A Vaccine Based on the A/ASIA/G-VII Lineage of Foot-and-Mouth Disease Virus Offers Low Levels of Protection against Circulating Viruses from the A/ASIA/Iran-05 lineage

Round 1

Reviewer 1 Report

This manuscript presents a trial of a FMDV vaccine tested in cattle to determine protection against challenge with a heterologous circulating strain of the virus. The authors demonstrate only modest protection, without protection from clinical disease in most animals, while viral shedding was reduced.  This is similar study to the authors' paper here (https://www.mdpi.com/2076-393X/9/10/1110) which seems to be cited incorrectly in reference 13.  However, the previously published work actually shows protection from clinical disease.  Therefore, enthusiasm for this manuscript is low, as it is presented as a rather failed attempt at a heterologous vaccine candidate.

Additionally, data presentation and statistical analyses lack clarity. As presented, the virus shedding data are difficult to make comparisons between groups to determine if the reduction in viral shedding is both statistically and biologically significant. 

Author Response

Response: We thank the Reviewers for spending their valuable time to review the manuscript and offer their comments.  We considered these comments very seriously and have made a genuine attempt to improve the manuscript.  We have made significant changes to the manuscript and our response is listed below. 

1. The details of the different virus isolates used in the study and their origin and genetic lineage is added to the Materials and Methods section 2.1.

‘The challenge virus, A/IRN/10/2018, was originally isolated from bovine tongue epithelium from a cow infected with FMDV in 2018 in Iran and identified as belonging to the FMDV A/IRN/05 lineage (SIS-13 sub-lineage).  The A/IRN22/2015 virus was also isolated from bovine tongue epithelium from a cow infected with FMDV in 2015 and was identified as a representative field isolate of the A/G-VII lineage.  Both the virus isolates were provided by the World Reference Laboratory for FMD (WRLFMD), The Pirbright Institute, United Kingdom.’

 

2. We have now provided a detailed explanation on the statistical methods used. Since these analysis were performed using R codes, it was written to explore several statistical options for model selection. Since the reviewers have considered it as confusing we have decided to remove the Supplementary Text file and references to different sections in the file i.e. Supplementary texts 1, 2, 3 and 4 in the Results and Discussion section of the manuscript.

This manuscript presents a trial of a FMDV vaccine tested in cattle to determine protection against challenge with a heterologous circulating strain of the virus. The authors demonstrate only modest protection, without protection from clinical disease in most animals, while viral shedding was reduced.  This is similar study to the authors' paper here (https://www.mdpi.com/2076-393X/9/10/1110) which seems to be cited incorrectly in reference 13.  However, the previously published work shows protection from clinical disease.  Therefore, enthusiasm for this manuscript is low, as it is presented as a rather failed attempt at a heterologous vaccine candidate.

Response: The reference cited is correct.  Whereas the A Malaysia 97 vaccine protected against A/G-VII lineage strains, the vaccine with A/G-VII did not protect serotype A/Iran-05 lineage strains.  Therefore, both A Iran 05 and A/G-VII vaccines must be a component of a multivalent vaccine for use in the Middle East.  The A Malaysia 97 vaccine strain is a component in vaccines used in South East Asia where there could be an incursion of A/G-VII lineage, whilst in the Middle East Asia, there is a threat of both the endemic circulation of A Iran-05 lineage and the incursions with A/G-VII lineages.

Secondly, as suggested in our introduction and aims, we aimed to test if A/G-VII vaccine would offer protection to A Iran-05 lineage thereby removing A Iran-05 component from the vaccines used in the Middle East Asia.  However, our results have shown that A Iran-05 cannot be replaced by this new vaccine strain.  Such a negative result is epidemiologically significant and must be available in the public domain and for countries considering the different combinations of vaccine strains in the region.  And we feel that our studies will help making the right choice of vaccine strains for the region.  In addition, for countries keeping vaccine banks, it is important information when choosing strains to include in their banks and when making decisions on the use of the strain when outbreaks occur.

Additionally, data presentation and statistical analyses lack clarity. As presented, the virus shedding data are difficult to make comparisons between groups to determine if the reduction in viral shedding is both statistically and biologically significant.

Response: Thank you for your comment. We have reanalysed the data to address the comments.  We did not observe a significant difference between groups in virus excretion, but we saw a group effect in PCR results, but we previously did not mention this in the paper. We have corrected this in the following sections in the revised manuscript.

• Materials and Methods: section 2.8 Quantitative analysis
• Results: Quantitative analysis of serological response to A/IRN/22/205 and A/IRN/18/2018
• Results: Virus isolation and detection of viral RNA in clinical samples

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript by Singanallur is well-written, easy to follow, and in general of good scientific quality (except for the major point raised here below). Its content is relevant for the scientific community, vaccine manufacturers, and those who purchase FMD vaccines.

 

Major point:

Figure 1, lines 95-96, and lines 324-329: The homologous titre in the full dose group is relatively low for a 43 PD₅₀ vaccine. With the remark in lines 95-96 that A/IRN/22/2015 was considered homologous to the vaccine strain, this reviewer wonders how homologous the strain used in the VNT assay really is. At least the name of the vaccine strain should be disclosed, and a % sequence identity (of P1) should be provided between the VNT and the vaccine virus. If the name cannot be disclosed, the sequence difference should be determined and mentioned.

It is the reviewer’s opinion that you should not perform a study with material of which the identity is unknown. That is scientifically not sound and ethically not acceptable for animal use. Also, if it is not exactly known what strain is in the vaccine, it is not known how heterologous the vaccination-challenge experiment really was. Right now, both the authors and the reader cannot judge how homologous and heterologous the strains were, and this makes interpretation of the study outcome difficult.

 

Minor points:

Line 45: reference ‘Paton et al., 2021’ is not numbered in the text, like the other references, and it is not listed in the reference list. Please correct.

Line 129: typo: catle

Lines 130-133: In principle, these criteria for disease generalisation usually only apply to the vaccinated animals. For control animals at least 3 feet should be affected. Correct? If so, please clarify in the text. This has also consequences for the results described in lines 240-141: what were the criteria for the controls?

Line 132: typo: catle

Line 136-137: Cattle scored positive on day 4 were allowed to stay until the end of the study, whereas usually they are removed from the experiment as these animals have reached the experimental endpoint. Maybe good to clarify.

Lines 151-152: On 2 locations: ‘the VNT assay’ instead of ‘the VNT’

Line 191: Please define ‘probang’.

Line 204: is ‘funcitonusing’ a typo? Should probably be ‘function using’.

Lines 224 and 234-235: Is the BVS described in these lines the same material as in lines 167-170? If so, please indicate in lines 167-170 that this material is referred as BVS and provided by BI. And please clarify whether the sera were generated by using monovalent vaccines?

Lines 271-272: Does this reviewer understand it correctly that additional VNT data were used to generate the plot as presented in Figure 2? Considering the significant lab-to-lab variation in VNT assay outcomes, it is worth to indicate whether the VNT data provided by BI were generated by WBVR or otherwise.

Lines 279-281: This reviewer does not understand why the inoculation route is the reason for nsp detection in all cattle post challenge. A good immunisation can limit FMDV replication in the animal such that a signal above the cut-off in a nsp ELISA cannot be observed. It is requested that the authors elaborate on what can be expected, possibly also based on literature references.

Line 288: For clarity, it is suggested to add ‘by PCR’ to the sentence: “Viral RNA could be detected by PCR in the serum of the…”

Line 318: typo: decreasesthe

Line 328: typo: sutdy

Line 343: typo: should read “…is still very valuable to estimate”

Table 2: It would benefit the reader to add a column indicating which animal was protected.

Author Response

Response: We thank the Reviewers for spending their valuable time to review the manuscript and offer their comments.  We considered these comments very seriously and have made a genuine attempt to improve the manuscript.  We have made significant changes to the manuscript and our response is listed below. 

1. The details of the different virus isolates used in the study and their origin and genetic lineage is added to the Materials and Methods section 2.1.

‘The challenge virus, A/IRN/10/2018, was originally isolated from bovine tongue epithelium from a cow infected with FMDV in 2018 in Iran and identified as belonging to the FMDV A/IRN/05 lineage (SIS-13 sub-lineage).  The A/IRN22/2015 virus was also isolated from bovine tongue epithelium from a cow infected with FMDV in 2015 and was identified as a representative field isolate of the A/G-VII lineage.  Both the virus isolates were provided by the World Reference Laboratory for FMD (WRLFMD), The Pirbright Institute, United Kingdom.’

 

2. We have now provided a detailed explanation on the statistical methods used. Since these analysis were performed using R codes, it was written to explore several statistical options for model selection. Since the reviewers have considered it as confusing we have decided to remove the Supplementary Text file and references to different sections in the file i.e. Supplementary texts 1, 2, 3 and 4 in the Results and Discussion section of the manuscript.

 

Comments and Suggestions for Authors

 

Reviewer 2:

The manuscript by Singanallur is well-written, easy to follow, and in general of good scientific quality (except for the major point raised here below). Its content is relevant for the scientific community, vaccine manufacturers, and those who purchase FMD vaccines.

 Major point:

Figure 1, lines 95-96, and lines 324-329: The homologous titre in the full dose group is relatively low for a 43 PD₅₀ vaccine. With the remark in lines 95-96 that A/IRN/22/2015 was considered homologous to the vaccine strain, this reviewer wonders how homologous the strain used in the VNT assay really is. At least the name of the vaccine strain should be disclosed, and a % sequence identity (of P1) should be provided between the VNT and the vaccine virus. If the name cannot be disclosed, the sequence difference should be determined and mentioned.

It is the reviewer’s opinion that you should not perform a study with material of which the identity is unknown. That is scientifically not sound and ethically not acceptable for animal use. Also, if it is not exactly known what strain is in the vaccine, it is not known how heterologous the vaccination-challenge experiment really was. Right now, both the authors and the reader cannot judge how homologous and heterologous the strains were, and this makes interpretation of the study outcome difficult.

Response: We fully agree with the reviewer and would have preferred to know more about the actual vaccine strain.  However, this information is commercial in confidence and cannot be published.  Despite this limitation, it is important for the FMD community to know how well new vaccine strains will perform in the field and that is why we have continued with this study.

One of the difficulties when testing commercial vaccine strains is the limitations on the detail the commercial company can provide.  We therefore had to use substitutes that to the best of our knowledge is related to the actual vaccine strain.  The sequenced A-GVII strains from 2015 (the BI strain is also from 2015, see Fraise et al. 2017; Ref 29) showed over 98% nucleotide similarities across the 1D region and therefore we don’t expect much antigenic difference between the vaccine strain and A/IRN/22/2015 used in our study.

Fraisse, F., Mouton, L., Dekker, A., Eblé, P.L, Coco-Martin, J., Hamers, C., Hudelet, P., Gaude, H., Goutebroze, S. New A/ASIA/GVII vaccine strain: a PD50 study with unexpected results in the controls. Global foot-and-mouth Disease research alliance symposium, 25-27 October 2017, Incheon, Korea.

In the revised version of the manuscript, we have addressed this comment in the conclusions.

Minor points:

Line 45: reference ‘Paton et al., 2021’ is not numbered in the text, like the other references, and it is not listed in the reference list. Please correct.

Response: The reference is numbered and added to the References section as the second reference.

 

Line 129: typo: catle

Response: Corrected the spelling as ‘cattle’

Lines 130-133: In principle, these criteria for disease generalisation usually only apply to the vaccinated animals. For control animals at least 3 feet should be affected. Correct? If so, please clarify in the text. This has also consequences for the results described in lines 240-141: what were the criteria for the controls?

Response: We have added the following to the text:

‘Control cattle are expected to have lesions on at least three feet to ensure the challenge was sufficient.’  And in the results section we added: ‘After challenge, all three unvaccinated controls developed clinical signs of FMD and showed generalised disease by 4 dpc with lesions on all four feet.’

Line 132: typo: catle

Response: Corrected the spelling as ‘cattle’

Line 136-137: Cattle scored positive on day 4 were allowed to stay until the end of the study, whereas usually they are removed from the experiment as these animals have reached the experimental endpoint. Maybe good to clarify.

Response: We wanted to keep the animals if ethically possible to ensure we could take samples from all animals until the end of the trial to allow statistically relevant comparisons between the groups.  In addition, since the animals were housed in tie stalls there was limited contact between the animals and therefore it was unlikely that they would have infected each other later during the trial to affect the comparisons.

Lines 151-152: On 2 locations: ‘the VNT assay’ instead of ‘the VNT’

Response: Since the expansion of the acronym VNT is Virus Neutralisation Test, there is no need to add the word ‘assay’.

Line 191: Please define ‘probang’.

Response:  The error is corrected since we did not collect probangs it has been deleted.  The sentence now reads ‘Total RNA from serum, nasal and oral swabs was isolated using the MagNA Pure 96 DNA and Viral NA Large Volume kit on the MagNA Pure 96 system (Roche® Life Science).                     

Line 204: is ‘funcitonusing’ a typo? Should probably be ‘function using’.

Response: The whole section is rewritten in the revised manuscript as follows:

‘The homologous and heterologous potency of the A/G-VII vaccine were calculated using the Spearman-Kärber method [23]  and logistic regression was done using R [24].  Confidence intervals for the binomial models were calculated using the delta method [25].  In the logistic regression, results from the present experiment were added to a dataset with results of 60 potency tests in 912 cattle previously generated by WBVR for BI and experiment is used as additional explanatory variable.  This way the slope of the dose response curve is estimated using all data and the position (the PD₅₀) on the data of the experiment.  Clinical protection based on count data were analysed using the two-sided Fischer exact test.  For statistical analysis VNT titres <0.6 were changed to 0.45. ANOVA was used to test the statistical differences between groups. If a statistical difference was found, a pairwise t-test (with Holm correction) was used to analyse differences between groups.  Group means and standard deviations were calculated and expressed as Mean ± SD. Titres against A/IRN/12/2015 were compared for the experiment with homologous and heterologous challenge using logistic regression using the slope of previous experiments as offset in the model. Longitudinal data on virus titres from serum, oral and nasal swabs were analysed using a linear mixed model, using the lme4 library in R [24], in which animal number as random variable and dpc and group as possible explanatory variables.  Using forward selection, the best model with the lowest AIC (Akaike’s Information Criterion) was chosen. In the linear mixed model samples from which no virus could be isolated were assigned a log­₁₀ titre of 0 and samples in which no RT-PCR curve was detected were assigned a Ct of 45. The duration of excretion was calculated as the time between first and last positive sample. Duration of virus and RNA detection was evaluated by normal linear regression.’

Lines 224 and 234-235: Is the BVS described in these lines the same material as in lines 167-170? If so, please indicate in lines 167-170 that this material is referred as BVS and provided by BI. And please clarify whether the sera were generated by using monovalent vaccines?

Response: The term vaccinated bovine sera (BVS) is added in the methods section.  Yes, they are generated using monovalent vaccines.  This information is already mentioned in the methods section following the names of the vaccine strains.

Lines 271-272: Does this reviewer understand it correctly that additional VNT data were used to generate the plot as presented in Figure 2? Considering the significant lab-to-lab variation in VNT assay outcomes, it is worth to indicate whether the VNT data provided by BI were generated by WBVR or otherwise.

Response: Yes, the VNT data was generated by WBVR for BI and BI kindly permitted us to use the data for the manuscript.  We have added this in the methods section that now reads:

‘In the logistic regression, results from the present experiment were added to a dataset with results of 60 potency tests in 912 cattle previously generated by WBVR for BI and experiment is used as additional explanatory variable.’

Lines 279-281: This reviewer does not understand why the inoculation route is the reason for nsp detection in all cattle post challenge. A good immunisation can limit FMDV replication in the animal such that a signal above the cut-off in a nsp ELISA cannot be observed. It is requested that the authors elaborate on what can be expected, possibly also based on literature references.

Response:  Taking on board the reviewer’s comments we have addressed this as part of the Major comment (please see above)

Since the potency tests are based on inoculating the virus on the dorsum of the tongue, primary virus replication occurs at the site of inoculation leading to generation of NSPs and induction of anti-NSP antibodies.  This local replication that often leads to minor lesions on the tongue is not considered vaccine failure, but only when secondary lesions appear on the limbs.  In other challenge models where the virus is either instilled in the nasal passage (intranasal instillation) or by direct contact challenge using donor animals, sterile immunity is more often observed in vaccine trials.

Line 288: For clarity, it is suggested to add ‘by PCR’ to the sentence: “Viral RNA could be detected by PCR in the serum of the…”

Response: The sentence is upended as follows ‘Viral RNA could be detected by real-time RT-PCR in the serum …’

Line 318: typo: decreasesthe

Response: Corrected as ‘decreases the’

Line 328: typo: study

Response: Corrected as ‘study’

Line 343: typo: should read “…is still very valuable to estimate”

Response: Corrected as ‘is still very valuable to estimate…’

Table 2: It would benefit the reader to add a column indicating which animal was protected.

Response: A column indicating protection is added to Table 2 and changes made to the title of the table.

 

 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors have addressed the questions and recommendations in a sufficient manner, and the manuscript can be published.