Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) first infects the host nasal mucosa, where the viral spike protein binds to angiotensin-converting enzyme 2 (ACE2) on the mucosal cells. This study aimed at searching host cell surface molecules that could contribute to the infection in two views; abundance on host cells and affinity to the spike protein. Since the nasal mucosa is lined by respiratory and olfactory epithelia, and both express an immunoglobulin superfamily member cell adhesion molecule 1 (CADM1), whether CADM1 would participate in the spike protein binding was examined. Immunohistochemistry on the mouse nasal cavity detected CADM1 strongly in the olfactory epithelium at cell-cell contacts and on the apical surface but just faintly in the respiratory epithelium. In contrast, ACE2 was detected in the respiratory, not olfactory, epithelium. When mice were administered intranasally with SARS-CoV-2 S1 spike protein and an anti-CADM1 ectodomain antibody separately, both were detected exclusively on the olfactory, not respiratory, epithelium. Then, the antibody and S1 spike protein were administered intranasally to mice in this order with an interval of 1 h. After 3 h, S1 spike protein was detected as a protein aggregate floating in the nasal cavity. Next, S1 spike protein labeled with fluorescein was added to the monolayer cultures of epithelial cells exogenously expressing ACE2 or CADM1. Quantitative detection of fluorescein bound to the cells revealed that S1 spike protein bound to CADM1 with affinity half as high as to ACE2. Consistently, docking simulation analyses revealed that S1 spike protein could bind to CADM1 three-quarters as strongly as to ACE2 and that the interface of ACE2 was similar in both binding modes. Collectively, intranasal S1 spike protein appeared to prefer to accumulate on the olfactory epithelium, and CADM1 was suggested to contribute to this preference of S1 spike protein based on the molecular abundance and affinity.

Angiotensin-converting enzyme 2 (ACE2) is the receptor of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1, 2]. At the molecular level, this receptor function is achieved primarily based on high binding affinity between the S1 receptor-binding subunit of SARS-CoV-2 spike protein (S1 protein) and ACE2 [3, 4]. Therefore, as well as the affinity strength between these two proteins, the abundance of receptor molecules present on host cells must be another major determinant that allows SARS-CoV-2 to infect the host effectively and finally causes coronavirus disease 2019 (COVID-19) in the host. In the present study, we aimed at searching cell surface molecules that could contribute to the infection from two points of view; affinity to S1 protein and abundance on the host cells.

Dysosmia and anosmia are recognized to be characteristic symptoms in the early phase of COVID-19 [5–7]. This suggests that nasal mucosa serves as the primary site of SARS-CoV-2 infection. The nasal mucosa is composed of two different types of, i.e., respiratory and olfactory, epithelia [8]. The respiratory epithelium occupies the anterior half of the nasal mucosa in mice and is a columnar cell lining composed of goblet and ciliated cells [9]. On the other hand, the olfactory epithelium occupies the posterior half of the mucosa in mice and is essentially a kind of peripheral neuronal tissue including sustentacular support cells and olfactory neurons [10]. In humans, the olfactory epithelium is located in the uppermost part of the nasal cavity and occupies just 8% of the nasal cavity surface area [11], but the epithelial surface cellular components are essentially identical to those of the murine olfactory epithelium [12].

ACE2 protein expression is detected in various organs including the lung, kidney, and heart [13]. In the lung, bronchial epithelial and alveolar cells are positive for ACE2, but the protein expression level was rather low in the entire respiratory tract including the nasal mucosa [13, 14]. Thus, nasal mucosal cells are assumed to have other receptors than ACE2 or to have co-factors that help ACE2 serve as the SARS-CoV-2 receptor. Actually, recent studies have identified new receptors and co-factors, such as transmembrane serine protease 2, neuropilin 1 [15], sialic acid-binding immunoglobulin-like lectin 1 [16], and leucine-rich repeat-containing protein 15 [17].

Cell adhesion molecule 1 (CADM1) is a member of the immunoglobulin superfamily and a single cell-membrane-spanning glycoprotein that functions primarily as an intercellular adhesion molecule [18, 19]. The distribution profile of CADM1 proteins is unique. It is expressed in a certain limited type of epithelial cells, such as respiratory, biliary, and endometrial epithelial cells [20–22], and is also expressed widely in central and peripheral neuronal cells including the olfactory epithelium [23, 24]. This unique feature of CADM1 distribution encouraged us to probe possible involvement of CADM1 in the process of SARS-CoV-2 infection in the nasal mucosa, which has both respiratory and olfactory epithelia.

In the present study, we performed immunohistochemistry on the mouse nasal mucosa to compare the distribution of ACE2 and CADM1 proteins. We also compared the degree of binding of S1 protein to ACE2 and CADM1 in the epithelial cell monolayer culture system that we established previously [25]. Next, we administered S1 protein to the mouse nasal cavity and examined where the proteins were distributed. In addition, we examined whether S1 protein bindings could be interfered with by an anti-CADM1 ectodomain antibody. Finally, we conducted computer simulation analyses of docking between S1 protein and CADM1.

Cells, Antibodies, and Mice

Madin-Darby canine kidney (MDCK) cells were purchased and cultured in Eagle’s minimal essential medium with 10% fetal calf serum, as described in our previous report [26]. Primary antibodies used are; anti-ACE2 (21115-1-AP, Proteintech, Rosemont, IL, USA; for Western blot (1:1,000) and immunohistochemical (1:100) analyses; and E-11, sc-390851, Santa Cruz, CA, USA, for immunofluorescence, 1:100), anti-CADM1 C-terminal (S4945, Sigma-Aldrich, Tokyo, Japan; for Western blot (1:1,000), immunohistochemical (1:200), and immunocytofluorescence (1:100) analyses), anti-CADM1 N-terminal (3E1, chicken IgY, CM004-3, Medical & Biological Laboratories, Nagoya, Japan), anti-S1 protein (MonoRab 4G6, A02053, GenScript, Piscataway, NJ, USA; 1:1,000), anti-chicken IgY (AffiniPure, AB_2339284, Jackson ImmunoResearch, West Grove, PA, USA; 1:100), anti-GAPDH (mAb-HRP-DirecT, M171-7, Medical & Biological Laboratories; 1:2,000), anti-β-actin (PM053, Medical & Biological Laboratories; 1:1,000), anti-neuron-specific enolase (M0873, DakoCytomation, Glostrup, Denmark; 1:1,000), and anti-E-cadherin (610181, BD Biosciences, Franklin Lakes, NJ, USA; 1:1,000). C57BL/6 mice were purchased from Japan SLC (Hamamatsu, Shizuoka, Japan). All mice used are male at 10–12 weeks of age.

Plasmid Construction and Transfection

The mammalian expression vectors pCX4pur carrying the human ACE2 full-length cDNA (pCXpur-ACE2) and pCX4bsr carrying the human CADM1 full-length cDNA (pCX4bsr-CADM1) were constructed previously [25, 27]. MDCK cells were grown to 60–70% confluence and were transfected with either pCX4pur-ACE2 or pCX4bsr-CADM1 using the Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Cells were selected by resistance to puromycin or blasticidin for 2 weeks.

S1 Proteins and Fluorescein Labeling

Recombinant protein SARS-CoV-2 S1 subunit tagged with mouse IgG2a Fc portion (S1N-C5257, ACROBiosystems, Newark, DE, USA) was used as S1 protein, as described in our previous report [25]. A Fluorescein Labeling Kit-NH2 (Dojindo, Kumamoto, Japan) and an indocyanine green (ICG)-Labeling Kit-NH2 (Dojindo) were used to conjugate fluorescein and ICG to the S1 protein, respectively, according to the manufacturer’s instructions.

Detection and Quantification of Fluorescein Fluorescence in MDCK Cell Cultures

Three thousand transfected or original MDCK cells were suspended in 10 μL of culture medium and were transferred separately into the bottom of the micro-Insert 4-Well (gasket) on a µ-Dish (35 mm, high, no. 80406; ibidi, Martinsried, Germany). After 3 h of incubation, 150 μL of medium were poured into the insert to fill all wells with identical medium. The next day, fluorescein-labeled S1 protein was added to the culture medium at a concentration of 3.0 μg/mL. At the same time, 3E1 or control IgY (Medical & Biological Laboratories) was added at a concentration of 1.0 μg/mL in some experiments. After cell cultures were incubated 24 h, the micro-insert gasket was gently removed, and cells were washed 3 times with culture medium. Then, the µ-Dish was filled with 1 mL of culture medium and was placed on the microscope stage of a C2+ confocal laser scanning system (Nikon, Tokyo, Japan). Fluorescein fluorescent images were captured with a ×40 objective lens and analyzed on the Nikon C2+ computer system. Fluorescein intensity (arbitrary units per unit area) was measured at five randomly selected high-power fields for each well using analysis controls tools. Cell cultures in wells on µ-Dishes were prepared and measured in triplicate for each experimental group, and the mean and standard deviation of fluorescein intensities were calculated using ROI Statistics. Experiments were independently repeated three times with similar results.

Intranasal Administration of S1 Protein and Antibodies to Mice

Under anesthesia by isoflurane, C57BL/6 mice at 12 weeks of age were administered with S1 protein, 3E1, and control IgY (Medical & Biological Laboratories). These protein solutions were prepared at a concentration of 1 μg/μL in phosphate-buffered saline (PBS). A P10 pipette tip (110–207C, Watson Bio Lab, Kobe, Japan) containing 5 μL of each solution was inserted 1–2 mm into the right external naris and the solution was released gently into the nasal cavity. Then, 5 μL of the solution was released via the left naris in the same way. After nasal administration, mice were recovered from anesthesia and allowed to move freely in the cage. At timepoints indicated, mice were subjected to histological and Western blot analyses. The protocol was approved by the Institutional Animal Experimentation Committee of Kindai University Faculty of Medicine (KDMS-2021-005).

Immunohistochemistry and Immunofluorescence

The immunohistochemical procedures were described in detail previously [28]. Briefly, formalin-fixed, EDTA-decalcified, paraffin-embedded mouse heads were cut into sections (4-µm thick) at the mid-sagittal plane, air-dried overnight at 37°C, deparaffinized in xylene, and rehydrated in a descending ethanol series. After the sections were autoclaved for 20 min at 95°C in 10 mm citrate buffer solution (pH 6.0), they were blocked with 2% bovine serum albumin (BSA) and incubated with the anti-CADM1 (C-terminal) or anti-ACE2 antibody overnight at 4°C, followed by incubation with a peroxidase-conjugated anti-rabbit antibody (Cytiva, Tokyo, Japan) for 2 h at 4°C. Secondary antibody staining was enhanced using the Histofine Simple Stain MAX-PO (R) kit (Nichirei Biosciences, Tokyo, Japan). The sections were incubated with ImmPACT™ AEC (Vector Laboratories, Burlingame, CA, USA) and were then counterstained with Mayer hematoxylin, dehydrated, and mounted. Negative controls were prepared by substituting control rabbit IgG for the specific primary antibody.

For immunohistofluorescence, mouse heads were embedded and frozen in Tissue-Tek O.C.T. compound (Sakura Finetek, Torrance, CA, USA), scraped off roughly, and cut into sections (5-µm thick) at the mid-sagittal plane in a cryostat using the Kawamoto’s adhesive films (Cryofilm type 2C(9); Section-Lab Co. Ltd., Yokohama, Japan) [29]. To detect IgY singly, sections were fixed with 4% paraformaldehyde in PBS for 15 min at 4°C, and after wash with PBS, were blocked with 2% BSA in PBS for 1 h at room temperature. Then, sections were incubated serially with the anti-IgY antibody overnight at 4°C and with Alexa Fluor 594 anti-rabbit IgG (AffiniPure, AB_2340621, Jackson ImmunoResearch) for 2 h at 4°C. Nuclei were labeled with DAPI (D523, Dojindo) for 15 min at room temperature. To detect IgY and S1 protein doubly, unfixed sections were blocked with 1% BSA in PBS for 30 min at room temperature. Then, sections were incubated serially with a mixture of the Alexa Fluor 488 anti-chicken IgY (AffiniPure, 703-545-155, Jackson ImmunoResearch) and anti-S1 protein antibodies overnight at 4°C and next with Alexa Flour 594 anti-rabbit antibody (Jackson ImmunoResearch) for 2 h at 4°C. Nuclei were labeled with Hoechst 33258 (H3569, Invitrogen) for 15 min at room temperature. Fluorescent images were captured using an All-in-One fluorescence microscope (BZ-X800, Keyence, Osaka, Japan). For mice administered with ICG-labeled S1 protein, a frozen forehead section at the mid-sagittal plane was air-dried and ICG signals were detected using a Cy7 fluorescent filter cube.

For immunocytofluorescence, MDCK cells cultured in µ-Dishes were fixed with methanol for 10 min at −20°C and blocked with 2% BSA for 30 min at room temperature. Then, cells were incubated with a mixture of antibodies against ACE2 and CADM1 (C-terminal) overnight at 4°C and were visualized with Alexa Flour 488 anti-mouse IgG and Alexa Flour 594 anti-rabbit antibodies (Jackson ImmunoResearch), respectively. After washing with PBS three times, nuclei were labeled with DAPI for 2 h at 4°C. Fluorescent images were captured using a C2+ confocal scanning system equipped with 488-nm argon and 543-nm helium-neon lasers (Nikon).

Western Blot Analysis

Mouse heads were embedded and frozen in Tissue-Tek O.C.T. compound (Tissue Finetek) and scraped off roughly till the near mid-sagittal section emerged. The respiratory and olfactory epithelium-lining regions were delineated on this section, and these regional tissues were separately cut out using a razor. After frost shattering, the tissues were lysed in a buffer containing 50 mm Tris-HCl (pH 8.0), 150 mm NaCl, 1% Triton X-100, and 1 mm phenylmethylsulfonyl fluoride. In other experiments, cultured cells were washed in PBS and were lysed in the same buffer. After removal of impurities by centrifugation, lysates were subjected to Western blot analyses, as described in our previous report [30]. Primary antibodies are described above. Peroxidase-conjugated secondary antibodies are against rabbits and mice (Cytiva).

Molecular Simulation

The 3D structures of the of SARS-CoV-2 spike protein trimer (PDB ID: 6Z97; https://www.rcsb.org/) and ACE2 (PDB ID: 6 M0J) and the receptor-binding domain of the spike protein trimer were described in our previous reports [25, 31]. The structure of CADM1 monomer was built by the homology modeling function of the MOE software (Molecular Operating Environment, Chemical Computing Group, Montreal, Canada) using the partial modeling structure (50–345 aa) of CADM1 (ModBase ID: 7f7df202e8d6aa361f9ca3847afe4608) and de novo structure of CADM1 (346–387 aa) (obtained from the QARK software: https://zhanggroup.org/QUARK/) as templates. After the homology modeling process, GROMACS software (https://www.gromacs.org, Sweden) was subjected to molecular dynamics simulation program. Using the obtained CADM1 monomer structure, the ZDOCK software [32] was applied to model the structure of CADM1 cis-homodimer. Then, the number of S1 protein dockings to CADM1 and ACE2 was analyzed and ZDOCK scores were calculated [31]. The S1 spike protein interface used for each docking was determined using the binding mode. The interface was defined as outer aspects of the amino acid residues which distance to ACE2 or CADM1 is less than 5 Å in each binding configuration. For reference, we calculated ZDOCK scores of cis-homodimerization of transthyretin (PDB ID: 1F41), a carrier protein of thyroid hormones in the plasma [33].

Statistical Analysis

Fluorescence intensity measurements were conducted in triplicate for every experimental group, and the mean and standard deviation were calculated for each experimental group. Comparative analyses were done on experiments consisting of more than two groups (Fig. 1a upper panel). We compared fluorescence intensities using one-way analysis of variance and used the Bonferroni correction of one-way analysis of variance between any of two groups. Comparisons between specific two groups were done with Student’s t test (Fig. 1a lower right panel). p values ≤0.05 were considered statistically significant.

Fig. 1.

Quantification of S1 spike protein binding to ACE2 and CADM1 in MDCK cell cultures. MDCK cells were transfected with cDNA for either ACE2 (MDCK-ACE2) or CADM1 (MDCK-CADM1). MDCK, MDCK-ACE2, and MDCK-CADM1 cells were grown to a confluence, and fluorescein-labeled S1 spike protein (3.0 μg/mL) was added to the culture along with either control IgY (upper) or 3E1 (lower) at a concentration of 1.0 μg/mL. After 1 day of incubation and wash, fluorescent intensity of fluorescein remaining on the cells was measured using a confocal laser microscopy system. a Representative photomicrographs are presented. At the bottom of each image, the mean and standard deviation of the intensity (arbitrary unit) are presented for the corresponding experimental group. #p < 0.001 versus the intensity in MDCK cell + IgY and in MDCK-ACE2 + IgY. *p < 0.001 versus the intensity in MDCK-ACE2 cell + 3E1 and in MDCK-CADM1 + IgY. b Cell lysates were prepared from original and transfected MDCK cells and subjected to Western blot analyses using antibodies indicated. The blot was reprobed with an anti-β-actin antibody to determine the amount of protein loading per lane. c Cells were also stained by immunocytofluorescence using a mixture of primary antibodies against ACE2 and CADM1, which were visualized with green and red fluorescence, respectively. Nuclei are counterstained with DAPI.

Fig. 1.

Quantification of S1 spike protein binding to ACE2 and CADM1 in MDCK cell cultures. MDCK cells were transfected with cDNA for either ACE2 (MDCK-ACE2) or CADM1 (MDCK-CADM1). MDCK, MDCK-ACE2, and MDCK-CADM1 cells were grown to a confluence, and fluorescein-labeled S1 spike protein (3.0 μg/mL) was added to the culture along with either control IgY (upper) or 3E1 (lower) at a concentration of 1.0 μg/mL. After 1 day of incubation and wash, fluorescent intensity of fluorescein remaining on the cells was measured using a confocal laser microscopy system. a Representative photomicrographs are presented. At the bottom of each image, the mean and standard deviation of the intensity (arbitrary unit) are presented for the corresponding experimental group. #p < 0.001 versus the intensity in MDCK cell + IgY and in MDCK-ACE2 + IgY. *p < 0.001 versus the intensity in MDCK-ACE2 cell + 3E1 and in MDCK-CADM1 + IgY. b Cell lysates were prepared from original and transfected MDCK cells and subjected to Western blot analyses using antibodies indicated. The blot was reprobed with an anti-β-actin antibody to determine the amount of protein loading per lane. c Cells were also stained by immunocytofluorescence using a mixture of primary antibodies against ACE2 and CADM1, which were visualized with green and red fluorescence, respectively. Nuclei are counterstained with DAPI.

Close modal

Expression of CADM1 and ACE2 in Mouse Nasal Mucosa

We conducted immunohistochemistry on the mouse nasal mucosa using antibodies against CADM1 and ACE2. CADM1 proteins were detected strongly in the olfactory epithelium on the cell membrane and were just faintly in the respiratory epithelium but absent from the cilia (Fig. 2a–c). In contrast, ACE2 proteins were easily detectable in the respiratory epithelium mainly on the apical surface of ciliated cells but were not in the olfactory epithelium (Fig. 2d, e), consistent with a previous report [34].

Fig. 2.

Immunohistochemistry of CADM1 and ACE2 on mouse nasal cavity. Formalin-fixed and paraffin-embedded mouse foreheads were cut into thin sections at the mid-sagittal plane and were immunostained with antibodies against CADM1 (a-c) and ACE2 (d, e) (red to brown). Nuclei are counterstained with hematoxylin (blue). b, c Enlarged images of the boxed areas in a. Boxed areas of c–e are enlarged in each inset. * Nasal glands are detached. a Bar = 200 μm. b–e Bar = 50 μm.

Fig. 2.

Immunohistochemistry of CADM1 and ACE2 on mouse nasal cavity. Formalin-fixed and paraffin-embedded mouse foreheads were cut into thin sections at the mid-sagittal plane and were immunostained with antibodies against CADM1 (a-c) and ACE2 (d, e) (red to brown). Nuclei are counterstained with hematoxylin (blue). b, c Enlarged images of the boxed areas in a. Boxed areas of c–e are enlarged in each inset. * Nasal glands are detached. a Bar = 200 μm. b–e Bar = 50 μm.

Close modal

This protein distribution profile was confirmed by Western blot analyses. We picked out the respiratory and olfactory epithelium-lining regions separately from a frozen mouse head halved at the mid-sagittal plane and extract proteins from each region (Fig. 3a). Correct regional sampling was verified by Western blotting with antibodies against neuron-specific enolase (neuronal marker) and E-cadherin (epithelial marker) (Fig. 3b). The respiratory epithelial extract contained more ACE2 proteins than the olfactory epithelial extract, whereas the olfactory epithelial extract contained more CADM1 proteins than the respiratory epithelial extract (Fig. 3b), consistent with the immunohistochemical results (Fig. 2).

Fig. 3.

ACE2 and CADM1 protein expression in the respiratory and olfactory epithelium-lining regions of mouse nasal cavity. a The two regions were delineated on the near mid-sagittal section of a frozen mouse head and proteins extracted from these regions separately. Two mice were analyzed (Nos. 1 and 2). b The extracts from each region were subjected to Western blot analyses using antibodies indicated. The blot was reprobed with an anti-GAPDH antibody to determine the amount of protein loading per lane.

Fig. 3.

ACE2 and CADM1 protein expression in the respiratory and olfactory epithelium-lining regions of mouse nasal cavity. a The two regions were delineated on the near mid-sagittal section of a frozen mouse head and proteins extracted from these regions separately. Two mice were analyzed (Nos. 1 and 2). b The extracts from each region were subjected to Western blot analyses using antibodies indicated. The blot was reprobed with an anti-GAPDH antibody to determine the amount of protein loading per lane.

Close modal

Detection of S1 Protein Bound to CADM1 on MDCK Cells

By cDNA transfection, we obtained MDCK cells that robustly expressed ACE2 (MDCK-ACE2) or CADM1 (MDCK-CADM1). Expression of ACE2 and CADM1 was confirmed by Western blotting (Fig. 1b) and immunofluorescence analyses. ACE2 and CADM1 proteins were clearly detected on cell membranes of transfected MDCK cells and were not in original MDCK cells (Fig. 1c).

We labeled S1 proteins with fluorescein and added them to the confluent monolayer cultures of the three types of MDCK cells. Taking the next planned experiments into consideration, we also added control antibody IgY at the same time. After three-time wash with PBS, we examined the fluorescein fluorescence localization and intensity. The S1 protein fluorescence was detected on the cell surface and periphery in MDCK-ACE2 and MDCK-CADM1 cell cultures (Fig. 1a, upper). The fluorescence intensity of fluorescein was high and low in MDCK-ACE2 and original MDCK cell cultures, respectively (Fig. 1a, upper), consistent with our previous report [25]. In MDCK-CADM1 cell cultures, the fluorescence intensity was intermediate, and significantly higher than that in original MDCK cell cultures (p < 0.001; Fig. 1a, upper).

Inhibition of S1 Protein-CADM1 Binding by 3E1 on MDCK Cells and in Mouse Nasal Mucosa

Since 3E1 is an anti-CADM1 antibody recognizing the ectodomain, this antibody may interfere with the binding between S1 protein and CADM1. We added fluorescein-labeled S1 proteins along with 3E1 to MDCK cell cultures. 3E1 co-addition did not change the fluorescence intensity of fluorescein in MDCK-ACE2 cell cultures (Fig. 1a, lower). On the other hand, 3E1 co-addition to MDCK-CADM1 cell cultures did decrease the intensity to the level as low as that in original MDCK cell + IgY cultures (Fig. 1a).

Then, we labeled S1 proteins with ICG and administered them to the nasal cavity of C57BL/6 mice. After 30 min, ICG signals were detected strongly on the olfactory epithelial surface but were not on the respiratory epithelial surface (Fig. 4a). We extracted proteins from the respiratory and olfactory epithelium-lining regions separately and subjected them to Western blot analyses. S1 proteins were detected in the olfactory region extract but were not in the respiratory region extract (Fig. 4b), consistent with the histological detection of ICG signals (Fig. 4a).

Fig. 4.

Localization of intranasal S1 spike protein and 3E1. a S1 spike protein was labeled with indocyanine green and administered intranasally to mice. After 30 min, a mid-sagittal section of the mouse forehead was observed through a fluorescence microscope. Indocyanine green is displayed in red. Green is autofluorescence. Note that indocyanine green signals are detected exclusively on the olfactory, not respiratory, epithelium. Bar = 500 μm. b Other mice were used for protein extraction. The respiratory and olfactory epithelium-lining regions were delineated on the near mid-sagittal section of a frozen mouse head and proteins extracted from these regions separately. Three mice (Nos. 1, 2, and 3) and one untreated mouse (NT) were analyzed. The extracts from each region were subjected to Western blot analyses using anti-S1 spike protein antibody. In the rightmost lane, recombinant S1 spike protein (75 ng) was loaded as a control. The blot was reprobed with an anti-GAPDH antibody to determine the amount of protein loading per lane. c Either control IgY (left; n = 3) or 3E1 (right; n = 3) was administered intranasally to mice. After 30 min, a mid-sagittal section of the mouse forehead was stained by immunohistofluorescence using the anti-IgY antibody, which was visualized with red fluorescence. Nuclei are counterstained with DAPI. Representative microphotographs are shown. Note that red fluorescent signals are detected exclusively on the olfactory, not respiratory (indicated by arrowheads), epithelium in the right panel. Bar = 200 μm.

Fig. 4.

Localization of intranasal S1 spike protein and 3E1. a S1 spike protein was labeled with indocyanine green and administered intranasally to mice. After 30 min, a mid-sagittal section of the mouse forehead was observed through a fluorescence microscope. Indocyanine green is displayed in red. Green is autofluorescence. Note that indocyanine green signals are detected exclusively on the olfactory, not respiratory, epithelium. Bar = 500 μm. b Other mice were used for protein extraction. The respiratory and olfactory epithelium-lining regions were delineated on the near mid-sagittal section of a frozen mouse head and proteins extracted from these regions separately. Three mice (Nos. 1, 2, and 3) and one untreated mouse (NT) were analyzed. The extracts from each region were subjected to Western blot analyses using anti-S1 spike protein antibody. In the rightmost lane, recombinant S1 spike protein (75 ng) was loaded as a control. The blot was reprobed with an anti-GAPDH antibody to determine the amount of protein loading per lane. c Either control IgY (left; n = 3) or 3E1 (right; n = 3) was administered intranasally to mice. After 30 min, a mid-sagittal section of the mouse forehead was stained by immunohistofluorescence using the anti-IgY antibody, which was visualized with red fluorescence. Nuclei are counterstained with DAPI. Representative microphotographs are shown. Note that red fluorescent signals are detected exclusively on the olfactory, not respiratory (indicated by arrowheads), epithelium in the right panel. Bar = 200 μm.

Close modal

Next, we administered 3E1 or control IgY intranasally to the mouse, and after 30 min, tried to detect the antibody localization by immunohistofluorescence. 3E1 was detected strongly on the olfactory epithelial surface, but rarely on the respiratory epithelial surface (Fig. 4c). Control antibody did not show any specific signals (Fig. 4c).

Finally, we administered either 3E1 or control IgY (10 μg each) to the mouse nasal cavity, and after 1 h, administered S1 protein (10 μg) intranasally. After 3 h, we tried to detect S1 protein in the nasal cavity by immunohistofluorescence. In the mice pre-administered with control IgY, S1 protein was detected on the olfactory epithelial surface (Fig. 5). In contrast, when the mice were pre-administered with 3E1, S1 protein was rarely detected on the olfactory epithelial surface but was detected as a protein aggregatefloating in the nasal cavity (Fig. 5).

Fig. 5.

Localization of intranasal S1 spike protein in mice pre-administered with 3E1. Either control IgY (left; n = 2) or 3E1 (right; n = 4) (10 μm each) and S1 spike protein (10 μg) were administered intranasally to mice in this order with an interval of 1 h. After 3 h, frozen mid-sagittal sections of the mouse forehead were stained by immunohistofluorescence using a mixture of antibodies against IgY and S1 spike protein, which were visualized with green and red fluorescence, respectively. Nuclei are counterstained with Hoechst. Representative microphotographs are shown. Arrowheads indicate S1 spike protein localization on the olfactory epithelial surface. Dotted lines circle red fluorescent aggregates corresponding to S1 spike proteinsfloating in the nasal cavity. Bar = 100 μm.

Fig. 5.

Localization of intranasal S1 spike protein in mice pre-administered with 3E1. Either control IgY (left; n = 2) or 3E1 (right; n = 4) (10 μm each) and S1 spike protein (10 μg) were administered intranasally to mice in this order with an interval of 1 h. After 3 h, frozen mid-sagittal sections of the mouse forehead were stained by immunohistofluorescence using a mixture of antibodies against IgY and S1 spike protein, which were visualized with green and red fluorescence, respectively. Nuclei are counterstained with Hoechst. Representative microphotographs are shown. Arrowheads indicate S1 spike protein localization on the olfactory epithelial surface. Dotted lines circle red fluorescent aggregates corresponding to S1 spike proteinsfloating in the nasal cavity. Bar = 100 μm.

Close modal

Docking Simulation of S1 Protein-CADM1 Binding

We conducted docking simulation analyses to examine how S1 protein binds to CADM1 cis-homodimer or ACE2. The highest ZDOCK score of S1 protein-CADM1 binding was three-quarters of that of S1 protein-ACE2 binding (see Y-intercepts of Fig. 6a). Notably, in the other docking configurations, the ZDOCK score of S1 protein-CADM1 binding was constantly higher than that of S1 protein-ACE2 binding (Fig. 6a). In addition, when S1 protein-CADM1 binding had the highest ZDOCK score, the interface of S1 protein was simulated to be fairly similar to the receptor-binding domain, i.e., the interface that S1 protein uses when S1 protein-ACE2 binding has the highest ZDOCK score (Fig. 6b, c). The highest ZDOCK score of transthyretin cis-homodimerization is 742.3 (Fig. 6a).

Fig. 6.

Molecular simulation of S1 spike protein binding to ACE2 and CADM1. a ZDOCK binding configurations were numbered in descending order of the ZDOCK scores, arranged in numerical sequence on the X-axis, and plotted with ZDOCK scores of S1–ACE2 (black line) and S1–CADM1 (red line) binding on the Y-axis. The larger the score is, the more stable the configuration is. The highest scores of S1–ACE2 and S1–CADM1 bindings are shown as Y-intercepts (x = 1) and indicated by black and red arrowheads, respectively. A red arrow indicates the highest score of cis-homodimerization of transthyretin. b Binding configuration of S1 spike protein and CADM1 cis-homodimer simulated when it has the highest ZDOCK score. c, d Receptor-binding domain of S1 spike protein (the interface used to bind ACE2 with the highest ZDOCK score) is colored pink in the 3D structure of the SARS-CoV-2 spike protein trimer (c). The interface of S1 spike protein used to bind CADM1 with the highest ZDOCK score is colored red (d).

Fig. 6.

Molecular simulation of S1 spike protein binding to ACE2 and CADM1. a ZDOCK binding configurations were numbered in descending order of the ZDOCK scores, arranged in numerical sequence on the X-axis, and plotted with ZDOCK scores of S1–ACE2 (black line) and S1–CADM1 (red line) binding on the Y-axis. The larger the score is, the more stable the configuration is. The highest scores of S1–ACE2 and S1–CADM1 bindings are shown as Y-intercepts (x = 1) and indicated by black and red arrowheads, respectively. A red arrow indicates the highest score of cis-homodimerization of transthyretin. b Binding configuration of S1 spike protein and CADM1 cis-homodimer simulated when it has the highest ZDOCK score. c, d Receptor-binding domain of S1 spike protein (the interface used to bind ACE2 with the highest ZDOCK score) is colored pink in the 3D structure of the SARS-CoV-2 spike protein trimer (c). The interface of S1 spike protein used to bind CADM1 with the highest ZDOCK score is colored red (d).

Close modal

In the present study, we found that in the mouse nasal cavity, ACE2 protein expression was higher in the respiratory epithelium than in the olfactory epithelium, while CADM1 protein expression was much higher in the olfactory epithelium than in the respiratory epithelium. Interestingly, when S1 protein was administered intranasally, its distribution pattern was in concurrence with that of CADM1, not of ACE2, suggesting that CADM1 might contribute to the binding of S1 protein to the olfactory epithelium. This speculation was supported by the in vitro experiments showing that S1 protein bound to CADM1 expressed on MDCK cells.

According to the data obtained by the cell culture experiments and computer simulation, the maximal strength of S1 protein-CADM1 binding was estimated to be half to three-fourths as strong as that of S1 protein-ACE2 binding. Though CADM1 appeared to have a substantially lower affinity to S1 protein than ACE2, we speculate that CADM1 significantly contributed to S1 protein-olfactory epithelium binding. There are two reasons. One is the abundance of CADM1 molecules on the olfactory epithelium, as suggested by immunohistochemical and Western blot analyses (Fig. 2, 3). Another is that the ZDOCK score of S1 protein-CADM1 binding is considered to be high enough for stable protein-protein binding in the plasma because the highest ZDOCK score of cis-homodimerization of transthyretin was 742.3 (Fig. 6a) and this dimer is stable in the plasma [33]. Therefore, Figure 6a suggests that CADM1 has multiple interfaces that enable S1 protein to bind CADM1 with significant affinity.

The olfactory epithelium is composed of sustentacular support cells and olfactory neurons. Both types of cells express CADM1 [24]. Olfactory epithelial CADM1 is primarily localized on the cell membrane at cell-cell contacts and is suggested to form homophilic binding in trans between neighboring cells, as is often the case for various types of epithelial cells expressing CADM1 [35, 36]. On closer look, CADM1 appeared not to be present on the cilia of olfactory neurons but to be present at the apical surface of sustentacular support cells (Fig. 2c), consistent with a past report [24]. These CADM1 molecules are likely to be the actual effector to mediate S1 protein-olfactory epithelium binding.

S1 protein-CADM1 and S1 protein-olfactory epithelium bindings appeared to be interfered with by 3E1, the anti-CADM1 ectodomain antibody that we developed originally [37]. Interestingly, when 3E1 was administered intranasally, it was localized exclusively on the olfactory epithelium, in concurrence with the intranasal distribution of S1 protein. CADM1 is highly homologous between mouse and human [38], and 3E1 has equivalent affinities for murine and human CADM1 [39]. Recently, we obtained a humanized 3E1 clone that had an affinity to CADM1 as strongly as the original 3E1 [25]. We are now developing a nasal spray containing the humanized antibody and examining whether it can be useful for COVID-19 prevention.

CADM1 is a member of the nectin-like molecule (Necl) family and is alternatively called Necl-2 [19, 40]. Another member Necl-5 has long been known to be the poliovirus receptor [41] and is recently revealed to be one of the potential co-receptors of SARS-CoV-2 [42]. Since Necl family members are structurally similar to each other [38], these molecules including CADM1 may be commonly involved in the initial step of virus-host interaction.

In conclusion, S1 protein appeared to prefer the olfactory epithelium rather than the respiratory epithelium as its binding site in the mouse nasal cavity, and CADM1 was suggested to contribute to this preference based on the molecular abundance and affinity. In addition, 3E1 appeared to inhibit this contribution of CADM1. Humanized 3E1 antibodies may be promising as preventive agents for COVID-19.

The authors thank Nobuyuki Mizuguchi (Kindai University Life Science Research Institute, Osaka, Japan) for his technical assistance.

This study protocol was reviewed and approved by Institutional Animal Experimentation Committee of Kindai University Faculty of Medicine, approval number [KDMS-2021-005].

The authors have no conflicts of interest to declare.

This study was supported by the Japan Society for the Promotion of Science Kakenhi (20K07434 to M.H., 22K07034 to A.Y., and 21K06978 to A.I.), the All-Kindai University support project against COVID-19 (to A.I., 2020 and 2021), and Advanced Research and Development Programs for Medical Innovation (A-178, A.I.).

M.H. and A.W. constructed expression vectors and performed transfection. F.T., M.H., and T.I. conducted cell culture experiments, confocal microscopic studies, and Western blot analyses. F.T. and A.Y. performed immunofluorescence and histological analyses. A.S. and Y.T. conducted simulation analyses. M.H. and A.I. conducted the statistical analyses. F.T., M.H., and A.I. confirmed the authenticity of all the raw data. F.T. and A.I. conceived and designed the study, and A.I. drafted the manuscript. All authors read and approved the final manuscript.

Additional Information

Fuka Takeuchi and Aki Sugano contributed equally to this work.

Data are not publicly available due to ethical reasons. Further inquiries can be directed to the corresponding author.

1.
Li W, Moore MJ, Vasilieva N, Sui J, Wong SK, Berne MA, et al. Angiotensin-converting enzyme 2 is a functional receptor for the SARS coronavirus. Nature. 2003;426(6965):450–4.
2.
Wang Q, Zhang Y, Wu L, Niu S, Song C, Zhang Z, et al. Structural and functional basis of SARS-CoV-2 entry by using human ACE2. Cell. 2020;181(4):894–904 e9.
3.
Hoffmann M, Kleine-Weber H, Schroeder S, Krüger N, Herrler T, Erichsen S, et al. SARS-CoV-2 cell entry depends on ACE2 and TMPRSS2 and is blocked by a clinically proven protease inhibitor. Cell. 2020;181(2):271–80.e8.
4.
Wrapp D, Wang N, Corbett KS, Goldsmith JA, Hsieh CL, Abiona O, et al. Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation. Science. 2020;367(6483):1260–3.
5.
Meng X, Deng Y, Dai Z, Meng Z. COVID-19 and anosmia: a review based on up-to-date knowledge. Am J Otolaryngol. 2020;41(5):102581.
6.
Butowt R, Bilinska K, von Bartheld CS. Olfactory dysfunction in COVID-19: new insights into the underlying mechanisms. Trends Neurosci. 2023;46(1):75–90.
7.
Tsukahara T, Brann DH, Datta SR. Mechanisms of SARS-CoV-2 associated anosmia. Physiol Rev. 2023;103(4):2759–66.
8.
Harkema JR, Carey SA, Wagner JG. The nose revisited: a brief review of the comparative structure, function, and toxicologic pathology of the nasal epithelium. Toxicol Pathol. 2006;34(3):252–69.
9.
Uraih LC, Maronpot RR. Normal histology of the nasal cavity and application of special techniques. Environ Health Perspect. 1990;85:187–208.
10.
Graziadei PP. The ultrastructure of vertebrate olfactory mucosa. In: Friedmann I, editor. The ultrastructure of sensory organs, 1st ed. 1973. p. 267–305.
11.
Lochhead JJ, Thorne RG. Intranasal delivery of biologics to the central nervous system. Adv Drug Deliv Rev. 2012;64(7):614–28.
12.
Dahl R, Mygind N. Anatomy, physiology and function of the nasal cavities in health and disease. Adv Drug Deliv Rev. 1998;29(1–2):3–12.
13.
Hikmet F, Méar L, Edvinsson A, Micke P, Uhlén M, Lindskog C. The protein expression profile of ACE2 in human tissues. Mol Syst Biol. 2020;16(7):e9610.
14.
Hou YJ, Okuda K, Edwards CE, Martinez DR, Asakura T, Dinnon KH3rd, et al. SARS-CoV-2 reverse genetics reveals a variable infection gradient in the respiratory tract. Cell. 2020;182(2):429–46.e14.
15.
Cantuti-Castelvetri L, Ojha R, Pedro LD, Djannatian M, Franz J, Kuivanen S, et al. Neuropilin-1 facilitates SARS-CoV-2 cell entry and infectivity. Science. 2020;370(6518):856–60.
16.
Loo L, Waller MA, Moreno CL, Cole AJ, Stella AO, Pop OT, et al. Fibroblast-expressed LRRC15 is a receptor for SARS-CoV-2 spike and controls antiviral and antifibrotic transcriptional programs. PLoS Biol. 2023;21(2):e3001967.
17.
Lempp FA, Soriaga LB, Montiel-Ruiz M, Benigni F, Noack J, Park YJ, et al. Lectins enhance SARS-CoV-2 infection and influence neutralizing antibodies. Nature. 2021;598(7880):342–7.
18.
Watabe K, Ito A, Koma YI, Kitamura Y. IGSF4: a new intercellular adhesion molecule that is called by three names, TSLC1, SgIGSF and SynCAM, by virtue of its diverse function. Histol Histopathol. 2003;18(4):1321–9.
19.
Zhang S, Lu G, Qi J, Li Y, Zhang Z, Zhang B, et al. Competition of cell adhesion and immune recognition: insights into the interaction between CRTAM and nectin-like 2. Structure. 2013;21(8):1430–9.
20.
Ito A, Okada M, Uchino K, Wakayama T, Koma Y, Iseki S, et al. Expression of the TSLC1 adhesion molecule in pulmonary epithelium and its down-regulation in pulmonary adenocarcinoma other than bronchioloalveolar carcinoma. Lab Invest. 2003;83(8):1175–83.
21.
Ito A, Nishikawa Y, Ohnuma K, Ohnuma I, Koma Y, Sato A, et al. SgIGSF is a novel biliary-epithelial cell adhesion molecule mediating duct/ductule development. Hepatology. 2007;45(3):684–94.
22.
Kimura R, Otani T, Shiraishi N, Hagiyama M, Yoneshige A, Wada A, et al. Expression of cell adhesion molecule 1 in human and murine endometrial glandular cells and its increase during the proliferative phase by estrogen and cell density. Life Sci. 2021;283:119854.
23.
Fujita E, Urase K, Soyama A, Kouroku Y, Momoi T. Distribution of RA175/TSLC1/SynCAM, a member of the immunoglobulin superfamily, in the developing nervous system. Brain Res Dev Brain Res. 2005;154(2):199–209.
24.
Tsukioka F, Wakayama T, Tsukatani T, Miwa T, Furukawa M, Iseki S. Expression and localization of the cell adhesion molecule SgIGSF during regeneration of the olfactory epithelium in mice. Acta Histochem Cytochem. 2007;40(2):43–52.
25.
Hagiyama M, Takeuchi F, Sugano A, Yoneshige A, Inoue T, Wada A, et al. Indigo plant leaf extract inhibits the binding of SARS-CoV-2 spike protein to angiotensin-converting enzyme 2. Exp Ther Med. 2022;23(4):274.
26.
Hosokawa Y, Hagiyama M, Iino T, Murakami Y, Ito A. Noncontact estimation of intercellular breaking force using a femtosecond laser impulse quantified by atomic force microscopy. Proc Natl Acad Sci USA. 2011;108(5):1777–82.
27.
Wu X, Azizan EAB, Goodchild E, Garg S, Hagiyama M, Cabrera CP, et al. Somatic mutations of CADM1 in aldosterone-producing adenomas and gap junction-dependent regulation of aldosterone production. Nat Genet. 2023;55(6):1009–21.
28.
Inoue T, Hagiyama M, Enoki E, Sakurai MA, Tan A, Wakayama T, et al. Cell adhesion molecule 1 is a new osteoblastic cell adhesion molecule and a diagnostic marker for osteosarcoma. Life Sci. 2013;92(1):91–9.
29.
Kawamoto T, Shimizu M. A method for preparing 2- to 50-micron-thick fresh-frozen sections of large samples and undecalcified hard tissues. Histochem Cell Biol. 2000;113(5):331–9.
30.
Mimae T, Okada M, Hagiyama M, Miyata Y, Tsutani Y, Inoue T, et al. Upregulation of notch2 and six1 is associated with progression of early-stage lung adenocarcinoma and a more aggressive phenotype at advanced stages. Clin Cancer Res. 2012;18(4):945–55.
31.
Takaoka Y, Sugano A, Morinaga Y, Ohta M, Miura K, Kataguchi H, et al. Prediction of infectivity of SARS-CoV2: mathematical model with analysis of docking simulation for spike proteins and angiotensin-converting enzyme 2. Microb Risk Anal. 2022;22:100227.
32.
Chen R, Li L, Weng Z. ZDOCK: an initial-stage protein-docking algorithm. Proteins. 2003;52(1):80–7.
33.
Hörnberg A, Eneqvist T, Olofsson A, Lundgren E, Sauer-Eriksson AE. A comparative analysis of 23 structures of the amyloidogenic protein transthyretin. J Mol Biol. 2000;302(3):649–69.
34.
Ahn JH, Kim J, Hong SP, Choi SY, Yang MJ, Ju YS, et al. Nasal ciliated cells are primary targets for SARS-CoV-2 replication in the early stage of COVID-19. J Clin Invest. 2021;131(13):e148517.
35.
Sakurai-Yageta M, Masuda M, Tsuboi Y, Ito A, Murakami Y. Tumor suppressor CADM1 is involved in epithelial cell structure. Biochem Biophys Res Commun. 2009;390(3):977–82.
36.
Hagiyama M, Kimura R, Yoneshige A, Inoue T, Otani T, Ito A. Cell adhesion molecule 1 contributes to cell survival in crowded epithelial monolayers. Int J Mol Sci. 2020;21(11):4123.
37.
Furuno T, Ito A, Koma Y, Watabe K, Yokozaki H, Bienenstock J, et al. The spermatogenic Ig superfamily/synaptic cell adhesion molecule mast-cell adhesion molecule promotes interaction with nerves. J Immunol. 2005;174(11):6934–42.
38.
Biederer T. Bioinformatic characterization of the SynCAM family of immunoglobulin-like domain-containing adhesion molecules. Genomics. 2006;87(1):139–50.
39.
Koma YI, Furuno T, Hagiyama M, Hamaguchi K, Nakanishi M, Masuda M, et al. Cell adhesion molecule 1 is a novel pancreatic-islet cell adhesion molecule that mediates nerve-islet cell interactions. Gastroenterology. 2008;134(5):1544–54.
40.
Takai Y, Miyoshi J, Ikeda W, Ogita H. Nectins and nectin-like molecules: roles in contact inhibition of cell movement and proliferation. Nat Rev Mol Cell Biol. 2008;9(8):603–15.
41.
Ren RB, Costantini F, Gorgacz EJ, Lee JJ, Racaniello VR. Transgenic mice expressing a human poliovirus receptor: a new model for poliomyelitis. Cell. 1990;63(2):353–62.
42.
Achom A, Das R, Pakray P. An improved Fuzzy based GWO algorithm for predicting the potential host receptor of COVID-19 infection. Comput Biol Med. 2022;151(Pt A):106050.