Development of DuoChol, a Thermostable Inactivated Whole-Cell/B-Subunit Oral Cholera Vaccine in Enteric Capsule

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript describes the development and preclinical evaluation of DuoChol, a thermostable oral cholera vaccine formulated as a lyophilized dry powder in enteric capsules. The candidate contains formalin-inactivated Vibrio cholerae O1 El Tor Inaba and Ogawa strains, recombinant cholera toxin B subunit, and sucrose as stabilizer. The study addresses an important public-health problem, namely the need for oral cholera vaccines with improved thermostability, simplified logistics, and potentially lower production costs. The work is technically interesting, especially in the construction of isogenic vaccine strains, the development of a mixed-mode chromatography process for rCTB purification, and the long-term stability assessment of the dry capsule formulation. However, several conclusions are currently overstated relative to the data presented. The manuscript mainly demonstrates antigen stability and immunogenicity in mice, but it does not provide direct in vivo protection data, true capsule delivery data in the animal model used, sufficient safety characterization, or robust evidence for manufacturing scalability and cost-effectiveness. Several key claims are also supported by “data not shown” or unpublished studies. Therefore, I recommend major revision before the manuscript can be considered further.

1. The manuscript repeatedly suggests that DuoChol has protective potential, improved efficacy, and suitability for outbreak response. However, the current animal data are limited to immunogenicity readouts, including serum IgG/IgA, fecal IgA, and vibriocidal antibody titers. No live V. cholerae challenge model, infant mouse colonization model, passive protection assay, or other direct protection experiment is presented.
2. The manuscript emphasizes the enteric capsule as a key innovation. However, in the mouse immunogenicity study, DuoChol capsules were dissolved in PBS and then administered intragastrically with sodium bicarbonate buffer. This design evaluates the antigenic content after storage, but it does not directly test intact capsule survival in gastric conditions, intestinal release, or delivery performance in vivo.
3. The reported 21-month stability at 4–40 °C is one of the main strengths of the paper, but the current evidence appears to rely on a limited prototype lot and a small number of capsules per condition and time point. The authors should clarify whether the stability study was performed with one manufacturing lot or multiple independent lots. If only one lot was tested, this limitation should be clearly acknowledged.
4. Important claims in the manuscript are supported by statements such as “data not shown” or “to be published,” including capsule acid resistance/dissolution, co-culture stability, mass spectrometry characterization, and pig capsule-delivery studies. These points are central to the manuscript’s novelty and translational relevance.
5. The authors state that DuoChol may offer affordable production and large-scale manufacturing advantages. However, the production described in the manuscript is still laboratory-scale, with 3 L fermentations, manual powder preparation, and prototype capsule filling. This is not sufficient to support strong claims about industrial scalability or cost parity.

Author Response

Responses to reviewer 1´comments

We thank the reviewer for the careful examination of our manuscript and for the criticism, comments and suggestions which we have taken due note of and aligned with in our revision of the manuscript, as specified in our point-by-point response to the comments.

Reviewer 1

Comments and Suggestions for Authors

The manuscript describes the development and preclinical evaluation of DuoChol, a thermostable oral cholera vaccine formulated as a lyophilized dry powder in enteric capsules. The candidate contains formalin-inactivated Vibrio cholerae O1 El Tor Inaba and Ogawa strains, recombinant cholera toxin B subunit, and sucrose as stabilizer. The study addresses an important public-health problem, namely the need for oral cholera vaccines with improved thermostability, simplified logistics, and potentially lower production costs. The work is technically interesting, especially in the construction of isogenic vaccine strains, the development of a mixed-mode chromatography process for rCTB purification, and the long-term stability assessment of the dry capsule formulation. However, several conclusions are currently overstated relative to the data presented. The manuscript mainly demonstrates antigen stability and immunogenicity in mice, but it does not provide direct in vivo protection data, true capsule delivery data in the animal model used, sufficient safety characterization, or robust evidence for manufacturing scalability and cost-effectiveness. Several key claims are also supported by “data not shown” or unpublished studies.

1. The manuscript repeatedly suggests that DuoChol has protective potential, improved efficacy, and suitability for outbreak response. However, the current animal data are limited to immunogenicity readouts, including serum IgG/IgA, fecal IgA, and vibriocidal antibody titers. No live V. cholerae challenge model, infant mouse colonization model, passive protection assay, or other direct protection experiment is presented.

Response: There is no useful animal model to assess protective activity of cholera vaccines, since infant mouse or rabbit infection models take place in animals that are too young to develop an active immune response. Passive protection in these models - or other adult animal models - using immune sera from vaccinated mice also give little useful information (about predicted protective efficacy as a vaccine in humans) since serum antibodies – in contrast to intestinal-mucosal antibodies - are neither responsible nor particularly predictive of in vivo protective efficacy. The best animal model prediction test of future protective efficacy in humans is in our opinion – and we may be regarded as world leading experts on these matters after having been much involved in the development of several of the currently used “previous generation” inactivated OCVs including Dukoral, OrcVax/Shanchol/Euvichol, and more recently India-licensed Hillchol – is what is used in the current study to compare the immunogenicity of a new “similar” OCV candidate with the closest existing already licensed and WHO-prequalified OCV (“similar” with regard to composition [inactivated whole-cells ± CTB], dosage, and planned schedule) which in our case is Dukoral with regard to both the relevant types of serum  antibodies and, even more important, the intestinal mucosal IgA anti-LPS and anti-CTB antibody responses. The fully comparable immunogenicity in these regards we demonstrate for DuoChol visavi the comparator Dukoral OCV is the best predictive measure of future “clinical” efficacy (and such “non-inferior” immunogenicity is probably together with safety what will also be mainly assessed by regulatory authorities for potential future licensure of this vaccine). We have now added in the revised manuscript a brief comment on these aspects (Discussion lines 652-662).

2. The manuscript emphasizes the enteric capsule as a key innovation. However, in the mouse immunogenicity study, DuoChol capsules were dissolved in PBS and then administered intragastrically with sodium bicarbonate buffer. This design evaluates the antigenic content after storage, but it does not directly test intact capsule survival in gastric conditions, intestinal release, or delivery performance in vivo.

Response: This criticism is valid, which is the reason why we mention about the follow-up work in collaboration with a veterinarian university - to be published separately – showing immunogenicity with the undissolved capsule formulation in a pig model. We have however done our own studies in vitro to validate/extend the capsule manufacturer´s studies on  “gastric conditions, intestinal release, and delivery performance in vivo” and are now including some of this data in the revised manuscript (Results lines lines 491-495) + that we have added 2 refs (refs 35,38 reporting clinical studies further supporting the expected properties of these capsules.

3. The reported 21-month stability at 4–40 °C is one of the main strengths of the paper, but the current evidence appears to rely on a limited prototype lot and a small number of capsules per condition and time point. The authors should clarify whether the stability study was performed with one manufacturing lot or multiple independent lots. If only one lot was tested, this limitation should be clearly acknowledged.

Response: Only one prototype lot has been tested for 21-month stability, but another later similar lot – prepared in 30-liter scale under GMP-like conditions - has proved to be equally stable both in vitro at different temperatures for >6 month as well as in vivo as tested after 3 months storage.  The limitation is now spelled out in the revision (Results lines 524-530) .

4. Important claims in the manuscript are supported by statements such as “data not shown” or “to be published,” including capsule acid resistance/dissolution, co-culture stability, mass spectrometry characterization, and pig capsule-delivery studies. These points are central to the manuscript’s novelty and translational relevance.

Response: The pig study was done with a different set of collaborators and will be published separately with these people as lead authors. With regard to the other “not shown” we now add dat  (Supplemental file, section C) on “co-culture stability” and “capsule acid resistance/dissolution” (Supplemental file, section D) as well further text info on the pre-pentamer CTB protein material (Results line 413-476, Figure3 legend).

 

5. The authors state that DuoChol may offer affordable production and large-scale manufacturing advantages. However, the production described in the manuscript is still laboratory-scale, with 3 L fermentations, manual powder preparation, and prototype capsule filling. This is not sufficient to support strong claims about industrial scalability or cost parity.

Response: We agree that this is an overstatement (only excused by our extensive previous experience in taking lab-scale OCVs to industrial products – Dukoral, OrcVAX, Schanchol, Hillchol). We have toned down and explained this “claim” in Discussion (lines 663-666) and deleted it in the Abstract.

 

We hope that with these responses and the revisions made the revised manuscript can now be accepted for publication.

Sincerely,

Manuela Terrinoni & Jan Holmgren

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript reports the development and preclinical evaluation of DuoChol, a thermostable formulation of an inactivated whole-cell/rCTB oral cholera vaccine. The concept of combining lyophilized inactivated vibrios with purified rCTB for cold-chain–independent delivery is worthwhile, and the long-term thermostability data at 40 °C are a genuine contribution. However, the manuscript has fundamental gaps in statistical methodology, method reproducibility, and evidence supporting the central equivalence claim, which precludes acceptance in its current form.

Major Comments

1. No statistical methods are described in the Methods section. The specific tests used, multiple comparison corrections applied, significance thresholds, and software should be fully reported so that readers can verify the p-values presented in the Results, Materials and Methods, and Figure legends (Methods).
2. The central claim that DuoChol elicits immune responses “fully comparable” to those of Dukoral is statistically unsupported.
3. The choice of F1 hybrid mice (CBA × C57BL/6) for oral cholera vaccine immunogenicity assessment was not justified. The rationale for selecting this crossover, commonly used inbred strains such as BALB/c or C57BL/6, should be explained with reference to the prior literature on their mucosal immune responses.
4. The dose selection was unexplained.
5. Figures 6 and 7 contain no statistical annotations (p-values, significance indicators or test identifiers). Without these, the reader cannot assess the claimed group differences or equivalences.
6. The stability data shown in Figure 5 are presented without error bars.
7. The SDS-PAGE gels in Figures 3 and 5 lack molecular weight markers, preventing the verification of band assignments. Appropriate molecular weight (MW) ladders should be included.
8. Individual animal data are not shown in any of the immunogenicity figure. Given n = 7 per group and the equivalence claim, individual data points should be displayed (e.g., as a swarm plot or overlaid on bar charts) so that readers can evaluate inter-animal variability.
9. The vibriocidal antibody assay protocol was not described.
10. Claims of “affordable production cost” and “cost-effective” are made without any cost data, bill-of-materials analysis, or comparison with existing vaccines. These statements should either be removed or supported with quantitative manufacturing cost estimates.
11. The title claims “Next-Generation Thermostable Oral Cholera Vaccine,” but no challenge experiment or functional neutralization assay was performed. While thermostability is supported by in vitro data, the term “Next-Generation” is an over-claim. A more conservative title, such as “Thermostable Formulation of an Inactivated Whole-Cell/rCTB Oral Cholera Vaccine with Preclinical Immunogenicity” would better reflect the evidence.

Minor Comments

1. In Section 2.2, several numerical subscripts in chemical formulas or compound names are rendered as regular text rather than subscripts.
2. The x-axis of Figure 5 (time) is non-linear, which may mislead readers into perceiving uniform-sampling intervals. Equal spacing or an explicit note of the nonlinear scale should be used.
3. The sucrose concentration range of 50–100 mg/mL stated on page 5 is too broad for a defined formulation. The final optimized concentration should be specified in the manuscript.

Author Response

Responses to reviewer 2´comments

We thank the reviewer 2 for the careful examination of our manuscript and for the criticisms, comments and suggestions which we have taken due note of and aligned with in our revision of the manuscript, as specified in our point-by-point response to the comments.

Reviewer 2

Comments and Suggestions for Authors

This manuscript reports the development and preclinical evaluation of DuoChol, a thermostable formulation of an inactivated whole-cell/rCTB oral cholera vaccine. The concept of combining lyophilized inactivated vibrios with purified rCTB for cold-chain–independent delivery is worthwhile, and the long-term thermostability data at 40 °C are a genuine contribution. However, the manuscript has fundamental gaps in statistical methodology, method reproducibility, and evidence supporting the central equivalence claim, which precludes acceptance in its current form.

Major Comments

1. No statistical methods are described in the Methods section. The specific tests used, multiple comparison corrections applied, significance thresholds, and software should be fully reported so that readers can verify the p-values presented in the Results, Materials and Methods, and Figure legends (Methods).

Response: This information has now been fully provided (Methods lines 328-333, Results 523-530 lines 557-575 and Figure legends Fig 5, 6, 7) together with the requested information (reviewer´s point 8) on values for individual animals that are shown in revised Figs 6 & 7.

1. The central claim that DuoChol elicits immune responses “fully comparable” to those of Dukoral is statistically unsupported.

Response: This has now been further described with the requested statistical support, see Results lines 556 and forward..

1. The choice of F1 hybrid mice (CBA × C57BL/6) for oral cholera vaccine immunogenicity assessment was not justified. The rationale for selecting this crossover, commonly used inbred strains such as BALB/c or C57BL/6, should be explained with reference to the prior literature on their mucosal immune responses.

Response: BALB/C and C57BL/6 are well known for displaying, respectively, Th2 and Th1 biased immune responses, a bias which is avoided in their heterozygous F1 hybrid off-spring. Citing Google´s A1 response on “F1 hybrid mice and Th1 versus Th2 immune response bias” summarizes this with 5 refs one of which is only cited together with the A1 response ref: “Inbred mouse strains exhibit strong genetic biases toward either Th1 (cellular) or Th2 (humoral) immune responses. Crossing these strains creates \(F_{1}\) hybrids (e.g., C57BL/6 \(\times \) BALB/c, generating CB6F1 mice) that typically display a co-dominant, balanced, or intermediate Th1/Th2 phenotype, which helps prevent excessive inflammation while maintaining broad, resilient immune protection. [1, 2, 3, 4, 5]” . (Methods, 252-259+ cited new refs)

 

1. The dose selection was unexplained.

Response: The dose, 1/20^th of a human dose, was chosen as pretests with both Duochol and licensed Dukoral, Shanchol and Hillchol OCVs had shown that this dose was well tolerated in mice while giving rise to good but not maximal antibody responses. The dose is similar to that previously used in the development of the now India-licensed Hillchol OCV (which used WHO prequalified Shanchol OCV as comparator vaccine). Although not mentioned in the current report it is an encouraging fact that the serum and fecal antibody responses now shown for Duochol (and Dukoral) if anything are higher than those obtained with identical assay methods in our lab with Shanchol and Hillchol.

5. Figures 6 and 7 contain no statistical annotations (p-values, significance indicators or test identifiers). Without these, the reader cannot assess the claimed group differences or equivalences.

Response: We are now describing these comparisons in greater detail, mainly as text information (Results lines 556-575) since it would disturb the already “loaded” figures (now also with individual animal data points as requested in item 8 below) to indicate by p-values or number of significance stars the fact that all vaccine groups and antibody types were significantly elevated compared to the Nil control group, while between the differently stored Duochol groups there were no observed significant differences in any antibody type, and between the different Duochol groups and the Dukoral comparator group there was only one observed difference reaching p<0.05 (in vibriocidal antibodies which were higher in DuoChol stored for 21 months at 40° C than in the Dukoral group).

1. The stability data shown in Figure 5 are presented without error bars.

Response: The error bars if added would be meaningless, small and mostly invisible. We have instead provided the requested information about variation by in the results line 523-530 specifying the mean and maximal coefficient of variation for the replicates.

1. The SDS-PAGE gels in Figures 3 and 5 lack molecular weight markers, preventing the verification of band assignments. Appropriate molecular weight (MW) ladders should be included.

Response: This has now been done.

1. Individual animal data are not shown in any of the immunogenicity figure. Given n = 7 per group and the equivalence claim, individual data points should be displayed (e.g., as a swarm plot or overlaid on bar charts) so that readers can evaluate inter-animal variability.

Response: Now changed as suggested.

1. The vibriocidal antibody assay protocol was not described.

Response: It was/is with a cited ref 30 in Methods line 327; now also mentioned which might have been reviewer´s observation in Figure 6 legend text .

10. Claims of “affordable production cost” and “cost-effective” are made without any cost data, bill-of-materials analysis, or comparison with existing vaccines. These statements should either be removed or supported with quantitative manufacturing cost estimates.

Response: We agree that this is an overstatement (only excused by our extensive previous experience in taking lab-scale OCVs to industrial products – Dukoral, OrcVAX, Schanchol, Hillchol). We havenow  as suggested softened this “claim” (Discussion lines 675-677) and deleted it in the Abstract.

11. The title claims “Next-Generation Thermostable Oral Cholera Vaccine,” but no challenge experiment or functional neutralization assay was performed. While thermostability is supported by in vitro data, the term “Next-Generation” is an over-claim. A more conservative title, such as “Thermostable Formulation of an Inactivated Whole-Cell/rCTB Oral Cholera Vaccine with Preclinical Immunogenicity” would better reflect the evidence.

Response: Based on our previous work in developing what we have referred to as 1^st generation inactivated OCVs Dukoral, Orcvax/Shanchol/Euvichol, and 2^nd generation simplified OCVs Euvichol-S and Hillchol, we felt it motivated to call the thermostable Duochol OCV 3^rd or next-generation inactivated OCV. However, we take notice of the reviewer´s criticism and have now given a more conservative title in line with the reviewer´s suggestion.

Minor Comments

1. In Section 2.2, several numerical subscripts in chemical formulas or compound names are rendered as regular text rather than subscripts. Response: Now corrected.
2. The x-axis of Figure 5 (time) is non-linear, which may mislead readers into perceiving uniform-sampling intervals. Equal spacing or an explicit note of the nonlinear scale should be used. Response: We have modified the X-axis with a clear interruption between 6 and 21 months.
3. The sucrose concentration range of 50–100 mg/mL stated on page 5 is too broad for a defined formulation. The final optimized concentration should be specified in the manuscript. Response: While we have not seen any difference in stability between 50-100 mg/ml, we have indicated that in further work we have used 60 mg/ml.

 

We hope that with these responses and the revisions made the revised manuscript can now be accepted for publication.

Sincerely

Manuela Terrinoni & Jan Holmgren

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Major Comments
1.    The Abstract claims that immune responses were “fully comparable” to Dukoral, but this remains statistically unsupported.
2.    Reference 32, cited to support the Th1/Th2 immune bias justification for F1 hybrid mice (Section 2.8), was misattributed. Lee and Tsai 1999 (ref 32) is a methods paper on LPS quantification by the Purpald assay (PMID 9918668), not a study on Th1/Th2 immune polarization. This may have resulted from the approach described in the response letter, where the authors relied on a Google AI-generated summary of the references rather than directly verifying the primary literature. Additionally, references 32 and 33 are identical to duplicate entries.
3.    The Abstract claims “potential for improved efficacy,” but no challenge experiment, neutralization assay, or functional correlate of protection was performed.
4.    The authors provided a reasonable explanation for the 1/20 human dose selection in their response letter (pretests showed that it was well tolerated and gave good but not maximal antibody responses, similar to the Hillchol development approach). However, this rationale has not been incorporated into the manuscript. The dose justification should be briefly stated in Section 2.8 so that readers can evaluate the experimental design without consulting the review.
5.    The authors state in their response that statistical methods were added to the “Figure legends Fig 5, 6, 7,” but this information is not present in any of the three figure legends. Additionally, Figures 6 and 7 do not display statistical annotations (asterisks, brackets, or p-value indicators) on figure panels. While p-values are reported in the Results section, readers should be able to identify significant comparisons directly from the figures without cross-referencing the text.

Minor Comments
1.    The section numbering jumps from 3.3 to 3.5; section 3.4 is missing from the Results.
2.    The post-hoc test was changed from Bonferroni (original submission) to Holm-Šídák (revision) without explanation in the manuscript.

Author Response

RESPONSES/CORRECTIONS TO 2^ND ROUND COMMENTS (REVIEWER 2)

We thank the reviewer for the several useful observations and comments which we respond to and have corrected as described point-by-point below (and indicated with light-blue marking in the revised manuscript).

Major Comments

1. The Abstract claims that immune responses were “fully comparable” to Dukoral, but this remains statistically unsupported. RESPONSE: We have now changed “fully comparable” to “similar” in the Abstract (line 29).
2. Reference 32, cited to support the Th1/Th2 immune bias justification for F1 hybrid mice (Section 2.8), was misattributed. Lee and Tsai 1999 (ref 32) is a methods paper on LPS quantification by the Purpald assay (PMID 9918668), not a study on Th1/Th2 immune polarization. This may have resulted from the approach described in the response letter, where the authors relied on a Google AI-generated summary of the references rather than directly verifying the primary literature. Additionally, references 32 and 33 are identical to duplicate entries. RESPONSE: We apologize for the confusion = our mistake, and have now changed to the intended: REF 31 IS THE REFERENCE TO GOOGLE AI response, REF 32 IS PREVIOUS REF 31 SCHULTE ET AL and REF 33 REFERS TO THE PURPALD METHOD FOR LPS QUANTIFICATION.
3. The Abstract claims “potential for improved efficacy,” but no challenge experiment, neutralization assay, or functional correlate of protection was performed. Response: We explained in the response to reviewer 1 (cited below in italics) that there is no useful small animal protection model for evaluating adaptive protective immunity of cholera vaccines and that the used non-inferior immunogenicity approach with a licensed similar OCV is currently the best approach to predict efficacy in a small animal model. We now -- in direct response to the present comment by reviewer 2 -- specify in Abstract that the potential for increased efficacy is in comparison with inactivated whole-cell only OCVs [lacking rCTB] (Abstract lines 30-31) and we have also added a sentence to this effect in Discussion (lines 623-626). “There is no useful animal model to assess protective activity of cholera vaccines, since infant mouse or rabbit infection models take place in animals that are too young to develop an active immune response. Passive protection in these models - or other adult animal models - using immune sera from vaccinated mice also give little useful information (with regard to predicted protective efficacy as a vaccine in humans) since serum antibodies – in contrast to intestinal-mucosal antibodies - are neither responsible nor particularly predictive of in vivo protective efficacy. The best animal model prediction test of future protective efficacy in humans is in our opinion – and we may be regarded as world leading experts on these matters after having been much involved in the development of several of the currently used “previous generation” inactivated OCVs including Dukoral, OrcVax/Shanchol/Euvichol, and more recently India-licensed Hillchol – is what is used in the current study to compare the immunogenicity of a new “similar” OCV candidate with the closest existing already licensed and WHO-prequalified OCV (“similar” with regard to composition [inactivated whole-cells ± CTB], dosage, and planned schedule) which in our case is Dukoral with regard to both the relevant types of serum antibodies and, even more important, the intestinal mucosal IgA anti-LPS and anti-CTB antibody responses. The fully comparable immunogenicity in these regards we demonstrate for  Duochol visavi the comparator Dukoral OCV is the best predictive measure of future “clinical” efficacy (and such “non-inferior” immunogenicity is probably together with safety what will also be mainly assessed by regulatory authorities for potential future licensure of   this vaccine). We have now in the revised manuscript added a brief comment on these aspects. (Discussion lines 662-676).”
4. The authors provided a reasonable explanation for the 1/20 human dose selection in their response letter (pretests showed that it was well tolerated and gave good but not maximal antibody responses, similar to the Hillchol development approach). However, this rationale has not been incorporated into the manuscript. The dose justification should be briefly stated in Section 2.8 so that readers can evaluate the experimental design without consulting the review. response: agreed, this is now added as suggested (lines 264-267 )
5. The authors state in their response that statistical methods were added to the “Figure legends Fig 5, 6, 7,” but this information is not present in any of the three figure legends. Additionally, Figures 6 and 7 do not display statistical annotations (asterisks, brackets, or p-value indicators) on figure panels. While p-values are reported in the Results section, readers should be able to identify significant comparisons directly from the figures without cross-referencing the text. RESPONSE: NOW CORRECTED AS REQUESTED, both in Figure 5, 6 and 7 legends and in Figs 6 and 7 with significance signs both for differences of vaccine groups versus the unvaccinated Nil group and for the non-significance between the DuoChol groups internally and versus the Dukoral comparison group. This meant that we deleted the description of Figure 5 CV percentages from the text and instead refer to the figure 5 legend.

Minor Comments

1. The section numbering jumps from 3.3 to 3.5; section 3.4 is missing from the Results. NOW CORRECTED
2. The post-hoc test was changed from Bonferroni (original submission) to Holm-Šídák (revision) without explanation in the manuscript. Response: HOLM- ŠÍDÁK was used as we were advised that it provides small, later introduced “IMPROVEMENTS” on BONFERRONI and slightly higher statistical power, although in most cases p values are closely similar (as they were in our analyses).

 

 

Round 3

Reviewer 2 Report

Comments and Suggestions for Authors

Major Comments
1.    Reference 31, cited to support the rationale for using F1 hybrid mice with “balanced, intermediate Th1/Th2 immune responses,” is a literal Google search URL (https://www.google.com/search?q=F1+hybrid+mice+and+Th1+versus+Th2+immune+response+bias). This is the same root problem flagged in the previous round, reliance on a search-engine rather than the primary literature, now formalized as a citation, and a search-engine query is not a citable source because it has no author, no fixed content, and no verifiable evidence. It must be replaced with a genuine primary reference.
2.    The broader reference list has not been verified against primary sources. Reference 32 (Schulte, Sukhova and Libby 2008) is a study of Th1/Th2 bias in atherosclerosis, not a justification for selecting C57BL/6×BALB/c F1 hybrids for vaccine immunogenicity testing; reference 26 is authored “Lee, Q.Y.C.” but the cited paper is Quesenberry and Lee 1996 (PMID 8742081). Each should be corrected, and the entire reference list should be systematically re-checked.
3.    The Discussion now states that DuoChol offers “potential for improved efficacy… and compared to the Dukoral OCV,” but the Results show only non-significantly different (similar) immunogenicity rather than improvement, and buffer-free administration is a convenience advantage rather than an efficacy advantage. This sentence should be corrected so that efficacy is claimed only relative to whole-cell-only OCVs, with the Dukoral comparison described as comparable immunogenicity plus a practicality benefit.
4.    Figure panel labeling is broken in two figures. The Figure 3 legend contains no panel E and labels two panels as “(F)” while the text cites “Figure 3E,” and the same one-letter shift affects Figure 2, where the text cites 2D/2E but the legend labels 2C/2D with no panel E. The panels must be renumbered.
5.    The Figure 6 legend describes a “small intestine IgA” panel and uses the label “(f)” twice, although Section 2.9.8 describes only serum and fecal samples and the actual figure shows six cleanly labeled panels. The legend should be corrected to match the figure, and any small-intestine data either documented with a methods subsection or removed.
6.    The increased rCTB yield (≥1.5 g/L) that underpins the high-yield framing is asserted to “significantly increase production” but is reported as “(not shown),” leaving the central manufacturing claim unsupported in the manuscript. These data should be provided as a supplementary table, or the quantitative “significantly increases” claim should be removed.
Minor Comments
1.    The Discussion still describes DuoChol’s elevated-temperature stability as “unprecedented among currently available OCVs” without supporting comparison data. It should be replaced with a measured comparative statement or supported by a comparison table.
2.    The Figure 5 legend contains a broken cross-reference, stating that oligomer formation “did not affect GM1-ELISA activity, as shown in (C),” whereas GM1-ELISA activity is panel (D); the panel letter should be corrected.
3.    The statistical test is named only in Methods 2.9.9 and is not restated in the figure legends, and the rationale for the change from Bonferroni to Holm-Šídák appears only in the response letter; a one-line test statement should be added to each of the Figure 5–7 legends, with a brief justification in the Methods.
4.    Numerous typographical and numeric errors persist, including “Noo significant” (line 558), “p>050” for p > 0.50 (line 568), “4–400C” and “0,0200 mbar,” “indistinguishible,” “OD600nmreached,” and “side.by-side”; a careful line-by-line proofread is needed before resubmission.

Author Response

Response to Reviewer 2 point by point

Major Comments
1. & 2   1. Reference 31, cited to support the rationale for using F1 hybrid mice with “balanced, intermediate Th1/Th2 immune responses,” is a literal Google search URL (https://www.google.com/search?q=F1+hybrid+mice+and+Th1+versus+Th2+immune+response+bias). This is the same root problem flagged in the previous round, reliance on a search-engine rather than the primary literature, now formalized as a citation, and a search-engine query is not a citable source because it has no author, no fixed content, and no verifiable evidence. It must be replaced with a genuine primary reference.
,,,,,,,,,,,,2.    The broader reference list has not been verified against primary sources. Reference 32 (Schulte, Sukhova and Libby 2008) is a study of Th1/Th2 bias in atherosclerosis, not a justification for selecting C57BL/6×BALB/c F1 hybrids for vaccine immunogenicity testing; reference 26 is authored “Lee, Q.Y.C.” but the cited paper is Quesenberry and Lee 1996 (PMID 8742081). Each should be corrected, and the entire reference list should be systematically re-checked.

Response: We agree that ref 31 Google A1 was not an appropriate ref and we have now deleted it. The reason for using it was that the A1 citation gave a fair summary of what many immunologists know that using F1 hybrid mice may be a way to compensate for some of the immune response biases known to exist in the parent mouse strains. Big EU immunological consortia on vaccines and adjuvants in the 1990s (in which JH/Univ of Gothenburg was a partner) used C57Bl6/BalbC F1 hybrid mice for their studies for this reason, even though the degree of “balancing” varies with antigen, type of immune response and context which makes it difficult to give a single original ref to support the claim. Our ref 32 Schulte et al was therefore given to underline the Th1 and Th2 bias of the parental strains rather than to directly support the choice of F1. So, as suggested we have now deleted the previous ref 31, we have kept the Schulte ref as example of the marked Th1/Th2 biases of the parent mouse strains, and added 2 new refs (out of several more possible, see the appendix to this response) exemplifying the “balancing” in the F1 hybrids, and further we have mitigated the text (l. 255-257) to simply say that “F1 hybrid mice were used to reduce the Th1 and Th2 immune response bias respectively of the parent mouse strains”. The motive for the choice of mouse strain is anyway not of key relevance for this paper so we think this issue has now been discussed beyond its worth.

 

3.    The Discussion now states that DuoChol offers “potential for improved efficacy… and compared to the Dukoral OCV,” but the Results show only non-significantly different (similar) immunogenicity rather than improvement, and buffer-free administration is a convenience advantage rather than an efficacy advantage. This sentence should be corrected so that efficacy is claimed only relative to whole-cell-only OCVs, with the Dukoral comparison described as comparable immunogenicity plus a practicality benefit.

RESPONSE: Agreed, and corrected as suggested (l. 635-640)

4. 
      Figure panel labeling is broken in two figures. The Figure 3 legend contains no panel E and labels two panels as “(F)” while the text cites “Figure 3E,” and the same one-letter shift affects Figure 2, where the text cites 2D/2E but the legend labels 2C/2D with no panel E. The panels must be renumbered.

RESPONSE: Agreed, and corrected.

 

5. 
      The Figure 6 legend describes a “small intestine IgA” panel and uses the label “(f)” twice, although Section 2.9.8 describes only serum and fecal samples and the actual figure shows six cleanly labeled panels. The legend should be corrected to match the figure, and any small-intestine data either documented with a methods subsection or removed.

RESPONSE: Agreed, and corrected – only fecal antibodies were measured.

6. 
      The increased rCTB yield (≥1.5 g/L) that underpins the high-yield framing is asserted to “significantly increase production” but is reported as “(not shown),” leaving the central manufacturing claim unsupported in the manuscript. These data should be provided as a supplementary table, or the quantitative “significantly increases” claim should be removed.
   
   

RESPONSE: We have as suggested removed the “significance claim” and simply told that these modifications resulted in increased 1.5 g/L rCTB levels (compared to cited previous 1.0 g/L).

 

Minor Comments

1. The Discussion still describes DuoChol’s elevated-temperature stability as “unprecedented among currently available OCVs” without supporting comparison data. It should be replaced with a measured comparative statement or supported by a comparison table.

RESPONSE: We have removed “unprecedented” and instead simply say (l. 676-677) “Such stability at elevated temperatures has not been reported for any of the currently available OCVs … “

2. The Figure 5 legend contains a broken cross-reference, stating that oligomer formation “did not affect GM1-ELISA activity, as shown in (C),” whereas GM1-ELISA activity is panel (D); the panel letter should be corrected.

RESPONSE: Corrected plus we note the Results text for the SDS-PAGE oligomer data had fallen out and has now been added as basis for the figure 5F results (l. 519-538).

3. The statistical test is named only in Methods 2.9.9 and is not restated in the figure legends, and the rationale for the change from Bonferroni to Holm-Šídák appears only in the response letter; a one-line test statement should be added to each of the Figure 5–7 legends, with a brief justification in the Methods.

RESPONSE: This has now been corrected as suggested in the figure 5-7 legends. The statistical methods used are clearly spelled out in Methods and we do not think the choice of these methods should be motivated more than any other methods.

4. 
      Numerous typographical and numeric errors persist, including “Noo significant” (line 558), “p>050” for p > 0.50 (line 568), “4–400C” and “0,0200 mbar,” “indistinguishible,” “OD600nmreached,” and “side.by-side”; a careful line-by-line proofread is needed before resubmission

RESPONSE: Corrected with apologies.

 

Appendix: Example of references reporting “balancing” of immune response biases in F1 hybrid mice (with citations from Abstracts)

• Jones JT, Kusel JR. The inheritance of responses to schistosomiasis mansoni in two pairs of inbred strains of mice. Parasitology. 1985 Apr;90 ( Pt 2):289-300. “The F1 hybrid from the C57BL/6/0la X BALB/c cross resembled the low-responding parental strain (C57BL/6/0la) with respect to faecal egg excretion, accumulation of eggs in the tissues and splenomegaly, and was intermediate in its pattern of antibody response.”
• Huygen K, Ljungqvist L, ten Berg R, Van Vooren JP. Repertoires of antibodies to culture filtrate antigens in different mouse strains infected with Mycobacterium bovis BCG. Infect Immun. 1990 Jul;58(7):2192-7. “F1 mice had the restricted antibody pattern of C57BL/6 after one injection of BCG and had a hybrid BALB/c-C57BL/6 phenotype following a boost injection of BCG 2 months after the initial infection.”
• Fernandez C, Mäkelä O, Möller G. Genetics of the anti-dextran B512 and the autoanti-idiotypic response: codominant expression in F1 hybrids and dichotomy of response and allotype-linked idiotype. Immunogenetics. 1980;10(6):573-82. “The anti-dextran antibodies in these strains lack the idiotypes characteristic of either CBA and C57BL antibodies to dextran, but they possess their own particular idiotype. F1 hybrids between two responder strains possessing different idiotypes on their antibodies against dextran, produce both idiotypes and two different autoanti-idiotypic antibodies.” 
• Sarlo K, Parris JS, Clark ED, Horn PA, Robinson MK, McCay JA, Peachee VL, Veloso YL, White KL Jr. Influence of MHC background on the antibody response to detergent enzymes in the mouse intranasal test. Toxicol Sci. 2000 Dec;58(2):299-305.“The mouse intranasal test (MINT) was developed to assess the immunogenic potential of [allergic responses to] detergent enzymes. …… The current data strongly support the use of the F1 hybrid as an appropriate strain for evaluating allergic [immunogenic] responses to enzymes.”
• Hedrick SM, Watson J. Genetic control of the immune response to collagen. III. Coordinate restriction of cellular cooperation and antigen responsiveness by thymus-directed maturation. J Immunol. 1980 Oct;125(4):1782-8.“The H-2 restriction of T helper cells from thymus-reconstituted nude mice was examined. Hybrid athymic mice were bred from BALB/c.nu and C57BL/6.nu parental strains and reconstituted with fetal thymus tissue from either parental strain. T helper cells from these mice, immunized to SRBC, were restricted to cooperation with B cells of the thymic H-2 haplotype. These T helper cells were shown to have originated from the F1 host..”