Virus-like and Virus Replicon Particles Targeting Multiple B-Cell Antigens Do Not Protect Against African Swine Fever Virus

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

Comments on vaccines-4175063

The revised version seems to be improved. But there are many issues remain to be addressed.

1.        The “SCORE screening” method (2.1) was referenced but not defined – no details on how fluorescence signals were normalized or converted to antigenic signal intensity (A.U.).

2.        Chemical coupling of p22 and CD2v to AP205 was described, but no quantitative data (e.g., densitometry from SDS-PAGE) were provided to confirm coupling efficiency – a key quality control for vaccine candidates.

3.        The “termination clinical score” (2.11/3.4) was repeatedly mentioned but not listed – readers cannot evaluate if endpoint criteria were objective or standardized.

4.        For some assays (e.g., flow cytometry in 3.6), the text stated “ANOVA with Dunnett’s correction” but did not specify how outliers were handled (critical for small sample sizes, n = 5 pigs/group).

5.        All in vivo pig groups used only 5 animals per group, with 1 animal euthanized early in the positive control group (ASFV-Est14) – this reduced statistical power and increases the risk of type II errors (false negatives) for vaccine efficacy.

6.        Mouse immunization studies (2.7) used 5 mice per group with no biological replicates – results on VLP immunogenicity were not validated independently.

7.        A core hypothesis of the study was that B-cell antigens induced functionally relevant antibodies (e.g., neutralization, ADCC, ADCP). However,

No ASFV neutralization assay was performed for serum from VLP/VSV-vaccinated pigs – the gold standard for evaluating antibody-mediated protection against viruses.

No ADCC/ADCP or hemadsorption inhibition (HAI) assays were conducted for CD2v/EP153R (antigens linked to hemadsorption) – the study only measures binding antibodies (ELISA) but not functional activity, a major gap in interpreting negative vaccine results.

8.        The study concluded that T-cell immunity was pivotal for ASFV protection but provided minimal phenotypic/functional T-cell data:

1)        Intracellular cytokine staining (ICS, 2.16) only measured IFN-γ and TNF (TNF-α?) – no assessment of key T-cell subsets (e.g., memory CD8+ T cells, Tfh cells) or cytotoxicity (e.g., granzyme B/perforin).

2)        No T-cell epitope mapping was performed for the selected B-cell antigens – the study cannot rule out that the antigens lack T-cell epitopes (a critical factor for vaccine-induced immunity).

3)        The VSV vaccine group showed “IFN-γ production by CD4+CD8+ T cells” (3.6) but no follow-up analysis (e.g., kinetics, antigen specificity) – this finding is unexplored and underinterpreted.

9.        The negative control group only received empty AP205 VLPs – there was no rVSV (empty) control group for the VSV-vaccinated pigs. This means the study cannot distinguish if the lack of protection is due to the ASFV antigens or the VSV platform itself.

10.    Montanide 28R is used for VLP vaccines (2.11) but no adjuvant-only group is included – the study cannot rule out adjuvant-induced nonspecific inflammation impacting results.

11.    The study concluded that the selected B-cell antigens were not protective (4. Discussion) but did not account for critical variables:

1)        CD2v/EP153R are glycosylated, and the VLP platform used only the extracellular domain of CD2v – the study does not test full-length CD2v/EP153R or confirm if conformational epitopes (critical for function) are preserved in the vaccine constructs.

2)        Pigs received three vaccine doses with a 3-week interval, but the challenge was at 65 dpi – no memory B/T-cell kinetics were measured (e.g., antibody avidity, memory T-cell frequency) to confirm if the vaccine induced long-term immunity.

3)        The vaccine antigens were derived from ASFV Georgia 2007 (genotype II), but the positive control used ASFV Estonia 2014 (genotype I) – the study did not address if antigenic differences between genotypes impact cross-protection, yet generalizes results to B-cell antigens broadly.

12.    The antigens (e.g., B169L, H171R) were selected based on recognition by SPF pig sera only (3.1), but the study did not confirm if this recognition is causally linked to protection (e.g., via passive antibody transfer). It only assumes correlation, a logical fallacy.

13.    The protein microarray used cell-free E. coli-expressed ASFV proteins (2.1) – these lack post-translational modifications (e.g., glycosylation, phosphorylation) present in native ASFV proteins. The study did not validate if the microarray detects physiologically relevant antibody-antigen interactions.

14.    The VSVDG-CD2v/EP153R vaccines induced low antibody levels (3.6), but the study provided only speculative explanations (e.g., glycan shielding) with no supporting data:

1)        No assessment of VSV vector replication in pig tissues – the replication-defective VSV may have poor in vivo persistence, leading to low antigen exposure.

2)        No measurement of VSV-specific immune responses (e.g., anti-VSV antibodies) – vector-induced immunity may have reduced the efficacy of booster doses (vector interference), a common issue with viral vectored vaccines.

15.    Key findings (e.g., VLP immunogenicity in mice, ASFV protein expression in VSV-infected cells) were presented with excessive figure references (e.g., Supplementary Figures 1igu but minimal textual context – readers must cross-reference figures constantly to follow the narrative.

16.    The “3.1 Identification of B-cell antigens” section confused immunodominance with protective potential – the text did not clearly distinguish between antigens recognized by antibodies and antigens that mediate protection, a critical conceptual distinction.

17.    The discussion briefly mentioned some limitations (e.g., lack of T-cell epitope selection) but failed to address major ones:

1)        No mention of the small sample size in pigs/mice.

2)        No acknowledgment that binding antibodies (ELISA) did not equate to functional antibodies.

3)        No discussion of the impact of using a single ASFV challenge strain (Georgia 2007) – the vaccine may have efficacy against other genotype II strains, which was not tested.

18.    Supplementary Figure 6 (ICS gating strategy) was referenced but no gating plots are provided for critical subsets (e.g., CD4+CD8+ T cells) – readers cannot verify the accuracy of cell subset identification.

19.    Supplementary Table 1 (antibodies for flow cytometry) lacks clone names and dilutions – a standard requirement for reproducibility in immunology studies. 20.    The study mentioned collecting spleen/liver/lymph nodes for pathological examination (2.11) but no pathological data (e.g., lesion scores, viral antigen staining) were presented – this is a key endpoint for ASFV, a hemorrhagic disease.

21.    The mouse immunization license (BE80/2021) was listed, but the study did not state humane endpoint criteria for pigs (e.g., temperature thresholds, clinical signs) – a requirement for animal research manuscripts.

 

Comments for author File: Comments.pdf

Author Response

POINT BY POINT REPLY TO THE REVIEWER’S COMMENTS; March 4, 2026

AUTHORS: We are very grateful to Reviewer 1 for taking the time to critically review our manuscript. Please find below our point-by-point replies.

1. The “SCORE screening” method (2.1) was referenced but not defined – no details on how fluorescence signals were normalized or converted to antigenic signal intensity (A.U.).

AUTHORS: We have added this information and the relevant references in the revised manuscript (lines 136-137).

2. Chemical coupling of p22 and CD2v to AP205 was described, but no quantitative data (e.g., densitometry from SDS-PAGE) were provided to confirm coupling efficiency – a key quality control for vaccine candidates.

AUTHORS: Densitometric analysis of the SDS-PAGE results is provided in Supplementary Figure 3A, where the fusion proteins used in this study were evaluated for their relative quantities.

3. The “termination clinical score” (2.11/3.4) was repeatedly mentioned but not listed – readers cannot evaluate if endpoint criteria were objective or standardized.

AUTHORS: The termination criteria are now defined in the manuscript (chapter 2.11, lines 283-289).

4. For some assays (e.g., flow cytometry in 3.6), the text stated “ANOVA with Dunnett’s correction” but did not specify how outliers were handled (critical for small sample sizes, n = 5 pigs/group).

AUTHORS: The identification of outliers is now mentioned in the manuscript (2.17) and ther reults for the dataset where it was performed is given in the Figure Legends.

5. All in vivo pig groups used only 5 animals per group, with 1 animal euthanized early in the positive control group (ASFV-Est14) – this reduced statistical power and increases the risk of type II errors (false negatives) for vaccine efficacy.

AUTHORS: SPF pigs from our breeding facility have a much higher homogeneity in their immune response compared to pigs purchased from a farm. Therefore, based on our previous experience with SPF pigs ¹, 5 pigs are sufficient to draw first conclusions on the protective value of a vaccine. When applying for animal licences the power calculations to estimate the number of aninmals are a requirement. In this case the power calculation resulted in 5 animals per group (power of 0.8; standard deviations less than to 55%). In addition, the 5 animals are conclusive for the data in the present study.

6. Mouse immunization studies (2.7) used 5 mice per group with no biological replicates – results on VLP immunogenicity were not validated independently.

AUTHORS: The group size of five mice was selected based on the previously published work. Specifically, in the study by Rothen et al., 2024 (doi: 10.3390/vaccines12080874), five mice per group were used to assess VLP immunogenicity, and this sample size was demonstrated to be sufficient to detect robust and statistically significant immune responses.

7. A core hypothesis of the study was that B-cell antigens induced functionally relevant antibodies (e.g., neutralization, ADCC, ADCP). However, no ASFV neutralization assay was performed for serum from VLP/VSV-vaccinated pigs – the gold standard for evaluating antibody-mediated protection against viruses.

AUTHORS: Neutralization of ASFV, at least when performed with macrophages, cannot be used to predict protection because it is atypical. ASFV exists in two infectious forms, each presenting a different antigenic landscape and susceptibility to neutralization. The first represents enveloped virions that contain CD2v which do not represent a target for neutralizing antibodies ^2-4. The second form represents the intracellular non-enveloped virion, which lacks the outer membrane and exposes the highly stable capsid, making it susceptible to neutralizing antibodies targeting p72 and minor capsid proteins (pB438L and pE120R). As we did not include capsid antigens in our vaccine, we do not expect any neutralization. Furthermore, such neutralizing antibodies have not been associated with protection ^5,6. Given that there is some controversy on this topic in the literature, we have added a recommendation of antibody functions to be tested for screening of antigens in future studies at the end of the Discussion in the concluding sentences. Unfortunately, although informative, none of these assays are established to predict protection.

No ADCC/ADCP or hemadsorption inhibition (HAI) assays were conducted for CD2v/EP153R (antigens linked to hemadsorption) – the study only measures binding antibodies (ELISA) but not functional activity, a major gap in interpreting negative vaccine results.

AUTHORS: We agree that these assays would be informative but at the same time, neither CD2v nor EP153R has been reported as a target for ADCC. We think that this needs to be addressed systematically and the assays need to be established with appropriate controls in and ideally monospecific porcine antibodies which are not available. Without this, the results will be difficult to interpret. We cannot provide this data in the frame of this work.

With respect ot HAI, these assays were not included as they do also not represent a clear correlate of protection ⁷. In addition, non-hemadsorbing LAV can be protective.

There are reports of antibody-enhanced infection caused by certain antibodies in particular those targeting p72 capsid⁶. To our knowledge the fate of immune complexes taken up by macrophages through FcR needs to be investigated (it is unclear if this results in infection or virus destruction). Accordingly, this goes beyond what is possible in the frame of this work.

In the end of the Discussion we have now pointed on this knowledge gap that should be addressed systematically in future studies.

8. The study concluded that T-cell immunity was pivotal for ASFV protection but provided minimal phenotypic/functional T-cell data:

1)        Intracellular cytokine staining (ICS, 2.16) only measured IFN-γ and TNF (TNF-α?) – no assessment of key T-cell subsets (e.g., memory CD8+ T cells, Tfh cells) or cytotoxicity (e.g., granzyme B/perforin).

AUTHORS: Many authors in the field only present INF-g ELISPOT data. In addition, INF-g and TNF (TNF is the official nomenclature) are the relevant cytokines for a Th1 reponse that is required for a virus infection. In vitro restimulation, by definition,  represent recall responses of memory T cells (no naïve T-cell responses can be induced in a overnight restimulation assay). Tfh cells are in the B cell folicle to which we had no access without killing the pigs. As there was no restimulation of CD8 T cells in the experimental vaccine groups we did not further explore cytotoxic activity. Given the negative results of our study a further exploration of T cells would not give additional insights.

2)        No T-cell epitope mapping was performed for the selected B-cell antigens – the study cannot rule out that the antigens lack T-cell epitopes (a critical factor for vaccine-induced immunity).

AUTHORS: This study focussed on B cell antigens, not on T-cell epitopes. The importance of T cells has been addressed in the Discussion.

3)        The VSV vaccine group showed “IFN-γ production by CD4+CD8+ T cells” (3.6) but no follow-up analysis (e.g., kinetics, antigen specificity) – this finding is unexplored and underinterpreted.

AUTHORS: CD2v is known to contain T-cell epitopes ⁸. It was not the aim of this study to further explore their identity and kinetics. We have added an few words (lines 575-576) for the interpretation of this observations in the Results describing Figure 8. 

 

9. The negative control group only received empty AP205 VLPs – there was no rVSV (empty) control group for the VSV-vaccinated pigs. This means the study cannot distinguish if the lack of protection is due to the ASFV antigens or the VSV platform itself.

AUTHORS: We were limited with the amount of space and stables as these experiments need to be performed under BSL3 conditions. Considering that empty VSV cannot induced protection they cannot serve as a positive control but only as a negative control. As the vaccines did not work, a negative control cannot help in the interpretation of this failure. If the vaccines would have worked, we would have performed follow-up experiments with a VSV as negative control to ensure correct interpretation.

 

10. Montanide 28R is used for VLP vaccines (2.11) but no adjuvant-only group is included – the study cannot rule out adjuvant-induced nonspecific inflammation impacting results.

AUTHORS: The Montanide oil-in-water adjuvants are very potent for antibody responses in pigs and we found no indication of such side effects in terms of increase in body temperature, swelling redening in the present study and previous work. Accordingly, this is a very unlikely explanation for the negative results. We have added this information to the Results (line 475).

 

11. The study concluded that the selected B-cell antigens were not protective (4. Discussion) but did not account for critical variables:

 

1) CD2v/EP153R are glycosylated, and the VLP platform used only the extracellular domain of CD2v – the study does not test full-length CD2v/EP153R or confirm if conformational epitopes (critical for function) are preserved in the vaccine constructs.

AUTHORS: We agree with this interpretation and mention this point in the Discussion (lines 639-644). It was not possible to express the full-length CD2v and EP153R on the VLP. This is why the VSV platfrom was used instead.

 

2) Pigs received three vaccine doses with a 3-week interval, but the challenge was at 65 dpi – no memory B/T-cell kinetics were measured (e.g., antibody avidity, memory T-cell frequency) to confirm if the vaccine induced long-term immunity.

AUTHORS: We agree that duration of immunity is important but evaluating is only worth doing with a good vaccine candidates which unfortunately was not developed in the present study. T-cells and antibody responses were measured shortly before challenge which is the relevant time point. Kinetics of immune responses were not considered essential for our aims.

 

3)  The vaccine antigens were derived from ASFV Georgia 2007 (genotype II), but the positive control used ASFV Estonia 2014 (genotype I) – the study did not address if antigenic differences between genotypes impact cross-protection, yet generalizes results to B-cell antigens broadly.

AUTHORS: The amino acid sequences for the selected antigens were identical between the ASFV-Est14 strain and the challenge virus.

 

12. The antigens (e.g., B169L, H171R) were selected based on recognition by SPF pig sera only (3.1), but the study did not confirm if this recognition is causally linked to protection (e.g., via passive antibody transfer). It only assumes correlation, a logical fallacy.

AUTHORS: We are aware that neither associations nor correlations mean causality, and do not make any statements in this direction. Nevertheless, in an immune responses is observed in protected animals, then such data give an indication that an antigen might be involved in protection and can be taken as a justification for further exploration. Performing antibody transfer experiments would be based on the identical assumptions.

 

13. The protein microarray used cell-free E. coli-expressed ASFV proteins (2.1) – these lack post-translational modifications (e.g., glycosylation, phosphorylation) present in native ASFV proteins. The study did not validate if the microarray detects physiologically relevant antibody-antigen interactions.

AUTHORS: This is of course a limitation of the approoach. Nevertheless, as stated in the Discussion most of these proteins have been identified by others using different approaches (see line 613-615).

 

14. The VSVDG-CD2v/EP153R vaccines induced low antibody levels (3.6), but the study provided only speculative explanations (e.g., glycan shielding) with no supporting data:

1)        No assessment of VSV vector replication in pig tissues – the replication-defective VSV may have poor in vivo persistence, leading to low antigen exposure.

AUTHOR: The VSV vector is known to be highly immunogenic in pigs based on our previous works ^9-12. Therefore, poor immunogenicity is unlikely to be caused by a lack of replication in pigs.

2)        No measurement of VSV-specific immune responses (e.g., anti-VSV antibodies) – vector-induced immunity may have reduced the efficacy of booster doses (vector interference), a common issue with viral vectored vaccines.

AUTHOR: This is a very well established vector system that as been used multiple times in pigs and many species . As the G-protein gene is deleted, from the vector anti-G protein reponses are absent or very weak. Accordingly, there is no interference using prime boost vaccination schemes. Nevertheless, it is possible that the third injection was not so effective. We have added this comment to the Discussion (line 668).

 

15. Key findings (e.g., VLP immunogenicity in mice, ASFV protein expression in VSV-infected cells) were presented with excessive figure references (e.g., Supplementary Figures 1igu but minimal textual context – readers must cross-reference figures constantly to follow the narrative.

AUTHOR: In our opinion, the Figures of the manuscript are clear and easy to follow. Of course if a reader is interested in more data such as sequence information, raw data, antibodies used, flow cytometry gating the Supplementary Figures need to be consulted. This, however, is only a minority of readers.

 

16. The “3.1 Identification of B-cell antigens” section confused immunodominance with protective potential – the text did not clearly distinguish between antigens recognized by antibodies and antigens that mediate protection, a critical conceptual distinction.

AUTHOR: We checked this chapter and it is unclear to us why the reviewer gets this impression. There is no statement linking antigenicity to protection. As the Title and text outlines, this chapter is only about antigens recognized by immune sera using the protein microarray. It is correct that we are listing them in terms of there use in previous experimental vaccines but not in terms of their protective potential. Independent of vaccinology, this data is imformative for possible serological diagnostic targets.

 

17. The discussion briefly mentioned some limitations (e.g., lack of T-cell epitope selection) but failed to address major ones:

1)        No mention of the small sample size in pigs/mice.

AUTHOR: We did not mention this because the number of animals were sufficient to answer the scientific questions. The local authorities have verified our power calculations for animal numbers and judged them to be sufficient. Further increasing the number of animals means not folowing the 3R principles which is illegal in Switzerland.

2)        No acknowledgment that binding antibodies (ELISA) did not equate to functional antibodies.

AUTHOR: We do mention ADCC, ADCP and complement-dependent cytotoxicity in the Discussion and now have included neutralization (last paragraph of the Discussion).

3)        No discussion of the impact of using a single ASFV challenge strain (Georgia 2007) – the vaccine may have efficacy against other genotype II strains, which was not tested.

AUTHOR: We do mention elaborate on possible protection against heterologous strains considering that there was no protection against homologous strains. We think that this would make a too lengthy discussion.

 

18. Supplementary Figure 6 (ICS gating strategy) was referenced but no gating plots are provided for critical subsets (e.g., CD4+CD8+ T cells) – readers cannot verify the accuracy of cell subset identification.

AUTHOR: Supplementary Figure 6 provides this information.

 

19. Supplementary Table 1 (antibodies for flow cytometry) lacks clone names and dilutions – a standard requirement for reproducibility in immunology studies.

AUTHOR: All clone names are given. The dilutions are not given because they are batch- and FCM machine-dependent. In fact, flow cytometry guidelines recommend that each lab determines optimal dilutions with every new batch of antibody. This is an important step in establishing the staining that should not be skiped even if suppliers sugge.st dilutions

 

20. The study mentioned collecting spleen/liver/lymph nodes for pathological examination (2.11) but no pathological data (e.g., lesion scores, viral antigen staining) were presented – this is a key endpoint for ASFV, a hemorrhagic disease.

AUTHOR: Collecting these tissues is a standard procedure in our lab but given the lack of any clinical and virological protection, no histopathological investigations were performed. We have modified this sentence from the text to avoid confusion of the reader.

 

21. The mouse immunization license (BE80/2021) was listed, but the study did not state humane endpoint criteria for pigs (e.g., temperature thresholds, clinical signs) – a requirement for animal research manuscripts.

AUTHOR: For the pig experiment the correct license number is BE46/2022. We have added this information regarding the scoring and the definition of discontinuation to the text of the revised manuscript Chapter 2.11, lines 283-289.

 

References

1. Radulovic E, et al. Development of protective immunity against African swine fever depends on host-environment interactions. Front Vet Sci 2025;12:1553310.
2. Walczak M, et al. ASF -survivors’ sera do not inhibit African swine fever virus replication in vitro. Journal of Veterinary Research 2022;66:21–27.
3. De Boer CJ, et al. Studies to determine the presence of neutralizing antibody in sera and kidneys from swine recovered from African swine fever. Arch Gesamte Virusforsch 1969;27:44–54.
4. Parker J, et al. Plaque formation by African swine fever virus. Nature 1968;219:524–5.
5. Gaudreault NN, et al. Subunit Vaccine Approaches for African Swine Fever Virus. Vaccines (Basel) 2019;7.
6. Munoz AL, et al. Characteristics of the major structural proteins of African swine fever virus: Role as antigens in the induction of neutralizing antibodies. A review. Virology 2022;571:46–51.
7. Lotonin K, Brito, F., Mehinagic, K., García-Nicolás, O, Liniger, K., Donzé, N., Python, S., Talker, S., Ploegaert, T., Ruggli, N., Benarafa, C., Summerfield, A. Correlates of protection against African swine fever virus identified by a systems immunology approach. eLife 2025;14:RP107579.
8. Schafer A, et al. Adaptive Cellular Immunity against African Swine Fever Virus Infections. Pathogens 2022;11.
9. Graaf-Rau A, et al. Reassortment incompetent live attenuated and replicon influenza vaccines provide improved protection against influenza in piglets. NPJ Vaccines 2024;9:127.
10. Avanthay R, et al. Evaluation of a novel intramuscular prime/intranasal boost vaccination strategy against influenza in the pig model. PLoS Pathog 2024;20:e1012393.
11. Ricklin ME, et al. Partial Protection against Porcine Influenza A Virus by a Hemagglutinin-Expressing Virus Replicon Particle Vaccine in the Absence of Neutralizing Antibodies. Frontiers in immunology 2016;7:253.
12. Ricklin ME, et al. Virus replicon particle vaccines expressing nucleoprotein of influenza A virus mediate enhanced inflammatory responses in pigs. Scientific reports 2017;7:16379.

 

Reviewer 2 Report (New Reviewer)

Comments and Suggestions for Authors

The authors have submitted the manuscript titled "Virus-like and virus replicon particles targeting multiple B cell antigens do not protect against African swine fever virus". I have given my concerns about the manuscript below:

1. The authors mention that the VLP and VSV vaccines generated detectable antibody responses, but they did not provide protection. To assess the true protection provided by the vaccines, they need to evaluate the antibody avidity and neutralizing antibody titers.
2. It is unclear how they decided the dose for mouse immunization. Also, why was the dose of VLP and VLP-CD2v different for mouse immunization?
3. They did not evaluate the baseline responses, prior to exposure.
4. Why was the concentration of the coating antigen and methods of development of ELISA different for mouse and pig sera?

Author Response

The authors have submitted the manuscript titled "Virus-like and virus replicon particles targeting multiple B cell antigens do not protect against African swine fever virus". I have given my concerns about the manuscript below:

1. The authors mention that the VLP and VSV vaccines generated detectable antibody responses, but they did not provide protection. To assess the true protection provided by the vaccines, they need to evaluate the antibody avidity and neutralizing antibody titers.

We respectfully disagree that performing such assays will be helpful. THere is no known association of antibody avidity to the selected antigens and protection. As the vaccines did not provide protection, it is not possible to assess the "true protection". In addition, neutralization of ASFV, at least when performed with macrophages, cannot be used to predict protection because it is atypical. ASFV exists in two infectious forms, each presenting a different antigenic landscape and susceptibility to neutralization. The first represents enveloped virions that contain CD2v which do not represent a target for neutralizing antibodies. The second form represents the intracellular non-enveloped virion, which lacks the outer membrane and exposes the highly stable capsid, making it susceptible to neutralizing antibodies targeting p72 and minor capsid proteins (pB438L and pE120R). As we did not include capsid antigens in our vaccine, we do not expect any neutralization. Furthermore, such neutralizing antibodies have not been associated with protection. For references on this topic we refer to:  Walczak M, et al. ASF -survivors’ sera do not inhibit African swine fever virus replication in vitro. Journal of Veterinary Research 2022;66:21–27.; De Boer CJ, et al. Studies to determine the presence of neutralizing antibody in sera and kidneys from swine recovered from African swine fever. Arch Gesamte Virusforsch 1969;27:44–54.; Parker J, et al. Plaque formation by African swine fever virus. Nature 1968;219:524–5.; Gaudreault NN, et al. Subunit Vaccine Approaches for African Swine Fever Virus. Vaccines (Basel) 2019;7.; Munoz AL, et al. Characteristics of the major structural proteins of African swine fever virus: Role as antigens in the induction of neutralizing antibodies. A review. Virology 2022;571:46–51.
Given that there is some controversy on this topic in the literature, we have added a recommendation of antibody functions to be tested for screening of antigens in future studies at the end of the Discussion were we are elaborating on possible functional assays in lines 679-682.

 

1. It is unclear how they decided the dose for mouse immunization. Also, why was the dose of VLP and VLP-CD2v different for mouse immunization?

AUTHORS: This dose of VLP is based on previous work from the Bachman lab which has 30 years of experience with VLP. CD2v VLP were used at a higher dose because these animals received only one VLP.

 

1. They did not evaluate the baseline responses, prior to exposure.

AUTHORS: These are SPF pigs and Switzerland is free of ASFV. It is impossible to have pre-existing immunity. We have used the negative control sera as negative controls

1. Why was the concentration of the coating antigen and methods of development of ELISA different for mouse and pig sera?

AUTHORS: These ELISA were performed in different labs that have different standard established protocols with different conjugates etc. 

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

Although the challenge test, as expected, yielded negative results, the overall experimental data looks very usefull.  The comprehensive data from these experiments is highly valuable for other research groups in this field.

The manuscript is very well written and, in my opinion, does not require any significant improvements. Additionally, the tool-box available to this group, together with the funding for such a massive and expensive experiment, is impressive.

I have a minor comment for the authors: the citations in the Introduction section could be updated or improved.

"The virus is spread through direct contact to infected animals, as well as through contam-48

inated feed, equipment, and vehicles. ASFV surveillance is particularly challenging due 49

to the virus’s ability to persist in the environment, as well as the lack of a safe vaccine"

 

The role of infected/contaminated pork-derived products is missing. Also, the role of the vaccine is overestimated. Infection can be relatively easily managed by minimal biosecurity measures, at least in domestic pigs. If the authors are highlighting the role of vaccines in ASF control in wild boar, the availability of a safe vaccine alone will likely not solve the problem.

Nevertheless, this is a very good manuscript providing a lot of useful information for the scientific community.

 

Author Response

I have a minor comment for the authors: the citations in the Introduction section could be updated or improved. The role of infected/contaminated pork-derived products is missing. Also, the role of the vaccine is overestimated. Infection can be relatively easily managed by minimal biosecurity measures, at least in domestic pigs. If the authors are highlighting the role of vaccines in ASF control in wild boar, the availability of a safe vaccine alone will likely not solve the problem.

AUTHOR: Many thanks for the positive comments. We have revised the first paragraph of the Introduction and now mention the importance of the contaminated pork products and biosecurity (line 49). We agree that vaccination alone cannot solve the problem but have not further elaborated on this complex issue. 

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses an important and timely question in African swine fever virus vaccine research, namely the identification of protective B cell antigens and the evaluation of subunit- and vector-based vaccine platforms in a stringent pig challenge model. The study was technically well executed, included an extensive antigen screening approach, and employed appropriate in vivo validation using a virulent genotype II ASFV strain. The negative protection outcome was clearly demonstrated and supported by multiple independent readouts. However, in its current form, the manuscript requires major revisions to improve the clarity of interpretation, strengthen the conceptual framework, and more clearly articulate the scientific significance of the findings. The main concerns relate not to data quality but to data interpretation, narrative structure, and positioning of the conclusions.

Major comments

1. The manuscript would benefit from a clearer articulation of its primary hypothesis. It remains somewhat ambiguous whether the study aims to identify protective B cell antigens, test the suitability of VLP and VSV platforms for ASFV vaccination, or demonstrate the limitations of humoral immunity-focused approaches. The authors are encouraged to explicitly state the working hypothesis in the Introduction and to align the Results and Discussion more clearly with this premise.

2. Although the lack of protection is clearly demonstrated, the discussion does not sufficiently contextualize these findings. The authors should expand on why robust antibody responses failed to confer protection, particularly in light of prior reports suggesting the partial or platform-dependent efficacy of similar antigens. A more balanced discussion on antigen selection, epitope conformation, antigen density, and immune hierarchy in ASFV infection is required.

3. The transition from microarray-based antigen identification to vaccine formulation and in vivo testing requires stronger justification. It remains unclear whether antigenicity, as detected by immune sera, correlates with protective capacity. The authors should discuss the limitations of antibody-binding-based screening approaches for predicting vaccine efficacy more explicitly.

4. The conclusion that effective protection correlates with T cell-mediated immunity is well supported by the data, particularly in the Estonia 2014 group. However, this conclusion was introduced relatively late and would benefit from earlier framing. The authors should consider restructuring the Discussion to emphasize this point as a key conceptual outcome of the study rather than a secondary observation.

5. Given that this study reports negative efficacy results, it is essential to clearly articulate its contribution to ASFV vaccine development. The authors should emphasize how these findings inform future vaccine design strategies, including antigen prioritization, immune correlates of protection, and limitations of B cell-focused subunit vaccines against virulent genotype II strains.

 

Minor comments

1. Please ensure consistent terminology when referring to vaccine platforms, antigens, and immune readouts throughout the manuscript.

2. Some figure legends would benefit from additional details to ensure that they are fully interpretable without reference to the main text.

3. The Discussion section is relatively long and could be streamlined to reduce repetition while strengthening the key messages.

Author Response

Comment 1: This manuscript addresses an important and timely question in African swine fever virus vaccine research, namely the identification of protective B cell antigens and the evaluation of subunit- and vector-based vaccine platforms in a stringent pig challenge model. The study was technically well executed, included an extensive antigen screening approach, and employed appropriate in vivo validation using a virulent genotype II ASFV strain. The negative protection outcome was clearly demonstrated and supported by multiple independent readouts. However, in its current form, the manuscript requires major revisions to improve the clarity of interpretation, strengthen the conceptual framework, and more clearly articulate the scientific significance of the findings. The main concerns relate not to data quality but to data interpretation, narrative structure, and positioning of the conclusions.

Author's answer 1: We are very grateful to Reviewer 1 for taking the time to critically review our manuscript.

Comment 2: The manuscript would benefit from a clearer articulation of its primary hypothesis. It remains somewhat ambiguous whether the study aims to identify protective B cell antigens, test the suitability of VLP and VSV platforms for ASFV vaccination, or demonstrate the limitations of humoral immunity-focused approaches. The authors are encouraged to explicitly state the working hypothesis in the Introduction and to align the Results and Discussion more clearly with this premise.

Author's answer 2: We have addressed this point and re-written large parts of the Introduction (lines 69 to 94) to formulate our hypothesis and justify our antigen selection. Also in the results text lines 378-389, we now explain and justify our selections.

Comment 3: Although the lack of protection is clearly demonstrated, the discussion does not sufficiently contextualize these findings. The authors should expand on why robust antibody responses failed to confer protection, particularly in light of prior reports suggesting the partial or platform-dependent efficacy of similar antigens. A more balanced discussion on antigen selection, epitope conformation, antigen density, and immune hierarchy in ASFV infection is required.

Author's answer: We have rewritten the discussion, focusing on the failures of the vaccines and possible explanations. We address antigen and platform selection as well as possible epitope confirmation.

Comment 4: The transition from microarray-based antigen identification to vaccine formulation and in vivo testing requires stronger justification. It remains unclear whether antigenicity, as detected by immune sera, correlates with protective capacity. The authors should discuss the limitations of antibody-binding-based screening approaches for predicting vaccine efficacy more explicitly.

Author's answer 4: The text describing this transition in the results was rewritten (lines 378-389). We also addressed the lack of association between antigenicity and protection in the discussion (lines 653-657).  

Comment 5: The conclusion that effective protection correlates with T cell-mediated immunity is well supported by the data, particularly in the Estonia 2014 group. However, this conclusion was introduced relatively late and would benefit from earlier framing. The authors should consider restructuring the Discussion to emphasize this point as a key conceptual outcome of the study rather than a secondary observation.

Author's answer 5: Good point. The fact that our study supports the important role of T cells is now a first discussion point (lines 597-601).

Comment 6: Given that this study reports negative efficacy results, it is essential to clearly articulate its contribution to ASFV vaccine development. The authors should emphasize how these findings inform future vaccine design strategies, including antigen prioritization, immune correlates of protection, and limitations of B cell-focused subunit vaccines against virulent genotype II strains.

Authors' answer 6: Obviously, the study was published to provide information on antigenicity, to show the limitations or unsuitability of the vaccine platforms selected, and to show that the selected proteins with the platforms do not even confer partial protection. When rewriting the discussion, we took this request into consideration and also summarizing sentence as a “take home message” for these negative data.

Comment 7 (minor): Please ensure consistent terminology when referring to vaccine platforms, antigens, and immune readouts throughout the manuscript.

Author's answer 7: We systematically checked this point and made the required corrections.

Comment 8 (minor): Some figure legends would benefit from additional details to ensure that they are fully interpretable without reference to the main text.

Author's answer 8: Additional text was added when necessary (in particular to Figure 7).

Comment 9 (minor): The Discussion section is relatively long and could be streamlined to reduce repetition while strengthening the key messages.

Author's answer 9: The discussion was rewritten to take the above major points into consideration. This was associated with large deletions of text. Nevertheless, to address the major comments of both reviewers, it is shortened by “only” 20 lines.

Reviewer 2 Report

Comments and Suggestions for Authors

Comments on vaccines-4086488-peer-review-v1

Title: Exploring ASFV B-cell antigens: VLP vaccines elicit antibody responses but fail to protect against the virulent genotype II strain

 

The present study by Kirill Lotonin et al. showed that VLP vaccines based on six proteins of African swine fever virus (ASFV) elicited antibody responses but failed to protect against the virulent challenge. The study highlights the pivotal role of cellular immunity in protection. Several major concerns, including antibody functions and T-cell response mechanisms, need to be addressed.

Major concerns

1. Insufficient analysis of antibody functions: The study only detected antibody titers (via ELISA) but lacks assessments of antibody functionality. For example: Whether the induced antibodies have viral neutralizing activity (critical for evaluating B-cell-mediated protection against ASFV)? Whether antibodies mediate effector functions like antibody-dependent cellular cytotoxicity (ADCC) or phagocytosis (ADCP)?—particularly relevant for surface antigens like CD2v and EP153R, which are expressed on infected cells. Without these data, it is impossible to determine if the "detectable antibody responses" are functionally irrelevant, which directly limits the interpretation of why antibody induction failed to confer protection.
2. Superficial investigation of T-cell response deficits: The core finding (VLP/VSV vaccines failed to induce protective T-cell responses) is not sufficiently explored. For VLP vaccines: The study attributed weak T-cell responses to "limited antigens" or "lack of T-cell epitopes," but does not test whether modifying the VLP platform (g., adding adjuvants to enhance antigen presentation to dendritic cells) could improve T-cell activation. For VSV vectors: The replication-defective design was hypothesized to reduce immunogenicity, but no data (e.g., viral vector replication kinetics in pig cells, antigen expression levels) is provided to support this. Additionally, T-cell subset analysis (e.g., CD4⁺ Th vs. CD8⁺ CTL activation) is insufficient to clarify which arm of cellular immunity is missing.
3. Incomplete rationale for antigen selection: While six antigens were selected for vaccine development, the functional relevance of newly identified antigens (g., B169L, H171R) is not fully validated. The study mentioned B169L has "viroporin-like activity" and H171R is a "virion component," but lacks data on whether these proteins are involved in key viral processes (e.g., entry, replication) or if they contain T-cell epitopes. This weakens the justification for their selection as vaccine candidates.
4. Gap between laboratory and field applicability: The study used SPF pigs, which have no prior pathogen exposure and a more robust immune system than commercial farm pigs. However, it does not discuss how this may affect the translational potential of the findings—g., whether the tested vaccines would perform differently in field conditions (where pigs may have pre-existing immunity or co-infections).
5. Writing errors: The manuscript has many mistakes and writing should be improved.

For more details, please refer to the attachment.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The manuscript writing should be improved.

Author Response

Comment 1: The present study by Kirill Lotonin et al. showed that VLP vaccines based on six proteins of African swine fever virus (ASFV) elicited antibody responses but failed to protect against the virulent challenge. The study highlights the pivotal role of cellular immunity in protection. Several major concerns, including antibody functions and T-cell response mechanisms, need to be addressed.

Author’s answer 1: We are very grateful to Reviewer 2 for taking the time to critically review our manuscript.

Comment 2. Insufficient analysis of antibody functions: The study only detected antibody titers (via ELISA) but lacks assessments of antibody functionality. For example: Whether the induced antibodies have viral neutralizing activity (critical for evaluating B-cell-mediated protection against ASFV)? Whether antibodies mediate effector functions like antibody-dependent cellular cytotoxicity (ADCC) or phagocytosis (ADCP)?—particularly relevant for surface antigens like CD2v and EP153R, which are expressed on infected cells. Without these data, it is impossible to determine if the "detectable antibody responses" are functionally irrelevant, which directly limits the interpretation of why antibody induction failed to confer protection.

Author’s answer 2: We agree that such antibody functions may be relevant for immune protection against ASFV. We have now addressed their importance in the Introduction, and formulate our hypothesis and justify our antigen selection (lines 79-94). Also in the results text lines 378-389, we now explain and justify our selections. In the final paragraph we also conclude that future studies should take such antibody functions into consideration when selecting antigens and platforms (lines 651-661).

Unfortunately, we are not able to perform functional analyses, as the project and funding is terminated. Furthermore, given the observation that the vaccination did not change the course of disease (VLP vaccines) or only induced weak antibody responses (VSV vaccines), we think that our sera collection is not suitable to demonstrate the contribution of such effector functions to protection.

 

Comment 3: Superficial investigation of T-cell response deficits: The core finding (VLP/VSV vaccines failed to induce protective T-cell responses) is not sufficiently explored. For VLP vaccines: The study attributed weak T-cell responses to "limited antigens" or "lack of T-cell epitopes," but does not test whether modifying the VLP platform (g., adding adjuvants to enhance antigen presentation to dendritic cells) could improve T-cell activation. For VSV vectors: The replication-defective design was hypothesized to reduce immunogenicity, but no data (e.g., viral vector replication kinetics in pig cells, antigen expression levels) is provided to support this. Additionally, T-cell subset analysis (e.g., CD4⁺ Th vs. CD8⁺ CTL activation) is insufficient to clarify which arm of cellular immunity is missing.

Author’s answer 3: In our opinion, what the reviewer is proposing is something for future studies that should be designed at also targeting the T-cell arm of immune defence against ASFV. Our approach was with a focus on antibody responses and we now more clearly articulate this in the Introduction (lines 69-94).  In fact, the VLP platform is “by design” excellent for antibodies but not for T cells.

For the VSV platform our previous work with HA, NA and NP of influenza virus showed that the VSV induces detectable T cell responses in pigs as well as very good antibody responses implying good in vivo antigen expression levels in pigs (we now cite those references). Accordingly, the lack of T cell responses is likely to be attributed to the ASFV antigens (although we had expected to detect responses).

With respect to T cell analyses, our data shows that neither VLP nor VSV induced any CD4 or CD8 T cell response detectable at the time of challenge. If CD8 T cells are not restimulated to produce cytokines, it is highly unlikely to find cytotoxic activity. Given the aim and outcome of the study, we do not think that it is a promising area of investigation with the vaccine candidates used. Having said that, our study underlines their importance and their consideration for future platform and antigen selection (see new Discussion line 595-606 and the final paragraph of the discussion).

 

Comment 4:  Incomplete rationale for antigen selection: While six antigens were selected for vaccine development, the functional relevance of newly identified antigens (g., B169L, H171R) is not fully validated. The study mentioned B169L has "viroporin-like activity" and H171R is a "virion component," but lacks data on whether these proteins are involved in key viral processes (e.g., entry, replication) or if they contain T-cell epitopes. This weakens the justification for their selection as vaccine candidates.

Author’s answer 4: We have addressed this point and rewritten large parts of the Introduction (lines 69 to 94) to formulate our hypothesis and justify our antigen selection. Also in the results text lines 378-389, we now explain and justify our selections. The reviewer is correct that the B169L and H171R are more difficult to justify, as their function is not very clear. We point on this problem in the discussion in lines 653-657.

Comment 5: Gap between laboratory and field applicability: The study used SPF pigs, which have no prior pathogen exposure and a more robust immune system than commercial farm pigs. However, it does not discuss how this may affect the translational potential of the findings—g., whether the tested vaccines would perform differently in field conditions (where pigs may have pre-existing immunity or co-infections).

Author’s answer 5: SPF pigs are often used for vaccination studies to reduce the impact of environmental factors and possibly pre-existing immunity. We agree that a more advanced vaccine candidate should also be tested in farm pigs, but sadely we are very far away from this with the present vaccine candidates.

Comment 6: Writing errors: The manuscript has many mistakes and writing should be improved.

Author’s answer 6: We have carefully checked the manuscript and made many corrections.  

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised manuscript has substantially improved compared to the previous version and adequately addresses the major concerns raised during the initial review.

The authors have clarified the primary working hypothesis in the Introduction and have more clearly aligned the experimental design, Results, and Discussion with this conceptual framework. In particular, the revised Discussion now provides a balanced and thoughtful interpretation of negative protection outcomes, appropriately contextualizing them within the current landscape of African swine fever virus vaccine research.

Importantly, the manuscript no longer presents the lack of protection as an isolated negative result but rather as informative evidence highlighting the limitations of antibody-focused subunit and vector-based vaccine approaches against virulent genotype II ASFV strains. The strengthened emphasis on cellular immunity, supported by a comparison with the ASFV Estonia 2014 group, represents a clear and well-substantiated conceptual conclusion of this study.

The transition from antigenicity-based screening to in vivo vaccine evaluation is now more transparently discussed, including explicit acknowledgment of the limitations of antibody binding assays in predicting protective efficacy. This significantly improves the scientific rigor and interpretability of this study.

Overall, although the study did not demonstrate protective vaccine efficacy, it provides valuable guidance for future antigen prioritization and vaccine design strategies in the ASFV field. The data are technically sound and clearly presented, and the conclusions are appropriately framed.

I consider this manuscript suitable for publication in its current form. Any remaining issues are minor and editorial in nature and do not require further experimental work to address them.

Author Response

response: Thank you for the time and effort you have dedicated to reviewing our manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript has been imroved marginally and should be revsied comprehensively.

The authors did not follow the corrections in the attachment.

1. Antibody levels against the CD2v and EP153R proteins in the VSV mix group were not significantly different from those of the empty VLP group, which indicates a failed immunization.
2. The title is inappropriate; this article presented two strategies for displaying antigens, but lacks logical coherence.
3. which domain of CD2v was expressed?
4. Section 3.2 did not explain why antigens screened in Section 3.1 were not selected.
5. The molecular markers, target proteins, AP205, and the VLPs formed should be displayed on a single nitrocellulose membrane in Figure 2.
6. Since various antigens in Figure 2 induced high-level antibodies, the piglets immunized with CD2v-VLP alone should be included as one group and the other four VLPs as another group.
7. Please justify the selection of EP153R in the VSV construct.
8. The labeling of the x-axis and y-axis in Figures 5, 7, and 8 should be checked.
9. What is the sample source for the cell types detected in Figure 7?
10. Figure 7 only showed antibody levels, without neutralizing antibodies against ASFV or the cell types involved.
11. TCID50 was not correct (TCID₅₀).
12. Section 3.1 needs references.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The manuscript writing should be improved.

Author Response

Comment 1:

The manuscript has been imroved marginally and should be revsied comprehensively.

The authors did not follow the corrections in the attachment.

Author’s answer 1: We apologize that we have missed to proposed correction in the pdf, and thank the reviewer for the effort in identifying the mistake. We have implemented most of them in the text and the Figures in particular with respect to unifying the nomenclature, improving the axis descriptions (either by chaing them or adding explanations in the Figure legends). Please note that we now use VLP as abbreviations for “Virus like particles” as well as PBMC instead of PBMCs throughout the text.

 

 

Comment 2

1. Antibody levels against the CD2v and EP153R proteins in the VSV mix group were not significantly different from those of the empty VLP group, which indicates a failed immunization.
2. The title is inappropriate; this article presented two strategies for displaying antigens, but lacks logical coherence.

Author’s answer 2: The reviewer is correct. We have changed the title to “Virus-like and virus replicon particles targeting multiple B cell antigens do not protect against African swine fever virus”. In the new version we explain our approach and its logic. Of course in hindsight given the outcome, this was a bad choice, but we think that it is  and ethical to publish also negative results. Regarding the use of the two platfomrs, it should be mentioned that the original plan of the work was to only use VLP but given the difficult to express the glycoproteins we included the VSV platform.   

Comment 3

3. which domain of CD2v was expressed?

Author’s answer 3: It was the extracellular domain. The sequence is available in the Supplementary Figure 2. We now refer to this information in line 159.  

Comment 4

4. Section 3.2 did not explain why antigens screened in Section 3.1 were not selected.
   
   

Author’s answer 4: In the new version of the manuscript we explain our selection criteria in the intrduction and refer to this paragraph at the beginning of Section 3.2. On one side we decided to test novel antigens that were recognized by serum from protected pigs only and for those we decided excluded many of the well-characterized major antigens (see Figure 1D), on the other side we focussed on antigens expressed in the membrane of infected cells and CD2v as a outer membrane protein.

Comment 5

5. The molecular markers, target proteins, AP205, and the VLPs formed should be displayed on a single nitrocellulose membrane in Figure 2.

 

Author’s answer 5: The original pictures of these membranes are found in Supplementary Figure 3. The cropping of the images in Figure 2 was done to reduce the space required for those plots.

Comment 6

6. Since various antigens in Figure 2 induced high-level antibodies, the piglets immunized with CD2v-VLP alone should be included as one group and the other four VLPs as another group.

 

Author’s answer 6: Yes, this grouping is depicted in Figure 4. .

 

Comment 7

7. Please justify the selection of EP153R in the VSV construct.

Author’s answer 7: The justification for the selection of EP153R is given in line 86-90 of the Introduction (its expression on the sruface of infected cells as well as its high genetic heterogeneity indicating the selection of possible antibody escape mutants).

Comment 8

8. The labeling of the x-axis and y-axis in Figures 5, 7, and 8 should be checked.

Author’s answer 8: We have revised the x- and y-axis labelling of these Figures, or added precise definitions of abbreviations for the axis in the Figure legends.

 

Comment 9

9. What is the sample source for the cell types detected in Figure 7?
   
   

Author’s answer 9: Figure 7 shows antibody responses and Figure 8 PBMC (as written in the Figure legends).

Comment 10

  10   Figure 7 only showed antibody levels, without neutralizing antibodies against ASFV or the cell types involved.

 

Author’s answer 10: We did not perform neutralization assays because sera from ASFV immune pigs to not neutralize infection of macrophages (at least in our hands and of many other labs). We are aware that many other assays could have been performed but due to the lack of any protective effects we think that further investigations are probably not rewarding with respect to the aim at hand. In addition, the poor funding for this project limited our possibilities to perform further investigations. Accordingly, the aim of this Figure was only to demonstrate which vaccine candidates were immunogenic.

 

Comment 11

10. TCID50 was not correct (TCID50).

 

Author’s answer 11: This is now corrected.

 

Comment 12

12. Section 3.1 needs references.

Author’s answer 12: We have added references as requested.