BCG is an attenuated form of the Mycobacterium bovis (M. bovis) bacteria, a close relative of Mycobacterium tuberculosis (M. tb), and is currently used to vaccinate infants against TB meningitis and miliary disease (Mahairas et al., 1996; Zimmermann et al., 2019). According to the recently updated BCG Atlas, BCG is the most universally given vaccine across the globe with over 160 million infants receiving the vaccine annually (Zimmermann et al., 2019), although long-term efficacy may vary globally (Angelidou et al., 2020). The benefits of BCG vaccination include its heat stability, easy mass production, low cost, and safe neonatal use, it is unaffected by maternal antibodies, and it is associated with induction of non-specific heterologous immunomodulatory effects (Kilpeläinen et al., 2019; Zimmermann et al., 2019). Indeed, BCG immunization at birth induces cross-protection for a number of infectious pathogens including M. tb, Candida albicans, Staphylococcus aureus, respiratory syncytial virus (RSV), influenza A virus, and herpes simplex virus type 2, due to its non-specific immunomodulatory, or adjuvant-like, effects (Starr et al., 1976; Spencer and Ganguly, 1977; Van’t Wout et al., 1992; Kleinnijenhuis et al., 2012; Nemes et al., 2018; Covián et al., 2019; O’Neill and Netea, 2020). While an in-depth discussion of the various immune-related mechanisms associated with BCG is beyond the scope of the present article, in recent years, it has been demonstrated that BCG is able to trigger the epigenetic reprogramming of innate immune cells (“trained immunity”) that may function as APC (antigen-presenting cells), as well as augmenting T-cell immunity by stimulating IFN-γ production and subsequent Th1 cellular responses (Kleinnijenhuis et al., 2012, 2015; Zimmermann et al., 2019).

The BCG bacterium now exists globally as a range of genetically distinct substrains that lack varying genomic regions relative to M. tb and M. bovis. The common deletion of RD1 (Region of Difference 1) that is present in all BCG strains is a major cause of attenuation (Conrad et al., 2017; Tiwari et al., 2019). The latter encodes a vital part of the type VII secretion system of M. tb, known as ESX-1 (Majlessi et al., 2005). Although BCG diversification has led to heterogeneous substrains, to date, no significant difference in efficacy has been reported in humans among the various substrains (Milstien and Gibson, 1990; Behr and Small, 1999). In animal studies, the immunogenicity of BCG appears to vary in a strain, dose, and delivery-dependent manner. In addition, comparative analyses in mice have reported differences in efficacy and virulence across 13 diverse BCG substrains (Zhang et al., 2016). Low doses of BCG elicit an almost exclusively Th1 response, while higher doses elicit a mixed Th1/Th2 response (Jiao et al., 2003). In addition, although BCG is delivered to infants intradermally, alternative delivery methods may enhance specific immune responses. For example, intranasal administration of BCG is a method of potentially inducing systemic and long-lasting secretory immune responses (Langermann et al., 1994; Angelidou et al., 2020). Therefore in the future design of BCG-based vaccines, the availability of BCG substrains already in use in the target population, specific environmental factors, and administration techniques may all be important variables to account for (Castillo-Rodal et al., 2006).

BCG’s ability to induce a robust Th1 response and non-specific immunomodulatory effects make it a successful “live-adjuvant” for combining with a variety of heterologous antigens. In the past, the direct co-administration of BCG mixed with an antigen of interest has been examined for protection against various diseases. For example, Pearce et al. have used paramyosin, an antigen derived from Schistosoma mansoni, for intradermal co-administration with BCG in mice. Partial protection of 26–33% against S. mansoni challenge occurred when BCG was co-administered with only 4–40μg of the paramyosin antigen, while 2mg of the antigen alone was required to obtain the same level of protection. Protection occurred through the induction of a dominant Th1 response (Pearce et al., 1988; Sher et al., 1991). In a related study in sheep, co-administration of BCG and paramyosin leads to enhanced immunogenicity and up to 48% worm reduction compared to BCG alone (Taylor et al., 1998). Furthermore, the use of BCG as an adjuvant alongside various autoclaved Leishmania has proven to be a promising combination and is widely used in Leishmania vaccine (Dube et al., 1998; Misra et al., 2001; Nahrevanian et al., 2013). More recently, the progression of recombinant DNA technology has led to the utilization of BCG as a vaccine platform via genetic engineering approaches. Originally described in 1991 in a landmark paper by Stover et al., foreign genes of interest can now routinely be introduced into BCG through the use of E.coli-mycobacterium shuttle vectors, such that the encoded proteins are expressed extracellularly, intracellularly, or remain bound to the BCG cell surface (Stover et al., 1991; Zheng et al., 2015). In an attempt to maintain a stable level of expression of foreign antigens in BCG, the mycobacterial hsp60 promoter is most commonly included in plasmid expression vectors that are capable of integrating at the non-essential mycobacteriophage L5 attachment site (attB; Stover et al., 1991; Zheng et al., 2015). Lastly, optimization of foreign gene sequences based on the codon-usage preferences observed in Mycobacteria may also prove beneficial to maximize the level of heterologous gene expression (Varaldo et al., 2004).

BCG is an extremely advantageous platform for vaccine development due to its long-history of being safe to use (it is approved for use in many countries worldwide), its stability, and ease of distribution, and it is a live-vaccine vehicle that provides sustained cellular immune responses. It also serves as an effective adjuvant by enhancing immunogenicity above that seen with just the antigen alone (Varaldo et al., 2004; Zheng et al., 2015) and is capable of “off-target” immune enhancing effects including trained immunity (Vierboom et al., 2021). When employing rBCG, surface expression or secretion of the protein of interest appears to be critical, partially due to increases in antigen presentation and improved cross-priming, both leading to increased cellular immunity (Deres et al., 1989; Reitermann et al., 1989; Grode et al., 2005; Sali et al., 2010). Therefore, as noted above and in Table 1, signal sequences, such as those derived from Ag85B and the 19kDa antigen, are commonly fused to the foreign genes of interest (Garbe et al., 1993; Gomez et al., 2000; Sixsmith et al., 2014; Oliveira et al., 2017). In many instances, it may be ideal to choose a signal sequence which utilizes the general secretory pathway (Bendtsen et al., 2005). The latter may be favorable since it transports proteins in an unfolded state, thus limiting the creation of non-functional misfolded and aggregated proteins and does not require the action of chaperone proteins or accessory molecules (Feltcher et al., 2010). Bacterial lipoprotein signal sequences, such as that associated with the 19kDa lipoprotein (LpqH), have also been used in rBCG research due to their strongly immunogenic properties (Stover et al., 1993; Dennehy et al., 2007). For example, an anti-rotavirus vaccine expressing VP6 only induced significant cellular immunity-based protection when linked to the 19kDa lipoprotein signal sequence that allowed transport of VP6 to the BCG outer membrane (Dennehy et al., 2007).

One potential limitation in the use of rBCG technology exists in the use of E. coli-mycobacterium shuttle vectors to transfer foreign genes of interest into BCG. These plasmids commonly contain an antibiotic resistance gene as a selectable marker. However, vaccines containing antibiotic genes are largely unsuitable for human use, predominantly due to the potential risk of horizontal transfer of antibiotic resistance to nearby microbial populations (Borsuk et al., 2007; Mignon et al., 2015). Therefore, methods, such as Cre-lox recombination, have been explored to subsequently remove antibiotic resistance markers following transformation and selection. Cre-lox recombination systems are very efficient for multiple gene replacements but to the best of our knowledge have yet to be applied to rBCG development, although they have been used in other Mycobacteria (Lambert et al., 2007; Murphy et al., 2015; Hurst-Hess et al., 2017). Alternate selection methods, such as auxotrophic complementation, allow in vivo selection in the absence of antibiotics through the use of a BCG strain that, for example, is auxotrophic for leucine (leuD knockout) in conjunction with a plasmid vector containing leuD that is then used to introduce the foreign antigen of interest (Borsuk et al., 2007). The creation of an unmarked rBCG vaccine in this manner has been described by Nascimento et al. (2009). In this particular study, a rBCG-pertussis vaccine, which elicited protection against B. pertussis challenge in mice, was generated through auxotrophic complementation based on lysine rather than leucine (Nascimento et al., 2009).

In summary, BCG has been shown to provide protection against a large range of infections beyond TB, including both bacterial and viral respiratory infections (Fine, 1995; Zheng et al., 2015). Even in the context of the current COVID-19 pandemic, a correlation has been suggested between COVID-19 case fatality rates and national rates of BCG vaccination (Rivas et al., 2021). Although it remains to be proven, protection against SARS-CoV-2 could potentially be associated with BCG’s trained immunity and non-specific Th1 responses, in line with its ability to reduce other respiratory tract infections in children (Rivas et al., 2021). In theory, any anti-SARS-CoV-2 activity displayed by BCG could potentially be enhanced by creating a novel rBCG-based vaccine candidate that expresses one or more protein components of the SARS-CoV-2 virus (Krammer, 2020) and our group is currently exploring this line of inquiry.

The use of BCG for a variety of vaccination purposes – both specific and non-specific – has opened the door for renewed enthusiasm to further explore the use of rBCG vaccine technology in future vaccine development and immunotherapy.

EM performed the literature searches and wrote the manuscript. MR, PD, and MN contributed to the writing and editing. All authors contributed to the article and approved the submitted version.

This manuscript was supported by funding obtained from the McGill Interdisciplinary Initiative in Infection and Immunity (MI4).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.