Mycobacterium tuberculosis (M.tb) remains among the leading causes of death from an infectious agent in the world and exacerbates disease caused by the human immunodeficiency virus (HIV) [1]. It is estimated that one-third to one-half of the 30 million AIDS deaths can be attributed to tuberculosis [2]. Approximately 28% of individuals living with HIV and tuberculosis reside in South Africa [3]. In these endemic regions, the footprint of co-infection is striking and hampers disease control. Numerous groups have examined the influence of M.tb on the development of HIV disease [4-6] as well as the converse - the impact of HIV on tuberculosis progression [6, 7]. An individual with HIV is more likely to reactivate latent M.tb infection and progress to active tuberculosis [8-10]. AIDS patients co-infected with M.tb often present with atypical radiographic findings and disseminated disease [11]. At present, dissection of the relationship between these two pathogens remains challenging.

M.tb strains of the W-Beijing family have been associated with multiple global outbreaks and may be coupled to the HIV epidemic. Cowley et al. [12] genotyped M.tb strains from archived post-mortem specimens from the Western Cape region of South Africa and found W-Beijing strains were rare or absent from samples collected between 1930 through the mid 1960s, but strains were recovered from the 1996-2005 samples. Cowley suggests the emergence of these strains is coincident with the spread of HIV infection. Middelkoop et al. and Caws et al. also reported an association between W-Beijing strains and HIV infection in South Africa [3] and Vietnam [13], respectively. In addition, W-Beijing outbreaks can involve multi-drug resistant (MDR) organisms [14]; some are “hypervirulent” compared to other virulent laboratory strains because infected mice fail to induce protective T_H1 responses needed to control bacterial replication, resulting in shortened survival [15]. Together, these reports suggest W-Beijing strains are found worldwide and are commonly associated with HIV-infected individuals.

While mice are used in tuberculosis research, this model is not utilized to study HIV because the virus does not replicate in this species. Whether any of the known viral proteins directly impact the host response to tuberculosis, independently of CD4 depletion, remains poorly defined. Tat, an HIV protein, is necessary and contributes to the transactivation of viral genes as well as some cellular genes [16-18]. Tat is released by infected cells, taken up by uninfected bystander cells [19], and is known to affect several host immune responses in vitro [20-24]. Our group and others have shown that Tat stimulates the expression of intercellular adhesion molecule 1 (ICAM1), endothelial selectin (E-selectin), and vascular cell adhesion molecule 1 (VCAM1) [21, 25, 26]. Still, other reports indicate Tat inhibits interferon-induced nitric oxide synthase gene activity in macrophages [27], which may impact the ability of the host to defend against infection with M.tb.

Since Tat is one of the best-studied HIV proteins and mediates many of the virus’s pathogenic effects, a transgenic mouse that expresses the Tat protein was used to study its effects in host defense against M.tb. Our laboratory has previously characterized the phenotype of the Tat transgenic (Tat-tg) mice [28] used in the present report. In this model, the Tat gene is under the control of the surfactant protein C (SPC) promoter that is active in type II alveolar epithelial and some bronchial cells of the lungs. We confirmed Tat expression by immunohistochemistry, reverse transcriptase quantitative PCR and by immunoblot analyses in lungs, but not heart or liver of transgenic mice. At baseline, the Tat-tg mice demonstrate notable phenotypic differences compared to the non-tg mice including “increased pulmonary cellular infiltration (but no visible differences in lung architecture), increased nitrotyrosine and carbonyl protein modifications, and increased levels of NFκB p65, manganese superoxide dismutase (MnSOD), and thioredoxin interacting protein” [28]. Although the Tat-tg mice showed increased indices of oxidative stress, there were no obvious signs of decreased viability, weight fluctuations, or accelerated aging compared to the non-tg animals. Also in the initial description of these mice, we found no significant changes in IL-1β, IL-12, or IL-6 at baseline, but the level of these inflammatory cytokines increase upon challenge of the Tat-tg mice with a sub-lethal dose of bacterial lipopolysaccharide (LPS).

To explore the role of HIV Tat on host susceptibility to M.tb, we infected Tat-tg and non-tg mice with a low-dose aerosol of a highly virulent W-Beijing strain of M.tb known as SA161 [29]. The pathogenesis of this strain has not been evaluated in this particular animal model. We found mortality was accelerated in the Tat-tg SA161 infected mice and disrupted granulomatous and inflammatory responses. We conclude this murine model is useful to address the basic pathogenesis of co-infection.

Tuberculosis has emerged over the past decade as the most prominent bacterial disease found in HIV-positive individuals worldwide [40, 41]. A relevant animal model of these co-infections would be highly useful; but whereas M.tb infects mice, HIV does not. An alternative approach is to derive a transgenic mouse that possesses organ-specific expression of key HIV proteins. Tat is a prime viral protein candidate because it is secreted from infected cells and taken up by uninfected bystander cells. Pulmonary complications cause morbidity and mortality in HIV-infected individuals even in the HAART era [42]. These complications are not only attributable to the direct presence of HIV in the lungs, but secreted viral proteins such as Tat. In the murine model used here, Tat is expressed in the lungs, providing a useful tool to study host immune responses to aerosol bacterial infection and to monitor its in vivo effects. Although it is difficult to accurately quantify the amount of Tat in the lung tissue of the Tat-tg mice, we have previously detected immunoreactive Tat protein from lung tissue via immunoblotting [28]. Using densitometry, we postulate the protein level of Tat in these mice is approximately 1-5ng/mg of total lung protein (unpublished). Westendorp et al. previously reported the circulating level of Tat from AIDS patients to be approximately 1.6ng/ml [43]. It is difficult to discern whether infection with W-Beijing SA161 impacts Tat expression or the levels of endogenous Tat protein; thus it would be sensible to monitor changes in the amounts of this viral protein in future studies. At present, we hypothesize Tat contributes to histopathological changes in the lungs and hampers efficient immune responses to M.tb.

Infection of animal models with laboratory strains of M.tb such as H37Rv may not adequately model the pathogenesis of naturally occurring tuberculosis isolates; thus we evaluated the immune response to M.tb W-Beijing SA161 infection. However, as a comparison, groups of non-tg and Tat-tg mice were also infected with H37Rv and their survival monitored. The significant finding from this study was that the Tat-tg mice were more susceptible to infection with W-Beijing SA161, as the mean survival time of Tat-tg mice was approximately 30 days, a considerably shorter survival time than the non-tg mice (Fig. 1a). Because necropsies were not completed, we do not know the exact cause of death, but it is possible lung consolidation is a major contributing factor. Also, because the lung was the only tissue in which bacterial load was monitored, the extent of extra-pulmonary dissemination of the bacteria in this experiment is not known. In contrast, all mice infected with H37Rv lived past the 100-day mark, suggesting infection with W-Beijing SA161 contributed to the quickened time to death in both groups of mice. A limitation to our survival assay was the small number of mice used per group, however we believe five mice per group was sufficient because a significant difference in survival between the Tat-tg and non-tg mice after W-Beijing infection (****, p=0.0039) was observed. These data are consistent with previously published work [44-46] that similarly reported increased mortality of mice infected with W-Beijing strains. Despite the differences in survival, the lung bacterial load at day 30 did not significantly differ between the two groups (Fig. 1b), however, the bacterial burden increased steadily up to this time point. These findings are consistent with previous reports of the rate of growth of other laboratory M.tb isolates after aerosol infection of non-tg mice [47]. These results suggest that bacterial burden is not the principal factor for mortality, but rather, we propose that Tat dysregulates the immune response and impairs immunity to bacterial challenge.

While histopathological analysis demonstrated that the uninfected Tat-tg mice show normal lung space with healthy alveoli (Fig. 2a), we did observe gross differences in the architecture of the lung granulomata-like lesions between the W-Beijing SA161-infected non-tg and Tat-tg mice (Fig. 2b-e). Non-tg lungs showed moderate, discrete foci of inflammatory infiltrates, suggestive of granulomata-like lesions. These findings are consistent with previous reports from co-author Ordway et al. regarding the pathology of C57BL/6 mice infected with a “hypervirulent” M.tb W-Beijing isolate, HN878, at this timepoint [30]. Per Ordway et al. macrophages and neutrophils likely comprise the inner most portion of the granuloma, surrounded by a “perivascular cuff” of newly expanded antigen-specific T cells [30]. Although it is difficult to do the side-by-side comparisons of our findings to Ordway et al., we believe Tat excessively recruits inflammatory and/or immunosuppressive cells to the site of infection, consequently resulting in compromised lung function. The Tat-tg granulomata were significantly less well-organized and showed considerable areas of diffuse inflammatory infiltrates, suggestive of lung consolidation, with very few sections of normal tissue. When the percentage of lung lesions at day 30 in each group was calculated, increased lung involvement in the Tat-tg group was confirmed (Fig. 2f). It is likely the bacteria disseminated to extra-pulmonary sites in the Tat-tg mice, however, since we did not culture the bacteria from locations other than the lungs, it is difficult to determine whether containment of the bacteria to the lungs was compromised; nonetheless dissemination is probable.

qRT-PCR was used to examine the changes in gene expression after M.tb W-Beijing SA161 infection. Analysis revealed increased expression of TNFα and IFNγ in the infected Tat-tg mice (Fig. 3), indicating that the combination of Tat and bacterial infection, but not Tat alone, alters the cytokine milieu. Although we did not quantify the protein levels of these pro-inflammatory cytokines, Ordway et al. confirmed that C57BL/6 mice infected with HN878 demonstrated significant production of IFNγ and TNFα early after infection [30]. TNFα and IFNγ responses have been shown to be important in host defense against M.tb [48, 49]. Although we observed increased expression of TNFα and IFNγ in the M.tb W-Beijing SA161-infected Tat-tg mice as compared to the non-tg group, we also noted a paradoxical decrease in survival. The reason for this is unknown, however, it is reasonable that infection with W-Beijing M.tb compromises lung function and architecture, concomitant to the onset of decline in pro-inflammatory responses.

Flow cytometry was used to elucidate the cellular populations after bacterial infection. While we did not observe significant differences in the number of macrophages, dendritic cells, or activated T cell populations in the lungs, we did find dramatically higher numbers of immunosuppressive cellular phenotypes including CD4+ CD25+Foxp3+ and IL-4 producing CD4+T cells in the Tat-tg mice 30 days post infection (Fig. 4b, d, respectively). Of note, the presence of Tat alone did not alter the cellular milieu in the absence of SA161 infection (Fig. 4b, d, “uninfected”). Our findings are again consistent to that of Ordway et al. who also observed a similar increase in the numbers of lung CD4+CD25+Foxp3+ T regulatory cells, dampened survival, and severe lung pathology in HN878-infected mice, despite a robust T_H1 response [29]. Our data do not allow us to discriminate where these infiltrating cells originate from, nor do they definitively prove increased numbers of CD4+CD25+Foxp3 T cells cause lung injury. It is possible that the heightened numbers of CD4+CD25+ Foxp3+ cells compromise lung function due, in part, to Tat-mediated increase of adhesion molecules that influence the adhesion of infiltrating cells in the lung such as ICAM1-expressing cells (Fig. 5a), specifically ICAM1+ CD8+ T cells (Fig. 5c). Liu et al. reported that Tat induces VCAM1 expression in human pulmonary artery endothelial cells [50] that facilitates lymphocyte attachment in the lung. D’Alessio et al. established an association between the presence of CD4+CD25+Foxp3 T cells and lung damage using a well-established model of acute lung injury via exposure to lipopolysaccharide (LPS) [51]. Moreover, this group detected the presence of T regulatory cells in the alveolar compartment of humans with acute lung injury.

Herein, we characterize the susceptibility and early immune responses (up to 30 days post infection) of non-tg and Tat-tg mice to low-dose aerosol of M.tb W-Beijing SA161. In parallel, we also examined the immune response to M.tb H37Rv; however we did not note any significant differences between the Tat-tg and non-tg mice infected with H37Rv in any of the parameters tested and were not pursued further. We show that M.tb W-Beijing SA161-infected non-tg mice form well-structured granulomata due to the organized and coordinated balance of the pro-inflammatory and immunosuppressive immune responses. This balanced immune response enables the non-tg mice to survive longer than the Tat-tg mice. Conversely, the Tat-tg mice succumbed to infection sooner, but this was not due to increased lung bacilli, since lung CFU counts were similar between the non-tg and Tat-tg mice. Rather, we posit that the presence of Tat in the lungs leads to increased ICAM1 expression resulting in the excessive recruitment of Foxp3 T regulatory and CD4+ producing IL-4+ cells to the lung. Influx of these cells into the lung results in disorganized and diffuse lesions that likely compromise lung function. While in the lung, these suppressive cells produce mediators that likely dampen pro-inflammatory responses at a time when these defenses are leveling off or decreasing. We postulate the aforementioned is mediated by TGF-β. In fact, Song et al. has documented that Tat enhances ICAM1 expression in astrocytes [25] and Windish et al. has reported ICAM1 is required for the induction of Foxp3 T regulatory cells through mechanisms mediated by TGF-β [39]. Taken together, these factors contribute to homeostatic imbalance, compromised lung function, and accelerated mortality. Although we did not directly assay for TGF-β in our model, Tat has been shown to induce TGF-β [52]. It is possible this disproportionate number of Foxp3 T regulatory cells and CD4+ IL-4 producing T cells likely suppresses ongoing T_H1 responses at a critical time at which these responses are abating, resulting in containment failure. In fact, a report by Sharma et al., demonstrated recovery of CD4+CD25+Foxp3+ T regulatory cells from bronchoalveolar lavage fluid from miliary tuberculosis patients [53].

At this time, we do not know the events that transpire in the W-Beijing SA161-infected mice after 30 days of infection. We anticipate that disease in the non-tg mice will progress in a manner consistent with previously published reports of murine tuberculosis using other M.tb W-Beijing strains. Previously, Ordway et al. demonstrated that at 100 days post infection, C57BL/6 mice infected with M.tb W-Beijing HN878, show larger and more numerous lung lesions that resulted in complete lung consolidation [29]. In terms of the immune response, Ordway et al. reported lung cells isolated from C57BL/6 mice infected with M.tb HN878 or H37Rv show comparable levels of IFNγ, TNFα, MCP-1, and IL-6 at timepoints taken later than 30 days post infection, but heightened T regulatory cell responses with HN878 infection. We expect that if we had performed longer-term studies, similar immunological responses would have developed in the M.tb W-Beijing SA161-infected non-tg mice used in our study at these later timepoints.

Based on our findings, we believe the presence of Tat in the lung dysregulates the normal pro-inflammatory and suppressive immune responses generated in response to infection with M.tb W-Beijing SA161.

The loss of CD4+ T cell responses in HIV-positive individuals is the basis of the increased susceptibility to multiple opportunistic infections. It is becoming increasingly evident that a range of end-organ pathologies, including the lung, cannot be directly ascribed to the presence of the virion at the tissue site. The new approach described here, using a mouse that expresses a key viral protein in the lungs, may provide new information about the pathogenesis of co-infections. We have been able to demonstrate that the presence of Tat disrupts the normal host immune response to M.tb. To our knowledge we are the first to report a link between HIV-Tat and CD4+CD25+Foxp3+ T regulatory cells and increased lung pathology. It will be prudent to confirm the role of ICAM1 in the recruitment of Foxp3 T regulatory cells using neutralizing antibodies or siRNA against ICAM1 in future studies. In addition, future work should define the molecular mechanisms that govern the association between Tat and this particular population of adaptive T cells. Taken together, these data suggest that Tat influences and disrupts the orderly emergence of acquired immunity to tuberculosis infection.