Generation of the humanized ACE2 lines

The mouse and human Ace2/ACE2 genes are located on the X chromosome in the region neighboring the pseudoautosomal region [32]. The mouse and human genes are remarkably similar, both in their structure and in the protein domains they encode (Fig 1A and 1B). The ectodomain of both proteins can be cleaved by ADAM17 to generate soluble ACE2 (sACE2) [8]. Differences have been noted in the regulatory elements that control expression of the mouse and human genes, with the most notable being the presence of an intragenic promoter upstream of a primate-specific exon within intron 8 of the human gene. Initiation of transcription from this internal promoter gives rise to a transcript encoding a truncated isoform of ACE2, designated dACE2 or short ACE2 [33]. The encoded protein lacks both the enzymatic activity of the full-length protein and its ability to bind SARS-Cov-2. Both human and mouse ACE2 can catalyze the formation of Ang1-7 from Ang II [34], although some differences have been noted in their enzymatic activity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Schematic showing the protein domains and exon intron organization of the mouse and human ACE2 locus and the structure of the loci in the two humanized mouse lines.

A. Protein structural and functional domains of the ACE2 protein. The ACE2 protein consists of a signal peptide (SP), an extracellular domain, a transmembrane domain (TM) and a cytoplasmic domain (CP). B. Exonic structure of human ACE2 gene (red) and flanking genes (pink) and mouse Ace2 gene (blue) and flanking genes (light blue). Regions of the human protein identified as interacting with the spike proteins of SARS-Co-V and SARS-CoV-2 are indicated by pink boxes above the human gene. Mouse ACE2 does not bind the spike protein in these two coronaviruses. The cleavage site for ADAM17 and TMPRSS11D/TMPRSS2 are shown. Tall boxes represent coding exons while shorter boxes represent 5’ and 3’ UTRs. In both species the start codon is located in the second exon, labeled here as 1b. Both mouse and human also express a transcript (not shown) which, although it lacks exon 1a, still gives rise to an ACE2 protein with the identical amino acid sequence. The distal promoter region (DP) and proximal promoter region (PP) are present in both mouse and human. The human gene contains a third transcriptional start site between exons 8 and 9, which gives rise to a shorter transcript encoding a protein expected to lack enzymatic function and the ability to bind SARS-CoV-2. Four interferon response elements (ISREs) are located in the promoter region upstream of the sequence that encodes the first exon of the shorter transcript (exon 1c). Exon 1c is only present in primates. C. Structure of humanized Ace2 locus. The locus in which only the coding sequence is humanized, MP-ACE2, is shown above. In this locus the human sequence extends from the start codon to approximately 4 kb downstream of the 3’ end of the final exon. In the locus in which humanization includes the promoter, hACE2, the human sequence extends from approximately 3.5 kb upstream of the start codon to approximately 4 kb downstream of the 3’ end of the final exon. The orange arrow in each humanized locus represents the floxed neomycin resistance marker, which is removed by transient expression of Cre.

https://doi.org/10.1371/journal.ppat.1011168.g001

We used syntenic replacement of the endogenous mouse Ace2 gene to generate two mouse lines that are expected to express human ACE2 (Fig 1C). The two lines differ only in the location of the 5’ crossover event, with MP-ACE2 leaving the mouse promoter and 5’ UTR in place, while hACE2 replaces the entire Ace2 gene, including 5’ regulatory regions, with a 51 kb syntenic segment of the human locus. This includes 3.5 kb of sequence upstream of the start codon. An ideogram of the structure of the ACE2 locus present in the mouse lines included in this study along with nomenclature used in referring to them is shown in Fig 2A.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Comparison of ACE2 expression in intestinal tissue of the humanized mouse lines.

A. Ideogram of the structure promoter/exon intron origin, and nomenclature/color scheme used in referring to the three mouse lines compared in these studies. mAce2 has the mouse promoter and coding sequence, MP-ACE2 has the mouse promoter but human coding sequence, and hACE2 has the human promoter and coding sequence. B. Droplet digital PCR (ddPCR) evaluation of ACE2 expression in cDNA prepared from male (♂) and female (♀) mice of the indicated genotype. Mean value for each is shown below. C. Western analysis of lysates prepared from mice of the indicted genotype/sex with the human-specific MAB933 antibody and ab15348, which recognizes both human and mouse ACE2. D. Comparison of the proximal to distal expression in the intestinal tract of ACE2/Ace2 in female mice heterozygous for the mouse locus (mACE2) and humanized locus (hACE2). Expression observed in the jejunum for the human and mouse gene was assigned a value of 1. ** p<0.01. For statistical comparison between all groups in B, see S1 Table.

https://doi.org/10.1371/journal.ppat.1011168.g002

In defining the transition sites from mouse to human DNA in the two lines, we considered information available concerning the promoter of the mouse and human genes and presence of transcription factor binding motifs [35]. For example, two evolutionarily conserved regions, referred to as the distal and proximal promoter regions, have been identified just upstream of the ACE2 translational start codon in both the mouse and human genes [36]. Mice and humans both express ACE2 transcripts that include an untranslated first exon (exon 1a) [37,38]. In the MP-ACE2 mouse line, the mouse/human transition occurs at the translational start site in exon1b, thus maximizing the length of the upstream mouse regulatory region and ensuring that splicing events involved in generation of the 5’UTR occur between mouse-derived sequences.

Expression of ACE2 in intestinal tract of humanized mice

ACE2/Ace2 is highly expressed in both the human and mouse intestinal tract, where it is located at the luminal membrane of small intestine enterocytes. Expression is also observed, albeit at lower levels, in intestinal crypt cells and in the colon [12]. We initiated characterization of the humanized mouse lines by examination of transcript levels in the jejunum (Fig 2 and S1 Fig). ACE2 is located on the X chromosome in both humans and mice. If X inactivation of the locus is random, approximately 50% of the cells in tissues from female offspring will express the human gene and approximately 50% will express the mouse gene. To simplify comparison of ACE2 and Ace2 expression between female and male mice throughout this report, the number of copies of Ace2 and ACE2 transcripts present in cDNA prepared from females heterozygous for the humanized locus (mAce2 x MP-ACE2 or mAce2 x hACE2) were multiplied by two.

Analysis of mRNA from the jejunum showed expression of Ace2 and ACE2 in the mAce2 and MP-ACE female mice, with mean values of 4415 and 3319 copies per ng of cDNA respectively (Fig 2B). Expression of ACE2 in the two MP-ACE2 males was slightly lower than Ace2 expression in mAce2 males (Fig 2B). This was surprising as, while sexual dimorphism for ACE2 has been reported, for the majority of tissues examined ACE2 is reported to be expressed at higher levels in males [39,40]. The expression of ACE2 in the mouse line carrying the fully humanized locus (hACE2) was also reduced by about 50% compared to mACE2 (Fig 2B). However, in this case no sexual dimorphism was observed. The human promoter drove reduced gene expression of ACE2 in both hACE2 males and females compared to their mAce2 count parts.

We determined if the difference we noted in ACE2 mRNA levels between lines correlated with differences in protein levels detectable by western analysis (Fig 2C). ACE2 levels were measured in lysates prepared from the jejunum of mAce2, MP-ACE2 and hACE2 mice using two different anti-ACE2 antibodies, MAB933 and ab15348. MAB933 is a species specific monoclonal antibody that recognizes the extracellular region of human ACE2, and ab15348 is a rabbit serum raised against epitopes present in the intracellular/carboxyl terminal domain of mouse and human ACE2. The anti-ACE2 monoclonal, MAB933, detected a 120 kDa protein in lysates prepared from male and female MP-ACE2 and hACE2 animals. As expected, the MAB933 did not identify any band corresponding to ACE2 in lysates prepared from the mAce2 mice. Consistent with the lower mRNA levels of ACE2 in female hACE2 mice compared to MP-ACE2 females, a noticeably less intense 120 kDa band was observed in the females carrying the fully humanized line.

Expression of ACE2 was then compared in lysates from the jejunum of mAce2, MP-ACE2 and hACE2 male mice. Consistent with the similar mRNA levels, the human specific antibody, MAB933, detected a 120 kDa band of similar intensity in the lysates from the MP-ACE2 and hACE2 mice, while no signal was observed in the lane corresponding to the sample from mAce2 animals (Fig 2C).

Lysates were also analyzed with anti-ACE2 serum, ab15348, which recognizes both human and mouse ACE2 proteins (Fig 2C). Similar to previous reports, we noted the differing mobility of mouse ACE2 and human ACE2 on western analysis [41,42]. Although human and mouse ACE2 are 805 a.a. in length (92,463 Da) and the predicted molecular weight of ACE2 from the two species is virtually identical, mouse ACE2 consistently displayed a lower molecular weight on western analysis than the human protein, perhaps reflecting the absence of two of the seven N-glycosylation sites present in human ACE2 [43]. Also apparent on analysis of the lysates with MAB933 and ab15348 were differences between these antibodies in there cross reactivity with other proteins present in the lysates. Despite this limitation, the ab15348 anti-ACE2 antibody identified ACE2 protein in lysates from both species, with the highest intensity corresponding to lysates prepared from the jejunum of the mAce2 mice. This is consistent with the 2 fold higher Ace2 cDNA copies in male mAce2 mice when compared to the two humanized lines (Fig 2B). Despite differences in the level of expression, human and mouse ACE2/Ace2 showed a remarkably similar pattern of expression, with the lowest levels observed in the stomach and colon (Fig 2D).

ACE2 expression in kidney and heart

Both human and mouse express high levels of ACE2/Ace2 in the kidney, and approximately 50% of ACE2 expression is associated with proximal tubule cells, where it is localized to the apical (luminal) brush border of the epithelial cells [44,45]. We examined ACE2/Ace2 mRNA levels in mAce2, MP-ACE2, and hACE2 mice by ddPCR. As predicted by previous studies, expression of Ace2 in the kidney of the mAce2 mice was lower than that observed in the intestinal tract, with approximately 500 copies/ng of RNA compared to 4,000 copies/ng in the intestinal tract (Compare Figs 2B and 3A). Consistent with previous reports, expression of Ace2 was slightly higher in males [46], but in our evaluation by ddPCR, this difference failed to reach significance. Robust expression of ACE2 was observed in the MP-ACE2 females; however, levels were significantly lower than those measured in mAce2 animals. This decrease in ACE2 expression was not observed in males. In contrast, although only two MP-ACE2 males were included in the study, expression of ACE2 in these animals was similar to that observed in mAce2 males, suggesting that regulatory elements contributing to sexual dimorphism in ACE2 expression may reside within the human coding segment of the ACE2 locus. However, a prominent role for the promoter in driving sexual dimorphism was supported by our examination of expression of ACE2 in mice with the fully humanized locus, as a four-fold increase in ACE2 transcripts was observed in male hACE2 kidneys compared to those from females (Fig 3A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Comparison of expression of ACE2 in kidney of the humanized mouse lines.

A. Droplet digital PCR (ddPCR) evaluation of cDNA prepared from male (♂) and female (♀) mice of the indicated genotype. Mean value for each is shown below the bar graph. B&C. Western analysis of lysates prepared from mice of the indicted genotype and sex with the human specific MAB933 antibody (B) and ab15348 (C), which recognizes both human and mouse ACE2. D. kidney ACE2 carboxypeptidase activity measured at one hour after initiation of the reaction. E. ACE2 carboxypeptidase activity in urine collected from mice of indicated sex and genotype determined one hour after initiation of the reaction. ** p<0.01. For statistical comparison between all groups in A see S1 Table. For statistical comparison between all groups in D and E see S1 Table.

https://doi.org/10.1371/journal.ppat.1011168.g003

We determined whether these differences in mRNA levels translated into differences in ACE2 protein levels in the kidney. Western blot analysis using the human specific anti-ACE2 antibody MAB933 of lysates prepared from MP-ACE2 and hACE2 mice showed an increase in protein levels in the kidney of the mice expressing the fully humanized locus. Consistent with the small increase in mRNA levels in the female hACE2 mice compared to their MP-ACE2 counterparts, an increase in ACE2 was observed in lysates from the females carrying the fully humanized locus (Fig 3B).

We next compared ACE2 protein levels in kidney lysates prepared from mAce2, MP-ACE2 and hACE2 males using the rabbit polyclonal antibody ab15348, which recognizes both human and mouse ACE2 (Fig 3C). The hierarchy of expression of ACE2/Ace2 mRNA in the three mouse lines was recapitulated on western blot analysis, with the lowest levels of expression observed in mAce2 mice, a slight increase observed in MP-ACE2 line, and a dramatic increase in the hACE2 kidney.

Together, our results suggest that elements within the human promoter amplify sexual dimorphism in ACE2 expression in the kidney. However, minor contribution by elements within the coding regions of the gene is also supported by the less dramatic difference we observed between the sexes in ACE2 expression in the MP-ACE2 mice compared to hACE2 mice.

Similar studies of expression of ACE2 in the mouse heart were carried out. Again, a very complex pattern of expression was observed, with the mouse promoter failing to drive levels of ACE2 expression observed in wild type mice. The lower levels of expression in the MP-ACE2 mice compared to the hACE2 mice suggested interaction between intergenic and upstream regulatory regions (S2 Fig).

Evaluation of ACE2 carboxypeptidase activity in humanized mouse lines

ACE2 encodes a carboxypeptidase, and thus enzymatic activity is commonly used as a surrogate measure of ACE2 protein levels [47]. Activity towards a known ACE2 substrate is assessed by carrying out the reaction in the presence and absence of an ACE2 inhibitor [41]. An important advantage of this method is the ability to measure soluble ACE2 (sACE2). The ectodomain of ACE2 is shed from the cell membrane [48], both constitutively and in response to physiological changes; however, sACE2 retains both its carboxypeptidase activity and ability to bind SARS-CoV-2 [49]. We therefore examined ACE2 activity in both kidney lysates and urine (Fig 3D and 3E). sACE2 was detectable in urine [50], likely reflecting shedding from epithelia of the proximal tubules [51], as its size makes it unlikely that it is excluded from the glomerular filtrate. ACE2 activity in both the kidney and the urine from the mAce2, MP-ACE2 and hACE2 mouse lines generally aligned with the number of Ace2/ACE2 transcripts observed by ddPCR. Sexual dimorphism was observed in all three lines, but the magnitude of the difference between male and females was increased with humanization of the locus.

ACE2 expression in the lung and airways

The central role of the lungs in COVID-19 related morbidity and mortality is surprising given the extremely low level of expression of this gene in the human lung, with a normalized expression level of 0.8 compared to 122.0 for the small intestine, 23.0 for the kidney, and 10.5 for the heart [52,53]. In contrast, a comparison of ACE2 activity between mouse organs reported that ACE2 activity in the mouse lung was relatively high, being approximately 25% that of the kidney, and no significant sex bias was observed [46]. We therefore determined whether this species difference in ACE2/Ace2 expression in the airways was reiterated in the two humanized mouse lines.

Lung, trachea, nasal epithelia were collected from mAce2, MP-ACE2, and hACE2 mice and evaluated for expression of ACE2/Ace2 by ddPCR (Fig 4A, 4C and 4E). In collection of the nasal epithelia, the respiratory and olfactory epithelia were isolated individually from the three mouse lines for mRNA expression analysis (Fig 4E). The enrichment for olfactory and respiratory epithelium was evaluated by examining expression of genes characteristic of these two tissues: Muc5a for respiratory epithelium and Ugt2a1/2 for olfactory epithelium (S3 Fig) [54,55]. As an additional indicator of levels of ACE2 protein levels, carboxypeptidase activity was determined in tissue homogenates for all tissues except olfactory epithelium (Fig 4B, 4D and 4F).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Comparison of ACE2 expression in lung and airways of humanized mouse lines.

A,C,E. ddPCR evaluation of cDNA prepared from male (♂) and female (♀): A, lung; C, trachea; E, nasal epithelium of mice of the indicated genotype. Mean number of copies per nanogram of cDNA for each group is shown in the table below the bar graphs for each tissue. B, D, F. ACE2 carboxypeptidase activity in samples evaluated for expression: B, activity in lung; D, activity in trachea; F, activity in nasal epithelium (respiratory). Activity in each sample was measured 1 hour after initiation of the reaction. For group sizes and statistical comparison between all groups in A, C, and E, see S1 Table. For group sizes and statistical comparison between all groups in B, D, and F, see S2 Table. OE, olfactory epithelium; RE, respiratory epithelium.

https://doi.org/10.1371/journal.ppat.1011168.g004

Robust expression of endogenous murine Ace2 gene was seen in the lung, with over 100 copies/ng cDNA, only four-fold lower than that observed in the kidney (~400 copies/ng). ACE2 carboxypeptidase activity paralleled this high expression, with activity approximately 25% of that observed in the mAce2 kidney (Figs 3D vs 4B). In contrast, expression of the humanized ACE2 locus in the hACE2 lung was approximately 100 fold lower than in the kidney, in agreement with the relative expression of ACE2 in these two organs in humans [52]. Slightly higher ACE2 expression was seen in the lung, trachea and nasal respiratory epithelia of male hACE2 mice relative to females, although this difference did not reach the level of statistical difference. In contrast, ACE2 expression in the olfactory epithelium of female hACE2 mice was approximately eight-fold higher than that measured in male hACE2 mice and more than 15-fold higher than that measured in other parts of the airway for either sex.

Expression in the airways of the MP-ACE2 mice was generally higher than that from the endogenous mouse Ace2 locus, except in the lungs, where expression from the MP-ACE2 locus was reduced by approximately 80% in female and 60% in male lungs relative to expression from the Ace2 locus (Fig 4A). These results were somewhat unexpected, since expression from both loci is driven by the same endogenous mouse upstream promoter elements. ACE2 specific carboxypeptidase activity was easily measurable in the airway epithelia of the MP-ACE2 and hACE2 mice, verifying translation of the human transcripts (Fig 4B, 4D and 4F). The decrease in MP-ACE2 expression in the lung was paralleled by an approximate fourfold decrease in human ACE2 protein levels (Fig 4B).

ACE2 expression in lung club and AT2 cells

A possible explanation for the low expression of ACE2 in the lung of the MP-ACE2 line is that the human gene is not expressed in club cells, which have been shown to be vulnerable to infection by the SARS-CoV-2 MA10 virus [56]. The sensitivity of mouse club cells to naphthalene because of their unique expression of Cyp2f2 provided a means of addressing this question [57]. Furthermore, carrying out the study with heterozygous female mice that carry one copy of Ace2 and one copy of the MP-ACE2 allowed assessment of expression from the human and mouse genes in the same tissue sample. Mice were treated for either 24 or 48 hours with naphthalene after which mRNA prepared from the lungs was evaluated for selective loss of club cells and expression of Ace2 and ACE2 was determined (Fig 5A) [58]. Within 48 hours, expression of two club cell mRNAs, Cyp2f2 (Cytochrome P450 2F2) and Scgb1a1 (Uteroglobin), could no longer be detected. Expression of Foxj1, which is expressed primarily in lung ciliated cells, was significantly increased. This is consistent with previous demonstration of the reparative role of ciliated cells, characterized by an increase in Foxj1 expression, as these cells lose their cilia and undergo squamous cell metaplasia in response to club death and detachment from the basement membrane [59]. Expression of the ciliated cell-specific Dnah6 (dynein axonemal heavy chain 6), essential in producing force for ciliary beating, is marginally reduced and then restored as the ciliated cells reestablish normal morphology and function. Expression of Ace2 could be detected 48 hour. after treatment, albeit at approximately 30% of the level observed in vehicle treated animals. In comparison, all MP-ACE2 expression was lost in the naphthalene treated MP-ACE2 animals, suggesting that expression of ACE2 in the MP-ACE2 mice is largely localized to the Cyp2f2-expressing, naphthalene sensitive club cell population. Differences in ACE2 expression between the MP-ACE2 and mAce2 mice could reflect the expression of Ace2 in one or more additional cell populations in the mouse.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Expression of ACE2 in club and alveolar type II (AT2) cells.

A. Expression of ACE2 and Ace2 in female mice heterozygous for mAce2/MP-ACE2 24 and 48 hours after exposure to naphthalene (Naph) normalized to expression of ACE2 and Ace2 in the vehicle treated animals. Club and ciliated cell resistance to naphthalene was evaluated by measuring change in expression of cilia and club cell specific marker genes in samples during this same time interval. B. ACE2 expression is unaltered in heterozygous mAce2/hACE2 treated with naphthalene. C. Expression of Sftpc verifies enrichment of AT2 population from lung of mAce2/hACE2 heterozygous females. D. ACE2 expression is increased six fold by in airway epithelium enriched for AT2 cells. E. Comparison of ACE2 carboxypeptidase activity in lung lysates and BALF of mAce2 and MP-ACE2 mice. For group sizes and statistical comparison between all groups in E see S2 Table. ** p<0.01.

https://doi.org/10.1371/journal.ppat.1011168.g005

Using a similar strategy, we measured ACE2 expression in female mice heterozygous for endogenous mouse Ace2 and hACE2 (Fig 5B). While a substantial decrease in the expression of Ace2 was observed 48 hours after naphthalene treatment, no change in the expression of the ACE2 was detected. These results indicate that the human and mouse promoters differ markedly in their ability to direct expression of Ace2/ACE2 to specific cell populations in the lung, with the mouse promoter driving high expression in club cells.

To determine the possible source of ACE2 expression in the hACE2 mice, populations enriched for alveolar type II (AT2) cells were isolated from heterozygous female mice. Enrichment was verified by assessing changes in expression of the AT2 specific transcript, Sftpc (Fig 5C). A six fold increase in expression of ACE2 was observed in the AT2 enriched population (Fig 5D). This is consistent with a model in which AT2 cells are the major cell type in which the human promoter drives ACE2 expression in the mouse lung.

Shedding of ACE2 into the airway lumen

ACE2 ectodomain shedding has been reported for human airway epithelia [49] and mouse lungs, where it is reported to play a role in regulation of neutrophil influx [60]. More recently, it has been suggested that the shed ectodomain (soluble ACE2) can mediate entry of SARS-CoV-2 into cells lacking ACE2 expression [61]. Certainly, loss of ACE2 from the cell membrane will impact it’s vulnerability to viral entry [62]. To determine the sensitivity of the human protein to shedding, we measured human ACE2 in the BALF of the humanized mouse line (Fig 5E). Whole lung lavage was carried out on mAce2 and MP-ACE2 mice prior to tissue collection. Carboxypeptidase activity was determined in both the bronchoalveolar lavage fluid (BALF) and lung homogenates. ACE2 enzyme activity was observed in the lavage and lung samples of both animals, indicating that human protein is actively shed by mouse epithelium.

Expression of the ACE2 isoforms in the humanized mouse line

The ACE2/Ace2 promoter in both human and mice is bipartite, with full length ACE2/Ace2 transcripts beginning with one of two initial exons, 1a or 1b [35]. Human RNA-seq studies indicate low levels of exon1a transcripts in most tissue examined, with the highest expression levels reaching approximately 10% of those of exon1b transcripts in the small intestine [63]. Primers specific for transcripts for ACE2/Ace2 that include exon 1a were generated and used to examine expression of 1a containing transcripts in tissues from the mAce2, MP-ACE2, and hACE2 mice. In mAce2 and MP-ACE2 mice, we found that exon1a containing transcripts comprised all of the transcripts present in the mouse lung and trachea (Fig 6A). In the fully humanized hACE2 mice (Fig 6B), similar to the MP-ACE and mAce2 mice, all the ACE2 transcripts in the lung included exon1a. However, in contrast to the MP-ACE2 and mAce2 mice, no exon 1a transcripts were observed in the tracheal samples from the hACE2 mice. Thus the human promoter only supports 1b transcripts. In the kidney and intestinal tract, exon1a containing transcripts were absent in all three mouse lines (Fig 6C). This is congruent with the human RNAseq evaluation of exon1a transcripts in that we found about 10% of the transcripts in the intestinal tract of the hACE2 mice included exon 1a [63]. These results indicate that inclusion of exon 1a in transcripts encoding ACE2 varies not only between species but also between tissues. This further supports a model in which expression of ACE2/Ace2 is regulated in a species- and tissue-specific manner.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Expression of ACE2 exon1a and exon1b mRNA transcripts in wild type and humanized mouse lines and induction of exon 1c transcripts after exposure to tilorone.

A-C Comparison of total ACE2/Ace2 transcripts measured by ddPCR (open bars) to those that include exon 1a in mRNA (shaded bars) isolated from the tissues and mouse lines indicated. D-E shows induction of exon1c transcripts in lung of humanized mice after tilorone (+ til) induced increase in interferon. ddPCR evaluation of ACE2 lung transcripts using PCR primers that recognized all ACE2 transcripts (open bars), primers that recognize either the short “Δ” transcript, or primers specific for the “long”, full length ACE2 mRNA (shaded bars). F. Induction of interferon responsive genes (IRG) was verified by demonstration of increased expression of Nlrc5 gene in the tilorone treated samples by qPCR. Expression is normalized to that measured in vehicle treated animals. For group sizes and statistical comparison between all groups see S1 Table. *p<0.05; ** p<0.01; *** p<0.005.

https://doi.org/10.1371/journal.ppat.1011168.g006

In vivo induction of dACE2/short ACE2 by interferon

Evaluation of changes in mRNA expression in airway epithelial cells from naïve and inflamed lungs has identified ACE2 as an interferon-stimulated gene (ISG) [64,65]. More recently, the effect of interferon on expression of both full length and variant transcripts has been extensively explored in primary epithelium and epithelial cell lines [33,63]. These studies report that interferon mediates the increase in an ACE2 mRNA transcript initiated in intron 8 from an intronic promoter, which results in the inclusion of an exon unique to primates, exon1c (see Fig 1). The transcript is predicted to encode a truncated ACE2 isoform (short or delta (d)ACE2) that lacks both the carboxypeptidase and SARS-CoV-2 binding domains of the longer protein. However, direct in vivo evidence supporting the induction of dACE2 expression in response to interferon is lacking, as is evidence for the dependence of its expression on an internal promoter and its specificity for airway epithelium. To this end, the three mouse lines, mAce2, MP-ACE2 and hACE2, were treated with tilorone, a potent interferon inducing drug [66], and expression was evaluated by ddPCR using three sets of primers: primers that recognize all human ACE2 transcripts, primers that detect only the full length transcript, and primers that detect only the dACE2 (short ACE2) transcript (Fig 6). The induction of ISG in all lines was verified by examination of induction of Nlrc5 (Fig 6F) [67]. No increase in the expression of Ace2 was observed in the tilorone treated animals. In contrast, we observed similar fold increases in the copy number of total ACE2 transcripts in both the MP-ACE2 and the hACE2 lines. The increase in expression of ACE2 in the fully humanized mouse was accompanied by increases in the dACE2 but not long ACE2 transcript. In contrast, an increase in the full length transcript was observed in the MP-ACE2 mice. Thus, while mouse Ace2 was not induced by interferon, the combined presence of the mouse promoter and the human exon/intron sequences in MP-ACE2 mice rendered the mouse promoter interferon sensitive.

SARS-CoV-2 infection of MP-ACE2 mice

To determine the susceptibility of the humanized lines to SARS-CoV-2 infections, seven MP-ACE2 (N2 BALB/cBy) male mice, 16 weeks of age, and two mAce2 littermates were exposed intranasally (i.n.) to 10⁵ pfu of virus (Fig 7A). No difference in the overall health of the mice was observed over the 5-day period, including no significant change in weight. At the end of this time, animals were euthanized, and homogenates were generated from the lung for evaluation of viral load and for isolation of mRNA. Virus was detected in five of the seven MP-ACE2 mice, although titers were relatively low compared to those achieved with the mouse-adapted virus [56] in all but one of the animals (Fig 7A). As expected, viral titers were below detection in the mAce2 littermate controls. As MP-ACE2 mice are generated by syntenic replacement, they do not express the mouse ACE2 receptor, and thus cell entry is mediated by human ACE2.

To verify that expression of ACE2 in the MP-ACE2 mouse line was sufficient to support viral proliferation, an additional experiment was carried out using MP-ACE2 mice on the C57BL/6N genetic background. In this case, C57BL/6N mice carrying the FOXJ1-ACE2 transgene were included as a positive control for infection and replication of SARS-CoV-2 in the lung. This mouse line, which has been used extensively for study of SARS-CoV and SARS-CoV-2 [27,29,68] carries a transgene in which human ACE2 expression is driven by the promoter of FOXJ1, a gene which in the lung is expressed by ciliated cells [69]. We verified expression of ACE2 in this line, comparing expression to that in MP-ACE2 mice and to Ace2 in wild type mice (Fig 7B). Robust expression was observed in the FOXJ1-ACE2 lung, intermediate between that of Ace2 in mAce2 mice and ACE2 in MP-ACE2 mice. Because of the rapid resolution of infection in the MP-ACE2 BALB/cBy mice, the lung SARS-CoV-2 load was examined 2 and 5 days after intranasal delivery of 1 x 10⁵ PFU. Mock infected FOXJ1-ACE2 and MP-ACE2 mice were included as controls. Again, no loss of weight was observed in either mock or viral infected MP-ACE2 mice. Despite the lower expression of ACE2 in the MP-ACE2 mice compared to the FOXJ1-ACE2 animals, the viral loads measured in the lung of the MP-ACE2 were approximately 20-fold higher than those of the FOXJ1-ACE2 mice 2 days post infection. However, because of the small group size, the difference did not achieve significance (Fig 7C). Interestingly, despite the observed higher titers of virus in the MP-ACE2 mice compared to the FOXJ1-ACE2 line 2 days after inoculation, by day 5 the MP-ACE2 mice had cleared the virus while measurable titers remained in the FOXJ1-ACE2 animals.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. SARS-CoV2 exposure of MP-ACE2, and mAce2 mice.

~16 week old mice of the indicated genotype were infected with 10⁵ PFUs of SAR-CoV-2 or vehicle (mock infected) and evaluated at either 2 days (DPI 2) or 5 days (DPI 5) for: A,C viral titer; B, abundance of ACE2 mRNA, determined by ddPCR in FOXJ1-ACE2 transgenic mouse line; C,D expression of SARS-CoV-2 nuclear capsid genes; E, expression of lung cell-type-specific genes; D-H, expression of the indicated pro-inflammatory genes. MP-ACE2 mice were N3/N4 generation C57BL/6N, N2 generation BALB/cBy, or ~N5 generation FOXJ1-ACE2 C57BL/6 (fuchsia). mACE2 (black) are littermates of the BALB/cBy MP-ACE2 animals. For group sizes and statistical comparison between all groups see S1 Table. * p<0.05; ** p<0.01; *** p<0.005; **** p<0.001.

https://doi.org/10.1371/journal.ppat.1011168.g007

mRNA was prepared from the lungs of the infected and control animals, and the levels of SARS-CoV-2 subgenomic and genomic nucleocapsid (N) RNA were determined (Fig 7D and 7E). Two days after infection of MP-ACE2 mice, high levels of N-RNA were observed, and, consistent with differences in viral load, these were 10-fold higher in the MP-ACE2 mice compared to the FOXJ1-ACE2 animals. At day 5, despite clearance of virus, high levels of N-RNA were still easily detected. N-RNA was also detected in all 7 MP-ACE2 BALB/cBy mice. As expected, given the species specificity of SARS-CoV-2, nuclear capsid RNA was not detected in RNA prepared from the lungs of the two mAce2 littermates exposed to virus.

Host response to viral infection

We determined whether SARS-CoV-2 infection led to measurable changes in major lung epithelial cell populations by qPCR analysis of mRNA prepared from mock and SARS-CoV-2 infected FOXJ1-ACE2 and MP-ACE2 mice using a panel of probes specific for alveolar type I (AT1), AT2, ciliated, or club cells. No changes were observed in the expression of AT1 or AT2 cell markers. The pattern of change in the expression of the two club cell specific genes, Cyp2f2 and Scgb1a1 in MP-ACE2 mice followed that observed with naphthalene mediated damage to club cells. Expression was decreased two days post infection, with rapid normalization by day 5. Similarly, changes in Foxj1 expression followed those observed in mice after damage to club cells. In contrast, no significant change in the expression of these genes could be measured in the FOXJ1-ACE2 mice. However, we cannot rule out the possibility that this reflects the smaller number of these mice included in the experiment.

An analysis of the expression of cytokines and chemokines in the infected MP-ACE2 mice revealed a robust immune response to SARS-CoV-2 infection, after which homeostasis was rapidly restored. A significant increase in Il1b was observed at day 2, with normalization by day 5 (Fig 7G). Tnf and Ccl2 followed a similar pattern, with elevated expression on day 2 that normalized with reduction of the viral load by day 5 (Fig 7H). A dramatic increase in Il6, which was found to be increased six-fold relative to mock infected animals, was detected at day 2 (Fig 7G). While transcripts for Infg were more abundant in the SARS-CoV-2 exposed animals at day 2, the increase in expression did not achieve significance. However, a robust and significant increase in the INF-γ sensitive Nlrc5 gene was observed (Fig 7H). As NLRC5 plays a key role in increased MHC class I gene expression secondary to interferon induction, we examined the expression of H2-K1 in the lung. A significant increase was observed in expression of this MHC antigen, an important component of both natural killer and cytotoxic T cell mediated removal of virally infected cells [70,71].

Histological evaluation of the lungs of the MP-ACE2 (Fig 8) mice supports a model in which viral replication in club cells leads to cell death and exfoliation of this airway epithelial population. The distal airways, the alveoli, and alveolar duct remained relatively unaffected. In contrast, loss of club cells as well as morphological changes in club cells consistent with both apoptosis and necrosis were seen throughout the length of the airway (Fig 8D–8F). However, repair was rapid, with little cell death, and repopulation of the airways was observed by 5 days post infection. In contrast, death and exfoliation of airway epithelial cells was less apparent in the Foxj-ACE2 mice, suggesting perhaps an increased tolerance to viral replication. Loss of ciliated cells was not apparent on examination of the larger airways where ciliated cells can be easily identified (Fig 8I). Surprisingly and perhaps in response to failure to clear the virus, histological changes were apparent in the distal lung, with septal thickening, increased cellularity and changes in the architecture of the alveolar unit.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Histological analysis of Foxj1 and MP-ACE2 SARS-CoV-2 infected lungs.

Tissue sections from left lobe of lungs from mock-infected and SARS-CoV-2 infected animals at 2 days post infection, fixed and stained with hematoxylin and eosin. (A-C) Mock infected MP-ACE2. (D-F) MP-ACE2 infected animals. (G-I) Foxj1-ACE2 infected animals. (A) Distal lung of a mock infected MP-ACE2 mouse showing a terminal bronchiole (tb) and the associated alveolar duct (ad) with club cells dominating the epithelium. (B) High magnification of the epithelium showing healthy club cells (arrowhead), which are easily identified by eosinophilic “club” shaped cytoplasmic protrusions into the airway. (C) Bronchi of mock-infected animal shows an epithelium dominated by ciliated cells (arrow) and submucosal smooth muscle (sm) and connective tissue (ct) characteristic of larger conducting airways. (D-F) Similar regions of lungs from MP-ACE2 SARS-CoV-2 infected mice show limited changes in architecture of the alveolar/capillary units. However, extensive damage to epithelium (particularly apoptosis/necrosis of club cells), is apparent extending from the terminal bronchiole to the bronchus. (D) Detached cells form a loosely adherent layer of cellular debris at the apical surface of the terminal bronchiole (red dashed line). Higher magnification of the epithelium throughout the airway (E,F) shows an abundance of detached epithelial cells with pyknotic nuclei and karyorrhexis, as well as cells with condensed eosinophilic cytoplasm consistent with apoptotic death (red star). (G-I) Limited damage to airway epithelium, club cells and ciliated cells was observed in the Foxj1-ACE2 mice. However, changes in the architecture of the distal lung is apparent (G, H) with loss of alveolar/capillary units, increased cellularity and septal thickening. (A,D,G) Bars represent 150 μm. Bars in remaining panels represent 50 μm.

https://doi.org/10.1371/journal.ppat.1011168.g008

Differential immune response of the Foxj1-ACE2 and MP-ACE2 mice to viral infection

An additional experiment was carried out to verify the robust proliferation of SARS-CoV-2 in the MP-ACE2 mice. A cohort of Foxj1-ACE2 mice was included to further examine the initial observation of differences between the two lines in the clearance of virus over the 5-day period following viral exposure (Fig 9). Mice were infected with 1 x 10⁵ PFUs of virus and weighed daily. A small drop in body weight was observed during the course of the experiment but failed to reach significance for either mouse line relative to mock infected animals (Fig 9A and 9B). No change in body condition score was noted in the mice over the course of the experiment, and the gross appearance of SARS-CoV-2 infected lungs did not differ from those of mock infected animals. Consistent with studies of the smaller Foxj1-ACE2 cohort shown in Fig 7, ACE2 levels in the Foxj1-ACE2 mice were over 2 fold higher than those measured in the MP-ACE2 animals.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 9. Comparison of SARS-CoV-2 clearance and immune response after exposure of Foxj1-ACE2 and MP-ACE2 mice to SARS-CoV-2.

(A and B) Weight loss of mice over the 5 day duration of the experiment did not differ (2 way ANOVA). (C) Expression of ACE2 determined by ddPCR in mice of the indicated genotypes. (D) Copies of SARS-CoV-2 sgRNA (nucleocapsid transcript) present in RNA prepared from the right inferior lobe. (E-H) Expression of the indicated proinflammatory gene. The level of expression in the control animals (mock or mACE2) is assigned a value of one. For MP-ACE2 mice, mAce2 littermates served as control animals, with both receiving an inoculum 10⁵ PFU SARS-CoV-2. For the Foxj1-ACE2 mice, mock (PBS) treated mice served as controls. For group sizes and statistical comparison between all groups, see S1 Table. * p<0.05; ** p<0.01; *** p<0.005; **** p<0.001.

https://doi.org/10.1371/journal.ppat.1011168.g009

Cohorts of animals were euthanized at two days and five days post infection, and ACE2 expression and viral load were assessed by measurement of nucelocapsid sgRNA transcripts present in mRNA prepared from the lung. Surprisingly but consistent with the results described above, a two log drop in sgRNA SARS-CoV-2 levels was observed in the MP-ACE2 mice, while sgRNA levels remained unaltered in the Foxj1-ACE2 animals five days post infection. A major difference in the pathogenesis of disease in the two ACE2 mouse lines was also apparent on evaluation of cytokine levels in the lung. As shown in the previous experiment, the clearance of virus in the MP-ACE2 mice was paralleled by a robust increase in transcripts for Ccl2, TNFα, and IL-6 (Fig 9E–9G). Elevated levels of Nlrc5 transcripts were observed in the MP-ACE2 mice, indicative of an interferon-mediated increase in expression of MHC antigens. In contrast, no significant increase in the levels of these markers was observed in the Foxj1-ACE2 mice. These changes in gene expression as well as the histological changes we observed on analysis of the MP-ACE2 infected mice are consistent with a model in which SARS-CoV-2 infection of lung ACE2 expressing club cells in the MP-ACE2 mice induces a rapid and protective immune response that limits viral spread. Together, the immune response and death of the infected club cells mediate clearance of virus from the lung and initiation of epithelial repair by 5 days post infection.

Ethics and biosafety

All animal work meets the standards of the Institutional Animal Care and Use Committee at University of North Carolina at Chapel Hill as set out in guidelines outlined by the U.S. Department of Agriculture as well as by the Association for the Assessment and Accreditation of Laboratory Animal Care. Safety conditions and approved standard operating procedures have been followed for all experiments involving SARS-CoV-2. BSL3 facilities at University of North Carolina at Chapel Hill were designed to meet safety requirements recommended by the U.S. Department of Health and Human Services, Biosafety in Microbiological and Biomedical Laboratories (BMBL), the National Institutes of Health (NIH), the Centers for Disease Control and Prevention (CDC), and the Public Health Service. We have submitted laboratory safety plans, and the CDC and the University of North Carolina at Chapel Hill Department of Environmental Health and Safety (EHS) have approved the facility for use.

Generation of mouse lines

The Ace2 displacer constructs were assembled using a standard recombineering approach. The mouse arms of homology were derived from the 129s7/AB2.2 bMQ BAC library clone bMQ306a09 (Source BioScience). The segment of human genomic DNA containing the ACE2 gene was derived from the human tile path BAC RP11-478H11 (BACPAC Resources). The resistance marker gene used for selection of embryonic stem (ES) cells in which the Displacer construct underwent genomic integration consisted of a PGK-neo cassette flanked by mutant loxP sites. Excision of the marker gene with transient expression of Cre leaves a nonfunctional lox site in its place. The recombination events for generation of both lines were carried out in the 129 ES cell line, Phnx43. ES cells used were derived from 129SvEv mice. ES cells carrying the correctly modified locus were used to generate chimeric mice, which were bred to 129S6/SvEv/Tac mice (Taconic Bioscience) to maintain the mutation on this genetic background. Chimeras were also bred to C57BL/6NCrl (Charles River Laboratories) and BALC/cBy, strain 000650 (The Jackson Laboratory) mice, and offspring were used for examination of susceptibility to SARS-CoV-2. The MP-ACE2 mouse line is designated by the MMRC as 069708-UNC, 069709-UNC and 069713-UNC; the hACE2 line is designated as 069710-UNC, 069714-UNC, and 069715-UNC; and the FOXJ1-ACE2 line is designated as 066719-UNC.

Enrichment for alveolar type II cells

Alveolar type II cells were isolated as described previously [81–83]. Briefly, lungs were perfused and airways lavaged to remove blood and airways cells, respectively. The lungs were instilled with one mL of 10 U/mL of Dispase II (Roche) in PBS, followed by 0.5 mL of 1% low melting agarose, removed and incubated in 0.5 mL of Dispase II solution for 45 minutes at room temperature. At the end of the incubation period, digested tissue was transferred to 7 mL of DMEM supplemented with 10% FBS and 0.01% DNase I, teased free of bronchi and bronchioles and further incubated. The resulting cell suspension was filtered through a 70 μm centrifuged at 300 x g for 10 minutes at 4°C. The cell pellet was resuspended in 1 mL of 4% isotonic Percoll (Sigma, St. Louis, MO) and the suspension overlaid on a Percoll step and centrifuged at 400 x g for 20 minutes at 4°C with no brake. AT2 enriched cells were collected at the 10–30% interface and washed once with PBS. To further remove contaminants, cells were stained with biotinylated antibodies against lineage markers (anti-CD45, anti-CD16/32, anti-CD31, and anti-integrin B4), labeled with streptavidin microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany), and then subjected to magnetic separation according to manufacturer’s instructions. The enriched type II cells were collected and used for isolation of RNA.

RNA preparation and expression analysis

Tissues were collected from male and female mice between 10 and 16 weeks of age. Total RNA was isolated using an RNA isolation solvent (Stat-60; Tel-Test, Friendswood, TX, USA) according to the manufacturer’s protocol. For qRT-PCR, RNA was reverse transcribed to cDNA using a high-capacity cDNA archive kit (Applied Biosystems) following the manufacturer’s recommended protocol. All probes and primers were purchased from a commercial vendor (Applied Biosystems). Amplification of DNA was carried out on the Applied Biosystems 7900 HT Fast RT-PCR System using qBio Blue (Genesee Scientific). Each sample was run in duplicate, and relative expression was determined by normalizing samples to 18S RNA (ΔΔCT). Expression of SARS-CoV-2 nucleocapsid in the lung was determines using the 2019-nCOV_N1 primers, and probes were as follows: Forward primer: GACCCC AAA ATC AGC GAA AT, Reverse primer: TCT GGT TAC TGC CAG TTG AAT CTG, Probe: AC CCC GCA TT ACG TTT GGT GGA CC (CDC N1 qRT-PCR assay) [84]. Comparison was to a CoV-2 standard curve RNA produced by PCR amplification of SARS-CoV-2 nucleocapsid by which a 5’ T7 polymerase promoter was introduced. This amplicon was used as template to generate in vitro transcribed RNA, which was then quantified and serially diluted (10⁸–10¹ copies/μl).

Droplet digital PCR (ddPCR)

RNA and cDNA were prepared as described above. ddPCR reactions were performed with QX200 Droplet Digital PCR System (Bio-Rad, Hercules, CA) according to the manufacturer’s instructions. Each FAM reaction mixture (20 μL) contained 10 μL ddPCR Supermix for Probes (No dUTP) (Bio-Rad), 9 μL of template cDNA (up 100 ng), and 1 μL of Primers/Probes (900 nM per primer and 250 nM probe). And SYBR Green 20 μL reaction mixtures contained 10 μL iTaq Universal SYBR Green Supermix (Bio-Rad), 1 μL of primers (250 nM each primer), and 9 μL of template cDNA (up to 100 ng). The reaction mixtures were loaded into a disposable droplet generator cartridge (Bio-Rad), and droplets were formed using Bio-Rad QX-100 emulsification device. The contents were transferred to a 96-well Eppendorf reaction plate (Genesee Scientific; Morrisville, NC) and sealed with foil using Eppendorf 96-well heat sealer. PCR amplification of the droplets was performed using a C1000 Touch Thermal Cycler (Bio-Rad) with the following parameters: 95°C for 10 min, followed by 40 cycles of 94°C for 30 sec and 60°C for 1 min, and a final 98°C for 10 min. After PCR amplification, the plate was scanned using a QX200 Droplet Reader (Bio-Rad). QX Manager Software (Bio-Rad) was used to analyze the data by calculating the absolute copy number of the target DNA (units of copies/μL) using Poisson distribution analysis. All TaqMan gene expression probes were purchased from Applied Biosystems. Primers for SYBR reactions were designed using the NCBI Primer Designing Tool.

ACE2 activity in tissues, urine and bronchial lavage fluid

ACE2 activity in tissue lysates was measured using specific fluorogenic ACE2 substrate (Mca-APK-(Dnp) (AnaSpec, San Jose, CA) in the presence or absence of the ACE2 inhibitor (MLN-4760) (Sigma-Aldrich, St. Louis, MO) as previously described [85]. Tissue samples were homogenized in lysis buffer (75 mM Tris-HCl, pH 7.5, 1 M NaCl, 0.5 mM ZnCl2, 0.01 mM Captopril, 0.1 mM Z-Pro-Prolinal, 1mM PMSF, EDTA-free inhibitor cocktail tablet from Roche, and 0.5% Triton X-100) and centrifuged at 14,000 x g for 10 minutes at 4°C. Protein concentration in tissue lysates was measured using the Bradford method. Tissue lysates (10 μg of protein for kidney extracts and 40 μg of protein for heart, lung, trachea, and sinus extracts) were pre incubated with 70 μL of assay buffer (75 mM Tris-HCl, pH 7.5, 1 M NaCl, 0.5 mM ZnCl2, 0.01 mM Captopril, 0.1 mM Z-Pro-Prolinal, and EDTA-free inhibitor cocktail tablet from Roche) with or without ACE inhibitor MLN-4760 (10 μM final) for 30 minutes at room temperature. After the incubation with ACE2 inhibitor, 30 μL of ACE2 substrate buffer (75 mM Tris-HCl, pH 7.5, 1 M NaCl, 0.5 mM ZnCl2, 0.01 mM Captopril, 0.1 mM Z-Pro-Prolinal, and 0.167 mM Mca-APK-Dpn) was added to each well to initiate the reaction. Samples were incubated in the dark for 1 hour at room temperature, and fluorescence values were measured at an excitation wavelength of 320 nm and emission wavelength of 420 nm using a BioTek Cytation5 plate reader (BioTek instruments, Winooski, VT). Results were expressed as ΔRFU (Relative Fluorescence Unit) after subtraction of RFU values obtained in the presence of MLN-4760. BALF was collected from mice as previously described, and urine was collected during necropsy.

Naphthalene treatment

Mice received a single intraperitoneal injection of naphthalene (Sigma-Aldrich, St. Louis, MO) solution (200 mg/kg) freshly prepared in corn oil. As controls, mice were injected with vehicle only. Mice were euthanized at 24 or 48 hours after the injections. Lungs were then surgically removed, snap frozen, and stored at -80°C for subsequent RNA extraction.

Virus strains

The virus strains icSARS-CoV-2 WT [86] and SARS-CoV-2 WT stocks were grown using Vero E6 cells and titered via plaque assay. Vero E6 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, GIBCO), 5% Fetal Clone II serum (Hyclone), and 1X antibiotic/antimycotic (GIBCO). Briefly, serially diluted virus was added to a monolayer of Vero E6 cells and overlayed with media containing 0.8% agarose. Plaques were counted after three days after visualization by staining with Neutral Red dye. All viral infections were carried out under biosafety level 3 (BSL-3) conditions under negative pressure. All personnel conducting viral experiments wore Tyvek suits equipped with personal powered air-purifying respirators.

In Vivo infection

All mice were bred at UNC at Chapel Hill. Anesthetized (ketamine/xylazine) mice were intranasally infected with 1 x 10⁵ PFU of SARS-CoV-2, whereas mock infected mice received only PBS. Mice were monitored daily for any weight loss or decrease in lung function. For determination of viral load and RNA analysis, samples were collected at indicated time points after euthanizing mice by isoflurane overdose.

Histology

After euthanasia, the left lung lobe was harvested inflated with 10% phosphate buffered formalin and then further fixed by submersion in fixative for 7 days. Lung lobes were embedded in paraffin and sectioned at 3 μm thickness. Sequential sections were stained with hematoxylin and eosin.

Quantification and statistical analysis

Build-in functions of GraphPad Prism were used for data analyses and visualization. Specific statistical tests as well as numbers of animals are included in respective figure legends and detailed in S1–S3 Tables. These include the number of mice in each group and comparisons between all mice of each sex and genotype.