The present experimental trial in goats was designed with the aim to determine the effects of Map vaccination on current TB diagnostic tests and assess the effects of PTB vaccination on TB infection.
All experimental goats became infected irrespective of vaccination status. However, a reduction in pulmonary pathology was observed in some vaccinated individuals compared with the unvaccinated group. Another remarkable finding was that all vaccinated goats showed only TB lesions at the site of infection (i.e., lungs and associated LN) in contrast to the increased dissemination frequency in non-vaccinated animals. This finding is analogous to the results of subcutaneous vaccination against PTB with the Map-killed vaccine in cattle subsequently challenged with Map. The vaccine in those studies induced systemic immunity, preventing bacteremia and, consequently, dissemination of mycobacteria from the primary infection site. However, they did not prevent the establishment of the initial infection . Similar results were obtained in another experimental infection with M. caprae in which three of six (50%) infected goats showed gross TB lesions in mesenteric LN . This outcome is also consistent with findings observed in field cases of TB in goats [30–32].
In this sense, the present results indicate a certain degree of containment of dissemination of the infection from the primary complex, even when a significant reduction in the bacterial load in pulmonary drainage LN has not been observed. Nevertheless, this containment may not be sufficient to effectively prevent excretion of mycobacteria and horizontal transmission within a herd.
The role of the immunological status of vaccinated animals, especially the T-cell response, can be critical in terms of control or spread of the infection from the primary respiratory focus. Similarly, Hope et al. found few lesions at necropsy and smaller bacterial loads in LN in calves inoculated with M. avium and subsequently challenged with M. bovis. The authors suggested that T-cell responses resulting from M. avium infection enhanced a protective secondary response after challenge with M. bovis.
After conducting cross-sectional analysis, we found positive correlations of the IFN-γ response to E/C with pathology severity, and more weakly with bacterial load. A similar result was obtained with the single ESAT-6-specific IFN-γ response compared with pathology scores in goats  and calves . Furthermore, a positive correlation between the E/C-specific IFN-γ response and bacterial load was also previously described in goats  and cattle . Moreover, after analyzing post-mortem data, clear differences in the pathological and bacteriological results were found within the vaccinated group. The IFN-γ responses of vaccinated goats with low and high post-mortem scores (VR and VNR, respectively) were compared. The significantly lower E/C-specific IFN-γ responses obtained in VR animals demonstrated the capacity of this immunological biomarker to predict the severity of the disease and, by default, to predict the vaccine outcome.
In terms of assessment of TB diagnostic tests, we have adapted the diagnostic tests routinely used in cattle in bovine TB eradication programs. We have also introduced new DIVA reagents to perform the IFN-γ assay. Sensitivities obtained in all tests were higher than those obtained in previous studies in naturally infected goats [20, 23]. Notwithstanding this, the enhanced sensitivity described in these works when combining skin tests and IFN-γ tests was also confirmed in the present experimental trial.
The sensitivities of skin tests performed in vaccinated animals (100% and 90% for SIT and CIT, respectively), were higher than those obtained in another study carried out with goat herds naturally co-infected with TB and PTB (71% and 42.7%, respectively) . In this work, the authors reported many animals with false-negative CIT results that showed higher skin-fold thickness to PPD-A than to PPD-B. It is important to note that the sensitivity of CIT in the present study was achieved using strict interpretation of the test. With the standard interpretation (ΔBov-ΔAv ≥ 3 mm), the sensitivity would be reduced to 60%.
By contrast, in our trial, only a slightly higher sensitivity was found when applying the SIT in comparison with the CIT. However, the results showed serious compromise of the specificity of the SIT in vaccinated animals prior to challenge (4 of 4 false-positive reactors), which completely disappeared when using the CIT. These results are in accordance with those found in another study performed in a PTB-vaccinated dairy goat herd free of TB . Similarly, compromise of the specificity of the SIT in PTB-vaccinated deer has also been reported [35, 36].
Compared with the SIT, the standard IFN-γ assay seemed to be more robust in terms of specificity. False-positive results were concentrated at the interval between weeks 12 and 14, and then disappeared at week 16 when there was not yet a response to M. caprae infection. After 12 wpv, the IFN-γ responses to PPD-A decreased faster than did those to PPD-B (see Figure 1). As a consequence, 60% of the animals were false-positive “bovine reactors” at week 14. More long-term trials must be performed to study the kinetics of the IFN-γ response to Map vaccination.
Encouragingly, the sensitivity obtained for the standard IFN-γ test after M. caprae challenge was very high considering that elevated whole-blood IFN-γ responses to PPD-A were previously observed in Map-vaccinated cattle . In this sense, from weeks 18 to 28 (4–14 wpi), we only detected a masking due to PPD-A in two vaccinated animals. However, the high IFN-γ responses observed in our study (measured shortly after experimental infection) may decrease with the progression of natural TB infection under field conditions.
In the last decade, much effort has been focused on the development of novel antigens for bovine TB diagnosis that are more sensitive and specific than the avian and bovine tuberculins. This research has already resulted in the identification of several antigens, such as ESAT-6, CFP-10, and Rv3615c, that reduce cross-reactive immune responses to different mycobacterial infections or vaccinations in cattle [26, 27, 38] and goats [39, 40]. In the present study, we showed the capacity of both E/C and Rv3615c to distinguish PTB-vaccinated and TB-infected goats. Furthermore, no differences in sensitivity were observed between the experimental groups. Importantly, when combining the positive results of E/C and Rv3615c IFN-γ assays, the sensitivity was identical to that obtained by the tuberculin-based IFN-γ assay, which is currently being employed as an ancillary test to the skin test in some eradication campaigns. A similar pattern was previously described in cattle experimentally infected with M. bovis, in which the sensitivity increased from 77.9% to 91% when considering the results of Rv3615c and E/C IFN-γ assay together .
By contrast, serological responses to the Map vaccine were moderate; only one and four vaccinated goats were seropositive to PTB at weeks 14 and 28, respectively. These data are not in agreement with the 50% positivity previously obtained in PTB-infected goats . Interestingly, a boost effect on the Map ELISA due to CIT might be found at week 28 (2 weeks after tuberculin testing) when comparing the results with the test performed at week 14. The boost effect on the IgG response due to the skin test was previously described in M. bovis-infected cattle  and M. caprae-infected goats .
Vaccination against Map represents an important advance in controlling PTB and improving the economic balance of affected farms. Therefore, the pros and cons of its application must be exhaustively evaluated. An attractive speculation is that the partial protection to TB infection observed in some PTB-vaccinated animals could indirectly facilitate the control of TB, although long-term field studies are required to confirm this perspective. Moreover, we have demonstrated that to some degree, PTB vaccination interfered (after M. caprae infection) with the sensitivity of tuberculin-based TB diagnostic tests. However, beyond the first 2 weeks post-infection, only 5% false-negative results were obtained in the IFN-γ assay. In addition, these animals reacted positively in more than 80% of the remaining post-challenge tests. Thus, considering the collective basis on which TB tests are usually applied, it is unlikely that an infected herd could be undiagnosed upon extrapolation to a larger population. On the other hand, our results confirm that the interference with the specificity can be fully overcome by using defined DIVA reagents. Thus, developing and subsequently introducing these reagents into routine diagnostics could represent an improvement in both strategies: control of PTB by vaccination and control of TB by rapid detection of infected animals.