Is interferon-gamma the right marker for bacille Calmette–Guérin-induced immune protection? The missing link in our understanding of tuberculosis immunology

Abstract

Bacille Calmette–Guérin (BCG), developed a century ago, is the only licensed tuberculosis (TB) vaccine in use to date. The protective efficacy of BCG against TB varies with no apparent protection in some population, and mechanisms of its immune protection is poorly known, and yet BCG is the most widely used vaccine, with more than 4 billion BCG-vaccinated children globally. BCG is probably the only licensed vaccine currently in use believed to mediate immune protection through the production of interferon (IFN)-γ by CD4 T cells, which in turn activates macrophages to kill Mycobacterium tuberculosis (Mtb). Currently, a number of new TB candidate vaccines are in different phases of clinical trial. The majority of these new vaccines are either recombinant forms of BCG or prime boosters of BCG (rBCG) and their immunogenicity is tested using BCG as a benchmark by measuring specific IFN-γ produced by CD4⁺ T cells as a protective immune marker. However, some recent studies that examined mechanisms of immune protection of BCG in animals and humans have reported a lack of correlation between IFN-γ production by CD4 cells and BCG-induced immune protection. These studies point to the fact that there is a missing link in our understanding of TB immunology. Conversely, there is emerging evidence that other T cell subsets (gammadelta, γδ), CD8⁺ T cells and natural killer (NK) cells may play a vital role in immune protection against Mtb infection and BCG-induced immune protection. γδ T cells and NK cells, which were considered to be part of the innate immunity in the past, have been shown to develop immunological memory upon re-encounter with the same pathogen. In this paper, the controversy over the role of IFN-γ as a marker for protective immunity against TB, and emerging data on the role of γδ T cells, CD8⁺ and NK cells in TB immunology, will be presented.

Tuberculosis

Tuberculosis (TB), caused mainly by Mycobacterium tuberculosis (Mtb), is one of the most important infectious diseases worldwide. An estimated one-third of the world's population is infected with Mtb, of which 9 million develop clinical TB and up to 2 million die annually [1]. Although a high burden of TB is observed in developing countries, especially subSaharan Africa, a significant increase in TB incidence has been reported in some cities in industrial countries [2]. Control of TB remains difficult because TB control programmes often have limited resources, due to TB and human immunodeficiency virus (HIV) co-infection, the emergence of multi-drug drug-resistant (MDR) strains of Mtb and to the lack of simple, accurate and inexpensive diagnostic tests [1,3].

It is generally acknowledged that an effective vaccine that prevents TB transmission is essential for cost-effective and sustainable long-term intervention [4]. However, BCG is the only licensed TB vaccine in use at present, and an estimated 100 million children receive BCG every year globally [5]. The ability of BCG to impart immune protection has been debated since the 1930s, because randomized controlled trials have shown efficacies ranging from 0 to 80% [6,7]. In addition, mechanisms of immune protection imparted by BCG are poorly understood, and the role of interferon (IFN)-γ as a marker of protective immunity imparted by BCG is controversial [4,8–12].

Like many other currently used vaccines, BCG was developed empirically without understanding the mechanisms of its immune protection [10]. In addition, the development of new TB vaccines continued, with no understanding of their protective immune mechanisms, and their immunogenicity is tested against BCG itself [9,13]. Development, testing and delivery of candidate vaccines require a thorough understanding of mechanisms behind immune protection, including an understanding of effector cells and molecules involved in protection, and how to measure these markers.

Bacille Calmette–Guérin (BCG)

BCG vaccine was developed, over a period of 13 years, by Calmette and Guérin by the attenuation of virulent Mycobacterium bovis through 230 passages [14]. It was first administered as an oral vaccine to a newborn in 1921 and was recommended for global use by the League of Nations in 1928 [11]. BCG was introduced into the Expanded Program of Immunization (EPI) in 1974 [14]. Since then an estimated 4 billion children have been vaccinated with BCG, with acceptable safety and at low cost [15]. Although early studies reported that children were protected against TB, subsequent studies since the 1930s have shown that the efficacy of BCG in imparting protection against TB has varied from 0 to 80%. Trials conducted in the 1940s and 1950s in some European countries demonstrated the vaccine to be highly effective (70–80%). However, trials conducted in North America actually showed that BCG does not confer any immune protection. This is why BCG has never been given routinely in the United States as it has in northern Europe [6,16]. One of the largest clinical trials, involving 250 000 individuals, conducted in the Chennai region of southern India, showed no apparent protection in adults and adolescents. In some cases, vaccinated individuals were more likely to develop tuberculosis compared to their placebo group [17]. Surprisingly, no clinical trial was conducted in Africa before integrating BCG into mass vaccination programmes [17]. However, studies on immune response profiles of BCG-vaccinated children is just beginning to emerge [13,18–20].

Finally, some recent studies have reported a lack of correlation between immune protection imparted by BCG and IFN-γ produced by CD4⁺ T cells [8,11,12].

Currently, approximately 12 TB vaccine candidates are in different phases of clinical trial [4,9,12]. Ten of these candidate vaccines are pre-exposure vaccines, whereas two candidate vaccines, the inactivated mycobacteria M. vaccae and the semi-purified Mtb fragments (RUTI), are considered post-exposure vaccines to be used as adjuncts to chemotherapy [4].

The pre-exposure candidate vaccines are either recombinant (rBCG30 and rBCGΔ UreC:Hly) or heterologous prime-boosters for BCG or rBCG [4,9,12]. The issue is: if BCG is the answer, what was the question?

Mechanisms of BCG-induced immune protection

In addition to controversies concerning the protective efficacy of BCG, very little is known about the mechanisms of immune protection imparted by BCG. Understanding how BCG confers protection is central to the development of new vaccines that aim at augmenting the efficacy of BCG or replacing it [21]. Like most vaccines currently in use, BCG was developed empirically without a clear understanding of how it induces protective immunity, and current efforts to develop new vaccines continued with the same lack of understanding [10]. Most current vaccines mediate their protective efficacy through the induction of vaccine-specific antibodies (http://www.who.int/immunization/document/elsevier-vaccine). BCG is the only vaccine whose protective immunity is believed to depend upon the induction of CD4⁺ T cells that produce IFN-γ, which in turn activates macrophages to kill Mtb[10,21].

One central concept that has dominated immunology of infectious diseases during the last two decades is the T helper type 1 (Th1)/Th2 paradigm, in which helper (CD4) T cells have been given an exclusive division of labour [22]. According to this paradigm, Th1 T cells protect the host from intracellular pathogens, including Mtb, while Th2 cells protect from extracellular pathogens [22]. Thus, in the case of TB, IFN-γ secreted by CD4 T cells (Th1 cells) has been used as a yardstick against which protective immunity against BCG or other related TB vaccines are evaluated. However, efforts to develop new vaccines that could augment or replace BCG have achieved moderate success, because there is a missing link in our understanding of BCG-induced immune protection and TB immunology. However, several studies have shown that the Th1/Th2 paradigm has major shortcomings and recommend a paradigm shift [23–26]. Therefore, a better understanding of how BCG protects against TB remains an important first step in order to develop better vaccines [21].

IFN-γ as protective immune marker?

Acquired immunity against TB, including that induced by BCG vaccination, is believed to be mediated by IFN-γ, as reported by several investigators, both in animal models and humans. One early study in an experimental model suggested that IFN-γ is critical for activation of macrophages [27]. This finding, along with the extreme susceptibility of people unable to produce IFN-γ[28] and mice with disruptions in the IFN-γ or p40 gene to Mtb infection [29,30], have led to the proposed use of IFN-γ as a correlate of protection for new vaccines against TB.

After the onset of acquired immunity, macrophages are activated by IFN-γ, mainly from T lymphocytes. As indicated earlier, one mechanism by which IFN-γ mediates mycobacterial killing is the activation of oxidative burst in macrophages [27]. Treatment of murine macrophages with IFN-γ prior to infection with Mtb activates cells to kill a significant fraction of the subsequently added bacteria by activating expression of nitric oxide (NO) synthase 2 and production of reactive nitrogen intermediates [31].

The second way in which IFN-γ mediates mycobacterial killing is by enhancing phagosome maturation in a manner dependent upon the immunity-related GTPases [32] and increased lysosomal delivery of mycobacteria by activating autophagy [33]. However, in addition to IFN-γ, cytokines tumour necrosis factor (TNF) and vitamin D have been shown to activate murine macrophages to kill Mtb[34].

An important point that needs to be underlined is the difficulty of differentiating between the role of CD4 T cells from that of secreted IFN-γ. CD4T cells are helper T cells that co-ordinate adaptive immune responses, including vaccine-based immunity. They help B cells to produce antibody and to undergo class-switching and affinity maturation; they recruit and activate CD8 T cells, macrophages, neutrophils, eosinophils, basophils and other effector cells (http://www.who.int/immunization/document/elsevier-vaccine). CD4 T cells probably activate DCs that present antigens to CD8 T cells and may also signal themselves directly to CD8 T cells via expression of CD40 or cytokines such as interleukin (IL)-2 [35,36].The diverse function of CD4 T cells is determined by their cytokine secretion patterns and their tissue locations [37].

Therefore, CD4 T cells are crucial for the generation of acquired immune responses against both intracellular and extracellular pathogens, and their central role is not restricted to secretion of IFN-γ to control Mtb infection (http://www.who.int/immunization/document/elsevier-vaccine) [38].

Although BCG vaccine has been in use for almost a century, few studies have examined the mechanism of its immune protection in animal models [39,40] and humans [41]. Many earlier studies in murine model using antibody depletion, gene-disrupted mice or adaptive transfer indicate that CD4 T cells are key to the control of Mtb infection [42–45]. However, as indicated below, some recent studies challenge the view that IFN-γ secreted by CD4 T cells is a marker for protective immunity imparted by BCG.

One study that examined the mechanisms of BCG-induced immune protection in IFN-γ knock-out mice has indicated that CD4 T cells control > 90% of intracellular Mtb growth in vitro in the complete absence of IFN-γ, via a NO-dependent mechanism [40]. Moreover, BCG-vaccinated IFN-γ-deficient mice exhibited significant protection against Mtb challenge that was lost upon depletion of CD4 T cells [40]. The above studies suggest that the levels of IFN-γ produced by a mouse in response to a candidate vaccine do not always correlate with the effectiveness of that vaccine during Mtb challenge [46].

These observations may imply that CD4 T cells possess IFN-γ-independent mechanisms that can limit the growth of intracellular pathogen and are dominant in secondary responses to Mtb[40]. Other studies in mice have provided strong evidence that IFN-γ does not correlate directly with BCG-induced protection [39].

Similarly, some recent studies that examined protective immune mechanisms of BCG in humans indicate a lack of correlation between BCG-induced immune protection and levels of IFN-γ. A study that evaluated immune responses against BCG in vaccinated humans, using several assays, found that mycobacterial growth inhibition did not correlate with IFN-γ responses [47]. Similarly, a study of newborns conducted in South Africa suggests that IFN-γ responses are not indicative of protection in humans; however, BCG vaccination induced a complex pattern of cytokine expression and phenotypes [41]. Among five phenotype patterns of CD4⁺ T cells, central memory T cells were more likely to produce IL-2 and effector T cells were more likely to produce IFN-γ[48]. A study that compared levels of 42 cytokines in UK and Malawian children show that seven of the Th1-related cytokines were higher in UK children, whereas 20 of the Th2 cytokines and chemokines were higher in Malawian children [18]. Furthermore, mycobacterial infection of human and mouse macrophages disrupts a pivotal part of IFN-γ intracellular signalling, resulting in the inhibition of IFN-γ-induced gene expression [49,50].

In fact there are studies, both in mice [51] and humans [52], that have shown increased IFN-γ during Mtb infection as being indicative of disease progression. A study carried out to compare cytokine levels during latent and slowly progressive TB in mice showed that the levels of IFN-γ increased significantly during slowly progressive TB but declined during latent TB [51]. Moreover, the VACSEL group, using early secreted antigenic target 6 kDa (ESAT-6), have shown that household contacts with increased expression of IFN-γ were at most risk of developing progressive disease [52]. Finally, a study by Bennekov et al. [53] suggests that IFN-γ levels during Mtb infection may simply be a measure of inflammatory status rather than a protective immune response.

Taken together, these studies suggest that immune responses that have so far been considered to be crucial for protection against TB, including IFN-γ production by CD4 T cells, are not sufficient for protection and do not represent usable correlates of risk or protection in the context of vaccine trials.

The role of CD8 T cell gammdelta (γδ) T cells and natural killer (NK) cells

In the past, the notion that IFN-γ-producing CD4 T cells are central to immune protection against TB has dominated TB immunology, while the role of other cells such as γδ T cells, NK cells and CD8 cells, and antibody responses, has been given little attention. However, emerging data from animals and humans suggest strongly that γδ T cells, NK cells and CD8 T cells and specific antibody isoptypes may play a crucial role in protection against TB. Importantly, immune cells, such as macrophages, γδ T cells and NK cells, that have been classified traditionally as components of the innate immune system, have been shown to develop immunological memory and expand upon re-encounter with the same pathogen [54–56].

In the past, immunological memory was thought to be limited to CD4, CD8 T cells and B cells. However, there is emerging evidence that γδ T and NK cells can develop immunological memory. Moreover, although lacking adaptive immune responses, plants and invertebrates are protected against reinfection with pathogens and, more interestingly, invertebrates even display transplant rejection. Various studies have also demonstrated cross-protection between infections independent of T and B cells [54]. The demonstration of immunological memory in lower animals and plants, and cells of the innate immune system that lack components of the adaptive immune system, challenge our current understanding of vaccine immunity and call for a paradigm shift. We have reviewed previously how the Th1/Th2 paradigm has misguided our understanding of infection immunity and vaccine development and the possible protective role of specific antibody isotypes against TB infection [26]. In the following section, the role played by γδ T cells, NK cells and CD8⁺T cells in immune protection against TB will be presented.

γδ T cells

γδ T cells represent a small subset of T cells that possess distinct T cell receptors (TCR) on their surface. These cells constitute a whole system of functionally specialized subsets that have been implicated in the innate responses against tumours and pathogens, the regulation of immune responses, cell recruitment and activation and tissue repair [57]. Immune responses of γδ T cells to Mtb were described as early as 1989 [58], and mycobacterial phosphoantigens were identified as potent stimulators of γδ T cells [59]. Direct evidence for the ability of Mtb to activate γδ T cells was provided by studies that have shown proliferation of peripheral blood γδ T cells in response to killed preparations of Mtb[60].

Later, several roles of these cells were described in protection against TB, such as phagocytosis, cytotoxicity, production of IL-17 and induction of maturation of dendritic cells (DCs) [56].

It was reported earlier that γδ T cells kill macrophages harbouring live Mtb through granule-dependent mechanisms, resulting in the killing of intracellular bacilli. Moreover, it has been reported that these cells reduce the viability of both extracellular and intracellular Mtb through granulysin and perforin, both detected in γδ T cells. A recent work has shown that human γδ T cells are also capable of phagocytosis, a function previously assigned exclusively to innate myeloid lineage cells such as neutrophils, monocytes and DCs [61]. These findings have suggested that γδ T cells contribute directly to a protective host response against Mtb infection [62].

Moreover, human γδ cells activated in vitro by phosphoantigens are capable of inducing maturation of monocyte-derived DCs [63], and these processes involve both membrane-bound (i.e. CD40L) and soluble (i.e. TNF-α and IFN-γ) T cell-derived signals [56].

Studies from mice and humans show that IL-17-producing γδ cells are the main source of IL-17, suggesting that γδ cells might play a significant role in the prevention of infection through the induction of mature granuloma formation during Mtb infection [56,64]. It has been reported that IL-17 is produced immediately after pulmonary BCG infection, and has also been detected at later stages of Mtb infection in mice [65]. A recent study that examined the protective effects of IL-17 in vivo showed that IL-17 induced protection in mice lacking the IFN-γ gene [66]. Th17 cells were associated with significant prolongation of survival compared to recipients of naive IFN-γ-deficient T cells [66].

CD8⁺ T cells

The role of CD8 T cells in the control of Mtb infection is not well established. Some studies indicate that mycobacteria do not induce CD8 T cell responses in humans [67–69], while animal studies suggest that CD8 T cell responses are important for the control of Mtb infection [70,71]. For instance, mice with disruption in the genes of B₂-microglobulin, TAP or CD8 T cells are more susceptible to Mtb infection than wild-type mice, implicating a role for CD8 T cells [70,71]. For instance, a study that investigated the relative role of CD4 T cells and CD8 T cells during acute and latent Mtb infection in mice has shown that anti-CD8 treatment resulted in a 10-fold increase in bacterial numbers in the lungs of Mtb infected mice, suggesting the role played by CD8 T cells in controlling Mtb infection [72]. CD8 T cells may modulate phagocytic activity or produce molecules such as granulysin that may be directly cytotoxic to the mycobacteria [73,74].

There is accumulating evidence from studies in animals and humans that molecules produced by CD8 T cells such as perforins or granzymes are critical for controlling Mtb infection [41,72,75]. However, the role of CD8 T cells in BCG-vaccinated children is limited. One study in BCG-vaccinated newborns in South Africa has shown that BCG indeed induces the production of CD8 T cells, but at quantitatively lower levels compared to CD4 T cells [76].

NK cells

In addition to γδ T cells, CD8 T cells and NK cells have been shown to be important in immune protection against TB. These cells have been classified traditionally as cells of innate immunity because they lack RAG-recombinase-dependent clonal antigen receptors. However, accumulating evidence in mice and humans suggest that, like the B and T cells of the adaptive immune system, NK cells are educated during development, possess antigen-specific receptors, undergo clonal expansion during infection and generate long-lived memory cells [54,57,77,78].

Emerging data from animal and human studies show that granulysin and perforin found in NK cells can kill intracellular mycobacteria and are implicated in protection against Mtb infection. For instance, a study by Semple et al. [78], that compared the cellular expression of granulysin and perforin cytolytic molecules in cord blood and peripheral blood from 10-week-old infants vaccinated at birth with BCG, showed that in cord blood only CD56⁺ NK cells expressed granulysin and perforin constitutively. These cytolytic mediators were up-regulated in CD4⁺ and CD8⁺ cord blood cells by ex-vivo stimulation with BCG, but not with purified protein derivative (PPD). Following BCG vaccination of neonates, both BCG and PPD induced increased expression of granulysin and perforin by CD4⁺ and CD8⁺ cells [78].

Conclusion

Currently, BCG is the only licensed vaccine in use to prevent TB. However, in some populations it has no immune protection. Although many studies have demonstrated the importance of many components of the immune system against Mtb infection, including IFN-γ, a reliable correlate of immune protection of BCG has remained elusive. Moreover, the mechanisms of immune protection imparted by BCG are poorly known, but BCG is not only the most widely used vaccine; it is also used as the benchmark against which immunogenicity of new vaccines is being evaluated. Although both CD4 T cells and specific IFN-γ may play an essential role in controlling Mtb infection, the use of IFN-γ as a protective immune marker for acquired immunity or vaccine immunity against TB needs to be re-evaluated. Conversely, effector cells such as CD8 T cells, NK cells and γδ T cells whose important roles have been ignored during the last few decades should be given due consideration. Finally, TB immunology and vaccinology must take into consideration the recent observation that immunological memory is not limited to CD4, CD8 and B cells.

Disclosure

I have no competing interest.