Severe Acute Respiratory Syndrome Coronavirus 2 Variant Infection Dynamics and Pathogenesis in Transgenic K18-hACE2 and Inbred Immunocompetent C57BL/6J Mice
, Talia S. Wong 1, Christopher M. Weiss 1,†, Kevin C. K. Lloyd 3,4, Qizhi Gong 2 and Lark L. Coffey 1,*
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis manuscript by Liu and colleagues gives an in-depth description of infection of K18 and C57BL/6J mice with multiple SARS-CoV-2 variants of concern (VOC). Clinical disease, viral titers, and histopathological changes in multiple tissues are thoroughly described. While presentation of the data is very thorough, and the overall scientific premise and experimental approach are sound, SARS-CoV-2 VOC infection in these mouse strains has already been published multiple times in various forms previously. What novel information this manuscript adds to the field is unclear.
Additional minor comments are as follows:
1. Figures 7 and 8: Visualization of SARS-CoV-2 antigen (in red) is difficult to discern in the images.
Author Response
We thank the reviewers for their helpful comments. In addition to modifying the text, we addressed major concerns by providing higher quality versions of figures. Although the figures inserted into the word file for review may appear fuzzy, each uploaded Figure on the journal website is high quality. We modified all figures containing histology images to increase the size of each panel for easier viewing. We added a new Figure (#10) that shows cell types in the brain co-localizing with SARS-CoV-2 and we added higher magnification images as requested to enhance viewing. We added many references and improved the discussion by including more information on related work from others. Please find a point by point rebuttal in the attached file.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe present study assessed K18 and C57Bl6 mice infected with different SARS-CoV-2 variants. It is a retrospective study that focusses on lungs and brain. It is valuable and in line with the 3R principles that the authors made use of material from studies undertaken for other purposes, to attempt a comparative assessment in particular of the type and extent of the pathological changes induced by different viral isolates in K18 and WT mice. However, the study falls unnecessarily short in several aspects. While making use of a large variety of virus isolates and a large number of animals, it does not provide new and/or in-depth information. Indeed, many other publications describing murine infectious with various SARS-CoV-2 isolates have, individually, provided more detailed information than the authors offer here. A look at the list of references shows that several of these have not been considered by the authors.
A particular shortfall is that the in situ comparison of the different infection models has apparently taken place at a very superficial level (viral NP expression is descibed as “in the lung” and “in the brain” which is not appropriate considering the IF and RNA-ISH efforts that were made). This does not need to be the case, as the material is available. Hence, the authors are encouraged to re-examine their slides and determine which cells express viral NP, and to substantiate the expression of viral antigen in leukocytes which the authors indicate, as well as which cells express the cytokine mRNA.
Another issue is that the authors repeatedly claim that K18 mice “express hACE2 in the brain – a feature not present in humans” (line 38), show “non-physiological expression of human ACE2 in the brain” (line 98) or “non-physiological expression of human ACE2 in extrapulmonary tissues” (line 625/626). As the only reference for these statement they offer their own previous paper in which they also speculate that “non-physiological expression of hACE2 in the brain” is the cause for brain infection in K18 mice. However, they do not reference a paper that has provided evidence of ACE2 protein expression in specific cells in the brain in K18 mice (which would be the prerequisite of its receptor function.
Specific items are listed below in a consecutive order:
Materials and Methods
Line 205: “embedding for histology” should be specified. Do the authors refer to the classical paraffin embedding?
Line 224 onwards: The authors describe at length how the pathologist approached the histopathological analysis. However, they do not state anywhere on which type of samples this was undertaken. These were presumably HE stained sections prepared from FFPE tissue specimens; this information needs to be provided.
Lines 233-241: The brain scoring is very difficult to follow. Are lines 233-235 and lines 239-241 not some form of repetition? And where do the additional 7 points for the unique lung changes come from that are mentioned in the results section (line 197/198)? This is not clear in the description.
Line 244: What does “sections” imply here?
Line 245/246: Here, the authors speak about decalicification of brain and lung sections. This is odd because neither brain nor lung can be considered calcified tissues…
Table 2: As an antigen “NP” should be defined a bit better… Also, it should be added to the table (else specified in the text) which staining protocols required antigen retrieval. Similarly, primary antibody incubation conditions (length, temperature) should be provided.
Results
Lines 281-296: Most of this text is already or should be in M&M (like the calculation of mouse numbers to allow statistical analysis, and if not it would rather belong there…
Line 562: Replace “immunohistochemistry” by “immunofluorescence”. This should be changed throughout the manuscript.
Lines 565-566: The authors describe NP expression in the lung in a very superficial way. They do not provide information as to the cell types that are infected; “lung” is not sufficient – are these pneumocytes? And if so, which type(s)? It can be expected that it is not too easy to provide good pictures at high magnification of infected pneumocytes in 14 µm thick sections but one can certainly be more specific (i.e. high mag insets) in Fig. 7 than the overview images allow. The NP expression pattern would, of course, be interesting to compare at cell type level for K18 and C57 mice.
Lines 582-588: Similar to the lungs, here the authors describe NP expression in the brain in a very superficial way. They do not provide information as to the cell type(s) that are infected. Such detail should be provided, its lack reduces the value of the current work.
Line 609: See comments above. Here the authors claim “diffuse staining” for NP “throughout the lung tissue”. This is insufficient information. Many other publications describing murine infectious with various SARS-CoV-2 isolates have provided more detailed information than the authors offer here…
Line 611/612: See comments above. The claim of “NP signals co-localizing with a subset of all immune cell types examined” is confusing/misleading. Do the authors claim that neutrophils, T cells and/or macrophages carry viral NP? And all this in addition to the “diffuse staining throughout the lung tissue”? This part of the results should be rewritten and much more detailed information be provided. See also the comment on Fig. 10.
Line 615/616: The authors claim that “CCL5 and IL-6 co-localized with SARS-CoV-2 infected cells…” but appear not to make any attempt at identifying/naming the cell types.
Discussion
Line 645: Replace “pathology” by “histopathological changes”.
Line 673-675: Unfortunately, to make this statement valid, the authors would need to specifically mention the cell types that they found to express viral NP (see comments above).
Line 675-677: The discussion of cytokine expression is too superficial. Which cell types are involved? What is the effect of expression etc.?
Line 684-685: The authors state that by not examining the brains of Omicron infected mice by histology (because they did not detect virus in the brain) they might have missed “indirect effects of SARS-CoV-2 infection on brain damage that may occur absent neuroinvasion”. This wording is a bit odd but the reader gets the idea. However, considering that in the present manuscript the authors did not even comment on the type of cells in the brain that were virus infected with the other variants, the reader would probably consider the above-mentioned argument as not very valid…
Line 687-691: There is literature to address (or provide information for) the line of thoughts regarding the distribution of histological lesions (at least on the distribution of perivascular infiltrates in infected brains). This should be cited. However, the question is what the authors consider as “histological lesions” in this context: these should be defined.
Line 693-695: Do mice with widespread SARS-CoV-2 infection of the brain show “neurological disease”/”clinical symptoms”/”neurological signs”? There is sufficient literature on studies in K18 mice infected with several virus variants at the same dose (104 PFU) that only reports weight loss (and euthanasia/clinical endpoint due to weight loss) but no clinical neurological signs in K18 mice. This part of the discussion should be revised.
Line 695-698: There are ample studies that have shown that K18 mice infected with SARS-CoV-2 variants that lead to widespread brain infection reach clinical endpoint by 6-7 dpi; hence one could not “follow infected mice over longer time period”; this should be considered in the discussion.
Line 707: Why would the C57Bl6 model be more appropriate to study long COVID?
Figures:
Figure 1: The graphs and images are not sufficiently in focus the way they are currently presented. The figure legend states that all mice received 104 PFU of the different virus isolates. However, in M&M (line 179) it is said that either 104 PFU or 105 PFU were used. This should be clarified.
Figures 3, 4 and 6: The images provided are very small. Since a focus of the work is on the histopathological changes, appropriately sized representative images should be shown. Also, why does the figure legend of Fig. 3, in line 414-416, include information on the changes in the brain without at least referring to Fig. 4? Also, the term “hypercellularity” is too vague in this context.
Figure 4: The neuronal necrosis in F is not convincing.
Figure 5: The graphs and images are not sufficiently in focus the way they are currently presented.
Figure 10: The legend should state what is shown in the top and bottom layer for each marker panel (mock infected controls included?). Also, to convincingly show or exclude co-localization of NP and the leukocyte markers, higher magnifications would be essential.
Figure 11: Higher magnifications would be essential to convincingly show co-localisation of signals for virus and cytokine (m)RNA.
Supplemental Table S1: It would be good if the authors would also add information as to the direction of the body weights (upp or down) that differed significantly (even if this is obvious in Fig. 1B).
Comments for author File: 
Comments.docx
Author Response
We thank the reviewers for their helpful comments. In addition to modifying the text, we addressed major concerns by providing higher quality versions of figures. Although the figures inserted into the word file for review may appear fuzzy, each uploaded Figure on the journal website is high quality. We modified all figures containing histology images to increase the size of each panel for easier viewing. We added a new Figure (#10) that shows cell types in the brain co-localizing with SARS-CoV-2 and we added higher magnification images as requested to enhance viewing. We added many references and improved the discussion by including more information on related work from others. Please find a point by point rebuttal in the attached file.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors present a thorough demonstration of the differences in infectivity and symptoms of several SARS-Cov-2 variants in two mouse susceptible backgrounds, in addition to defining the utility of the K18 mouse for studying severe SARS-CoV-2 infections and characteristics of variant-specific virulence versus the non-lethal C57BL/6J model for studying immune responses, the dynamic of infection among specific infective variants, and perhaps a model that reflects long-COVID. Clearly, this study provides valuable criteria of these models that should improve vaccine development programs in addition to relevant therapeutic development for COVID and long-COVID. The models would also appear to facilitate the assessment of the mechanism of infectivity and virulence of new VOCs that arise in the future.
The manuscript is extremely well written in its over-all structure and with organization of ideas and conclusions, including a very proactive a thorough "limitations" explanation in the Discussion. I found the presentation of results and statistical analyses to be very clear and solid that very well support the text narrative.
In conclusion, it is rare to find such a solid and impactful manuscript upon first submission, of which it is difficult to find any concerns, even typos.
Author Response
We thank the reviewers for their helpful comments. In addition to modifying the text, we addressed major concerns by providing higher quality versions of figures. Although the figures inserted into the word file for review may appear fuzzy, each uploaded Figure on the journal website is high quality. We modified all figures containing histology images to increase the size of each panel for easier viewing. We added a new Figure (#10) that shows cell types in the brain co-localizing with SARS-CoV-2 and we added higher magnification images as requested to enhance viewing. We added many references and improved the discussion by including more information on related work from others. Please find a point by point rebuttal in the attached file.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThis revised manuscript by Liu and colleagues provides a thorough characterization of SARS-CoV-2-induced disease in K18 and B6 mice. The authors have done a commendable job evaluating multiple SARS-CoV-2 variants of concern (VOC) in K18 mice and mouse adapted (MA) SARS-CoV-2 in C57BL/6 mice. However, concerns regarding the extent to which this information adds to the field remain. Additional minor comments are as follows:
1. The resolution of most of the figures (Figs 1, 2, 5, 6, 7, 8, 9, 10, 11, and 12) remains poor. The figures should be replaced with high-quality images so that the reader can accurately evaluate the findings.
2. With the exception of Figures 3E and 3F ("4 days" panels) and Figures 4F and G, all of the H&E images need to undergo white balancing.
Author Response
Please see attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe manuscript has significantly improved in quality with its revision and now represents a valuable contribution to the literature on COVID-19 animal models. Thank you. Given that the authors have the knowledge, skills and exertise, one is surprised why they did not make the additional effort in the first place; it would have saved all of us quite a bit of time …
There seems to be some lack of transparency regarding examination of the brains. The authors apparently collected the left hemisphere for PCR analysis (line 274) but then do not specify which part(s) of the brain was/were fixed in formalin (line 282: “tissues”) and processed for histology (line 306: “brain tissue”). However, in the results, the paragraph on viral antigen expression in the brain (line 743 onwards) does only mention the cortex (and refers to coronal sections later, in the figure legends). This reviewer considers it as likely that the entire right hemisphere was paraffin wax embedded and initially examined histologically and by IF for viral antigen expression to determine whether viral antigen was expressed, i.e. to complement the PCR results. Therefore, somewhere, information on the presence and distribution of viral antigen in the brain should be provided – and put into context with the PCR results (instead of the cortex alone). This would also reduce the mentioned limitations of the study
There a few remaining issues. These are listed below in a consecutive order:
Introduction
Line 119: Add “viral” after “active”.
Line 165: Delete the “d” in “passaged”.
Line 191: Replace “increased histopathological disease severity” by more appropriate wording, such as “more severe histopathological changes” or similar.
Results
Lines 748: When stating that representative patterns are shown, a reference to the figures should be included.
Line 753/754: Reword “increased in intensity” and “the highest NP signal” as this implies a degree of signal intensity in individual cells rather than the number of positive cells which the authors appear to refer to.
Line 755: Specify what is meant with “only rare NO staining” (presumably that there were only few positive neurons).
Line 757: The authors claim that “microglia are recruited to infected neurons” and they they show this in Figure 10. This statement is not proven. The image shows Iba1+ microglial cells with the morphology of activated microglial cells. This would be consistent with microglial activation but does not prove recruitment. The text needs to be amended.
Line 779: Add “mouse” after K18.
Line 780: Add “variant” after beta (or reword in another appropriate way).
Line 785: The sentence “The infection pattern is consistent NP staining in the lung.” needs amending. NP staining suggest a target cell spectrum (or infection pattern if the authors prefer) but the sentence does not make sense as it is now.
Line 788/789: The sentence “Arrowheads point to example immune cells containing NP protein within the cell.” lacks a clear reference to a figure. Please add. Also, delete “protein” as the P stands for protein anyway.
Line 791-793: This sentence is odd. The author state that they observed “qualitative increases” “which were more pronounced” but then do not comment on what the refer to. Suggest to reword so that this bit makes more sense to the reader.
Discussion
Line 866/867: The statement “show non-physiological expression of human ACE2 in extrapulmonary tissues including brain.” is too superficial considering the references provided. In particular the paper by Carossino and co-authors in Viruses, 2022 (doi:10.3390/v14030535) emphasised on page 25 that hACE2 mRNA expression in the absence of protein expression indicate that hACE2 is not solely responsible for neuroinvasion in K18 mice. The other cited publication (Dong et al., 2022; doi:10.1128/JVI.00964-21) has used an antibody against ACE2 that they state to react with hACE2 (without proof that it does not cross-react with mACE2); the authors claim “expression of hACE2 in the brain” without specifying which cells exhibit expression, and showing an overview image with a slight diffuse brownish staining of some part of a mouse brain at low magnification (Fig. 5B) which cannot be considered as convincing in light of the paper by Carossino and co-authors. Therefore, if the authors want to keep reference to ACE2 expression in their discussion, it should definitiely consider the flaws of reference 41 and better refer to reference 40 which far better reflects what other researchers have observed in the brain of K18 mice when staining for ACE2.
Line 919: Replace “pathology” by more appropriate wording.
Line 923: S.th.is missing between “of” and “are”…
Line 931: Change “of cytokines” to “on cytokine transcription”
Line 934: The wording “artifacts of the histology approach” should be altered. Specify what is referred to here; the term artifacts does not seem to be appropriate here. A discordance of results could be due to differences in virus distribution in the lung tissue subjected to PCR and to histology, respectively…
Line 946: Which “histological lesions” do the authors refer to here? If it is a perivasular infiltrate, then this will likely be true, since the inflammatory response is often restricted to the brainstem (there are references for this). Therefore, a limitation of the present study might be that it only examined the cortex instead of the whole brain of the mice…
Line 957/958. The authors claim “significant individual differences in susceptibility and response to SARS-CoV-2” as the reason for within-group differences. What about difference in the amount of virus that reaches the lung or the nose despite using the same inoculation dose?
Figures:
Figures 3: In the legend, please reword “Representative lung histopathology” by a more appropriate wording, such as representative images of the histopathological features”. Since the previous round of review appears not to have made it clear that the term “pathology” should be used appropriately, the authors are advised to please check the meaning of the term (a simple definition is: “the science of the causes and effects of diseases”)…
Figure 4: Thanks for higher magnifications of the images to illustrate what is interpreted by the authors as neuronal necrosis.
Author Response
Please see attachment.
Author Response File: Author Response.pdf
