Abstract
Background Anosmia is a frequent symptom in patients with the coronavirus disease 2019 (COVID-19) driven by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and mostly recovers within weeks. This clinical figure is significantly different from that of anosmia after upper respiratory infection, which occurs in only a small proportion of patients and does not recover or requires months to recover. The background mechanisms of COVID-19 induced olfactory dysfunction have not been elucidated.
Methods To address the unique pathophysiology of olfactory dysfunction associated with COVID-19, we examined the existence and distribution of ACE2 (virus binding receptor), TMPRSS2 and Furin (proteases to facilitate virus entry) in the nasal mucosa, composed of the respiratory mucosa (RM) and olfactory mucosa (OM), and the olfactory bulb (OB) in mouse and human tissues by immunohistochemistry and gene analyses.
Results Ace2, Tmprss2, and Furin gene expressions were confirmed in the nasal mucosa and OB. ACE2 was widely expressed all in the RM, OM and OB. Co-expression of ACE2, TMPRSS2, and Furin was observed in the RM including the RE and subepithelial glands and in the OM, especially in the supporting cells on the olfactory epithelium and the Bowman’s glands. Notably, the olfactory receptor neurons (ORNs) in the OM were positive for ACE2 but almost negative for TMPRSS2 and Furin. The cells in the OB expressed ACE2 strongly and Furin weakly and did not express TMPRSS2.
Conclusions ACE2 was widely expressed in the RM, OM and OB, but TMPRSS2 and Furin were expressed in certain types of cells and were absent in the ORNs. These findings, together with clinically reported ones, suggest that COVID-19 related anosmia can occur due to mainly sensorineural and central dysfunction and, to some extent, conductive olfactory dysfunction. That the ORNs express ACE2 but not TMPRSS2 or Furin may explain the early recovery of anosmia.
Short Summary Protein expression patterns of ACE2, TMPRSS, and Furin suggest that COVID-19 related anosmia can occur due to mainly sensorineural dysfunction without olfactory neuronal damage.
Introduction
The recent international spread of the coronavirus disease 2019 (COVID-19) driven by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) poses a serious health emergency. COVID-19 usually begins with simple respiratory symptoms such as fever, shore throat, and cough for 2-3 days (1, 2). Notably, chemosensitive disorders, such as loss or decline of smell (anosmia or hyposmia), and loss of taste (ageusia or dysgeusia), have been repeatedly reported as a unique clinical feature of COVID-19 (3, 4), and now is considered as typical symptoms of the early stage of the SARS-CoV-2 infection (4, 5). Recently published meta-analysis demonstrated a 52.73% prevalence of olfactory dysfunction among 1,627 COVID-19 patients (6). Of particular, a high rate of anosmia (a complete loss of smell) has been well documented (3, 5, 7, 8). In contrast, only a smaller proportion (up to 20%) of patients with upper respiratory infection (URI) exhibits olfactory dysfunction (9). The prognosis of olfactory dysfunction due to URI is generally poor, with a majority of patients showing no or slight recovery within a few months (10), whereas olfactory dysfunction in patients with COVID-19 recovers relatively early, with reported duration being from 1 to 4 weeks (3, 8, 11–13). In a cohort of COVID-19 patients with olfactory dysfunction, loss of smell was reported to be the first symptom of COVID-19 in 27% of patients (12). These three clinical features in COVID-19 are significantly different from those found in patients with common URI, which occurs due to various virus including rhinovirus, picornavirus, and parainfluenza virus (9, 14–16).
Cell entry of SARS-CoV-2 depends on binding of the viral spike proteins to a cellular receptor, angiotensin-converting enzyme 2 (ACE2) (17–19) and depends on spike protein cleavage by Furin (19–21) and spike protein priming by host cell proteases including transmembrane protease serine 2 (TMPRSS2) (19, 22). Thus, high expressions of ACE2, TMPRSS2, and Furin are considered to enhance SARS-CoV-2 entry and the resulting clinical symptoms. To date, the expression of ACE2 (23, 24) and TMPRSS2 (24), but not Furin, has been reported in the nasal epithelium by using gene analyses. However, histological evaluation of their expression has been limited; only one study reported expression of ACE2 and TMPRSS2 proteins in the nasal mucosa and respiratory sinus, but it did not demonstrate immunostaining images of the nasal mucosa (25).
The nasal mucosa is divided into the respiratory mucosa (RM) and olfactory mucosa (OM) based on its histological components and functions. The RM consists of various types of epithelial cells including the ciliated columnar and goblet cells. The OM serves olfaction and consists of the olfactory epithelium (OE) and subepithelial tissues (26). The degree of olfaction is closely related to the number of mature olfactory receptor neurons (ORNs) in the OE. The olfactory system consists of peripheral compartments such as the OM and central structures, such as the olfactory bulb (OB) and piriform/entorhinal cortex (27).
In the present study, in order to elucidate the background mechanisms of olfactory dysfunction and the pathogenesis of high viral load in the upper airway in COVID-19 patients, we investigated the expression of ACE2, TMPRSS2, and Furin in the RM, OM, and OB in human and mouse tissues.
Methods
Experimental samples
Animal tissue samples were all obtained from the mice examined in the previous published studies (26, 28), because purchasing new animals had been prohibited in our facility due to the epidemic spread. The samples from 6 eight-week-old male C57BL/6 mice (26) and an eight-week-old-male Sprague Dawley rat (28) were used, and the following paraffin-embedded tissues were collected; the RM area and OM area of the nose, the OB area, and the kidney and prostate for positive controls of immunostaining (Figure 1A). Human tissues were obtained from patients undergoing surgery for the treatment of chronic sinusitis or olfactory neuroblastoma. These included the OM (n = 3), the middle turbinate (n = 5), and inferior turbinate (n = 6); the latter two were used for the evaluation of the RM. Routine morphology was evaluated in haematoxylin and eosin-stained sections by a qualified pathologist and otolaryngologists. Tissue evaluation was performed only in the parts characterized as non-diseased. All experiments were conducted in accordance with institutional guidelines and with the approval of the Animal Care and Use Committee of the University of Tokyo (No. P14-051, P15-115) and of the Research Ethics Committee of the Graduate School of Medicine and Faculty of Medicine, the University of Tokyo, Japan (12009, 2019073NI). Since archived specimens were used, written informed consent was waived.
Histological analyses
To detect the expressions of ACE2 and TMPRSS2 in the RM, OM, and OB, histological analyses were performed by immunostaining. Four-μm-thick serial paraffin sections were deparaffinized in xylene and dehydrated in ethanol before immunostaining. Prior to immunostaining, deparaffinized sections were treated with 3% hydrogen peroxide to block endogenous peroxidase activity and were incubated in Blocking One (Nacalai Tesque, Kyoto, Japan) to block non-specific immunoglobulin binding. After antigen activation, primary antibodies against ACE2 (1:300 dilution; rabbit monoclonal, Abcam, ab108252; Cambridge, UK), TMPRSS2 (1:1000 dilution; rabbit monoclonal, Abcam, ab92323; Cambridge, UK), Furin (1:100 dilution; rabbit monoclonal, Abcam, ab183495; Cambridge, UK), and PGP9.5 for a neuronal marker (1:500 dilution; guinea pig polyclonal, Abcam, ab10410; Cambridge, UK) were detected with peroxidase conjugated appropriate secondary antibodies and a diaminobenzidine substrate. The mouse kidney and rat prostate were stained for positive controls for ACE2 and for TMPRSS2 and Furin, respectively (Figure 1B). All samples were stained under the same condition and protocol as the positive control staining. Images of all sections were captured using a digital microscope camera (Keyence BZ-X700) with 4×, 10x, 20×, and 40x objective lenses.
Gene expression analyses
Our previous DNA microarray data from the nasal mucosa and OB (NCBI Gene Expression Omnibus database under the series number GSE 103191) was used to examine the expressions of ACE2, TMPRSS2, and Furin. The expression levels of each gene were normalized against the expression level of Rps3 (encoding ribosomal protein S3) in each sample.
Results
The immunohistological data is summarized in Table 1. ACE2, TMPRSS2, and Furin were present in the human and mouse nasal mucosa and in the mouse OB, though the expression pattern of ACE2, TMPRSS2, and Furin varied among tissues. Remarkably, co-expression of ACE2, TMPRSS, and Furin was detected in the supporting cells and the Bowman’s glands in the OM and diffusely in the RM, but not in the ORNs of the OM or the OB.
In the mouse RM, ACE, TMPRSS2, and Furin were all strongly expressed in the cytoplasm of the RE cells and the subepithelial glands (Fig. 2a, b). ACE2 and TMPRSS2 were highly co-expressed in the RE. The villous brush border of the respiratory columnar epithelium was strongly positive for TMPRSS2 expression. Moderate cytoplasmic staining for TMPRSS2 and Furin was observed in the subepithelial tissue (Fig. 2a, b). In the OM, only the supporting cells and the Bowman’s glands expressed all of ACE2, TMPRSS2, and Furin. The olfactory nerve bundles were moderately positive for ACE2 and TMPRSS2. Notably, all cells in the OE including the supporting cells, ORNs, and basal cells, were definitely positive for ACE2, while the ORNs was negative for TMPRESS2 and Furin (Fig. 2c, d). In the OB, no cells expressed ACE2, TMPRSS2, or Furin simultaneously (Fig. 3a-c). ACE2 positive cells were recognized in the glomerular layer, mitral cell layer, and granule cell layer, whereas those cells were all negative for TMPRSS2. On the contrary, some cells in the glomerular layer and mitral cells were positive for Furin. TMPRSS2 was strongly expressed in the cells in the core of the OB (Fig. 3b).
To reinforce the above histological result, we investigated the presence or absence of gene expressions of Ace2, Tmprss2, and Furin in the mouse nasal mucosa and OB. From the database of the previous study, expression of the three genes in the nasal mucosa and OB were confirmed (Fig. 3d).
In human nasal mucosa, PGP9.5 antibody clearly visualized the OE containing olfactory neurons. ACE2 was localized in PGP9.5-positive ORNs. While Furin were not present in the OE, TMPRSS2 was weakly expressed in the apical layer of the OE. (Fig. 4a, b). In the RM, ACE2, TMPRSS2, and Furin were widely co-expressed in the epithelium (Fig. 4c, d). These findings were basically identical to those found in mouse tissues.
Discussion
COVID-19 has been observed to present with numerous clinical symptoms, including a deteriorated sense of taste and smell, respiratory and digestive disorders (3, 4, 29). The results of the present study explain why olfaction is frequently impaired in COVID-19 patients. We observed the immunolocalization of ACE2, TMPRSS2, and Furin in the nasal tissue and the OB, which were considered to play a pivotal role in the manifestation of olfactory dysfunction induced by SARS-CoV-2 infection. Histologically, ACE2, TMPRSS, and Furin were co-expressed in the supporting cells and the Bowman’s glands on the OM, and the RM, but not in the ORNs of the OM or the OB. Confirmation of Ace2, Tmprss2, and Furin gene expressions in the nasal mucosa and the OB supported these immunohistochemical findings.
Olfactory dysfunction is defined into three types according to anatomical location; conductive, sensorineural, and central dysfunction (30). Conductive dysfunction results from blockage of odorant airflow to the OE. Sensorineural dysfunction is caused by damage of the ORNs and olfactory nerve, impaired olfactory adaptation, and/or odorant transport. The supporting cells and the Bowman’s glands are involved in olfactory adaptation and neurotrophic and physical support for the OE (31–33). Thus, if the function of the supporting cells and Bowman’s glands is deteriorated, odor adaptation is impaired and subsequent sensorineural dysfunction occurs. Central dysfunction occurs by the damage of the olfactory processing pathways in the central nervous system (30). Based on the present histological results, olfactory dysfunction induced by SARS-CoV-2 can be either conductive, sensorineural, or central dysfunction, but be mainly sensorineural dysfunction without olfactory neuronal damage.
In the olfactory mucosa, co-expression of ACE2, TMPRSS2, and Furin in the supporting cells and the Bowman’s glands suggest that COVIT19 may induce deterioration of mucus production and OE support, resulting in impairment of odor adaptation and transduction. Moreover, co-expression of ACE2 and TMPRSS2 in the olfactory nerve bundle implies odor transduction can be impaired by neuronal dysfunction. It is unlikely that SARS-CoV-2 directly damages the ORNs in the OM, because the ORNs expressed ACE2, but not TMPRSS2 or Furin. Also, the severe damage of the ORNs cannot explain the early recovery of olfaction in COVID-19 patients with anosmia, since the turnover rate of ORNs is approximately 30 days (34). Centrally, in the OB, co-expression of ACE2 and Furin in the mitral cells, which have large cell body and secondary dendrites, suggests that central olfactory dysfunction may occur due to synaptic inhibition from the ORNs through to processing at the olfactory bulb. Although most patients with COVID-19 are likely to improve olfactory dysfunction, some patients did not recover their olfaction in a few months (8). COVID-19 patients with prolonged olfactory dysfunction may suffer from continuing sensorineural and central olfactory dysfunction.
Considering that high expression of all ACE2, TMPRSS2, and Furin in the RE and the subepithelial glands, SARS-CoV-2 possibly induces conductive olfactory dysfunction due to hypersecretion and goblet cell hyperplasia (14, 35). In fact, not a few patients with COVID-19 are reported to suffer from nasal obstruction and rhinorrhea (3). However, olfactory symptoms due to conductive olfactory dysfunction may fluctuate and rarely become completely lost. Also, certain proportion of COVID-19 patients with hyposmia and anosmia did not exhibit nasal obstruction or rhinitis symptoms (3, 36). Thus, the contribution of conductive olfactory loss may be limited.
The present study demonstrated high expression of ACE2 with TMPRSS2 and Furin in the nasal mucosa and OB. Their expression patterns suggest that olfactory dysfunction in COVID-19 patients may be mainly of sensorineural and central and, to some extent, conductive.
Footnotes
Funding: This work was supported by JSPS KAKENHI Grant-in-Aid for Scientific Research (C) [grant number 24791749, 16K20231] and by the Smoking Research Foundation (Tokyo, Japan).