High-Avidity Anti-Filovirus IgG Elicited Using Protein Subunit Vaccines Does Not Correlate with Protection

Round 1

Reviewer 1 Report

Authors have presented a study of antibody responses to recombinant protein subunit vaccines adjuvanted with a novel oil in water emulsion adjuvant. A comparison of avidity of antibodies elicited by survivors versus non-survivors was done. Authors show that avidity profiles are not a predictor of survival in context of this vaccine, and therefore cannot be used as a corelate of protection. While the findings are clearly presented and conclusions are fair, authors must include the conclusion in the title of the manuscript. An example can be:

"High Avidity Anti-Filovirus IgG elicited by Protein Subunit Vaccines is not a predictor of corelates of protection."

Minor comment:

Authors mention in the discussion that the waning response towards MARV-GP, over the period of one year could be due to the immune response being directed towards conserved regions on the ebolavirus glycoproteins. It is not very clear what is implied by this. Authors must elaborate on this. While the initial responses are similar, is the MARV-GP less immunogenic? Why will the animals' immune system prefer to have a more prolonged response against EBOV GP and SUDV GP?

There are some gaps and formatting errors along with minor typographical errors. This can be improved and will probably be taken care of by the publication team.

Author Response

The authors thank the reviewer for their time and consideration. We hope to have addressed the reviewers’ concerns.

While the findings are clearly presented and conclusions are fair, authors must include the conclusion in the title of the manuscript. An example can be:

"High Avidity Anti-Filovirus IgG elicited by Protein Subunit Vaccines is not a predictor of corelates of protection."

            We have updated the title to include the main findings:

“High Avidity Anti-Filovirus IgG elicited by Protein Subunit Vaccines Does Not Correlate with Protection.”  

 

Minor comment:

Authors mention in the discussion that the waning response towards MARV-GP, over the period of one year could be due to the immune response being directed towards conserved regions on the ebolavirus glycoproteins. It is not very clear what is implied by this. Authors must elaborate on this. While the initial responses are similar, is the MARV-GP less immunogenic? Why will the animals' immune system prefer to have a more prolonged response against EBOV GP and SUDV GP?

We believe the differences in antibody levels between ebolavirus spp and marburgvirus lies in the identity shared between these three glycoproteins and have added phrasing to clarify this point (line 414-417)

Reviewer 2 Report

In this article the authors wanted to assess and investigate vaccine elicited antibody avidity as a potential correlate of protection, with particular emphasis on vulnerable populations. They then wanted to further characterise the role of antibody in protective and non-protective vaccine responses. The authors state that understanding vaccine correlates of protection is an important and vital component moving towards the development of alternative vaccine strategies for pregnant women, young children and immunocompromised individuals.

The authors point out that there are currently very limited vaccine options available to prevent Ebola virus disease and importantly, the duration of protective efficacy for these particular vaccines is currently unknown. This is because there are no known vaccine correlates of protection defined as “measurable immunological responses which correlate with statistical significance to a second measurable aspect of protection such as decreased viremia, or survival day after infection.” The authors highlight the point that they have previously conducted several pre-clinical studies in non-human primates, and they state that they have established a vaccine platform capable of fully protecting these animals upon viral challenge with a lethal dose of the Kikwit strain of Zaire ebolavirus. Therefore, in order to assess the vaccine correlates of protection and to be able to further characterise the humoral response elicited by their vaccines, the authors compared the avidity of IgG developed by animals which survived a viral challenge and those that did not.

In the present study, the authors wanted to investigate the avidity and binding of various antibodies under low pH and chaotropic conditions of protective, partially protective, and non-protective humoral responses. They conclude that while protein subunit vaccines elicit high avidity antibody, differential binding under chaotropic conditions was not a predictor of survival. Antibody binding strength was very important for neutralising the virus, avidity varied with the number of immunisations and the time that had elapsed since the last vaccination and was also elevated following increased antibody maturation. The authors highlight that multiple vaccine doses are able to enhance the avidity maturation as well as a cross-reactive response to Sudan ebolavirus Glycoprotein. However, any antibody that remained one year post challenge was of moderate avidity. Vaccination against Zaire ebolavirus has demonstrated significant efficacy in the past, but the protective cellular and humoral responses involved are still very poorly defined and poorly understood.

This manuscript has been well written with clearly presented data and sensible conclusions based on the results that have been generated. It was a pleasure to read. I am also delighted to see that serum samples were used from previously completed studies as this demonstrates good use of resources without culling more non-human primates. The data in this article highlight just how complicated this area of research is and how variable the results can be depending on treatment type, sampling time, dosing regimen and pH.

I am slightly surprised that this article only cites 15 references which seems a little slim considering the field of research. However, the references that are cited are relevant.

Main points and comments:

1.     How many replicate assays were used for each data point in Figure 2a?

2.     Can the authors comment on whether they expected to see more of a change in the curves in Figure 3c? Obviously, there is a greater area-under-the-curve for Figure 3c than 3b but did they expect a bigger distance between the grey area and the top of the curves for the survivors?

3.     How do the authors account for the fact that the 2 surviving animals showed similar results in Figure 4c as animal number CA881A? A similar situation appears to occur in Figure 4b as well.

4.     Animal C22081 does seem to have a completely different reactivity in comparison with the other animals shown in Figure 4. Can the authors make any comments to explain why this may be the case? What accounts for the differences between CA881A and C22081 bearing in mind they were both non-survivors?

5.     Can the authors comment on why there are lower antibody levels and lower avidity after three doses of vaccine for the Marburg marburgvirus Glycoprotein as demonstrated in Figures 6 and 7?

6.     The placement of the surviving animals on Figure 6 seems to be a little bit unexpected. Can the authors comment on the results shown in this Figure please (with regards to the survivors)?

7.     Figure 8 and it’s multiple components are a little bit too small to be able to see the individual symbols with any clarity. Please can this be addressed?

8.     Why do the authors think the responses generated by the three-dose regimen start to wane after 20 weeks?

9.     If strongly bound antibodies do not correlate with survival, what is this actually telling us? What conclusions can be drawn from the data demonstrating that animals with similar relative avidity have no real correlation with survival or survival day?

10.  The low numbers of animals that are used in these types of studies may have an affect on the outcome of the comparisons. Do the authors think that using larger numbers of smaller animals (ie guinea pigs for example), would give more robust data or would the likely outcomes be similar to the data seen here with non-human primates?

11.  Statistical analyses used within this article were suitable for this type of research.

Minor points

12.  Materials and Methods section 2.2.  Lines 104 and 121 appear to have an issue with the degrees centigrade symbol (although it was fine on line 108).

13.  Line 101 and 109 have uL but line 106 has µL. Can these be made consistent please? Also, please ensure they either all have a space between the number and the units (as in line 106) or they are close together (as in lines 101 and 109).

14.  Somewhere in lines 142 and 143, there appears to be a word or two missing. Please rectify this.

15.  Results. Table 1. Please ensure the ug symbol ties in with the uL/ µL symbol mentioned above (ie µg or ug).

16.  The words “Study Week” are missing from Figure 2a.

17.  Figure 9 legend is incorrect in that Figure 9b should be listed as pH 5.5 and Figure 9c should be listed as pH 6 (assuming the Figure itself is correct).

18.  Lines 354 to 356 appear to have a word missing.

This manuscript has been well written with clearly presented data and sensible conclusions based on the results that have been generated.  It was a pleasure to read. 

Author Response

Reviewer #2

We thank the reviewer for their time and consideration. Below please find responses to each comment, changes made to the manuscript are highlighted in yellow.

Main points and comments:

1. How many replicate assays were used for each data point in Figure 2a?

We have updated the figure to include the total number of animals in each group and for clarity further indicated that samples were run in duplicate.

 

2. Can the authors comment on whether they expected to see more of a change in the curves in Figure 3c? Obviously, there is a greater area-under-the-curve for Figure 3c than 3b but did they expect a bigger distance between the grey area and the top of the curves for the survivors?

We thank the reviewer for noting this and have added a sentence to the discussion (Line 404-407) to further highlight findings at this timepoint. We would like not to speculate too much and hope this addresses the reviewers’ concerns.

 

3. How do the authors account for the fact that the 2 surviving animals showed similar results in Figure 4c as animal number CA881A? A similar situation appears to occur in Figure 4b as well.

We believe the responses found in animal CA881A highlight the need for a combination approach of strongly binding antibody with specificity for neutralizing epitopes. Given that our analysis is of polyclonal responses and did not differentiate between neutralizing and non-neutralizing epitopes we believe this can partially explain these differences. We have added phrasing to the discussion to clarify this point (Line 424-427).

 

4. Animal C22081 does seem to have a completely different reactivity in comparison with the other animals shown in Figure 4. Can the authors make any comments to explain why this may be the case? What accounts for the differences between CA881A and C22081 bearing in mind they were both non-survivors?

We are not sure of why these responses varied, C22081 has overall lower antigen specific IgG 20 weeks post immunization which may account for these differences. We are unable to define the mechanisms behind the different rates of waning. We have included the untreated antigen specific levels over the study as a supplemental figure.

 

5. Can the authors comment on why there are lower antibody levels and lower avidity after three doses of vaccine for the Marburg marburgvirus Glycoprotein as demonstrated in Figures 6 and 7?

The authors believe the difference in identity between these three spike proteins is responsible for the levels of similarity/differential rates of waning (Line 414-417)

 

6. The placement of the surviving animals on Figure 6 seems to be a little bit unexpected. Can the authors comment on the results shown in this Figure please (with regards to the survivors)?

The authors believe that the cross-reactive antibodies are targeting conserved regions of the glycoproteins which may not play an active role in mitigating viral neutralization of the infecting virus.

 

7. Figure 8 and it’s multiple components are a little bit too small to be able to see the individual symbols with any clarity. Please can this be addressed?

This figure has been resized for readability.

 

8. Why do the authors think the responses generated by the three-dose regimen start to wane after 20 weeks?

This follows typical vaccine response kinetics also observed for other vaccines. See for example To et al. 2022, doi=10.1021/acsinfecdis.1c00600

 

9. If strongly bound antibodies do not correlate with survival, what is this actually telling us? What conclusions can be drawn from the data demonstrating that animals with similar relative avidity have no real correlation with survival or survival day?

We have updated the existing sentences in the discussion section that addressed this concern to increase clarity. Indeed avidity seems to mostly be correlated with indication of a mature immune response, but less indicative of protection as we now also clearly state in the title. We hope this addressed the reviewers’ comment (Line 424-427)

 

10. The low numbers of animals that are used in these types of studies may have an affect on the outcome of the comparisons. Do the authors think that using larger numbers of smaller animals (ie guinea pigs for example), would give more robust data or would the likely outcomes be similar to the data seen here with non-human primates?

The authors have previously published on small animal experiments using rodent-adapted EBOV. Unfortunately, we and others in the field of filovirus vaccines have repeatedly experienced discordant rates of protection between rodents and non-human primates with concordant antigen specific antibody levels. Therefore, we believe that only non-human primates can serve as an ideal animal model defining final human vaccine candidates. The authors would recommend further analysis with either an increased number of non-human primates, or an analysis of convalescent serum from survivors to further study the relevance of our observations in a human clinical context.

11. Statistical analyses used within this article were suitable for this type of research.

Minor points

11. Materials and Methods section 2.2.  Lines 104 and 121 appear to have an issue with the degrees centigrade symbol (although it was fine on line 108).

This has been corrected.

 

12. Line 101 and 109 have uL but line 106 has µL. Can these be made consistent please? Also, please ensure they either all have a space between the number and the units (as in line 106) or they are close together (as in lines 101 and 109).

These have been adjusted to utilize the “µ” symbol with uniform spacing.

 

13. Somewhere in lines 142 and 143, there appears to be a word or two missing. Please rectify this.

This sentenced has been corrected.

 

14. Table 1. Please ensure the ug symbol ties in with the uL/ µL symbol mentioned above (ie µg or ug).

This has been corrected.

 

15. The words “Study Week” are missing from Figure 2a.

The figure has been updated.

 

16. Figure 9 legend is incorrect in that Figure 9b should be listed as pH 5.5 and Figure 9c should be listed as pH 6 (assuming the Figure itself is correct).

The figure legend has been updated.

 

17. Lines 354 to 356 appear to have a word missing.

The sentence has been corrected.