Evolution of nasal and olfactory infection characteristics of SARS-CoV-2 variants

Abstract

SARS-CoV-2 infection of the upper airway and the subsequent immune response are early, critical factors in COVID-19 pathogenesis. By studying infection of human biopsies in vitro and in a hamster model in vivo, we demonstrated a transition in tropism from olfactory to respiratory epithelium as the virus evolved. Analyzing each variants revealed that SARS-CoV-2 WA1 or Delta infects a proportion of olfactory neurons in addition to the primary target sustentacular cells. The Delta variant possesses broader cellular invasion capacity into the submucosa, while Omicron displays longer retention in the sinonasal epithelium. The olfactory neuronal infection by WA1 and the subsequent olfactory bulb transport via axon is more pronounced in younger hosts. In addition, the observed viral clearance delay and phagocytic dysfunction in aged olfactory mucosa is accompanied by a decline of phagocytosis related genes. Furthermore, robust basal stem cell activation contributes to neuroepithelial regeneration and restores ACE2 expression post-infection. Together, our study characterized the nasal tropism of SARS-CoV-2 strains, immune clearance, and regeneration post infection. The shifting characteristics of viral infection at the airway portal provides insight into the variability of COVID-19 clinical features and may suggest differing strategies for early local intervention.

Extended Data Fig. 1. Detection of SARS-Cov-2 in the olfactory neuroepithelium.

a, SARS-Cov-2 antibody testing. 1 anti-spike S and 3 different anti-NP were verified to be reliable for frozen section immunohistochemistry. Hamster olfactory tissue was examined at 4dpi. All 4 antibodies stained in the same pattern showing intensive viral load mainly located in the apical sustentacular cell layer. No signal could be detected in mock control. Catalog number for each antibody is presented accordingly. b, RNAscope analysis showing SARS-Cov-2 viral RNA on 4dpi in hamster olfactory epithelium. c,d, Co-staining of NP and sustentacular cell marker Krt18. Image was captured from the boxed area in panel (b) of Figure 1. e, Confocal image of NP and Tuj1 staining in mock control. Scale bars, 20 μm.

Extended Data Fig. 2. Decreased Omicron variant infection in hamster olfactory epithelium.

a, Co-staining of neuronal marker Tuj1 and respiratory epithelium marker Foxj1 in mouse nasal cavity. Scale bar, 200 μm. b, Representative image shows OMP and rabbit anti ACE2 co-staining in hamster turbinate horizontal section. Intense ACE2 expression is seen in OMP⁺ olfactory epithelium. The green arrows show the respiratory-olfactory transition area with lower ACE2 expression. Scale bar, 100 μm. c, Confocal images show the distribution of NP and Tuj1 in a coronal section at L1 of the nasal cavity. Tissues were examined on 4dpi, boxed areas were highlighted at bottom. Note that NP was dramatically declined from Tuj1 negative respiratory epithelium (RE) in hamsters infected with WA1 or Delta. The respiratory infection in Omicron group was markedly increased. Scale bars = 500 μm.

Extended Data Fig. 3. Tropism of SARS-CoV-2 variants in the posterior nasal mucosa.

a, Confocal images showing the distribution of NP and Tuj1 in posterior nasal cavity sections at 4dpi. Coronal sections at L3 were examined, where the proportion of olfactory epithelium is predominant. The olfactory epithelium infection in the Omicron group was decreased remarkably.Scale bar = 500 μm. b, Co-staining of NP and Iba1 (macrophage marker) or NP and Vimentin (mesenchymal cell and olfactory ensheathing cell marker) in Delta infected hamster. Scale bars = 20 μm. The white dotted line in (b) indicates the basement membrane.

Extended Data Fig. 4. SARS-Cov-2 WA1 and Delta variants infect olfactory sensory neurons.

a, Co-staining of NP and OMP in olfactory epithelium. 5 weeks-old hamsters were infected with SARS-CoV-2-Delta variant (1 × 10⁷ TCID50) and were examined on 4dpi. Arrowhead indicates an infected neuron. b,c, Confocal images show co-localization of NP with Tuj1⁺ or OMP⁺ axon. (b) shows a larger view of Figure 2f. Boxed areas in (b) were highlighted at bottom. 1m (b) or 7–8 weeks-old (c) hamsters were infected with WA1. d, Representative image shows NP signal does not colocalize with Vimentin in axon bundles. e,f, Co-staining of NP and OMP in olfactory epithelium. g, Confocal image shows Tuj1⁺ immature olfactory neurons in mock group. 7–8 week-old hamsters were infected WA1 (d, e) or Delta variant (f) at 1 × 10⁵ TCID50 and were examined at 4dpi. Scale bars, 20 μm.

Extended Data Fig. 5. Expression of Nrp1 in mouse olfactory epithelium and bulb

a, qPCR analysis of Nrp1 expression in mouse olfactory mucosa at ages 2 weeks, 2 months, and 19 months. Each data point represents an individual mouse (n=3). b,c, Immunostaining analysis shows the expression of Nrp1 in the Tuj1⁺ immature olfactory neurons and axon bundles. Confocal images were acquired from horizontal section of young (1 month) and old (8 month) mice. Boxed areas were highlighted on right. In young mice, a few mature neurons above the Tuj1⁺ cells also express a low level of Nrp1. d, Percentage of Nrp1 expressing cells in Tuj1⁺ immature neurons. Olfactory tissues from wildtype mice were examined at the indicated age groups. e, Confocal images show the expression of Nrp1 in young and old mouse olfactory bulb. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. Arrowheads highlight Nrp1⁺ cells in glomerular layer. Scale bars, 20 μm (b,c); 50 μm (e).

Extended Data Fig. 6. Increased brain transport of SARS-CoV-2 in young hamsters

a-f, Confocal image capturing a cross section of olfactory epithelium and olfactory bulb. SARS-CoV-2-infected young or old hamsters were examined at 6 dpi. Boxed areas highlight the infected lateral olfactory axons crossing the cribriform plate and projecting to the olfactory bulb.Images were captured with 4 μm Z-stack and exported by maximum intensity projections. OE, olfactory epithelium; OB, olfactory bulb. g, RNAscope analysis shows viral RNA in SARS-Cov-2 infected hamster OB glomeruli at 4dpi. h, Co-staining of Caspase-3 and Iba1 in olfactory bulb at 4dpi. h,i, Confocal image shows co-staining of endothelial cell marker CD31 and ACE2 in mouse (h) or hamster (i, ACE2 only) olfactory bulb. j, Immunostaining of CD45 and microglia marker Iba1 in the olfactory bulb of hACE2 mice at 6 dpi. Arrowheads indicate Iba1 negative immune cells. In the hACE2 strain, human ACE2 overexpression was driven by mouse Krt18 promoter. Scale bars, 100 μm (a,d); 50 μm (h-j).

Extended Data Fig. 7. Regeneration of the olfactory epithelium

a, qPCR analysis of Ifng expression in turbinate tissues. SARS-CoV-2 infected hamsters were examined at indicated time points. b, Dynamic of Iba1⁺ macrophages infiltration and CXCL10 expression in hamster olfactory epithelium. c, Representative images show Krt5⁺ basal cells expressing proliferation marker Ki67 on 4dpi in hamster olfactory epithelium. d, qPCR analysis of Sox2, Lgr5, and Tubb3 expression in turbinate samples at indicated time points. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. Each data point represents an individual mouse. Scale bars, 20 μm.

Fig. 1.. SARS-Cov-2 WA1 selectively targets human olfactory neuroepithelium

a,b, Confocal images of SARS-CoV-2 viral antigen NP (red, Novus, NB100–56576) and olfactory neuronal marker β-III Tubulin (Tuj1, green) in superior turbinate biopsies from 2 separate patients. Images were obtained under tile scan mode, which covered olfactory and adjacent respiratory epithelium in the same piece of tissue. Boxed area in (b) was highlighted in Extended Data Fig. 1c, d. c, Co-staining of NP and Tuj1 in human biopsy collected from the olfactory cleft. d, Representative image of NP overlapped with Tuj1-negative ciliated cell (brightfield). Confocal image was obtained from a biopsy which contains only respiratory epithelium. e, Quantification of NP⁺ cells per mm tissue. 24 independent specimens have exclusively respiratory epithelium (RE), while 7 specimens contained both respiratory and olfactory epithelium (OE). Arrowheads (a-c) indicate the detachment of infected cells. Data in (e) are represented as mean ± S.D. p value was calculated by one-way ANOVA. Scale bars, 50 μm (a and b); 20 μm (c,d).

Fig. 2.. Omicron variant shows tropism transition from olfactory to respiratory epithelium

a, Scheme of the tissue section. To avoid variability across different animals, frozen sections were collected and examined at three consistent levels (L1–3) representing the anterior (mainly respiratory epithelium), middle (respiratory + Olfactory epithelium), and posterior (mainly olfactory epithelium). b, Confocal images of NP and Tuj1-labeled hamster nasal sections at L2. WA1, Delta, and Omicron infected hamsters were examined on 4 dpi. Boxed areas are highlighted at bottom. Scale bars = 500 μm. c, Percentage of the infected olfactory epithelium. The total length of Tuj1⁺ or NP⁺/Tuj1⁺ epithelium in each section at L1–3 were quantified using Image J. d, Quantification of NP⁺ cells in nasal respiratory epithelium. The total NP⁺ cells in Tuj1⁻ respiratory epithelium including paranasal sinuses of each section were counted. e, qPCR analysis of ACE2 expression in mouse nasal respiratory or olfactory epithelium at age of 2 weeks, 2months, and 19months. The entire nasal respiratory or olfactory epithelium from the same animal were isolated separately. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. Each data point represents an individual animal.

Fig. 3.. Delta variant infects cells in submucosa of the nose

a, Representative image shows NP⁺/Pan-Keratin⁺ Bowman’s glands in Delta treated hamsters. b, Quantification of infected Bowman’s glands. The average number of NP⁺ Bowman’s glands in one 14 μm section was calculated. 3 sections per animal were counted. c, Confocal image shows NP⁺/α-SMA⁺ myofibroblasts. Hamsters infected with Delta variant on 4dpi were examined. d, Co-staining of Tuj1 and NP in nasal sections at 7dpi. Whole nasal cavity images were captured using a tile scan and z stack mode on a 14 μm section. Boxed area in Omicron infected hamster is highlighted on the right. Scale bars = 500 μm. e, Quantification of NP⁺ respiratory epithelial cells in paranasal sinuses. 3 sections per animal was counted. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. Each data point represents an individual animal.

Fig. 4.. Age associated SARS-Cov-2 infection in olfactory sensory neurons

a-c, Confocal image showing WA1-infected hamster olfactory epithelium at 4dpi. Insert in (a) highlighting an NP stained axon bundle (horizontal section). Arrowheads indicate virus infected Tuj1⁺ immature (b) or OMP⁺ mature (c) sensory neurons (coronal sections). White line indicated the basal layer of epithelium. d, e, NP⁺ axon travel from neuroepithelium to laminar propria and merge into Tuj1⁺ axon bundle. f-h, Quantification of NP⁺ axons in young and old hamsters at 6dpi. Representative images show horizontal (f) or coronal sections (g). NP⁺ axons were quantified per μm of the diameter of axon bundle. i,j, Representative images showing NP located in Tuj1⁺ human olfactory neurons (j) and the percentage of NP⁺ cells in Tuj1⁺ population (i). Dotted line in (j) indicates virus infected NP⁺ axon. Arrowheads denote NP⁺/Tuj1⁺ neurons compared to uninfected cells (empty arrowhead). Infected biopsies from 3 young donors (age 25–33 years) and 5 biopsies from older donors (age 54–72 years) were quantified for Tuj1⁺ neuronal infection. k,l, Representative images of Nrp1 expression in human olfactory epithelium (k) and quantification of Nrp1⁺ cells in Tuj1⁺ population (l). 3 biopsies from young (age 20–30 years) and 4 biopsies from older donors (age 68–79 years) were examined for Nrp1 expression. Images in (f) were captured with 3 μm Z-stack and exported by maximum intensity projections. Each data point represents an individual sample from hamster (h), or human (i and l). Details of human biopsies can be found in Supplementary Table 1. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. Scale bars, 20 μm.

Fig. 5.. Increased olfactory bulb transport of SARS-CoV-2 in young hamsters

a-c, Confocal images of Iba1 and NP co-staining in hamster olfactory bulbs. Arrowheads indicate infected axon. d, Co-staining of NP and Tuj1 in a serial section next to panel (b). e, Quantification of Iba1⁺microglials in hamster olfactory bulb. Each data point in (d) represents an individual hamster sample. Data are represented as mean ± S.D. Statistical significance was determined by unpaired two-tailed t-test. f, Confocal image of cleaved caspase-3⁺/NeuN⁻ apoptotic cells (arrowheads) in the glomerular layer at 4dpi. Images were captured with 3 μm (a-d) or 4 μm (f) Z-stack and exported by maximum intensity projections. Olfactory bulb tissues were collected from young and old hamsters on 6dpi (a-d) or from mock control. Scale bars, 50 μm, (a-d); 20 μm, (f). ONL, olfactory nerve layer; GL, glomerular layer; EPL, external plexiform layer; MCL, mitral cell layer. Boxed areas are highlighted at bottom. Dotted circles indicate glomeruli.

Fig. 6.. Age-associated delay in viral clearance in olfactory mucosa

a, Representative images showing CD45 and Iba1 co-staining in olfactory mucosa. Mock or maSARS-Cov2 infected wildtype mice were examined at 6dpi. b, Co-immunostaining shows Iba1⁺ macrophages engulfing NP⁺ debris in hamster olfactory mucosa at 4dpi. c, Representative image of Iba1 and cleaved caspase-3 co-staining in hamster at 4dpi. Arrowheads highlight the Iba1⁺ macrophages undergoing apoptosis. d, Representative images showing Iba1 or NP staining in serial sections. Each panel combines 6 40x images acquired under tile scan mode. Young or old hamsters’ olfactory tissues were examined at 6dpi. e,f, Quantification of Iba1⁺ (e) or NP⁺ (f) cells in hamster nasal olfactory lumen at 6dpi. Serial sections (d) from 4 different levels were quantified. g, Violin plots showing the differentially expressed Clec4n (Dectin2) or Fpr2 in young and old macrophage/dendritic lineage. h, Confocal images of Iba1 and Dectin2 co-staining in mouse olfactory mucosa. Each data point represents an individual hamster sample. Statistical significance was determined by unpaired two-tailed t-test. The white dotted line in (a-c) indicates the basement membrane. Scale bars, 20 μm (a-c, h); 50 μm (d).

Fig. 7.. Regeneration of olfactory epithelium and re-expression of ACE2

a, Confocal images showing ACE2 (red) and Krt5⁺ horizontal basal cells (green) in olfactory epithelium of mock or SARS-CoV-2 infected hamster at 4 dpi. b,c, qPCR analysis of ACE2 (b) or OMP (c) mRNA expression in SARS-CoV-2 infected hamster turbinate lysate at indicated time points. d,e, Representative images of Krt5⁺ cells in newly regenerated olfactory epithelium (d) on 6dpi, and quantification of epithelium thickness (e). The thickness of septal olfactory epithelium was measured using Zen lite “line” function. For each section, 8 spots were measured randomly. f, Confocal image showing regenerated hamster olfactory epithelium expression of ACE2 at 28dpi. g, Representative image shows ACE2 and Tuj1⁺olfactory neurons in an olfactory biopsy from a COVID-19 patient on day 12 post diagnosis. Dots in graph represent independent animal. Data are represented as mean ± S.D. p value was calculated by unpaired two-tailed Student’s t test. Scale bars, 20 μm.