Tuberculosis (TB) is a leading cause of infectious death worldwide. TB is caused by the bacterium Mycobacterium tuberculosis (Mtb) and transmission happens via inhalation of droplets and aerosols coughed by an individual with active disease. There are about 10 million cases of active TB annually, and it is estimated that approximately 2 billion people are latently infected. Although most latently infected individuals are not sick and do not transmit the disease, in about 5%–10% of these people the disease reactivates. TB kills about 1.5 million people each year and resistance to current treatments increases steadily. Prevention of TB via a vaccine is considered the optimal solution by the World Health Organization. Presently, the only licensed vaccine against TB, Bacillus Calmette-Guérin (BCG), prevents childhood versions of TB but affords low efficiency in the prevention and transmission of the disease in adults. Hence, an improved vaccine is urgently required to combat TB.
In May 1796, the English doctor Edward Jenner intentionally transferred cowpox material from the hand of a milkmaid to an 8-year-old boy. Jenner, who reasoned that milkmaids’ frequent exposure to cowpox protected them from smallpox, wanted to prove his theory and successfully prevented the boy from developing smallpox disease following a deliberate infection. This moment commonly marks the discovery of vaccination.
However, the concept of cross-protection between cowpox and smallpox was already established in many farming communities before this pivotal moment. The farmer Benjamin Jesty had protected his wife and two sons from a smallpox epidemic in 1774 by using cowpox lesions from the udder of an infected cow and similar practices were known in India for centuries. Although it was still unknown at that time that infectious diseases are caused by microorganisms, Jenner’s and Jesty’s experiments provided the foundations for the development of live vaccines.
The germ theory (that pathogens cause disease) was proven towards the end of the 19th century. From this it became increasingly clear that some pathogens only cause disease in animals and are non-virulent or only cause a mild disease in humans (such as the virus causing cowpox). It also emerged that pathogens can be artificially weakened (attenuated) to no longer cause disease in humans, but to still induce a protective immune response. Based on these discoveries, in the early 20th century, the French scientists Albert Calmette and Camille Guérin started a series of experiments that would culminate in the development of a live attenuated vaccine against tuberculosis (TB).
Similar to smallpox, TB has plagued mankind for millennia and over the last 200 years, more than 1 billion people have died from TB, far more than from any other infectious disease. In 1882, Robert Koch succeeded in identifying the bacterium Mycobacterium tuberculosis (Mtb) as the causative agent for TB and in 1901 he showed that TB in cattle is caused by Mycobacterium bovis, a related but different species of mycobacterium. Inspired by Jenner’s experiments with cowpox, Calmette and Guérin started their experiments with M. bovis in 1908. Because M. bovis can cause TB-like disease in humans, they passaged the bacteria more than 200 times and tested each generation on animals including guinea pigs, rabbits and cows to monitor virulence. Eleven years later, their cultures contained M. bovis bacteria that no longer caused disease in many animal models, including cows. Two years after that, the experimental vaccine, termed Bacillus Calmette-Guérin (BCG), was deemed ready for humans. On July 18, 1921, doctors in Paris gave BCG to an infant whose mother had died of TB only hours after giving birth. Over the following years the safety of BCG was further confirmed and it was rapidly adopted in many countries as a vaccine for infants.
To date, BCG remains the only licensed and widely used vaccine against TB, and 2021 marks its 100th anniversary. In most TB-endemic regions of the world, BCG is still given to infants shortly after birth. The vaccination prevents childhood versions of TB and, hence, saves thousands of children’s lives annually. However, the efficacy of BCG wanes over time, and protection against TB is often lost by young adulthood.
Because of the rapid decline of TB in many western countries after BCG was introduced and a general improvement of hygiene and living conditions, research to find a more effective TB vaccine largely stopped for most of the 20th century. As such, TB is now largely confined to low- and middle-income countries.
Towards the end of the 20th century it became clear that the development of an improved TB vaccine is critical in order to reach the World Health Organization goal of TB elimination. Renewed research efforts and better technology, including genomic sequencing and gene technology, identified that BCG lacks multiple genomic regions compared to the ancestral virulent M. bovis and to Mtb. Among others, the loss of the genetic region of difference 1 (RD1) was identified as a major attenuation event. RD1 encodes a unique mycobacterial type VII secretion system (ESX-1), a type of molecular syringe, required to inject effector molecules into the host cell and to allow escape from the phagosome. Due to the lack of ESX-1, BCG cannot escape to the cytosol and remains restricted to the phagosome (Figure 1). While these scientific breakthroughs have shed light on the limitations of BCG that occurred during the sequential passaging of M. bovis by Calmette and Guérin, they also opened up new avenues for improvement of BCG and the development of other live attenuated vaccines for TB.
Live attenuated vaccine candidates for TB
Over the last 20 years, approximately only 15 new TB vaccine candidates have entered clinical trials. The majority of those are subunit or viral-vectored vaccines designed to replace BCG or to boost the immune response after a primary BCG vaccination (Figure 2). Unfortunately, many of the most advanced subunit and viral-vectored TB vaccine candidates have not shown superiority over BCG. In fact, some of these clinical trials have identified that BCG re-vaccination provides better outcomes than the newly developed vaccine candidates. Consequently, there is renewed interest in live attenuated vaccines for TB and several new live vaccine candidates have been developed over the last decade. However, the enormous costs of bringing a vaccine to market (often more than US $1bn) and the limited interest from pharmaceutical companies to invest into a disease that is largely confined to low-income countries meant that very few of those candidates have progressed to clinical trials. Live attenuated TB vaccine candidates can broadly be classified into recombinant BCG (rBCG) strains, attenuated Mtb strains and other bacterial species (Figure 3).
Recombinant BCG strains
BCG improvement is based on the idea that inclusion of additional genetic elements and/or deletion of particular pathways will increase immunogenicity and efficacy of the rBCG strain without increasing virulence. Currently, only one rBCG strain is in advanced clinical trials. VPM1002, also known as BCG ΔureC::hly, expresses listeriolysin O, a pore-forming and haemolytic protein from the bacterium Listeria monocytogenes. To facilitate phagosome acidification and optimal function of listeriolysin O in BCG the ureC gene encoding urease C has been deleted. The combined deletion of ureC and the insertion of listeriolysin O allows access of bacterial DNA and proteins to the cytosol of immune cells, which in turn is associated with increased activation of cytosolic immune responses. VPM1002 is more effective than BCG in pre-clinical animal models, has passed phase I and II safety and immunogenicity trials in adults and infants and is currently undergoing phase III clinical trials.
A number of next-generation derivatives of VPM1002 have also been generated, including strains that (i) are auxotroph for vitamin B6 (BCG ΔureC::hly Δpdx1); (ii) show increased autophagy and apoptosis (BCG ΔureC::hly ΔnuoG); or (iii) secrete human cytokines (BCG ΔureC::hly_hIL7/18). While some of these strains have shown further increased efficacy in animal models, none of them have entered clinical trials.
Another promising rBCG strain is BCG Δzmp1. Zmp1 encodes the zinc metalloprotease 1, a protease utilized by virulent mycobacteria to arrest phagosome maturation. Deletion of zmp1 is associated with increased phagolysosome delivery and enhanced immunogenicity, safety and protection in various animal models, and BCG Δzmp1 is undergoing early trials.
A further concept of BCG improvement is based on the partial reversion of attenuation. When the loss of the ESX-1 secretion system of M. bovis was identified as a major reason for the attenuation of BCG, it was hypothesized that the incorporation of the ESX-1 secretion system of Mtb could increase immunogenicity and efficacy of BCG. The resulting BCG::ESX-1Mtb strain, indeed, provided superior immunity and protection against Mtb challenge in mice and guinea pigs but has been deemed unsafe as a human vaccine, due to increased persistence and virulence in immunocompromised hosts. More recently, a rBCG strain expressing the ESX-1 system of Mycobacterium marinum (BCG::ESX-1Mmar) has been developed as a potential alternative. In addition, my own research focuses on guiding ESX-1-dependent Mtb proteins, such as ESAT-6, to the related ESX-5 secretion system of BCG to overcome the detrimental effects of ESX-1Mtb inclusion. This work has culminated in the development of the rBCG strain BCG::ESAT6-PE25SS. Both BCG::ESX-1Mmar and BCG::ESAT6-PE25SS combine improved protection and immunogenicity with low virulence in animal models and await further evaluation.
In addition, several other rBCG strains have shown some advantage over BCG in small animal models, but have not reached clinical trials (Figure 3). To name a few, these include strains with increased biofilm formation (BCGΔBCG1419c), secretion of cytokines (e.g., rBCG-IL2, rBCG-IFNγ, rBCG-IL18) or Mtb-derived proteins (e.g., (H)PE-ΔMPT64-BCG, rBCG E6, rBCG30), expression of latency genes (e.g., rBCG:XB) or genetic pathogen resistance loci (e.g., BCGi) or a combination of the above (e.g., AERAS-422, rBCG-Ag85B-IL15).
Attenuated Mtb strains
Vaccine candidates that are based on Mtb require at least two attenuating mutations to guarantee safety and to prevent reversion to the virulent ancestral phenotype. In most cases this is achieved by deletion of genes in distinct metabolic pathways, such as amino acid synthesis, cofactors and/or transcription factors. Attenuated Mtb strains usually are non-virulent in animal models but provide increased immunogenicity and protection compared to BCG due to the larger antigenic repertoire.
Only one attenuated Mtb strain is currently in advanced clinical trials. MTBVAC (Mtb ΔphoP ΔfadD26) carries deletions in the transcription factor phoP and the lipid biosynthesis gene fadD26. While phoP is involved in intracellular survival and ESAT-6 secretion, fadD26 plays an important role in the synthesis of virulence-associated cell wall lipids. MTBVAC is as safe as BCG and has successfully passed several clinical trial stages. Other attenuated Mtb strains that have shown superior protection in pre-clinical animal studies include Mtb ΔRD1 ΔpanCD, Mtb ΔleuD ΔpanCD, Mtb ΔlysA ΔpanCD, Mtb ΔlysA ΔsecA2, Mtb ΔargB/F, Mtb ΔLprG, Mtb ΔsigH and Mtb ΔsigE ΔfadD26, to name just a few (Figure 3).
However, similar to rBCG strains containing ESX-1 or its secreted proteins, attenuated Mtb strains suffer from an obstacle – that is cross-reactivity with the gold-standard diagnostic tool for TB, the Interferon Gamma Release Assay (IGRA). IGRAs detect specific immune responses to proteins that are found in Mtb but not in BCG, including ESX-1-dependent proteins (Figure 1). Because attenuated Mtb strains, in most cases, still contain these proteins, vaccination with attenuated Mtb strains can lead to positive IGRA results and hence difficulty to distinguish vaccine-mediated immune responses from immune responses directed against Mtb infection. These challenges are currently addressed by the development of new diagnostic tools.
Given that BCG was derived from M. bovis, targeted attenuation of M. bovis has also been trialled, but the resulting strain, MbΔp27-p55, is more virulent than BCG. Using non-pathogenic mycobacterial species and other bacteria as a live attenuated vaccine for TB is also being considered. For example, the fast growing mycobacterium M. smegmatis has been used to generate the TB vaccine candidate IKEPLUS. Based on the replacement of the M. smegmatis esx3 locus with the esx3 ortholog of Mtb, IKEPLUS provides similar protection to BCG in mice. Furthermore, M. microti, an animal-adapted mycobacterium, showed strong attenuation in immunocompromised mice and demonstrated similar protective efficacy as BCG. Modified strains of Listeria ivanovii, Listeria monocytogenes and recombinant attenuated Salmonella typhimurium vaccines (RASV) have also been used to deliver Mtb-derived antigens (Figure 3). In most cases improved or equivalent protection to BCG was observed in animal models, but none of these vaccine candidates has progressed to clinical trials.
The year 2021 marks a special occasion in the history of live attenuated TB vaccines. Hundred years after its first use, BCG remains the gold standard for TB prevention in children. While this anniversary is an important milestone, TB remains a global threat and the development of a new and effective vaccine is crucial if TB is to be significantly reduced, let alone eradicated, within the coming decades. With the largely disappointing results of subunit vaccine candidates in recent clinical trials, the various live attenuated vaccine candidates described earlier provide a glimmer of hope that an improved TB vaccine will be found. Importantly, despite being 100 years old, the mechanism by which BCG actually works is still largely unknown, and it is unclear why BCG only confers protection against childhood versions of TB and why protection often wanes in adolescence. If further research leads to a better understanding of the correlates of BCG-mediated protection, a more targeted improvement of BCG will likely be possible.
World Health Organization. Global Tuberculosis Report 2020. http://www.who.int/tb/publications/global_report/en/ [Accessed 1 July 2020]
Gengenbacher, M., Nieuwenhuizen, N.E. and Kaufmann, S.H.E. (2017) BCG - old workhorse, newskills. Curr. Opin. Immunol.47, 8–16. DOI: 10.1016/j.coi.2017.06.007
Nieuwenhuizen, N.E. and Kaufmann, S.H.E. (2018) Next-generation vaccines based on Bacille Calmette-Guerin. Front. Immunol.9, 121. DOI: 10.3389/fimmu.2018.00121
Gröschel, M.I., Sayes, F., Simeone, R. et al. (2016) ESX secretion systems: mycobacterial evolution to counter host immunity. Nat. Rev. Microbiol. 14, 677–691. DOI: 10.1038/nrmicro.2016.131
Ginsberg, A.M. (2019) Designing tuberculosis vaccine efficacy trials - lessons from recent studies. Expert Rev. Vaccines.18, 1–10. DOI: 10.1080/14760584.2019.1593143
McShane, H. (2019) Insights and challenges in tuberculosis vaccine development. Lancet Respir. Med.7, 810–819. DOI: 10.1016/2213-2600(19)30274-7
Andreas Kupz is a group leader at the Australian Institute of Tropical Health and Medicine at James Cook University. His interests include microbiology, immunology and infectious disease epidemiology with specific focuses on basic immunology and vaccine development for tuberculosis. He is the co-chair of the Live Attenuated Vaccines research community at the Collaboration for Tuberculosis Vaccine Discovery hosted by the Bill and Melinda Gates Foundation. Email: firstname.lastname@example.org