Olfactory immunology: the missing piece in airway and CNS defence

Abstract

The olfactory mucosa is a component of the nasal airway that mediates the sense of smell. Recent studies point to an important role for the olfactory mucosa as a barrier to both respiratory pathogens and to neuroinvasive pathogens that hijack the olfactory nerve and invade the CNS. In particular, the COVID-19 pandemic has demonstrated that the olfactory mucosa is an integral part of a heterogeneous nasal mucosal barrier critical to upper airway immunity. However, our insufficient knowledge of olfactory mucosal immunity hinders attempts to protect this tissue from infection and other diseases. This Review summarizes the state of olfactory immunology by highlighting the unique immunologically relevant anatomy of the olfactory mucosa, describing what is known of olfactory immune cells, and considering the impact of common infectious diseases and inflammatory disorders at this site. We will offer our perspective on the future of the field and the many unresolved questions pertaining to olfactory immunity.

Introduction

The ongoing coronavirus disease 2019 (COVID-19) pandemic has highlighted the importance of vaccines that prevent severe illness, hospitalization and death^1,2. At the same time, the widespread prevalence of breakthrough infections and reinfections^3–5, even in previously vaccinated individuals, illustrates the shortcomings of vaccines. Breakthrough infections typically present with milder symptoms that are contained to the upper respiratory tract, but these still come with serious consequences. Upper respiratory severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection can lead to smell loss^6–8, long COVID^8,9 and potential systemic viral dissemination. Perhaps most critically, upper airway infection allows for continued pathogen transmission^10,11. This not only presents a danger to immunocompromised individuals and the unvaccinated but also provides an opportunity for viral evolution leading to immune evasion. Data from a plethora of SARS-CoV-2 animal studies consistently indicate that nasal tissue is less well protected than the lung from reinfection following prior immunization or infection^11–23. Indeed, across airborne infection models, the nasal mucosa is generally unprotected even in the presence of systemic immunity^24–28.

Moving forward, the most pressing challenge for vaccinology is to design vaccines that generate sterilizing immunity at all portals of infection. Defending the entire upper airway from infection should be a key correlate for vaccine-induced protective responses, both against SARS-CoV-2 and future respiratory pandemics. The nasal airway is the entry point for many pathogens, and establishing protective immunity in this tissue is essential to break the chain of transmission. Many approaches have been suggested to orchestrate locally protective immunity at the nasal surface, including mucosal immunization routes, specific antigen formulations and distinct adjuvant signalling^29–35. The efficacy of these mucosal vaccination strategies is currently limited by two considerations: what immune parameters are required to protect the nasal passages? And how can this tissue-specific immunity be generated? Common wisdom has held that a tissue-tailored mucosal immune response is required for upper airway protection more so than in the lower respiratory tract^36–38. Numerous hypotheses have been offered insofar as to what constitutes this mucosal response: secretory IgA antibodies, tissue-resident T cells and mucosal cytokines are frequently mentioned^{22,23,38–41}. Moreover, how to best elicit these protective responses is unclear^{14,18,32–35,42} (Box 1). On top of that, how do we determine efficacy? Although peripheral blood is easily sampled, studies of mucosal tissue, particularly the nasal mucosa, are impeded by difficulty in tissue acquisition. Most prior work relies on nasal washes^39,43–47 that disproportionately sample the lower nasal turbinates and cannot capture the complete mucosal antibody response, particularly in the superior nasal turbinates. Therefore, determining how vaccination or infection impacts local nasal mucosal immunity and viral control is a technical challenge we have yet to overcome. But one consideration looms over the above questions: the nasal mucosa contains at least two distinct tissue types that require protection, namely the olfactory mucosa and the respiratory mucosa, and each possesses unique immune considerations (Fig. 1a).

Box 1. Vaccine approaches to protect the upper airway.

What immunization approaches most effectively protect the olfactory mucosa? Vaccines variably protect the olfactory mucosa and respiratory mucosa, and the tropism of many infectious diseases also differs between these tissues, but recent data measuring general infection of the nasal passages may provide clues. Prior infection has been shown to generally providew more complete protection of the URT than immunizations, and accordingly, live attenuated vaccines show improved nasal passage protection in animal models compared with inactivated antigen^100,310, although the olfactory mucosa (OM) can still be left exposed¹⁴. Mucosal immunization, particularly intranasal dosing, is one putative approach to induce local immunity. Intranasal vaccines often struggle to induce strong antibody titres³¹¹, probably owing to poor antigen and adjuvant retention, but can reduce nasal replication better than parenteral immunizations in some cases^14,18,312. ‘Prime and pull’ strategies try to synergize the efficacy of parenteral and mucosal immunization³¹³ and can improve nasal protection^314,315. Regardless of approach, immunization should engender lymphocytes that home to the olfactory mucosa to confer local protection. This homing is imprinted in the lymph node by the inflammatory signals induced either by infection or by a vaccine adjuvant. If adjuvants can successfully mimic the lymphocyte activation induced by infection, olfactory homing and protection would occur regardless of the route of immunization. Indeed, several studies show that parenteral immunization with non-conventional adjuvants can induce superior protection of the nasal passages^17,48,316, consistent with data from other mucosal tissues³¹⁷. More analysis of the adaptive response in the lymph node, and dissection of olfactory-specific protection, is needed to produce the best immunization formula.

Fig. 1 |. Heterogeneity of the upper respiratory tract: the olfaction fraction.

a, The upper respiratory tract consists of two distinct tissues with important implications for immunity, namely the olfactory mucosa (blue) and the respiratory mucosa (beige). The olfactory mucosa must balance olfaction with immune defence and acts as a barrier to the CNS. The olfactory and respiratory mucosae have different requirements for immune protection, including local humoral protection from resident plasma cells in the olfactory mucosa. b, Prior studies of intranasal infection and immunity often indicate that upon rechallenge with a pathogen, pathogen replication is reduced. However, these studies treat the nasal passages as a homogeneous tissue. In actual fact, the overall reduced pathogen replication that is observed could represent many different scenarios depending on the tropism of the pathogens used and the quality of the immune response they induce. The lower panel indicates some hypothetical examples, considering whether a pathogen shows tropism for the olfactory mucosa (OM) alone, the respiratory mucosa (RM) alone or for both the olfactory mucosa and respiratory mucosa. Differences between the olfactory mucosa and respiratory mucosa, both in pathogen tropism and protective immune parameters, must be carefully considered and analysed to yield interpretable data regarding consequences of infection or immunization.

Our recent work has provided an essential new insight into these conundrums⁴⁸. By focusing on infection in the olfactory mucosa, we identified a novel endothelial barrier, termed the blood–olfactory barrier (BOB), that prevents circulating serum antibodies from accessing the olfactory mucosa. Therefore, even in situations in which highly neutralizing blood-borne antibody is present, the olfactory mucosa is still vulnerable to infection. Interestingly, circulating antibody accesses and protects the respiratory mucosal surface in the nasal passages, and only the olfactory mucosa is strictly segregated from serum antibodies. This gap in immunity at the olfactory barrier can be overcome by a population of extravascular plasma cells that are driven to reside in the olfactory mucosa following infection, secreting antibodies that directly reach the mucosal surface. Although many of these plasma cells were IgA⁺, protection did not depend on IgA, consistent with other studies of nasal IgG⁴⁹. Plasma cell-mediated olfactory protection is dependent on signals given to B cells in the lymph node, but intriguingly, these mucosal plasma cells are not always generated following immunization. Using multiple mouse models of viral infection, we demonstrated that these olfactory plasma cells are absolutely required to protect the olfactory mucosa, and furthermore, to protect the CNS from neuroinvasive viruses. These results answer several outstanding questions about upper airway protection from infection. Blood-derived antibodies provide a critical layer of protection for the respiratory epithelium, but the BOB prevents circulating antibody from accessing and protecting the olfactory mucosa⁴⁸. This difference in antibody transudation may resolve conflicting studies on upper airway antibodies and infection^{11–23,39,43–47}, in which conclusions were confounded by nasal washes failing to distinguish olfactory and respiratory antibody and by pathogens infecting olfactory mucosa and respiratory mucosa to differing levels (Fig. 1b). Olfactory protection can, however, be achieved by locally protective mucosal plasma cells, and the generation of these plasma cells is dependent on the signals engendered by infection or immunization. Differences in vaccine formulation and delivery may at least partially explain why different vaccination strategies could protect the upper airway from infection to variable degrees (Box 1). These findings have illuminated the significance of a previously neglected field: olfactory immunology.

Perhaps most importantly for immunologists and vaccinologists, these discoveries emphasize the heterogeneity of the upper respiratory tract; with distinct respiratory and olfactory mucosae that exhibit different correlates for immune protection, the nasal tissue cannot be treated as a monolithic tissue (Fig. 1). Although protecting the olfactory mucosa is critical to breaking the transmission chain for respiratory pathogens, it is perhaps even more essential in the context of neuroinvasive pathogens. The olfactory nerve acts as a direct portal to the brain, forming a single-cell connection from the airway to the CNS that is essential for the sense of smell but can be subverted by pathogens. The olfactory mucosa is, therefore, a CNS mucosal barrier, the sole line of defence between the external environment and catastrophic neurological disease^50,51. Neuroimmunologists have subjected CNS barrier tissues such as the blood–brain barrier (BBB) and meninges to detailed examinations across neuroimmune diseases. These studies have yielded rich and comprehensive characterizations of their anatomical minutiae⁵², and yet, the olfactory barrier has been largely overlooked. This can be likened to a castle guard preparing for siege defence: the walls have been fortified, cracks in the battlements repaired, secret entrances have been sealed — but the front gate has been left open! Breaching other CNS barrier tissues, such as the tissues of the eye or meninges, requires penetrating several cell layers and structures to reach the CNS;^52,53 only the olfactory mucosa contains neurons that directly interface with the outside world and the CNS. Fortunately, despite our relative neglect, the olfactory mucosa has evolved several structural and immune barriers that safeguard this entryway (Fig. 2).

Fig. 2 |. Cell types and effector mechanisms in the olfactory mucosa.

a, The figure indicates the parenchymal cell types that compose the olfactory mucosa. The luminal side is coated in mucus and directly exposed to the airway. The neuroepithelium contains olfactory sensory neurons (OSN), sustentacular cells, microvillar cells and the Bowman’s glands. Lining the basal lamina are horizontal basal stem cells (HBCs) and globose basal stem cells (GBCs). Within the lamina propria, OSN axon tracts run directly towards the olfactory bulb of the brain. Olfactory ensheathing cells (OECs) are interwoven within these axon bundles. Also, within the lamina propria are lymphatic and blood endothelial cells. The blood–olfactory barrier (BOB) prevents antibodies and other large circulating molecules from entering the olfactory mucosa. The exact composition of the BOB is unknown; beyond endothelial cells, pericytes, macrophages or olfactory ensheathing cells could contribute to barrier integrity. b, Innate immune cells of the olfactory mucosa are indicated in dark red. At homoeostasis, macrophages can be observed within the neuroepithelium and lamina propria with several distinct morphologies. During inflammation, dendritic cells (DCs), neutrophils, monocytes and eosinophils can infiltrate the tissue and contribute to the immune response. c, Following infection or immunization, T cells and B lineage cells migrate to the olfactory mucosa and take up long-term residence. These lymphocytes (shown in purple) can provide protective local immunity against future challenge. d, Intracellular olfactotropic pathogens can be neuroinvasive or non-neuroinvasive. Viruses and bacteria may infect non-neuronal epithelial cells or OSNs. Neurotropic pathogens that infect OSNs may either be cleared before reaching the CNS or migrate through OSN axons into the olfactory bulb. Extracellular pathogens, such as bacteria and eukaryotes, can migrate along axon bundles to reach the brain. These pathogens are better able to infect the olfactory mucosa when the tissue has been previously damaged, compromising existing structural impediments. IFN, interferon; iNOS, inducible nitric oxide synthase; OM, olfactory mucosa; TNF, tumour necrosis factor.

And yet, to this point, little work has been done to characterize the olfactory mucosa as an immune tissue (Box 2), and several outstanding questions remain. Which haematopoietic populations reside within this tissue? How does infectious disease uniquely impact the olfactory mucosa compared with the rest of the upper respiratory tract (URT)? What mechanisms does the olfactory mucosa use to protect the brain from invasive threats? How can we augment olfactory mucosal tissue defence to lessen disease burden imposed by airborne pathogens? In this Review, we highlight the existing literature on immunology of the olfactory mucosa. This is an emerging field, and we will offer our perspective on the many unexplored questions in olfactory mucosal, URT and CNS immunity.

Box 2. Sniffing out the historical link between olfaction and immunity.

Emerging data have stirred memories of a long-suspected link between olfaction and immunity. Fundamentally, these two systems attempt to perform similar functions, namely, to recognize a foreign substance and coordinate a rapid physiological response. Olfaction can lead to changes in behavioural immunity³¹⁸, such as the ‘disgust’ response which is known to trigger a ‘prepared’ immune state in response to noxious and potentially threatening stimuli^319,320. Pleasant odours can also impact the immune system: exposure to soothing fragrances following stress leads to a decrease in inflammation³²¹. Removal of the olfactory bulbs leads to depression and a dysfunctional immune system in a manner not fully understood^312,322, but the neuroimmune signalling pathways that coordinate olfactory–immune crosstalk have been profiled in drosophila³²³. An additional structural parallel between olfactory chemosensation and immune recognition has been hypothesized to explain the tendency for mammals to choose mates with dissimilar major histocompatibility complex (MHC) polymorphisms. Peptides that bind to particular MHC molecules have been shown to bind receptors in the olfactory and vomeronasal organs^324–326, demonstrating a mechanistic link between olfactory and immune non-self-discrimination that could serve to increase MHC repertoire diversity and, thereby, disease resistance at the population level. Another structural analogue between olfactory and immune cells has recently been demonstrated, as olfactory receptors expressed by leukocytes have important roles in immunity³²⁷. Olfactory–immune communication is not unidirectional; infection and subsequent immune activation in drosophila led to altered olfaction³²⁸ and there is even growing evidence that mammals, such as dogs, can smell when a person is ill^329–333. This olfactory discernment is responsive to common immune or inflammatory substrates, especially volatile organic compounds, but can identify molecules specific to particular diseases³³⁴. Although poorly understood, a diverse body of literature indicates that the immune and olfactory systems can dynamically interact.

Immune barriers in the olfactory mucosa

Immunologic analyses of the URT have primarily addressed the respiratory regions of the nasal mucosa⁵⁴, especially in humans, neglecting olfactory regions almost entirely. The reasons for this are probably twofold. The first reason is a technical hurdle: inferior nasal turbinates or nasal polyps can be sampled with relative ease compared with more superior olfactory regions. The second reason is that there persists a false assumption that URT mucosae are homogeneous — that any sample is representative — ignoring that olfaction exists and requires special equipment. Thus, the human respiratory mucosa has been extensively characterized in contexts including rhinosinusitis, allergy and viral diseases such as COVID-19. Far less work has been done on immune cell populations and functions within the olfactory mucosa. In animal models, analysing the tissue-resident cells of the olfactory mucosa is further complicated by cell isolation from the nasal turbinates. When generating single-cell suspensions, tissue mincing causes turbinate bone marrow to intermix with mucosal cells. We have overcome this complication by developing an intranasal antibody labelling technique that distinguishes true olfactory mucosa CD45⁺ cells⁴⁸. In the following sections, we will review the constituent olfactory mucosal cells and structures and discuss how they contribute to immunity.

Anatomical barriers

Situated within the upper turbinates of the superior nasal cavity, the anatomical structure of the olfactory mucosa provides intrinsic obstacles to an incoming pathogen. In some species, the olfactory mucosa dominates the nasal space, but in humans, olfactory tissue is present only on the upper of three nasal turbinate pairs and is proportionally limited in size, covering approximately 5 cm². Air is needed to sample environmental odourants, but only 15% of air passing across the lower respiratory turbinates reaches the upper ‘olfactory recess’ in which air slows to increase odourant detection⁵⁵. Limiting air volume exposure may restrict pathogen exposure, but the olfactory system has several physiological barriers that support olfaction and counteract environmental threats.

Olfactory mucus.

The first structural barrier encountered by an olfactory pathogen is the surface mucus layer loaded with antimicrobial peptides^56–58. Although the entire nasal mucosa is lined by mucus, olfactory mucus is characterized by a tissue-specific combination of mucin 1, mucin 5AC and mucin 5B proteins that distinguish it from the neighbouring respiratory mucosa⁵⁹. Mucin specialization suggests that antimicrobial molecule production within the URT may also have spatially regulated expression patterns between respiratory mucosa and olfactory mucosa. Indeed, other proteins secreted into the olfactory mucus have unique immunomodulatory properties, such as olfactory binding proteins, which are essential for receptor-mediated olfactory chemosensation but also possess antimicrobial functions^60,61. Moreover, mucus collected from the human olfactory cleft has increased enzymatic activity compared with respiratory mucus, which may contribute to interactions with olfactory microfauna and their metabolites⁶². Future studies on olfactory mucus may yield further insight into tissue-specific properties that facilitate both odourant detection and immune defence.

Tissue structure.

Beneath the mucus layer lies the olfactory neuroepithelium^63,64, which has been extensively characterized by neurobiologists⁶³. This avascular pseudostratified columnar epithelium is packed full of olfactory sensory neuron (OSN) cell bodies that project dendrites into the mucus layer, wherein their specialized cilia access airway odourants. The neuroepithelium itself (Fig. 2) contains large numbers of structurally supportive sustentacular cells, as well as at least two distinct microvillar cell types⁶⁵ and Bowman’s gland ducts. Beneath these cells lie basal stem cell populations, globose basal cells and horizontal basal cells, that repopulate the neuroepithelial layer⁶⁶. To speedily convey sensory information, OSN axons extend basally from the cell body, pass through the basal lamina⁶⁷, converge in fascicular bundles and tunnel through olfactory lamina propria en route to the olfactory bulb of the brain. These fascicular bundles contain immune cells and olfactory ensheathing cells (OECs, refer to the following sections)⁶⁸. Between axon tracts, secretory Bowman’s glands and endothelial vessels, the olfactory lamina propria (which is proportionally larger in humans than in mice) contains numerous immune cell types and fibroblasts (Fig. 2a).

Endothelial barriers.

Endothelial cells within the nasal mucosa have distinct and unusual phenotypes, as demonstrated by a recent study that has identified atypical venous sinusoids and lymphatic vessels in respiratory and olfactory nasal tissue⁶⁹. Olfactory antigen delivery to draining lymph nodes has not been studied directly; however, unconventional LYVE1⁻VEGFR3⁺ collecting vessels in the olfactory regions probably have significant roles in tissue surveillance⁷⁰ (Fig. 2a). Several animal studies have demonstrated cerebrospinal fluid drainage through cribriform plate lymphatics before connecting to nasal lymphatics^71–77. Although the CNS is primarily drained by a cranial lymphatic network^52,78–80, the olfactory route may sometimes have an important physiological role as well. Supporting this hypothesis, a recent study has demonstrated that lymphangiogenesis induced by vascular endothelial growth factor C during autoimmune disease promotes CNS drainage through the olfactory lymphatics, emphasizing that olfactory lymphatics can also contribute to battling neuroinflammation in certain contexts⁷³. In humans, it is unclear how the olfactory lymphatic route may complement the CNS drainage known to occur through meningeal lymphatics⁷⁰, but post-mortem tissues^81,82 and in vivo nasal cerebrospinal fluid studies^82–84 suggest that material from the CNS may egress through the cribriform plate into the olfactory mucosa under certain conditions.

Blood–olfactory barrier.

Our group recently reported a novel endothelial barrier, the BOB, that has important implications for olfactory mucosa and CNS protection from airborne infections⁴⁸. This barrier restricts the movement of large circulating molecules, including antibodies, into the olfactory mucosa. Prior studies have sought to detect a blood–nerve barrier within the olfactory mucosa^85,86, and although these studies have concluded that some lower molecular weight compounds can access the olfactory mucosa from the bloodstream, these studies lacked the granularity and ability to distinguish between the endothelial and nerve fascia barriers. What might be the teleological purpose of the BOB? We propose that the BOB is a functional BBB extension, preventing circulating factors and pathogens from entering the olfactory tissues and travelling along the ‘nose-to-brain’ axis. Indeed, exploiting the olfactory nerve tracts to circumvent canonical CNS barriers has been used to intranasally deliver therapeutics into the brain^87,88, but also suggests a CNS vulnerability that the BOB may protect. The BOB probably also protects olfactory neurogenic potential and function from harmful circulating substances, much like the blood–retina barrier in the eye.

Like the BBB, the BOB dramatically restricts local tissue availability of larger serum proteins, requiring a fundamental reconsideration of what can and cannot support URT (olfactory) immune protection. We propose that upper airway breakthrough SARS-CoV-2 infections in vaccinated individuals could in large part be owing to the inability of circulating antibodies to protect olfactory tissues. Large molecular weight serum proteins such as antibodies and complement are probably excluded from the olfactory mucosa, but what are the exact molecular size restrictions imposed by the BOB? Can some molecules — especially drugs targeting the olfactory mucosa — be actively transported across? Do BOB formation and persistence depend upon the presence of neurons, and would it be retained in cases of tissue metaplasia in which neuroepithelium is replaced with respiratory cells? Are multiple cell types involved in BOB integrity, as with the neurovascular unit in the brain? These and other fundamental questions remain to be answered.

Adaptive immunity

B cells.

MHC class I expression within the olfactory mucosa is variable across cell types⁸⁹, yet OSNs are effectively devoid of MHC class I, making them especially reliant on humoral immunity. As previously noted, the BOB prevents serum antibody from accessing the olfactory mucosa, but local antibody secretion in the olfactory mucosa is highly protective. Without this pre-existing antibody protection directly at the mucosal surface, intracellular pathogens can infect OSN dendrites and translocate through axons into the brain without ever encountering other immune cells⁴⁸. Humoral immunity is, therefore, vital for protection against neuroinvasive microbes, but because the BOB prevents serum antibody from protecting all the cells of the olfactory mucosa, local antibodies would be essential to defending against even non-neurotropic airborne pathogens with olfactory tropism, such as SARS-CoV-2 (refs. 48,90–92). Local antibody production may also prevent early replication and continued pathogen transmission, while also preventing the olfactory mucosa from serving as a foothold for further pathogen selection. Influenza infection studies in ferrets indicate that nasal passage viral replication, but not lung replication, leads to transmission between individuals⁹³. Although this work has detected virus in both respiratory and olfactory mucosa, no study has yet to directly test whether olfactory infection alone permits transmission. Nevertheless, passive antibody transfer experiments suggest that olfactory viral replication may be sufficient for transmission. The BOB means that antibody transfer should protect the respiratory mucosa, but not the olfactory mucosa, and studies of murine influenza infection indicate that passive antibody transfer does not block viral transmission^94,95. Formally testing this possibility will require researchers to carefully design experiments to analyse transmission using viruses with known respiratory and olfactory tropism (Fig. 1b). But these studies are critical because selective replication in olfactory and not respiratory tissues could drive evolution of variants with enhanced olfactotropic and neuroinvasive qualities.

Olfactory B cells are detected in the lamina propria of human biopsies, often localized near secretory Bowman’s glands that stain positive for Ig molecules, perhaps indicating that these glands may assist in luminal antibody secretion⁹⁶ (Fig. 2c). However, intranasal antibody administration in mice demonstrates that IgG antibodies can freely diffuse throughout the tissue without being impeded by structural barriers such as the basal lamina⁴⁸. These B lineage cells appeared to produce all antibody isotypes, consistent with concurrent studies in salamanders and rats that demonstrated multiple antibody isotypes differentially distributed across the tissue⁹⁷. In response to olfactory infection, B cells from fish respond to challenge by producing IgT, a mucosal antibody that protects the olfactory surface^98,99. In addition to recruitment following infection, local live attenuated immunization increases the frequency of IgA⁺ plasma cells in the olfactory mucosa^48,100. Single-cell RNA sequencing studies have found B lineage populations, particularly class-switched plasma cells, in samples from mice and humans^89,101,102.

Our recent work suggests critical non-redundant roles for these olfactory plasma cells in olfactory mucosal defence against viral infection⁴⁸. These protective plasma cells appear to secrete several antibody isotypes, provide long-term protection, and intriguingly, can be driven to olfactory mucosa residence following non-local priming in distal lymph nodes, suggesting that parenteral immunization can imprint an olfactory mucosa-homing phenotype⁴⁸ (Fig. 2c). In addition to further elucidating the signals that dictate olfactory mucosa migration, numerous outstanding questions about olfactory plasma cells remain. Which local cells provide tissue retention signals and what is their relationship to long-lived plasma cells in the bone marrow or other mucosal sites? Our efforts have focused on neutralizing antibodies, but studies of olfactory mucosa murid herpesvirus infection indicate that antibodies can limit local viral replication in an Fc-dependent manner^26,103. Plasma cell-derived pre-existing antibody is critical for preventing olfactory mucosa infection, but memory B cell populations may also help maintain the plasma cell pool or respond to re-infection, as has been found in the lower respiratory tract¹⁰⁴. Fully understanding these local B lineage populations is paramount for vaccination against respiratory and neurotropic pathogens.

T cells.

Adaptive immunity in the olfactory mucosa also includes contributions from T cells. T cell populations (CD4⁺, CD8⁺ and γδ T cells) have been described in mouse nasal passages, but whether these were truly olfactory or contained respiratory or bone marrow contaminant is unclear^105,106. However, following influenza virus infection, nasal CD8⁺ resident memory T (T_RM) cells accumulated within the olfactory mucosa⁴¹. These antigen-specific T cells were extravascular long-term resident cells that, upon rechallenge, provided superior and more durable viral control than lung T_RM cells. Furthermore, the nasal passage T_RM cells alone were able to reduce viral dissemination to the lung, emphasizing the role olfactory CD8⁺ T cells can perform in respiratory virus defence⁴¹ (Fig. 2c). CD8⁺ olfactory T cells have also been observed during an olfactotropic viral infection in fish¹⁰⁷. Post-COVID-19, patients with smell loss were found to have olfactory infiltration of interferon-γ (IFNγ)-producing T cells, indicating that T cell-mediated inflammation can persist long after olfactory viral clearance¹⁰⁸. Similarly, another study has found that patients with long COVID had elevated olfactory mucosal interferon signatures and T cell-associated genes¹⁰⁹. Similar long-term changes in the olfactory chemokine expression, interferon-stimulated genes and T cell markers were observed in SARS-CoV-2-infected hamsters¹⁰⁹ (Fig. 2c). These studies and similar transcriptional analyses from humans suggest that T cell pathogenesis may have a role in dysosmia (altered sense of smell)¹¹⁰. In agreement with this, CD8⁺ T cell and natural killer T cell signatures are associated with human age-related olfactory loss and have been proposed to directly signal to the olfactory stem cell niche to disrupt neurogenesis¹¹⁰, and T cell cytokine production is observed in mouse models of chronic nasal inflammation^111,112. Together, these data suggest that olfactory T cells can control infections but may also instigate chronic olfactory changes. More work to characterize olfactory T cell subsets, including CD4⁺ T cells and γδ T cells, is needed.

Innate immunity

Macrophages.

Macrophages represent the most abundant immune cell type in mouse and human olfactory mucosa^89,101,102 (Fig. 2b). Tissue-resident macrophages are known to mediate various functions — including neuronal maintenance, wound repair and infection responses — and macrophages have been shown to perform each of these roles in the olfactory mucosa. Macrophages can be found in close association with OSNs, both within the neuroepithelium and within olfactory nerve fascicles¹¹³. Olfactory macrophages express Cx3cr1, and deficiency in this receptor leads to a reduction in dendritic morphology for intraepithelial macrophages¹¹⁴, a phenotype that mirrors morphological changes seen in CX₃CR1-deficient microglia¹¹⁵. Following OSN death, clodronate-mediated macrophage depletion reduced neurogenesis¹¹⁶ and expression of immune response genes such as Cxcr4 (ref. 117). Similar analyses implicate macrophage Lif and Msr1 expression in tissue regeneration¹¹⁸, perhaps directed by IL-1β signalling from dying OSNs and subsequent expression of Ccl2 and Ccl3 (refs. 119–121). Macrophages have also been shown to be important for defence against olfactotropic pathogens. Macrophage numbers are increased in mice following porcine hemagglutinating encephalomyelitis virus (PHEV) infection, correlating with upregulation of inflammatory mediators such as IFNs, IL-6 and tumour necrosis factor (TNF)¹²². Influenza A virus-infected OSNs became apoptotic and were engulfed by macrophages, preventing viral spread to the brain¹²³. Similarly, when Staphylococcus aureus was delivered intranasally after OSN damage, macrophages within the nerve fascicle were able to phagocytose the bacteria¹²⁴. Macrophages are further implicated in chronic olfactory mucosal inflammation, as patients with chronic rhinosinusitis (CRS) who present with smell loss have elevated macrophage numbers compared with controls^125,126. Mouse inflammation models recapitulate the elevated macrophage numbers^127,128 and macrophage skewing to an immune defence (IL-6 expressing) phenotype¹¹¹. How macrophages balance these conflicting roles in neuronal support and immune defence remains to be thoroughly examined, but these data suggest that there are functionally or ontologically distinct olfactory mucosal macrophage subsets.

Other innate immune cells.

Although less well studied than macrophages, other innate leukocytes can be detected within the olfactory mucosa, including dendritic cells^89,101,102 (Fig. 2b). Circulating myeloid cells, although probably not residing in the olfactory mucosa long-term, can impact the olfactory mucosa during inflammation. Monocytes may differentiate into macrophages in the olfactory mucosa following recruitment from blood, and monocytic inflammation has been observed in an olfactory listeriosis model¹²⁹. Neutrophilic olfactory mucosal inflammation occurs in numerous contexts and is often severe. Mice given intranasal poly(I:C) treatments see a rapid neutrophil influx¹¹² that subsequently launch themselves into the nasal airway¹²⁷, a phenotype also witnessed in amoeba infection¹³⁰. Intranasal lipopolysaccharide administration similarly led to neutrophil influx¹³¹, but the role of neutrophils in human CRS is less clear. Although elevated neutrophil levels have been observed¹²⁵, neutrophilia is not associated with smell loss in CRS¹³². Neutrophils may also have either beneficial or pathogenic roles in acute infection. Olfactory neutrophils are elevated in mouse and hamster SARS-CoV-2 infections^90,133. Upon neutrophil depletion or blockade, SARS-CoV-2 titres were actually decreased and olfactory mucosal damage was mitigated, suggesting that neutrophils contribute to olfactory mucosa destruction and permit increased viral replication¹³³. Among the other granulocytes (mast cells, basophils and eosinophils), only eosinophils have been definitively reported within the olfactory mucosa, typically in the context of CRS or amoebic infection¹³⁴ (Fig. 2b). Although eosinophil numbers in the olfactory mucosa seem to be elevated in human rhinosinusitis^126,135, they are not associated with impaired olfaction. Natural killer cells have been observed in human olfactory biopsies and may express inflammatory genes that signal to basal progenitors and OSNs, inhibiting their ability to properly regenerate¹¹⁰. To our knowledge, no studies have directly searched for other innate-like lymphocytes in the olfactory mucosa.

Stromal immune barriers

In addition to haematopoietic immune cells, parenchymal cells in the olfactory mucosa may contribute to immunity through cell-autonomous pathogen clearance, the production of antimicrobial and inflammatory compounds, and communication with the haematopoietic compartment (Fig. 2a).

Olfactory sensory neurons

OSNs represent a curious case in intrinsic immunity. Compared with CNS neurons, OSNs are accustomed to much higher rates of death and regeneration, suggesting that they may respond differently to inflammation. As OSNs are a single-cell gateway to the CNS, evidence suggests that they use unique antiviral signalling pathways to stifle intracellular infections that attempt to invade the brain parenchyma (Fig. 2d). Type III IFNs are a critical component of early mucosal responses to infection, and indeed, IFNλ reduces murid herpesvirus infection at the olfactory mucosa¹³⁶. Similarly, IFNλ signalling prevents influenza virus spreading from the olfactory mucosa to the lung¹³⁷. Conversely, vesicular stomatitis virus, a virus highly sensitive to type I IFN signalling, replicates aggressively in OSNs¹³⁸, suggesting inherent vulnerabilities in OSN type I IFN responses. Yet, other studies indicate that type I IFN has a critical role in combating URT viral infection prophylactically and after disease onset^139–142, although differences in olfactory and respiratory infection have not been quantified in these studies. Our work has demonstrated that OSNs can non-lytically clear influenza B virus infection more quickly than neighbouring respiratory epithelial cells by using a rapidly induced antiviral response¹⁴³. OSNs can also prevent virus from reaching the brain by quickly inducing apoptosis^107,123. In summary, OSNs certainly exhibit vulnerabilities to infection but can respond swiftly to some pathogens, and their antimicrobial capabilities remain to be carefully characterized.

Other epithelial cells.

Sustentacular cells provide structural support within the olfactory neuroepithelium and make up the largest portion of non-OSN cells. These cells are the target of SARS-CoV-2 infection in humans owing to their ACE2 expression^90–92,144. Following sustentacular cell infection, chemosensory function is impaired and the entire neuroepithelial layer appears to slough off, as the olfactory mucosa structure is compromised. Also within the neuroepithelium, olfactory microvillar tuft-like cells express Il25 and genes for cysteinyl leukotriene production, which they produce upon airway allergen exposure resulting in eosinophilia¹⁴⁵. An ensuing study demonstrated that microvillar Ltc4 expression induced by allergens stimulated olfactory stem cell proliferation⁶⁵, suggesting these cells may coordinate the immune response and neurogenesis (Fig. 2a).

Olfactory ensheathing cells.

OECs surround OSN axon bundles as they pass through the olfactory lamina propria into the CNS. OECs are related to astrocytes and Schwann cells, acting as an important glial component of the olfactory nerve^68,86,146. OECs are promising cellular therapies for treating brain and spinal cord injuries, probably owing to their neuroprotective and neurogenic functions, but studies also suggest that they have potent immune-modifying abilities¹⁴⁷. OEC phenotypes within the olfactory bulb are geared towards axon regeneration, whereas olfactory mucosa OECs express genes associated with the defence response, inflammation and immunomodulation¹⁴⁸. Advantageously poised to patrol olfactory nerve tracts (Fig. 2a), olfactory mucosa OECs can produce inducible nitric oxide synthase in response to bacterial invasion of the damaged olfactory nerve¹⁴⁹ and phagocytose infected or dying olfactory axons^150,151. In the context of OSN death, OECs recognize phosphatidylserine produced by dying axons, phagocytosing a greater OSN number than olfactory mucosa macrophages¹⁵⁰ in a process that may be enhanced by MIF and HTRA1¹⁵¹.

Together, these adaptive, innate and stromal cells coordinate to maintain olfactory function and combat disease. But what are the specific threats, infectious or otherwise, that impact the olfactory system? Next, we will review disease pathogenesis within the olfactory mucosa, with an emphasis on how the local immune response ameliorates or exacerbates disease.

Disease in the olfactory mucosa

Olfactotropic pathogens

Airborne pathogens initiate infection in the upper airway, in which they first encounter host defences. However, the specific impact pathogens have on the olfactory mucosa is poorly described, in large part owing to the technical difficulty in measuring microbial replication and corresponding inflammation of the human superior nasal turbinates (nasal swabs sample the lower respiratory turbinates of the nose). Consequently, the olfactory tropism of many common airborne pathogens is unknown, and we probably drastically underestimate the number of airway infections that impact the olfactory mucosa. Pathogens currently known to infect the olfactory mucosa, which we refer to as olfactotropic infections, are reviewed in Table 1. Here, we will highlight infections that have special implications for olfactory immunity. We can think of olfactotropic pathogens in two broad categories: neuroinvasive and non-neuroinvasive (Fig. 2d). The olfactory mucosa is heavily innervated by OSNs and pathogens can hijack OSNs for direct CNS invasion, resulting in potentially lethal meningitis or encephalitis. Many pathogens are known to exploit this entryway to the CNS^50,152, but ascertaining the proportion of meningitis and encephalitis cases that originate from olfactory infection is difficult. At the same time, many non-neuroinvasive respiratory pathogens infect both the olfactory mucosa and the respiratory mucosa, and olfactory immune defence must limit viral dissemination, break the chain of community transmission, and prevent olfactory mucosal damage and smell loss.

Table 1 |.

Infection stinks: pathogens in the olfactory mucosa

Pathogen	Notes	Refs.
Naegleria fowleri	Free-living amoeba that causes catastrophic lethal disease known as primary amoebic meningoencephalitis through the olfactory nerve in humans134,165; similar pathogenesis seen in animal models168	134,165,168
Balamuthia madrillaris	Free-living amoeba, opportunistic agent of lethal granulomatous amoebic encephalitis in humans and animals228; olfactory neuroinvasion seen in immunocompromised mice229 but not in immunocompetent organisms230	228–230
Ichthyophthirius multifiliis	Trout parasite that infects the olfactory epithelium but with overt CNS invasion	98
Mucorales spp.	Fungal pathogens have not been formally shown to invade the olfactory nerve, but Mucorales order fungi commonly cause rhino-orbital-cerebral mucormycosis, suggesting fungal invasion along the olfactory pathway	231
Streptococcus pneumoniae	Common human respiratory bacterial infection displays olfactory neuroinvasion in immunocompromised mice232,233; in immunocompetent mice, it traverses along the olfactory nerve to the meninges but does not infect the brain parenchyma234	232–234
Chlamydia spp.	Chlamydia muridarium and Chlamydia pneumoniae invade the olfactory nerve during intranasal infection in mice227,235, with C. pneumoniae upregulating brain inflammatory genes, suggesting olfactory contribution to neurodegenerative disorders associated with human C. pneumoniae infection227,236	227,235,236
Neisseria meningitidis	Bacterial human pathogen shown in transgenic mice to cause olfactory mucosa damage and neuroinvasion without blood dissemination	237
Listeria monocytogenes	Bacterial pathogen thought to access the human brain primarily through the trigeminal nerve but can enter via the olfactory epithelium in neonatal mice	129
Burkholderia pseudomallei	Intracellular bacterium that infects the olfactory mucosa of mice238, causing OSN death, exposing the axon tracts and allowing for stark neuroinvasion189; B. pseudomallei can shelter within the nerve fascicles and neuroinvasion is enhanced following prior olfactory mucosa damage190	189,190,238
Staphylococcus aureus	Common human bacterial infection that in mice can opportunistically invade the CNS through the olfactory nerve following olfactory mucosa damage	124
Flavobacterium columnare	Bacterial pathogen that invades the trout olfactory organ and provokes an immune response	99
SARS-CoV-2	Coronavirus that infects sustentacular cells of human olfactory mucosa, frequently causing inflammatory damage and loss of smell90–92,144,153; olfactory neuroinvasion does not seem to occur in humans, but animal models, such as K18-hACE2 mice, display CNS invasion via the olfactory nerve158,159,161; in Syrian golden hamsters, some later-emerging variants infect the brain through the olfactory nerve162,164	90–92,144,153,158,159,161,162,164
SARS-CoV-1	Does not penetrate CNS in wild-type mice but shows olfactory neuroinvasion in the K18-hACE2 mouse model239; autopsy study in humans detected CNS virus but did not sample nasal mucosa240	239,240
Human coronavirus OC43 (HCoV-OC43)	Intranasal inoculation in mice leads to CNS infection that focally radiates from the olfactory bulb241,242; frequently detected in human post-mortem brain samples243	241–243
Middle Eastern respiratory syndrome coronavirus (MERS-CoV)	Neurologic symptoms in humans;244,245 transgenic mouse studies indicated possible olfactory involvement in CNS infection246	244–246
Porcine hemagglutinating encephalomyelitis virus (PHEV)	Following mouse intranasal infection, invades the brain via the olfactory and trigeminal nerves	122
Influenza virus	In mice, several strains infect the nasal passages but not the brain41,137,143,247–250; ferret studies indicated olfactory mucosa infection in highly pathogenic strains all without reaching the brain179,181; olfactotropic CNS neuroinvasion occurs for some strains in either mice or ferrets178–182	41,137,143,178–182,247–250
Vesicular stomatitis virus (VSV)	A cytopathic rhabdovirus that infects livestock but rarely humans251; in mice, VSV rapidly infects OSNs and transmits to the brain252 before being cleared by T cells253; excellent olfactory immunity model owing to its strict olfactory tropism during intranasal infection and its ability to generate robust adaptive immunity254	251–254
Rabies	A case of fatal rabies encephalitis reported via airborne exposure, and with rabies virions detected in the olfactory bulb255; intranasal rabies challenge in mice demonstrated OSN infection and transsynaptic transmission into the brain256	255,256
Vaccinia	Vaccinia virus has been shown to infect animals intranasally and reach the brain, possibly through olfactory transmucosal invasion	257–259
Cytomegalovirus (CMV)	Human CMV causes olfactory bulb lesions in infants185 and childhood loss of olfaction186; murine-adapted CMV (MCMV) spreads between animals nasally, and inactive MCMV binds to OSNs, suggesting potential intranasal transmission260	185,186,260
Herpes simplex virus 1 (HSV-1)	Invades the brain through olfactory nerve in rats261; infects olfactory mucosa of mice but is not neuroinvasive262; HSV1 may infect human OSNs and reach the brain263, given the large number of associated neurological sequelae226, post-mortem HSV1 detection264,265 and long-term loss of olfaction in convalescent herpetic meningoencephalitis patients266; HSV1 accesses the brain through the trigeminal nerve and may enter these cells through neuron termini at the olfactory surface	226,261–266
Human herpesvirus 6 (HHV-6)	May infect human olfactory mucosa and invade the brain based on post-mortem antigen detection in the brain and in the nasal cavity	267
Murid herpesvirus 4 (MuHV-4)	Infects mouse OSNs before reaching B cells in the draining lymph node but does not undergo CNS neuroinvasion	26,268
Aujeszky’s disease virus	Known as pseudorabies, this herpesvirus infects pigs and invades along the olfactory tract following intranasal inoculation	269,270
Poliovirus	Intranasal instillation in primates leads to olfactory CNS neuroinvasion271–274, although biopsy analysis indicates this is unlikely to occur in humans275; mice expressing a poliovirus receptor transgene are susceptible to olfactory neuroinvasion following intranasal challenge276	271–276
Eastern equine encephalitis virus (EEEV)	Highly fatal but poorly transmitted virus in humans; brain invasion appears to occur via the olfactory nerve in guinea pigs, mice and rhesus macaques	277–279
Venezuelan equine encephalitis virus (VEEV)	Invades the brain via the olfactory route in both mice and macaques	28,280–282
Chikungunya virus	Infects olfactory neurons in immunodeficient mice	283
Zika virus	Zika virus infects olfactory ensheathing cells in neonatal mice, causing transient olfactory deficits; broad olfactory tropism seen in immunodeficient mice, and human olfactory organoids are also susceptible to infection	284
West Nile virus (WNV)	The leading global cause of arboviral encephalitis, but the CNS entry route is unknown; experimental aerosol administration leads to mouse olfactory neuroinvasion	285,286
Other flaviviruses	Japanese encephalitis virus, Murray Valley encephalitis virus and St. Louis encephalitis virus infect the olfactory epithelium and olfactory bulb	287–293
Parainfluenza	A large portion of patients with olfactory dysfunction tested positive for human parainfluenza virus 3 (ref. 294); mice intranasally infected with parainfluenza virus 1 (Sendai virus) had persistent viral RNA in the olfactory epithelium and bulb long after infection, suggesting viral immune evasion in the olfactory system295,296	294–296
Paramyxoviruses	Nipah and Hendra virus cause acute, severe respiratory disease and encephalitis in humans, and although haematogenous spread to the brain is probably involved, animal models demonstrate olfactory neuroinvasion for both viruses297–300; measles virus can be highly neuroinvasive and invades the olfactory nerve in hamsters301	297–301
Borna disease virus	Devastating CNS disease of livestock following faecal matter inhalation; when administered intranasally to rats, OSN neuroinvasion drives fatal encephalitis	302–304
Hepatitis virus	Murine hepatitis virus enters the CNS via the olfactory pathway305–308, causing significant OSN death and rapidly inducing brain chemokines and cytokines309	305–309

CoV, coronavirus; MERS, Middle East respiratory syndrome; OSN, olfactory sensory neuron; SARS, severe acute respiratory syndrome.

SARS-CoV-2.

The COVID-19 pandemic brought the impact of olfactory mucosal viral infection to the forefront of much scientific and public discourse. SARS-CoV-2 directly infects both the olfactory and respiratory epithelia in humans^91,92,144 (Fig. 2d). Fortunately, evidence suggests that OSN infection and subsequent CNS neuroinvasion do not occur^91,92,144. Instead, SARS-CoV-2 mediates olfactory pathology by infecting sustentacular cells leading to transient damage, inflammation and subsequent tissue structure loss^{90,91,108,144}. Without the structural support provided by sustentacular cells, olfactory neurons die or become dysfunctional, and smell is compromised to varying degrees^8,153. This partial or complete smell loss (clinically, hyposmia or anosmia) is typically short-term, as the olfactory epithelium is regenerated by the underlying stem cell populations⁶⁶, although persistent inflammation can lead to long-term hyposmia or anosmia¹⁰⁸. Interestingly, chemosensory deficits also strongly predict the humoral response in SARS-CoV-2 infection¹⁵⁴, suggesting a functional link between olfactory infection and immunity induction. The olfactory pathogenesis of SARS-CoV-2 is mirrored in rhesus macaques, as the typical URT viral replication is observed in the absence of frank CNS neuroinvasion^15,23. Replicating virus has not been detected in long-term hyposmic or anosmic olfactory biopsies, indicating that innate and adaptive immune responses can clear virus from the olfactory system¹⁰⁸. However, prolonged viral shedding has been observed from nasal swabs^155,156, suggesting that the olfactory mucosa could harbour virus in some individuals¹⁵⁷. Patients with long COVID often present with neurological symptoms, including olfactory deficits⁸, but whether this occurs because of viral persistence, cell-intrinsic OSN alterations or continued olfactory mucosal inflammation is unknown.

In contrast to human infections, animal models of SARS-CoV-2 and other coronavirus infections are characterized by olfactory neuroinvasion, raising concerns that future variants could gain neurovirulent capabilities as they repeatedly passage through olfactory tissues. In the commonly used K18-hACE2 mouse SARS-CoV-2 model, nearly ubiquitous epithelial hACE2 expression directs OSN infection and consequent CNS pathology is observed^158–160, resulting in lethal neuroinvasion across multiple SARS-CoV-2 variants¹⁶¹. Much like in humans, SARS-CoV-2 was initially only believed to infect olfactory mucosal sustentacular cells in hamsters¹⁶², but more recent variants have been shown to infect OSNs and invade the CNS^163,164. Hamsters also have lasting olfactory perturbations following SARS-CoV-2 infection, indicating that they may be useful for post-COVID olfaction studies¹⁰⁹. Given the frequency of coronavirus epidemic outbreaks this century (Table 1), more research on their olfactotropism is needed.

Naegleria fowleri.

Naegleria fowleri, the ‘brain-eating amoeba’, is perhaps the most notorious pathogen capable of olfactory transmucosal infection. This free-living amoeba is ubiquitously present in warm bodies of freshwater but only drives disease when it contacts olfactory tissue in the nasal turbinates¹⁶⁵. N. fowleri crosses the olfactory epithelium and quickly rampages through the olfactory nerve into the CNS (Fig. 2d) in which it causes an almost universally fatal inflammatory condition known as primary amoebic meningoencephalitis. Olfactory immunity against N. fowleri is complex: immunization against N. fowleri in animal models offers limited protection¹⁶⁶, but neutrophils and other myeloid cells slow disease progression but also contribute to disease pathogenesis^134,167. Because N. fowleri is only pathogenic across the olfactory mucosa, and olfactory neuroinvasion is conserved across mammals including mice, this infection serves as a powerful model system to highlight the unique immune properties and vulnerabilities of the olfactory mucosa¹⁶⁸. Seropositivity studies suggest that many humans may have some protection owing to subclinical exposure^169–171; but does the BOB prevent antibodies and/or complement from slowing pathogenesis? Are neutrophils or other cells able to respond to the amoeba more quickly in non-olfactory tissues?

Influenza virus.

Analysis of patients with influenza has shown that subjective olfactory dysfunction increases as vaccination rate decreases¹⁷², suggesting not only that frequent olfactory influenza virus infection occurs, but also a preventative immune capacity. Olfactotropism seems to depend on strain, but influenza virus infection is associated with neurologic symptoms and sequelae^173,174, and in some cases, influenza infection coincides with meningitis or encephalitis^175–177. Direct CNS neuroinvasion has been reported in several mammalian infection models. Influenza A/WSN/33 (H1N1) infects OSNs and translocates to the CNS in mice¹⁷⁸, and highly pathogenic and pandemic strains are predisposed to olfactory neuroinvasion in ferrets^179–182. In addition, influenza virus-derived antigen has been identified in human post-mortem olfactory nerves, lending credibility to an olfactory route of CNS infection¹⁸³. The recent highly pathogenic avian H5N1 strain of influenza virus has been shown to infect the brains of some animals¹⁸⁴, raising concerns that olfactory neuroinvasion may contribute to future emergent pandemics. Whereas such highly pathogenic cases are rare, the olfactotropism of more common, lowly pathogenic strains is understudied, and much remains to be learned about the viral and host factors that determine the neuroinvasive potential of various influenza virus strains. More common influenza virus strains may infect the olfactory mucosa, and potentially even reach the brain, but olfactory tropism is generally not explored unless a patient develops severe CNS disease. Supporting this, some influenza viruses infect the olfactory mucosa but innate mechanisms allow OSNs to quench viral replication to prevent neuroinvasion^123,143. The olfactory mucosa can also have a role in limiting influenza virus dissemination to the lung, as type I IFNs and type III IFNs were shown to be crucial for containing two different strains of influenza virus to the nasal passages of mice¹³⁷. Interestingly, comparing nasal and lung infection across multiple influenza virus strains in ferrets indicates that only nasal infection supports airborne transmission between organisms, whereas lung infections are not spread⁹³. These data reinforce the importance of preventing influenza virus infection in the upper airways to limit propagation.

Opportunistic infections.

Some pathogens may act opportunistically to infect the olfactory mucosa and reach the CNS. For example, cytomegalovirus (CMV) is a congenital disease in humans that frequently leads to neurological disorders such as hearing and smell loss, but whether CMV uses OSNs to infect the brain is unknown. Olfactory bulb lesions have been observed in infants with CMV¹⁸⁵, and olfactory defects are reported throughout childhood¹⁸⁶ in a manner that is decoupled from hearing loss¹⁸⁷. Furthermore, a human olfactory receptor was identified as a CMV entry receptor¹⁸⁸, opening the possibility that olfactory invasion may explain some CNS pathologies. Damage of the olfactory mucosa, whether acute or chronic, can expose it to opportunistic pathogens that may infect the tissue and penetrate the CNS. Olfactory mucosa damage in mice allows for bacteria such as Burkholderia pseudomallei^124,189,190, Streptococcus agalactiae¹⁹¹ and S. aureus¹²⁴ to subsequently colonize the olfactory mucosa and invade the brain, either intracellularly through OSNs or extracellularly along the axon tract⁵⁰ (Fig. 2d). The olfactory mucosa is damaged throughout life, from pathogenic and other insults, and maintaining structural and cellular olfactory mucosal barriers is critical for preventing opportunistic infections.

Post-viral olfactory dysfunction.

Post-viral olfactory dysfunction is one of the most frequently reported acute and chronic side effects of upper respiratory illness^{153,192–195}, including well-documented short-term and long-term olfactory loss following COVID-19^196–201. Potential mechanisms driving post-viral olfactory dysfunction include direct OSN infection and death, infection of other olfactory mucosal cells leading to inflammation and/or neuroepithelial damage, CNS consequences resulting in olfactory bulb dysfunction, or respiratory infection leading to nasal inflammation and airflow blockage. The immune response is heavily involved in all these pathologies, ranging from the rapid antiviral responses of epithelial cells to haematopoietic cell recruitment during sustained inflammation. Indeed, a recent study of olfactory biopsies following clearance of SARS-CoV-2 has identified that long-term smell loss was associated with immune cell infiltration and inflammatory gene expression¹⁰⁸. Persistent T cell infiltration was accompanied by a shift in myeloid cell populations away from an anti-inflammatory, wound-healing phenotype, reflecting a disruption in the balance between productive and deleterious immune responses to infection. This immune dysfunction extended to an inflammatory gene signature in sustentacular cells of the olfactory epithelium, as well as lower OSN numbers in dysosmic patients¹⁰⁸. Exactly how other pathogens mediate the loss of chemosensation warrants further study.

Damage of the olfactory mucosa

Olfactory injury, inflammation and regeneration.

Respiration brings a constant stream of airborne environmental pollutants, microbial toxins or inorganic compounds that can damage the olfactory mucosa and drive inflammation²⁰². CRS is a persistent inflammation of the upper airways that frequently coincides with dysosmia. Biopsies from patients with CRS indicate that inflammation in the olfactory mucosa is associated with hyposmia¹²⁶, and comparison of healthy control and CRS samples indicates increased olfactory metaplasia that can be characterized histologically by the type of epithelial deformation^125,135. Interestingly, one study of olfactory function found that type II cytokines were associated with worse olfaction prior to corrective CRS surgery, but improved olfaction postoperatively²⁰³. By contrast, type III cytokines correlated with better olfactory scores preoperatively, but corresponded with worse scores after surgery, suggesting the immune response may shape olfactory potentiation. Indeed, immune and glial cell activation supports phagocytosis of apoptotic debris and tissue regeneration after OSN ablation²⁰⁴, and infiltrating immune cells facilitate OSN regeneration¹²⁸. Furthermore, deficiencies in TNFR1 or basal progenitor cell NF-κB signalling led to defective olfactory regeneration, emphasizing the importance of immune crosstalk in neurogenesis¹²⁸. Overall, resident stem cells in the olfactory mucosa display remarkable regenerative capacity and therapies targeting these cells are currently being explored to combat dysosmia^205,206.

To understand the olfactory implications of chronic inflammatory conditions such as CRS and neurodegeneration, mouse lines that inducibly express inflammatory mediators such as TNF¹¹¹ and IL-13 (ref. 207) in the olfactory mucosa have been developed^131,208. The inducible olfactory inflammation mouse drives TNF expression from CYP2G1⁺ sustentacular cells²⁰⁹ to recapitulate CRS-induced olfactory loss and progressive olfactory neurodegeneration. Studies in inducible olfactory inflammation mice have indicated cognitive functional defects, epithelial reorganization, macrophage-mediated and T cell-mediated cytokine production, and reprogramming of basal progenitor cells from a proliferative to an inflammatory phenotype^111,209,210. Likewise, olfactory damage and inflammation can drive acute cytokine expression in the olfactory mucosa, neutrophil infiltration and tissue deformation^112,125,211, but critically, this peripheral inflammation can be communicated into the CNS¹³¹. In zebrafish, olfactory epithelial damage leads to rapid neutrophil recruitment into the olfactory organ of the brain²¹². Similarly, in hamsters, SARS-CoV-2-mediated olfactory mucosa inflammation is sufficient to induce changes in the olfactory bulb⁹⁰. The connection between mucosal and brain inflammation has important consequences, as frequent olfactory inflammation may contribute to neurodegenerative pressure²¹³ over time. Further work is needed to understand how olfactory mucosa inflammation not only drives immune-mediated olfactory disorders but also supports inflammatory communication between the olfactory mucosa and the CNS.

Presbyosmia and neurodegeneration.

Age-associated olfactory loss, or presbyosmia, is extremely common in elderly patients (occurring in >50% of adults over 65 years and in 60–80% of those aged over 80 years^214–217). Presbyosmia is associated with olfactory metaplasia, the replacement of olfactory epithelium with respiratory epithelium^218,219. This tissue conversion mirrors that observed in models of chronic olfactory mucosal inflammation²⁰⁹ and is probably owing to inflammation-related changes in olfactory stem cell progenitors¹¹⁰. Transcriptional evidence of strong cytokine stimuli and elevated immune cells in a presbyosmic cohort compared with controls suggest a direct role for sustained inflammation in age-associated olfactory loss¹¹⁰. Concordant with age-associated neuroinflammation, impaired olfaction is associated with (and is one of the strongest predictors of) cognitive decline²²⁰, Alzheimer disease²²¹, dementia²²² and Parkinson disease²²³. Myriad factors probably contribute to this association²²⁴, but prior olfactory infections may accelerate cognitive decline through repeated inflammatory stimuli. To this end, individuals with familial Alzheimer disease were found to express signatures of antiviral inflammation in the olfactory bulb and olfactory tract²²⁵. Corroborating this, HSV1, which can infect sensory fibres within the olfactory neuroepithelium, has been implicated in Alzheimer disease²²⁶. More directly supporting the olfactory infection hypothesis, in a mouse model, Chlamydia was shown to invade the CNS through the olfactory nerve and upregulate Alzheimer disease-associated gene signatures²²⁷. Much remains to be investigated about the bidirectional link between olfaction and neurodegeneration and the role infection might have in both.

Conclusion

The olfactory mucosa must be considered as a critical component of pathogen defence, both as part of the respiratory tract and as a mucosal barrier for the brain. Immune protection of the olfactory mucosa is vital for protection against continued respiratory pathogen transmission and neurotropic microbial invasion. Previously, it was difficult to reconcile our assumptions about peripheral immunity with many studies revealing incomplete URT immunity. We believe that appreciating the unique immunological considerations of the olfactory mucosa is not only critical for vaccine-induced URT immunity but also provides clarity into prior data and informs better experimental design. For instance, nasal washes and swabs insufficiently capture the olfactory mucosal immune response, and increased upper nasal turbinate sampling may reveal that many URT pathogens have distinct olfactotropism. An improved understanding of pathogen-induced olfactory dysfunction and neurodegeneration could lead to better and more targeted therapeutics for diseases such as COVID-19. A key consideration in drugs that target the olfactory system will be the BOB. Additional characterization of the BOB, and the role it has in immune defence, is also of interest for intranasal drug delivery and CNS anatomy. The BOB confers a degree of immune isolation and privilege to the olfactory mucosa, but very little is known about the tissue-resident immune cells in this tissue. Analysing these cells, and how they interact with stromal cells in both health and disease, will shed new light on URT and CNS defence. The olfactory mucosa deserves more attention as a mucosal immune barrier, and we smell the dawning of a new age in its study (pun intended).

Glossary

Anosmia

The complete loss of smell, typically defined clinically by the University of Pennsylvania Smell Identification Test scores <19

Blood–olfactory barrier

(BOB). A blood–endothelial barrier that prevents the movement of large molecules from circulation into the olfactory mucosa

Dysosmia

A general term for an altered sense of smell

Hyposmia

A reduced sense of smell, typically defined clinically by the University of Pennsylvania Smell Identification Test scores in the 19–33 range, although scoring can be adjusted by age and sex

Olfactory binding proteins

Soluble proteins in the nasal mucus that bind to odourants to facilitate recognition by olfactory receptors. They have also been shown to have antimicrobial effects

Olfactotropic

A pathogen that is capable of infecting cells within the olfactory mucosa

Presbyosmia

An age-associated loss of smell