Mycobacterium tuberculosis bacteriology and immunity necessitate a multi-pronged immune strategy for vaccination

Abstract

Mycobacterium tuberculosis (Mtb) continues to kill more individuals each year than any other individual pathogen. Advances in bacteriology and immunology signify that Mtb vaccines need to induce responses from multiple components of the immune system, over and above Th1 responses targeted by most current vaccine candidates. This review provides an update on key features of Mtb bacteriology, host immunity, and discusses how new information is driving novel vaccine designs.

Introduction

Around one-quarter of the global population is infected with Mycobacterium tuberculosis (Mtb) and the life-time risk of reactivation in these individuals is between 5 and 10%¹. Tuberculosis (TB) impacts all countries and all age-groups and was the cause of 1.25 million deaths in 2023, including 161,000 deaths in individuals with human immunodeficiency virus (HIV) infection^2,3. The Bacille Calmette-Guérin (BCG) vaccine remains the only licensed TB vaccine, and while effective in preventing miliary and meningeal TB in children, provides little protection against pulmonary TB in infected adults⁴. BCG’s protection appears to be linked, at least partly, to trained innate immunity, a memory-like process for innate immune cells⁵. Replacement of BCG with improved live-attenuated Mtb-based approaches (such as MTBVAC, Table 1) is being investigated and is hoped will lead to enhanced protection compared to BCG. As the prevalence of multi-drug resistant Mtb continues to expand worldwide, the availability of safe and effective vaccines to reduce the global TB burden is paramount^6,7. Accordingly, the development of effective pre-and post-exposure vaccines is a key goal in the World Health Organization’s (WHO) End TB strategy^3,8. While six vaccine candidates are currently in Phase 3 development, only M72/AS01 has today demonstrated efficacy (50%) in preventing pulmonary TB in humans with Mtb infection (Table 1)³

Table 1.

Overview of Mtb vaccines currently in Phase 3 clinical development

Name (sponsors)	Components	Design	Features of the immune response	Evidence of efficacy	Target indication	Target population	Ref
M72/AS01E (Bill & Melinda Gates Medical Research Institute, GSK)	M72 fusion protein (Mtb32A and Mtb39A)	Protein with AS01 adjuvant	M72-specific antibodies and diverse, polyfunctional M72-specific CD4 + T cells. No CD8 + T cell responses.	50% protection against progression to active pulmonary TB in Mtb-infected, HIV-negative adults for 3 years	Disease prevention	Adolescents, adults, with/without HIV, with/ without Mtb infection	Tait et al., 201957; Day et al., 201356
MTBVAC (Biofabri, Bharat Biotech, University of Zaragoza, International Vaccine Initiative, Tuberculosis Vaccine Initiative, HIV Vaccine Trials Network)	Mtb	Live attenuated	Th1 response with proliferation of effector γδ T cells with an IFNγ and cytotoxic profile	Preclinical	Disease prevention	All ages	Felgueres et al., 2025134
Immuvac (Indian Council of Medical Research, Cadila Pharmaceuticals)	Mtb	Inactivated	-	-	Disease prevention	Adolescents, adults	-
GamTBvac (Gamaleya Federal Research Center for Epidemiology and Microbiology, Ministry of Health of the Russian Federation)	g85a and ESAT6-CFP10 fusion protein	Protein with Th1 GcP adjuvant	Antigen-specific IFNγ release, Th1 CD4 + T cells, and IgG responses	-	Disease prevention	Adults	Tkachuck et al., 2020135
VPM1002 (Serum Institute of India Private Limited, Vakzine Projekt Management GmbH)	BCG strain	Live attenuated	Lower polyfunctional CD4+ and CD8 + T cells than BCG	Preclinical	Prevention of infection	All ages and people cured of active TB	Cotton et al., 2022136
BCG travel vaccine (Henry M. Jackson Foundation for the Advancement of Military Medicine)	BCG strain	Live attenuated	-	-	Prevention of infection	Adults	-

Γδ gamma-delta, BCG Bacille Calmette-Guérin, GcP glucan-chitin particles, HIV human immunodeficiency virus, IFNγ interferon-gamma, Mtb Mycobacterium tuberculosis, TB tuberculosis.

The success of Mtb as a human pathogen relies on is ability to evade and control host immunity. The extraordinary complexity of interactions between Mtb and the host immune response evolves during the course of the infection, presenting unique challenges for vaccine development⁹, particularly since much about the Mtb organism and these interactions remains opaque. Here I provide a contemporary update summarizing the most recent findings describing Mtb bacteriology and its importance to vaccinology. The key role of cluster of differentiation (CD)4 + T cell-directed immune responses that contribute to the moderate efficacy of the M72/AS01 vaccine candidate is highlighted. Improving efficacy beyond that seen with M72/AS01 will require activation of other aspects of the immune responses over and above CD4 + Th1 T cell responses, with attention to induction of CD4 + Th17 T cell responses by, for example, macrophage-inducible C-type lectin (Mincle) agonist adjuvants, CD8 + T cells by RNA vectors, or toll-like receptor (TLR)7 agonist adjuvants. Other arms of the immune response that could also be harnessed include CD1-restricted CD4 + T cells by glycolipid antigens, and B cell/antibody responses using the glycoconjugate principle that invokes T cell dependent responses.

An update in Mtb bacteriology

Genomics

The first complete Mtb genome sequence was published in 1998 and updated in 2022, providing important insights into Mtb biology^10,11. Surprisingly, around 10% of the genetic code is dedicated to two families of acidic glycine-rich protein virulence factors: PE (Pro-Glu: 99 proteins) and PPE (Pro-Pro-Glu: 68 proteins) that are transported across the cytoplasmic membrane by the ESX family of type VII early secretory systems¹¹. They can be associated with the outer membrane or fully secreted, but how they move from the periplasmic space and traverse the outer membrane is largely unknown^12,13. PE/PPE proteins interact in numerous ways with the host immune system, and while their subcellular locations and functions remain incompletely understood, they have been implicated in adherence and in modulating immune responses that are favorable to bacterial and intracellular survival ¹⁴.

Another 5% of the genome is allocated to the production of kinases, methyltransferases, acyltransferases, succinyltransferases, glycosyltransferases, deacetylases, and phosphatases that can directly or indirectly modify host chromatin, potentially affecting drug resistance, intracellular survival, and host immune evasion. The functions of a handful of modifying enzymes have been determined and most remain to be characterized¹⁵.

Genomic analysis also revealed a range of secreted proteins such as lipases, phospholipases, esterases, and proteases that could act as virulence factors, as well as storage proteins and ferritin-like proteins thought to contribute to successful intracellular survival¹¹. Compared to other bacteria, a much larger proportion of the coding capacity (9.3%) of Mtb is dedicated to the production of enzymes involved in lipid metabolism^11,16. Mtb effectively degrades host cell lipids for its own metabolism and has at least two separate enzyme systems for fatty acid synthesis, used for construction of mycolic acids and lipids unique to its cell wall. The distinctive features of the Mtb genome provide critical insights into its biology and the host interactions that contribute to its success as a human pathogen.

Understanding Mtb’s biology through its’ genomic diversity as expressed by the existence of different lineages and evolution is an ongoing effort¹⁷. Mtb however, appears predominantly to be antigenically conserved with some rare variable antigens appearing to be relevant in Th17 immunity.

The structurally unique mycobacterial cell wall

The Mtb cell envelope follows the same structural principles as Gram-negative bacteria, having a double membrane structure and a capsule consisting of polysaccharides. In contrast to other bacteria however, the cell envelope is extremely rich in a diverse array of lipids, including glycolipids, phospholipids, aminolipids, phosphoglycolipids, phenolic-glycolipids, and sulfolipids, representing up to 40% of its cellular dry mass, which is double that recorded for the highest lipidated Gram-negative bacteria (20%)¹⁸.

The five compartments comprising the Mtb cell envelope are the cytoplasmic membrane made up of phospholipids, glycolipids, and proteins, a periplasmic space, the cell wall skeleton of peptidoglycan and arabinogalactan layers, an outer membrane composed of an inner layer of mycolic acid and an outer lipid layer of trehalose mycolates, lipoarabinomannans and phosphatidylinositol mannosides, and the outer capsule made up of three main polysaccharides, mostly α-D-glucan, as well as α-D-arabino-d-mannan, and α-D-mannan (Fig. 1)^18,19. The outer membrane glycolipids intercalate with mycolic acids and lie perpendicular to the cell wall (Fig. 2)¹⁹. Besides creating a physical barrier, several outer layer trehalose-containing lipids, mannose-containing lipids, and phenolic glycolipids (PGL) are recognized by host cell receptor or mimic host cell epitopes and act as immuno-escape, and immunomodulating factors¹⁹. The outer membrane is 7–8 nm thick as a result of folding and perpendicular arrangement of the long alkyl chains of mycolic acids. The Mtb cell envelope is thus complex, thick, and waxy in nature, and is fundamental to Mtb pathogencity¹⁹. The family of Mtb envelope lipids and glycolipids as well as α-D-glucan, the primary component of the outer capsule, have important immuno-modulatory activities that are activated on binding to host receptors (Table 2). These actions result in inhibition of phagocytosis and inflammasomes, apoptosis, T cell activation, and help to establish infection, granuloma formation, cavitation, and tranmissability¹⁹. α-D-glucan is a virulence factor that promotes bacterial persistence after initial infection by inducing differentiation of dendritic cells with an altered phenotype and reduced functionality^20,21. Intracellular glucan, produced by the same pathway, is used as a nutrient source during periods of latency. Blockage of key enzymes in the glucan synthesis pathway reduces Mtb survival²².

Fig. 1. Structure of the Mtb cell envelope and predicted location of known and unknown transporter systems.

Reproduced with permission and adapted from Kalscheuer et al.²³. The layers of the Mtb cell envelope are the inner (or cytoplasmic) membrane made up of mainly phosphatidyl-myo-inositol mannosides (PIMs), the periplasmic space and cell wall skeleton of peptidoglycan and arabinogalactan layers (peptidoglycan–arabinogalactan complex or AGP) shown here as one layer, an outer membrane composed of an inner layer of mycolic acids which are covalently linked to the AGP, and an outer lipid layer of non-covalently attached (glyco)lipids and lipoglycans including phthiocerol dimycocerosates (PDIMs), PIMs, trehalose mono- and di-mycolates (TMM and TDM), diacyl trehaloses (DATs), polyacyltrehaloses (PAT), sulfoglycolipids (SGL), lipomannan (LM) and lipoarabinomannan (LAM), and the outer capsule made up of α-glucan, LAM, arabinomannan, mannan, PIM and TMM. The ESX and SEC (membrane-bound SecYEG translocon is shown) transporter systems span the inner membrane. Secretory proteins make their way through the layers of the cell envelope by unknown means^12,13. Transporter systems hypothesized to be PE/PPE proteins or surface porins/structures possibly mediated by EspB are predicted to span the outer membrane but remain unidentified¹³¹. Secreted proteins include virulence factors such as PE-PPE, PE with multiple polymorphic tandem repeats (MPTR), ESAT-6, EsxB (CFP-10), proline-E domain and polymorphic GC-rich repetitive sequences (PE-PGRS), protein tyrosine phosphatase (PtpA), secretion of an acid phosphatase (SapM), and protein kinase G (PknG) ¹³².

Fig. 2. Intercalated surface glycolipids of the outer membrane of the Mtb cell envelope.

Reproduced with permission from Garcia-Vilanova et al.¹⁹ The outer membrane of the Mtb cell envelope sits above the covalently linked peptidoglycan–arabinogalactan (AGP) complex that sits in the periplasmic space. Lipids in the outer membrane intercalate among mycolic acids that lie perpendicular to the plasma membrane. The figure shows the surface location of phosphatidyl-myo-inositol mannosides (PIMs) and lipoglycans (e.g., the mannose capped lipoarabinomannan [ManLAM]), trehalose containing lipids (trehalose-dimycolate [TDM], sulfolipid-1 [SL-1], and diacyl-trehalose [DAT]). Other lipids and outer materials (see Fig. 1) are not shown. Relative number of molecules and size are not accurately depicted, reflecting published experimental data¹⁹.

Table 2.

Summary of selected, important, Mtb surface adhesins^{12,32,33,37,40,47,137}

Surface adhesin	Host receptor	Effect
Surface lipids, glycolipids and α-glucan
Trehalose dimycolate	• Monocyte-inducible C-type lectin scavenger receptor A • Macrophage receptor with collagenous structure	Inhibition of phagosome arrest, production of TNF-α, IL-6, and MCP-1 with recruitment of monocytes and neutrophils
Lipomannan	• TLR2 • CD36	IL-12 production, TLR2 and TLR4 agonist
Phosphatidylinositol mannosides	• DC-SIGN • Macrophage mannose receptor (MR)
• CD1b	TLR2 and type-C lectin agonist
Mannose-capped lipoarabinomannan	• DC-SIGN • MR • CD1b • CD36	Interferes with phagocytosis and arrests phagocyte maturation
α-glucan	• Dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN)	IL-10 production, inhibits CD1 expression, induces phagocytosis
Surface proteins and structures
Antigen 85 complex	• Fibronectin • Tropoelastin and elastin	Cell attachment and invasion
Alanine-proline-rich antigen	• Fibronectin • Pulmonary surfactant protein-A • DC-SIGN • MR
19 kDa (LpqH) lipoprotein	• MR • Toll-like receptor (TLR2)	TLR2 agonist, promotes uptake of bacteria, IL-12 and TNF-α production, T cell proliferation, apoptosis
PstS-1 transporter protein	• MR • TLR2 • TLR4	Pro-inflammatory response
Glyceraldehyde-3-phosphate-dehydrogenase	• Human epidermal growth factor • Fibrinogen, collagen, plasminogen, fibronectin and laminin • DC-SIGN	Promotes pathogen growth
Malate synthase	• Laminin • Fibronectin	Unknown
Chaperone Cpn60.2	• TLR2 • TLR4 • CD43	Blocks macrophage apoptosis
ESAT-6	• Laminin • Beta-2-microglobulin	Attachment, cell lysis, reduced maturation of cellular MHC-I molecules
Heparin binding hemagglutinin antigen	• Epithelial cells • Heparin (extra cellular matrix and membrane bound)	Extrapulmonary dissemination
Protein kinase D	• Brain endothelium-associated laminin	Facilitates transfer across the blood-brain barrier
PE_PGRS33	• TLR2	Maturation of dendritic cells
PE-PGRS11, PE-PGRS17	• TLR2	Maturation of dendritic cells, proliferation of CD4 + T cells
PPE26 PPE57 PPE60	• TLR2	Macrophage activation, Th1/Th17 polarization
PE-PGRS60	• fibronectin	Enhanced adhesion and invasion
Type IV like pili	• Unknown	Unknown
Curli pili	• Laminin	Adherence

The characteristics of the cell envelope result in very low permeability to nutrients and antibacterial drugs^18,23. As a consequence, Mtb invests significantly in production of a range of secretion systems.

Secretion systems

Mtb employs the general secretion pathway (Sec) secretion system, and the twin-arginine translocation (TAT) export systems, both used by bacteria to transport proteins across the cytoplasmic membrane to the periplasmic space²⁴. The Sec pathway transports unfolded proteins. Unusually, Mtb has two SecA proteins; SecA1 transports proteins that are tagged with a specific signal sequence, whereas SecA2 does not require a signal sequence. The SecA2 ATPase contributes to pathogenicity of Mtb by suppressing macrophage responses and promoting intracellular growth ²⁵.

The TAT system transports pre-folded proteins such as phospholipases, that play a role in pathogenesis. Mtb isolates with tat deletions show defective growth, suggesting that a functional TAT system is essential for survival²⁴.

In addition to these, five type VII early secretory systems (ESX-1–5) unique to mycobacteria are used to export effector immune evasion proteins across the inner membrane¹³. They appear to have four-helix bundle structures that are similar to bacterial chaperones²⁶. ESX-1 works together with phthiocerol dimycocerosates, complex outer membrane lipids known to be virulence factors that interfere with phagocytosis, and contributes to the Mtb cell wall barrier. Key substrates of ESX-1 are early secretory antigen target 6 system (ESAT-6 or ESX-A), EspA, EspB, EspC and culture filtrate protein [CFP]-10. ESAT-6 is a well known virulence factor and EspB and EspC may transport proteins across the outer membrane and are potentially involved in direct cell-to-cell contact with host cells¹³.

ESX-1, ESX-3 and ESX-5 all transport PE/PPE proteins and are thought to modify cell wall permeability to allow nutrient uptake. All PPE proteins have a hydrophobic tip close to the N-terminal which interacts with its chaperone and determines secretion system specificity¹². ESX-1, ESX-3 and ESX-5 are required for virulence, while the structure and function of ESX-2 and ESX-4 are unknown^13,24.

Transport of nutrients across the outer wall is effected by porins and PE/PPE small molecule selective channels, as well as other cell transport systems yet to be identified^12,18,23,26. Surprisingly, the transporter mechanisms spanning the Mtb outer wall which are critical for Mtb virulence and possibly survival, remain unidentified (Fig. 1)^27,28. Proteins transported to the periplasm by the Sec and TAT systems have been found in the extracellular environment, indicating that transporters spanning the thick outer cell wall must be present²⁷. By contrast, substrates of the ESX transport systems have not been identified in periplasm, indicating that the ESX system is used for transportation across the whole Mtb cell envelope. However, since the ESX membrane complex is too small to span the entire cell envelope, other secondary mechanisms, such as specific ESX substrates (PE/PPE or WXG proteins) are thought to effect transport across the outer wall^29,30.

Secreted proteins

Mtb secretes a plethora of proteins via its various secretory mechanisms that are potent virulence factors directed against all aspects of the host immune response (comprehensively reviewed by Pal et al., 2022) (Fig. 3)²⁴. The ESX-5 system alone is predicted to secrete more than 150 proteins, many involved in immune evasion and interactions with the host³¹. Attack mechanisms used by secreted proteins include interference with antigen presentation, alteration of inflammasome and cytokine production, inhibition of phagosomal maturation, interference with cell death mechanisms, downregulation of oxygen species production, and epigenetic reprogramming of host genes involved in the inflammatory response²⁴. Several important examples are discussed below.

Fig. 3. Host immune responses mediated by Mtb secretory proteins.

Reproduced with permission from Pal et al.²⁴ Mtb secretory proteins modulate numerous host defense mechanisms including antigen presentation, inflammation, cell death, cytokine induction and phagosome maturation to favor proliferation of the bacilli and successful infection. The secretory proteins have abilities including phosphorylation, ubiquitination, histone modification and regulation of transcription which affect various host immune cell functions. CFP-10: EsxB, ESAT: early secretory antigenic target, PE: proline-glutamine, PknG: protein kinase G, PPE: proline-proline-glutamine, PtpA: protein tyrosine phosphatase A, SapM: secreted acid phosphatase, Zmp1: zinc metalloprotease-1.

ESAT-6 and EsxB (or CFP-10) form a tight dimer that is dependent on the ESX-1 transport system for secretion³². ESAT-6:CFP-10 binds to host beta-2-microglobulin, inhibiting surface expression of major histocompatibility complex (MHC)-I-beta-2-microglobulin complexes on the host cell surface, which in turn interferes with MHC-I-mediated antigen presentation and iron metabolism³². ESAT-6 is secreted during early infection and downregulation of MHC-I antigen presentation may be one means by which Mtb subverts adaptive CD8 + T cell responses. ESAT-6 may also be membrane-bound and provide anchorage by binding to host laminin³³.

The PE/PPE family are exclusive to mycobacteria and considered necessary for Mtb infection, persistence, and dissemination. Many PE and PPE proteins interact to form PE/PPE heterodimers that may be surface-bound at various locations within the cell wall, or transported by the ESX secretion systems. Known functions of these proteins include interactions with innate immune cells that interfere with antigen presentation, inducing both Th1 and Th2-polarised responses. Th1-inducing PE/PPE proteins may distract the host by initiating a Th1 response directed against ‘smoke screen’ proteins while the infection is being established, later promoting a more Th2-directed response to support intracellular survival¹². Similarly, different PE/PPE proteins have other contrasting pro- and anti-inflammatory effects likely to contribute to bacterial survival during different phases of infection. Other functions of PE/PPE proteins include calcium binding which promotes immunomodulatory and intra-cellular effects that support Mtb survival, inhibition of phagosome-lysosome fusion within macrophages, resistance to intracellular stress caused by acidification, nutrient depletion, hypoxia, nitric oxide, and reactive oxygen species, and modulation of host cell death. Early on during infection, apoptosis and autophagy are inhibited to promote the establishment of infection, whereas apoptosis is encouraged late in infection to allow bacterial dissemination¹². Of note, PPE18 is included in the M72/AS01 candidate vaccine. PPE18 stimulates interleukin (IL)-10 secretion and suppresses production of proinflammatory cytokines, contributing to a Th2-biased host immune response that favors intracellular survival³⁴.

Mtb adenylate cyclase is implicated in pathogenesis and elucidation of the transmembrane part of its structure suggests that this segment may have dual roles, firstly in assembly of the helical domain, which links the transmembrane and catalytic domains, and potentially as a receptor for metals and other unknown ligands³⁵. This implies that the transmembrane component can directly regulate catalytic activity in response to extracellular ligands.

Mtb also secretes tuberculosis necrotizing toxin (TNT) to kill host cells by way of secretion into the macrophage cytosol, hydrolysis of NAD+ and subsequent necroptosis. The EsxF and EsxE WXG-family of small proteins (whereto ESAT-6 and CFP-10 also belong) form a water-soluble heterodimer EsxEF complex that assembles into oligomers enabling pore formation to allow secretion of outer membrane protein CpnT (channel protein with necrosis-inducing toxin) and its TNT C-terminal domain²⁶.

The Mtb genome encodes around 130 methyltransferases and 47 N-acetyltransferases, of which a large number are secreted¹⁵. Several have been identified as potent modifiers of the host epigenome, acting via a range of mechanisms to direct the innate immune response to provide an intracellular survival advantage, promote necrosis, and alter cell death programs¹⁵. The function of most of these enzymes remains unknown, and large gaps exist in our understanding of Mtb’s ability to manipulate the host genome to its advantage.

Surface adhesins and pili

Mtb cell wall proteins that have been identified as adhesins include pili, heparin-binding hemagglutinin, antigen 85 (Ag85) complex proteins, alanine- and proline-rich antigenic glycoprotein, 19 kDa lipoprotein, malate synthase, protein kinase D, chaperone and glyceraldehyde-3-phosphate dehydrogenase, as well as the ESAT-6 and PE/PPE proteins discussed above (Table 2)³³.

Mtb was considered for many years to lack pili but at least two types of pili have been identified recently, with potentially more yet to be described^33,36,37. Mtb produces Type IV and curli pili that are similar to those found on Gram-positive and Gram-negative bacteria^33,36,38. Mtb curli pili are surface adhesins that interact with the host immune response, facilitating bind to host laminin and macrophages, biofilm production, and metabolic changes that promote virulence^39,40. Mtb type IV pili appear to be produced on contact with host inflammatory cells and continue to be expressed within macrophages³⁶. As yet the role of each type of pilus in Mtb virulence and drug sensitivity has not been characterized and their potential as vaccine targets is not known.

It is evident that Mtb uses an array of diverse molecules and structures to adhere to a multiplicity of host cell receptors and proteins in the extracellular matrix, allowing it to evade host defenses, directly modulate the host immune response, and promote bacterial survival. However, identification and in-depth characterization of their contribution to infection is lacking for most.

Biofilm production

Mtb is assumed to form biofilms based on its tendency to aggregate in rope-like formations and within granulomas, but in vivo evidence is lacking⁴¹. Mtb forms biofilms on pulmonary bronchial epithelial cells in vitro, which are characterized by high biomass, high polysaccharide content, and retention of metabolic activity within the biofilm in contrast to other Mycobacterial species. While Mtb was shown to disrupt epithelial cell barriers and cause non-apoptotic cell death, these actions have not been linked conclusively to biofilm formation. The challenging conditions for growth inside biofilms may select for metabolic changes associated with persistence and antibiotic tolerance^42,43.

Iron and other metals homeostasis

Restriction of the availability of iron for infecting pathogens is a form of ‘nutritional immunity’ regulated by the host. Mtb produces both membrane-bound (mycobactin) and secreted (carboxymycobactin) siderophores allowing it to sequester iron from a range of host environments and store it as ferritin⁴⁴. The transcription of genes involved in iron metabolism is managed by IdeR, a metal-dependent regulator that when absent, leads to iron intoxication and failure of infection.

Mtb has evolved effective strategies to remain viable in iron-poor environments. The absence of iron initiates a series of metabolic modifications that produce a non-replicative, persistent phase with increased expression of pathways leading to increased antibiotic resistance. This state reverses when access to iron is restored. These data suggest that control of iron metabolism is essential for Mtb survival and resistance to antibiotics, making IdeR, or bacterial ferritin potential vaccine candidates by inducing iron intoxication⁴⁴.

Other metals, such as copper and zinc, are also involved in Mtb metabolism, replication and virulence⁴⁵. Mtb has evolved specific responses to upregulate or downregulate copper and zinc uptake and release according to environmental pressures. Host strategies to restrict the availability of metal ions to Mtb are therefore poorly effective. Copper is used as an antibacterial agent by macrophages, and at high concentrations, zinc ions bind to Mtb proteins resulting in their dysfunction. Targeted delivery of metal ions could potentially contribute to growth restriction of Mtb. Alternatively, blocking the mechanisms used by Mtb for metal acquisition could also restrict growth and promote clearance^45,46.

The host immune response

Initial interactions between Mtb and the host usually occur in alveoli where alveolar macrophages and interstitial dendritic cells become infected, triggering innate and adaptive immune responses and acting as bacterial reservoirs (Fig. 4)⁴⁷. On contact with Mtb, macrophages undertake a range of activities that aim to eliminate Mtb and recruit the inflammatory response. These include phagocytosis, autophagy, apoptosis, activation of the inflammasome, and production of reactive oxygen species. Depending on the macrophage : pathogen interaction, macrophages are activated to produce proinflammatory cytokines such as IFNγ and IL-12 that recruit immune cell mediators to the infection site, potentially leading to the formation of granulomas. Phagocytosis of Mtb by macrophages activates a range of T cell responses including MHC-restricted responses that induce peptide-specific T cells, MHC-independent responses that induce unconventional gamma-delta (γδ) T cells that detect non-peptide phospho-antigens, CD1-restricted lipid-specific T cells, and MHC-I-related gene protein (MR1)-restricted T cells (known as mucosal-associated invariant T [MAIT] cells)⁴⁸.

Fig. 4. Overview of pulmonary Mycobacterium tuberculosis (Mtb) infection and its potential outcomes.

Reproduced with permission from Rahlwes et al.⁴⁷ 1 Mtb is inhaled and engages alveolar macrophages and the innate immune system within alveoli. 2 Bacteria are ingested by alveolar and interstitial macrophages and dendritic cells, activating a wide range of T cell responses. 3 Immune activation leads to recruitment of new cells and the adaptive immune response is triggered. 4 Innate and adaptive immune responses act to achieve bacterial eradication, or to restrict its active replication within a granuloma. 5 Active pulmonary tuberculosis either after primary infection or after reactivation occurs when bacteria gain the upper hand over restrictive immune mechanisms, leading to symptomatic disease and transmission to a new host.

What leads to eradication of infection in some individuals, formation of a granuloma in others, and active pulmonary TB in others (Fig. 5)? Individuals exposed to Mtb who clear the bacteria before an adaptive immune response develops are called ‘resisters’ and there is considerable interest in defining the resister phenotype and leveraging the immune mechanisms that appear to protect these individuals from infection. Infection appears more likely to occur when bacterial ligation occurs via the mannose receptor, which inhibits pathogen recognition and inhibition of phagosomes. Macrophage apoptosis then releases bacteria and attracts other immune cells that become infected⁴⁹. Mtb virulence mechanisms discussed above swing into action to promote intracellular survival, achieve the desired T-helper cell immune bias, and reduce the likelihood of successful clearance. Therefore, initial macrophage interactions with Mtb are thought to be critical in determining the outcome of exposure⁵⁰.

Fig. 5. Human resistance to infection and disease caused by Mycobacterium tuberculosis (Mtb).

The outcome of exposure to Mtb depends on the duration and intensity of exposure, as well as immunogenetic mechanisms of resistance. The most common outcome is persistent asymptomatic infection under successful immune control. A few individuals will develop active TB after initial exposure, and several will remain uninfected (resisters). Described differences in immunity Resisters vs latent TB infection are: increased levels of CD8+ glucose monomycolate + granzyme B + T cells¹³³: Presence of mycobacteria-reactive MR1-restricted T cells⁶⁹, presence of alveolar macrophages with pre-activated baseline M1 phenotype, poly-cytotoxic T cells expressing granzyme B and alveolar macrophages with a stronger TNF profile and less pronounced IFNγ profile⁷², CD4 + T cells enriched in Th17 T cells and regulatory T cell profile⁵³, IgM and class-switched IgG antibody responses, and non-IFNγ T cell responses⁷⁰.

The central role of macrophages

Mycobacteria engulfed by macrophages survive and replicate in phagosomes by preventing fusion with lysosomes and inhibiting the process of acidification needed to activate lysosomal proteases. Macrophages can be activated by Th1 CD4 + T cells to increase lysosome phagosome fusion, stimulate production of oxygen radicals, nitric oxide and other antimicrobial proteases, increase antigen presentation and recruitment of T cells, and secretion of IL-12, which promotes further Th1 T cell differentiation.

Studies in mice suggest that within primary TB lesions, macrophages altered after exposure to TNF express proinflammatory cytokines and induce TNF, IL-6, and complement pathways. These altered macrophages, together with impaired regeneration of damaged lung epithelium and local hypoxia, produce a microenvironment that over time, generate the conditions needed for bacterial dissemination and cavitating pulmonary TB⁵¹.

Macrophages are the initial contacts for Mtb and conduct the subsequent immune response. However, they are affected by Mtb virulence factors that can skew their activities to favor bacterial survival and growth⁵⁰. Optimization of initial macrophage interactions with Mtb might be an important goal for vaccination strategies.

The critical role of CD4 + T cells in human resistance to Mtb infection

IFNγ-expressing Th1 cells are considered a primary mediator of protection, but by themselves are not sufficient to prevent Mtb infection or disease⁵². CD4 + T cells can control Mtb infection without producing IFNγ, suggesting the existence of other protective immune methods^53–55. Such mechanisms could include induction of Th17 responses driven by exposure to rare variable antigens characterized by variable T cell epitopes. These antigens stimulate CD4 + T cells to express RORyT (a Th17 lineage-defining transcription factor) and IL-17, which contrasts to the Th1 responses with high IFNγ production induced by classical immunodominant and highly conserved Mtb antigens⁵⁵. Additionally, MHC-II-mediated interactions between infected monocyte-derived macrophages and CD4 + T cells led to expression of glycolysis genes needed for Mtb control, resulting in clearance of bacteria from macrophages independently of IFNγ⁵⁴.

M72/AS01 as a proof of concept for the importance of CD4 + T cells in immunity

The investigational M72/AS01 TB vaccine contains M72, a fusion protein of two antigens present in Mtb and BCG: Mtb32A (thought to be a serine protease) and Mtb39A (PPE18) combined with the Th1-inducing adjuvant AS01⁵⁶. In a Phase 2b study conducted in Kenya, South Africa, and Zambia, vaccine efficacy in preventing bacteriologically-confirmed pulmonary TB was 50% (95% confidence interval 2–74) in HIV-negative adults with Mtb infection (defined as a positive IFNγ release assay without evidence of active disease), and persisted for at least 3 years^57,58. A Phase 3 study of M72/AS01 commenced in sub-Saharan Africa in 2024⁵⁹.

In BCG-primed adults, vaccination with M72/AS01 induced M72-specific CD4 + T cell subsets that produced numerous combinations of Th1 cytokines^56,60–62. These responses were observed in purified protein derivative (PPD) negative and positive individuals, and have been shown to persist for several years^56,57,60,61. CD8 + T cell responses producing IFNγ or TNF-α have only been observed in one study conducted in PPD-positive individuals, although other studies in PPD-positive individuals have not detected CD8 + T cells after vaccination^56,57,61.

A comparison of six TB candidate vaccines found that M72/AS01 induced the highest levels of Th1-expressing CD4 + T cells out of all the tested vaccines in both Mtb-infected and non-infected participants, implying the best vaccine ‘take’⁶³. Few CD4 + T cells producing IL-17 were detected and antigen-specific CD8 + T cells were not detected after any of the vaccines. This study showed that apart from the higher magnitude of the immune response after M72/AS01, the functional profiles of the responses were similar across all of the tested vaccines, reflecting that to date, Mtb vaccine antigens have been largely selected on their capacity to induce IFNγ in peripheral blood mononuclear cells (PBMCs).

The H4:IC31 vaccine contained an Ag85B and TB.10.4 (from the ESAT-6 protein family) fusion protein and the Th1 adjuvant, IC31. In a Phase 2 study, efficacy of H4:IC31 in preventing sustained Mtb infection in BCG-vaccinated adolescents was 30.5% (p = 0.16), whereas efficacy of re-vaccination with BCG was 45.4% (p = 0.03), raising hopes that BCG re-vaccination could be used to prevent infection in adolescents living in high-risk settings⁶⁴. Enhanced Th1 and some Th17 CD4 + T cells were observed after BCG re-vaccination⁶⁵. Nevertheless, in a second study, the findings were unable to be replicated⁶⁶, suggesting that prevention of infection may present a higher bar than prevention of disease.

The characteristics of “resisters”

Between 9% to 14% of individuals exposed to Mtb remain persistently negative to tuberculin skin testing and IFNγ release assays (“resisters”) (Fig. 5)⁶⁷ Clinical and behavioral factors associated with the resister phenotype have not been identified⁶⁸. However, characterization of cell subsets and cytokine profiles from resisters indicate qualitative differences in antibody and T cell responses compared to individuals who develop latent infection. Compared to individuals with latent TB, resisters had higher frequencies of circulating CD8 + T cells expressing glucose monomycolate and granzyme B⁶⁸. This recently identified cell subset from patients with active TB is proposed to contribute to clearance of infected host cells. Mtb-stimulated PBMCs from resisters also developed higher proportions of polyfunctional conventional and nonconventional T cells expressing programmed cell death protein 1 and/or granzyme B⁶⁸. Resisters display higher numbers of circulating MR1-restricted T cells that lyse infected cells⁶⁹, and higher levels of Mtb-specific T cells with a Th17 phenotype compared to individuals with latent TB⁵³. Studies in high risk settings (household contacts and gold miners in South Africa) have shown IFNγ-independent T cells and IgG antibody responses on ex vivo exposure to Mtb antigens in resisters^70,71. Resisters also displayed higher titers of Mtb-specific IgG1 antibodies to Mtb protein antigens and lipoarabinomannan with enhanced avidity and Fc (antibody constant)-glycan profiles unique to the resistance phenotype^70,71.

The notion that alveolar macrophages might play a critical role in determining the response to Mtb exposure is supported by a study of bronchioalveolar lavage samples from individuals with HIV, in which alveolar macrophages from resisters differed both in number and type compared to individuals with latent TB⁷². Levels of alveolar macrophages in resisters were in the upper limit of the normal range and higher than infected individuals, and included a cluster of poly-cytotoxic CD8 + T cells that expressed granzyme B, perforin and granulysin that all play roles in the anti-Mtb immune response. Resisters showed constitutively higher levels of baseline IFNγ production by pre-activated baseline M1 phenotype alveolar cells, and a TNF-dominated response after ex vivo challenge with Mtb that was not observed in cells from individuals with latent TB⁷².

Resisters therefore appear to possess a mix of constitutive (genetic, epigenetic or related to other factors such as the microbiome), innate and adaptive immune attributes that differentiate them from individuals who develop infection on exposure to Mtb. Functionally, these differences are expressed as a stronger TNF response to early infection, higher levels of functional, class-switched specific antibodies with unique Fc structures, a Th17 CD4 + T cell profile, and recruitment of nonconventional and conventional CD8 + T cells. Surprisingly, IFNγ is not strongly linked to protection/resistance, while genetic deficiencies in IFNγ and evidence from M72/AS01 which induces robust IFNγ responses, appear to show the opposite. A further complication is the impact of high intensity exposure which appears to be able to ‘override’ the resister phenotype⁶⁷. At present it is not known if there are any essential factors needed to achieve resister status. Given that there is unexpectedly high diversity in the repertoire of MR1-restricted T cells between individual resisters⁶⁹, it seems more likely that there is more than one resister phenotype, with varying contributions of genetic, innate and adaptive responses that can potentially result in successful bacterial clearance up to certain levels of exposure intensity.

Identification of immune features that are common prerequisites for resistance could have a major bearing on vaccine design. Identifying resisters and unraveling consistent features that differentiate them from individuals who develop infection, while also controlling for the extent of exposure, remains a formidable task, but one that could provide critical insights into human protection against Mtb.

What genetic deficiencies can tell us about Mtb immunity

The genetic makeup of the host is an important factor that influences susceptibility to infections and their outcome. Within the multitude of described inborn errors of immunity, many are associated with an increased risk of mycobacterial infection⁷³. They include deficiencies that reduce the numbers or functioning of T cells or their receptors, particularly CD4+ and Th1 T cells, and deficiencies that lead to reduced IFNγ^73,74. The latter causes Mendelian susceptibility to mycobacterial disease (MSMD), which refers to individuals who are specifically at high risk of mycobacterial diseases caused by weakly virulent strains and BCG⁷⁵. Identification of MSMD provided the first inkling that the response to Mtb exposure had a genetic component and underscored the role of IFNγ as a key actor influencing the outcome of Mtb infection⁷⁴.

Anti-TNF immunotherapy is also associated with an increased risk of Mtb infection⁷⁶, but associations between TNF gene polymorphisms and increased susceptibility to TB in humans have been inconsistently observed, and mainly restricted to Asian populations, suggesting that other factors may modulate the impact of these polymorphisms^77,78. Identification of a TNF gene variant causing loss of TNF function in family members with recurrent pulmonary TB led to the unexpected discovery that TNF appears to have an isolated role in the prevention of Mtb disease and is required for respiratory-burst-dependent immunity leading to maturation of alveolar macrophages necessary for control of inhaled Mtb⁷⁹. Neither of the individuals carrying the TNF variant showed heightened susceptibility to infections by other bacteria. The authors concluded that TNF appears to have few unique physiological functions except its involvement in production of reactive oxygen species, and was otherwise largely redundant in inflammation and infection⁷⁹. This finding would explain the absence of associations between TNF polymorphisms that do not result in loss of function and an increased risk of TB.

Mtb downregulates TNF production and releases soluble TNF receptors that neutralize TNF activity to avoid TNF-mediated cell death⁸⁰. Correction of TNF functioning could enhance the host response to Mtb⁸¹. However, excess TNF and reactive oxygen species production contributes to inflammation and disease progression. Vaccine strategies targeting TNF will need to maintain a careful balance between the functional versus destructive potentials of TNF in TB⁸⁰. It is surprising that no link has been found with respect to IL-17 and/or IL-17 receptor deficiencies. IL-17 receptor deficiencies are linked with Staphylococcus aureus and Candida skin infections and also with bacterial respiratory infections⁸² .

Moving beyond M72/AS01

The demonstrated vaccine efficacy of M72/AS01 of 50% supports a critical role for robust Th1 CD4 + T cell responses in protection against TB but also highlights the need for additional vaccine components to broaden the immune response and further improve protection against disease. Vaccines currently in clinical development include multi-antigen vaccines targeting Ag85 complex proteins, PE proteins, Esp and ESAT-6 proteins, and others, formulated as live-attenuated vaccines, RNA, viral vector or adjuvanted vaccines (Tables 1 and 3). Adjuvants under investigation include TLR4 agonists (glucopyranosyl lipid adjuvant [GLA]), TLR7 agonists, TLR9 agonists (CpG2006 or central unmethylated CG dinucleotide plus flanking regions), Th1/Th17 adjuvants (targeting Mincle), and Th2 adjuvants (aluminum hydroxide) (Table 3).

Table 3.

Overview of Mtb preventative vaccines currently in early clinical development (https://newtbvaccines.org/tb-vaccine-pipeline/clinical-phase/)

Name (sponsors)	Antigen component	Design	Features of the immune response	Target indication	Target population	Phase	Reference
H107e/CAF®10b (Serum Statens Institute)	PPE68, ESAT-6, EspI, EspC, EspA, MPT64, MPT70, and MPT83	Multi-antigen with liposomal adjuvant containing Mincle (Th1/Th17) and CpG2006 (TLR9 agonist)	Th1/Th17 responses in mice and non-human primates	Prevention of infection and disease	Adults	1	Woodworth et al., 202492
BNT164a1 and BNT164b1 (BioNTech SE)	M72, VapB47 (virulence associated proteins B and C), hypoxic response protein 1, resuscitation-promoting factors A and D, ESAT-6, Antigen 85 A, heparin-binding hemagglutinin	RNA	CD4 + , CD8 + T cell responses, reduced bacterial load in mice	Disease prevention	Adolescents, adults	1b/2a	Ates, 202584-
ChAdOx1.85A (University of Oxford)	Antigen 85A	Adenoviral and modified vaccinia virus Ankara vectors	Polyfunctional CD4+ and CD8 + T cell responses, IgG anti-antigen 85 A responses in humans	Prevention of infection and disease	All ages	2a	Wilkie et al., 2020103
AEC/BC02 (Anhui Zhifei Longcom Biopharmaceutical Co., Ltd.)	Antigen 85B, ESAT6-CFP10	BC02 adjuvant: BCG-derived unmethylated CpG DNA fragments, Al(OH)3	Antigen-specific Th1responses in mice. Disease prevention in guinea pigs	Disease prevention	Adults	2a (currently suspendeda)	Lu et al., 2015138
QTP101 (Quratis, NIAID/NIH)	ID93 (Rv2608, Rv3619, Rv3620 [ESX protein family], and Rv1813 [ESX protein family])	Fusion protein with TLR4 agonist GLA	Th1 responses in humans. No CD8 + T cell responses	Disease prevention	Adolescents, adults	2a	Choi et al., 2023139
TB/FLU-05E (Smorodintsev Research Institute of Influenza)	TB10.4 (ESAT-6 protein family) and HspX (latency-associated antigen)	Recombinant attenuated influenza vector (Flu/THSP)	IFNγ-producing CD8 + T cells in mice. BCG then TB/FLU-05E induced antigen-specific effector and central memory CD4+ and CD8 + T cells, protection against severe disease in mice.	Prevention of infection	All ages	2a	Vasilyev et al., 2021140
DAR-901 (Dartmouth, St. Louis University)	-	Inactivated whole cell vaccine	IFNγ responses and anti- lipoarabinomannan antibodies in humans. Failed to prevent TB infection in BCG-immunized adolescents	Disease prevention	Adolescents, adults	2b	Munseri et al., 2020141, von Reyn et al., 2017142

BCG Bacille Calmette-Guérin, GcP glucan-chitin particles, GLA glucopyranosyl lipid adjuvant, IFNγ interferon-gamma, Mtb Mycobacterium tuberculosis, TB tuberculosis, TLR toll-like receptor.

^astated reason for suspension: It is planned to restart the clinical study after the completion of the clinical trial protocol modification (https://clinicaltrials.gov/study/NCT05284812).

Other potential immune mechanisms that could be leveraged by vaccination are the induction of CD8 + T cells, unconventional γδ, CD1, and MAIT cells, antibodies, and B cells, and the use of alternative vaccine technologies such as glycoconjugate vaccines. Live-attenuated vaccines can induce many of these immune mechanisms and have other potential benefits such as inducing longer-lasting immune stimulation than subunit vaccines, with potential activation of pattern recognition receptors to modulate the innate immune response. However, their development is challenged by the need to balance their various mechanisms in appropriate proportions to avoid unintended immune inhibition or impacts on efficacy⁸³.

BNT164 (BioTech SE) contains four RNA constructs expressing M72, VapB47(virulence associated proteins B and C), hypoxic response protein 1, resuscitation-promoting factor (Rpf) A and D, ESAT-6, Ag85A, and heparin-binding hemagglutinin⁸⁴. In mice, BNT164 candidates elicited CD4+ and CD8 + T cell responses, IgG responses against all vaccine antigens except ESAT6 and RpfD, and reduced the bacterial burden after a low-dose Mtb aerosol challenge⁸⁴. Two BNT164 candidates are currently in Phase 1b/2a (NCT05547464).

Induction of Th17

Polymorphisms in Th17-related genes impact susceptibility to TB and disease severity, and Mtb resisters have enriched Th17 T cell responses^53,85. The exact role by which IL-17 controls Mtb remains elusive, and like other aspects of the host immune response, the activities and effects of Th17-producing cells differ according to the stage of TB infection^86,87. During primary infection, Th17 cells acting at the lung epithelium secrete chemokines to recruit neutrophils to the site, promoting phagocytosis and bacterial killing. Chronic infection is associated with reduced production of Th17 cytokines, whereas excess IL17 is associated with reactivation and pulmonary inflammation⁸⁸. In the murine model, Th17 cytokine production was independent of IFNγ production, and was actively suppressed by Mtb using ESX-1, favoring a Th1 response⁵². In view of evidence of the role of Th17 responses in mycobacteria control, several second-generation vaccine candidates include Th17 adjuvants in efforts to improve protective efficacy.

Adjuvants that target Mincle are potent inducers of Th1 and/or Th17 in animal models, but appear to be less efficient in inducing Th17 in humans^89,90. The Th17 polarization capability of Mincle agonists increased in mice when administered with silica nanoparticles⁹¹. Coadministration of Mincle and a TLR9 agonist (CpG2006) given as the adjuvant CAF® (cationic adjuvant formulation)10b, improved Th17 responses in mice and non-human primates⁹². CAF®10b combined with the H107e fusion protein is a vaccine candidate being developed by the Statens Serum Institut. H107e is comprised of PPE68, ESAT-6, EspI, EspC, EspA, MPT64, MPT70, and MPT83, modified to improve protein expression, stability, and remove cross-reactivity with BCG^93,94. In mice, coadministration of BCG and a H107 prototype had synergistic effects on the immune response, inducing a broad repertoire of T cells, including fewer Th1-differentiated T cells, less IFNγ expression, and increased Th17 responses⁹³. H107e and BCG given to BCG-vaccinated mice provided significantly higher long term protection against TB compared to BCG or H107e alone⁹⁴. H107e/CAF10b is currently in Phase 1 (NCT06050356).

A critical aspect will be the induction of tissue-resident CD4 + /CD8+ memory T cells beyond primary and effector immunity. Mtb-specific IL-17-producing tissue resident cells subsets have been identified in human lungs from individuals with TB and their induction may be a key target for TB vaccines⁹⁵. However, tissue-resident memory cell induction may require the use of alternative immunization routes (intranasal, intradermal, or epidermal), beyond traditional systemic immunization⁹⁶.

Induction of CD8 + T cells

Animal studies suggest that CD8 + T cells and their products, granulysin, granzyme B, and perforin are required for immunity against Mtb, and that CD8 + T cell depletion weakens early control of Mtb and results in an increased bacterial burden during chronic infection^97–99. Studies in mice have also shown that adaptive CD8 + T cells are critical in vaccine-mediated protection against TB reactivation, undergoing re-positioning to actively contain Mtb in lymphatics and infected lymph nodes¹⁰⁰. Activation of conventional CD8 + T cells depends on antigen processing by the MHC-I pathway; however, relatively few Mtb antigens recognized by restricted T cells have been identified¹⁰¹. Another study showed that the immune phenotype of T cells induced by vaccination differs from that induced by infection, and cooperative interplay between two normally opposing transcription programs (T-bet and FOXO1/TCF1) is needed for robust CD8 + T cell responses after vaccination. The failure of many vaccines to induce detectable CD8 + T cell responses may be in part due to conditions of excessive or prolonged post-vaccination inflammation, such as that induced by emulsion or precipitate-based vaccines, under which this cooperative relationship breaks down¹⁰².

Efficacious vaccines to prevent TB infection or reactivation will need to be able to harness CD8 + T cells. Viral vector vaccines, live-attenuated vaccines, RNA vaccines, live attenuated intracellular bacteria, dendritic cell vaccines, and novel adjuvants (such as TLR7 agonists) are all potential strategies to induce CD8 + T cell recruitment^98,99. A chimpanzee adenoviral vector (ChAdOx1) and modified vaccinia Ankara virus (MVA) vaccine expressing Ag85A (ChAdOx1.85A + MVA85A) was investigated in BCG-vaccinated adults in the United Kingdom. ChAdOx1 induced polyfunctional Ag85A-specific CD4+ and CD8 + T cells, and Ag85A-specific IgG that were boosted by MVA85A, but this did not translate into vaccine efficacy^103,104. TLR7 is expressed on mucosa and promotes innate anti-virus responses with production of IFN, TNF, IL-1 and IL-6, and regulation of CD8 + T cell responses. A TLR7 agonist (gardiquimod) nanoparticle adjuvant administered to mice enhanced and broadened B cell responses and induced strong antigen-specific CD8 + T cell responses, suggesting that CD8 + T cell recruitment may be achievable using subunit vaccines combined with an TLR7 agonist adjuvant¹⁰⁵.

Unconventional innate-like CD8 + T cells with roles in Mtb protection are as γδ T cells, MAIT cells, HLA-E-restricted T cells, and lipid-reactive (CD1-restricted) T cells^106,107. CD8+ γδ T cells have innate and adaptive immune functions. Individuals with latent TB appear to have an expanded population of CD8+ γδ cells with enhanced cytokine release and cytolytic responses, and strong antibody-dependent cell-mediated cytotoxicity (ADCC), supporting a role for antibodies in Mtb protection¹⁰⁸. MAIT cells recognize antigens via MHC I-related molecules, have cytotoxic effects and produce cytokines including IFNγ, TNF and IL-17, making them important in the innate responses to intracellular pathogens^106,107. Nonconventional CD8 + T cell subsets can react more quickly to infection than conventional T cells, and there is evidence in animal models of their importance in early Mtb clearance. However, much more needs to be known about the actions of these cell subsets before specific epitopes targeting these cells are included in TB vaccines¹⁰⁶.

CD1-restricted CD4 + T cells

Lipid-reactive (CD1-restricted) T cells, are activated by presentation of lipid antigens by CD1 molecules on dendritic cells and macrophages¹⁰⁹. There are four isoforms of CD1(a, b, c, d) with distinct distributions on dendritic cells, and distinct structural features that bind specific sets of lipids¹¹⁰. Structural analyses of CD1b antigen presentation complexes show lipid chains buried in CD1, with protrusion of the saccharide head part which is recognized by the T cell receptor^109,111.

Unlike MHC-restricted T cells, CD1-restricted T cell receptors are highly conserved¹¹². For example, trehalose monomycolate-specific T cells are detected in cord blood and in uninfected and BCG-naive blood donors, indicating that a pre-fabricated Mtb-specific CD4 + T cell subset exists in unrelated human populations¹⁰⁹.

CD1-restricted T cells variously secrete IFNγ, TNF, IL-2 and IL-17 when stimulated, depending on the glycolipid antigen¹¹². Functionally, these cells are thought to transition to the lung early on during infection, before MHC-II-restricted T cells, where they kill Mtb-infected cells and control bacterial replication. However, it is not yet known how CD1-restricted T cells are activated or the effects of antigen exposure on memory phenotype, cell proliferation or activation¹¹². The diversity of antigen receptors and effector functioning of CD1-restricted T cells is greater than innate cells but less than MHC-restricted T cells, placing them somewhere between classical innate and MHC-restricted T cells¹¹².

The presence of a conserved antigen-specific CD1-restricted CD4+ memory T cell subset in humans capable of reacting rapidly to antigen challenge is highly suggestive of a role of these cells in protection against Mtb. Further delineation of the function of these cells could support the investigation of glycolipid antigens in vaccines to leverage the activities of specific CD1-restricted CD4 + T cell populations. Additionally, the structural features of the CD1 antigen presentation complex could provide an opening to create T cell immunity to more conventional bacterial polysaccharides by linking them to lipids. Or perhaps, it may be feasible to find innovative ways to translate structural insights about CD1 glycolipid presentation into MHC-II presentation of saccharides towards CD4-T cell receptors to create more classical T cell responses to saccharide epitopes.

Antibodies and Fc receptors

The importance of T cells in protection against TB has led to a degree of under-recognition of the contribution of antibodies in protection against infection. A growing body of evidence supports a role for the humoral immune response in protection against TB¹¹³. Antibody signatures in resisters, individuals with latent TB, and those with active TB are markedly different^70,71,113.

In latent TB, antibodies have been implicated in opsonization, macrophage functionality, activation of inflammasome, and restriction of intracellular growth. Murine antibody transfer studies show that some IgG antibodies against arabinomannan and lipoarabinomannan from humans exposed to Mtb or with latent TB reduced the bacterial burden in infected mice¹¹⁴. In macaques, animals with preexisting arabinomannan-specific IgA in airways and plasma were more likely to develop latent TB than Mtb infection¹¹⁵. Controlled infection was also associated with increased IgG responses to Mtb proteins during early infection¹¹⁵.

Investigation of human antibodies to arabinomannan showed marked variability in binding specificity and functionality according to the stage of infection. Anti-arabinomannan antibodies from individuals who were BCG-vaccinated, Mtb exposed or who had latent TB (but not active TB), had increased opsonophagocytic activity, decreased Mtb intracellular growth rates, and enhanced fusion of phagosomes and lysosomes in Mtb-infected macrophages. Antibodies from these individuals conferred protection to Mtb-infected mice¹¹⁴.

Aside from antigen specificity, antibody function in TB is influenced by the Fc domain and its interactions with immune cells¹¹⁶. α-glucan is abundantly expressed on the Mtb capsule and α-glucan antibodies are found in high levels in patients with active or treated TB¹¹⁷. However, α-glucan-specific antibodies do not restrict Mtb growth in vitro¹¹⁸. Specific Fc profiles can drive restriction of Mtb, and antibodies targeting α-glucan with engineered Fc domains optimized to promote neutrophil survival and functioning, were able to drive neutrophil-dependent growth restriction of Mtb in vitro¹¹⁸. Additionally, serum from individuals with various stages of latent, active and treated TB that was able to control BCG growth in vitro was associated with higher phagocytic activity, suggesting an association between the presence of antibodies and Fc-mediated phagocytosis¹¹⁹. In view of the enormous heterogeneity of the antibody response and the diverse characteristics of the Fc domains, considerable work is needed to identify epitope specificity and Fc-mediated effects that could increase vaccine efficacy^116,120.

B cells

The role of B cells in TB control continues to be debated. Like many other parts of the immune response, both beneficial and detrimental effects have been attributed to B cells during TB exposure, infection and disease. Studies in mice suggest that marginal zone B cells that accumulate in the lungs and spleen during infection undergo changes in their effector functions, producing multiple cytokines with restrictive effects on infection and regulatory effects on innate and adaptive immune cells¹²¹.

Glycoconjugate vaccines

Glycoconjugate vaccines comprise a carbohydrate antigen conjugated to a protein carrier, and typically induce strong T cell-mediated antibody responses against the carbohydrate antigen, with long lasting B cell immune memory. The use of glycoconjugates in TB vaccines would allow the immune response to Mtb polysaccharides and glycolipids to go beyond T cell-independent responses, with the potential for improved functionality of antibodies and improved B cell memory. Efforts to develop conjugate vaccines against TB have explored the capsular polysaccharides lipoarabinomannan and arabinomannan, as well as a phenolic glycolipid (PGL-tb1) hapten, conjugated to a variety of potentially active protein carriers targeting TB (Ag85 complex), rotavirus (spike protein ΔVP8*), as well as tetanus toxoid and cross-reacting material from diphtheria toxoid (CRM₁₉₇), already used in licensed conjugate vaccines¹²². A recent study confirmed the potential of this approach¹²³. Given the abundance of α-glucan on Mtb capsule and recent work that allows their chemical synthesis¹²⁴, α-glucans could be a useful target for conjugate vaccines if antibody function can be optimized. Mtb conjugate vaccines have shown promising immunogenicity in mice but remain in preclinical development.

Challenges in the preclinical evaluation of new TB vaccines

As well as the lack of clear immune correlates of protection against TB infection and disease, a major barrier in the development of TB vaccines is the lack of animal models that are predictive of vaccine efficacy in humans. Preclinical evaluations have been conducted in mice, guinea pigs, rabbits, non-human primates, and cattle¹²⁵. These species differ in their susceptibility to infection and disease, the spectrum of organ involvement, and in different aspects of their immune response. Thus, while preclinical studies can provide useful information about safety, immunogenicity and efficacy, their applicability to the human context is limited. This dilemma was vividly demonstrated when the virus vector MVA85A vaccine failed to show efficacy in clinical trials despite robust humoral and cellular responses observed in preclinical evaluations^104,126. An pertinent recent systematic review of murine studies of investigational TB vaccines tested against BCG found that the level of protection achieved by the candidate vaccines was influenced by vaccine type, vaccination route, the type and duration of the Mtb challenge. C57BL/6 mice were the most frequently tested strain, although more recently available Collaborative Cross and Diversity Outbred mice strains may allow the assessment of genetic diversity more closely reflective of the human population¹²⁷.

Attempts are underway to improve current animal models and vaccine selection criteria¹²⁶. One strategy is to use an ultra-low challenge dose of Mtb in murine models (1 colony-forming unit instead of 100) based on evidence that low-dose challenge more closely replicates levels of human exposure and immune responses than conventional doses¹²⁸. Others include humanized murine models, zebrafish models, and three-dimensional in vitro cell culture models, namely organoids (tissue-engineered cell-based three-dimensional in vitro models) and spheroids (simpler single-cell cultures that aggregate in a spherical shape)^125,129. In vitro cell cultures have been explored for their ability to reproduce granulomas, either from PBMC collected from individuals at different disease stages incubated in a collagen-coated matrix, generation of multicellular lung tissue, or cell encapsulation with microspheres. Advanced technologies that amalgamate human cells and synthetic culture medium manufactured using microfabrication technologies could potentially produce an “organ on a chip” and be used to investigate complex physiological functions at the organ level¹²⁹. In vitro cell culture models could allow deeper assessment of CD4 + , CD8 + , and CD1-restricted T cell functions in TB, immune responses to glycoconjugates, functions of macrophages, epithelial cells, and dendritic cells in response to vaccination, and the elimination of Mtb driven by T cell immunity and antibodies. Despite the current absence of predictive models, the 50% efficacy of M72/AS01 observed in humans will serve as a benchmark for validation of future animal challenge models.

Where to next?

Mtb genomics and structural research has demonstrated that Mtb shares structural similarities with Gram-negative bacteria in terms of its cytoplasmic, outer membranes, and clear capsular structure, but differ in the dedication of a large proportion of the genome to lipid metabolism and the production of dozens of PE/PPE proteins involved in virulence, particularly macrophage-mediated killing mechanisms. The unique set of glycolipid structures within the cell membrane, folded in such a way as to increase the space they occupy, leads to impermeability. This impermeability necessitates construction of specialized type VII secretory systems that are involved in translocation of outer membrane proteins and a large and diverse set of secretory proteins involved in virulence. An array of adhesins allow binding of Mtb to numerous host receptors that can variously facilitate infection, granuloma formation, cavitation, and transmissibility. While knowledge about virulence mechanisms, immunomodulatory actions, proteins and lipids that are essential for Mtb survival is increasing, much still remains unknown about Mtb interactions with its host.

Research into Mtb immunity established that CD4 + T cell immunity is critical for control of Mtb disease. Consequently, most Mtb vaccine antigens were originally selected on their capacity to induce IFNγ in PBMC cultures from persons latently infected with Mtb. M72/AS01 is the first new TB vaccine that has showed a degree of protection against the development of pulmonary TB in individuals with evidence of TB infection, confirming that Th1 CD4 + T cells play a role in protection against active TB. Investigational vaccines currently in clinical development contain multiple protein antigens and a range of mainly Th1 adjuvants (Tables 1 and 3). Given the complexity of Mtb:host interactions and Mtb’s extraordinary immune-controlling capabilities, success of these vaccines over and above M72/AS01 seems unlikely without activation of other components of the immune response. Possible exceptions are H107e/CAF®10b which contains Mincle and a TLR9 agonist, shown to induce strong Th17 responses in animal models, and BNT164 that induces IFNγ, and CD4+ and CD8 + T cell responses in mice but whose capacity to induce IL17 or be modified to induce IL17 is not yet reported^84,92. At least four additional components of the immune system could be tapped to achieve a high efficacy vaccine against TB infection or disease. These are a CD4 + Th17 response, CD8 + T cells, inclusion of glycolipid antigens aiming for CD1-restricted CD4 + T cell responses, and glycoconjugates of arabinomannan and glucan aiming for improved antibody function and B cell memory.

RNA vaccines can induce CD4+ and CD8 + T cell responses, but it is less certain if they can successfully direct a Th17 response, and the use of glycoconjugates and glycolipids requires a subunit approach. Furthermore, achieving correct protein folding can be challenging, possibly more-so when multiple antigens are required. RNA technology alone is therefore unlikely to be able to meet the needs of the next generation of TB vaccines, whereas subunit approaches combined with appropriate adjuvants may prove more amenable to a multi-pronged immune activating approach. A novel nanoparticle platform successfully combined a lipid and protein (mycolic acid and Ag85B) and induced glycolipid- and peptide-specific T cells in vaccinated mice¹³⁰. Live attenuated vaccines also have the potential to induce a multi-pronged immune response.

Numerous questions remain: is a there a limit on the number of CD4 + T cell antigens able to be included in vaccines due to spatial competition for CD4 T cells reaching their various targets? What is the incremental improvement in efficacy with addition of a Th17 response? Can rare variable Mtb antigens be leveraged in vaccine design to induce Th17? How can we take advantage of the protective effects of rapid responder CD1-restricted T cells? Can we identify essential immune features for protection from resisters? Since most discovered CD4 T cell antigens are derived from Mtb’s secretome, are other target antigens critical for Mtb survival yet to be identified, such as those capable of inducing protective CD8 T cell responses? Despite large gaps in our understanding of Mtb, ongoing technical advances and research and the first positive proof-of-concept for a new TB vaccine, suggest that effective prevention of TB is within reach.

Author contributions

J.T.P. conceived the review and prepared the original draft.

Competing interests

The author declares no competing interests.