A defined syphilis vaccine candidate inhibits dissemination of Treponema pallidum subspecies pallidum

Syphilis is a prominent disease in low- and middle-income countries, and a re-emerging public health threat in high-income countries. Syphilis elimination will require development of an effective vaccine that has thus far remained elusive. Here we assess the vaccine potential of Tp0751, a vascular adhesin from the causative agent of syphilis, Treponema pallidum subsp. pallidum. Tp0751-immunized animals exhibit a significantly reduced bacterial organ burden upon T. pallidum challenge compared with unimmunized animals. Introduction of lymph nodes from Tp0751-immunized, T. pallidum-challenged animals to naive animals fails to induce infection, confirming sterile protection. These findings provide evidence that Tp0751 is a promising syphilis vaccine candidate.

In this manuscript the authors show that rabbits immunised with the antigen, Tp0751, resulted in attenuated lesion development (moderately convincing), inhibition of T.pallidum dissemination (convincing) and increased cellular infiltration at lesion sites (convincing at the tissue level). There is still an urgent need for a better vaccine for syphilis and this manuscript does make a significant contribution in this area.
However, while some of the findings were convincing at the tissue level, others were less so, as evidenced by comments such as " …difficult to reach a definitive conclusion concerning the true lesion burden …" or "It is likely that both of these proposed mechanisms contributed to the proposed mechanisms …". One aspect that would significantly strengthen the manuscript would be if the authors could link their whole animal and tissue level observations with some underlying immunological mechanisms. While they did observe the broad types of cells at the primary lesion sites, it would have been more useful if they had analysed specific Tp0751 antigen responses in defined T cell types.
The authors also suggest that their positive observation may have been " …suggestive of local production of Tp0751-specific antibodies". Why didn't they measure these antibody levels and perhaps even test for their functional role; neutralisation or binding. By including these additional levels of analyses the study results would be more likely to have interest and impact in the field.
Reviewer #3 (Remarks to the Author): Summary of the key results The authors have undertaken an experimental model to investigate immunological, clinical, and microbiological correlates of protection against syphilis conferred by a newly developed vaccine. They have found that immunized rabbits have a reduced bacterial load measured by PCR in the organs as compared to unimmunized rabbits, but the vaccine apparently did not provide sterile protection which would be ideal. In the second part of the study, however, the rabbit infectivity testing negative results point towards the desired sterile protection.
Originality and interest: The claims of the paper are novel and the vaccine may have an important role in mitigation of syphilis complications, but would not work to prevent the disease and ultimately stop transmission. Although the data seems not to be conclusive, the results are most interesting and timely for the development of a fully protective syphilis vaccine.
Data & methodology: -Would the authors consider having used a group of only three animals as a limitation given the discrepant results between animals? -Despite the authors' argument that PCR may have detected dead organisms may be true, the discrepant result for PCR vs microscopic assessment of primary lesions (Rabbit 1) needs to be interpreted with caution.
-Could the lower number of Treponema in the inoculation site of Rabbit 3 be related to a decreased proliferation? The current explanation (i.e. greater number of cells migration) seems contradictory to the greater immune cell infiltration in the skin of Rabbit 3. -qPCR results in Extended Data Table 1 do not seem to correspond to Figure 3. There might be a typo for the Liver and Spleen Ct1 and Ct2 results: where it says "10 squared" should read "103". Also Liver Im2 seems to have the same mistake. Please confirm since these are the most critical findings of the study.
-I would recommend that Extended Table 1 is fitted into the manuscript to facilitate reading comprehension.
-Could the popliteal lymph node for Rabbit 3 have been tested using PCR to confirm the hypothesis of greater load of dead treponemes draining to the lymph node?
Appropriate use of statistics: -Authors may want to revise the statistical analysis of lesions progression (i.e. ulceration and diameter). Currently individual lesions on each rabbit are analyzed as independent items. However, the lesion sites in one rabbit may have correlated results in terms of lesion progression compared to other rabbits. The reason is that lesions in a single rabbit are exposed to the same protective immunity; therefore these are not fully independent units and may need specific statistical methods for comparison (e.g. multilevel model).

Conclusions:
The overall results are very promising but with some contradictions that seem difficult to overcome with a small sample (n=3) study like this, the authors may want to discuss what are the next steps.
It would be good that the researchers explain the path towards development of a syphilis vaccine that can be trialed in humans and possibly to acknowledge certain gaps, including that the current vaccine would not be effective to prevent chancre development and that the current study does not ensure the vaccine is broadly protective against different T pallidum strains.
Clarity and context: -The researchers should more clearly explain why it is important that Tp0751 belongs to the same protein family as the meningococcus b vaccine. If one reads the abstract, as it is currently explained, some people could think that the meningococcus vaccine is protective for syphilis.
-Discrepant results of experiments could be more clearly written -Readers outside the discipline would benefit of a schematic of the main result to accompany publication.
1. "…parts of the Material and Methods need additional information e.g. it is not even introduced which T. pallidum strain has been used to challenge the rabbits." We thank the Reviewer for bringing this omission to our attention. We have now fixed this error, please see Lines 414 and 424-425. 2. Animal numbers and statistics are the major weakness of the study. The authors explanation provided in the Reporting Checklist for Life Sciences Articles "The number of animals included in the protection experiments was reflective of a number sufficient to investigate the protective effect of the vaccine candidate while adhering to budgetary constraints." is not satisfying. (A) with only 2(ctrl)+3 animals results will not be normal distributed. However, throughout the MS parametric tests are applied. This should be seen critical and reanalysis should be made using non-parametric tests. Currently, the risk is high that significance levels provide a false assumption because of a lack of power. (B) I recommend to compute a post hoc power analysis of data to identify the weakness. We agree with the Reviewer and recognize the limitations of our animal numbers. (A) Statistical analyses have been re-done using non-parametric tests for situations where n>2 (nonparametric tests cannot be performed with n=2). For Fig. 2a (Fig. 3e) to show this normalized data and this has been explained in the Data Analysis section of the materials and methods section (Lines 539-554). This same approach was used for analysis of cellular infiltrates in primary lesions (Figure 4), and an additional panel was added to display this normalized data (Figure 4e). Modifications have been made throughout the results and discussion sections to reflect these new analyses. (B) It is our understanding the power post-hoc analysis is useful to determine the required n value for a future experiment based upon a pilot study, but not to evaluate the statistical significance of a collected dataset. Furthermore, p-values and power are directly related to one another. Since we have re-analyzed our data using non-parametric comparisons, we do not believe post-hoc power analysis is necessary for this study. However, we will be using such calculations to design our future experiments, and this has been mentioned in the Discussion section (please see Lines 312-314). 3. The results are promising and provide a pretty good idea about the usefulness of tp0751 as a component in vaccination against syphilis. However, the presented data are currently based on low animal numbers and must be interpreted carefully. We agree with the Reviewer and have included a statement in the Discussion explaining this, see Lines 312-314.
4. Ideally, animal numbers need to be upgraded. I would say group size must be at least 4-5/group. It looks like some more animals were used for vaccination against tp0751 (e.g., animal Im26 pops-up all over sudden in Supplementary Fig. 2). I wonder if some analysis can be supported by additional data sets to increase animal numbers. However, in case upgrading animal number is not possible, the statistical analysis needs a thorough revision. We agree that the ideal situation would be to perform additional experiments with more animals. However, considering that the experiment took approximately 40 months to complete, this addition is not possible within the timeline of this manuscript submission or the confines of the budget that supports this research. The proposed plan is to perform a scaled-up experiment within the next one to two years using larger animal group sizes and a longer duration to test the extent of immunity generated. Additional points:  L. 31: Normally tp0751 should be written in italics (pl. check w/ Journal stile). The standard convention is for genes to be italicized and proteins to remain in standard font. This is also the standard for articles published in Nature Communications, and we have followed that convention in this manuscript.  L. 46 and 47: Suggest to add (%). This has been added, please see Lines 46 and 47.  L. 61: Is 'desensitization' a real issue in the clinic? It is often recommended by CDC and others, but how often does it really occur. I think it is not necessary to make this point here. There are enough plausible reasons why a vaccine is needed.  Table 1) has been altered to include the time points for the collection of each dataset and, where possible, values are now presented as mean +/-SEM. We hope that the dataset now includes enough detail.  L. 153: Was analysis also supported by qPCR? qPCR was also performed on primary lesions, however, a different result was observed. This is explored in the discussion Lines 242-252.  Fig. 3: I suggest to change the ordinate in B. to D. into scientific scaling similar to graph A. The ordinate has been changed to scientific scaling for Fig. 3b-d.  I may have missed it, but Im1 has a huge variation. It would be good to explain this in the Discussion as well as the reason why only Im2 was tested positive in bone tissue. Please refer to Table 1, which shows the mean and SE for each individual rabbit. While Im1 displays higher variation than the other immunized rabbits, the variation is similar to that of the control rabbits. The reason underlying the positive bone result for Im2 is unknown, but this finding relates back to the difficulties of working with a small sample size and an outbred animal model. This has been discussed in Lines 312-314.