Immunology of tuberculosis

Author
Lee W Riley, MD
Section Editor
C Fordham von Reyn, MD
Deputy Editor
Elinor L Baron, MD, DTMH
Disclosures: Lee W Riley, MD Nothing to disclose. C Fordham von Reyn, MD Nothing to disclose. Elinor L Baron, MD,
DTMH Nothing to disclose.
Contributor disclosures are reviewed for conflicts of interest by the editorial group. When found, these are addressed by
vetting through a multi-level review process, and through requirements for references to be provided to support the content.
Appropriately referenced content is required of all authors and must conform to UpToDate standards of evidence.
Conflict of interest policy

All topics are updated as new evidence becomes available and our peer review process is complete.
Literature review current through: Dec 2015. | This topic last updated: Sep 30, 2015.
INTRODUCTION The human host serves as the only natural reservoir for Mycobacterium tuberculosis.
The ability of the organism to efficiently establish latent infection has enabled it to spread to nearly onethird of the world's population [1]. According to the 2014 World Health Organization report, an estimated 9
million people developed tuberculosis and 1.5 million died [2]. The progression from latent tuberculosis
infection to active disease remains poorly understood. Given the magnitude of the health problem and the
emergence of drug-resistant strains of the organism, a better understanding of the immunology of this
disease and the development of an effective vaccine are highly desirable. (See "Epidemiology of
tuberculosis".)
The immunology of M. tuberculosis will be reviewed here. The microbiology and pathogenesis of this
infection, including virulence factors, tropism for the lungs, and latency factors, are discussed separately.
(See "Natural history, microbiology, and pathogenesis of tuberculosis".)
IMMUNOLOGY The majority of individuals in the general population who become infected with M.
tuberculosis never develop clinical disease [3]. This demonstrates that the innate and adaptive immune
response of the host in controlling tuberculosis (TB) infection is effective. Mycobacterial and host factors
that adversely affect these two arms of the immune system contribute to latent tuberculosis infection
(LTBI) and active disease.
Host factors
Innate immunity The pathophysiology of innate immune response during first encounter of M.
tuberculosis with lung cells remains poorly characterized. In the average human alveolus, there are more
than 28,000 epithelial cells (pneumocytes) and about 50 macrophages [4,5]. Mouse studies have shown
that after about 14 days of infection, the predominant cell type infected with M. tuberculosis is the myeloid
dendritic cell rather than the alveolar macrophage [6]. Thus, during the very early phase of lung infection,
the interaction of M. tuberculosis with lung epithelial cells may affect later dendritic cell and alveolar
macrophage migration and ultimately clinical outcome. Little is known about what happens during this
early phase.
Once M. tuberculosis comes into contact with dendritic or alveolar macrophages, the interaction of these
cells with M. tuberculosis first involves recognition by these cells of microbe-associated molecular

patterns (MAMPs) by pattern recognition receptors (PRRs) located on the cell surface or in the cytosol [7].
Distinct sets of macrophage PRRs recognize distinct sets of MAMPs of M. tuberculosis. These PRRs
serve to trigger innate immune response against molecules recognized to be foreign to the host cell. The
recognition of M. tuberculosis by a group of PRRs called toll-like receptors (TLRs) triggers cell signal
transduction that induces a proinflammatory response that is supposed to control the infection [8].
However, M. tuberculosis has evolved to subvert these host responses for its own survival in the host.
Toll-like receptors TLR is a mammalian homologue of the toll receptor in the insect Drosophila that
plays a role in conferring immunity of the insect against yeast infections [9]. TLRs exhibit pattern
recognition of organism group characteristic molecules, such as the lipopolysaccharide of gram-negative
organisms, peptidoglycan of gram-positive organisms, double-stranded RNA of viruses, and lipoteichoic
acids of yeasts [10-12]. (See "Toll-like receptors: Roles in disease and therapy".)
TLR2 and TLR4 are important for the recognition of M. tuberculosis MAMPs [10,13]. M. tuberculosis is
killed by activation of the TLR2 by the bacterial lipoprotein (19-kD) in both mouse and human
macrophages [14]. However, the killing in the mouse macrophage was dependent on the intracellular
nitric oxide pathway, while in the human macrophage, it was independent of this pathway [14]. That is, the
human TLR2 activation by the 19-kD lipoprotein killed M. tuberculosis, but no nitric oxide production could
be demonstrated.
Further characterization of the mechanism of killing of M. tuberculosis in human macrophages has
identified that TLR1/2 activation up-regulates expression of the vitamin D receptor as well as vitamin D-1hydroxylase [15]. The expression of vitamin D receptorrelated genes leads to an increased expression of
cathelicidin, an antimicrobial peptide, which is then responsible for inhibition of growth of M. tuberculosis.
TLR2 and TLR4 require MyD88, an intracellular adapter protein required for inducing early innate immune
response to pathogens [16]. However, in murine macrophages, M. tuberculosis can activate macrophages
via an MyD88-independent pathway [17]. Furthermore, M. tuberculosis infection of MyD88deficient C57BL/6 mouse is fatal despite an intact adaptive immune response [18]. When the mice were
vaccinated with M. bovisBacillus Calmette-Gurin (BCG), they were protected from M
tuberculosis infection, indicating that the adaptive immune response functioned normally in this MyD88deficient mouse. The adaptive immune response was not sufficient to compensate for the innate immune
defect in unvaccinated mice, suggesting that the innate immune response plays a substantive role in
protection against tuberculosis, at least in the mouse model [18].
Genetic polymorphism found in TLR2/4 has been shown to affect infection outcomes in some populations.
TLR4 polymorphism among Asian Indian population has been reported to be associated with increased
severity of tuberculosis, while, in other populations, no such association has been found [19,20]. Such
differential host response may also be affected by M. tuberculosis strain differences. Human studies have
shown that some M. tuberculosis strains preferentially activate TLR2, whereas others activate TLR4 [21].
Evasion mechanisms After the tubercle bacterium gets taken up by macrophages, it uses several
strategies to evade early intracellular killing mechanisms inside these target cells. Some of the
mechanisms previously thought to contribute to these strategies include:
Resistance to reactive oxygen intermediates (ROIs)
Inhibition of phagosome-lysosome fusion
Inhibition of phagosome acidification

Escape from the phagosomal compartment into the cytoplasmic space

During the initial stages of infection, tubercle bacilli stimulate the migration of neutrophils and
mononuclear phagocytes to the site of infection. M. tuberculosis possesses a variety of products that
enable it to constitutively resist ROIs produced by these cells, which are usually toxic to other pathogens.
Lipoarabinomannan (LAM) serves as a scavenger for oxygen intermediates [22]. Entry of the organism
into macrophages via complement receptors (CR1 and CR3) does not stimulate ROI production [23,24].
Cyclopropanated mycolic acids of the cell wall may help the organism resist hydrogen peroxide [25].
Furthermore, mycolic acids can inhibit the expression of interleukin (IL)-12, monocyte chemotactic protein
1 (MCP-1), tumor necrosis factor (TNF)-alpha in a toll-like receptor 2 (TLR-2)dependent manner in
macrophages [26]. Thus, drugs targeted against mycolic acid biosynthesis can exert an antibacterial
effect early in the phase of M. tuberculosis infection or during transition from latent infection to active
disease, when mycolic acid synthesis is most active.
In Escherichia coli and Salmonella, the oxyR gene serves as a regulatory element in response to
oxidative stress. In contrast, the oxyR homologue in M. tuberculosis contains numerous deletions and is
believed to be nonfunctional [27,28]. The constitutive anti-ROI products of M. tuberculosis described
above may obviate the need for such a gene in this organism.
M. tuberculosis taken up by macrophages can interfere with phagosome maturation. Phagosome
maturation requires the conversion of Rab5 into GTP-bound Rab7 and the production of
phosphatidylinositol 3-phosphate (PI3P) in the phagosomal membrane [29]. Two secreted bacterial
products PtpA (a tyrosine phosphatase) and NdkA (a GTPase) are believed to be involved in Rab7
inactivation [30,31].
Fusion of lysosomes with phagosomes leads to the release of microbicidal lysosomal contents.
However, M. tuberculosis can inhibit phagosome-lysosome fusion inside murine peritoneal macrophages
more than one week after infection [32,33]. When the macrophages were observed only two hours after
infection in another study, 85 percent of the phagosomes containing M. tuberculosis or BCG strains were
fused, but the number of such phagosomes declined over time in cells infected with live H37Rv or Ra
strains of M. tuberculosis [34]. M. tuberculosis strains incubated with an antibody raised against BCG did
not inhibit fusion, but they remained viable inside macrophages [33]. Thus, while there is evidence that M.
tuberculosis can inhibit phagosome-lysosome fusion, the relevance of this mechanism in ensuring
intracellular survival is not clear.
Both M. avium and M. tuberculosis can inhibit vacuolar acidification [35,36]. They do this by selectively
excluding the proton-ATPase from the phagosome. While this inhibitory mechanism may be important
during the initial encounter of the organism with macrophages, it is not clear how much of a role this plays
in the ultimate intracellular fate of the pathogen. Macrophages activated by cytokines can acidify
phagosomes, but the acidification itself may not contribute to significant killing of mycobacteria inside
such cells [37].
M. tuberculosis is also known to interfere with antigen presentation by major histocompatibility complex
(MHC) class II molecules, which is important for priming CD4 T cells [38]. Such interference could lead to
resistance of M. tuberculosis to acquired immunity in the host and, hence, may contribute to bacterial
persistence, a characteristic feature of this organism. M. tuberculosis products suggested to inhibit
macrophage presentation of MHC class II molecules include lipoarabinomannan, 25 kDa glycoprotein,
and a 19 kDa lipoprotein [22,39,40].

The traditional view is that when M. tuberculosis is engulfed by macrophages or dendritic cells, it resides
inside the phagosomal compartment and that it does not escape into the cytosol. However, data suggest
that the organism can enter the cytosol and that this is dependent on two major secreted M.
tuberculosis proteins, early secretory antigenic target 6 (ESAT-6) and culture filtrate protein (CFP)-10,
encoded by the Esx-1 locus [41]. ESAT-6 has been shown to have a membranolytic effect in vitro [42,43].
In the cytosol, mycobacterial cell wall component N-glycolyl muramyl dipeptide is believed to induce the
production of type I interferons by binding to the intracellular pattern recognition receptor called nucleotide
oligomerization domain 2 (NOD2) and that the induction of type I interferons is dependent on intact Esx-1
system [44,45]. ESAT-6 may promote access to the cytosol of other M. tuberculosis immunostimulatory
products that activate the NLRP3 inflammasome complex that promotes caspase-I-dependent secretion
of type I interferons [46]. The effect of type I interferon production on outcome M. tuberculosis infection
outcome is not clear.
Reactive nitrogen intermediates Several studies have focused on the role of reactive nitrogen
intermediates (RNI) in clearing infection before acquired immunity develops. In mice, RNIs are the only
macrophage effector molecules associated with killing of M. tuberculosis. A murine macrophage cell line
expressing RNI but defective in ROI expression was able to kill a virulent strain of M. tuberculosis after
stimulation with interferon (IFN)-gamma [47].
Additional evidence for the importance of RNI in controlling M. tuberculosis infection in mice is illustrated
by the following observations:
Mortality in mice infected with M. tuberculosis is increased when inducible nitric oxide synthase
(NOS2 or iNOS), which leads to RNI expression, is blocked by nitric oxide synthase inhibitors [48].
IFN-gamma is a potent stimulus of NOS2 expression in mice. Mice with disrupted IFN-gamma
or NOS2 genes become highly susceptible to M. tuberculosis infections [49,50].
Human macrophages do not express NOS2 when exposed to IFN-gamma, and NOS2 expression in
peripheral blood-derived macrophages is difficult to demonstrate. However, several studies have provided
evidence for NOS2 expression by alveolar macrophages obtained from bronchial alveolar lavage (BAL)
fluid of patients with TB [51-53].
Thus, human NOS2 may require factors other than those that stimulate NOS2 expression in murine
macrophages. RNIs may play an important role in the initial (innate immunity) as well as the later control
(acquired immunity) of M. tuberculosis infection.
There is epidemiologic evidence supporting the importance of RNIs in TB control. A study in New York
City found that the most common drug-susceptible strain of M. tuberculosis circulating in the city (C strain)
during the early 1990s was resistant to RNIs generated in vitro and was associated with injection drug
use [54]. This strain was used to identify a novel gene called noxR1. Expression of this gene in
nonpathogenic strains of E. coli or M. smegmatis led to increased resistance to RNIs and ROIs in vitro
[55].
Another protein, alkyl hydroperoxidase subunit protein, expressed by gene ahpC of M. tuberculosis also
protects human cells from necrosis and apoptosis caused by RNIs [56]. AhpC is believed to detoxify
peroxynitrite, a potent oxidant generated from the reaction of nitric oxide and superoxide anion (O 2-) [57].
Peroxynitrite-induced damage has been suggested to be repaired by methionine sulfoxide reductase
encoded by msrA [58]. Another gene noxR3 of M. tuberculosis has been reported to protect Salmonella
typhimuriumagainst oxidative and nitrosative stress [59].

Interferon-gamma In human macrophages, the effect of IFN-gamma in killing M. tuberculosis has not
been definitively demonstrated. In one study, for example, cells preincubated with IFN-gamma enhanced
the intracellular proliferation of M. tuberculosis; in comparison, killing of another intracellular
organism, Leishmania, was increased [60]. However, the coadministration of IFN-gamma
with calcitriol (the most active metabolite of vitamin D), which causes monocytes to mature in vitro, leads
to intracellular killing ofM. tuberculosis [61]. Calcitriol independently stimulates NOS2 production and
suppresses growth of M. tuberculosis inside the human macrophage-like cell line, HL-60 [62].
Other cytokines In addition to the role that IFN-gamma plays in M. tuberculosis control described
above, other cytokines also contribute. One important cytokine is TNF-alpha. Its role in controlling latent
infection with M. tuberculosis in humans has been dramatically demonstrated by multiple reports of
reactivation TB occurring in individuals treated for rheumatoid arthritis or Crohn's disease with anti-TNFalpha (infliximab) or TNF-alpha receptor (etanercept) antibodies [63-65]. TNF-alpha is also important for
proper granuloma formation (see below). Persistently infected mice treated with neutralizing anti-TNFalpha antibody develop granuloma disorganization, which eventually kill the mice [66]. IFN-gamma and
IL-12 levels in these mice are not affected. Mice deficient in this cytokine or TNF-alpha receptor are highly
susceptible to M. tuberculosis infection [67,68].
The role of other cytokines in outcome after M. tuberculosis infection is less clear. IL-10 produced by
macrophages has an antiinflammatory effect. However, IL-10disrupted mice are not any more resistant
to M. tuberculosis infection than wild-type mice [69]. The role of another suppressive cytokine
transforming growth factor (TGF)-beta in infection outcome in mice or humans remains unresolved [70].
Nutrient deprivation Nutrient limitation by macrophages is another innate immunity-related tool for
control of M. tuberculosis. M. tuberculosis engulfed into a phagosomal compartment must access
nutrients and carbon source for its persistence. Macrophages limit access of glucose and lipids to M.
tuberculosis residing inside phagosomal compartments, but M. tuberculosis has developed a strategy to
utilize lipids derived from host cell's triacylglycerol and cholesterol, which accumulate in the phagosome
as lipid droplets [71]. Cells containing such lipid droplets are often referred to as foamy macrophages.
Vitamin D, however, can prevent M. tuberculosisinduced lipid droplet accumulation [72].
Adaptive immunity Macrophages as well as other phagocytic cells such as dendritic cells carry
phagocytized M. tuberculosis to draining lymph nodes where they present M. tuberculosis antigens to T
cells. The presentation of antigens to T cells induces cellular immune response.
Cellular immunity Two to six weeks after infection with M. tuberculosis, T cells that specifically
recognize M. tuberculosis antigens appear; this response can be demonstrated clinically by the
development of a delayed-type hypersensitivity (DTH) response to intradermally injected tuberculin or
purified protein derivative (PPD). One study of T cell responses in persistently anergic patients with
documented pulmonary TB demonstrated that T cells produced IL-10 but not IFN-gamma and failed to
proliferate in vitro following stimulation with PPD [73]. By contrast, T cells from PPD-positive patients
produced both IL-10 and IFN-gamma and displayed a dramatic proliferative response to PPD stimulation.
The DTH response per se does not correlate with protection against TB, since numerous BCG
vaccination trials have demonstrated that disease can occur in those who mount a DTH response [74]. As
a result, the protective T cell response must be distinguished from the T cell response associated with
DTH.

Interferon-gamma release assays have been developed; these are in vitro, whole blood-based tests to
measure T cell activation. The assays are an alternative to the tuberculin skin test (TST) for detection of
latent M. tuberculosis infection in human hosts [75-79]. The test measures interferon-gamma released
into blood from T cells when they are activated by M. tuberculosis antigens in vitro. The tests use antigens
specific to M. tuberculosis including ESAT-6 and CFP-10 [75,76]. These proteins are encoded by genes
located within the region of difference 1 (RD1) of the M. tuberculosis genome and are absent in vaccine
strain BCG or M. bovis. This enables the test to differentiate those latently infected with M.
tuberculosis from those vaccinated with BCG. (See "Interferon-gamma release assays for diagnosis of
latent tuberculosis infection".)
The importance of T cells in the protective immune response against TB was first demonstrated in mice;
adoptive transfer of T cells from BCG-immunized mice protected irradiated recipient mice from infection
[80,81]. Other animal studies showed that this protective response was mediated by CD4-bearing T cells
[82,83]. If it is assumed that the DTH response is mediated by CD4+ Th1 cells, the wide range of
protection (0 to 80 percent) demonstrated by numerous BCG trials suggest that CD4+ T cells are not
sufficient for protection and that other cells must be involved [74]. However, the greatly increased risk of
TB with HIV infection, in which CD4+ T cells become depleted, suggests that these cells are important for
protection against TB in humans. CD4+ T cells exert their effector function by producing IFN-gamma,
which activates macrophages. This response is important, particularly during early phase of an infection.
In one study, in CD4-disrupted mice, levels of IFN-gamma in the lungs, while diminished early in infection,
reached levels found in wild-type mice after about three weeks, suggesting that other cell types (CD8+
cells) can compensate for the decreased cytokine expression by CD4 T cells [84]. Finally, in addition to
the role of cytokines produced by CD4+ cells, apoptosis of infected cells by CD4+ T cells may contribute
to controlling infection. However, published reports on the role of apoptosis in M. tuberculosis infection
control remain equivocal [85,86].
Cytotoxic T lymphocytes (CTLs) have been implicated in protection against M. tuberculosis, and active
investigation into the details of this mechanism is ongoing.
Mice with disruption of the beta-2 microglobulin gene fail to control infection with a virulent strain
of M. tuberculosis (Erdman) despite having intact CD4+ and cytolytic gamma-delta T cells [87]. This
gene inhibits expression of functional class I MHC-I molecules, which is a feature of CD8+ CTLs.
Another study using different strains of gene-disrupted mice found that perforin contributed only
partially to the protective ability of CTLs, there were beta-2 microglobulin-dependent T cells that
exerted a protective effect distinct from CTLs, and transporter associated with antigen processing
(TAP) pathways were predominant in mediating protection against M. tuberculosis, but TAPindependent mechanisms also played a role [88].
Mice immunized with Mycobacterium vaccae can generate CD8+ T cells that express IFN-gamma
and that are lytic to macrophages infected with M. tuberculosis [89].
Live mycobacteria activate more CD8+ T cells than dead organisms or PPD [90].
Antigen-presenting pathways other than MHC class I or class II that stimulate this type of CTL response
have been explored. One such pathway, CD1-restricted CTL stimulation, involves MHC-like cell surface
molecules that process and present nonpeptide antigens to T cells [91]. In patients with active TB, two
types of T cells that recognize M. tuberculosis lipid and lipoglycan antigens presented by CD1b-bearing
cells were found, "double negative" (DN) CD4-/CD8- T cells and CD8+ T cells [86]. These cells were both
capable of lysing macrophages (CD-1 bearing cells) infected with M. tuberculosis [86].

However, cell lysis would not necessarily lead to protection in the absence of bacterial killing. CTLmediated cell lysis involves two pathways: a degranulation pathway that generates perforin and
granzymes, and a Fas-FasL-dependent pathway that induces apoptosis of the target cell [92,93]. Studies
with perforin gene-disrupted mice showed that this pathway was not essential for early protection
against M. tuberculosis infection [94,95]. Another study showed that perforin-disrupted mice eventually did
succumb to infection at a later time point, suggesting that the protection is partially dependent on perforin
[88]. Coculturing DN and CD8+ T cell lines with CD1 cells infected with M. tuberculosis revealed that DN
CD1-restricted cells had no effect on the viability of the organism, while CD8+ CD1-restricted T cells
reduced the number of colony forming units (CFU) by 35 to 50 percent [86].
This bacterial killing is mediated by granulysin, a protein found in the granules of human CTLs and natural
killer (NK) cells, but not in murine cells [80]. Granulysin is present in CD1-restricted CD8+ T cells but not
in CD1-restricted DN T cells. Granulysin, a member of the saposin-like protein family, induce blister-like
lesions on the surface of M. tuberculosis [96].
There also appears to be a role for CD8+ alpha beta TCR+ cells, which recognize antigens bound to
MHC class I. In one study, for example, two human TCR alpha beta+ CD8+ T cell lines specific for M.
tuberculosis antigens recognized lipid antigens when presented by CD1a or CD1c antigen-presenting
cells and displayed both cytotoxicity and cytokine responses [97].
5'-adendomsinephosphosulfate reductase (CysH), an enzyme essential for the production of reduced
sulfur-containing metabolites, has been shown to be important for M. tuberculosis during the chronic
infection phase [98]. Resistance to nitrosative and oxidative stress (RNI and ROI) may be the mechanism
of this protection [98].
Granuloma formation In addition to the specific cell-mediated protective response involved in M.
tuberculosis elimination, granuloma formation is an important mechanism of the host to control infection.
Granuloma formation requires balanced expression of cytokines and chemokines, including RANTES,
M1P1-alpha, M1P1-beta, MCP-1, MCP-3, MCP-5, and IP10 [99,100]. Chemokine receptors also
determine proper formation of granulomas and, with M. tuberculosis infection, the expression of CCR5
(receptor for RANTES, M1P1-alpha, and M1P1-beta) increases in macrophages [101]. CCR2-disrupted
mice are more susceptible to M. tuberculosis than the wild-type mice [102]. CCR2 is a receptor for MCP1, -3, and -5. MCP-1disrupted mice, however, are not susceptible [103].
M. tuberculosis disrupted in an operon called mce1 was shown to become hypervirulent in the mouse
model of TB [104]. This hypervirulence was associated with aberrant proinflammatory cell migration to the
site of infection in the mouse lungs and poor granuloma formation. The mce1 operon mutant was unable
to stimulate MCP-1 and TNF-alpha expression by murine macrophages [104]. Disruption of
the mce1 operon may alter the cell wall architecture and thus an effect on host granulomatous response
[105]. Indeed, an untargeted metabolomics analysis of the cell wall lipids of the mce1 operon mutant
showed more than 400 lipids that were significantly altered from those of wild-type M. tuberculosis [106],
indicating that changes in cell wall lipid contents affect host immune response. The cma2A mutant, noted
above, induced larger sized granulomas in mouse lungs than did the wild-type M. tuberculosis [107].
These observations suggested that alteration in the cell wall lipid composition or its remodeling can
greatly affect host immune response and that a certain level of proinflammatory response induced by M.
tuberculosis itself is necessary for proper granuloma formation that is protective both to the host and the
bacterium. Thus, the granuloma is not only a host protective factor but may serve as a shelter constructed

by the tubercle bacterium itself for its long-term survival in the host and that bacterial cell wall lipids may
play an important role in this host-pathogen homeostatic interaction.
Humoral immunity The role of the humoral immune response in protection against M. tuberculosis is
controversial. The range of findings can be illustrated by the following observations:
Passive transfer experiments of serum from BCG-vaccinated animals or M. tuberculosis-infected
animals and humans to other animals have provided conflicting evidence of protection [108].
Antibodies against a variety of mycobacterial antigens, including lipid and carbohydrate products,
can be demonstrated in both asymptomatic PPD-positive persons and in patients who develop
active disease.
M. tuberculosis strains incubated with an antibody raised against BCG promoted phagosomelysosome fusion, but the viability of such strains inside macrophages was not affected [33].
Transgenic mice unable to make immunoglobulin M become more susceptible to M.
tuberculosis [109].
Thus, unequivocal evidence for the protective role of humoral immunity against TB does not exist. In the
absence of detailed studies, it is difficult to conclude how much of the protective response against TB, if
any, is mediated by humoral immunity.
Pathogen factors
Mycobacterial lipids M. tuberculosis lipid products have been associated with tuberculosis
pathogenesis since 1947, when Middlebrook suggested that a growth morphology called "cording" was
associated with virulent tubercle bacilli [110]. Subsequently, the toxic effect of petroleum ether extractable
"cord factor" from M. tuberculosis was characterized in a murine model [111]. This cord factor was
eventually identified as glycolipid trehalose dimycolate (TDM) [112]. Although studies have shown that
"cording" is not necessarily restricted to virulent Mycobacteria, structural changes in TDM elicit distinct
mammalian host immune responses.
TDM is comprised of mycolic acids covalently bound to trehalose. Mycolic acids are beta-hydroxyl fatty
acids with an alpha-alkyl side chain; the M. tuberculosis cell wall contains three classes of this fatty acid:
alpha-, keto-, and methoxy-mycolates [113]. Alpha-mycolic acid has two cyclopropane rings, which are in
the cis configuration, while the keto- and methoxymycolates have one ring in either cis or trans
configuration [114]. The cyclopropanation status of mycolic acids is important for pathogenesis in the
mouse model as illustrated by the following findings:
M. tuberculosis strains lacking keto-mycolates grow poorly inside THP-1 cells [115].
Absence of keto- or methoxy-mycolates in M. tuberculosis leads to attenuation of infection in a
murine model [116].
M. tuberculosis strains disrupted in the cyclopropane synthase gene, pcaA, are attenuated in
mice. pcaA is required for cord formation and for the synthesis of the proximal cyclopropane ring of
alpha-mycolic acid. TDM prepared from this mutant is less inflammatory [117].
M. tuberculosis disrupted in cyclopropane-mycolic acid synthase 2 (cmaA2), a gene responsible
for the trans-cyclopropanation of mycolic acid in the cell wall, is hypervirulent in BALB/c mice [107].
Lipid extracted from this mutant is hyper-proinflammatory.
In addition, phthiocerol dimycocerosates (PDIM) are important for growth of M. tuberculosis in mouse
lungs (although not in mouse spleen or liver) [118]. PDIM is thought to protect M. tuberculosis against the

bactericidal activity of reactive nitrogen intermediates [119]. A mutation in a lipoprotein (LppX), needed for
the export of PDIM, led to attenuation of the mutant in a mouse infection model [120].
M. tuberculosis grown in detergent-free medium produces biofilm and extracellular matrix of the biofilm
has been shown to include mycolic acids [121]. Whether M. tuberculosis makes biofilm in vivo during its
natural course of infection is not known. The mce1operon of wild-type M. tuberculosis is repressed during
the early phase of infection in mice [122] and a strain of M. tuberculosis disrupted in this operon overexpresses mycolic acids [106,123,124]. Thus, it is conceivable that M. tuberculosis makes biofilm in vivo.
Mycolic acids have been shown to inhibit proinflammatory cytokine expression by human alveolar
epithelial cells (A549) and murine RAW cells [21]. Furthermore, the mce1 operon mutant shows reduced
expression of mmpL8, mmpL10, stf0, pks2, and papA2 genes involved in transport and metabolism of
lipids associated with proinflammatory response [106].
These observations suggest that M. tuberculosis cell wall lipids play an important role in the early
interaction of this organism with the host immune response and that the cell wall lipids ultimately
determine clinical outcomes. Clinical isolates of M. tuberculosis with changes in lipid composition can
affect the distribution of organisms in a community, as illustrated by a large outbreak of tuberculosis in
Tennessee and Kentucky in 1994 through 1996 caused by strain CDC1551 [125]. CDC1551 produced an
enhanced proinflammatory cytokine response in mice, which was traced to changes in its cell wall lipids
[126].
Another outbreak-associated strain, HN878, expresses a highly active lipid species called phenolic
glycolipid (PGL), not detected in M. tuberculosis strain CDC1551 or laboratory strain H37Rv [127]. PGL
inhibits proinflammatory cytokine release by macrophages [127,128], but a later study showed that PGL
itself does not mediate hypervirulence [129].
VACCINE DEVELOPMENT The World Health Organization guidelines recommend that Bacille
Calmette-Guerin (BCG) be administered as part of routine newborn immunization to protect against
tuberculosis (TB), although its efficacy is limited. The development of new vaccines has been hampered
by lack of good understanding of correlates of protection and difficulty identifying surrogate endpoints of
protective immunity against active disease that can be used in vaccine trials. Experimental animal models
have not been reliable in predicting vaccine protection against human infection. (See "BCG vaccination".)
Despite these obstacles, there are many candidate vaccines undergoing clinical trials [130]. Trial results
for the modified vaccinia virus Ankara-expressing mycobacterial antigen 85A (MVA85A) were
disappointing [131-133]. The failed trial in South Africa unmasked a major deficiency in our understanding
of human immunological correlates of protection against tuberculosis and the limitation of murine models
to predict these correlates [133]. Other candidates include AERAS-402/Crucell Ad35, based on
replication-deficient adenovirus expressing M. tuberculosis antigen 85A, 85B, and TB10.4 [134],
recombinant BCG expressing M. tuberculosis antigens (VPM1002, rBCG30), recombinant subunit or
fusion proteins (M72, Hybrid-1+IC31; Hybrid-1+CAF01, Hyvac4, AERAS-404 + IC-31), fragmented M.
tuberculosis cells (RUTI), and non-tuberculosis attenuated mycobacteria (M. vaccae) [135]. Many of these
candidates advanced to human trials after they were shown in nonhuman animal models to confer better
protection than BCG.
Many of the new vaccines are designed to be given through "heterologous prime-boost strategy" in which
the new vaccine expressing M. tuberculosis antigens is given as a booster in those primed with BCG or a
recombinant BCG vaccine. The M. tuberculosisantigens are delivered in a viral or some non-

mycobacterial vector. Other strategies for TB vaccine application include therapeutic vaccine or
immunotherapy, designed to be administered to those already infected with M. tuberculosis to prevent
them from progressing to active disease, and adjunctive vaccine designed to be given together with a
short-course drug regimen [136,137]. The idea of an adjunctive vaccine is to reduce the bacterial burden
with standard drugs to a level that can then be cleared by vaccine-induced immunity, which may facilitate
shorter-course drug regimens [137].
Major obstacles to development of safe, effective, and affordable TB vaccines remain. These include our
need to better understand the immunology of TB, especially the component associated with protection
from a vaccine and natural infection; identification of biomarkers for vaccine efficacy and protection;
selection of clinical trial sites with appropriate target populations; regulatory and ethical constraints
associated with clinical trials; and limited resources and funding available for vaccine development,
clinical trials, and implementation.
SUMMARY
The majority of individuals in the general population who become infected with Mycobacterium
tuberculosis never develop clinical disease; this demonstrates that the innate and adaptive immune
response of the host in controlling tuberculosis (TB) infection is effective. Factors that adversely
affect the immune system contribute to development of latent tuberculosis infection and active
disease. (See 'Immunology' above.)
Components of innate immunity thought to be involved with defense against TB infection include
killing via reactive nitrogen intermediates, interferon-gamma and other cytokines, use of toll-like
receptors for recognition of M. tuberculosis molecules, and other factors. (See 'Innate
immunity' above.)
The role of cellular immunity is important in the protective response against TB, and diagnostic
tools have been developed based upon this cellular response including the tuberculin skin
test (TST) and the interferon-gamma release assay (IGRA). The TST measures the delayed-type
hypersensitivity response to intradermally injected tuberculin. The IGRA is an in vitro, whole bloodbased test that measures T cell activation by M. tuberculosis antigens. The role of humoral immunity
in the protective response against TB is uncertain. (See 'Adaptive immunity' above.)
Granuloma formation is also an important mechanism of the host to control infection; it requires
expression of cytokines and chemokines. In addition, it may play a role in protecting the tubercle
bacilli for its long-term survival in the host. (See 'Granuloma formation' above.)
M. tuberculosis cell wall lipids play an important role in the early interaction of the organism with
the host immune response and determination of clinical outcome. (See 'Mycobacterial lipids' above.)
The efficacy of BCG vaccination is limited; the development of new vaccines has been hampered
by lack of good understanding of correlates of protection and difficulty identifying surrogate
endpoints of protective immunity against active disease that can be used in vaccine trials. In
addition, experimental animal models have not been reliable in predicting vaccine protection against
human infection. (See 'Vaccine development' above.)