The tuberculosis vaccine Bacillus Calmette-Guérin trains the innate immune system

Data from studies conducted over the last 60 years have shown that The Bacillus Calmette-Guerin (BCG) vaccine protects against a variety of subsequent heterologous infections, including viral, bacterial and fungal infections. This has given renewed interest and hope that the BCG vaccine can be used in the fight against COVID-19. The first clinical trials have started in Australia and The Netherlands with 4,000 and 1,000, respectively, health care professionals being vaccinated. In this COVID-19 perspective Kristoffer Jarlov Jensen describes the potent immune-stimulating properties of BCG.

This article was originally published in Danish in BestPractice Nordic Almen Praksis August 2018.

While vaccines are typically developed for and administered against a single disease via antigen-specific recognition, there is gradually strong evidence that a number of frequently used vaccines may also affect susceptibility to a wider spectrum of diseases. This phenomenon is called nonspecific effects of vaccines (sometimes referred to as off-target effects or heterologous effects).¹ 

The tuberculosis vaccine Bacillus Calmette-Guérin (BCG) is one of the most widely studied cases of nonspecific effects. A large number of epidemiological studies in different populations have found a correlation between the BCG vaccine and a lower overall mortality rate. Most recently, three randomized clinical trials in infants in West Africa have all shown a markedly reduction in neonatal (first 28 days) mortality following early BCG vaccination, primarily as a result of reduced sepsis and respiratory infection mortality (the meta-analysis for the three studies showed a mortality reduction of 38 % (95% CI: 17-54%)).^2,3

The innate immune system is trained

Very noticeably, the mortality-reducing effect has been observed already a few days after the vaccination in the randomized studies.⁴ This indicates that the vaccine’s non-specific effects are at least partly due to the innate immune system rather than adaptive (antigen-specific) mechanisms, which takes longer to develop. Animal studies and in vitro studies support this assumption.

The potent immune-stimulating properties of BCG and other mycobacteria are well known and have been widely used in the form of vaccine adjuvants, including in Freund’s complete adjuvant. Furthermore, numerous mouse model studies conducted over the past 60 years have shown that BCG protects against a variety of subsequent heterologous infections, including viral, bacterial and fungal infections.^5,6 A ground-breaking study in 2012 from Nijmegen in the Netherlands showed that this protective effect depends on a functional innate immune system, while to a lesser extent on an adaptive immune system. This is because BCG modulates the innate immune cell monocytes and NK cells to respond more strongly to subsequent second stimulation of the innate immune system.^7,8

Thus, the innate immune cells are said to be trained by BCG, and the mechanism is therefore called ‘trained immunity‘. This has led to a genuine paradigm shift in immunology, where ”trained immunity” has superseded the previous immunological understanding of the innate immune system as static and without the ability to learn (as opposed to the adaptive immune system).⁹ The memory of the innate immune system is shown to be embedded through epigenetic modifications of the DNA of monocytes and NK cells in the form of acetylations and methylations of promoter and enhancer regions to relevant immunological and metabolic genes. These modifications (H3K27ac and H3K4me3) typically lead to increased or faster transcription of the associated genes.¹⁰

BCG protects against viral infection in human model

Although the biological mechanisms behind the observed beneficial effects of BCG are far from fully elucidated, the trained immunity hypothesis seems to have considerable explanatory power. However, the direct clinical translation for the observed striking mortality-reducing effects of BCG from the human studies has been lacking, as such experiments will often involve invasive and possibly damaging procedures.

However, the same Dutch research group that created the ‘trained immunity’ hypothesis for BCG has now, in collaboration with Danish researchers from the Bandim Health Project at the Statens Serum Institute and the University of Southern Denmark, published a robust proof-of-concept study that cleverly uses an approved live attenuated virus vaccine, yellow fever vaccine, as a human infection model to study BCG’s protective immunological effects.

The yellow fever vaccine virus causes inflammation and replicates in the body after immunization, thus resembling a natural viral infection, though without causing serious illness. The study, published in Cell Host & Microbe this year (2018), shows that prior BCG vaccination reduces the rate of infection after yellow fever vaccination in the form of reduced viremia and inflammation, while the formation of the specific immunity against yellow fever is not inhibited.¹¹ 

The training is mediated by the cytokine IL-1beta

As previously shown, BCG vaccination causes widespread epigenetic reprogramming of the monocytes and an increased innate cytokine response to ex vivo stimulation (for example, against common pathogenic microorganisms). However, not all BCG vaccinated individuals experience the ’trained immunity’. In the yellow fever vaccination study, the IL-1beta cytokine response to ex vivo stimulation one month after BCG vaccination was strongly correlated with reduced viremia (five days) after receiving the yellow fever vaccine. This correlation was not found for other cytokines such as TNF-alpha or IL-6. At the same time, there was no correlation between the specific vaccine response (antibody formation or cellular adaptive responses) and the acute viremia.

In a larger cohort study, the authors found evidence that carriers of a particular genetic polymorphism in the promoter region of IL-1beta have reduced ability to express IL-1beta by direct in vitro stimulation, and that their monocytes have reduced ability to be trained by BCG in vitro (measured as increased innate cytokine response to secondary stimulation). The reduced training effect is stronger in homozygous carriers of the variant than in heterozygous carriers, that is, there is a dose-response effect of the polymorphism.

Overall, the results indicate that the BCG training is via IL-1beta signalling, which the authors finally elegantly demonstrate by culturing human monocytes with artificial IL-1beta in vitro, and then demonstrating training in the form of stronger cytokine responses to secondary innate immune system stimulation. The authors collectively suggest that the BCG training of the monocytes is mediated by IL-1beta.¹¹

The study is extremely interesting in that it establishes a proof-of-concept in humans of the protective effect of BCG against heterologous infection, as well as proposing a pathway of how the BCG vaccine trains the innate immune system via IL-1beta signalling.

The effect of the vaccine takes place in the bone marrow

In the present study, BCG was administered four weeks before the yellow-fever vaccine, and previously the training effect of the peripheral monocytes has been demonstrated up to 12 months after BCG.¹² Because monocytes in the bloodstream have a life span of a few days, the long-term effect after BCG immunization can hardly be explained by the direct training effect of the monocytes upon encounter of BCG in periphery blood.

In contrast, a recent study in mice has shown that BCG can be transported from the blood to the bone marrow via phagocytic cells, resulting in an expansion of hematopoietic stem cells directed toward the myeloid cell line. The resulting macrophages have a changed gene expression profile and possess enhanced anti-mycobacterial properties. This effect on the hematopoietic stem cells could be observed in the mice up to five months after BCG immunization.¹³

BCG’s nonspecific effects and complicated interactions

While the clinical studies from Guinea-Bissau clearly show a strong nonspecific protective effect of BCG against neonatal mortality, not all studies have been able to demonstrate beneficial nonspecific effects on mortality or morbidity. However, the apparently contradictory findings may prove compatible when important background factors are included in the equation, as these may interact with the BCG vaccine effects (that is, effect modifiers).

For example, it has been suggested that vaccine immunity transferred from the parents interact positively with the BCG effect. A recent clinical trial of BCG vaccination of infants in three Danish maternity departments (Calmette study 2012-2015) showed overall no beneficial effects of BCG.¹⁴ However, in the subgroup of infants whose mothers reported being previously BCG vaccinated, BCG was associated with significant reduction in infectious hospitalization as well as the number of visits to general practice.^15,16

Thus, while the Calmette study did not provide evidence for a reintroduction of BCG into the Danish health program, the study provided important new data for a better understanding of BCG’s complex immune training. The understanding of how transmitted maternal (perhaps also paternal) immunity can enhance the beneficial immune-training effect of BCG is unknown. Epigenetic programming can be transmitted horizontally to the offspring,¹⁷ although it has not yet been shown for precise innate immune defence training, while placental transfer of antibodies has been known for a long time.¹⁸ If internalization of BCG in antigen-presenting cells and transport to the bone marrow is essential for the persistent training condition, the maternal immunity component may contribute in the form of anti-mycobacterial antibodies that promote the opsonization of BCG bacteria.¹⁹ However, as most BCG vaccinated mothers received the vaccine in childhood, the transgenerational transfer of immunity requires a rather long-lasting parental BCG effect.

Other interesting identified effect modifiers can be briefly mentioned in this context:

• Gender: In the Guinean clinical trials, boy children experienced a markedly faster reduction in mortality after vaccination compared to girls.⁴
• Other routine vaccines: BCG’s beneficial effects appear most pronounced before the children are given inactivated vaccines, after which the effect may be cancelled or reverted to a negative effect.^16,20 Thus, the time of evaluation after BCG is critical.
• The BCG type: BCG occurs in a number of subtypes, so-called strains, which are genetically different and also manufactured by different vaccine companies with differences in production technology and concentration (colony forming units). These vaccine genetic and manufacturing factors may also influence both specific and nonspecific immunological and clinical effects.^21,22

The interacting and modulating parameters mentioned here naturally increase the complexity of identifying BCG’s mechanisms of action.

While efforts to understand the effects of BCG on the innate immune system are partly driven by the lack of previous explanation of how a tuberculosis vaccine could lead to markedly reduced sepsis and respiratory-related infectious mortality in infants, it should also be noted that not even the specific tuberculosis-protective mechanism of BCG is fully understood. The BCG vaccine is thus still the only approved vaccine against tuberculosis. It is not inconceivable that the effect is largely mediated through innate immune mechanisms.²³ A number of clinical trials of new candidate vaccines for tuberculosis also attempt to learn from BCG’s good stimulating properties of the innate immune system.^24,25 

The nonspecific effects and the vaccination programs

While BCG’s tuberculosis-specific protection is far from optimal, our epidemiological and clinical studies from Bandim Health Project in Guinea-Bissau and Denmark emphasize the importance of also assessing the overall health effects of vaccines. Other routine vaccines that are about to be phased out or have already been phased out have shown beneficial nonspecific effects in population studies, including oral polio vaccine,^26,27 measles virus vaccine,^28,29 and the smallpox vaccine.^30,31 The immunological studies mentioned in this article have been crucial to prompt WHO to recognize and show interest in nonspecific effects of vaccines.³² From previous neglect and oblivion, the agenda has now moved on to address to what extent and under what circumstances nonspecific effects are important and should be taken into account in public health programs.

Conclusion

The above-described study “BCG Vaccination Protects against Experimental Viral Infection in Humans through the Induction of Cytokines Associated with Trained Immunity” has showed that BCG can train the immune system to increase readiness for a viral infection. Also, the study provides a strong indication of, that this training is mediated by the BCG vaccine stimulation of the central cytokine IL-1beta.