Immunodeficiency in individuals co-infected with the human immunodeficiency virus type 1 (HIV-1) causes a breach in host containment of Mtb resulting in increased incidence of reactivation of TB. With the spread of human immunodeficiency virus (HIV) and dramatic increase in the number of cases of multidrug-resistant (MDR-TB) and extremely drug-resistant (XDR-TB) infections, the WHO in the early 1990's declared TB as a re-emerging infectious disease. WHO estimates that there are more than 8.7 million new cases of active TB each year (WHO, 2012) and about 2.6 million new cases of HIV infection with 1.8 million AIDS-related deaths per year (UNAIDS, 2010). Alarmingly, about 14 million individuals worldwide are estimated to be dually infected with TB-HIV (Getahun et al., 2010). TB is the largest cause of death in AIDS patients, (~29%), most of which (99%) occur in developing countries (Pawlowski et al., 2012). In HIV/TB co-infection, both pathogens combat the host's immune responses through various mechanisms. HIV depletes CD4+ T cells of the lamina propria of the gastrointestinal tract (Clayton et al., 1992; Brenchley and Douek, 2008) resulting in T cell dysfunction in cytokine production and cytotoxic activity, and therefore compromising the immune response. HIV/TB co-infection dramatically increases the risk of latent TB reactivation by 20- to 37-fold, and is therefore the most powerful risk factor for latent TB reactivation (Getahun et al., 2010).

Antimicrobial therapy with various regimens is used to treat TB. However, their inadequate and inappropriate use has triggered an increase in MDR-, XDR-TB, and total drug-resistant (TDR-TB) cases worldwide, adding to the challenges of TB eradication and healthcare cost (Caminero, 2010). Based on drug resistance data from 114 countries and two-regions of China, it was estimated that ~500,000 MDR-TB cases emerged in 2006, representing 4.8% of all cases of TB (WHO, 2007). While drug-sensitive strains of Mtb in immunocompetent patients are treated for at least 6 months with multiple antibiotic regimens to achieve efficient clearing of the mycobacteria by the host, in MDR-TB or XDR-TB cases second-line drugs are used, with a higher risk of adverse effects and lower potency, raising the costs of treatment for patients carrying MDR strain dramatically (Dye et al., 2002; Mukherjee et al., 2004). In addition, although multidrug combinations treat TB successfully, the used multidrug regimens are ineffective against MDR-, XDR-, and TDR-TB and have not prevented transmission of disease in endemic regions due to high infectivity rate and lack of rapid diagnostic tests to detect active TB infection.

The ability of the mycobacterium to enter latency with no clinical symptoms of disease leads to poor, untimely diagnosis and treatment, which negatively impacts TB elimination strategies in the United States and many other countries (ATS/CDC/IDSA, 2000, 2005). The tuberculin test (TST), which is a conventional method for detection of Mtb infection, is known to have important limitations. TST was shown to cross-react with Bacille-Calmette-Guerin (BCG) vaccine and with non-TB mycobacterial (NTM) species and is particularly difficult to interpret in BCG-vaccinated foreign-born individuals from TB-endemic countries (Agger and Andersen, 2002; Shah et al., 2012). Also, inter-reader variability in result interpretation affects the outcome of TST specificity, affecting the treatment schedule and therefore the overall control of disease.

After inhalation of Mtb through aerosol particles, the mycobacteria are phagocytosed by alveolar macrophages and dendritic cells (DC) and subsequently transported to the site of infection, mainly the lungs. Inside the macrophage, Mtb is trapped in the phagosome where it replicates until the innate immune response activates the macrophage in an effort to control the pathogen. During this time, the bacteria are capable of arresting phagolysosome formation, through a few known mechanisms, preventing bacterial death from high pH and hydrolytic enzymes present in phagolysosomal compartment. In DC, the pathogen is carried to the draining lymph nodes, where Mtb–derived antigens are presented to T lymphocytes and Mtb antigen-specific T cells that are generated. These cells recirculate and induce formation of granulomas in the lung. The granuloma is comprised of a large number of immune cells that altogether with a fibrous cuff restrain bacterial cells from spreading. Although the innate immune response slows down the physiological dynamics of Mtb, it rarely leads to complete killing of the bacteria, and most often it promotes latency, whereby the bacteria is contained in the granuloma.

TB vaccine candidates should elicit cellular immune responses that are important in controlling Mtb. T lymphocytes are generally believed to mediate immunity against TB based both on animal models of infection and human data. Interferon-γ (IFN-γ) producing CD4+ T helper cells (Th1) and antigen-specific CD8+ T cells clearly play an important role in immunity against TB (Cooper et al., 1993; Flynn et al., 1993; Munk and Emoto, 1995; Bastian et al., 2008). The MHC class I restricted T cells also participate in resistance to Mtb infection (Flynn et al., 1992). This resistance appears to be mediated by grazymes A and B and by granulysin (Canaday et al., 2001). Tumor necrosis factor-α (TNF-α) and p55 receptor expression are associated with TB resistance (Flynn et al., 1995). CD4+ Th2 cells do not generally participate in mediating protection against TB, but a role for B cells in host defense is becoming apparent (Maglione and Chan, 2009) as B cell deficient mice were found to be more susceptible to TB (Vordermeier et al., 1996; Maglione et al., 2007).

IL-17-producing T helper cells (Th17 cells) have been implicated in the early phase adaptive immunity and host defense against Mtb (Khader et al., 2007). In addition, Th17 cells were also implicated in human and murine responses to BCG vaccination (Scriba et al., 2008; Cruz et al., 2010). Mtb challenge of mice vaccinated with BCG resulted in an increase in interleukin IL-17, TNF-alpha, IL-6, and MIP-2 expression along with lung tissue damage (Cruz et al., 2010). Based on the possible role of Th17 in Mtb infection, it is proposed that a protective TB vaccine will need to induce both Th1 and Th17 memory responses. It is hypothesized that the vaccine-elicited Th17 cells will be sequestered in the lung and upon subsequent challenge with Mtb, the Th17 memory cells will proliferate rapidly, producing IL-17 and triggering the local expression of chemokines that will recruit protective Th1 cells for the clearance of Mtb in the lungs. Therefore, a Th17 response will clearly be beneficial since it will accelerate the recruitment of Th1 cells for rapid bacterial clearance. Current vaccine studies are now focusing on IL-17-producing memory cells in the hopes of improving vaccine strategies against TB in humans.

In the next sections we will emphasize novel approaches and strategies in TB vaccine development, highlighting scientific issues that need to be addressed in this complicated field of vaccine development (Figure ​(Figure11).

An enormous challenge to studying latent TB infection and evaluating vaccine candidates against this form of infection is the unavailability of an adequate animal model that can mimic the human LTBI infection. The latency model generally referred to as the Cornell model was first described in the late 1950s (McCune et al., 1956, 1966a,b; McCune and Tompsett, 1956). In this model mice are inoculated intravenously (i.v.) with ~2× 10⁶ viable bacilli of virulent Mtb and the resultant infection is treated for 12 weeks with the anti-mycobacterial drugs isoniazid (INH) and pyrazinamide (PZA) beginning within 20 min after infection. After the 12-week antibiotic treatment, no tubercle bacilli can be cultured from the animals' organs for many months. However, at this time point, administration of cortisone (at immunosuppressive doses) at 2–3 months after the interruption of the antibiotic therapy reverts this condition, and Mtb can be cultured from lungs and spleens of ~50% of the animals. Despite having the advantage of achieving and maintaining for many weeks very low numbers of the tubercle bacilli within the tissues of infected mice, this model has three major limitations: (1) Dormancy is difficult to standardize because the optimal antibiotic concentration and duration of treatment to achieve low numbers of bacilli varies from experiment to experiment; (2) Only 50% of the animals successfully treated with antibiotic develop dormant infection, which imposes a major complication in the interpretation of the experiments (Lenaerts et al., 2004); and (3) Most variants of the Cornell model use high doses of immunosuppressive agents to achieve reactivation, which by definition constitutes a complication for studies designed to evaluate the host immune response during reactivation of the disease. An interesting alternative of animal model of latent TB in both mice and guinea pigs has been recently proposed (Kashino et al., 2006, 2008a). A streptomycin-dependent (auxotroph) Mtb strain 18b that was isolated in 1955 in Japan from a patient with streptomycin-resistant TB (Hashimoto, 1955) was utilized in these studies. The Mtb 18b strain, which was demonstrated to be entirely streptomycin-dependent, was unable to grow unless streptomycin is present in the growth media (Hashimoto, 1955; Honore et al., 1995). In this model of TB latency, the Mtb 18b strain was shown to replicate in the lungs and spleens of mice and guinea pig treated with streptomycin and upon antibiotic treatment withdrawal the bacteria stopped replicating concomitant with the over-expression of α-crystallin (Kashino et al., 2006, 2008a) which is a protein member of dormancy regulon (DosRS) and one of the most abundantly produced protein during exposure to hypoxia and nutrient starvation (Cunningham and Spreadbury, 1998; Sherman et al., 2001), leading to the conclusion that Mtb 18b mouse infection model mimics latent TB. The Mtb 18b infection was characterized by formation of granuloma and the persistence of low numbers of viable non-replicating bacilli in mice and guinea pigs for at least 6 months. It was also shown to induce resistance to reinfection with virulent Mtb and potent T-cell responses to native and purified recombinant Mtb proteins (Kashino et al., 2006). Intriguingly, after starvation for longer periods of time, the bacilli resume replication once streptomycin is added back to the culture, confirming that streptomycin is required for exiting the dormant state in this particular strain and was used recently for screening of drugs against non-replicating bacteria (Sala et al., 2010).