Albert Calmette and Camille Guèrin developed BCG between 1908 and 1921 by successfully passing an isolated strain of M. bovis in vitro. In the first half of the 20^th Century, BCG vaccines were prepared and preserved by different manufacturing laboratories. This resulted in genotypic and phenotypic differences in the daughter strains with variations in tuberculin conversion and the frequency of adverse reactions. During this time, BCG showed a progressive decrease in virulence, the most important during the first 15 passes 78. Over the next 40 years, until the freeze-dried process of the Pasteur strain in 1961 and after more than 1,000 passes, it was clear that the standardization and stability of vaccine strains needed to be enforced. This led to the adoption of the seed-lot system by the International BCG Technical Conference in 1956 8. Furthermore, Colditz, et al 3 independently quantified the detrimental or protective benefit of the vaccine in several prospective trials from case control studies and concluded that BCG significantly reduced the risk of active TB development by an average of 50%. The different criteria applied to interpret the confounding variables in the human trials make it difficult to identify the impact of BCG vaccines on protective efficacy 4.

In 1999, Behr et al 9 elucidate some of the molecular events that occur during BCG attenuation. Under laboratory conditions used for bacterial passes, and using microarray technology and genome sequence of M. tuberculosis H37Rv as a framework, they inferred the genealogy of BCG strains in an evolutionary approach, which was supported by BCG historical records 10. These studies only led to the understanding of the genes missing from M. bovis leaving deeper research to be carried out to explore the association between the 61 ORF genes deleted from BCG during evolution and virulence attenuation, and may be more importantly, research into the loss in protective efficacy over the last eight decades 11. What remains unknown is the irrefutable role that the smallest deletions, duplications and polymorphisms in the nucleotide sequence play to induce protective efficacy, as well as the production of antigens in the overall attenuation of BCG 912.

Following adoption of the seed-lot system, several BCG sub-strains of the original vaccine strain are now in circulation. It is known that these sub-strains vary in certain characteristics (Table 1), such as their antigenic structure, secreted protein profiles 131415, IS6110 copy number 11, mycolic acid pattern 16, and the characteristics and quantity of exoquelins 17. Alterations in the mycolic acid subclass directly affect not only cell wall fluidity and permeability to hydrophobic agents, but also the ability of strains to grow within macrophages. In addition, strains with an altered cord factor have altered virulence 1618.

Today, the most commonly used BCG sub-strains are the Pasteur, Japanese and the Glaxo (Figure 1) but other daughter strains are/or were in use (Moreau, Montreal, Russian, Prague, Danish, Birkhaugh, Australian) 12.

Recently, Lagranderie et al 20 evaluated the capacity of five BCG sub-strains to trigger and maintain the immune response in mice and found statistically significant differences in their protective efficacy. However, the data are subject to the particular experimental conditions used in their mice model. In reality, vaccine candidates are tested in several models and under different evaluation criteria. It is clear that the strategies for the development and evaluation of the new vaccine candidates need to be standardized under shared criteria in order to understand their ability to induce protection 21 in animal models that mimic the key aspects of naturally occurring human tuberculosis 22. Currently, work is being carried out to understand the differences in the major BCG vaccines being used in human trials. In these studies, BCG vaccines are being studied in a well-characterized mouse model of pulmonary tuberculosis, which tries to mimic natural human infection as much as possible 23. Under these conditions 10 different daughter strains shown to induce a wide range of protection against M. tuberculosis intratracheal infection displaying differences in terms of colony forming units (CFU's) reduction, delayed type hypersensitivity response and proportion lung surface affected by pneumonia going to be difficult correlate the protective response offered by the current disposable vaccines with the protective efficacy determined in previous human clinical trials (Figure 2 shown representative lung section of mice unvaccinated or vaccinated with these daughter strains) [unpublished data].

The availability of the complete genome sequence of two M. tuberculosis strains and the partial sequence of the BCG Pasteur sub-strain, may lead to the understanding of the principal source of antigenic variations, which would be significant for vaccine design and for inducing protective immunity in tuberculosis 242526. As seen in BCG sub-strains, variation in colony morphology among isolates of M. tuberculosis from patients is extremely common and multiple phenotypes are often apparent even from a microbial culture from a single sputum sample. However, attempts to correlate such differences with pathogenic potential cannot be sustained.

The development of a new vaccine with improved protective immunity against M. tuberculosis depends on the efficient recruitment of antigen-specific T-cells principally CD4+ in the lungs, as well as on the cytokines that are released particularly IFN-γ, which is important for inducing the macrophage killing activation mechanism 27.

Cell-mediated immunity plays the principal role in containing infection, and the routes of vaccine administration and immunization influences immune response development. In infants, BCG vaccination generally induces a Th1 cytokine response and stimulates cytotoxic T-lymphocyte activity in neonates 28. An alternative route of BCG administration, which does not induce the side effects associated with subcutaneous immunization, is via rectal delivery. This method induces a similar immune response and protection in several animal models 2930 without altering the recruitment patterns of activated T-cells 31. Intranasal immunization induces higher protection by rapid induction of IFN-γ and T-cell response in the lung tissue but there are some who have serious misgivings in using live bacilli. However, nasal administration of recombinant BCG as a means to deliver immune dominant antigens to the mucosa shows some promise for therapeutic use in HIV infection 3233.

M. tuberculosis has a complex multiplicity of antigens with a diverse chemical and immune reactive nature as has been demonstrated in lipids, polysaccharides, and proteins. Some of these induce the granuloma formation, macrophage activation, and adjuvant activity while others are immunosuppressive and enhance host toxicity 2734. It should be remembered that the T-cell mediated immune response can result in protective immunity but it may also lead to pathological immunity that is detrimental to the host. The understanding of specific antigens expressed during early infection, disease, latency or reactivation and their immunological characterization are key for the development of new vaccines.

Looking back in the history of tuberculosis research, during the second half of 20^th Century, various types of vaccines have been developed against mycobacterial infection to control bacilli replication and dissemination in order to stop tuberculosis transmission and future eradication. This vaccine development has been included the use of secreted proteins and surface exposed proteins, nucleic acid vaccination, rational attenuation of Mycobacterium sp strains, and recombinant BCG vaccines, all of which are currently being tested in a wide variety of animal models before future use in human clinical trials 25353637.

Now a day M. bovis BCG vaccine is the only live bacterial vaccine in use, which shows no major side effects. This is the most used vaccine and has been administered to more than two billon people worldwide and has showed long-lasting immunity.

Until now, medical staffs and health care workers accepted that BCG strains have been able to induce protection in the first year of life as well in all animal models that have been used; something that other vaccine candidates have still failed to achieve. Recent data of protective efficacy coming from more than 180 vaccine candidates put in perspective the evaluation of existing BCG vaccines in a way to identify potential candidate for a rationally designed recombinant BCG vaccine or an auxotroph vaccine by comparing protective efficacy 21. These new designed recombinant vaccines have been shown to induce Th1 and long-lasting protective response with a smaller number and size of granulomas and a reduced load of colony forming units.

The development of new acellular vaccines composed of one or more antigens may offer a faster route for creating an alternative to BCG vaccines. However, this approach needs a distasting research to identify and defining those potential proteins recognized by human T-cells which are capable of inducing cell-mediated and long-lasting immunity. On last ten years attention has mainly been focused on the group of proteins secreted by the bacteria in culture media during the active replication of M. tuberculosis in early growth phase. These have been recognized in the early stages of tuberculosis infection in several animal models and may be the reason that viable mycobacteria are needed for effective vaccination against M. tuberculosis infection 2535.

Various approaches are currently being analyzed using sub-unit vaccines based on culture filtrate proteins of M. tuberculosis, which when are administrated with an adjuvant have been shown to induce protective immunity in mice, guinea pigs and non-human primate models 40. Antigen isolation has depended on the use of monoclonal antibodies 39 or by direct chromatography of culture filtrates. Such isolation procedures are characterized by biochemical and immunological techniques, with the corresponding genes being sequenced and cloned for further use as a vaccine. Several antigens have been cloned and purified as recombinant proteins, due to the large-doses required for immunization. The choice of an antigen or several antigens to formulate optimal immunological cocktails for vaccine use needs to be tested through in vitro and in vivo studies. In in vitro human studies, ESAT-6, CFP-10, MPT64, MPB70 and fusion proteins of the ESAT-6 and antigen 85B have been tested to show their ability to induce protective T-cell response. The most used criteria for tuberculosis vaccine designs are antigen-specific proliferation and IFN-γ secretion assays 35404142. ESAT-6 was a promises in this area because is an immune dominant antigens most frequently recognized by TB patients, which contains a great number of T-cell epitopes. In animal models, ESAT-6 has been recognized in the first phase of infection and has been demonstrated to be a strong T-cell immunogen inducing prime memory immunity, which persisted in individuals who had recovered from disease 2543. The 30/32-kDa complex is among the most important secretor proteins of M. tuberculosis and has been shown to be protective in the guinea pig model of pulmonary tuberculosis. Their abundant production either extracellulary in broth culture or intracellulary in human monocytes, suggests a vital role in the physiology of the bacterium.

The recent elucidation of complete genome sequence of two M. tuberculosis strains and accelerated development of genetic tools for the study of M. tuberculosis has enabled the development of DNA vaccination, which has been tested in several animal models 4445. Over the last few years, genetic immunization has become the most popular strategy in vaccine development and may be a successful alternative for the delivery of M. tuberculosis antigens that drive the cellular immune response. The DNA vaccines that expressed the 85B, MPT64 and ESAT-6 antigens has been demonstrated that they reduced the level of pulmonary infection following M. tuberculosis challenge, by the specific CD4+ and CD8+ T-cell response 444647. It was also shown that combined vaccination with three antigens was more effective than vaccination with a singe vector. The use of cytokines to increase vaccine protection enhanced the specificity of the immune response triggered against these antigens, thereby potentially maintaining the TH1 response 46. These results concur with previous studies concerning sub-unit vaccines that have shown them to be only equal at best to BCG vaccines. In these way, immunization using a DNA vaccine of the 32-kDa protein stimulated protective immunity against BCG and M. tuberculosis infection 4748. Plasmid vectors expressing the 38-kDa glycoprotein induce a comparable level of protection with other DNA vaccines and develop a strong antigens-specific Th1 response in human and murine lymphocytes, characterized by IFN-γ secretion 45. However, genetic vaccination with the 19-kDa lipoprotein resulted in a non-protective antibody response 49. Even though 38 kDA and 19 kDa are important antigens that induce a Th1 type immune response, they do not protect mice from M. tuberculosis infection. Similarly, no protection was observed when immunization used chaperon-like proteins, in contrast with early works in which the heat-shock proteins like HSP60 and HSP70 resulted in a protective effect. Similar results to those using 38 kDa and 19 kDa have been obtained by immunizing with MPB83 DNA vaccines 5051.

The major advance in DNA immunization is the notion that intramuscular vaccination is better for inducing a protective Th 1 type immune response against tuberculosis without sensitizing the animal to tuberculin testing. An alternative to DNA immunization is the use of recombinant vaccine viruses that express immune dominant antigens 52.

Although DNA vaccines are an attractive alternative for the development of a new vaccine, they do not surpass the protection conferred by BCG vaccination 53 and need further safety evaluation prior to testing in human populations.

The stable expression of foreign DNA in BCG on a plasmid vector established a basis for the construction of polyvalent recombinant BCG vaccine but the antibiotic resistance markers are not appropriate for the selection or maintenance of recombinant plasmid containing antigens. An alternative would be to develop auxotroph mutants that would maintain stability, as they offer a genetic function required for the survival of the mycobacteria in their host. For this purpose defining mycobacterial promoter sequences is critical for expressing high levels of selected antigens responsible for cell-mediated immunity.

Recombinant BCG vaccines will have the excellent adjuvant activity of BCG and thus, significant levels of T-cell reactivity but it is uncertain whether they will be capable of generating high levels of cytotoxic T-cell response. The use of BCG as a live vaccine vector 60 for the presentation of heterologous and homologous antigens can be a powerful method for driving the immune response to the Th1 phenotype or for replacement of BCG target genes by homologous recombination 55.

Auxotroph and avirulent mutants of M. tuberculosis and M. bovis 56575859 have been developed showing wide variability in their avirulent character giving equivalent protection between them and comparable to that of BCG, which have also proved safe in individuals with immunodeficiency disease.

With the complete genomic sequence and database comparisons, it is now possible to attribute tentative functions to roughly 40% of the 3924 protein coding genes for the genome of M. tuberculosis 24. In addition, this information reveals new members of family proteins and potential variation between M. tuberculosis, M. bovis and M. bovis BCG 60. Until recently, there have been two principal ways to learn more about the role of new antigens in the immune pathology of tuberculosis. All primary results come from some biochemical or immune proliferative assays induced by individual proteins with further characterization under specific conditions in animal models. Studies using such tools have resulted in the identification of immune dominant antigens, which are strongly recognized in humans and with the potential for developing of a novel TB vaccine. An applied approach allowed the identification of deleted genes and the development of antigens that can distinguish between M. tuberculosis infection and BCG vaccination 61. This offers the prospect of starting vaccine design using the genetic information by reverse vaccinology 26.

Comparative analysis of the complete genome sequence of M. tuberculosis, M. bovis, BCG strains and M. leprae and other mycobacterial species may allow the identification and development of subunit vaccines with an exceptional specificity, however the high similarity of these strains limit this potential window of antigens.

The recent development of recombinant BCG vaccines that secrete cytokines or specific antigens is one of the most secure routes in the development of improved vaccines against tuberculosis 3662. However, as seen in the sub-unit vaccine approach, adverse results from antigens which are capable of enhancing protection efficacy of the current BCG vaccine could be found. Using recombinant BCG that secretes, IFN-γ or IL-2, which have been shown to reduce disease in a pre-infection vaccine model, as well as TNF-alpha which is indispensable for the formation of tuberculous granulomas, leading to an increase in lung pathology without reducing bacillary loads 4863.

Other recombinant live vaccines that produce pore-forming cytolysin 64, Interferon alpha 2B 62 demonstrate enhanced immune genecity but not improved protection. Recently, Horwitz 65 reported enhanced protective response to challenges with virulent M. tuberculosis following vaccination with recombinant BCG that secreted 30-Kda antigen of M. tuberculosis.

The use of closely related species of Mycobacterium may be the second most interesting route to follow for vaccine improvement. Mycobacterium microti is a naturally attenuated strain with a narrow host range. In mice and rabbit models, vaccination with this strain resulted in significant reduction in the M. tuberculosis load and histopathology lesions 66. Another approach is the use of different host vectors that express shared antigens 6467 such as Salmonella thyphimurium that secretes ESAT-6, which is used as a vaccine and reduces the load of mycobacteria throughout the course of infection. M. vaccae is a saprophytic bacillus which, as other members of the Mycobacterium genera, share the major antigens and this was used recently as alternative immunotherapy in addition to chemotherapy in MDRTB cases 68.

Finally, little is known about the effect of vaccine boosters using some antigens and how this strategy may be able to drive the response to maintain a prorogued Th1-type memory T-cell response].

The use of atypical mycobacteria as hosts for the production of M. tuberculosis antigens needs to be designed with the same precautions as for BCG. Post et al 69, demonstrated the inefficacy of recombinant strains of M. vaccae and M. smegmatis that expressed the 19-kDa lipoprotein, and the abrogation of the protection conferred by host strains alone 49. On the other hand, the deletion of these genes in M. bovis has a significant effect on the ability of BCG to protect against M. tuberculosis challenge 70.

The rational mutation of virulence genes or putative virulence factors is being evaluated for their biological function 71. In order to develop better vaccines, several mutants have been produced by transposon mutagenesis or illegitimate recombination. The current strategy is to use antibiotics to select mutant strains that better express antigens than induce protection. However, the use of antibiotics in this way and the possibility to revert their virulence or recombination events between these and the wild type strains means that this alternative is a long way from human trials 37. The exciting aspect of this early research is the specificity of the response elicited by each one of these strains, which plays a key role in protection 59 through homologous or analogous challenge with the virulent strain.