Previous infection with seasonal coronaviruses does not protect male Syrian hamsters from challenge with SARS-CoV-2

SARS-CoV-2 variants and seasonal coronaviruses continue to cause disease and coronaviruses in the animal reservoir pose a constant spillover threat. Importantly, understanding of how previous infection may influence future exposures, especially in the context of seasonal coronaviruses and SARS-CoV-2 variants, is still limited. Here we adopted a step-wise experimental approach to examine the primary immune response and subsequent immune recall toward antigenically distinct coronaviruses using male Syrian hamsters. Hamsters were initially inoculated with seasonal coronaviruses (HCoV-NL63, HCoV-229E, or HCoV-OC43), or SARS-CoV-2 pango B lineage virus, then challenged with SARS-CoV-2 pango B lineage virus, or SARS-CoV-2 variants Beta or Omicron. Although infection with seasonal coronaviruses offered little protection against SARS-CoV-2 challenge, HCoV-NL63-infected animals had an increase of the previously elicited HCoV-NL63-specific neutralizing antibodies during challenge with SARS-CoV-2. On the other hand, primary infection with HCoV-OC43 induced distinct T cell gene signatures. Gene expression profiling indicated interferon responses and germinal center reactions to be induced during more similar primary infection-challenge combinations while signatures of increased inflammation as well as suppression of the antiviral response were observed following antigenically distant viral challenges. This work characterizes and analyzes seasonal coronaviruses effect on SARS-CoV-2 secondary infection and the findings are important for pan-coronavirus vaccine design.


Introduction
• "Immunological imprinting, previously known only for its negative effects" -not true, imprinting of innate immunity has been shown to be able to contribute to protection • "imprinting has been most associated with sequential influenza virus infection" -not most associated, but like "best studied" • "past spillover coronaviruses" -not sure if all this evolutionary and epidemiological info is needed for understand; also very long complicated sentences in first paragraph of intro • "immunologically dominant surface Spike (S) protein" -S may be dominant protective antigen in vaccines, but not sure that is also immunologically dominant in natural infection; infections seems to mount B-and T-cell immunity against many other viral proteins such as N; to our knowledge dominance of S over other antigens has never been demonstrated • "emergence of more immune evasive variants" -emergence of variants evading (or escaping) immunity • "represented more closely related virus exposures" -more importantly also represent what happened recently to large parts of the human population; subsequent exposure by prototype and/or beta, then omicron

• Results
• "available for this project" -availability should not be relevant, rather rationale • "Previously published studies of seasonal coronavirus infection mainly leveraged the mouse model while studies of SARS-CoV-2 and variant infections by our group and others have used Syrian hamsters to assess pathogenesis, host responses, and vaccines 30-36.Since Syrian hamsters are outbred, we hypothesized these animals to be a more representative model of human genetic variability for our study.To give insight into the relatedness of the human coronaviruses and the potential of using hamsters as a single animal species for this work, we first conducted analysis on the S protein sequences for the coronaviruses of interest and the corresponding Syrian hamster sequences of the specific viral host cell receptors (Fig. 1A)." -please condense • "This viral sequence analysis suggested that the seasonal coronaviruses were antigenically distant to the SARS-CoV-2 Wuhan and VOC viruses which were more similar to each other."-trivial, please condense • "We were not able to acquire HKU1 for this study, therefore the susceptibility to HKU1 was not investigated."-delete • "Three groups of Syrian hamsters were intranasally inoculated" -please provide group size n, age and sex • "Although live virus was not identified in the lungs, histopathological analysis of lung tissue indicated immune cell infiltration and mild pathology."-please highlight some charatceristic pathological changes in the histopathological pictures in Figure -1D,E e.g. by arrows or other symbols.Reporting lung weight may not be sufficient.Adding histological scores may enhance the data.
• "To understand the immune responses in the respiratory tissues of NL63 and OC43 inoculated animals, we assessed host gene expression of select immune genes via qPCR.In the nasal turbinates (Extended Data Fig. 3)" -these key data on differential expression profiles should be considered as main figure instead of supplementary • "we went on to determine the impact and mechanisms of seasonal coronavirus imprinting on a SARS-CoV-2 Wuhan secondary challenge 26-28 -these cited references are reporting conflicting results; please mention that this is to be experimentally assessed and unveiled • "were imprinted with" -see above; were infected with to possibly imprint • "nadir" -not commonly used • "psc" -abbreviation not intraodroduced before • "remained statistically above" -somewhat weird expression • "Histopathological analysis of the lungs after secondary Wuhan challenge showed characteristic interstitial pneumonia, hemorrhaging, and mononuclear cell infiltration into the alveolar space peaking on day 6 psc of the mock-Wuhan group which was similar to our previous study of Wuhan infected younger Syrian hamsters 34.Both the NL63-Wuhan and OC43-Wuhan groups had significantly less evidence of interstitial pneumonia in the lungs following Wuhan challenge throughout the time course."-these data seem not to be presented • "during Wuhan challenge, the NL63-Wuhan group had a significant increase or a back-boosting effect of NL63 neutralizing antibodies following Wuhan exposure with titers increasing above 1:320 by day 14 psc" -very small group size and relatively small changes may not allow to draw these conclusions • "However, both seasonal virus imprinted groups had trends of higher neutralizing antibody titers against the antigenically distant variants of Delta, Beta, and Omicron compared to the mock primary group."-too small groups and no proper statistics done to support this conclusion • "Imprinting with Wuhan protects against secondary challenge with antigenically divergent SARS-CoV-2 variants (…) These results suggest Beta and Wuhan to be antigenically closer to each other than to Omicron and that the heterologous exposure of Wuhan followed by Beta may be an optimal antigenic distance to induce a greater breadth of antibody cross-reactivity."-these data have been presented by others before and may not be result of imprinting but need to be rather called cross-protection amongst SARS-CoV-2 variants.
• Entire section "Host response analysis during secondary challenge indicates immune regulation is based on antigenic distance" is very interesting and may be considered key of the entire study, however proper cluster analysis of groups and statistics lacking in Figure 6 are needed.
• "More predicted T cell epitopes on OC43 shared with SARS-CoV-2 Wuhan" -this section is irrelevant for the hamster model as no MHC-I nor II predictions are possibly in lack of data for this model.Modelling humans HLA-restricted peptides does not support finding in the previous key paragraph.
The discussion should be compacted and focus on the proposed interference between primary and secondary infections.
Reviewer #2 (Remarks to the Author): In their manuscript titled, "Signatures of Original Antigenic Sin or protection following coronavirus imprinting in Syrian hamsters challenged with SARS-CoV-2 or variants", Francis et al characterize the impact of prior infection with seasonal coronaviruses on infection, pathogenesis, and immune responses to heterologous zoonotic coronavirus secondary infections.In doing so, they identify differential impacts of NL63 and OC43 primary infections on virological, disease, and immune-response outcomes to CoV2 secondary infections.Interestingly, the differential disease outcomes appeared to correlate with backboosting of NL63 neutralizing antibody responses and T-cell recall responses to OC43, with the former resulting in poor control of, and the latter improving host response to, secondary exposure.The methods and study designs were appropriate, and the results support the major conclusions.A few minor notes described below.
Figure 3 describes the neutralizing antibody responses to both the seasonal and zoonotic coronaviruses, with panel A providing an overview of the kinetics of response, only to the primary infecting virus, throughout the entire study while panel B provides a snapshot of the days 9 and 14 post-secondary infection, only to the zoonotic secondary infecting virus and its variants.To gain a better understanding of the recall responses to the secondary infection, it would improve the interpretation of the study if the authors could include either a reproduction of panel Ai and Aii against Wuhan virus or just the day 0 post-secondary challenge in panel B so that fold-change nAb responses can be calculated for NL63, OC43, and Wuhan.This would allow readers to interpret the relative focusing of the recall responses between the primary and secondary infecting viruses.While panel Ai clearly shows back-boosting of NL63 responses, it is unclear how that boosting response compares to the Wuhan nAb response in those same animals since only day 9 and 14 responses are shown in panel B. Also in figure 3, the homologous wuhan-wuhan sera show remarkable breadth of nAb responses against VoCs with only a 2-fold drop in nAb titers against Omicron and even a potential increased titer against Beta.This is surprising given observations by others showing much larger differences in heterologous nAb titers between Wuhan and Omicron/Beta VoCs.A discussion on this with some potential explanations would be helpful.
Finally, this study only evaluated recall responses over a short period following primary infection (56 days) with minimal waning of primary nAb responses prior to secondary infection.A discussion on this limitation and impact on conclusions is warranted.
Reviewer #3 (Remarks to the Author): In this manuscript, Francis and colleagues use the Syrian hamster model to assess the relative role previous infection with seasonal coronaviruses contributes to subsequent infection and host response elicitation with SARS-CoV-2 virus.Authors conduct primary challenge experiments with several seasonal coronaviruses, followed by rechallenge with an early SARS-CoV-2 virus, or primary challenge with the early Wuhan virus followed by rechallenge with different variants of concern.A range of virological and host response parameters were conducted.Collectively the authors find that primary infection with seasonal coronaviruses did not confer robust protection against SARS-CoV-2 virus, but do identify subtle differences in host responses using several parameters, providing information towards understanding the diversity of responses possible when assessing correlates of protection towards coronaviruses moving forward.The manuscript is well-written, well-referenced, logically organized, and presents a clear rationale for experiments performed.There are some areas that would nonetheless benefit from additional clarity and justification as discussed below.

Comments:
-Page 4, 2nd paragraph of introduction, it would be helpful to the reader if you mentioned what relative percentage of the population had been infected with seasonal coronaviruses prior to the start of the SARS-CoV-2 pandemic, to better contextualize why preclinical models with preexisting antibody to seasonal CoV is important to emulate.
-Page 7, it is understandable (and not a detriment to the study) that HKU1 could not be included.However, can the authors speculate (here or in the discussion) how this virus might have behaved in the analyses included (e.g. based on the spike protein identity analysis shown in Fig 1 ) if this virus would have behaved relative to the other seasonal coronaviruses examined?-In rechallenge experiments, animals were rested from day 14 p.i. to day 56 p.i., at which point secondary challenge took place.Why was this interval chosen and do the authors think that a longer duration of resting between challenges would have modulated the rechallenge results?-Page 12, please specify which text maps to Figure 2D and 2E; it is currently unclear how the lung weight differences presented in Figure 2E align with the text as currently written in this section.
-Page 14, authors state that both seasonal virus imprinted groups had "trends of higher neutralizing antibody titers" against SARS-CoV-2 variants.However, the data shown in Figure 3B (to my eye) shows generally comparable antibody titers among all mock, NL63, and OC43 primary-challenged animals; are the authors drawing from additional data to make this statement, or comparing mean values that can be reported which reflect this trend?As currently presented in the graphs, without additional supporting information, I think the higher "trends" alluded to here are a bit spurious.
-page 16/Figure 5, the current graphs as presented make it a bit challenging for the reader to compare trends present from specific groups of animals that are present between different assessments of neutralizing antibodies (the graph are designed for, and effectively show, augmentation of neutralizing antibody titers against a single variant over time in different treated groups of animals).If the authors want to state, as they do in the text, that specific animal groups displayed trends across multiple assays shown in Figure 5, an additional panel in this figure (e.g.where the legend is currently present), showing comparative levels of one timepoint assessed for the presence of different neutralizing antibodies between different groups, would make it much easier for the reader to quickly see the data the authors are alluding to in making this statement.
-Figure 6, please specify what fold change is in relation to (mock animals?) and what blue in the scale bar refers to (values with a negative fold-change relative to mock?).Also please specify what normalization (if any) took place in these specimens; authors show in earlier figures that lungs had different weights so I presume some sort of normalization to a housekeeping gene was conducted?-Figure 1C, the coloration between live virus and qPCR in these graphs is very, very subtle; recommend a different differentiation strategy (e.g.colors with greater separation, or filled vs open symbols) to show this difference.Also please clarify why the LOD for these data is different than in figure 2C."inflamma5on and an5viral response suppression" to "increased inflamma5on as well as suppression of the an5viral response were signatures in an5genically distant viral challenge" on line 49 and 50.
Reviewer 1, Comment 9: "These results are important for calcula5ng the an5genic distance between coronaviruses and for determining posi5ve or nega5ve cross-reac5ve signatures."-one means of assessing distances complementary to phylogene5c, an5genic cartography etc.
Response to Reviewer 1, Comment 9: The sentence has now been modified to read "These results are important for assessing an5genic distances between coronaviruses complementary to phylogene5cs and an5genic cartography" on lines 52 to 53.
Reviewer 1, Comment 10: "matrix of protec5on" -not a generally used term, thus not unambiguous what meant here Response to Reviewer 1, Comment 10: We have removed this sentence.

Introduc5on:
Reviewer 1, Comment 11: "Immunological imprin5ng, previously known only for its nega5ve effects" -not true, imprin5ng of innate immunity has been shown to be able to contribute to protec5on Response to Reviewer 1, Comment 11: This sentence has been modified to "Immunological imprin5ng, previously described as 'Original An+genic Sin', is the powerful mark, posi5ve or nega5ve, that a first viral infec5on makes on the host's immune system and response to subsequent an5genically related viral exposures."On lines 57 to 59.
Reviewer 1, Comment 12: "imprin5ng has been most associated with sequen5al influenza virus infec5on" -not most associated, but like "best studied" Response to Reviewer 1, Comment 12: We appreciate this discrepancy.We have corrected "imprin5ng has been most associated with sequen5al influenza virus infec5on" to "imprin5ng has been best studied in the context of sequen5al influenza virus infec5on" on lines 61 and 62.
Reviewer 1, Comment 13: "past spillover coronaviruses" -not sure if all this evolu5onary and epidemiological info is needed for understand; also very long complicated sentences in first paragraph of intro Response to Reviewer 1, Comment 13: The sentence has now been modified to "Human coronaviruses (HCoV) belong to either the alphacoronavirus or betacoronavirus genera sugges5ng these genera to be of human concern" on lines 72 to 74.
Reviewer 1, Comment 14: "immunologically dominant surface Spike (S) protein" -S may be dominant protec5ve an5gen in vaccines, but not sure that is also immunologically dominant in natural infec5on; infec5ons seems to mount B-and T-cell immunity against many other viral proteins such as N; to our knowledge dominance of S over other an5gens has never been demonstrated Response to Reviewer 1, Comment 14: We appreciate this comment.We have added three cita5ons demonstra5ng S as the immunologically dominant protein in the context of infec5on to line 74.
Reviewer 1, Comment 15: "emergence of more immune evasive variants" -emergence of variants evading (or escaping) immunity Response to Reviewer 1, Comment 15: At the reviewer's sugges5on, we have modified the sentence to "emergence of variants of concern (VOCs) evading immunity" on lines 93 and 94.
Reviewer 1, Comment 16: "represented more closely related virus exposures" -more importantly also represent what happened recently to large parts of the human popula5on; subsequent exposure by prototype and/or beta, then omicron Response to Reviewer 1, Comment 16: We have modified this sentence to include the sen5ment that these exposures were relevant to a large por5on of the current global popula5on.The sentence now reads "The results were then compared to imprin5ngchallenge studies performed with Ancestral SARS-CoV-2 virus and the VOCs, Beta and Omicron, as these combina5ons represented more closely related virus exposures as well as recent exposures affec5ng large popula5ons of people."On lines 106 to 109.

Results:
Reviewer 1, Comment 17: "available for this project" -availability should not be relevant, rather ra5onale Response to Reviewer 1, Comment 17: We appreciate this comment.We have removed HKU1 availability from the text.
Reviewer 1, Comment 18: "Previously published studies of seasonal coronavirus infec5on mainly leveraged the mouse model while studies of SARS-CoV-2 and variant infec5ons by our group and others have use Syrian hamsters to assess pathogenesis, host responses, and vaccines 30-36.Since Syrian hamsters are outbred, we hypothesized these animals to be a more representa5ve model of human gene5c variability for our study.To give insight into the relatedness of the human coronaviruses and the poten5al of using hamsters as a single animal species for this work, we first conducted analysis on the S protein sequences for the coronaviruses of interest and the corresponding Syrian hamster sequences of the specific viral host cell receptors (Fig. 1A)." -please condense Response to Reviewer 1, Comment 18: Thank you for the comment.We have condensed the sec5on to now read, "Previously published studies of seasonal coronavirus infec5on mainly leveraged the mouse model; however, we chose to use Syrian hamsters for our study of human coronaviruses due the u5lity it has shown in SARS-CoV-2 pathogenesis studies."On lines 120 to 123.
Reviewer 1, Comment 19: "This viral sequence analysis suggested that the seasonal coronaviruses were an5genically distant to the SARS-CoV-2 Wuhan and VOC viruses which were more similar to each other."-trivial, please condense Response to Reviewer 1, Comment 19: We have now revised this sentence to "This viral sequence analysis suggested that the seasonal coronaviruses were an5genically distant to the SARS-CoV-2 Ancestral strain while VOCs were more an5genically similar" on line 127 to 129.
Reviewer 1, Comment 20: "We were not able to acquire HKU1 for this study, therefore the suscep5bility to HKU1 was not inves5gated.-delete Response to Reviewer 1, Comment 20: "We were not able to acquire HKU1 for this study, therefore the suscep5bility to HKU1 was not inves5gated."has been deleted from the text.
Reviewer 1, Comment 21: "Three groups of Syrian hamsters were intranasally inoculated" -please provide group size n, age and sex Response to Reviewer 1, Comment 21: We apologize for not including this informa5on."Male hamsters, aged 8-weeks were u5lized.A total of 12 animals were used in each group (three animals sampled on days 3 and 6 post inocula5on and six animals sampled on day 14 post inocula5on)."Has been added to lines 143 to 145; "A total of 18 male Syrian hamsters were used in each group (three animals sampled on days 0, 3, 6 and 9 post secondary challenge and six animals sampled on day 14 post secondary challenge)."Similar text has been added in appropriate places throughout the manuscript.
Reviewer 1, Comment 22: "Although live virus was not iden5fied in the lungs, histopathological analysis of lung 5ssue indicated immune cell infiltra5on and mild pathology."-please highlight some characteris5c pathological changes in the histopathological pictures in Figure -1D,E e.g. by arrows or other symbols.Repor5ng lung weight may not be sufficient.Adding histological scores may enhance the data.
Response to Reviewer 1, Comment 22: We appreciate this sugges5on.To support our histological data, we have included pathological scoring in the new Extended Data Fig. 2, 7 and 9.
The descrip5on of this data to support Fig. 1D and E now reads "Although live virus was not iden5fied in the lungs, histopathological analysis of lung 5ssue indicated immune cell infiltra5on and mild pathology.Mononuclear cells infiltrated the alveoli with associated inflamma5on on day 6 and 14 pi following NL63 inocula5on and small clusters of macrophage-like cells were observed las5ng un5l day 14 (Fig. 1D).The lungs of NL63 inoculated animals scored for inflamma5on peaking at an average of 1.3/4 on day 6 and somewhat resolving to .33/4 on average on day 14 (Extended Data Fig. 2).The OC43 inoculated animals had significant mononuclear cell infiltra5on (mainly macrophage-like cells and neutrophils) and hemorrhaging on day 6 pi.Edema was also noted around a group of blood vessels in one of three animals on day 14 post OC43 inocula5ons (Fig. 1D).The same OC43 inoculated animal scored a 2/4 on lung inflamma5on with less than 25% of the lung being affected (Propor5on score of 1) (Extended Data Fig. 2).Not surprisingly the lungs from the 229E inoculated animals, which did not show evidence of ac5ve viral infec5on even in the upper respiratory tract, retained a similar architecture compared to control uninfected 5ssue (Fig. 1D) and scored 0/4 on lung histology (Extended Data Fig. 2).The SARS-CoV-2 Ancestral virus inoculated animals displayed evidence of bronchopneumonia, significant peribronchiolar thickening, hemorrhaging, inters55al pneumonia, and mononuclear cell infiltra5on on day 6 pi (Fig. 1D) as we have previously reported 41 .Ancestral virus inoculated animals scored on all histology parameters, with over 50% of the lung being affected on day 6 and 14, and highest scoring in any category being on day 6 with an average score of 3 in bronchiole, corresponding to marked epithelial lesions (Extended Data Fig. 2)."On lines 172 to 191.
Reviewer 1, Comment 23: "To understand the immune responses in the respiratory 5ssues of NL63 and OC43 inoculated animals, we assessed host gene expression of select immune genes via qPCR.
In the nasal turbinates (Extended Data Fig. 3)" -these key data on differen5al expression profiles should be considered as main figure instead of supplementary Response to Reviewer 1, Comment 23: Thank you for this sugges5on.To address this, we have made an addi5onal main text figure, now Fig. 2 to include a succinct representa5on of the data from Extended Data Fig. 3 and 4 (Now Extended Data Fig. 4 and 5), which remain in the extended data, as represented on lines 206 to 219.
Reviewer 1, Comment 24: "we went on to determine the impact and mechanisms of seasonal coronavirus imprin5ng on a SARS-CoV-2 Wuhan secondary challenge 26-28 -these cited references are repor5ng conflic5ng results; please men5on that this is to be experimentally assessed and unveiled Response to Reviewer 1, Comment 24: Thank you for poin5ng this out.We have addressed that these references are conflic5ng with the addi5on of "given the conflic5ng results reported in human clinical studies 21,34,35 " on lines 229 and 230.
Reviewer 1, Comment 25: "were imprinted with" -see above; were infected with to possibly imprint Response to Reviewer 1, Comment 25: "were imprinted with" has been adjusted to "were intranasally inoculated with" on lines 232 and 233.
Reviewer 1, Comment 27: "psc" -abbrevia5on not intraodroduced before Response to Reviewer 1, Comment 27: The abbrevia5on is described in its first use on line 242.
Reviewer 1, Comment 29: "Histopathological analysis of the lungs aoer secondary Wuhan challenge showed characteris5c inters55al pneumonia, hemorrhaging, and mononuclear cell infiltra5on into the alveolar space peaking on day 6 psc of the mock-Wuhan group which was similar to our previous study of Wuhan infected younger Syrian hamsters 34.Both the NL63-Wuhan and OC43-Wuhan groups had significantly less evidence of inters55al pneumonia in the lungs following Wuhan challenge throughout the 5me course."-these data seem not to be presented Response to Reviewer 1, Comment 29: We apologize for this confusion.To support our histological data, we have included pathological scoring in the new Extended Data Fig. 7 and refined the conclusions made from the histology datasets.Fig. 2D (Now Fig. 3D) is described below.
The descrip5on of this data now reads "Histopathological analysis of the lungs aoer secondary Ancestral virus challenge showed characteris5c inters55al pneumonia, hemorrhaging, and mononuclear cell infiltra5on into the alveolar space peaking on day 6 psc of the Mock-Ancestral virus group (Fig. 3D) which was similar to our previous study of Ancestral virus infected younger Syrian hamsters 41 .Histology scoring correspondingly peaked on day 6 when there was an overall severity score of 3/4 on average and scoring of 1 or greater in all animals for inflamma5on, pneumonia, and bronchioles (Extended Data Fig. 7).NL63-Ancestral animals but not OC43-Ancestral animals had an increased pneumonia scoring in the lungs following Ancestral challenge throughout on day 6 (Extended Data Fig. 7).Of note, the NL63 primary group had marked accumula5on of red blood cells in the epithelium, blood vessels, and alveolar space on day 3 psc (Hemorrhage score of 2/4).Small clusters of mononuclear cells were evident in the parenchymal inters55um and, to a lesser extent, in the alveoli on day 3 and 6psc (Fig. 3D, Extended Data Fig. 7).In general, the OC43 imprinted animals had lungs with less evidence of immune cell infiltra5on and pathology compared to NL63 imprinted animals, with lower inflamma5on scoring and propor5on of the lung affected in OC43 imprinted animals on days 6 and 14 psc (Extended data Fig 7 ), although epithelial sloughing in the bronchioles was evident on day 3 psc and mild inters55al pneumonia was present on day 6 psc.Areas of pneumocyte hypertrophy were also observed las5ng un5l day 14 psc (Fig. 3D).The Ancestral-Ancestral virus control group had liple evidence of alveolar wall thickening and the alveolar space remained clear; however, there was significant presence of red blood cells in the alveolar wall and mononuclear cell infiltrates in the bronchiolar walls on day 3 psc.Some hyperplasia of alveolar pneumocytes were also noted on day 3 and 6 psc (Fig. 3D).No animal in the Ancestral-Ancestral group scored over a 2 on any lung histology severity marker, with an overall peak on day 3 at 1.3/4 (Extended Data Fig. 7)."On lines 274 to 296 Reviewer 1, Comment 30: "during Wuhan challenge, the NL63-Wuhan group had a significant increase or a back-boos5ng effect of NL63 neutralizing an5bodies following Wuhan exposure with 5ters increasing above 1:320 by day 14 psc" -very small group size and rela5vely small changes may not allow to draw these conclusions Response to Reviewer 1, Comment 30: Thank you for this comment.We have clarified that despite the small n in Fig. 4Ai (current figure lineup), these differences were found to be significantly different: "The day 14 psc NL63 neutralizing an5body 5ters were significantly increased compared to day 14 ppi (1:160), as well as just prior to the secondary inocula5on, on day 55 ppi, (1:80) (Fig. 4Ai)."lines 318 to 320.
Reviewer 1, Comment 31: "However, both seasonal virus imprinted groups had trends of higher neutralizing an5body 5ters against the an5genically distant variants of Delta, Beta, and Omicron compared to the mock primary group."-too small groups and no proper sta5s5cs done to support this conclusion Response to Reviewer 1, Comment 31: We appreciate the comment.We have addressed that the trends observed were non-significant, "both seasonal virus imprinted groups had non-significant trends of higher neutralizing an5body 5ters against the an5genically distant variants of Delta, Beta, and Omicron compared to the mock primary group" on lines 341 to 343.
Reviewer 1, Comment 32: "Imprin5ng with Wuhan protects against secondary challenge with an5genically divergent SARS-CoV-2 variants (…) These results suggest Beta and Wuhan to be an5genically closer to each other than to Omicron and that the heterologous exposure of Wuhan followed by Beta may be an op5mal an5genic distance to induce a greater breadth of an5body cross-reac5vity."-these data have been presented by others before and may not be result of imprin5ng but need to be rather called cross-protec5on amongst SARS-CoV-2 variants.
Response to Reviewer 1, Comment 32: We have modified the text to reflect the phenomenon of cross-protec5on.In lines 345 and 346 the text is now "Imprin5ng with ancestral SARS-CoV-2 induces cross-protec5on against secondary challenge with an5genically divergent variants".Also, in lines 398 and 399 the text now reads "These results suggest cross-protec5ve responses where Beta and Ancestral SARS-CoV-2 are an5genically closer to each other than to Omicron".
Reviewer 1, Comment 33: En5re sec5on "Host response analysis during secondary challenge indicates immune regula5on is based on an5genic distance" is very interes5ng and may be considered key of the en5re study, however proper cluster analysis of groups and sta5s5cs lacking in Figure 6 are needed.
Response to Reviewer 1, Comment 33: We appreciate this sugges5on.To accommodate this, we have included the fully detailed bar graph representa5on of this data in newly added Extended Data Fig. 10, 11 and 13, described on lines 1504 to 1544.
Reviewer 1, Comment 34: "More predicted T cell epitopes on OC43 shared with SARS-CoV-2 Wuhan" -this sec5on is irrelevant for the hamster model as no MHC-I nor II predic5ons are possibly in lack of data for this model.Modelling humans HLA-restricted pep5des does not support finding in the previous key paragraph.The discussion should be compacted and focus on the proposed interference between primary and secondary infec5ons.
Response to Reviewer 1, Comment 34: Thank you for poin5ng out the caveat that we should have iden5fied previously.We have now shortened the T cell analysis sec5on and highlighted the caveat of using the human MHC molecules for the analysis.The modified sec5on is from lines 513 to 541.

Reviewer #2
Remarks to the Author: In their manuscript 5tled, "Signatures of Original An5genic Sin or protec5on following coronavirus imprin5ng in Syrian hamsters challenged with SARS-CoV-2 or variants", Francis et al characterize the impact of prior infec5on with seasonal coronaviruses on infec5on, pathogenesis, and immune responses to heterologous zoono5c coronavirus secondary infec5ons.In doing so, they iden5fy differen5al impacts of NL63 and OC43 primary infec5ons on virological, disease, and immune-response outcomes to CoV2 secondary infec5ons.Interes5ngly, the differen5al disease outcomes appeared to correlate with back-boos5ng of NL63 neutralizing an5body responses and T-cell recall responses to OC43, with the former resul5ng in poor control of, and the laper improving host response to, secondary exposure.The methods and study designs were appropriate, and the results support the major conclusions.A few minor notes described below.

Response to Reviewer #2's Remarks:
We thank Reviewer #2 for their comments as they have improved the interpreta5on and discussion of the manuscript.The comments are addressed as described below.

Comments:
Reviewer 2, Comment 1 : Figure 3 describes the neutralizing an5body responses to both the seasonal and zoono5c coronaviruses, with panel A providing an overview of the kine5cs of response, only to the primary infec5ng virus, throughout the en5re study while panel B provides a snapshot of the days 9 and 14 post-secondary infec5on, only to the zoono5c secondary infec5ng virus and its variants.To gain a beper understanding of the recall responses to the secondary infec5on, it would improve the interpreta5on of the study if the authors could include either a reproduc5on of panel Ai and Aii against Wuhan virus or just the day 0 post-secondary challenge in panel B so that fold-change nAb responses can be calculated for NL63, OC43, and Wuhan.This would allow readers to interpret the rela5ve focusing of the recall responses between the primary and secondary infec5ng viruses.While panel Ai clearly shows back-boos5ng of NL63 responses, it is unclear how that boos5ng response compares to the Wuhan nAb response in those same animals since only day 9 and 14 responses are shown in panel B.
Response to Reviewer 2, Comment 1: This is a very important point, and we apologize the data was confusing.To address this, we have expanded upon Fig. 3B (Now Fig. 4B) to include all 5me points post secondary challenge, including day 0 (labelled 'pre').This change has been described in the figure legend outlined on lines 1316 to 1318.We originally leo the other 5me points out because there were no detectable neutralizing an5bodies, but we see that led to confusion.We hope the inclusion of the other 5me points improves the interpreta5on of the data.

Reviewer 2, Comment 2
: Also in figure 3, the homologous wuhan-wuhan sera show remarkable breadth of nAb responses against VoCs with only a 2-fold drop in nAb 5ters against Omicron and even a poten5al increased 5ter against Beta.This is surprising given observa5ons by others showing much larger differences in heterologous nAb 5ters between Wuhan and Omicron/Beta VoCs.A discussion on this with some poten5al explana5ons would be helpful.
Response to Reviewer 2, Comment 2: Thank you for the comment.We have added some points in the discussion on possibili5es for differences among studies to explain Fig. 3B (Now Fig. 4B).The following lines were added to the discussion at lines 710 to 715, "Addi5onally, we found a high level of cross-neutralizing an5bodies among SARS-CoV-2 Ancestral virus and the VOCs.Previous studies have also inves5gated cross-protec5on and cross-neutraliza5on among SARS-CoV-2 VOCs with some studies showing high levels of cross-neutraliza5on and others with lower levels 101,105 .Differences in cross-neutraliza5on or protec5on compared to the findings from our study could be due to lower infec5ous doses at inocula5on or 5me between primary infec5on and secondary challenge or blood collec5on".

Reviewer 2, Comment 3
: Finally, this study only evaluated recall responses over a short period following primary infec5on (56 days) with minimal waning of primary nAb responses prior to secondary infec5on.A discussion on this limita5on and impact on conclusions is warranted.
Response to Reviewer 2, Comment 3: Thank you for asking about this limita5on.To address this, we have added a discussion on the ra5onale on using this 5meline, in addi5on to hypotheses as to what would happen at altera5ve secondary infec5on 5mepoints on lines 598 to 617.
The following was included in the Discussion: "Our study design had a rela5vely short recovery 5me of 56 days between primary infec5on and secondary challenge.We leveraged this design as it was long enough for allow innate responses to return to baseline and adap5ve responses to begin contrac5ng while also being short enough for data acquisi5on and comple5on of our many groups.Due to this design our conclusions can only reflect responses in this 5meframe.Since the acute infec5on phase of COVID-19 in humans is considered to be 2-4 weeks 77,78, to understand the development of long-lived memory cells and the recall response upon secondary exposure, a longer recovery 5me should be studied in the future.In our study we observed a decrease in an5bodies against NL63 and OC43 over the recovery period which was not observed for the SARS-CoV-2 ancestral virus infected group.It would be interes5ng to inves5gate longer 5me periods for the seasonal coronaviruses as we expect that an5body levels would con5nue to decline, as has been seen in humans 17,79,80 .Given this, upon exposure to SARS-CoV-2 at a later 5mepoint when an5body responses are waning, it is possible that less of an an5genic sin-like response towards NL63 would be induced.Several studies have now inves5gated long-term memory development against SARS-CoV-2 in humans [81][82][83][84][85] .Dan et al. demonstrated that 8-months aoer infec5on, B cell responses were rela5vely maintained in pa5ent sera, while the CD4 and CD8 T cell responses decreased over 5me 81 sugges5ng that if we extended our recovery period aoer SARS-CoV-2 exposure the humoral response may be sustained while the decrease of CD4 cells over 5me may affect the germinal center response through a T follicular helper cells dependent mechanism [86][87][88] .More work is needed to understand how cross-reac5vity among coronaviruses may change over 5me."

Reviewer #3
Remarks to the Author: In this manuscript, Francis and colleagues use the Syrian hamster model to assess the rela5ve role previous infec5on with seasonal coronaviruses contributes to subsequent infec5on and host response elicita5on with SARS-CoV-2 virus.Authors conduct primary challenge experiments with several seasonal coronaviruses, followed by rechallenge with an early SARS-CoV-2 virus, or primary challenge with the early Wuhan virus followed by rechallenge with different variants of concern.A range of virological and host response parameters were conducted.Collec5vely the authors find that primary infec5on with seasonal coronaviruses did not confer robust protec5on against SARS-CoV-2 virus, but do iden5fy subtle differences in host responses using several parameters, providing informa5on towards understanding the diversity of responses possible when assessing correlates of protec5on towards coronaviruses moving forward.The manuscript is well-wripen, well-referenced, logically organized, and presents a clear ra5onale for experiments performed.There are some areas that would nonetheless benefit from addi5onal clarity and jus5fica5on as discussed below.
Response to Reviewer #3's remarks: We thank reviewer 3 for their interest and apprecia5on of the study.Addi5onally, we thank the reviewer for sugges5ons that will improve the manuscript.Comments are addressed below.

Comments: Reviewer 3, Comment 1
: Page 4, 2nd paragraph of introduc5on, it would be helpful to the reader if you men5oned what rela5ve percentage of the popula5on had been infected with seasonal coronaviruses prior to the start of the SARS-CoV-2 pandemic, to beper contextualize why preclinical models with preexis5ng an5body to seasonal CoV is important to emulate.
Response to Reviewer 3, Comment 1: We appreciate this point and think it is very important to include.Thank you!To beper contextualize why preclinical models including the seasonal coronaviruses are important we added context as to what percentage of the popula5on have been exposed to seasonal coronaviruses and what the reinfec5on dynamics.This descrip5on is on lines 80 to 87.
The sec5on included in the introduc5on is as follows: "It is es5mated that adults experience 2-3 colds annually while children can have even more common cold virus events each year 18 .Addi5onally, several human studies suggest that immunity to common colds is short-las5ng, and reinfec5on can occur yearly. 17While several different viruses can cause the common cold, coronaviruses are es5mated to be the cause of 15-30% of common colds 19 .." Reviewer 3, Comment 2 : Page 7, it is understandable (and not a detriment to the study) that HKU1 could not be included.However, can the authors speculate (here or in the discussion) how this virus might have behaved in the analyses included (e.g. based on the spike protein iden5ty analysis shown in Fig 1 ) if this virus would have behaved rela5ve to the other seasonal coronaviruses examined?
Response to Reviewer 3, Comment 2: Thank you for this ques5on.Given the spike iden5ty analysis in Fig. 1A as well as characteris5cs of HKU1 we have speculated on how HKU1 imprin5ng may affect a secondary exposure to SARS-CoV-2 on lines 637 to 644.
The sec5on in the discussion now reads "In our present study we did not include analysis of HKU1 due to availability of the virus.However, since HKU1 is also a betacoronavirus and is second only to OC43 in terms of spike amino acid iden5ty when compared to SARS-CoV-2, we an5cipate that animals imprinted with HKU1 would have similar outcomes and poten5al cross-reac5ve T cell responses as we have found in the OC43 imprinted animals.This may be further supported by the fact that HKU1 uses the same entry receptor as OC43, 9-O-acetylated sialic acid."

Reviewer 3, Comment 3
: In rechallenge experiments, animals were rested from day 14 p.i. to day 56 p.i., at which point secondary challenge took place.Why was this interval chosen and do the authors think that a longer dura5on of res5ng between challenges would have modulated the rechallenge results?
Response to Reviewer 3, Comment 3: Thank you for this ques5on.We feel the conversa5on to be very important for the field (and wish we could have compared many 5me points!).To address this ques5on (which was also posed by Reviewer #2) as well as hypothesize what a longer dura5on of res5ng between imprin5ng and challenge, we have expanded upon our discussion on lines 598 to 617.
The inserted paragraph is as follows: The following was included in the Discussion: "Our study design had a rela5vely short recovery 5me of 56 days between primary infec5on and secondary challenge.We leveraged this design as it was long enough for allow innate responses to return to baseline and adap5ve responses to begin contrac5ng while also being short enough for data acquisi5on and comple5on of our many groups.Due to this design our conclusions can only reflect responses in this 5meframe.Since the acute infec5on phase of COVID-19 in humans is considered to be 2-4 weeks 77,78, to understand the development of long-lived memory cells and the recall response upon secondary exposure, a longer recovery 5me should be studied in the future.In our study we observed a decrease in an5bodies against NL63 and OC43 over the recovery period which was not observed for the SARS-CoV-2 ancestral virus infected group.It would be interes5ng to inves5gate longer 5me periods for the seasonal coronaviruses as we expect that an5body levels would con5nue to decline, as has been seen in humans 17,79,80 .Given this, upon exposure to SARS-CoV-2 at a later 5mepoint when an5body responses are waning, it is possible that less of an an5genic sin-like response towards NL63 would be induced.Several studies have now inves5gated long-term memory development against SARS-CoV-2 in humans [81][82][83][84][85] .Dan et al. demonstrated that 8-months aoer infec5on, B cell responses were rela5vely maintained in pa5ent sera, while the CD4 and CD8 T cell responses decreased over 5me 81 sugges5ng that if we extended our recovery period aoer SARS-CoV-2 exposure the humoral response may be sustained while the decrease of CD4 cells over 5me may affect the germinal center response through a T follicular helper cells dependent mechanism [86][87][88] .More work is needed to understand how cross-reac5vity among coronaviruses may change over 5me."

Reviewer 3, Comment 4
: Page 12, please specify which text maps to Figure 2D and 2E; it is currently unclear how the lung weight differences presented in Figure 2E align with the text as currently wripen in this sec5on.
Response to Reviewer 3, Comment 4: We apologize for the confusion.We have expanded our descrip5on of Fig. 2E (Now Fig. 3E) and specified those results dis5nct from the Fig. 2D (now Fig. 3D) results on lines 296 to 302.
The text now reads: "Some hyperplasia of alveolar pneumocytes were also noted on day 3 and 6 psc (Fig. 3D).No animal in the Ancestral-Ancestral group scored over a 2 on any lung histology severity marker, with an overall peak in pathological scoring on day 3 at 1.3/4 (Extended Data Fig. 7).The damage to the lung as noted in the histopathology in the Mock-Ancestral group was accompanied by an increase in lung weight which peaked on day 6 post inocula5on at 5% of animal body weight (Fig. 3E).Ancestral-Ancestral animals had a significant decrease in lung weight compared to Mock-Ancestral animals on all days psc, whereas the seasonal coronavirus imprinted groups had significant decrease in lung weight compared to Mock-imprinted animals on days 3 and 14, but not day 4 psc (Fig. 3E).

Reviewer 3, Comment 5
: Page 14, authors state that both seasonal virus imprinted groups had "trends of higher neutralizing an5body 5ters" against SARS-CoV-2 variants.However, the data shown in Figure 3B (to my eye) shows generally comparable an5body 5ters among all mock, NL63, and OC43 primary-challenged animals; are the authors drawing from addi5onal data to make this statement, or comparing mean values that can be reported which reflect this trend?As currently presented in the graphs, without addi5onal suppor5ng informa5on, I think the higher "trends" alluded to here are a bit spurious.
Response to Reviewer 3, Comment 5: We apologize for the lack of clarity in this figure .To improve the interpreta5on of this data, we have included the results from days 0, 3 and 6 post-secondary inocula5on, in addi5on to the day 9 and 14 data that was originally presented in Fig. 3B  (Now Fig 4B).We have also specified this trend is nonsignificant and applies only to Delta and Beta, as indicated on lines 341 to 343.We hope the data is beper understood now.The text now reads "both seasonal virus imprinted groups had non-significant trends of higher neutralizing an5body 5ters against the an5genically distant variants of Delta and Beta, compared to the mock imprinted group."

Reviewer 3, Comment 6
: page 16/Figure 5, the current graphs as presented make it a bit challenging for the reader to compare trends present from specific groups of animals that are present between different assessments of neutralizing an5bodies (the graph are designed for, and effec5vely show, augmenta5on of neutralizing an5body 5ters against a single variant over 5me in different treated groups of animals).If the authors want to state, as they do in the text, that specific animal groups displayed trends across mul5ple assays shown in Figure 5, an addi5onal panel in this figure (e.g.where the legend is currently present), showing compara5ve levels of one 5mepoint assessed for the presence of different neutralizing an5bodies between different groups, would make it much easier for the reader to quickly see the data the authors are alluding to in making this statement.
Response to Reviewer 3, Comment 6: Thank you for this comment.We have now expanded Fig. 5 (Now Fig. 6) to include a graph with just day 14 post-secondary inocula5on, allowing to sta5s5cally show differences between these groups (Fig. 6B).
The descrip5on of Fig. 6B is on lines 391 to 399 and reads "The Ancestral-Omicron group had neutralizing 5ters on day 14 psc with geometric means >1:320, >1:320, >1:320, and >1:640, for the Ancestral strain, and Beta, Alpha, and Delta variants, respec5vely (Fig. 6B).The Ancestral-Omicron group had 5ters of 1:1280 against the Omicron virus itself.Ancestral-Ancestral and Ancestral-Beta groups were above 1:640 for all variants with the excep5on of the Omicron neutraliza5on assay which were lower (Fig. 6B).Interes5ngly, the Ancestral-Beta group had a higher median than the Ancestral-Ancestral group for the neutraliza5on of Beta, Omicron, Alpha, and Delta viruses (Fig. 6B)."

Reviewer 3, Comment 7
: Figure 6, please specify what fold change is in rela5on to (mock animals?) and what blue in the scale bar refers to (values with a nega5ve fold-change rela5ve to mock?).Also please specify what normaliza5on (if any) took place in these specimens; authors show in earlier figures that lungs had different weights, so I presume some sort of normaliza5on to a housekeeping gene was conducted?
Response to Reviewer 3, Comment 7: To clarify these points, we have added a descrip5on of fold change calcula5ons, the scale bar as well as normaliza5on to a housekeeping gene in the figure legend for Fig. 6 (Now Fig. 7), on lines 1361 to 1364.
The text now read "Fold-change was calculated via ΔΔCt against baseline (Day 0) with BACT as the housekeeping gene.The legend depicts fold change in which upregula5on (greater than 1) is represented ranging in from white (=1) to red.Any downregula5on (less than 1) is represented in blue." Reviewer 3, Comment 8 : Figure 1C, the colora5on between live virus and qPCR in these graphs is very, very subtle; recommend a different differen5a5on strategy (e.g., colors with greater separa5on, or filled vs open symbols) to show this difference.Also please clarify why the LOD for these data is different than in Figure 2C.
Response to Reviewer 3, Comment 8: We apologize for this oversight.We have updated these colours in Fig. 1C so that the data can more easily be differen5ated between groups.The LOD in these assays was lower than other assays as we decreased the dilu5on series by star5ng with an undiluted sample rather than a 1:10 dilu5on.This strategy was taken so that we could increase the possibility of detec5ng lower amounts of virus that may be in the sample.This descrip5on has been added to the methods for clarifica5on in lines 810 and 811.
The authors make a great effort to adapt their narrative according our suggestion; to clarify and correct possibly misleading expressions.For me my main criticism regarding the lack of clear experimental evidence for 'Original Antigen Sin' remains.Fig- 3 may be key here.In brief, NL63: no effect (neither on viral loads nor on pathology; similar to mock); OC43: minimal effect on lung viral loads; none on pathology); effect only seen when animals are second time exposed to (any) SARS-CoV-2 isolate.This cannot be called Antigenic Sin this is adaptive immunity.Some level of cross-protection between ancestral SARS-CoV-2 and VoCs is well documented before.In my opinion the authors should consider focussing on the OC43-SARS-CoV-2 interaction to substantiate their hypothesis.
Minor Comments: 1-The Abstract may be by far too long.
2-Old Comment 14: The references provided show that S contain distinct B-and T-cells epitopes that dominate S-specific immune responses; not that responses to S are dominating during natural infection.
3-I suggest to write "ancestral" with a small letter 4-Old Comment 21: group sizes appear too small to draw quantitative conclusions.5-Old Comment 23: The authors are encouraged to improve their analysis of the DEG analysis e.g.clustering, statistics, etc. to go beyond plain reporting of primary results.
6-Line 274-276: It is misleading to call mock-infected (thus previously non-infected) animals that are infected for the first time with SARS-COV-2 exposed to a "secondary ancestral virus challenge".7-Line 345: This describes consecutive infection by two SARS-CoV-2 viruses."Imprinting" is wrong (as it implies already the to be proven hypothesis); what is actually meant is "prior infection".Nevertheless, cross protection from prior SAR-CoV-2 is now long standing knowledge and thus an almost trivial finding.8-Old Comment 34: This extensive narrative immunoinformatic discovery of T-cell epitopes is misleading because related to the human model and not the hamster model as used to generate the original results in the MS.I suggest to cut down, make this narrative much less suggestive, and/or move to Supplementary.
Reviewer #2 (Remarks to the Author): Thank you for responding to my requests.I found the responses to be appropriate and the revised version of the manuscript is easier to interpret (at least for me).