Tuberculosis (TB) in goats is a chronic infectious disease, mainly caused by Mycobacterium bovis and Mycobacterium caprae, members of the Mycobacterium tuberculosis complex (MTBC). This disease entails important economic costs for livestock industries (1) and could be a source of TB for cattle (2), other domestic animals (3, 4), wildlife (5), and humans (6).

Spain has the second-highest goat census of the EU, with 2.7 million goat heads (data extracted from FAOSTAT on 17/02/2020). Besides, the high TB burden in goats could explain a number of new bovine TB breakdowns, hampering the goal of TB eradication in cattle (7). Therefore, some regions with a high concentration of caprine herds carry out TB eradication campaigns in caprine flocks (8); however, goat herds are still not subjected to a national eradication program, except for those epidemiologically linked with cattle (9).

The cornerstone of an efficient caprine TB eradication program is the diagnosis. The Spanish bovine TB eradication program effectiveness is highly dependent on the routine tuberculin skin testing (10). Current bovine TB testing is based on the single intradermal tuberculin test (SIT) and single intradermal cervical comparative tuberculin tests (SICCTs), and the interferon-gamma release assay (IGRA). However, in goats under certain epidemiological contexts, those diagnostic tests have some drawbacks in terms of sensitivity (Se) and specificity (Sp) (8, 11).

Another concern for TB diagnostics is the vaccination against Mycobacterium avium subsp. paratuberculosis (MAP), which has been largely implemented in small ruminants, to prevent the development of clinical disease (12). Nevertheless, even though MAP vaccines are authorized (e.g., Gudair^® vaccine), it has been demonstrated that paratuberculosis (PTB) vaccination interferes with STs and IGRA used for TB diagnosis (13, 14). In addition, the efficacy of M. bovis bacille Calmette-Guérin (BCG) vaccine has also been assessed in goats during the last decade in different vaccination trials (15–19). Even though these trials showed that BCG conferred certain protection to experimentally and naturally infected goats, it was evidenced that vaccination interfered with current TB diagnostic tests (16, 20).

To overcome diagnostic interferences due to BCG vaccination, defined antigens to differentiate infected from vaccinated animals (DIVA) have been developed (14, 21); nevertheless, those antigens have shown lower Se compared to tests based on standard tuberculins (22). Recently, a new multi-protein complex called P22, obtained from purified protein derivative of M. bovis (PPD-B) by affinity chromatography, has been developed (23), yielding high Se and variable Sp, depending on the animal species and epidemiological contexts (24). To date, this antigen has been tested to detect humoral response against MTBC in different species (25–30); however, there is a lack of information regarding its performance for cell-mediated immunity (CMI)-based diagnostics.

The aim of this study was to evaluate the performance of different cell-mediated and humoral immunodiagnostic tests, based on the P22 antigenic complex, for the diagnosis of TB in goats under different epidemiological and control scenarios.

The results of Se of herd A and Sp of herds B and C are summarized in Table 3. The TB-positive status of all animals from herd A was confirmed by positive mycobacterial culture and/or positive lesions in histopathological analysis.

Efficient and accurate diagnosis is of paramount importance for the success of eradication programs based on test and slaughter strategy. Here, the performance of new P22 antigenic complex-based cell-mediated and humoral tests for the diagnosis of TB in goats was assessed under different epidemiological and TB control scenarios.

Recently, the P22 antigenic complex has been evaluated for the detection of IgG in ELISA tests in different species: cattle goat, sheep, pigs, and wild boar (24–27), red deer (28), badgers (29), and alpacas and llamas (30). In the present study, the performance of the P22 antigenic complex for diagnostic tests based on CMI, namely, STs and IGRA, was evaluated for the first time in goats. Indeed, the use of P22 for IGRA tests has only been reported in red deer experimentally infected with M. bovis (37).

The combined interpretation of tests leads to a substantial improvement of Se at the expense of a variable loss of Sp. As expected, in the conventional context, the SICCT alone or combined with the STAND-IGRA (8, 11, 38) showed the best performances by DOR analysis. The performances of tuberculin-based tests were followed by the combinations of SICCT with the P22-IGRAs, which increased Se at the cost of a certain loss in Sp. Moreover, the combination of cP22-ST + P22-ELISA clearly increased the Se with the benefit of a minimal decrease of Sp, showing similar results than the combination of SICCT + P22-ELISA. These findings are in concordance with previous studies of P22-ELISA and tuberculin-based skin testing. In cattle, the combination of SIT + P22-ELISA showed an improvement of Se of 30 and 6 p.p. compared to the SIT and the P22-ELISA alone, respectively (25). In another study conducted in goats (39), the same combination improved the Se of the SIT and the P22-ELISA in 19 and 9.5 p.p., respectively. Also, in the same study in goats, the combination of SICCT + P22-ELISA improved the Se of the SICCT in a 24 p.p. These results confirmed the benefits of the strategic use of serological and CMI-based diagnostic tests in parallel to maximize the Se in infected settings.

In the TB-VAC context, the combination of P22-ELISA with the two P22-based STs showed similar performances than the DIVA-IGRA. However, the latter showed considerably lower Se than the combinations of P22-ELISA with P22-based STs (reduction in 23–24 p.p.). Previous studies reported the excellent Sp (16) and the lack of Se (40) of DIVA-IGRA, although the DOR analysis tended to overestimate the Sp in this study. The Se of vaccine-associated diagnostic tests is an essential requirement for the development of an integral vaccination strategy (41), and the combination of cP22-ST + P22-ELISA showed an efficient and innovative diagnostic approach in the TB-VAC context, showing the highest combined Se and Sp values (94 and 95%, respectively).

Concerning the use of the ST in solitary, the P22-based STs showed lower Se compared to both the SIT and the SICCT tests, although previous studies in dairy goat flocks, with larger samples and different epidemiological situations, have shown lower Se for SIT (65%, 95% CI: 63.3–68.2) (8) and SICCT (44.5%, 95% CI: 35–55) (42). However, the Se of the cP22-ST (74%, 95% CI: 63–83) was similar to Se observed in two previous studies using DIVA STs (based on the peptide cocktails ESAT-6, CFP-10, and Rv3616c) developed for the diagnosis of TB in cattle: 76%, 95% CI: 59–93 (43) and 75%, 95% CI: 47.7–97.7 (44). In the latter, the addition of the Rv3020c peptide improved the Se to reach 87.5% (95% CI: 61.7–98.5), being similar to the Se of sP22-ST (87%, 95% CI: 74–94) obtained in the present study. On the other hand, in BCG-vaccinated animals, the Sp of SIT and SICCT decreased dramatically (27 and 20 p.p. of reduction, respectively), whereas the Sp of sP22-ST and cP22-ST remained high (97 and 98%, respectively). These findings again highlight the suitability of P22-based STs as TB vaccine-associated diagnostic candidates, although improvements to increase the Se should be necessary.

Moreover, herd PTB status and MAP vaccination may also affect the interpretation of the results. MAP infection was not reported in farms B and C, and no recent clinical history of PTB was observed by the veterinarians. Despite this, vaccination against MAP is a common practice in small ruminants in Spain (12), and diagnostic interferences due to MAP vaccination on TB diagnosis cannot be ruled out in these two MAP-vaccinated herds. In this sense, strong reactions to PPD-A were observed at skin testing (Supplementary Data), but the results of comparative tests (cP22-ST and SICCT) showed higher Sp compared to their respective single STs (i.e., sP22-ST and SIT). These findings indicate that some degree of cross-reactivity due to MAP vaccination was still maintained. Similarly, interferences of MAP vaccination on TB diagnosis, mainly in CMI-based diagnostic tests, were previously observed in MAP-vaccinated goats (14, 45).

Surprisingly, the P22-based IGRAs, particularly the sP22-IGRA, showed higher Se compared to STAND-IGRA and even higher compared to DIVA-IGRA. However, the Se of sP22-IGRA was similar to that previously observed by the STAND-IGRA (92%, 95% CI: 84–96) in other studies conducted in goats (26). The results of Se of the cP22-IGRA in the present study were also similar to those previously observed in experimentally M. bovis-infected red deer (37). However, a slight loss of Sp in the P22-IGRAs was detected compared to STAND-IGRA. Even so, the Sp was within ranges (95–100%) described for STAND-IGRA in previous studies (11, 38, 45). This mild reduction in Sp could be explained by the high concentration of P22 used for stimulation of whole blood (20 μg/ml) and by the fact that P22 complex contains 21 proteins also present in M. avium (23), which can cause cross-reactivity with MAP vaccination and/or infection. Indeed, the interference of MAP vaccination on STAND-IGRA has been previously observed in adult MAP-vaccinated goats (13, 14, 45). The Sp of P22-IGRAs considerably decreased in BCG-vaccinated herds compared to that previously described for the STAND-IGRA (16). Overall, the results of sP22-IGRA suggest that this test could be a potentially valuable tool for TB eradication in endemic areas, although further studies to determine the optimal concentration of P22 are required to improve its Sp with a minimal loss of Se.

Serological diagnostics is a cost-effective alternative for TB diagnostics. However, the Se of antibody-based diagnostic tests was generally lower compared to tests based on CMI (46, 47). In the present study, the Se of P22-ELISA was slightly lower than that in previous studies in goats and cattle (25, 39). This loss of Sp might be explained by the fact that animals from herd A were not vaccinated against MAP nor subjected to frequent intradermal testing, factors that could enhance humoral responses against MTBC antigens (48). Interestingly, the Se was significantly enhanced when using P22 and MPB83 ELISAs in parallel. Thus, even though MPB83 is a major component of the P22 complex, specific IgGs of some infected animals were only detectable using the MPB83 purified recombinant protein alone, while others were only detected using the P22 complex, which contains additional serodominant epitopes (23).

Finally, Sp of the P22-ELISA reached considerably higher Sp in MAP-vaccinated (and BCG-unvaccinated, i.e., herd B) goats (93%) compared to that previously found in Spanish (78%) and Norwegian MAP-vaccinated goats (58%) (24). In the latter study, besides MAP vaccination, MAP coinfection and/or contact with environmental mycobacteria was not discarded as a source of diagnostic interference. Interestingly, in the present study, the Sp was also high in BCG- and MAP-vaccinated goats (96%), suggesting that BCG vaccination does not induce antibody responses that cause interference on the diagnosis by the P22-ELISA. The absence of antibody responses was consistent with the fact that the BCG Danish strain used for vaccination expresses low levels of MPB83 and MPB70 (49), which are the most abundant proteins of the P22 antigenic complex (23). Moreover, tuberculin skin testing after 42 days of MAP or BCG vaccination caused a boosting effect on humoral responses against tuberculin antigens, resulting in false-positive cattle for an MPB83-based ELISA (50). Here, minimal or no boosting effects of MAP/BCG vaccination due to skin testing were observed on the P22-ELISA. Indeed, goats from herd B were sampled around 2 years after vaccination against MAP, and ST was performed once or twice after MAP vaccination. Also, 34/68 goats from herd C were vaccinated with BCG and Gudair^® at 9–10 months before the sampling, whereas the rest of the animals were vaccinated more than 1 year before, and no ST was performed since. Based on the results herein, the P22-ELISA seemed to be a useful ancillary diagnostic tool, either in BCG or MAP vaccination context, although it should be confirmed in further studies with larger sized herds.

In conclusion, this study reinforces the applicability of the P22 antigen complex as a complementary instrument for TB diagnostics in goats under different control scenarios. The P22 serological diagnostic is a cost-effective alternative, and combined interpretation with STs, either with PPD-B or P22, showed promising results. Moreover, the use of P22 antigenic complex in CMI-based diagnostic tests showed encouraging results, being suitable for further research on the improvement of TB diagnostics.

All datasets generated for this study are included in the Supplementary Material.

All animals included in this study belonged to commercial farms and were not experimental animals. All sampling and handling procedures were carried out by authorized veterinarians according to standard farm methods and in conformity with Spanish legislation (Royal Decree 2611/1996 and amendments) and European Union laws for the protection of animals used for scientific purposes (2010/63/EU). Test and slaughter of positive animals, as well as post-mortem sampling to confirm the disease, were conducted according to the regulations defined by the Catalan Government (Resolution AAM/1314/2014). Written informed consent was obtained from the owners for the participation of their animals in this study.

BP and JB contributed to conceptualization. CA-V, BP, and MG contributed to data curation. CA-V and BP performed the formal analysis. BP and LJ acquired funding. CA-V, BP, MG, EV, MD, IM, JI-L, and JB contributed to the investigation. CA-V, BP, JI-L, EV, and MD contributed to the methodology. BP and LJ contributed to project administration. JI-L and MS acquired resources. CA-V and BP wrote the original draft. JB, LJ, JI-L, EV, and MD contributed to writing, reviewing, and editing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.