Bovine tuberculosis is the fourth most important livestock disease worldwide . The benefits of developing, applying and maintaining improved control and eradication strategies for BTB are manifold, and directly impact on human and animal health . The specific immune cell signalling pathways involved in the immune response to BTB are highly complex and poorly characterised in cattle. This has obvious limitations for the understanding of, and design of improved diagnostics and effective therapeutics. However, studies on tuberculosis in the human and murine models have highlighted the involvement of cell regulatory signalling pathways in the immune response that are also likely to be relevant in BTB. Whereas traditionally, studies of BTB have focused on adaptive immunity, the findings from these studies are pointing toward a critical role for signalling via the innate immune system, including TLRs in initiating and directing the subsequent immune response and determining the outcome of infection [8–10].
The timing and potency of the cellular and immunological events that occur immediately post-infection are crucial determinants governing infection . Pathogen-induced phenotypic changes in host cells are often accompanied by marked changes in gene expression due to host- and/or pathogen mediated reprogramming of the transcriptome during infection . Previous work by our group compared the gene expression differences between bovine tuberculin-stimulated and non-stimulated PBMC from BTB-infected animals  and demonstrated that stimulation with bovine tuberculin induced significant gene expression changes that can be useful for dissection of the immune response to BTB. Subsequently we identified a novel gene expression profile indicative of innate immune gene repression in heavily infected cattle in vivo . Expression clustering of these data yielded a gene infection signature for disease and highlighted genes and regulatory pathways, including the TLR cell signalling pathway .
In the present study we have shown that after overnight recovery in vitro, PBMC from BTB-infected cattle are significantly more responsive to bovine tuberculin stimulation than control animal PBMC. Gene expression levels were estimated in PBMC from BTB-infected and healthy controls either non-stimulated or at 3 h and 12 h after bovine tuberculin stimulation using a common reference microarray approach (Fig. 1). Significant gene expression changes were observed in response to bovine tuberculin stimulation for both animal groups over a 12 h time course. The microarray data indicated that substantially different gene expression profiles were evident in the BTB-infected animals relative to the control animals (Fig. 2a and Fig. 2b). Following a 12 h incubation with bovine tuberculin, these analyses also showed that the immune response in the infected animal group was both rapid and transient (Fig. 2a). Analysis of gene expression differences across the time course showed that differences between groups were most evident in the period between 0 h and 3 h after bovine tuberculin stimulation-indicating that the early response to bovine tuberculin is substantially different between the two animal groups.
Comparative analysis of the PBMC gene expression program in response to bovine tuberculin identified a panel of 18 genes that were significantly differentially expressed in both animal groups. Interestingly, all of these genes were expressed in opposite directions in the two groups. Expression of 15 of the 18 genes was increased in PBMC from BTB-infected animals including genes encoding proteins involved with transcriptional regulation (TLE3, TAF6, HCFC1 and GATA4), a chemokine (CCL1), and a chemokine receptor(CSF2RA). In contrast, only three genes were decreased in expression in the BTB-infected group. Conversely, the opposite pattern was observed in PBMC from the control animals with 15 genes decreased, and three genes increased in expression relative to the BTB-infected group (Fig. 3 and Table 1). These data suggest a number of important gene targets for further study, as well as identifying cell regulatory pathways that may be differentially regulated in BTB-infected animals.
Of the 250 microarray spot features that were differentially expressed in the BTB-infected animals, those displaying increased expression in response to bovine tuberculin outnumbered those with decreased expression by a factor of two. We have previously shown that the expression of a number of the genes represented by these spot features, including TLR2, TLR4 and NFKB, was significantly repressed in heavily infected cattle in vivo . The transformation associated with gene repression in vivo to gene activation detected in the same animals in vitro was evident in the reversal in direction of expression of a number of well characterised genes including those encoding TLRs, MHC molecules, and cytokines. Furthermore, the 41 genes examined by real time qRT-PCR confirmed the BOTL-5 microarray results and supported a trend towards a proinflammatory immune response to bovine tuberculin in PBMC from BTB-infected cattle. Expression of a gene encoding the key transcription factor and mediator of the immune response (NFKB1) indicated a distinctive proinflammatory gene expression program characterised by a significant two-fold increase in IL8 expression and 16-fold increase in IFNG expression (Fig. 4 and Table 2).
While the differentially expressed genes detected in this study provide evidence of a pre-existing gene expression program most likely caused as a result of M. bovis infection in the diseased cattle, other confounding factors could be responsible for some of the changes detected. Changes or fundamental differences in cell subpopulations between animals and the separation of PBMC from an in vivo immunosuppressive environment  into culture media may have affected some gene expression patterns. Although there was no evidence at post-mortem examination of clinical disease in the cattle caused by other infectious agents, the presence of undetected pathogens may have generated host immune responses in either group of cattle that could have accounted for some of the transcriptional changes detected. In addition, it is important to acknowledge that the control and infected cattle were sampled from different herds, and that this may represent a confounding factor in the analysis of gene expression differences between the groups.
Genes encoding adaptor and mediator molecules of the TLR activation pathway were also profiled by real time qRT-PCR to examine cellular pathways contributing to the differential response between BTB-infected and control animal groups. The gene encoding the Toll-interacting protein (TOLLIP), a negative regulator of TLR signalling  was examined and gene expression was significantly increased in both the control and BTB-infected animal groups (4.32 fold and 2.33 fold, respectively) in response to bovine tuberculin stimulation. TOLLIP has previously been found to impair TLR-2 and TLR-4 activation of NF-κB ; therefore, it might be expected to prevent the downstream signalling from TLR-2 and TLR-4 in healthy control and BTB-infected cattle samples respectively. However, significant increased expression of NFKB1 and genes encoding mediators of NF-κB (CHUK and IKBKB, Fig. 4 and Table 2), coupled with differential expression of over 250 spot features on the microarrays suggested that proinflammatory signalling was not inhibited in the BTB-infected animal samples. Interestingly, another TLR-4-specific accessory molecule, Toll-like receptor adaptor molecule 2 (TICAM2), discovered in studies of M. tuberculosis infection in mice can bypass the inhibitory effects of TOLLIP to transmit cell signals  leading to increased expression of the NFKB1 gene . Expression of the TICAM2 gene was examined using real time qRT-PCR and was found to be increased in the BTB-infected animal samples only (1.28 fold, Fig. 4 and Table 2). The increased expression of TICAM2 may possibly provide a mechanism through which proinflammatory gene activation is regulated in BTB-infected animals (Fig. 2a).
Our results are consistent with proposed mechanisms that drive an ineffective TH2 type response and contribute to the outcome of BTB infection in cattle. It has been suggested that the TH1/TH2 bias of the immune response can be determined by specific TLRs . Furthermore, TLR-2 activation is a less efficient method of proinflammatory gene activation and may play a role in TLR-2 mediated immunosuppression .
In the present study, increased expression of the IL10 and TGFB1 genes was detected in the BTB-infected group in response to bovine tuberculin stimulation (Fig. 4) and elsewhere this has been associated with decreased ability of PBMC and macrophages to restrict mycobacterial growth in both humans [38, 39] and mice [40, 41]. Furthermore, IL-10 has been implicated in the suppression of the proinflammatory immune response and subversion of the host bacteriocidal immune response . The results of the present study suggest that the 16-fold increased expression of the IFNG gene (Table 2) may drive the changes in gene expression detected in the BTB-infected animal group. In addition, although the IL10 gene is significantly increased in expression by 1.23 fold (Table 2), this may be insufficient to compete with the proinflammatory effects of IFN-γ. In this regard, it has been previously shown that the ratio of pro- and anti-inflammatory cytokines will determine the overall outcome of the immune response and subsequent correlated gene activation and/or repression [43, 11].
In many countries the presence of M. bovis-infected wildlife can act as reservoirs of BTB infection for livestock and there is increased information available on the host response to infection in a variety of these species [44, 45]. In our previous study , transcriptional profiling of infected and non-infected control animals in the absence of exogenous antigen stimulation demonstrated decreased expression of MHC class II molecules, and similar findings have been reported in deer in response to natural TB infection .Thacker and colleagues characterised the immunological responses of peripheral blood leukocytes (PBL) from M. bovis-infected and non-infected white-tailed deer to infection by monitoring cytokine gene expression after exogenous antigen stimulation . The infected deer displayed a significant 75-fold increase in the expression of the proinflammatory cytokine, IFN-γ, comparable to the response detected in cattle during the present study. One notable difference was that the increase in IL10 and TGFB1 gene expression in response to bovine tuberculin stimulation of PBMC from infected cattle was not detected in infected deer. However, post-stimulation time points differed between the two studies and this could account for the differences observed.
The results of the present study are consistent with work carried out on Johne's disease in cattle caused by M. avium subsp. paratuberculosis (MAP) [47, 48]. Those studies also detected a novel gene expression profile in PBMC from MAP-infected animals that was both rapid and transient across time [47, 48]. In a separate study, stimulation of PBMC with MAP was shown to suppress the proinflammatory immune response , and the authors also found evidence of a dynamic TH1/TH2 type response, which eventually gives way to a predominantly TH2 like response in MAP-infected animals . Such a temporal expression pattern, where peak production of IL-10 lags behind that of IFN-γ, has also been demonstrated in relation to MAP stimulation of bovine PBMC . An important outcome from these studies involving both MAP-infected and BTB-infected cattle, is that the host gene expression profiles resulting from infection by these two related mycobacteria are remarkably similar.