Vaccines against Tuberculosis: Where Are We and Where Do We Need to Go?

In this review we discuss recent progress in the development, testing, and clinical evaluation of new vaccines against tuberculosis (TB). Over the last 20 years, tremendous progress has been made in TB vaccine research and development: from a pipeline virtually empty of new TB candidate vaccines in the early 1990s, to an era in which a dozen novel TB vaccine candidates have been and are being evaluated in human clinical trials. In addition, innovative approaches are being pursued to further improve existing vaccines, as well as discover new ones. Thus, there is good reason for optimism in the field of TB vaccines that it will be possible to develop better vaccines than BCG, which is still the only vaccine available against TB.


Introduction
It has proven challenging to develop vaccines against pathogens whose control depends on the cellular immune response. Tuberculosis (TB) is no exception. Most successful vaccines today target pathogens against which humoral immunity suffices to achieve protection and often sterile eradication. But TB vaccines need to drive primarily the cellular arm of the immune system. Mycobacterium tuberculosis (Mtb), the intracellular pathogen that causes TB, was discovered in 1882 by Robert Koch (Nobel Laureate in 1905) and is responsible for more human deaths than any other single pathogen today [1][2][3]. The combination with HIV co-infection, which dramatically compromises host resistance to TB, leads to high disease prevalence in affected endemic populations. Currently, each year more than 1.5 million people die of TB, and more than 9 million newly develop TB [4]. This dire situation is further compromised by the increasing prevalence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) Mtb strains [5], and the more recent occurrence of TDR (totally drug-resistant) Mtb strains, which are virtually untreatable [6].
TB, known also as the white plague, has been around for millennia. Early last century, hopes were that TB could be conquered by vaccination with the newly developed M. bovis BCG vaccine, isolated by and named after Calmette and Guérin in Lille, France [7]. These hopes were further boosted by the development of the first anti-tuberculous drugs during WWII by Selman Waksman (Nobel Laureate in 1952), who discovered that streptomycin was bacteriostatic for Mtb [8]. Initially, treatment with streptomycin appeared highly efficacious, but the tide turned when drug resistance rapidly developed, an early testimony of Mtb's ability to acquire drug resistance when treated by single antibiotics. Despite this early writing on the wall, the misconception that TB could be conquered by antibiotics and BCG vaccination led to complacency for several decades. This situation dramatically changed only in the early 1990s, when the World Health Organization (WHO) declared TB a global emergency.
From that time onwards TB scientists, who had been focusing much of their efforts on other areas of research and development due to a lack of interest in and funding for TB, were able to reorient efforts and initiate significant activities in the study of TB [9]. Soon, researchers determined Mtb's genome sequence [10], and began to dissect TB's immunology and cell biology.
Although in most cases, TB can be treated successfully with multidrug combinations of antibiotics (except for MDR-, XDR-, and TDR-TB), treating TB cases is clearly insufficient to interrupt disease transmission in highly endemic populations, because an active TB case will typically infect some 10 to 15 contacts [4] per year. Better preventive measures that block Mtb transmission need to be developed, including vaccines that prevent establishment of Mtb infection in the susceptible human host (infection-blocking vaccines), or alternatively, vaccines that prevent progression of established infection towards active TB disease (disease-progression-blocking vaccines). The only available vaccine against TB today remains the almost one-century-old BCG. It is routinely administered to infants in many countries worldwide and provides significant protection against severe forms of TB, mostly disseminating and meningeal forms. However, the protective efficacy of BCG against pulmonary TB in adults is inconsistent and incomplete, and BCG vaccination campaigns have had little impact on the occurrence of pulmonary TB, which represents the transmissible form of this disease. It is still unclear why the protective effect of neonatal or early-age BCG vaccination often begins to wane in early adolescence, at least in TB-endemic areas, as will be discussed further below. Nevertheless, in a North American study, BCG's protective efficacy was found to last for over 50 years [11], thus pointing to the importance of possible environmental modulators (co-infections, co-morbidity, nutrition, genetics, TB exposure intensity, etc.; see below). Another issue of concern that compromises BCG's utility is that infants with HIV have an increased risk of developing disseminated BCG-osis [12], bearing similarities to individuals with genetic deficiencies in the interleukin (IL)-12/IL-23/interferon-gamma (IFNc)/signal transducer and activator of transcription (STAT)1 pathway [13,14]. This implies that TB vaccines need to be developed that not only have a superior ability to induce protective immunity against TB, but also have a better safety profile compared to BCG [12,15].
There has been significant progress in TB research over the last 20 years. From a pipeline empty of new TB vaccine candidates in the early 1990s, more than 12 novel TB vaccine candidates now have been or are being evaluated in clinical trials in humans. Moreover, many new approaches and strategies are under evaluation for further improvement of these and other vaccines. This raises hope that vaccines can be developed which are better than BCG. It will be critical to mobilize the needed funding, political, and supranational support, and increase public awareness of the pressing needs in TB in order to step up efforts towards controlling this resilient disease. The recent declaration by the European Parliament to support TB vaccine research and development (http://www.tbvi.eu/news-agenda/ news/news-message/eu-parlement-discusses-the-importance-of-tbvaccines.html) is a laudable step in this direction.

TB Infection Control
Infection with Mtb does not necessarily lead to TB disease. Only 3%-10% of immunocompetent individuals that are infected will progress towards active disease during their life-time [16]. It is unknown if-and what proportion of-infected individuals are able to possibly achieve complete or ''sterile'' eradication of bacteria following primary pulmonary infection. Existing yet indirect evidence suggests that this may be the case and could be dependent on innate rather than adaptive immune defense mechanisms, as persons have been described who despite high intensity exposure-and thus likely infection-remain tuberculin skin test (TST)-negative and do not develop disease [17,18]. The TST is a classical indicator of delayed type hypersensitivitydependent T-cell activity towards Mtb antigens in the case of previous contact. TST surveys suggest that around one-third of the world's population, amounting for over 2.2 billion people, is infected with Mtb [4]. Although .90% are able to contain infection in a latent or subclinical stage, this huge reservoir of latently infected individuals fuels the high numbers of new active TB cases [19]. The development of active TB is dramatically accelerated by co-infection with HIV, which increases Mtb reactivation rates from 3%-10% per life-time to 5%-10% per life-year [20]. Latent Mtb infection also poses iatrogenic clinical challenges, due to increases in TB reactivation following administration of biologicals, such as tumor necrosis factor-alpha (TNFa)/IL-12/IL-23 blockers in treating inflammatory diseases like rheumatoid arthritis, Crohn's disease, and psoriasis [21]. New diagnostic tests that distinguish more specifically between Mtb infection and BCG vaccination or infection with non-tuberculous mycobacteria (NTM) than the classical TST have been developed recently, based on Mtb antigen-specific release of IFNc in whole blood or isolated leucocytes. These assays have been coined interferon gamma release assays (IGRAs) and have utility in clinical management, although sensitivity issues in borderline IGRA-positive and in remotely infected persons, which may result in false-negative test results, still need to be addressed [22].
A precise understanding of protective immunity and protective host defense is still needed. The study of individuals who are able to control Mtb infection in the long term may be particularly informative in this respect. Despite two decades of intensified research, we have to acknowledge that we still do not fully understand what exactly constitutes protective immunity and host defense against TB (see below section on TB immunology). If we could decipher essential mechanisms, pathways, and biomarkers of protection, then pathways of host defense could be targeted by rationally designed vaccines, and protection measured by accurate biomarkers [23,24]. Obviously, TB is much more complex than, for example, many viral infections, in which highly pathogenspecific mechanisms and correlates of protection exist in the form of neutralizing antibodies. An additional challenge in TB is the need to do better than achieve natural immune control of longterm infection: notably, by inducing immune responses that not only mediate life-long control of Mtb infection (which poses the inherent risk of endogenous reactivation), but also achieve sterile eradication of bacteria from the infected host. First-generation new TB vaccines in clinical trials are designed to achieve stronger immunity but may not necessarily succeed at sterile eradication (discussed below) [25,26].
Besides mediating protection on its own, the host immune response is also known to synergize with chemotherapy through incompletely defined mechanisms [27]. Thus not only preventive, but also therapeutic, TB vaccines that improve the host response to chemotherapy will be of significance; the latter may help shortening treatment, as well as reducing and perhaps treating drug resistance.

M. bovis BCG: The Current TB Vaccine
BCG is the one of the most widely administered vaccines worldwide, having been given over 4 billion times. BCG has been part of the expanded program on immunization (EPI) since the early 1970s and features relatively few serious adverse events. Only fairly recently has it become apparent that individuals with genetic defects in key immune genes or infants with clinically active HIV infection are highly susceptible to developing disseminating BCG disease, posing a significant risk in HIV-burdened populations where TB is often highly endemic [12]. This has led to changes in policy from the WHO Global Advisory Committee on Vaccine Safety, recommending that BCG should not be used in HIVpositive children.
Despite the relative efficacy of BCG in infants, the major unanswered question is why BCG fails to prevent pulmonary TB in adolescents. It has been proposed that immune memory wanes in adolescence, which is the most critical period for TB infection and/ or its progression to active disease. The fact that immunological memory is induced by BCG at an early age (neonates or infants) at which the immune system is not yet fully mature may provide one explanation. However, other factors may also contribute to decreasing BCG efficacy and/or enhanced susceptibility of young adults to TB. These factors include co-infections with helminths; viruses that compromise host immunity (an obvious, extreme example is HIV); and NTM that can induce immune regulation [28]. Helminth infection has been associated with increased TB incidence and reduced BCG vaccine efficacy in affected populations [29]. Such modulatory effects may be mediated by immune deviation, e.g., the induction of T helper 2 (Th2) or regulatory Tcell (Treg) responses, which dampen the protective efficacy of Th1dependent protection. A recent study in mice documented that helminth infection can impair innate pulmonary host defense against Mtb through the IL-4 receptor pathway [30].
Whatever the explanation for its inadequacy, BCG is clearly insufficient for worldwide TB control. Thus, there is a strong need to develop vaccines that can either boost BCG's initial priming and protective effects, or replace BCG by superior vaccines. Both strategies and approaches will be discussed below, after a short introduction to the key immunological features of TB.
The Immunology of TB Mtb organisms are transmitted by aerosols originating from the lungs of persons with active TB. Mtb-loaded aerosols are inhaled by TB contacts and reach the alveoli of their lungs, where they are taken up by various cell types, including alveolar macrophages, interstitial macrophages, local dendritic cells (DCs), and likely also epithelial cells (Figure 1: Infection). In humans and mice, Mtb infection is followed by a delayed onset of adaptive immune responses when compared to other infections [31]. Recent studies in mice have shown that Mtb possesses the unique ability to establish infection while delaying the onset of an adaptive immune response by 2-3 weeks. Using animals carrying genetically engineered T-cell receptors specific for Mtb antigens, it has been shown that priming of specific T-cell responses against Mtb takes place only in the local draining lymph nodes of the lung, and that transport of Mtb-infected cells from the lung parenchyma to the lymph nodes is significantly delayed for a period of at least 1-2 weeks [32][33][34][35][36]. It remains unknown which cell types are involved in Mtb antigen presentation to naïve T-cells, either DCs that crosspresent antigens from infected cells, or perhaps also DCs migrating from the lung [37]. It is thought that Mtb can actively inhibit the migration of infected cells from the lung to the draining lymph nodes, or that Mtb can infect cells with only a poor migratory potential like epithelial cells. In both cases, however, initiation of Mtb-specific T-cell immunity is delayed. This ''procrastination'' strategy likely allows Mtb to establish infection and build up a critical mass of infecting organisms, from which two types of populations can be derived: (i) ''dormant'', ''nonreplicating'', or persisting organisms that remain in a metabolic low activity state in infected cells, and present a potential reservoir of bacteria that can resuscitate later on in life (Figure 1: Latency); and (ii) actively replicating metabolically active bacteria that stimulate the host immune response, which confers protection but can also precipitate TB pathology. The latter is hallmarked by the formation of central caseous necrotic lesions, filled with extracellular infectious Mtb and cell debris [38].
Mtb is able to persist inside otherwise hostile host cells, notably professional phagocytes, by employing different immune evasion strategies. These will not be discussed here in detail, but include the inhibition of phagosome maturation, the inhibition of autophagy, the inhibition of apoptosis, egression into the cytosol, blocking of MHC antigen processing and presentation, and the inhibition of IFNc-receptor signaling, all of which help Mtb to evade host defence [39][40][41][42][43].
Following initial deposition of Mtb in the lung, macrophages and other immune cells are recruited during the early innate response to infection. The resulting cellular infiltrates become organized as primary granulomas. These are highly dynamic structures from which cells can rapidly in-and efflux. At a later stage, when adaptive immunity has been initiated, specific Tlymphocytes start infiltrating the granuloma; this leads to the formation of larger, well-organized, solid granulomas in which Mtb organisms typically are located mostly centrally. If control of infection and inflammation is not balanced optimally, the central regions of the granulomas will become necrotic and later caseous as just discussed [44], allowing outgrowth of high numbers of Mtb organisms. When these liquefying granulomas also damage airway linings, infected material is discharged into the airways. At this stage, TB infection has become contagious ( Figure 1: Active TB disease).
Thus immunity to Mtb is a two-edged sword: it protects the human host against disseminating infection, but also facilitates transmission of TB to contacts. Therefore, TB vaccines not only need to induce optimal immunity to Mtb, but also a balanced response that favors protective and avoids pathogenic mechanisms.
The Mtb-specific immune response comprises a multitude of different cell types. This ranges from classical and non-classical Tcells to neutrophils, B-cells, and natural killer (NK) cells. The role of CD4 + Th1-cells in TB is best understood. Th1-cells, which secrete IFNc and TNF as signature cytokines, are crucial for protective immunity against mycobacterial infections. Their induction is strictly dependent on the secretion of phagocyteproduced IL-12 [13]. Deleterious mutations in genes encoding proteins essential to the induction (IL-12p40 subunit, IL-12Rb1) or function (IFNcR1, IFNcR2, Stat1) of Th1-cells, or their depletion due to HIV infection, greatly enhance susceptibility to severe infections due to otherwise poorly pathogenic mycobacteria or Mtb [13,45]. This is also highlighted by the enhanced incidence of TB reactivation during treatment with biologicals targeting TNF, which are used in treating inflammatory disorders such as rheumatoid arthritis, Crohn's disease, and psoriasis [21].
CD4 + Th1-cells recognize Mtb antigens presented by MHCclass II molecules on DCs and macrophages [46][47][48], the dominant habitats of Mtb. As mentioned above, CD4 + T-cells are initially primed in the draining lymph nodes of the infected lung. Endosomal/phagosomal antigens like Mtb naturally access the MHC-class II processing route, which leads to the presentation of Mtb peptide-loaded MHC-class II complexes, providing ''signal 1'' at the DC surface. Only in the presence of a second signal, which is provided by co-stimulatory molecules, such as CD80, CD86, CD40, CD27, and 4-1BB/CD137, and a third signal, provided by essential cytokines like IL-12, however, full activation and differentiation of Th1-cells takes place. In the absence of any of these three essential components, T-cell activation may either not be initiated, or be preliminarily aborted; e.g., in the absence of signal 2, T-cell tolerance/anergy can result. Recent studies suggest that T-cell activation in TB granulomas is far from optimal but the limiting factors, which may be antigen/MHC-class I/II availability, lack of co-stimulation, or lack of supportive cytokines, still need to be elucidated [49].
Next to CD4 + Th1-cells, more recent evidence supports the prominent activation of Th17-cells in TB. This is a recently discovered T-cell subset that produces IL-17A, IL-17F, in many cases IL-22 and TNF, and in some cases also IFNc. Th17-cells are pro-inflammatory and mediate anti-microbial immunity against extracellular bacteria and fungi, particularly at mucosal surfaces. The exact mechanisms leading to Th17 activation and differentiation are less clear than for Th1-cells but are believed to involve phagocyte-dependent production of IL-1b or IL-6, and transforming growth factor-beta (TGFb). Other studies have pointed towards the essential role of IL-23, a closely related family member of IL-12, in the induction and expansion of Th17-cells or Th17 memory cells [50]. Current evidence suggests that Th17-cells play a significant role in the early phase of protection against high-dose aerosol Mtb infection in mice [51,52]. Th17-cells emerged in the early phase of adaptive immunity and contributed to host defense against intracellular pathogens, including Mtb [53]: IL17A has been found to be essential in forming mature granulomas in the lung following BCG or virulent Mtb infection [54]. Their role during chronic and latent infection is less clear [55]: hyperactivity of Th17-cells in previously Mtb-infected animals induced by multiple BCG revaccinations was detrimental and led to increased immunopathology with IL17-MIP2-dependent influx of neutrophils and tissue destruction rather than containment of infection in the lung [55,56]. The major source of early IL-17 during infection, however, may not be from CD4 + Th17-cells but rather from T-cell receptor gamma delta (cd) cells [54,56].
In addition to CD4 + T-cells, CD8 + T-cells also contribute to optimal immunity and protection against TB [48,57]. The mechanisms underlying CD8 + T-cell activation in TB are incompletely defined. It is clear that DCs possess multiple pathways to load MHC-class I molecules, either by classical cytosolic processing, or by alternative processing of phagosomelocated pathogens and endosome-located antigens. Active transmembrane transport, or leakage through micro-damaged membranes of Mtb phagosomal antigens to the classical cytosolic proteasome/MHC-class I presentation pathway, may also lead to MHC-class I loading with Mtb peptides. The recent suggestion that virulent mycobacteria can escape from the phagosome into the cytoplasm and thereby directly access MHC-class I processing/presentation may provide a new mechanism [43]. Thirdly, Mtb-infected cells can undergo apoptosis, leading to the formation of apoptotic vesicles that are taken up by DCs, after which the antigenic cargo is cross-presented through MHC-class I and class II molecules [58,59]. Fourthly, autophagy, which plays a prominent role in cellular homeostasis and in bacterial sequestration in vacuolar organelles, is involved in antigen presentation to, and cross-priming of, T-cells in the response to intracellular pathogens, including Mtb [60,61]. Thus, there are multiple pathways for activation of CD8+T-cells by phagosomal antigens.
Besides classical MHC-class Ia-restricted CD8 + T-cells, also MHC-class Ib-restricted CD8 + T-cells have received attention recently [62,63] . These include HLA-E-restricted cells [64,65]; lung mucosal-associated invariant T-cells (MAIT) that recognize Mtb antigens in the context of the non-classical molecule MR1 [66]; and CD1-restricted T-cells that mostly recognize lipid antigens derived from Mtb. CD1b molecules follow the MHCclass II presentation route, but relatively little is known for MR1 and HLA-E. HLA-E molecules are enriched in the Mtb phagosome, thus providing a plausible mechanism for Mtb peptide loading [66]. CD8 + T-cells are endowed with multiple mechanisms to attack Mtb-infected cells. Importantly-and in contrast to CD4 + Thcells-CD8 + T-cells are able to recognize non-phagocytic cells such as epithelial cells, which can also be infected by Mtb [67]. All nucleated cells express MHC-class I molecules while only professional phagocytes-or IFNc-activated cells-express MHCclass II molecules and can be recognized by CD4 + T-cells. Thus CD8 + T-cells are able to survey larger numbers of cells and a broader range of cell types compared to CD4 + T-cells. In addition, as already mentioned above, CD8 + T-cells are able to survey the cytoplasmic compartment of infected cells for pathogen-derived peptides through various antigen processing/presentation routes. Thus CD8 + T-cell immunity offers clear advantages and complementarity to CD4 + T-cell immunity.
CD8 + T-cells are able to secrete granules that contain cytotoxic molecules such as perforin, granzymes, and granulysin. These molecules can lyse host cells, and in the case of granulysin, also directly kill Mtb and other bacteria. In addition, CD8 + T-cells can induce apoptosis of infected target cells through ligating Fas or other TNF-R family-related cell-death receptors. Apoptosis may be activated to control Mtb infection (and is also counteracted by virulent Mtb [68]. Thirdly, CD8 + T-cells are also able to release Th1 cytokines such as IFNc, TNF, and in many cases also IL-2. These different effector pathways have been well documented in humans [69][70][71][72][73][74][75]. These functions are also probably shared by MHC-class Ib-restricted CD8 + T-cells, suggesting a role in protection for classical as well as non-classical CD8 + T-cells in TB.
Besides T-cells, new evidence also suggests accessory roles for NK-cells and B-cells in optimal immunity against TB [76]. Emerging evidence strongly suggests that the activation and infiltration of granulocytes is associated with TB disease activity and may play an important role in TB pathogenesis [77,78]. Recent genome-wide gene expression studies in TB patients revealed a striking neutrophil-associated expression pattern that was dominated by a type-1 interferon signaling signature. This profile normalized after curative treatment of infection, and had a core Mtb-specific component. Thus, other cells certainly contribute to host defense in TB, and may provide new biomarker signatures of pathogenesis, disease activity, and protective immunity.
Despite the more limited knowledge on the role of CD8 + T-cells in TB compared to CD4 + T-cells, paradoxically there seems to be greater consensus on ''multifunctionality'' as a marker of protection in TB for CD8 + T-cells than for CD4 + T-cells. Multior poly-functionality is a term used to describe the simultaneous production of multiple cytokines (mostly IFNc, IL-2, TNF) or the expression of multiple effector functions (perforin, granulysin, cytolysis, etc.) by one single T-cell clone. For example, the coexpression of IFNc and IL-2 by Mtb-specific CD8 + T-cells was associated with protective host defense following curative TB treatment [70,79]. Conversely, there is some controversy whether poly-functional CD4 + T-cells are associated with protection, or even with disease. Some reports, mostly derived from vaccination studies in animals, have suggested that poly-functional (IFNc + IL-2 + TNF + ) CD4 + T-cells are associated with protective immunity [80,81], but more recent studies have suggested that these cells may not indicate protection, and even mark TB disease activity [82][83][84]. This issue clearly needs further investigation. In this context, it needs to be realized that even during active TB disease not all lesions in the human lung are reactivating. At least some sites of infection are probably still controlled by the host immune response. Thus, much akin to recent new models that describe latent TB as a graded continuum of infection rather than as a single static entity [85], it is likely that TB disease follows a similar spectrum, ranging from the reactivation of only few lesions up to full-blown activation of many lesions such as seen in rapidly disseminating disease. It follows that in mild and less severe forms of TB, in which only part of lesions is reactive, others may be adequately controlled by the host response, which may involve multi-functional CD4 + T-cells. In any case, further studies to resolve these controversies are eagerly awaited.
Besides T-cells endowed with protective effector functions, Mtb also induces activation and expansion of regulatory T-cell (Treg) populations. These range from naturally occurring, broadly reactive CD4 + CD25 + FoxP3 + Tregs, to antigen-specific-induced CD4 + Tregs [86,87] and newly identified CD8 + Tregs, which have been mostly identified in the human response to mycobacteria [64,88,89]. These Treg have multiple inhibitory effects: they can directly suppress CD4 + Th-cell activity through secretion of IL-10, CCL4, or the expression of membrane-bound TGFb; they can deactivate APCs and thus inhibit optimal priming of CD4 + responses; and they can inhibit the influx of CD4 + T-cells into draining lymph nodes, and/or inhibit their proliferation and expansion [86], thus delaying and inhibiting onset of adaptive immunity to Mtb.
Thus a multitude of interlinking cells and mechanisms are involved in the immune response to Mtb infection, including activating/effector, and inhibitory/regulatory functions. The balance of the response and the ability to deliver effective hits to Mtb-infected target cells dictates whether the host response will lead to control of infection and possibly bacterial eradication, or whether infection will progress with increasing numbers of bacteria, lung tissue damage, and transmission of Mtb organisms to new susceptible hosts [44].

Designing Better TB Vaccines: Two Basic and Complementary Approaches
In the absence of reliable TB biomarkers of protection and a human challenge model, TB vaccine design by necessity will have to rely on empirical testing of candidate vaccines in long and costly efficacy trials. The identification of protective biomarker signatures would greatly facilitate vaccine discovery and testing [23,24]. If it were possible to identify a correlate with the same potency as neutralizing antibodies against viruses, discovery of high profile TB vaccine candidates and clinical testing could be markedly accelerated. As of 2009, more than a dozen TB vaccine candidates entered clinical trials, and many more are in the pre-clinical pipeline to be considered for testing in phase I clinical trials [3]. Five vaccines are currently in phase I/IIa (safety/immunogenicity) clinical trials and two of these are in larger phase II clinical trials. At least 16 vaccine candidates are in advanced pre-clinical development and well over 20 next-generation candidates are in the discovery pipeline.
The development of new TB vaccines follows two basic avenues [1,25,90]. The first aims at replacing BCG by either improved recombinant (r)BCG or by genetically attenuated Mtb. The profile of genetically improved rBCG should be: a) safer; b) more immunogenic; c) inducing longer lasting protection; and d) inducing protection against highly virulent clinical isolates such as Mtb Beijing strains, as well as MDR, XDR, and TDR Mtb strains. One way to improve BCG is either by introducing immunodominant Mtb-specific antigens that are absent from BCG, such as RD1 locus-encoded antigens (ESAT6, CFP10); or by over-expressing antigens that BCG already expresses by itself (cognates of Ag85 complex), but probably not sufficiently high throughout all phases of infection. Another way to improve BCG is by introducing genetic modifications for superior targeting of essential immune pathways, for example, by enhancing or facilitating cross-priming; and to inhibit its ability to neutralize phagosomal maturation. Obviously, both strategies can be combined [38]. Besides these two rBCG vaccine approaches, a second major strategy to develop better live vaccines against TB has been to attenuate Mtb. These strategies include the deletion of essential metabolic genes to create auxotrophic mutants; or the deletion of major virulence genes and their regulators (see below). In either case, at least two independent genetic loci need to be deleted to avoid reversion to virulence. One study engineered M. microti to express RD1 Mtb antigens, which indeed resulted in better experimental protection [91]. A recently published third type of live mycobacterial vaccine against TB is recombinant M. smegmatis. When a locus called esx-3 was removed from the M. smegmatis genome, a strongly enhanced innate immune response was seen. Unexpectedly, when this attenuated strain was complemented by the transfer of the Mtb esx-3 locus, strong bactericidal and protective activity against Mtb was observed in the mouse model, and in some cases even leading to sterile eradication [92].
The second major avenue to develop better TB vaccines relies on the development of subunit vaccines. These are non-live, or in the case of viral vectors, non-replicating vaccines, which can be delivered safely into the human host regardless of immunocompetence. Subunit vaccines in TB are mostly based on recombinant proteins admixed with proper adjuvants, or the use of attenuated viral vectors. Although subunit vaccines theoretically could be used as priming vaccines, current views are that they may be mostly used as booster vaccines on top of BCG-, recombinant BCG-, or attenuated Mtb-priming vaccines. The subunit vaccines are then anticipated to boost strong, long-lived immune responses in already primed individuals, such that these will persist to levels high enough to protect the 3%-10% vulnerable individuals against TB disease.
The various types of vaccines and their intended use are summarized in Table 1.

Live Vaccines, Aiming to Replace BCG as Priming Vaccines
The first attempts to create improved BCG vaccines were made by Horwitz et al., who over-expressed antigen 85B (Ag85B) in BCG, called rBCG30 [93]. This vaccine resulted in improved protection against TB in guinea pigs, and appeared to be immunogenic in humans [94]. This vaccine has successfully passed a phase I clinical safety trial but is currently on hold. Following a similar strategy, selected RD1 antigens or larger parts of the RD1 region were transferred into BCG [95]. The rBCG's profile indeed showed better immunogenicity, but this was compromised by its enhanced virulence in SCID mice. Thus, deleting the virulence determinants of this region may offer a viable strategy to improve immunity while maintaining safety. More recent studies engineered a recombinant BCG strain to over-express Ag85A, Ag85B, and TB10.4 [96]. While appearing to be safe and to induce high levels of IFNc-producing T-cells in mice, its protective efficacy against TB was not improved substantially over wild-type BCG.
Another approach to improve BCG has been the expression of listeriolysin, while deleting expression of ureC in BCG [97]. Listeriolysin is derived from Listeria monocytogenes, the organism in which it is required for phagosomal escape and translocation to the cytosol of the infected cell. Listeriolysin likely perforates the phagosomal membrane, allowing leakage of enzymes (such as cathepsins and other proteases) and bacterial components into the cytoplasm of the infected cell. Listeriolysin is only active at acidic pH 5.5 and is degraded immediately in the cytosol because this protein comprises a PEST sequence (i.e., a peptide sequence containing the amino acids proline [P], glutamic acid [E], serine [S], and threonine [T]), which induces its own degradation [98,99]. Hence, its activity is restricted to the phagosome. These features ensure safety of this construct. BCG normally expresses UreC, which is able to counteract phagosomal acidification; deletion of UreC from BCG prevents this, such that the resulting acidic pH of the BCG phagosome facilitates the activity of listeriolysin. The recombinant BCG lacking ureC and expressing listeriolysin caused superior protection as compared to parental BCG [97]. Interestingly, rBCGDUrec:Hly appeared to increase apoptosis of infected host cells, and the resulting apoptotic vesicles carrying vaccine components facilitated cross-priming by DCs, inducing strong CD4 + and CD8 + T-cell responses. In addition, this strain was found to activate Th17-cells, which next to the induced Th1-cells, probably facilitated optimal immunity against Mtb [59,100,101]. Importantly, this improved rBCG induced superior protection not only against laboratory strains of Mtb, but also against clinical Mtb isolates with high virulence, including Mtb Beijing family strains. The rBCGDureC:hly vaccine VPM1002 has now successfully passed a phase I clinical safety trial and has entered a phase IIa trial in newborns in 2011 (Table 1). Over-expression of perfringolysin in BCG had rather similar effects as listeriolysin [96]. However, perfringolysin has a relatively broad pH optimum and lacks the PEST sequence [38], such that it can act from inside and outside the cell and remains active in the cytosol. Indeed, a phase I clinical safety trial with rBCG-expressing perfringolysin and Ag85A, Ag85B, as well as antigen Rv3407 (instead of TB10.4) [96,102], had to be terminated because of side effects (reactivation of shingles [103]) (Aeras-422, see Table 1).
The second class of live Mtb vaccines comprises genetically attenuated Mtb derivatives. Attenuation can be achieved by inducing auxotrophy, or by deleting essential virulence genes. Auxotrophy relies on deletion of genes involved in the metabolism of essential nutrients for Mtb, such as amino acids or pantothenate. The first auxotrophic Mtb strains were generated as early as the mid-1990s [104]. More recently, promising candidates for clinical evaluation have been generated, including the Mtb DRD1DpanCD strain, which contains deletions in the major Mtb RD1 virulence gene cluster as well as in the biosynthetic pathway of pantothenate [105]. This strain induced efficient protection in normal and immunocompromised mice, and featured promising safety. Another promising attenuated Mtb derivative is the DPhoPDfad Mtb strain [106]. The two-component transcriptional regulator system PhoP/PhoR regulates the expression of a large number of genes, including Mtb virulence genes. The highly attenuated laboratory strain Mtb H37Ra was found to contain an inactivating mutation in the phoP gene [107]. The fad gene is involved in the synthesis of the cell wall lipid phthiocerol dimycocerosate (PDIM). This vaccine candidate is aimed at entering a phase I clinical safety trial in 2012.

Subunit Vaccines, Primarily Aiming to Boost (Recombinant) BCG-Induced Responses
Subunit vaccines against TB are delivered either through viral vector systems or as recombinant proteins mixed with adjuvants. Two viral delivery systems for TB antigens are currently in clinical trials: Modified vaccinia virus Ankara (MVA) expressing Ag85A has proved extremely immunogenic in both naïve as well as BCGprimed individuals, inducing high Ag85A-specific CD4 + T-cell responses in humans [108,109]. Replication-deficient adenoviral systems have either used Ad5 or Ad35 platforms. Ad35 carrying Mtb antigens Ag85A, Ag85B, and TB10.4 was highly immunogenic in humans, inducing strong CD4 + and CD8 + T-cell and IFNc responses [110,111]. Both the MVA over-expressing Ag85A and the Ad35 expressing Ag85A, Ag85B, and TB10.4 are currently in phase IIb clinical trials. The interference of preexisting human antibodies that could neutralize and negatively impact on adenoviral delivery needs further investigation, but may not be a major impediment to at least some adenoviral vectors [112]. New viral factors are currently being considered, such as replication-defective lymphocytic choriomeningitis virus vectors, which stimulate CD4 + and CD8 + T-cell as well as B-cell responses [113].
Significant progress has been made in the construction and testing of recombinant proteins, mostly fusion proteins that combine two or more immunodominant antigens of Mtb. These are invariably admixed with potent Th1-activating adjuvants. One of the most advanced products is the Hybrid 1 fusion protein, designed by Statens Serum Institute, which consists of Ag85B fused to ESAT6. When administered together with the potent IC31 adjuvant, which activates human Toll-like receptor (TLR)9 and facilitates antigen uptake by DCs, thereby providing a slowrelease depot [90], this product leads to strong CD4 + Th1 IFNc induction in humans. This was first reported in naïve human volunteers [114]. An encouraging observation was the long-lasting high-level immune response still detectable 2.5 years after the last vaccination. More recently, the same vaccine was found to boost immune responses previously induced by either BCG or latent Mtb infection [115]. The Hybrid 1 fusion protein in IC31 adjuvant is aimed at entering phase IIa trial in 2012. A very similar fusion protein, Hyvac 4, which consists of Ag85B fused to the antigen TB10.4 (which in contrast to ESAT6 is expressed also by BCG), is in a parallel clinical development program (Table 1).
A similar approach was followed by GlaxoSmithKline (GSK) that designed the M72 fusion protein, consisting of Mtb antigens Rv1196 and Rv0125 [116]. The M72 fusion protein was admixed with different GSK synthetic adjuvants and had a favorable clinical profile in terms of safety and immunogenicity [116]. The currently favored adjuvant is the AS01 E , which contains monophosphoryl lipid A, mixed with QS21; the first stimulates TLR4 and drives CD4 + Th1 as well as Th17 responses, and the latter primarily triggers CD8 + T-cell responses. This vaccine is currently undergoing phase IIa trial. Both AS01 E and IC31 are liposomal formulations. Currently other adjuvants are being constructed, such as CAF01 [117,118]. This liposome-based formulation combines dimethyldioctadecylammonium (DDA) with trehalose 6,69-dibehenate (TDB), a synthetic analogue of the mycobacterial cord factor trehalose 6,69-dimycolate (TDM), which ligates the innate immune receptor monocyte-inducible C-type lectin (Mincle) and has strong adjuvant properties [119].
Other antigens close to phase I clinical trial include heparinbinding hemagglutinin (HBHA) [120]. HBHA is expressed by both Mtb and BCG and is heavily methylated at its C-terminal portion, against which T-cell responses seem to be specifically directed [121]. The protein's methylation is essential for tethering to and invading human cells, including non-phagocytic cells [122]; this may relate to its ability to facilitate extrapulmonary dissemination of infection. Besides HBHA many other antigens have been discovered that could be interesting vaccine candidates. These include TB latency antigens such as DosR regulon-encoded proteins and starvation antigens like Rv2660, discussed below.

New Antigens
Until now virtually all TB vaccine strategies have focused on pre-exposure vaccination, which ideally induces strong, longlasting memory T-cell responses that are rapidly mobilized after infection with Mtb and mediate protection. Recent work has revealed that Mtb alters its metabolic state from active replication to slow or non-replication during infection. This is accompanied by significant changes in Mtb's gene expression profile. The resulting changes in protein expression also have an impact on which antigens are available to the immune system as a result of Mtb's alterations in antigen expression when entering into starvation or dormancy. Interestingly, this process is partly immune pressure-dependent [123]. The expression of ''early secreted antigens'' produced by replicating Mtb such as Ag85B decrease whereas the expression of ''latency antigens'' is induced or increased. Such Mtb latency antigens are, for example, encoded by (i) the Mtb DosR regulon, which mediates Mtb's response to hypoxia; (ii) starvation antigens that are upregulated by Mtb upon depletion of nutrients; and possibly also (iii) the broader enduring hypoxic response (EHR) genes, which include many genes of the DosR regulon [124][125][126][127][128][129].
Incorporation of starvation/latency antigens into further improved (live or subunit) vaccines may enhance the impact of these vaccines. The potential success of this approach was recently reported by Peter Andersen and colleagues, who showed that vaccination with a fusion protein linking the Hybrid 1 (Ag85B-ESAT6) backbone to Mtb starvation/latency antigen Rv2660c (termed Hybrid 56) induced superior protection against TB, particularly against late chronic infection, when compared to BCG or H1 vaccination. The Rv2660c gene had been selected because of its strong upregulation under starvation conditions [130]. Indeed, this antigen was recognized by T-cells from immune donors in a South African population [131]. H56 was also efficient in preventing TB reactivation in a non-human primate model, in which BCG priming was followed by fusion protein boosting. Even when anti-TNF mAb was used in an attempt to trigger TB reactivation, the vaccine-induced protection remained intact [132]. This vaccine is entering phase I clinical safety testing in 2012.
It is also possible to consider alternative strategies, for example, to include Mtb antigens that are expressed by early reactivating bacilli, such as resuscitation-promoting factor antigens or downstream antigens [75,102,126]. The availability of a pool of memory T-cells able to respond to these rapidly expressed reactivation antigens might provide a means of promptly detecting and controlling reactivated Mtb.

The Complex Milieu of the Vulnerable Human Host: Immune Modulation by External, Environmental, and Endogenous Factors
In real life, Mtb infection occurs in individuals who are often coinfected by environmental microbes or pathogens that can modulate immune responses, including responses to vaccination such as BCG vaccination. Non-tuberculous (environmental) mycobacteria (NTM) can significantly downregulate and alter the profile of BCG-induced immune responses [15,133]. Nutritional factors such as iron status and vitamin D, which influences innate immunity by inducing anti-microbial peptides, such as cathelicidin (LL-37) and autophagy in association with IFNc, also impact TB host immunity [134,135]. Different Mtb lineages or genetic differences in virulence between Mtb strains are also important in this context. Clinical programs will need to compare vaccination efficacy in different geographic and ethnic settings, and evaluate the impact of differences in Mtb epidemiology and strains or clades on vaccine-induced immune responses and protective efficacy [1].
Mtb itself is a master regulator of immune regulatory circuits. It induces strong CD4 + and CD8 + Treg activity that suppresses Th1 and Th17 responses (see above). Mtb exploits these regulatory mechanisms, which are normally used to prevent excessive inflammatory responses and accompanying tissue damage. In addition, Mtb downregulates key pathways of protection such as apoptosis, autophagy, and IFNc production as well as IFNcR signaling. Of note, highly discriminative biomarkers upregulated in T lymphocytes from TB patients as compared to latently infected individuals comprise genes involved in inhibitory signaling [136]. The likely synergistic effects of blocking multiple pathways at multiple different levels, added to the delay in its transport from the lung to the lymph node and the ensuing delay in priming Tcell immunity, collectively create an environment in which Mtb can establish a safe niche of infection from which it can seed the human body with a new wave of bacteria whenever immune pressure is alleviated, for example, during co-infection or malnutrition. Thus, the challenge for new TB vaccines is not only to induce strong memory immune responses with long-term efficacy in ''clean animal laboratory settings'', but also to do so in the face of the highly complex environmental, co-infection, and immunoregulatory background of the vulnerable human host.

Therapeutic Vaccines
New TB vaccines should not only induce immune responses that control existing or new infections at the latent stage, but ideally also eradicate Mtb organisms from the human body and achieve sterile eradication [25]. This would avoid the threat of TB reactivation from latency that can occur decades after initial infection. Such vaccines must also operate in the huge reservoir of latently infected individuals. Vaccines considered to be useful in such settings need to express antigens that are expressed by Mtb organisms during latent-stage infection. A major unknown is the precise mechanism that mediates protective immunity and possibly sterile eradication of Mtb. Recent results obtained with the above-discussed modified M. smegmatis vaccine are promising [92] with respect to the latter. However, in the absence of detailed knowledge, it would seem fortuitous to develop vaccines that stimulate all arms of the immune system, including Th1 cells; Th17 cells (that play a role in the early response, possibly through the recruitment of neutrophils); CD8 + T-cell populations with cytolytic and tuberculocidal activity (e.g., through the activity of perforin and granulysin; apoptosis; autophagy), as well as TCRcd cells and NK T cells, which bridge innate and adaptive immunity, and provide IL-17 and IFNc early in the response to infection [137]; and perhaps, lipid reactive T-cells, notably CD1-restricted T-cells that can express a myriad of relevant effector functions against Mtb.

A Role for B-Cells and Fc Receptors?
Vaccines that prevent productive Mtb infection are needed. Only very few bacilli are thought to reach the lung and establish the primary complex. Theoretically, the availability of mucosal antibodies at the site of infection could be valuable. It is interesting to consider that recent studies have documented a hitherto unappreciated role for B-cells in potentiating protective immunity against Mtb [76]. What is the potential role of such antibodies? First, they could target Mtb to macrophages for destruction, e.g., through FccR pathway activation. In this context it is interesting to note that FccR is among the most discriminative markers upregulated in latently infected individuals as compared to TB patients [78,138]. Second, they might exert chemotactic functions and attract additional immune cells such as monocytes, neutrophils, and NK cells to the site of infection. And third, by targeting Mtb to Fc receptors, antibodies might promote activation and antigen processing/presentation by DCs and macrophages and promote microbicidal activity in these cells [139,140].
Other TB vaccination strategies should address the targeting of the mucosal immune system. Much research is focusing on the potential of mucosal immunity and its enhancement by local vaccine delivery. Recent studies documented that pulmonary aerosol delivery of BCG or subunit vaccines in animal models was highly effective in inducing rapid immunity at the site of infection [141][142][143][144]. Exploring new routes of vaccination and targeting unique mucosal immune response pathways, which include mucosal cells such as intraepithelial cells, MAIT, and T-cells expressing homing receptors for the lung, such as CXCR6 [145], have been relatively unexplored in TB and may have high potential. Even though such cells are unlikely to prevent TB infection stricto sensu, their rapid early response is likely to contribute to early control of infection, particularly because these cells may respond to infection during a critical period in which Mtb delays onset of adaptive immunity as discussed above.

Concluding Remarks
In striking contrast to only 10 years ago, the TB vaccine candidate pipeline is now filled with a number of high-profile candidates, some of which have already entered or are soon likely to enter clinical trials. First immunogenicity profiles of vaccines tested in humans look encouraging. Because we do not know the mechanisms of protection against TB and are lacking surrogate end-point biomarkers of protection, we cannot predict whether these vaccines will have protective efficacy in humans. Nevertheless, many of these vaccine candidates have proven records in animal models, ranging from mice to guinea pigs to non-human primates, raising hopes they may also be protective in humans. Vaccines may be used either as stand-alone vaccines or-more likely-in heterologous prime-boost settings.
The next 10-15 years will be critical for TB vaccines to demonstrate protective efficacy as well as safety. At the same time, it must be realized that these first-generation vaccines are neither designed to prevent infection nor to achieve sterile eradication, but rather to better prime and/or boost infection control. Thus at best, such vaccines will only inhibit or delay TB reactivation. Nextgeneration vaccines should be designed aiming at preventing infection or achieving sterile eradication, perhaps by redirecting immune responses at Mtb antigens expressed during latency [124][125][126]130] or by using new platforms [92]. The new vaccines, especially subunit vaccines, most likely will need to be administered multiple times in order to maintain adequate life-long immune memory. Virtually nothing is known about the timing of such boosting schedules when it comes to vaccines aimed at inducing primarily cellular immunity. Finally, the impact of coinfection, immune regulation, and environmental factors (nutrition; type-2 diabetes) that enhance susceptibility to developing TB disease remain ill-defined, and need further research.
In summary, there is reason for optimism in the field of TB vaccines: there is a substantial portfolio of interesting TB vaccine candidates in clinical phase-I/II testing, some of which are already fairly advanced in the TB vaccine pipeline. This is complemented by numerous promising candidates further upstream in the pipeline. In fact, an emerging bottleneck may not be the number of pre-clinical TB vaccine candidates that TB researchers can produce, but rather the number of vaccines that can be tested clinically in efficacy trials, given the limited clinical trial capacity worldwide, i.e., a shortage that exists not only in Africa but also in Asia [1,3]. Finally, the identification of TB surrogate end-point biomarkers or ''correlates of protection'' may drastically reduce the need for the current long-term large-scale clinical trials, and thus will speed up TB vaccine discovery and clinical testing [1,9].