Xinqidi,Your biological research partner
Xinqidi Biotech Co.,Ltd,Wuhan,China 2008-2021
R&D 13th year

In-vivo expressed Mycobacterium tuberculosis antigens recognised in three mouse strains after infect

Issuing time:2021-06-09 15:32


Novel tuberculosis (TB)-vaccines preferably should (i) boost host immune responses induced by previous BCG vaccination and (ii) be directed against Mycobacterium tuberculosis (Mtb) proteins expressed throughout the Mtb infection-cycle. Human Mtb antigen-discovery screens identified antigens encoded by Mtb-genes highly expressed during in vivo murine infection (IVE-TB antigens). To translate these findings towards animal models, we determined which IVE-TB-antigens are recognised by T-cells following Mtb challenge or BCG vaccination in three different mouse strains. Eleven Mtb-antigens were recognised across TB-resistant and susceptible mice. Confirming previous human data, several Mtb-antigens induced cytokines other than IFN-γ. Pulmonary cells from susceptible C3HeB/FeJ mice produced less TNF-α, agreeing with the TB-susceptibility phenotype. In addition, responses to several antigens were induced by BCG in C3HeB/FeJ mice, offering potential for boosting. Thus, recognition of promising Mtb-antigens identified in humans validates across multiple mouse TB-infection models with widely differing TB-susceptibilities. This offers translational tools to evaluate IVE-TB-antigens as diagnostic and vaccine antigens.


Tuberculosis (TB) remains the most widespread and deadly infectious disease from a single bacterial pathogen worldwide1. The transmission of its causative agent, Mycobacterium tuberculosis (Mtb), cannot be sufficiently controlled -in adolescents and adults- by childhood vaccination with Bacillus Calmette-Guérin (BCG), the only currently licensed TB vaccine2. Therefore, numerous efforts in the last decades have attempted to develop better TB vaccines3. Two recent and independent clinical efficacy trials showed that BCG re-vaccination led to a decrease in sustained Mtb infection, and that multiple injections of an adjuvanted TB subunit vaccine, M72/AS01E, were able to prevent development of TB in latently Mtb infected individuals4,5. Although the efficacy of M72/AS01E still needs to be corroborated in larger cohorts and across different geographical areas6, this finding underscores the clear potential of using selected Mtb antigens as BCG booster vaccines2. However, immune correlates of TB protection remain unknown, posing a challenge to the identification of such protective Mtbantigens for vaccine design7.

In recent years, we have identified a class of Mtb antigens, named IVE-TB antigens, encoded by Mtb genes that are highly and consistently expressed in the lung of TB susceptible (C3HeB/FeJ) as well as TB resistant (C57BL/6J) mice following aerosol Mtb (Erdman) challenge8,9. Because these Mtb genes are highly expressed in vivo, the encoded IVE-TB proteins represent an interesting set of candidate antigens for several reasons. First of all, they are preserved among at least 219 Mtb clinical isolates and other pathogenic mycobacteria, thus offering the potential to elicit immune responses against multiple human virulent Mtb strains. Secondly, they share high homology with BCG and are thus likely implementable in booster vaccine strategies. Additionally, they contain multiple predicted HLA-Ia and HLA-II binding motifs covering 85% of the human population. Lastly but most importantly, many IVE-TB antigens were found to be recognised by immune cells of latently Mtb infected (LTBI) subjects and TB patients9,10. Interestingly, these T cells were capable of producing multiple cytokines, in addition to or even in the absence of IFN-γ, a cytokine known to be necessary but not sufficient in conferring protection against TB9,10.

Due to the absence of human Mtb challenge models and validated correlates of protection, promising candidate vaccines are commonly tested in mice before being evaluated in larger animal models and eventually clinical trials11. Many Mtb antigens, such as those included in the most advanced candidates in the TB vaccine pipeline12,13,14,15,16, have been screened previously in C57BL/6 and BALB/c17. Both strains (particularly C57BL/6) are relatively resistant to TB and do not display the full spectrum of human TB pathology, such as the development of centrally necrotic lesions18. By contrast, the less frequently used C3HeB/FeJ strain (also known as the Kramnik model) presents with necrotic lesions and lung cavitation resembling those found in human TB19.

In order to translate our previous findings from human cohorts towards highly controlled experimental in vivo TB models in which protective efficacy of candidate vaccines can be explored, we aimed to: (I) evaluate promising IVE-TB antigens across different strains of Mtbinfected mice, including not only TB resistant (C57BL/6, BALB/c), but also TB susceptible (C3HeB/FeJ) strains, reasoning that the recognition of antigens in mice with different genetic backgrounds, and widely different TB susceptibilities, would better reflect the TB heterogeneity found in the human population; (II) study whether IVE-TB antigens induce IFN-γ as well as non-IFN-γ cytokine responses (TNF-α and IL-17) also in mice and in different organs (i.e., spleen, lung and mediastinal draining lymph nodes), to enable identification of those antigens that, though recognised in all mouse strains, would induce different cytokines profiles in resistant and susceptible mice. Upon asserting that an antigen is recognised in the susceptible strain, the cytokine and cellular responses can then be redirected towards a protective response (e.g., the response found in the resistant strains) by vaccination using appropriate adjuvants. (III) To determine which IVE-TB antigens recognised during Mtb infection are also recognised following vaccination with BCG, with the rationale to discriminate between highly specific Mtb antigens that could be useful candidates for TB diagnostics in areas with standard neonatal BCG vaccination programs, versus Mtb/BCG shared antigens that can be used as BCG booster vaccines; (IV) and validate findings in a multi-laboratory approach across two different laboratories and experimental settings to strengthen confidence in the results, especially in light of the large heterogeneity distinctive of human TB. The latter conforms to recent recommendations, which encourage researchers to perform pre-clinical discovery screenings in different mouse strains, routes of infection, Mtb strains and laboratories, before moving to more costly and resource demanding animal and human studies17.

We find 19 Mtb antigens that are significantly recognised by cells isolated from different murine tissues, presenting a wide array of immune responses with respect to the cytokines produced and the three mouse genetic backgrounds investigated. Of note, the recognition of 11 of these antigens was corroborated in an independent setting. Thus, our multi-laboratory approach identified promising Mtb antigenic targets for potential application as tools in TB control.


IVE-TB antigen recognition in multiple organs of three genetically different mouse strains by cells producing different cytokines

Promising TB candidate vaccine antigens first need to be validated in small animal models before they can be selected for evaluation in larger animal models such as non-human primates, and eventually for testing in human clinical trials11. Here, we aimed to verify which IVE-TB antigens, previously shown to be recognised by T cells from TB patients and LTBI subjects9,10, were consistently and broadly recognised during the course of murine Mtbinfection in different genetic backgrounds, given the fact that human populations are genetically highly diverse. To include different genetic backgrounds with widely diverse TB-susceptibility phenotypes, we selected C57BL/6, BALB/c and C3HeB/FeJ mice and challenged these with Mtb20,21. Different cell pools from each organ (spleens, lungs and mediastinal lymph nodes (medLN)), strain and time-points (early (5 weeks) or later (9–12 weeks) stage of Mtbinfection) were tested in parallel by stimulation with single or fusion IVE-TB antigens (n = 22) (Table 1) as well as with positive controls (ConA and PWM) (Fig. 1a, Supplementary Data 1). Besides IVE-TB antigens, also several other, stage specific- and secreted-Mtb antigens (n = 5) were included in the screening since several studies previously described these as promising (diagnostic) TB biomarkers or as putative antigens for TB vaccination22,23,24,25 (Table 1). The concentrations of IFN-γ, TNF-α and IL-17 were measured in the cell supernatants after 3 days of stimulation. Cytokine levels in response to each Mtb antigen were compared between naïve mice and Mtb challenged mice for each organ and mouse strain, after subtracting background concentrations from unstimulated negative control samples (Supplementary Data 1).

Table 1 List of Mtb antigens included in this study.
Fig. 1: Most IVE-TB antigens are recognised by cells from Mtb infected mice with differences among mouse strains, organs and cytokines.

a C57BL/6 (BL6), BALB/c (BALB) and C3HeB/FeJ (C3H) mice were challenged with 1 × 104 CFU Mtb-lux strain intratracheally and killed at an early (5 weeks) or later (9–12 weeks) stage of Mtb infection. Uninfected naïve control mice were killed and tested in parallel. Pooled cells from lungs, mediastinal lymph nodes (medLN) or spleens were stimulated with single or fusion Mtb antigens or with positive control stimuli ConA and PWM. The concentrations of IFN-γ, TNF-α and IL-17 were measured in the cell supernatants after 3 days of stimulation and corrected for background. b Heatmaps of the specific Mtb antigen responses for each cytokine are shown. Colour gradients indicate the cytokine concentration, while asterisks indicate significant differences between the naïve and the Mtbchallenged groups. White *p-value < 0.05, computed with Kruskal–Wallis test; black asterisk marks differences that remained significant after multiple test correction (q-value < 0.1, i.e., FDR-adjusted p-values of Mann–Whitney.test). c 3D plot of the Variable Importance in Projection (VIP) scores obtained from each PLS-DA model computed on the basis of the following known classes: (i) time points (i.e., naïve mice vs. mice killed at early or late point after Mtb infection); (ii) mouse strains (i.e., C57BL/6 (BL6), BALB/c (BALB) and C3HeB/FeJ (C3H) mice); (iii) organs (i.e., lungs, mediastinal lymph nodes or spleens); (iv) cytokines (i.e., IFN-γ, TNF-α and IL-17). Antigens with a VIP score >1 were considered to have an above average influence on the model and are depicted by solid dots. Antigens with a VIP score >1 for the cytokine PLS-DA model are depicted by solid red dots. Antigens with a VIP score <1 are depicted by empty dots and their names are omitted in the 3D plot.

Several trends emerged when grouping the cytokine responses against the 27 Mtb antigens included in this study (Supplementary Fig. 1). High TNF-α responses were found to the majority of the selected Mtb antigens, most predominantly in the spleen, in all three mouse strains. Secondly, the cells analysed from the C3HeB/FeJ mice showed an overall reduced ability to produce TNF-α against Mtb antigens, especially cells from the medLN and lungs whose TNF-α response was low or absent even after mitogen stimulation (Fig. 1b and Supplementary Fig. 1). Antigen-specific IFN-γ and IL-17 responses were more diversely distributed and varied strongly between organs and mouse strains (Fig. 1b). Interestingly, especially in BALB/c and C57BL/6 mice, significant IL-17 responses were mainly induced in cells isolated from the lungs and medLN, while IL-17 responses were almost absent among the splenocytes (Supplementary Fig. 1). For splenocytes and medLN, the highest peak of cytokine production was mostly observed at an early stage of Mtb infection, while for lung cells this was observed at the later stage (Supplementary Fig. 1). IFN-γ responses tended to be highest in C3HeB/FeJ mice, the most TB susceptible strain, confirming that like in humans IFN-γ in itself is not sufficient for protective immunity.

Among the Mtb antigens that induced a significant specific increase in cytokine production, the highest number was recognised by medLN cells from C57BL/6 mice (18/27 antigens recognised), followed by C3HeB/FeJ (7/27 antigens recognised) and BALB/c (4/27 antigens recognised) mice (Fig. 1b).

From these antigens, only one was recognised by all mouse strains (Rv0826), five by C57BL/6 and C3HeB/FeJ (Rv1886, Rv0440, Rv1980, Rv1791 and Rv1733), and 3 by C57BL/6 and BALB/c mice (Rv1131, Rv1221 and Rv0470). In the C57BL/6 mice and, to a lesser extent, in the other two mouse strains, medLN cells produced multiple cytokines in response to single Mtbantigens. The different combinations found were: IFN-γ, TNF-α and IL-17 against 3 antigens (in C57BL/6 mice: Rv1886, Rv1131 and Rv1980); IFN-γ and TNF-α against 5 antigens (in C57BL/6 mice: Rv0287/88, Rv0470, Rv1791, ESAT6/CFP10 and Rv0440. For Rv0440, the same cytokine profile was found in C3HeB/FeJ mice); IFN-γ and IL-17 against two antigens (in C3HeB/FeJ mice: Rv0826 and Rv1980); TNF-α and IL-17 against 2 antigens (in C57BL/6 mice: Rv3616; in BALB/c mice: Rv1131).

Antigen-specific splenocyte responses to Mtb antigens were mainly detected – as measured by either one or more of the three cytokines evaluated – in C57BL/6 mice (11 antigens recognised) and to a lesser extent in BALB/c (3 antigens recognised) and C3HeB/FeJ (2 antigens recognised) mice. Among these, only the recognition of one antigen, Rv1886, was shared between C57BL/6 and BALB/c mice, and one other (Rv1791) between C57BL/6 and C3HeB/FeJ. Only splenocytes from C57BL/6 mice produced multiple cytokines, i.e., IFN-γ and TNF-α, against 5 antigens (Rv1886, Rv0287/88, Rv0470, Rv1791 and Rv3615).

Lung cells consistently recognised fewer antigens than cells isolated from other tissues (Fig.1c). However, some of these antigens were recognised by lung cells of different mouse strains (Rv0287/88 was recognised by all mouse strains; Rv1886 by C57BL/6 and BALB/c; ESAT6/CFP10 and Rv1791 by C57BL/6 and C3HeB/FeJ). Significant levels of IFN-γ and IL-17 were secreted by lung cells in response to 4 antigens (in C57BL/6 mice: Rv0287/88 and Rv0470; in BALB/c mice: Rv1886; and in C3HeB/FeJ mice: Rv1791). Only against one antigen (Rv0287/88) a combined IFN-γ, TNF-α and IL-17 response was found in BALB/c mice.

Most of the differences listed above were no longer significant after the very stringent FDR correction, but importantly this included also responses to controls and to previously described antigens known to be highly immunogenic such as ESAT6/CFP10 (Figs 1b and 2c). Therefore, the results are shown both with and without FDR correction, to avoid the risk of missing biological relevant candidates at this early stage of discovery.

Fig. 2: IVE-TB antigen-specific responses after Mtb infection and BCG vaccination in C3HeB/FeJ (C3H) mice.

a C3HeB/FeJ mice were either challenged intranasally with H37Rv-Mtb strain (105 CFU) (10 mice), vaccinated subcutaneously with BCG (5 mice) or left untreated (5 mice). After 6 weeks, splenocytes were harvested and stimulated per mouse for 6 days with a set of eighteen Mtb antigens or positive controls (ConA, PPD, bead disrupted BCG (BCG bub) and Mtb lysate), or negative control (HPV16E6 recombinant protein). The IFN-γ concentrations detected in the unstimulated samples were subtracted from the stimulated samples for each condition within each mouse. b IFN-γ production after antigen stimulation in splenocytes from naïve unimmunized mice (left panel), and BCG immunised versus Mtb infected mice (right panel). Dots display the median IFN-γ production. The (Rv) codes of the antigens are indicated if the median IFN-γ production exceeded 500 pg/ml in both BCG immunised and Mtb infected mice. The horizontal dotted line indicates the cut-off at 500 pg/ml, which is twice the median IFN-γ production found in response to the negative control in BCG immunised and Mtb infected mice (right panel). c IFN-γ responses against Mtb antigens in splenocytes from naïve, BCG immunised and Mtb infected mice are shown. Dots represent a single mouse (10 mice in the Mtbinfected group and 5 mice in the naïve and BCG immunised group in total) and bars indicate the median IFN-γ response of each group. The Mtb challenged and BCG immunised groups were compared to the naïve group using the Kruskal–Wallis test and a p-value < 0.05 was considered significant. Asterisk marks differences that remained significant after multiple test correction (q-value < 0.1, i.e., FDR-adjusted p-values of Mann–Whitney test).

Interestingly, in the resistant mouse strains, there were a few antigens (Rv0470, Rv0826, Rv0991, Rv1131, Rv1221, Rv1733, Rv1846, Rv1980, Rv2034, Rv2873, Rv3052, Rv3353 and Rv3616) that induced statistically significant TNF-α but not IFN-γ responses. Except for Rv1846, this lack of overall significant IFN-γ responses seemed to be the result of the observed heterogeneity in donor responses rather than the complete absence of this cytokine (Supplementary Fig. 3).

To validate the results in a subsequent independent multivariate analytical approach, we used PLS-DA modelling to determine the key antigens that drove the discrimination in cytokine responses between (i) time points (E/C, Rv0642, Rv0287/88, Rv0991, Rv1791, Rv1886, Rv2029 and Rv3616); (II) mouse strains (E/C, Rv0287/88, Rv0440, Rv0470, Rv0642, Rv1221, Rv1791, Rv1886, Rv2029, Rv2873 and Rv3615); (iii) organs (E/C, Rv0287/88, Rv1131, Rv1791, Rv1886, Rv2029, Rv3048, Rv3615 and Rv3616); and (iv) type of cytokine (E/C, Rv0287/88, Rv0440, Rv1131, Rv1791, Rv1886, Rv1980, Rv2029 and Rv3615), based on the Variable Importance in Projection (VIP) values (VIP > 1). The PLS-DA protein clusters (Supplementary Fig. 2a) were very similar to those obtained by unsupervised PCA (Supplementary Fig. 2b) confirming their validity. (Supplementary Data 1 and Fig. 1c). Of note, this independent multivariate approach confirmed 13 of the 19 proteins that were found as being interesting targets in our first analysis.

Overall, nineteen out of 27 selected Mtb antigens (Fig. 1c, in bold) elicited significant cytokine production from cells isolated from Mtb infected mice compared to those found in uninfected animals. Of note, although as expected these responses were somewhat heterogeneous, 13 antigens appeared to drive discriminatory cytokine responses between the different mouse strains, time points, type of cytokine and tissues analysed in this study. But perhaps most importantly, the results show that human T cell recognition profiles translate well to mouse models, which was a major aim of this study.

IVE-TB antigen-specific responses after live BCG vaccination

A TB subunit vaccine, in order to be effective, should preferably boost immune responses induced by prior vaccination26. BCG, despite its insufficient protective efficacy against adult TB, remains the only licensed vaccine against TB and is highly effective in protecting against severe childhood TB2. To evaluate whether immune responses against IVE-TB antigens could be induced by BCG immunisation, and to compare those responses to those induced by Mtbinfection, C3HeB/FeJ mice were either challenged with H37Rv-Mtb strain (105 CFU) (10 mice), vaccinated subcutaneously with BCG (5 mice) or left untreated (5 mice) (Fig. 2a). After 6 weeks, splenocytes were harvested and stimulated for 6 days with a set of 18 Mtb antigens selected from our first screen (Fig. 1c). Positive (ConA, PPD, bead disrupted BCG, Mtb lysate) and negative (HPV16E6 recombinant protein) controls were also included. The IFN-γ production in response to the different stimuli was measured and analysed, and as above concentrations detected in the unstimulated samples were subtracted from each condition within each mouse. TNF-α and IL-17 responses were only present at low levels in C3H mice in the first experiment (Fig. 1b) and therefore not assessed in this screening. Three antigens (Rv0470, Rv1733 and Rv3616) were strongly recognised by Mtb infected as well as BCG vaccinated C3HeB/FeJ mice (i.e.: IFN-γ production >500 pg/ml) (Fig. 2b). Additionally, despite the different experimental setup and laboratory setting, eleven Mtb antigens (Rv0287/88, Rv1886, Rv0470, Rv1131, Rv0440, Rv1980, ESAT6/CFP10, Rv3616, Rv1733, Rv1221 and Rv3052) were confirmed to be significantly recognised after Mtb challenge in C3HeB/FeJ mice (Fig. 2c). Interestingly, besides ESAT6/CFP10 (E/C), of which peptides are already widely used as TB diagnostic tool in IGRA tests27, Rv0287/88, Rv1886, Rv1131, Rv1221, Rv1846 and Rv3052 were the only antigens for which IFN-γ responses differed highly between the Mtb infected and the BCG vaccinated group (Fig. 2c). Prior to FDR correction, differences between the Mtb infected and the BCG vaccinated group were significant only for E/C (as expected), Rv3052 and Mtblysate.

Collectively, these findings confirm, extend and partially validate the immune recognition of eleven Mtb antigens previously identified in humans across three genetically unrelated mouse TB-infection models and two different independent laboratory settings.


An important requirement for novel TB candidate vaccines is that they evoke immunity to Mtbproteins that are expressed, processed and presented to the host immune system during infection. Over the last years, we identified multiple Mtb proteins recognised by immune blood cells from latently Mtb infected individuals and TB patients3,9. During the antigen selection process, we selected proteins encoded by genes highly expressed either during the entire course of infection, or during specific stages of in vivo Mtb infection3. Before entering clinical studies, candidate vaccines need to be evaluated and proven to be immunogenic in animal studies11. Therefore, we here examined a selection of Mtb antigens identified in our previous in vitro human studies and backtranslated this to three different, genetically and TB-susceptibility diverse mouse TB models. We found that among a set of 27 Mtb antigens recognised by human cells, 19 were able to induce the production of at least one of three cytokines (IFN-γ, TNF-α, IL-17) in at least one of the tissues examined in each of the three mouse strains after challenge with Mtb. Further confirmation of our previous data in human cohorts was that several Mtb antigens induced considerable cytokine responses other than IFN-γ9,28. Importantly, and in accordance with current recommendations17, the results were partially validated in a second independent laboratory and experimental setting (Supplementary Table 1), showing recognition of 11 Mtb antigens in infected C3HeB/FeJ mice. Among those, three antigens (i.e.: Rv0470, Rv1733 and Rv3616) were found to be recognised also after BCG vaccination. The latter finding is essential when considering the design of BCG-booster vaccines against TB2.

Overall, in the experiments performed in the first laboratory, we observed highly significant responses to Mtb antigens which varied by magnitude, cytokine combinatorial patterns, tissues and mouse strains (Fig. 1). As previously observed in human studies9, the differences in the levels and types of cytokines induced by an antigen, or the fact that certain antigens did not elicit detectable immune responses, is not linked to a particular protein function. Interestingly, non-IFN-γ responses were detected in the spleen and the mediastinal draining lymph nodes of the resistant C57BL/6 mouse strain after stimulation with a substantial number of antigens (Rv0470, Rv0826, Rv0991, Rv1131, Rv1221, Rv1733, Rv1846, Rv1980, Rv2034, Rv2873, Rv3353 and Rv3616), which agrees with our previous findings9 and also with the recent description of non-canonical IFN-γ-independent immune responses found in the peripheral blood of Mtbexposed “resistant” individuals28. Cells from C57BL/6 mice recognised significantly more Mtbantigens than those from BALB/c and C3HeB/FeJ mice. Although C57BL/6 and BALB/c mice are equally resistant to TB29, several genetic determinants might have contributed to the differences found in the antigens recognised. For instance, the adaptive type 1 immune response to pulmonary mycobacterial infection appears later and with lower magnitude in BALB/c compared to C57BL/6 mice30,31. In line with our findings, the lower magnitude of IFN-γ and TNF-α production levels in BALB/c mice was already described in splenocytes and lung lymph node cells upon antigen stimulation31. Unexpectedly to us, we found that lung and mediastinal lymph nodes cells from C3HeB/FeJ mice failed to produce detectable TNF-α, not only after Mtb antigens exposure, but also upon some of the positive mitogen controls. During mycobacterial infection, TNF-α deficient mice develop necrotic granulomas, pathogenic lesions which are also characteristic of the C3HeB/FeJ mice32,33. Although cells isolated from lung of C3HeB/FeJ mice can produce TNF-α, they do that to a lesser extent than C57BL/6 mice, potentially explaining why they develop necrotic lesions. The lower TNF-α production in C3HeB/FeJ mice might be caused by a greater and premature accumulation of a functionally exhausted T cell subsets, characterised by a limited capacity to secrete IL-2 and TNF-α34. Demonstrating the recognition of Mtb antigens in this TB susceptible strain was particularly important from a corrective vaccine perspective. Redirecting the cellular response in such strains towards a protective response (e.g., that found in the resistant strain) could be achieved by immunising mice with recognised antigens and appropriate adjuvants (e.g. CAF01), as has been done for instance with ESAT6, which is a strongly antigenic protein in active TB patients, but is also a powerful antigen for vaccination when delivered in CAF0135.

Only two antigens were consistently recognised by cells isolated from the same tissue across all mouse strains: Rv0287/88 and Rv0826. Rv0287/88 is a heterodimer secreted by the ESX-3 secretion system, implicated in Mtb virulence and survival36,37. Lung cells from all three mouse models produced IFN-γ when stimulated with Rv0287/88, while TNF-α and IL-17 were secreted concomitantly in BALB/c mice and TNF-α in C57BL/6 mice. In the latter strain, also splenocytes and mediastinal lymph nodes cells responded to Rv0287/88 by producing IFN-γ and TNF-α. These findings validate and extend what has been shown for lung T cells of CB6F1 (BALB/c × C57BL/6) and B6C3F1 (C57BL/6 × C3HeB/FeJ), which are able to recognise Rv0287/88 after 21 days of Mtb infection38. Although evidence for the antigenicity of Rv0287/88 accumulates, susceptible C3HeB/FeJ mice immunised with BCG did not recognise this fusion protein. This contrasts with another study, performed in a TB-resistant strain, wherein splenocytes of BCG vaccinated mice responded to specific Rv0288 epitopes39. Rv0288 constitutes part of the H4:IC31, an adjuvanted polyprotein vaccine that, though immunogenic, failed to boost neonatal BCG vaccination in a recent prevention of (sustained) infection clinical trial4. Testing this vaccine in BCG immunised mice with different TB susceptibility phenotypes might provide improved models that are more predictive of humans.

At an early stage of Mtb infection, Rv0826 was recognised by medLN cells from all mice, mainly inducing TNF-α in C57BL/6 and BALB/c mice and IFN-γ and IL-17 in C3HeB/FeJ mice. Rv0826 is a conserved protein with unknown function, highly expressed during enduring hypoxia and in extensively drug resistant Mtb strains40,41. These characteristics, together with the strong cellular recognition found in humans42 and in mice, would make this antigen an attractive target. However, we observed consistent unspecific immune responses against Rv0826 in the splenocytes of all mice, verified in an independent experiment (Fig. 2b), which deserves further investigation. This might be due to cross-reactivity to other microbial antigens, e.g., microbiota derived, but this remains currently speculative.

Besides Rv0287/88 and Rv0826, other antigens were consistently recognised across tissues and mouse strains, and better able to elicit cells producing multiple cytokines (Rv1886, Rv1791, Rv0470, Rv1131, Rv0440, Rv1980) than other antigens (ESAT6/CFP10, Rv3615, Rv3616, Rv1733, Rv1221, Rv1846, Rv0991, Rv2034, Rv2873, Rv3052 and Rv3353). Among those, Rv0470c was particularly interesting because it is an IVE-TB antigen consistently recognised across mice both after Mtb challenge and after BCG vaccination, in independent experiments and in different laboratories. Additionally, though recognised in susceptible mice, Rv0470 induced different magnitudes and type of cytokine responses in the resistant strains (Fig. 1b, c). Rv0470c is an enzyme essential for both the synthesis and cyclopropanation of mycolic acids and the cording morphology of both BCG and Mtb43,44. In addition, Rv0470c prevents phagosome maturation in human monocyte-derived macrophages, while in mice, its deletion alters the persistence and pathology of Mtb at late stages of infection43,45. Of note, immunisation with BCG_0470 increased the delayed-type hypersensitivity as well as Th1 responses in BCG primed BALB/c mice46. All these features, supported by our data in mice and humans9,10, point to Rv0470c as an attractive BCG booster vaccine candidate when delivered with a properly selected adjuvant47.

Interestingly, Rv0287/88, Rv1886, Rv1131, Rv1221, Rv1846 and Rv3052, though highly homologous to M. bovis proteins, were the antigens, besides ESAT6/CFP10, for which IFN-γ response highly differed between Mtb infected and BCG vaccinated groups (Fig. 2c)9. Comparing the specific responses between TB patients and non Mtb exposed BCG vaccinees might reveal the potential of these antigens as adjunct TB diagnostic tools.

In conclusion, this study demonstrates the consistent recognition of several promising IVE-TBMtb antigens among multiple genetically and TB susceptibility diverse Mtb infected mouse strains, across various infection models and tissues and time points evaluated, thus offering translational tools to evaluate Mtb antigens in appropriate animal models.

Article classification: Biological abstract
Share to:
Add:Room A11-329, 1st Floor, No.1, SBI Venture Street, Optics Valley, East Lake
New Technology Development Zone, Wuhan, China.
Certificate NO.:U18Q28010569R0S