAPI (blue). Insets illustrate odor signal transmission from ciliated apical projections of OSNs in the OE, through the cribriform plate (dotted line), to higher-order neurons, such as mitral cells, in the glomerulus of the OB.\\u003cbr\\u003e\\n \\u003cstrong\\u003eb–c.\\u003c/strong\\u003e RT-qPCR quantification of SARS-CoV-2 subgenomic N (SgN) RNA in the OE (b) and lung (c) at indicated days post-infection, expressed as fold change relative to mock-infected controls. Means and SEMs from 3 independent biological replicates per group are shown.\\u003cbr\\u003e\\n \\u003cstrong\\u003ed.\\u003c/strong\\u003e Immunofluorescent staining of the OE at 3 days post-infection (dpi) for SARS-CoV-2 nucleocapsid protein (N; green), spike (S; red), OMP (cyan), and DAPI (blue). Merged image (All) highlights co-localization and epithelial disruption compared to mock.\\u003cbr\\u003e\\n \\u003cstrong\\u003ee–h.\\u003c/strong\\u003e \\u0026nbsp;RT-qPCR quantification of immune response genes expressed over time in the OE of SARS-CoV-2-infected hamsters, including \\u003cem\\u003eIsg15\\u003c/em\\u003e(e), \\u003cem\\u003eCcl5\\u003c/em\\u003e (f),\\u003cem\\u003e Iba1\\u003c/em\\u003e (g), and \\u003cem\\u003eCd3e\\u003c/em\\u003e (h). Data represent fold change versus mock at each time point. Means and SEMs from 3 independent biological replicates per group are shown.\\u003cbr\\u003e\\n \\u003cstrong\\u003ei.\\u003c/strong\\u003e Immunofluorescent staining of OE sections for MxA (red), a downstream interferon effector, in OE sections from mock- and SARS-CoV-2–infected hamsters at 3, 5, 10, 14 and 30 dpi. Cell nuclei are stained with DAPI (blue).\\u003cbr\\u003e\\n \\u003cstrong\\u003ej.\\u003c/strong\\u003e Heat map of type I interferon response-associated gene expression in the OE at 4 dpi following infection with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Binder41.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/4b150d08c278685204c50021.png\"},{\"id\":87980237,\"identity\":\"ab9558eb-39f2-46f0-9af6-6cc7729eb512\",\"added_by\":\"auto\",\"created_at\":\"2025-07-31 05:49:14\",\"extension\":\"png\",\"order_by\":2,\"title\":\"Figure 2\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":944620,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eSARS-CoV-2 infection causes profound disruption of olfactory epithelium architecture and initiates month-long repair response\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e Representative images of the olfactory epithelium (OE) from mock- and SARS-CoV-2–infected hamsters at the indicated days post-infection (dpi), stained for Cy5-labeled olfactory marker protein (OMP, magenta) and nuclei (DAPI, blue).\\u003cbr\\u003e\\n \\u003cstrong\\u003eb.\\u003c/strong\\u003e Quantification using QuPath (0.5.1) software of OE morphology and OSN metrics across time points: OE thickness (top left), OE cellular density (nuclei per mm of OE surface length) (top right), and number (OMP+ cells per mm of OE surface length)(bottom left) and fractional makeup (fraction of all cells stained as OMP+) (bottom right) of OMP+ cells (OSNs) within the OE.\\u003cbr\\u003e\\n \\u003cstrong\\u003ec.\\u003c/strong\\u003e Immunofluorescent staining of OE from mock- or SARS-CoV-2-infected hamsters at the indicated dpi for Ki67 (red) and nuclei (DAPI, blue). Quantification using QuPath (0.5.1) software of Ki67+ nuclei as a percentage of total OE nuclei (lower left panel) and number per mm length of OE surface (lower right panel).\\u003cbr\\u003e\\n \\u003cstrong\\u003ed.\\u003c/strong\\u003e Single-cell RNA sequencing analysis of longitudinal SARS-CoV-2-infected and control hamster OE, processed in Seurat v5.1.0 (R) and visualized by UMAP. Cells are color-coded by annotation, including mature OSNs, immature OSNs, neural precursors, globose basal cells, horizontal basal cells, and other, non-neuronal cell populations (Other).\\u003cbr\\u003e\\n \\u003cstrong\\u003ee–f.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) showing changes in interferon-stimulated gene (ISG) expression (e) and axonemal assembly programs (f) at the indicated dpi across the cell populations defined in (d). Significance is shown as –log10(p-value) × sign(NES); dashed lines represent false discovery rate thresholds for significant values (with padj=0.05).\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Binder42.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/98eca37727fbd7772e67b7db.png\"},{\"id\":87980778,\"identity\":\"9eb628e6-777c-4351-960b-58cf21875ed3\",\"added_by\":\"auto\",\"created_at\":\"2025-07-31 05:57:14\",\"extension\":\"png\",\"order_by\":3,\"title\":\"Figure 3\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":892623,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eSARS-CoV-2 infection attenuates neuronal activity and triggers sustained inflammatory signatures in the olfactory bulb beyond viral clearance.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e Heat map displaying differential expression of immediate early genes (IEGs) in olfactory bulb (OB) tissue from hamsters infected with SARS-CoV-2 at varying doses (100, 1000, or 10000 plaque-forming units [PFU]) and days post-infection (dpi), relative to mock-infected controls.\\u003cbr\\u003e\\n \\u003cstrong\\u003eb–c.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) quantifying enrichment of hallmark inflammatory pathways—response to type I interferon (b) and chemokine activity (c)—in OB tissues from hamsters infected at the indicated doses and dpi. Enrichment scores are shown with dot size proportional to statistical significance (padj; adjusted p-values).\\u003cbr\\u003e\\n \\u003cstrong\\u003ed.\\u003c/strong\\u003e Confocal images of OB tissue sections from mock- and SARS-CoV-2-infected hamsters at the indicated dpi, stained for SARS-CoV-2 spike (S; red) and nucleocapsid protein (N; green) and DAPI (blue). Rightmost panel shows staining of OE at 3 dpi as a control for infection.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Binder43.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/7ef12862b4953c7f690af004.png\"},{\"id\":87980241,\"identity\":\"d63f0ec1-ae79-4ad5-b7fd-94fbd0822635\",\"added_by\":\"auto\",\"created_at\":\"2025-07-31 05:49:14\",\"extension\":\"png\",\"order_by\":4,\"title\":\"Figure 4\",\"display\":\"\",\"copyAsset\":false,\"role\":\"figure\",\"size\":2082271,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eDelayed clearance of olfactory neuronal debris triggers sustained microglial activation and immune signaling in the olfactory bulb.\\u003c/strong\\u003e\\u003cbr\\u003e\\n\\u003cstrong\\u003ea.\\u003c/strong\\u003e\\u0026nbsp;Gene set enrichment analysis (GSEA) showing transcriptional signatures of microglial activation in olfactory bulbs (OB) from SARS-CoV-2–infected hamsters over time and at varying infection doses (PFU; plaque-forming units). Enrichment scores and adjusted p-values are indicated.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eb.\\u003c/strong\\u003e\\u0026nbsp;Immunofluorescence staining of OB sections at indicated days post-infection (dpi), showing expression of MxA (red) and DAPI (blue).\\u003cbr\\u003e\\n\\u003cstrong\\u003ec–f.\\u003c/strong\\u003e\\u0026nbsp; Mean fluorescence intensity of MxA in OMP+, OMP+/IBA1+, or OMP-/IBA1+ OB tissue subsets (c, e-f), and fraction of IBA1+ tissue within OMP+ regions (d).\\u003cbr\\u003e\\n\\u003cstrong\\u003eg.\\u003c/strong\\u003e\\u0026nbsp;Fluorescent microscopy of OB sections from mock- and SARS-CoV-2-infected animals at 14 dpi stained for MxA (red), IBA1 (green), OMP (magenta), and DAPI (blue). Overlays (All) illustrate co-localization and spatial distribution of interferon activity and microglia in relation to OSNs in the outer (OMP+) and inner (OMP-) regions of the olfactory bulb.\\u003cbr\\u003e\\n\\u003cstrong\\u003eh.\\u003c/strong\\u003e\\u0026nbsp;Heat map of differentially expressed genes involved in core microglial functions in OBs from SARS-CoV-2-infected hamsters at multiple timepoints (dpi) and infection doses (PFU), demonstrating log\\u003csub\\u003e2\\u003c/sub\\u003e(fold change) of each gene compared to mock.\\u003cbr\\u003e\\n\\u003cstrong\\u003ei–j.\\u003c/strong\\u003e\\u0026nbsp;GSEA showing sustained enrichment of gene sets related to apoptotic cell clearance (i) and phagocytosis (j) in OBs following infection.\\u003cbr\\u003e\\n\\u003cstrong\\u003ek.\\u003c/strong\\u003e\\u0026nbsp;Whole-head coronal sections from mock- and SARS-CoV-2-infected hamsters at 10dpi stained for OMP (magenta) and DAPI (blue), highlighting differential loss of OMP cell bodies and their axonal projections (insets 1 and 2).\\u003cbr\\u003e\\n\\u003cstrong\\u003el.\\u003c/strong\\u003e\\u0026nbsp;Quantification of mean OMP intensity within outer regions of the OB (OMP+) across time points.\\u003cbr\\u003e\\n\\u003cstrong\\u003em.\\u003c/strong\\u003e\\u0026nbsp;Plot showing average log₂(fold change) in OSN-associated gene expression over time in the OB and olfactory epithelium (OE) of SARS-CoV-2-infected hamsters compared to mock at 4 dpi.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u0026nbsp;\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"Binder44.png\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/f2be93cf40b43d7ab78542c1.png\"},{\"id\":89557568,\"identity\":\"5d355292-d833-4439-bbde-dd12b88027e4\",\"added_by\":\"auto\",\"created_at\":\"2025-08-21 09:44:52\",\"extension\":\"pdf\",\"order_by\":0,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"manuscript-pdf\",\"size\":6741974,\"visible\":true,\"origin\":\"\",\"legend\":\"\",\"description\":\"\",\"filename\":\"manuscript.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/9d522e79-8c99-4bac-8448-2d250ad89814.pdf\"},{\"id\":87980239,\"identity\":\"ad2a88a9-cf06-47d0-88cb-b9a2346d1463\",\"added_by\":\"auto\",\"created_at\":\"2025-07-31 05:49:14\",\"extension\":\"pdf\",\"order_by\":1,\"title\":\"\",\"display\":\"\",\"copyAsset\":false,\"role\":\"supplement\",\"size\":818764,\"visible\":true,\"origin\":\"\",\"legend\":\"\\u003cp\\u003e\\u003cstrong\\u003eExtended Data Figure 1 | Antiviral and immune gene signatures in the olfactory epithelium are variant-independent.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e Immunofluorescence staining of olfactory epithelium (OE) slices from mock- and SARS-CoV-2–infected hamsters at the indicated days post-infection (dpi), showing detection of viral spike (S; red) and nucleocapsid (N; green) proteins. DAPI (blue) marks nuclei. Insets show digitally magnified views of boxed areas.\\u003cbr\\u003e\\n \\u003cstrong\\u003eb–e.\\u003c/strong\\u003e RT-qPCR analyses of antiviral response genes expressed over time in the lungs of SARS-CoV-2-infected hamsters, including \\u003cem\\u003eIsg15\\u003c/em\\u003e (b), \\u003cem\\u003eCcl5\\u003c/em\\u003e(c),\\u003cem\\u003e Iba1\\u003c/em\\u003e (d), and \\u003cem\\u003eCd3e\\u003c/em\\u003e (e). Data represent fold change versus mock at each time point. Means and SEMs from 3 independent biological replicates per group are shown.\\u003cbr\\u003e\\n \\u003cstrong\\u003ef.\\u003c/strong\\u003e Immunofluorescence staining for SARS-CoV-2 S protein (red) and CD3+ T cells (magenta) in the OE at 5 dpi versus mock. Nuclei are stained with DAPI (blue).\\u003cbr\\u003e\\n \\u003cstrong\\u003eg-h.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) of positively enriched antiviral and immune pathways (g) and cell-types (h) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant. Pathways (labeled A-E or A-D) are defined at the bottom of the figure.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eExtended Data Figure 2 | Neuronal loss following SARS-CoV-2 infection disrupts homeostatic gene expression and olfactory signaling in the olfactory epithelium.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e Quantification of “bare” olfactory epithelium (OE), defined as the fraction of OE devoid of olfactory sensory neurons (OSNs; OMP+) at the indicated days post-infection (dpi).\\u003cbr\\u003e\\n \\u003cstrong\\u003eb.\\u003c/strong\\u003e Schematic of olfactory neurogenesis and cell type hierarchy, illustrating differentiation from horizontal basal cells to mature OSNs.\\u003cbr\\u003e\\n \\u003cstrong\\u003ec.\\u003c/strong\\u003e Pseudotime projection of single-cell transcriptomic data showing a differentiation trajectory from basal cells to mature neurons. Cellular trajectory analysis was performed on neuronal and precursor populations using Monocle 3 in R. Visualizations were generated using ggplot2, Seurat, Monocle3, and fgsea.\\u003cbr\\u003e\\n \\u003cstrong\\u003ed.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) showing changes in response to cytokines at the indicated time points. Significance is shown as –log10(padj-value) × sign(NES); dashed lines represent false discovery rate thresholds of padj=0.05.\\u003cbr\\u003e\\n \\u003cstrong\\u003ee.\\u003c/strong\\u003e Violin plot showing olfactory receptor (OR) gene expression in mature OSNs at 1, 3, and 10 dpi compared to mock.\\u003cbr\\u003e\\n \\u003cstrong\\u003ef.\\u003c/strong\\u003e Heat map showing differential OR gene expression (average log₂ fold change) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant, compared to mock.\\u003cbr\\u003e\\n \\u003cstrong\\u003eg.\\u003c/strong\\u003e GSEA of negatively enriched pathways (left) and cell types (right) in the OE at 4 dpi with the WA1 or BA.5 variants of SARS-CoV-2, or the furin cleavage site-deficient WA1 ΔPRRA mutant. Pathways (labeled A-F) are defined at the bottom of the figure.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eExtended Data Figure 3 | SARS-CoV-2 suppresses neuronal activity and elicits sustained immune signatures in the olfactory bulb in the absence of local infection.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e RT-qPCR analysis of \\u003cem\\u003eFos\\u003c/em\\u003e expression in the olfactory bulb (OB) from mock- and SARS-CoV-2–infected hamsters at the indicated time points. \\u003cbr\\u003e\\n \\u003cstrong\\u003eb.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) quantifying immediate early gene (IEG) programs in OB tissues from hamsters infected at the indicated doses (PFU; plaque-forming units) and days post-infection (dpi). Enrichment scores are shown with dot size proportional to statistical significance (padj; adjusted p-values).\\u003cbr\\u003e\\n \\u003cstrong\\u003ec–d.\\u003c/strong\\u003e RT-qPCR analysis of \\u003cem\\u003eIsg15\\u003c/em\\u003e (c) and \\u003cem\\u003eCcl5\\u003c/em\\u003e (d) expression in OB from mock- and SARS-CoV-2–infected hamsters at the indicated time points. \\u003cbr\\u003e\\n \\u003cstrong\\u003ee.\\u003c/strong\\u003e SARS-CoV-2 subgenomic N (SgN) RNA levels in the OB of mock- and SARS-CoV-2-infected animals, measured by RT-qPCR and expressed as the inverse of the cycle threshold (1/(Ct value)).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eExtended Data Figure 4 | Localized interferon signaling and microglial activation within the olfactory bulb are spatially enriched in OSN projection zones.\\u003c/strong\\u003e\\u003cbr\\u003e\\n \\u003cstrong\\u003ea.\\u003c/strong\\u003e RT-qPCR analysis of \\u003cem\\u003eIba1\\u003c/em\\u003e in olfactory bulb (OB) tissue from mock- and SARS-CoV-2-infected hamsters over time post-infection. \\u003cbr\\u003e\\n \\u003cstrong\\u003eb.\\u003c/strong\\u003e Immunofluorescent images of OBs stained for IBA1 (green) and DAPI (blue) at the indicated days post-infection (dpi). Lower panels show digitally magnified views of the boxed regions.\\u003cbr\\u003e\\n \\u003cstrong\\u003ec.\\u003c/strong\\u003e Schematic illustrating stratification of the OB based on olfactory marker protein (OMP) expression into outer (OMP+) and inner (OMP-) regions. Outer region: olfactory nerve layer (ONL), glomerular layer (GL). Inner region: granule cell layer (GCL), internal plexiform layer (IPL), mitral cell layer (MCL), external plexiform layer (EPL).\\u003cbr\\u003e\\n \\u003cstrong\\u003ed.\\u003c/strong\\u003e Mean fluorescence intensity of MxA in OMP- (inner) OB regions in mock- and SARS-CoV-2 infected hamsters at the indicated dpi. \\u003cbr\\u003e\\n \\u003cstrong\\u003ee-f. \\u003c/strong\\u003ePercentage of IBA1+ cells within OMP+ outer regions (e) and OMP- inner regions (f) of the OB in mock- and SARS-CoV-2-infected hamsters across time points.\\u003cbr\\u003e\\n \\u003cstrong\\u003eg. \\u003c/strong\\u003eFraction of IBA1+ tissue within OMP- inner OB regions in mock- and SARS-CoV-2-infected hamsters at the indicated dpi.\\u003cbr\\u003e\\n \\u003cstrong\\u003eh.\\u003c/strong\\u003e Gene set enrichment analysis (GSEA) of synaptic pruning programs in OBs from infected animals across all time points and infection doses.\\u003cbr\\u003e\\n \\u003cstrong\\u003ei.\\u003c/strong\\u003e Analysis of full time-course sequencing data showing changes in significance of GSEA for olfactory sensory neuron (OSN) marker genes at the indicated time points. Significance is shown as –log10(p-adjusted value) × sign(NES); dashed lines represent false discovery rate thresholds.\\u003cbr\\u003e\\n \\u003cstrong\\u003ej.\\u003c/strong\\u003e Heat map displaying differentially abundant type I interferon signaling and antigen presentation proteins in the OB at early and late time points post-infection as measured by mass spectrometry. Scale representative of log2(fold change) in abundance. \\u003cbr\\u003e\\n \\u003cstrong\\u003ek.\\u003c/strong\\u003e Heat map displaying fold change of genes associated with microglial activation and phagocytic activity in human post-COVID OB tissue compared to healthy donor OB tissue as determined via RNA-sequencing.\\u003c/p\\u003e\",\"description\":\"\",\"filename\":\"FrereetalExtDataFig14.pdf\",\"url\":\"https://assets-eu.researchsquare.com/files/rs-7179724/v1/cdf1a23ebb4ea643ae8d8275.pdf\"}],\"financialInterests\":\"There is \\u003cb\\u003eNO\\u003c/b\\u003e Competing Interest.\",\"formattedTitle\":\"A Peripheral Neuron-to-Microglia Signaling Axis Connecting Transient Viral Infection to Persistent Neuroinflammatory States\",\"fulltext\":[{\"header\":\"Full Text\",\"content\":\"\\u003cp\\u003eAcute viral infections are increasingly recognized for their capacity to induce sustained symptoms that persist well beyond the period of viral clearance\\u003csup\\u003e2\\u003c/sup\\u003e. Among the most prominent examples is SARS-CoV-2, the causative agent of COVID-19, which has drawn attention to the poorly understood relationship between acute infection and long-term clinical sequelae. Despite being largely restricted to the respiratory tract, SARS-CoV-2 has consistently demonstrated the ability to affect distal tissues, both during active infection and after viral resolution\\u003csup\\u003e3\\u003c/sup\\u003e. Indeed, while prevalence estimates vary, the World Health Organization reports that approximately 10–20% of individuals experience symptoms beyond three months after resolution of SARS-CoV-2 infection\\u003csup\\u003e4\\u003c/sup\\u003e. Together these post-viral complications have collectively come to be referred to as post-acute sequelae of SARS-CoV-2 infection (PASC), or more colloquially as long COVID.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eNotably, neuropsychiatric symptoms rank among the most frequently reported features of long COVID\\u003csup\\u003e5-7\\u003c/sup\\u003e. Meta-analyses indicate that more than 20% of post-COVID patients report ongoing issues such as sleep disturbance, cognitive impairment, fatigue, anxiety, depression, pain, and prolonged anosmia, or loss of smell\\u003csup\\u003e8\\u003c/sup\\u003e. The underlying causes of these persistent symptoms remain elusive, in part because multiple studies have shown that SARS-CoV-2 is rarely seen to infect neural tissues \\u003cem\\u003ein vivo\\u003c/em\\u003e\\u003csup\\u003e9\\u003c/sup\\u003e. Both clinical and experimental studies into alternative methods of virus-induced damage have implicated a range of potential contributory processes, including thromboembolic events\\u003csup\\u003e6,10\\u003c/sup\\u003e, complement dysregulation\\u003csup\\u003e10-12\\u003c/sup\\u003e, altered serotonergic signaling\\u003csup\\u003e13\\u003c/sup\\u003e, reactivation of latent herpesviruses\\u003csup\\u003e14,15\\u003c/sup\\u003e, sustained inflammatory activity\\u003csup\\u003e1,16\\u003c/sup\\u003e, immune dysregulation\\u003csup\\u003e17,18\\u003c/sup\\u003e, residual viral reservoirs\\u003csup\\u003e19\\u003c/sup\\u003e, and lasting damage incurred during acute infection\\u003csup\\u003e20,21\\u003c/sup\\u003e; however, despite this evidence, it still remains unclear how a transient viral infection may be able to mediate these processes at times following viral clearance and throughout tissues spared from direct infection.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eInterestingly, recent evidence has further implicated olfactory disruption as a potential driver of long COVID neurocognitive symptomology\\u003csup\\u003e1,22\\u003c/sup\\u003e. Olfactory dysfunction is a common feature of SARS-CoV-2 infection, with anosmia affecting approximately 40% of unvaccinated and infection-naïve individuals with COVID-19\\u003csup\\u003e23\\u003c/sup\\u003e. Anosmia is thought to be due to robust SARS-CoV-2 infection of the OE in the upper respiratory tract during early infection. The OE is home to OSNs that express olfactory receptors on their highly ciliated apical surfaces and mediate primary olfaction by binding odorant molecules in the nasal lumen\\u003csup\\u003e24\\u003c/sup\\u003e. OSNs, which are largely supported in the OE by sustentacular cells, transmit these odorant signals through long axonal projections which extend from their cell bodies in the OE through the cribriform plate to the glomerular layer (GL) of the OB, where they synapse with mitral cells and other second-order neurons (\\u003cstrong\\u003eFig. 1a\\u003c/strong\\u003e)\\u003csup\\u003e24\\u003c/sup\\u003e. Although the sustentacular cells of the OE are highly susceptible to SARS-CoV-2, multiple human and animal studies have shown that OSNs are largely resistant to infection due to minimal expression of the viral entry receptor angiotensin-converting enzyme 2 (ACE2)\\u003csup\\u003e25-27\\u003c/sup\\u003e. Anosmia is therefore believed to result from epithelial disruption and local inflammation, rather than direct neuronal infection, a model supported by the transient nature of smell loss in most patients\\u003csup\\u003e28\\u003c/sup\\u003e.Despite some evidence suggesting limited infection of neuronal cells, recent studies have shown that COVID-associated olfactory disturbance is associated with significant alteration of neuronal biology in the central nervous system, and these changes can be observed to persist to time points well beyond viral clearance\\u003csup\\u003e1,22\\u003c/sup\\u003e. Again, however, while these pathologic changes have been noted, the particular mechanism by which a largely non-neurotropic SARS-CoV-2 virus can drive these phenotypes remains unclear. Given the significant viral load that accumulates in the OE and the functional link that infection of this tissue appears to have with central nervous system dysfunction, we here sought to investigate whether a mechanistic link may exist to connect these phenomena. \\u0026nbsp;\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eTo investigate how SARS-CoV-2 infection of the OE might initiate biological changes leading to persistent neurological symptoms, we utilized the golden hamster model, widely regarded as the most physiologically relevant small animal system for COVID-19\\u003csup\\u003e29\\u003c/sup\\u003e. To this end, we characterized the temporal dynamics and tissue specificity of viral replication following intranasal challenge of SARS-CoV-2 over a 30-day period. Viral RNA levels in the OE increased sharply within 24 hours post-infection and remained elevated for ten days before returning to baseline (\\u003cstrong\\u003eFig. 1b\\u003c/strong\\u003e). In parallel, the lower airways of the same animals exhibited similarly high viral RNA levels early after infection, but these declined more rapidly, dropping by three orders of magnitude by day 7 and becoming undetectable by day 10. (\\u003cstrong\\u003eFig. 1c\\u003c/strong\\u003e). Infection of the OE could also be confirmed in an independent cohort of animals using immunohistochemistry (IHC) for SARS-CoV-2 nucleocapsid (N) and spike (S) proteins (\\u003cstrong\\u003eExtended Data Fig. 1a\\u003c/strong\\u003e). Notably, viral proteins were cleared from the OE more rapidly than RNA, with near-complete resolution by day 5, suggesting either higher sensitivity of RNA-based detection or delayed clearance of viral RNA following protein clearance (\\u003cstrong\\u003eFig. 1b and Extended Data Fig. 1a\\u003c/strong\\u003e). Co-staining with the olfactory marker protein (OMP), an OSN marker, revealed minimal colocalization with viral proteins, corroborating past studies suggesting that SARS-CoV-2 infection in the OE is largely restricted to non-neuronal sustentacular cells (\\u003cstrong\\u003eFig. 1d\\u003c/strong\\u003e)\\u003csup\\u003e25\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eTo assess the host response to SARS-CoV-2 infection in the OE, we next examined the temporal dynamics of antiviral and immune signaling. RT-qPCR analyses revealed robust induction of immune-related genes, including \\u003cem\\u003eIsg15\\u003c/em\\u003e, a canonical interferon-stimulated gene (ISG); \\u003cem\\u003eCcl5\\u003c/em\\u003e, a chemokine involved in immune cell recruitment; \\u003cem\\u003eIba1\\u003c/em\\u003e, a marker for microglia and macrophages; and \\u003cem\\u003eCd3e\\u003c/em\\u003e, a T cell-specific marker (\\u003cstrong\\u003eFig. 1e-h\\u003c/strong\\u003e). All four genes were strongly upregulated during the early phase of infection, with peak expression occurring between 1 and 5 days post-infection (dpi), coinciding with maximal viral replication. These trends were also observed in the lung, although immune gene expression in the OE was more pronounced and slower to resolve (\\u003cstrong\\u003eExtended Data Fig. 1b-e\\u003c/strong\\u003e). Protein-level validation confirmed these findings, with IHC detecting increased expression of the interferon-inducible MxAprotein and CD3 T cell marker in the OE (\\u003cstrong\\u003eFig. 1i and Extended Data Fig. 1f\\u003c/strong\\u003e).\\u003c/p\\u003e\\n\\u003cp\\u003eTo assess whether these olfactory immune responses were conserved across SARS-CoV-2 variants or when viral biology was muted, we performed transcriptional profiling of OE tissue from hamsters infected with the original Washington strain (WA1), a cleavage site-deficient mutant (WA1 ΔPRRA), or the BA.5 Omicron variant at peak infection (4 dpi). Despite differences in viral genotype and pathogenesis, all three strains elicited nearly identical type I interferon (IFN-I) responses (\\u003cstrong\\u003eFig. 1j\\u003c/strong\\u003e), with comparable patterns of immune gene expression and pathway enrichment, including markers of immune cell infiltration (\\u003cstrong\\u003eExtended Data Fig. 1g-h\\u003c/strong\\u003e). These data suggest that SARS-CoV-2-induced OE inflammation represents a conserved host response that is not variant or viral fitness dependent.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eGiven the extent of SARS-CoV-2 infection in the OE, we next assessed its impact on OSNs and overall epithelial architecture. Immunostaining for OMP on fixed cross-sections of the OE revealed marked structural disruption during peak infection (\\u003cstrong\\u003eFig 2a\\u003c/strong\\u003e). OSNs exhibited a progressive loss of their characteristic ciliated morphology, accompanied by detachment from the epithelial surface tracking with viral infection and immune response. Notably, OMP⁺\\u0026nbsp;cell bodies were frequently observed within the nasal lumen, consistent with the accumulation of sloughed neuronal debris and widespread epithelial disorganization. Quantification of changes to OE structure demonstrated a significant thinning of the epithelial layer during this period, with many areas of OE showing near complete OSN ablation between 5-10dpi \\u003cstrong\\u003e(Fig. 2b and Extended Data Fig. 2a)\\u003c/strong\\u003e.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eFollowing viral clearance, evidence of epithelial repair became apparent in stained OE sections. Quantification of OE thickness and OMP⁺\\u0026nbsp;OSN cell counts revealed a gradual recovery that eventually approached baseline levels, but only by 30dpi (\\u003cstrong\\u003eFig. 2a-b\\u003c/strong\\u003e). This regenerative process was mirrored by an increase in Ki67 expression, a marker of cellular proliferation, localized to the basal layer of the OE, where progenitor and stem cell populations reside (\\u003cstrong\\u003eFig. 2c\\u003c/strong\\u003e)\\u003csup\\u003e30\\u003c/sup\\u003e. Notably, Ki67 expression returned to baseline by 14 dpi, while the recovery of mature OSNs lagged behind, indicating a temporal gap between stem cell activation and neuronal differentiation, consistent with previous observations\\u003csup\\u003e30\\u003c/sup\\u003e. This delayed repopulation of the olfactory neuroepithelium may help explain why some individuals experience persistent anosmia following SARS-CoV-2 infection\\u003csup\\u003e31\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eTo further investigate the functional consequences of SARS-CoV-2 infection on OSNs and their progenitors, we compared mock to SARS-CoV-2 infected OE by single-cell RNA sequencing (scRNA-seq). This approach resolved distinct OSN populations and their lineage-associated precursors, with trajectory analysis confirming expected differentiation hierarchies (\\u003cstrong\\u003eFig. 2d and Extended Data Fig. 2b-c\\u003c/strong\\u003e). Across cell types, infection elicited a broadly uniform transcriptional response, marked by a peak in ISG and chemokine expression at 3dpi (\\u003cstrong\\u003eFig. 2e)\\u003c/strong\\u003e. These immune signatures inversely correlated with the expression of genes involved in axoneme assembly, the structural foundation of motile cilia, suggesting that inflammation may suppress ciliogenic programs during peak infection (\\u003cstrong\\u003eFig. 2f\\u003c/strong\\u003e \\u003cstrong\\u003eand Extended Data Fig. 2d\\u003c/strong\\u003e). Given that odorant receptors are localized to the cilia of OSNs\\u003csup\\u003e32\\u003c/sup\\u003e, we assessed their expression levels and observed significant downregulation at 1, 3, and 5 days post-infection in both scRNA-seq and bulk RNA-seq datasets (\\u003cstrong\\u003eExtended Data Fig. 2e–g\\u003c/strong\\u003e). Notably, while transcriptional repression of odorant receptors was evident following infection with both the original SARS-CoV-2 strain (WA1) and the Omicron variant, the extent of downregulation was consistently reduced in Omicron-infected samples, consistent with reports indicating that this variant induces smell loss at a reduced rate compared to other variants\\u003csup\\u003e33\\u003c/sup\\u003e\\u003csup\\u003e,\\u003c/sup\\u003e\\u003csup\\u003e34\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eWhile these findings provide a molecular explanation for infection-induced anosmia and align with previous studies\\u003csup\\u003e25\\u003c/sup\\u003e, they also reveal that these epithelial abnormalities largely resolve by 14dpi. This transient nature of OE disruption therefore fails to account for the persistent neuroinflammation observed both in this same small animal model as well as some individuals long after viral clearance, which occurs within 7dpi\\u003csup\\u003e1,35,36\\u003c/sup\\u003e. To explore this disconnect, we assessed the impact of SARS-CoV-2 infection on neuronal biology in the OB, the first central nervous system structure to receive input from OSNs and a potential site for longer-term immune engagement. To this end, we first examined the expression of Fos, a canonical immediate early gene (IEG) and established marker of neuronal activity\\u003csup\\u003e37\\u003c/sup\\u003e. IEGs are rapidly and transiently induced in response to cellular stimuli and play essential roles in synaptic plasticity, learning, and memory\\u003csup\\u003e37\\u003c/sup\\u003e. RT-qPCR analysis revealed a marked reduction in Fos transcript levels in the OB coinciding with the onset of OE damage and OSN loss. Notably, this suppression persisted throughout the 30-day post-infection period, indicating a sustained depression of OB neuronal activity following SARS-CoV-2 exposure (\\u003cstrong\\u003eExtended Data Fig. 3a\\u003c/strong\\u003e). To further characterize this response, we performed bulk RNA sequencing of the OB across the same time course. Differential expression analysis confirmed that, like \\u003cem\\u003eFos\\u003c/em\\u003e, other odorant-evoked neuronal IEGs that have been reported, such as: \\u003cem\\u003eArc, Egr1, Egr3, Fosb, Npas4\\u003c/em\\u003e, and \\u003cem\\u003eNr4a1\\u003c/em\\u003e\\u003csup\\u003e38\\u003c/sup\\u003e, were also significantly downregulated (\\u003cstrong\\u003eFig. 3a and Extended Data Fig. 3b\\u003c/strong\\u003e). Together, these findings suggest that SARS-CoV-2 infection induces broad and long-lasting suppression of neuronal activity in the OB, potentially disrupting neurobiological functions well beyond the period of viral clearance.\\u003c/p\\u003e\\n\\u003cp\\u003eTo better define the molecular underpinnings of sustained OB dysfunction, we performed gene set enrichment analysis (GSEA) on our longitudinal OB RNA-sequencing data. This analysis revealed persistent upregulation of IFN-I and chemokine signaling pathways, with a notable late-phase intensification emerging at 14dpi and persisting through at least 30dpi (\\u003cstrong\\u003eFig. 3b-c\\u003c/strong\\u003e). These prolonged immune signatures in the OB stood in sharp contrast to the transient and steadily resolving responses observed in the OE (\\u003cstrong\\u003eFig. 2e, Extended Data Fig. 2d\\u003c/strong\\u003e), indicating a decoupling of peripheral viral clearance from central immune activity. These data were further validated by RT-qPCR within an additional cohort of infected animals using \\u003cem\\u003eIsg15\\u003c/em\\u003e and \\u003cem\\u003eCcl5\\u0026nbsp;\\u003c/em\\u003eas proxies for the innate antiviral response (\\u003cstrong\\u003eExtended Data Fig. 3c-d\\u003c/strong\\u003e).\\u003c/p\\u003e\\n\\u003cp\\u003eTo determine whether this prolonged transcriptional activity reflected ongoing viral presence, we next examined OB tissues for evidence of SARS-CoV-2. RT-qPCR targeting subgenomic nucleocapsid (sgN) RNA was negative except for trace positivity at 4dpi, the peak of systemic viral burden and well before the late-phase intensification of OB immune activation; further IHC for S and N protein revealed an absence of viral protein in the OB throughout the observed period (\\u003cstrong\\u003eFig. 3d and Extended Data Fig. 3e\\u003c/strong\\u003e). These findings align with previous reports indicating that direct infection of the central nervous system by SARS-CoV-2 is minimal to absent and point instead to a non-infectious trigger of OB inflammation\\u003csup\\u003e1,35,39\\u003c/sup\\u003e.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eGiven the absence of direct OB infection, we next sought to identify the cellular mediators of this delayed inflammatory response. We interrogated our GSEA datasets for signatures of immune cell infiltration and observed significant enrichment for transcriptional profiles of microglia, the resident macrophage population of the brain (\\u003cstrong\\u003eFig. 4a\\u003c/strong\\u003e). Notably, microglial signatures peaked at 4 dpi, declined by 5 dpi, and then rose again to a second peak at 14 dpi—following viral clearance from the olfactory epithelium—and remained elevated through 31 dpi, closely paralleling the temporal dynamics of interferon and chemokine pathway activation (\\u003cstrong\\u003eFig. 1b, 3b-c and Extended Data Fig. 1a)\\u003c/strong\\u003e. RT-qPCR and IHC validation further confirmed sustained upregulation of the macrophage/microglia marker gene \\u003cem\\u003eIba1\\u003c/em\\u003e beyond 30dpi in an independent cohort (\\u003cstrong\\u003eExtended Data Fig. 4a-b\\u003c/strong\\u003e).\\u003c/p\\u003e\\n\\u003cp\\u003eTo further\\u0026nbsp;characterize IFN-I signaling in the OB, tissue\\u0026nbsp;sections from SARS-CoV-2-infected hamsters were stained for MxA, a canonical ISG\\u003csup\\u003e40\\u003c/sup\\u003e, showing robust MxA induction throughout the 30dpi period (\\u003cstrong\\u003eFig. 4b)\\u003c/strong\\u003e. Importantly, these responses were most prominent at the\\u0026nbsp;outermost layers of the OB, particularly within the\\u0026nbsp;olfactory nerve layer (ONL)\\u0026nbsp;and\\u0026nbsp;glomerular layer (GL), regions enriched for\\u0026nbsp;OMP\\u003csup\\u003e+\\u003c/sup\\u003e axonal projections\\u0026nbsp;from OSNs in the OE (\\u003cstrong\\u003eFig. 4b\\u003c/strong\\u003e)\\u003csup\\u003e41\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eTo refine these observations, OB sections were stratified into\\u0026nbsp;OMP\\u003csup\\u003e+\\u003c/sup\\u003e (ONL and GL)\\u0026nbsp;and\\u0026nbsp;OMP\\u003csup\\u003e–\\u003c/sup\\u003e (external plexiform layer (EPL), mitral cell layer (MCL), internal plexiform layer (IPL), and granule cell layer (GCL)),\\u0026nbsp;regions, and levels of \\u003cem\\u003eIba1\\u003c/em\\u003e and \\u003cem\\u003eMxA\\u0026nbsp;\\u003c/em\\u003eexpression were independently quantified in each compartment (\\u003cstrong\\u003eExtended Data Fig. 4c\\u003c/strong\\u003e). These analyses revealed significantly elevated expression of both gene products through 14dpi, with the strongest signals consistently localized to the OMP\\u003csup\\u003e+\\u003c/sup\\u003e regions (\\u003cstrong\\u003eFig. 4c-f, Extended Data Fig. 4d-g\\u003c/strong\\u003e). Notably, fluorescent co-localization of MxA and IBA1 demonstrated that\\u0026nbsp;microglia were the predominant source of \\u003cem\\u003eMxA\\u003c/em\\u003e expression\\u0026nbsp;at later time points, and that this activation was sustained in microglia well beyond viral clearance (\\u003cstrong\\u003eFig. 4g\\u003c/strong\\u003e). Moreover, while MxA\\u003cem\\u003e+\\u003c/em\\u003e microglia were observed throughout the OB, their staining intensity was markedly higher in the OMP+ region, suggesting that proximity to OSN axons may shape long-lasting microglial activation.\\u003c/p\\u003e\\n\\u003cp\\u003eHaving implicated microglia as the probable drivers of sustained immune activity in the OB, we next aimed to more fully define the nature of their activation. To this end, we analyzed OB RNA-Seq time course data to profile transcriptional signatures associated with core microglial functions. These analyses revealed robust and persistent upregulation of genes involved in phagocytosis and antigen presentation (e.g., \\u003cem\\u003eAif1, Cd68, Cd14, Fcgr1a, Fcgr2b, Itgam, Itgb2, Ptprc, Cd74, Cd209a, B2m\\u003c/em\\u003e), innate immune sensing and inflammatory signaling (e.g., \\u003cem\\u003eTlr2, Tlr4, Siglec1, Syk, Il1b, Tnf, Il18, Cx3cr1\\u003c/em\\u003e), immune cell adhesion and trafficking (e.g., \\u003cem\\u003eItgb1, Lyx\\u003c/em\\u003e), and apoptotic cell clearance (\\u003cem\\u003eTrem2, Cd36, Mrc1, Cd163\\u003c/em\\u003e), reflecting a broad immune activation state associated with post-injury homeostasis (\\u003cstrong\\u003eFig. 4h-j\\u003c/strong\\u003e).\\u003c/p\\u003e\\n\\u003cp\\u003eIn addition to the above microglia signatures, arguably the most striking feature was the sustained upregulation of genes associated with the classical complement cascade, a pathway increasingly recognized for its role in activity-dependent synaptic pruning\\u003csup\\u003e42\\u003c/sup\\u003e. Notably, we observed persistent elevation of \\u003cem\\u003eC1qa, C1qb, C1qc\\u003c/em\\u003e, and downstream components such as \\u003cem\\u003eC3, C4a, C3ar1\\u003c/em\\u003e, and \\u003cem\\u003eC5ar1\\u003c/em\\u003e, all of which have been implicated in tagging synapses for elimination during development and in response to injury (\\u003cstrong\\u003eFig. 1h; Extended Data Fig. 4h\\u003c/strong\\u003e). This transcriptional profile suggests that microglia in the OB remain engaged in complement-mediated remodeling well beyond the resolution of acute infection, potentially contributing to circuit refinement or, conversely, maladaptive synaptic loss. Given this, we next sought to identify a potential trigger for these activities. Here, the localization of microglial recruitment and immune signaling to the outer OB provided an important clue that OSN projections may be serving as this trigger. Following initial SARS-CoV-2 infection in the OE, OSN cell bodies were observed to be damaged and cleared from the OE. Despite this clearance, OSN axonal projections were still notably visible feeding into the outer OB, suggesting a potential delayed clearance of these OSN components (\\u003cstrong\\u003eFig. 4k\\u003c/strong\\u003e). Quantification of OMP staining intensity in the outer OB confirmed this delayed clearance, as OMP signal was largely sustained in the OB until ~14dpi, at which point marker intensity decreased significantly. This significant decrease was sustained out to 30dpi (\\u003cstrong\\u003eFig. 4l\\u003c/strong\\u003e) and notably aligned with the upregulation of inflammatory genes and genes important to microglial debris clearance (\\u003cstrong\\u003eFig. 4h-j\\u003c/strong\\u003e). Further evidence of this delayed clearance of OSN components in the OB could be seen in comparison of bulk sequencing data from the OB and OE. These data showed that at early time points following SARS-CoV-2 infection (4dpi), marker genes of OSN cells decreased at significantly greater rates in the OE than the OB, suggesting delayed degradation of OB OSN components (\\u003cstrong\\u003eFig. 4m\\u003c/strong\\u003e). Analysis of full time-course sequencing data showed that maximum negative enrichment of OSN marker genes in the OB was delayed until at least 8dpi \\u003cstrong\\u003e(Extended Data Fig. 4i\\u003c/strong\\u003e). Together, these data suggest that OSN axonal projections remain fixed in tissue despite loss of OSN cell bodies due to SARS-CoV-2 infection; microglia may thus be recruited and activated to clear this neuronal debris.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eAlthough molecular profiling of the human OB following COVID-19 remains challenging, primarily due to the difficulty of obtaining this tissue post-mortem, several datasets are available. To explore whether the immune signatures observed in our animal models are recapitulated in humans, we conducted a meta-analysis of publicly available mass spectrometry data from post-COVID donor OBs. This analysis revealed consistent upregulation of IFN-I signaling and antigen presentation pathways, closely mirroring the transcriptional patterns observed in our preclinical studies (\\u003cstrong\\u003eExtended Data Fig. 4j\\u003c/strong\\u003e). To further support these findings, we examined RNA-sequencing data from OB tissue collected from two additional post-COVID donors, which likewise showed elevated expression of genes associated with microglial activation and phagocytic activity, including \\u003cem\\u003eTREM2, FCGR1A, CD209, CD68, ITGAM, FCGR2B,\\u003c/em\\u003e and \\u003cem\\u003eTLR4\\u003c/em\\u003e (\\u003cstrong\\u003eExtended Data Fig. 4k\\u003c/strong\\u003e). While the limited number of human samples tempers definitive interpretation, these converging signals suggest that the same immune processes observed in model systems may also be operative in the human brain following SARS-CoV-2 infection, offering a potential mechanistic link to the lingering neurological symptoms reported in long COVID.\\u003c/p\\u003e\\n\\u003cp\\u003eAlthough SARS-CoV-2 primarily targets the respiratory tract, clinical and imaging studies have shown that even mild infections can produce persistent structural and functional brain changes detectable long after viral clearance\\u003csup\\u003e39\\u003c/sup\\u003e. These findings have led to the hypothesis that a persistent viral reservoir underlies post-acute symptoms, but evidence of a sustained and functionally replicative viral reservoir in the brain is currently lacking\\u003csup\\u003e43\\u003c/sup\\u003e. Challenging this concept, here, we propose an alternative mechanism by which a transient respiratory infection leads to lasting neurological symptoms. We show that SARS-CoV-2-induced damage to the OE causes bystander loss of OSNs. Due to the unique anatomy of OSNs, whose axons traverse the cribriform plate to the OB, this peripheral injury extends centrally. As a product of their length, degenerating OSNs take time to be cleared, triggering sustained microglial activation and persistent immune signaling. This phenomenon has notably also been observed in other rodents and zebrafish, suggesting induction of a conserved damage response to OSN injury\\u003csup\\u003e44-46\\u003c/sup\\u003e.\\u003c/p\\u003e\\n\\u003cp\\u003eTogether, these findings outline a conserved mechanism by which a self-limited peripheral infection can drive prolonged, localized neuroinflammation. By uncovering a damage-mediated glial response in the absence of direct viral invasion and temporally removed from acute peripheral infection, this work provides insight into the pathophysiology of long COVID and broader principles by which acute peripheral insults can durably alter brain function. This work therefore highlights new therapeutic opportunities for post-viral and neuroinflammatory syndromes.\\u003c/p\\u003e\"},{\"header\":\"Declarations\",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eAcknowledgments\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eWe would like to thank the Zegar Family Foundation and the\\u0026nbsp;Conestoga Road Foundation\\u0026nbsp;for enabling these studies.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eAuthor contributions\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe study was designed by J.T.F. and B.T. The experiments were performed J.T.F, J.D., L.H., C.A., S.U., R.A.S., T.T.W, and C.C.C. Data analysis was performed by J.T.F. Preparation and writing of the manuscript was performed by J.T.F, J.M., and B.T.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eCompeting interests\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eThe authors declare that they have no competing interests.\\u003c/p\\u003e\"},{\"header\":\"Materials and Methods \",\"content\":\"\\u003cp\\u003e\\u003cstrong\\u003eViruses and cells\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eSARS-CoV-2 (USA-WA1/2020) (NR-52281) (Biodefense and Emerging Infections Research Resources Repository, BEI Resources) was propagated in Vero-E6 cells (American Type Culture Collection) that were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco) supplemented with 2% fetal bovine serum (FBS) (Millipore Sigma), 1 mM Hepes (Lonza Bioscience), and 1% penicillin/streptomycin (Thermo Fisher Scientific). Propagation culture supernatants were filtered using Amicon Ultra-15 centrifugal filter units (Sigma-Aldrich). Viral stocks were sequenced to ensure that the furin cleavage site was maintained and then stored at -80°C until use. Following single use, stocks were discarded to prevent freeze-thaw cycling of viral particles. Infectious viral titers were determined via plaque assay in Vero-E6 cells with an overlay of DMEM supplemented with 2% FBS, 1 mM HEPES, and 0.7% Oxoid agar (Thermo Fisher Scientific). Plaque assays were fixed with 4% paraformaldehyde (PFA) (Thermo Fisher Scientific) prior to staining with crystal violet solution (Sigma-Aldrich). All infections in these studies were performed using passage three or four SARS-CoV-2 virus.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eHamster Studies\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eMale Syrian golden hamsters (\\u003cem\\u003eMesocricetus auratus\\u003c/em\\u003e) aged six to seven weeks were obtained from Charles River Laboratories. Hamsters were allowed to acclimate to U.S. Centers for Disease Control and Prevention (CDC)- and U.S. Department of Agriculture (USDA)-approved biosafety level 3 (BSL-3) facilities at either Icahn School of Medicine at Mount Sinai (ISMMS) or New York University Langone Health (NYULH) for at least 7 days prior to inclusion in experiments. Hamsters were randomly assigned into treatment groups and treated intranasally with 100uL of either phosphate-buffered saline (PBS) (Gibco) or a viral stock containing 1000 plaque-forming units (PFU) of SARS-CoV-2 (diluted as needed in PBS) under ketamine/xylazine anesthesia. Hamsters were euthanized at various time points following infection via intraperitoneal injection of pentobarbital and cardiac perfusion with either 60 mL of cold PBS (for experiments where tissues were taken for RNA-based assays) or 30 mL of cold PBS followed by 30 mL of cold 4% PFA (for experiments where tissues were taken for exclusively histology-based assays). Tissues set to be analyzed via RNA-based assays were harvested into TRIzol (Thermo Fisher Scientific) and homogenized in Lysing Matrix A homogenization tubes (MP Biomedicals) using a Fast-Prep-24 5G bead grinder and lysis system (MP Biomedicals). Samples were homogenized for 40 seconds at 6 m/s for two cycles. Tissues harvested for histological assays were collected into 4% PFA and allowed to fix for at least 72 hours before being washed three times in PBS to prepare for downstream processing. All animal experiments were performed according to protocols approved by the IACUCs and Institutional Biosafety Committees at ISMMS and NYULH. All infections and handling of infectious material took place in BSL-3 facilities.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eQuantitative Reverse Transcription Polymerase Chain Reaction (RT-qPCR)\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eRNA was isolated from tissues homogenized in TRIzol via standard phenol-chloroform extraction. One microgram of extracted total RNA was reverse transcribed using the SuperScript III reverse transcriptase system and oligo(dT) primers (Thermo Fisher Scientific). The resulting cDNA was assayed via qRT-PCR using the KAPA SYBR Fast Master Mix (KAPA Biosystems) and a Thermo Fisher Quantstudio 7 Instrument (Thermo Fisher Scientific). Delta-delta cycle threshold or inverted raw Ct values were used to compare transcriptional changes between SARS-CoV-2- and mock-treated hamsters as detailed in the primary text. Graphical representations of these data were generated using GraphPad Prism (version 10) (GraphPad Software).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eDecalcification of Snout Tissues and Histological Processing\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003ePrior to paraffin embedding, formalin-fixed head and snout tissues were processed as detailed above (see “Hamster Studies”) and were washed in PBS three times. These tissues were then dissected and trimmed to leave only connected bony snout and anterior cranial cavity (containing the olfactory bulb). These snout and anterior cranial cavity tissues were then placed into at least five-fold volumetric excess of EDTA Bone Decalcifier Buffer (pH 7.4) (alternatively labeled “Buffered Versenate, pH 7.4”) (StatLab) and placed on a rocker (10 rpm) at room temperature for 10 days. EDTA Bone Decalcifier Buffer was replaced with fresh buffer every two days during this incubation. Following the tenth day of incubation, snout tissues were rinsed three times in PBS and then embedded in paraffin. Paraffin blocks were cut into 5uM sections using a microtome and mounted onto charged glass slides for downstream processing.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eImmunofluorescent Staining and Imaging\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eSlide-mounted sections were processed for immunofluorescent staining as described previously\\u003csup\\u003e1,20\\u003c/sup\\u003e. In brief, sections were deparaffinized via submersion in xylene and rehydrated in gradated ethanol solutions. Antigen retrieval was performed by submerging slides in IHC-Tek Epitope Retrieval Solution (IHC-Tek cat. IW-1100) and steaming for 45 minutes in the IHC-Tek Epitope Retrieval Steamer (IHC-Tek cat. IW-1102). Tissue was blocked with a solution of tris-buffered saline (TBS) (Fisher Scientific) supplemented with 10% goat serum (Millipore Sigma) and 1% bovine serum albumin (Thermo Fisher Scientific). Primary antibodies were diluted in TBS supplemented with 1% bovine serum albumin and added to slides. Slides were incubated overnight at 4° C and then washed with a solution of TBS supplemented with 0.025% Triton X-100 (Thermo Fisher Scientific). Fluorophore-conjugated secondary antibodies (goat anti-rabbit IgG AlexaFluor 568 and goat anti-rabbit IgG AlexaFluor 647 [Thermo Fisher Scientific, Cat# A-11011 and A-21245, respectively]) were diluted to a concentration of 1:1000 in TBS with 1% bovine serum albumin and added to slides as appropriate given primary antibody staining. Slides were incubated at room temperature in the dark for 1 hour and were then washed in TBS supplemented with 0.025% Triton X-100. Section nuclei were then stained with 4′,6-diamidino-2-phenylindole (DAPI) and rinsed in PBS prior to coverslipping with ProLong Diamond Antifade Mountant (Thermo Fisher Scientific). Slides were imaged using a Zeiss Axio Observer epifluorescence and crossed-polarization inverted microscope platform with initial processing and stitching using Zeiss ZEN (Blue Edition) software (Carl Zeiss Microscopy).\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eBulk RNA Sequencing (FDA, retrieve our own data)\\u0026nbsp;\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eRNA-sequencing FASTQ files analyzed in these experiments were retrieved from NCBI GEO datasets, GSE161200 and GSE203001. Each dataset contained longitudinally sampled olfactory tissues from SARS-CoV-2- and control-treated animals. Following data retrieval, RNA-sequencing reads were aligned to the \\u003cem\\u003eMesocricetus auratus\\u003c/em\\u003e transcriptome (assembly BCM_Maur_2.0) using Salmon 1.10 software. Differential expression analysis was then conducted on each experiment independently using DESeq2 to compare SARS-CoV-2-infected tissues to control tissues. All samples were compared to internal experimental controls, and all log2(Fold Change) results comparing infected to uninfected subjects were obtained from experimentally matched samples. All genes with P\\u003csub\\u003eadj\\u003c/sub\\u003e \\u0026lt; 0.1 were considered “differentially expressed genes” (DEGs). Graphical representations of data were generated using the R programing language and the ggplot2 and heatmap packages.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eSingle-Cell RNA-Sequencing Analysis\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eRaw FASTQ files generated from single-cell RNA sequencing of longitudinal SARS-CoV-2-infected and control hamster olfactory epithelium (OE) tissues were retrieved from the 4DN data portal\\u003csup\\u003e25\\u003c/sup\\u003e. These files were processed and aligned to the \\u003cem\\u003eMesocricetus auratus\\u003c/em\\u003e genome (assembly and annotation BCM_Maur_2.0) using Cell Ranger (10x Genomics). The resulting files were processed in Seurat 5.1.0 using the R programming language (The R Foundation). Cells with mitochondrial gene expression surpassing 5% of the total transcriptome and with greater than 6,000 or less than 500 uniquely expressed genes were excluded from analysis in Seurat. Samples were normalized using the SCTransform method and subsequently integrated. Cellular clusters were defined with 30 dimensions using a resolution of 0.5. Conserved cluster-specific markers were used to determine cluster identity. Differential expression analysis was then conducted using the FindMarkers function in Seurat to compare cell-type-specific transcriptional changes in all genes between infected and uninfected conditions. Genes in each respective differential expression analysis were ranked by Log2(Fold Change) and analyzed with preranked Gene Set Enrichment Analysis (GSEA) conducted by the fgsea R package. Gene sets assessed included the pre-curated C5 (ontology) and C2 (curated) gene set collections maintained by the Molecular Signatures Database (MSigDB) (UC San Diego, Broad Institute). Additional analyses included cellular trajectory (pseudotime) analysis, which was conducted on identified neuronal and precursor populations using the Monocle 3 package in R, and statistical analysis of individual or clustered gene expression, which was conducted with the ggpubr R package. Visualizations of single-cell RNA-sequencing data were created using the R programming language and the ggplot2, Seurat, Monocle3, and fgsea packages.\\u003c/p\\u003e\\n\\u003cp\\u003e\\u003cstrong\\u003eQuantification of Microscopy Data\\u003c/strong\\u003e\\u003c/p\\u003e\\n\\u003cp\\u003eFollowing stitching and initial processing of microscopy images in the Zeiss ZEN (Blue Edition) software, imaging files were imported into QuPath (0.5.1) for further processing and quantification. All images that were part of a staining set were combined into a QuPath project and processed analogously to allow consistency in visualization. For quantification of OE thickness, distance was measured from the luminal surface of the epithelium to the lamina propria on slide sections stained for olfactory marker protein (OMP), a marker for olfactory sensory neurons (OSNs). Measurements were taken at 100mm intervals on the epithelial layer overlaying the septum or medial-most wall of each nostril, and these values were averaged for each subject to give a mean value of OE thickness. For measurement of all cells and of OSNs per specific length of OE, the epithelial layer of the septum or medial-most wall of each nostril was encircled using the QuPath “Closed Polygon Annotation” tool, capturing all cells between the luminal surface of the OE and the lamina propria. Nuclei and OMP positivity within this capture area were algorithmically quantified using the QuPath “Positive Cell Detection” tool. For OE quantification images, nuclei were labeled with DAPI, and OMP was labeled with Cy5. “Positive Cell Detection” tool settings reflected this labeling and cells were designated as OSN positive if the mean Cy5 intensity in the nucleus plus a 2um expansion zone surpassed a threshold of 1,000 Cy5 brightness intensity units. This threshold was verified visually by the experimenter to appropriately capture cells that were OMP positive (OMP+) and exclude cells that were OMP negative (OMP-). The luminal surface of the encircled olfactory epithelial layer was also traced to measure the span of OE captured. Reported values (cells per length of OE; OSNs per length of OE; fraction of OMP+ OE cells) were calculated from these measurements. Analogous protocols were followed for analysis and quantification of Ki67 (rather than OMP) positivity.\\u0026nbsp;\\u003c/p\\u003e\\n\\u003cp\\u003eFor quantification of areas of “bare” OE, the entire length of the luminal surface of all visible nostril epithelium was traced and measured in OMP-labeled sections. These same surfaces were then subsequently re-traced, but only in areas where OMP+ cells could be visually observed in the epithelium underlying the traced surface. “Bare” lengths of membrane were calculated by subtracting the OMP+ luminal surface length from the total luminal surface length. Quantification of OMP (Cy5), MxA (Alexafluor 568), and IBA1 (Alexafluor 488) in images of labeled olfactory bulb tissues was performed by manually outlining the olfactory bulbs in the images using the QuPath “Closed Polygon Annotation” tool. These outlines were refined through the use of the QuPath “Pixel Classifier” tool, which was set to only retain areas where tissue was present within the encircled region. These areas were further segmented into an OMP+ “outer region” (containing outer areas of the olfactory bulb including the glomerular layer and olfactory nerve layer) and an OMP- “inner region” (containing inner areas of the olfactory bulb including the external plexiform layer, mitral cell layer, internal plexiform layer, and granule cell layer) by using the QuPath “Pixel Classifier” tool. Areas above and below a threshold of 400 Cy5 fluorescence intensity units (here marking OMP staining) were annotated as OMP+ and OMP-, respectively. A schematic displaying this process can be seen in Extended Data Fig. 4a. Mean fluorescence intensities for each channel and tissue area were calculated within OMP- and OMP+ regions, and the QuPath “Positive Cell Detection” and “Pixel Classifier” tools were used as described above to quantify cells positive for IBA1 and MxA in each region. The area of these annotations, as well as the mean fluorescence intensity of all markers within these annotations, were also quantified to enable multivariable measurements such as MxA staining intensity within the confines of IBA1+ regions. All image quantification measurements were calculated from these raw values. Each olfactory bulb was quantified independently and reported as an independent data point.\\u003c/p\\u003e\"}],\"fulltextSource\":\"\",\"fullText\":\"\",\"funders\":[],\"hasAdminPriorityOnWorkflow\":false,\"hasManuscriptDocX\":true,\"hasOptedInToPreprint\":true,\"hasPassedJournalQc\":\"\",\"hasAnyPriority\":true,\"hideJournal\":true,\"highlight\":\"\",\"institution\":\"\",\"isAcceptedByJournal\":false,\"isAuthorSuppliedPdf\":false,\"isDeskRejected\":\"\",\"isHiddenFromSearch\":false,\"isInQc\":false,\"isInWorkflow\":false,\"isPdf\":false,\"isPdfUpToDate\":true,\"isWithdrawnOrRetracted\":false,\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true},\"keywords\":\"\",\"lastPublishedDoi\":\"10.21203/rs.3.rs-7179724/v1\",\"lastPublishedDoiUrl\":\"https://doi.org/10.21203/rs.3.rs-7179724/v1\",\"license\":{\"name\":\"CC BY 4.0\",\"url\":\"https://creativecommons.org/licenses/by/4.0/\"},\"manuscriptAbstract\":\"\\u003cp\\u003eAcute viral infections can cause lasting symptoms in anatomically distant, uninfected tissues, a phenomenon that challenges traditional notions of viral pathogenesis. A leading example is long COVID, a condition in which neurological and other clinical symptoms can materialize long after viral clearance\\u003csup\\u003e1\\u003c/sup\\u003e. In investigating the mechanisms underlying this phenomenon, we found that despite evading direct infection, olfactory sensory neurons (OSNs) demonstrate a progressive slow decline following SARS-CoV-2 infection, triggering a prolonged neuroinflammatory response that persists for weeks to months post viral resolution. Using both small animal models and human clinical samples, we demonstrate that the virus selectively infects sustentacular cells in the olfactory epithelium (OE), leading to structural disruption and secondary OSN loss. Axonal debris from degenerating OSNs accumulates in the olfactory bulb (OB), where it triggers sustained activation of resident microglia cells and persistent inflammatory signaling. These immune responses are spatially restricted to OB regions innervated by the damaged neurons and are marked by transcriptional programs involved in phagocytosis, synaptic remodeling, and debris clearance. Together, these findings delineate a conserved neuron-to-glia injury axis in which peripheral neuronal damage initiates a protracted cascade that, despite the absence of direct central nervous system infection, culminates in delayed and persistent neuroinflammation. This mechanism offers a unifying framework for how transient respiratory infections can lead to persistent neurological sequelae, including those seen in long COVID.\\u003c/p\\u003e\",\"manuscriptTitle\":\"A Peripheral Neuron-to-Microglia Signaling Axis Connecting Transient Viral Infection to Persistent Neuroinflammatory States\",\"msid\":\"\",\"msnumber\":\"\",\"nonDraftVersions\":[{\"code\":1,\"date\":\"2025-07-31 05:49:10\",\"doi\":\"10.21203/rs.3.rs-7179724/v1\",\"editorialEvents\":[],\"status\":\"published\",\"journal\":{\"display\":true,\"email\":\"info@researchsquare.com\",\"identity\":\"researchsquare\",\"isNatureJournal\":false,\"hasQc\":true,\"allowDirectSubmit\":true,\"externalIdentity\":\"\",\"sideBox\":\"\",\"snPcode\":\"\",\"submissionUrl\":\"/submission\",\"title\":\"Research Square\",\"twitterHandle\":\"researchsquare\",\"acdcEnabled\":true,\"dfaEnabled\":false,\"editorialSystem\":\"\",\"reportingPortfolio\":\"\",\"inReviewEnabled\":false,\"inReviewRevisionsEnabled\":true}}],\"origin\":\"\",\"ownerIdentity\":\"9c8beeaf-98dd-46ac-b46f-2abec25164ef\",\"owner\":[],\"postedDate\":\"July 31st, 2025\",\"published\":true,\"recentEditorialEvents\":[],\"rejectedJournal\":[],\"revision\":\"\",\"amendment\":\"\",\"status\":\"posted\",\"subjectAreas\":[{\"id\":52385236,\"name\":\"Biological sciences/Microbiology/Virology/SARS-CoV-2\"},{\"id\":52385237,\"name\":\"Biological sciences/Immunology/Neuroimmunology\"}],\"tags\":[],\"updatedAt\":\"2025-08-22T22:20:11+00:00\",\"versionOfRecord\":[],\"versionCreatedAt\":\"2025-07-31 05:49:10\",\"video\":\"\",\"vorDoi\":\"\",\"vorDoiUrl\":\"\",\"workflowStages\":[]},\"version\":\"v1\",\"identity\":\"rs-7179724\",\"journalConfig\":\"researchsquare\"},\"__N_SSP\":true},\"page\":\"/article/[identity]/[[...version]]\",\"query\":{\"redirect\":\"/article/rs-7179724\",\"identity\":\"rs-7179724\",\"version\":[\"v1\"]},\"buildId\":\"XKTyCvWXoU3ODBz1xrDgd\",\"isFallback\":false,\"isExperimentalCompile\":false,\"dynamicIds\":[84888],\"gssp\":true,\"scriptLoader\":[]}","source_license":"CC-BY-4.0","license_restricted":false}