Survival perspectives from the world's most successful pathogen, Mycobacterium tuberculosis.

Abstract

Studying defined mutants of Mycobacterium tuberculosis in the mouse model of infection has led to the discovery of attenuated mutants that fall into several phenotypic classes. These mutants are categorized by their growth characteristics compared with those of wild-type M. tuberculosis, and include severe growth in vivo mutants, growth in vivo mutants, persistence mutants, pathology mutants and dissemination mutants. Here, examples of each of these mutant phenotypes are described and classified accordingly. Defining the importance of mycobacterial gene products responsible for in vivo growth, persistence and the induction of immunopathology will lead to a greater understanding of the host-pathogen interaction and potentially to new antimycobacterial treatment options.

Full text

Page 1

REVIEW
NATURE IMMUNOLOGY
VOLUME 4 NUMBER 10 OCTOBER 2003
949
Tuberculosis remains the leading cause of mortality due to a bacterial
pathogen, and infects approximately 32% of the world’s human popu-
lation. Tuberculosis was declared a global health emergency by the
World Health Organization in 1993 and annually causes approxi-
mately 1.8 million deaths1, which translates into 4,931 deaths per day
(the equivalent of more than 10 Boeing 747 airplanes per day full of
people succumbing to this global health crisis). The emergence of
multi-drug-resistant strains of Mycobacterium tuberculosis, the
causative agent of tuberculosis, and the susceptibility of patients
infected with human immunodeficiency virus to tuberculosis have
fuelled the spread of the disease. M. tuberculosis was isolated by Koch,
who conclusively proved the causal association between the bacillus
and disease. The availability of the complete genome sequence of
M. tuberculosis2and advances in techniques for gene transfer and dis-
ruption3–5have led to the discovery of several survival strategies used
by this complex human pathogen.
Antituberculous immunity
The main route of infection for the tubercle bacillus is the respiratory
tract, where the bacteria are inhaled in airborne droplets that proceed
distally to the lung to establish an infection. After entering the lung,
the first cell type encountered by the bacteria is the alveolar
macrophage, which has the microbicidal armory to destroy most
potential invaders. However, the tubercle bacillus has the extraordi-
nary ability to persist and even to replicate in this extremely hostile
environment, where most other pathogens perish.
Tuberculous immunity relies mainly on cell-mediated immunity
rather than humoral immunity. This immunity specificity is demon-
strated by the considerably increased risk of tuberculosis in patients
with reduced cell-mediated immunity, such as those infected with
human immunodeficiency virus or individuals undergoing immuno-
suppressive therapy, compared with patients with defective humoral
immunity, such as those with multiple myeloma, who show no
increased predisposition to tuberculosis6. The macrophage has multi-
ple functions in tuberculosis, including antigen processing and pre-
sentation, transport to deeper tissues and hence dissemination of
infection7, and effector cell functions. However, at least in the mouse
model of infection, M. tuberculosis has the ability to evade the
onslaught of innate immunity, as virulent bacilli replicate exponen-
tially within mouse lung during the first few weeks after infection, yet
after the onset of acquired immunity, the growth of the bacilli plateaus
(Fig. 1). Protective acquired immunity to M. tuberculosis is dominated
by CD4+and CD8+T cells with the T helper type 1 cytokine profile8.
The idea of the importance of the key T helper type 1 cytokines inter-
feron-γ (IFN-γ) and interleukin 12 is supported by the high suscepti-
bility to tuberculosis of individuals with defects in interleukin 12, its
receptor and the IFN-γ receptor9–11.
M. tuberculosis can survive and replicate intracellularly, where most
other invaders perish (Fig. 2). It is generally accepted that the tubercle
bacilli’s main niche is the host macrophage, and M. tuberculosis seems
to have evolved effective mechanisms to survive most macrophage
effector functions. These defensive mechanisms include the inhibition
of phagosome-lysosome fusion12, where M. tuberculosis was found to
retard phagosomal maturation13, and the use of complement recep-
tors 1 and 3 for cell entry, which do not trigger the oxidative burst14,15.
However, M. tuberculosis can enter the macrophage through multiple
receptors without adversely effecting its survival16. Other effector
mechanisms include the production of catalase and superoxide
Suzanne M. Hingley-Wilson, Vasan K. Sambandamurthy and William R. Jacobs
are in the Howard Hughes Medical Institute, Albert Einstein College of Medicine,
Bronx, New York 10461, USA. Correspondence should be addressed to W.R.J.
(jacobsw@hhmi.org).
Published online 26 September 2003; doi:10.1038/ni981
Survival perspectives from the
world’s most successful pathogen,
Mycobacterium tuberculosis
Suzanne M Hingley-Wilson, Vasan K Sambandamurthy & William R Jacobs, Jr
Studying defined mutants of Mycobacterium tuberculosis in the mouse model of infection has led to the discovery of attenuated
mutants that fall into several phenotypic classes. These mutants are categorized by their growth characteristics compared with
those of wild-type M. tuberculosis, and include severe growth in vivo mutants, growth in vivo mutants, persistence mutants,
pathology mutants and dissemination mutants. Here, examples of each of these mutant phenotypes are described and classified
accordingly. Defining the importance of mycobacterial gene products responsible for in vivo growth, persistence and the
induction of immunopathology will lead to a greater understanding of the host-pathogen interaction and potentially to new
antimycobacterial treatment options.
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 2

REVIEW
dismutase, which are capable of degrading reactive oxygen intermedi-
ates17,18. The tubercle bacilli have even been reported to down-regu-
late some modulators of host immunity such as interleukin 12 (refs.
19,20), major histocompatibility complex class II (ref. 21), IFN-γ-
mediated activation of the macrophage22, the IFN-γ-induced gene
gamma.1 (ref. 23) and host cell apoptosis24.
Classification of in vivo growth mutants
Studies of defined constructed mutants of M. tuberculosisin the mouse
model of infection have led to the classification of attenuated mutants
into several phenotypic classes (Table 1). These mutants are catego-
rized by their growth characteristics, such as severe growth in vivo
(sgiv) mutants, which show a very severe reduction in colony-forming
units with time; growth in vivo (giv) mutants, which do not grow as
robustly as wild-type M. tuberculosis in the lungs of immunocompe-
tent mice, yet still grow better than sgiv mutants; persistence (per)
mutants, which fail to grow or persist after the onset of acquired
immunity; and mutants that have the same growth characteristics as
per mutants, but show altered pathology (pat) compared with that of
wild-type M. tuberculosis (Fig. 3).
Attenuated mutants of M. tuberculosis are of interest as potential
vaccine candidates, with each of the different classes described above
providing a different approach to vaccine development and potential
disease prevention. A new vaccine for tuberculosis is urgently needed,
as the present vaccine, Mycobacterium bovis bacille Calmette-Guerin
(M. bovis BCG), shows vast differences in efficacy, with one study in
India reporting an efficacy of 0% (ref. 25). In addition, despite being
administered to over three billion people26, the current vaccine has
failed to curtail the spread of the disease. It would also be more advan-
tageous to have a vaccine capable of blocking the initial infection, in
contrast to BCG, which mainly protects against the uncontrolled repli-
cation and dissemination of the bacilli from the primary foci of infec-
tion to the rest of the lungs and body27. One approach for improving
BCG as a vaccine vector would be to add antigens or functions that
enhance its immunogenicity. For example, the ability of BCG to elicit
CD8+T cell responses was improved by the addition of the ability to
secrete listeriolysin from Listeria monocytogenes28. The immunogenic-
ity of BCG has also been improved by the overexpression of key
antigens29,30, or the addition of the primary attenuating region of
deletion of BCG, RD1, back to BCG31.
A chief goal of new tuberculosis vaccine development is to have a
vaccine derived directly from M. tuberculosis, the causative agent of
tuberculosis in humans, rather than M. bovis, which causes tuberculo-
sis in cattle. Recent advances in the molecular biology of mycobacteria
have resulted in the construction of several attenuated mutants of
M. tuberculosis as candidates for a vaccine32–34. The objective for gen-
erating a M. tuberculosis–derived vaccine is that it would be more effi-
cacious and induce a qualitatively better protective immune response
than a M. bovis–derived BCG vaccine. The decrease in protective qual-
ity of BCG may be due to the decrease in the immunogenicity of
BCG35. Thus, M. tuberculosis mutants should offer better protection
because of the expression of an antigenic profile that is nearly identical
to that of virulent M. tuberculosis. Identifying and characterizing these
classes of mutants could also provide insights into the interplay
between the tubercle bacilli and the host. This knowledge could help to
define the strategies this highly successful intracellular pathogen uses
to subvert the hostile environment within the host.
Severe growth in vivo (sgiv) mutants
The sgiv mutants were initially considered to be ideal vaccine candi-
dates. They are highly attenuated because of their inability to replicate
and survive in vivoand would most likely be safer than BCG, in partic-
ular for immunocompromised individuals. However, the ability to
replicate and persist may be a desirable trait in an antituberculous vac-
cine, as it would allow for a prolonged immune response, similar to
that obtained with the live attenuated BCG vaccine. Most sgivmutants
cause considerably reduced pathology as a result of their severe growth
defect. Many are defective in their ability to obtain the essential nutri-
ents required for successful intracellular replication from the hostile
phagosomal environment in which M. tuberculosis replicates. For
example, the ability to acquire magnesium while residing in the
phagosome seems to be essential for the virulence of both M. tubercu-
losis and another intracellular pathogen, Salmonella enterica. Thus,
mutations in mgt, which is involved in magnesium transport, result in
reduced replication in vitro in response to a combination of low mag-
nesium and mildly acidic pH, representative of the M. tuberculosis–
containing phagosome36. These M. tuberculosis mutants show sub-
stantially reduced growth in the lungs and spleens of mice, characteris-
tic of an sgiv mutant.
Many auxotrophic mutants of M. tuberculosis also show the charac-
teristic sgiv phenotype. Indeed, M. tuberculosis leucine (leuD)32, lysine
(lysA)37, proline (proC)38, tryptophan (trpD)38and purine (purC)33
auxotrophs fail to grow in the mouse lung. As for the potential use of
sgiv auxotrophic mutants as vaccine candidates, both M. tuberculosis
proC and trpD gave the same protection as BCG after challenge with
intravenously administered M. tuberculosis in mice38, whereas the
purC auxotroph33provided similar protection to that provided by
BCG in an aerosolized guinea pig model of infection. These data indi-
cate that sgivmutants are potential candidates for an improved vaccine
and warrant further investigation.
Growth in vivo (giv) mutants
Most constructed mutants fall into the giv class, showing reduced
growth and pathology, resulting in an attenuated phenotype and asso-
ciated increase in host survival time. These mutants may make ideal
vaccine candidates because they maintain replication and are more
likely to result in a long-lasting host immune response than would an
sgiv mutant. However, a potentially dangerous situation could arise if
the mutant regains its rate of growth, as may happen in immunocom-
promised individuals, resulting in disease. Thus, continued character-
ization of specific mutants of M. tuberculosis is required to develop
strains that are able to elicit a strong protective immune response, but
are unable to reactivate in individuals with severe immunodeficiency.
950
VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY
Figure 1 Growth of M. tuberculosis in the lungs of immunocompetent mice.
C57BL/6 mice were infected intravenously with 1 × 106colony-forming units
(CFU) of M. tuberculosis, and growth kinetics were determined in the lungs.
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 3

REVIEW
An example of a giv phenotype is caused by deletion of the gene
encoding exported repetitive protein (erp) in both M. bovis BCG and
M. tuberculosis39. The deletion of this gene, which has no ascribed func-
tion and is specific to mycobacterial species40, impaired the growth of
the bacilli both in the lungs and spleens of BALB/c mice, and in cul-
tured macrophages. Therefore, Berthet et al. postulated that virulence
is dependent on the ability of the bacilli to multiply39. Another example
of a giv mutant of M. tuberculosis is the glutamine synthetase gene
(glnA1) mutant. This gene product is important in nitrogen metabo-
lism, and the glnA1 mutant is attenuated both for growth in
macrophages in vitro and in survival studies in the guinea pig model of
infection41. The acquisition of iron by intracellular M. tuberculosis has
been proposed to be mediated by mycobactins (2-hydroxyphenyloxa-
zoline-containing siderophore molecules), and an mbtB deletion
mutant, is restricted for growth in iron-limited media but not iron-
replete media, and shows reduced growth in the human monocyte-
derived cell line THP-1 (ref. 42). From these studies alone it seems that
M. tuberculosis has evolved a variety of ways to synthesize or acquire
essential nutrients for successful intracellular replication from the
nutrient-limited phagosome in which the bacilli resides.
A means of identifying virulence-associated genes is the technique of
signature-tagged mutagenesis43, a powerful screening method for the
identification of genes required for the growth of a pathogen in an ani-
mal model. This technique has been successfully used for many human
pathogens, including M. tuberculosis44,45. In both studies44,45, mutants
were obtained that were in the giv category (showed a growth defect in
the mouse lung) and that emphasized the importance of the distinctive
cell wall–associated lipids of the tubercle bacillus. Camacho and col-
leagues44identified 16 attenuated mutants, most of which were
involved in lipid metabolism or transport across the membrane.
Several of the mutants had disruptions in a cluster of genes involved in
the synthesis of the complex cell wall–associated lipid components
pthiocerol and phenolpthiocerol derivatives, which are found only in
pathogenic mycobacterial species. Similarly, in the study by Cox et al.45,
mutants were identified that had disruptions in the synthesis and trans-
port of phthiocerol dimycocerosate. All the phthiocerol dimycocerosate
mutants showed a change in colonial morphology, with overlapping
tubular structures rather than the characteristic flat ‘corded’ structures
of wild-type colonies, consistent with an alteration in the cell wall.
To sense and adapt to the changing environment encountered in the
various stages of the host’s immune response, systems such as two-
component regulatory proteins of M. tuberculosis are believed to
be important for intracellular survival. The phoP and phoQ two-
component regulatory proteins control the transcription of genes
important in the survival of many key human pathogens, including
Salmonella sp., Shigella sp., Yersinia sp. and Mycobacteria sp. Indeed, a
M. tuberculosis phoPdeletion mutant is attenuated in the mouse model
of infection. It has reduced growth in the liver, lung and spleen, which is
characteristic of a giv mutant, and also shows reduced intracellular
multiplication in bone marrow–derived macrophages in vitro46. The
phoP mutant differed from wild-type in the relative abundance of
monoacylated and triacylated mannosylated lipoarabinomannans
(ManLAMs) and monoacylated ManLAMs. ManLAMs, which are
found in the pathogenic slow-growing mycobacterium species, are
considered to be key virulence factors and powerful anti-inflammatory
mediators47. This may explain the giv phenotype of the phoP mutant,
although the specific mechanisms remain to be elucidated.
As most bacterial virulence factors are exported, it can also be postu-
lated that M. tuberculosis uses previously unknown secretion systems
with which to secrete virulence factors. The deletion of the accessory
secretion factor secA2, resulting in the attenuation of virulence of this
mutant in mice, is again characteristic of a giv phenotype48. This obser-
vation indicates that M. tuberculosis can secrete immune mediators
capable of counteracting the host innate immune response. As men-
tioned earlier, M. tuberculosiscan partially overcome the potent toxicity
of a key macrophage effector mechanism, the production of reactive
oxygen intermediates, by secreting superoxide dismutase18. Superoxide
dismutase lacks a classical signal sequence for protein export, yet the
secretion of superoxide dismutase is secA2 dependent, and many other
such immune mediators have been postulated to be secreted by this
previously unknown pathway48.
The importance of the de novo biosynthesis of pantothenate (vita-
min B5) in the virulence of M. tuberculosis was demonstrated by the
deletion of the pantothenate-synthesizing genes panC and panD34.
The double panCD deletion mutant of M. tuberculosis was highly
attenuated in survival studies, and much safer in immunocompro-
mised SCID mice than the BCG vaccine. The ∆panCD mutant of
M. tuberculosis retained the ability to undergo limited replication in
immunocompetent mice and conferred substantial protection to mice
after aerosol challenge with virulent M. tuberculosis.
Persistence (per) mutants
The tubercle bacillus is extremely successful in persisting within the
human host, where it is generally considered to remain walled off
within granulomas of the lung7, although a recent study has sug-
gested that bacilli may also persist in ‘nonprofessional’ phagocytes as
well49. Indeed, most people infected with the bacillus show no signs
of disease and only develop active disease when their immune system
is perturbed—for example, after the onset of AIDS, with aging and
with severe alcohol abuse. This asymptomatic infection is called
latent disease and is notoriously difficult to treat, mainly because the
bacilli are dormant and most antibiotics are directed against replicat-
ing bacteria. Indeed, the current protocol for treating active tubercu-
losis is a 6-month course of ‘frontline’ drugs to reach noncultivable
stasis (in which no bacilli can be cultured ex vivo). However, dor-
mant bacilli are likely to be noncultivable, and hence the disease
potentially could still reactivate. Indeed, in the mouse model of
latent infection (the Cornell model), general immunosuppressive
treatment can reactivate M. tuberculosis infection after treatment to
noncultivable stasis50,51.
NATURE IMMUNOLOGY
VOLUME 4 NUMBER 10 OCTOBER 2003
951
Figure 2 Electron micrograph of intracellular M. tuberculosis infection.
Human A549 alveolar epithelial cell infected for 72 h with M. tuberculosis
Erdman (arrow) at a multiplicity of infection of 10:1. Original
magnification, ×5000.
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 4

REVIEW
To withstand host adaptive immune responses, M. tuberculosis may
have adapted its persistance traits. Mutants would therefore be
expected to be deficient in persisting in the face of this immunological
onslaught (that is, per mutants). Characterization of this class of
mutants could lead to the elucidation of the mechanisms M. tuberculo-
sis uses to survive and persist while in this stage.
Koch first described the association of mycobacterial virulence with a
colonial morphotype called ‘cording’; that is, the formation of braided
microscopic bundles by the bacteria52. A genetic screen for M. bovis
BCG mutants that failed to cord led to the identification of a class of
mutants that failed to persist53. These mutants, which replicated nor-
mally in the initial stages of infection, were disrupted in the gene
encoding proximal cyclopropane of α mycolates (pcaA), which is
required for cording and for the synthesis of the mycolic acid cyclo-
propane ring, thus defining a function for cyclopropanated lipids in
mycobacterial pathogenesis. The same was also found for the M. tuber-
culosis pcaA mutant. The lungs of mice infected with M. tuberculosis
pcaA mutant are characterized by large aggregations of lymphocytes,
indicating that the cyclopropanation of mycobacterial lipids may be
important in reducing the influx of lymphocytes, which are essential to
the host’s acquired immune response.
Other per mutants have mutations in the gene encoding isocitrate
lyase (icl)54, which is essential for the metabolism of fatty acids; muta-
tions in the genes encoding phospholipase C (plc)55, which are
responsible for the cleaving of the phospholipid phosphatidylcholine
and are associated with the most virulent mycobacterial species56; and
the relMtb mutant, which is compromised in the maintenance of
long-term viability57. The icl mutant bacilli show normal growth in
the acute phase of infection in mice, but very reduced growth after the
onset of acquired immunity54, indicating that persistence relies at
least in part on fatty acid metabolism.
M. tuberculosis has four putative phospholipase C genes, plcA, plcB,
plcC and plcD, and mutants with single and multiple gene deletions
have been constructed55. All of the mutants
showed a decrease in phospholipase C activ-
ity, with each gene product contributing
equally to overall phospholipase activity55.
The triple (plcABC-deficient) and quadruple
(plcABCD-deficient) mutants showed the
classical per phenotype; they were attenuated
in the late phase of infection, thus indicating
that phospholipase C activity has an as-yet-
unidentified function in enabling M. tubercu-
losis to survive the onslaught of the T cell
mediated immune response of the host.
After nutrient deprivation, all bacilli begin
to grow more slowly and mount what is
known as the ‘stringent response’. This
response is characterized by a reduction in the
amounts of protein synthesis and also the
synthesis of rRNA and tRNA. In M. tuberculo-
sis, the stringent response is mediated by
hyperphosphorylated guanine nucleotides,
or (p)ppGpp, which are synthesized and
hydrolyzed by the M. tuberculosis RelMtb
enzyme58. Mutants, which are deficient in the
activity of this enzyme, are defective in long-
term survival in vitro59. The phenotype of the
mutant in the lungs and spleen of BALB/c
mice is that of a per mutant, showing an
impaired ability to grow during acquired
immunity. This indicates that the stringent response is required for the
persistence of M. tuberculosis in vivo.
Tubercle bacilli also modulate the stress-inducible ‘SOS’ response in
adverse conditions. This response can alter the genetic mutation rate
as a means of generating mutants more fit to survive than the parental
strain, and Boshoff et al. hypothesized that the main replicative DNA
polymerase (DnaE) in M. tuberculosis may be involved in error-prone
DNA repair synthesis60. Indeed, dnaE2 is upregulated in vitro in
response to the DNA damaging agents, ultraviolet light and hydrogen
peroxide. A dnaE2 mutant showed reduced survival after ultraviolet
irradiation and reduced colony counts in the mouse lung 9 months
after infection and was attenuated for virulence in survival studies60.
Drug resistance emerged more frequently in the wild-type and com-
plemented strains than in the mutant, showing that the action of this
polymerase contributes to the emergence of drug resistance in vivo.
This extensive study establishes a function for DnaE2-mediated DNA
repair during persistent infection.
Overexpression of the heat-shock protein 70 (Hsp70) proteins in M.
tuberculosis also resulted in considerable impairment of persistence
during the chronic stage of infection61. No obvious difference in
growth was noted in bone marrow–derived macrophages in vitro, orin
mice during the early phase of infection. However, the survival and
growth of the mutant in the persistent stage of infection was substan-
tially reduced and coincided with an increased number of CD8+IFN-
γ-secreting T cells in the spleen. Overexpression of Hsp70 may favor
the host rather than the pathogen during the chronic phase of infec-
tion, and thus the induction of Hsp expression may provide a means of
fortifying host defenses during latent disease.
Pathology (pat) mutants
The pat mutants show no growth defect in the face of either innate or
acquired immunity, but they do not elicit a typical antituberculous
immune response and hence show a distinct difference in host
952
VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY
Table 1 Classification of the described M. tuberculosis mutants
Class M. tuberculosis mutantGene function
sgiv leuD32
lysA37
mgt36
proC38
trpD38
purC33
erp39
phoP46
secA2 (ref. 48)
fadD28, pps-promoter, mmpL7 (ref. 45) PDIM synthesis and transport
glnA1 (ref. 41)
panCD34
pcaA53
icl54
plcABCD56
relMTB57
dnaE2 (ref. 60)
Hsp70 (ref. 61)
sigH63
rpoV(sigA)/whiB3 (ref. 65)
hbhA66
Leucine synthesis
Lysine synthesis
Magnesium transport
Proline synthesis
Tryptophan synthesis
Purine synthesis
Exported repetitive protein (function unknown)
Two-component regulatory protein
Accessory secretion factor (secretes SOD)
giv
Glutamine synthetase (nitrogen metabolism)
Pantothenate synthesis
Proximal cyclopropanation of a-mycolates
Isocitrate lyase (involved in cording phenotype)
Phospholipase C
(p)ppGpp synthesis and hydrolysis
DNA polymerase
Hsp70
Sigma factor H
Sigma factor A and putative transcription factor
HBHA
per
pat
dis
PDIM, phthiocerol dimycocerosate; SOD, superoxide dismutase.
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 5

REVIEW
pathology. The identification of this group of mutants emphasizes the
existence of genes within the M. tuberculosisgenome that are required
for increased immunopathology of the host. However, the fact that
the tubercle bacillus has such pathology-escalating genes may be
related to transmission of the bacilli, which is dependent on progres-
sive lung damage and necrosis. Indeed, for M. tuberculosis to infect a
new host, the bacteria must be coughed out into the surrounding
atmosphere and hence must be extracellular to aid effective dissemi-
nation as droplets. The bacilli are thought to be extracellular only in
the center of a liquefied necrotic lesion associated with extensive lung
pathology and damage. The identification of pat mutants also
emphasizes that colony-forming unit analyses alone are insufficient
and that survival studies and pathology should also be examined
when characterizing in vivo bacterial mutants in general62.
A M. tuberculosis extracytoplasmic function sigma factor H gene
(sigH) mutant was found to be attenuated for virulence in the mouse
model of infection. However, this mutant showed no in vivo growth
defect63. The attenuated phenotype correlated with fewer and smaller
foci of infection, which do not progress as rapidly as in wild-type
M. tuberculosis–infected mice. There was no difference in the virulence
in immunocompromised SCID mice, indicating that T cell–mediated
immunity is involved in this mutant phenotype. Indeed, flow cytometry
showed that at 4 weeks after infection of immunocompetent mice, the
mutant bacteria recruited only 10% the number of CD4+and CD8+
T cells into the lung, with fewer cells producing IFN-γand tumor necro-
sis factor-α. From these results, it seems that the modest T cell response
elicited by the mutant bacteria is sufficient for control and that the
overly exuberant response seen in wild-type infection is the cause of the
increased tissue pathology and earlier death noted in wild-type
M. tuberculosis–infected mice.
Another sigma factor mutant was discovered when genetic com-
plementation analyses of attenuated and virulent strains of M. bovis
identified a gene, rpoV (also called sigA), which restored virulence
to the attenuated strain64. This gene showed a high degree of
homology with principal transcription factors and sigma factors
from other bacteria, and a more recent study identified the putative
transcription factor with which it interacts, called whiB3 (ref. 65). A
M. tuberculosis whiB3 deletion mutant showed no difference in bac-
terial burden compared with that of wild-type, yet showed an atten-
uated phenotype in immunocompetent mice, as observed from
survival studies. The pathological changes noted in the wild-type
infected mice were much more severe than those in the mutant
infected mice, indicating that RpoV-WhiB3–mediated transcription
influences the immune response of the host. Both the SigH and
whiB3 pat mutants lack the control of a regulated set of genes, indi-
cating that such regulons are required for the induction of typical
tuberculous immunopathology.
Further elucidation of differences in the pathology induced by dif-
ferent pat mutants is required to fully define the function of the host
immune response. For example, does the mutant-induced immuno-
pathology involve different subtypes and percentages of T cells, as
noted in the sigH and Hsp70 mutants61,63, or other cell types? Also, is
there a difference in the production or secretion of immunomodula-
tory molecules induced by these mutants? Dissecting these interac-
tions will lead to an increased understanding of advantageous and
disadvantageous immunopathological responses and could result in
the development of strategies that boost the hosts immune response
during active disease. A pat mutant showing negligible pathology, but
persisting at a high level of infection may also be a good vaccine candi-
date, stimulating the immune response but causing no harmful
pathology or disease. In immunocompromised individuals, such a
mutant might pose little threat, as it is unable to cause disease in the
individual and also fails to undergo transmission to a new host.
Dissemination (dis) mutants
Although this classification system is a useful means of cataloging the
various mutants encountered, there may be variations. One example is
the heparin-binding hemagglutinin (hbhA) mutant, which shows no
apparent growth defect when administered intravenously, but shows a
defect in dissemination to extrapulmonary sites when administered
intranasally66. This mutant constitutes a new class of mutants, the dis-
semination (dis) mutants. The development of a vaccine that cannot
disseminate would be of particular utility in immunocompromized
individuals. There have been many reports of disseminated BCG dis-
ease in children vaccinated with the current vaccine who also subse-
quently developed AIDS67–69, and also in adults with AIDS who
received the vaccine as an infant70,71.
M. tuberculosis can persist within the host in cell types other than
macrophages, including alveolar epithelial cells49, in which the bacillus
can survive and replicate in vitro72. Disruption of M. tuberculosishbhA
affected the adherence, invasion and survival in alveolar epithelial
cells, but not macrophages, and showed reduced extrapulmonary col-
onization in BALB/c mice66. This indicates that passage through the
alveolar epithelial cell layers is required for extrapulmonary dissemi-
nation of the bacillus. Incubating wild-type bacilli with antibodies to
heparin-binding hemagglutinin adhesin (HBHA) can inhibit dissemi-
nation, which, as well as confirming this hypothesis, has implications
for the potential use of such antibodies therapeutically. However,
Mueller-Ortiz et al.73, also working with a hbha mutant of M. tubercu-
losis, reported that there was an overall growth defect noted with this
mutant in the lungs, liver and spleen of the more resistant mouse
strain, C57BL/6. These differences in observation could be related to
the different mouse strains used by the researchers. Nevertheless, the
growth defect in C57BL/6 mice was still more pronounced in the livers
and spleens than in the lungs, which may be the result of a defect in
dissemination as well as growth. Further definition of the function of
HBHA and other bacterial factors in dissemination is needed for a
greater understanding of the host-pathogen interaction during this
key phase of the disease.
Conclusion
Knowledge of the intracellular ‘lifestyle’ of the tubercle bacillus will
lead to a greater understanding of the host-pathogen interaction and
the mechanisms by which M. tuberculosis evades both the innate and
NATURE IMMUNOLOGY
VOLUME 4 NUMBER 10 OCTOBER 2003
953
Figure 3 Classification of mutants based on their growth in the lungs of
immunocompetent mice. C57BL/6 mice were infected intravenously with
1 × 106colony-forming units (CFU) of wild-type (WT) or mutant bacilli.
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 6

REVIEW
954
VOLUME 4 NUMBER 10 OCTOBER 2003 NATURE IMMUNOLOGY
adaptive immune responses. Such knowledge should lead to the devel-
opment of new, safe and efficacious vaccines and antimycobacterial
treatment strategies.
An absolute requirement of a live M. tuberculosis–derived vaccine is
that it be as safe or safer than BCG in humans. The sgivmutants, which
fail to grow in vivo, such as the leucine32, proline38, tryptophan38and
purine33auxotrophs, show protection similar to that induced by BCG
vaccination. The combination of two or three such deletions would
provide a considerable safety advantage for use in humans. However, it
remains to be determined whether limited replication of such a mutant
would provide optimal immunogenicity and long-term protection.
Another possibility is that such mutants would be too attenuated to
elicit such protection.
The giv mutants are also defective for growth in vivo, yet not to the
same extent as noted with the sgiv mutants. The giv mutants seem to
be defective in their ability to interact with the innate immune
responses, such as the phthiocerol dimycocerosate mutants45, and
plcABCD56and secA248mutants. A combination of such mutations
may reduce the virulence of a M. tuberculosis strain enough to provide
safety in immunocompromised animals, as noted with the panCD34
mutant, yet allow for limited replication and the elicitation of optimal
immunogenicity. The addition of a pat mutation, with reduced
immunopathology, or a dis mutation, with reduced extrapulmonary
dissemination, would also increase the safety of such a vaccine by fur-
ther reducing the risk of disease and dissemination.
More knowledge about per mutants, which are defective in persist-
ing in the face of an adaptive immune response, will improve our
understanding of the mechanisms used by M. tuberculosis to survive
the adaptive mycobactericidal mechanisms. This knowledge should
lead to new immunotherapies, as well as to strategies for enhancing
the immunogenicity of an attenuated M. tuberculosis vaccine. It may
even be possible to rationally design effective therapeutic vaccines that
could eliminate a pre-existing M. tuberculosisinfection. Thus, the ideal
vaccine candidate would probably be a mix of these classes, providing
a safer and more efficacious alternative to BCG. Because the develop-
ment of new therapeutic strategies and vaccines are urgently required
to curtail this global health crisis, efforts focusing on elucidating the
strategies M. tuberculosis uses to be a successful pathogen should be
actively pursued.
ACKNOWLEDGMENTS
Supported by National Institutes of Health grant AI 26170.
1. Corbett, E.L. et al. The growing burden of tuberculosis: global trends and interactions
with the HIV epidemic. Arch. Intern. Med. 163, 1009–1021 (2003).
2. Cole, S.T. et al. Deciphering the biology of Mycobacterium tuberculosis from the com-
plete genome sequence. Nature 393, 537–544 (1998).
3. Smith, I. Mycobacterium tuberculosis pathogenesis and molecular determinants of
virulence. Clin. Micro. Rev. 16, 463–496 (2003).
4. Young, D.B. Ten years of research progress and what’s to come. Tuberculosis (Edinb)
83, 77–81 (2003).
5. Jacobs, W.R. Jr. in Molecular Genetics of Mycobacteria (ed. Hatfull, G.F. & Jacobs,
W.R. Jr.) 1–19 (ASM Press, Washington D.C., 2000).
6. Barnes, P.F., Modlin, R.L. & Ellner, J.J. in Tuberculosis; pathogenesis, protection and
control (ed. Bloom, B.R.) 417–430 (ASM Press, Washington D.C., 1994).
7. Dannenberg, A.M. & Rook, G.A.W. in Tuberculosis pathogenesis, protection, and con-
trol (ed. Bloom, B.R.) 459–483 (ASM Press, Washington, DC, 1994).
8. Flynn, J.L. & Chan, J. Immunology of tuberculosis. Ann. Rev. Immunol. 19, 93–129
(2001).
9. Doffinger, R. et al. Inherited disorders of IL-12- and IFN-γ-mediated immunity: a mol-
ecular genetics update. Mol. Immunol. 38, 903–909. (2002).
10. Fieschi, C. et al. Low penetrance, broad resistance, and favorable outcome of inter-
leukin 12 receptor β1 deficiency: medical and immunological implications. J. Exp.
Med. 197, 527–535 (2003).
11. Lichtenauer-Kaligis, E.G. et al. Severe Mycobacterium bovis BCG infections in a large
series of novel IL-12 receptor β1 deficient patients and evidence for the existence of
partial IL-12 receptor β1 deficiency. Eur. J. Immunol. 33, 59–69 (2003).
12. Armstrong, J.A. & Hart, P.D. Response of cultured macrophages to Mycobacterium
tuberculosis, with observations on fusion of lysosomes with phagosomes. J. Exp. Med.
134, 713–740 (1971).
13. Xu, S. et al. Intracellular trafficking in Mycobacterium tuberculosis and
Mycobacterium avium-infected macrophages. J. Immunol. 153, 2568–2578 (1994).
14. Schlesinger, L.S., Bellinger-Kawahara, C.G., Payne, N.R. & Horwitz, M.A.
Phagocytosis of Mycobacterium tuberculosis is mediated by monocyte complement
receptors and complement C3. J. Immunol. 144, 2771–2780 (1990).
15. Wright, S.D. & Silverstein, S.C. Receptors for C3b and C3bi promote phagocytosis
but not the release of toxic oxygen from human phagocytes. J. Exp. Med. 161,
2016–2023 (1983).
16. Zimmerli, S., Edwards, S. & Ernst, J.D. Selective receptor blockade during phagocy-
tosis does not alter the survival and growth of Mycobacterium tuberculosis in human
macrophages. Amer. J. Resp. Cell. Mol. Biol. 15, 760–770 (1996).
17. Jackett, A.S., Aber, V.R. & Lowrie, D.B. Virulence and resistance to superoxide, low
pH and hydrogen peroxide among strains of Mycobacterium tuberculosis. J. Gen.
Micro. 104, 37 (1978).
18. Middlebrook, G. Amer. Rev. Tubercul. 69, 471–472 (1954).
19. Hickman, S.P., Chan, J. and Salgame, P. Mycobacterium tuberculosis induces differ-
ential cytokine production from dendritic cells and macrophages with divergent
effects on naive T cell polarization. J. Immunol. 168, 4636–4642 (2002).
20. Nau, G.J. et al. Human macrophage activation programs induced by bacterial
pathogens. Proc. Natl. Acad. Sci. USA 99, 1503–1508 (2002).
21. Noss, E.H., Harding, C.V. & Boom, W.H. Mycobacterium tuberculosis inhibits MHC
class II antigen processing in murine bone marrow macrophages. Cell. Immunol.
201, 63–74 (2000).
22. Ting, L.-M., Kim, A.C., Cattamanchi, A. & Ernst, J.D. Mycobacterium tuberculosis
inhibits IFN-γ transcriptional responses without inhibiting activation of STAT1.
J. Immunol. 163, 3898–3906 (1999).
23. Chan, J., Fan, X., Hunter, S.W., Brennan, P.J. and Bloom, B.R. Lipoarabinomannan,
a possible virulence factor involved in persistence of Mycobacterium tuberculosis
within macrophages. Infect. Immun. 59, 1755–1761 (1991).
24. Sly, L.M., Hingley-Wilson, S.M., Reiner, N.E. & McMaster, W.R. Survival of
Mycobacterium tuberculosis in host macrophages involves resistance to apoptosis
dependent upon induction of antiapoptotic Bcl-2 family member Mcl-1. J. Immunol.
170, 430–437 (2003).
25. Fine, P.E. BCG: the challenge continues. Scand. J. Infect. Dis. 33, 243–245 (2001).
26. Fine, P.E.M., Carneiro, I.A.M., Milstein, J.B. & Clemens, C.J. Issues relating to
the use of BCG in immunization programmes. http://www.who.int/vaccines-docu-
ments/DocsPDF99/www9943.pdf (2002).
27. Huebner, R.E. BCG vaccination in the control of tuberculosis. Curr. Top. Micro.
Immun. 215, 263–282 (1996).
28. Hess, J. et al. Mycobacterium bovis Bacille Calmette-Guerin strains secreting listeri-
olysin of Listeria monocytogenes. Proc. Natl. Acad. Sci. USA 95, 5299–5304 (1998).
29. Horwitz, M.A., Harth, G., Dillon, B.J. & Maslesa-Galic, S. Recombinant bacillus cal-
mette-guerin (BCG) vaccines expressing the Mycobacterium tuberculosis 30-kDa
major secretory protein induce greater protective immunity against tuberculosis than
conventional BCG vaccines in a highly susceptible animal model. Proc. Natl. Acad.
Sci. USA 97, 13853–13858 (2000).
30. Horwitz, M.A. & Harth, G. A new vaccine against tuberculosis affords greater survival
after challenge than the current vaccine in the guinea pig model of pulmonary tuber-
culosis. Infect. Immun. 71, 1672–1679 (2003).
31. Pym, A.S. et al. Recombinant BCG exporting ESAT-6 confers enhanced protection
against tuberculosis. Nat. Med. 9, 533–539 (2003).
32. Hondalus, M.K. et al. Attenuation of and protection induced by a leucine auxotroph of
Mycobacterium tuberculosis. Infect. Immun. 68, 2888–2898 (2000).
33. Jackson, M. et al. Persistence and protective efficacy of a Mycobacterium tuberculo-
sis auxotroph vaccine. Infect. Immun. 67, 2867–2873 (1999).
34. Sambandamurthy, V.K. et al. A pantothenate auxotroph of Mycobacterium tuberculo-
sis is highly attenuated and protects mice against tuberculosis. Nat. Med. 8,
1171–1174 (2002).
35. Behr, M.A. & Small, P.M. Has BCG attenuated to impotence? Nature 389, 133–134
(1997).
36.Buchmeier, N. et al. A parallel intraphagosomal survival strategy shared by
Mycobacterium tuberculosis and Salmonella enterica. Mol. Micro. 35, 1375–1382
(2000).
37. Pavelka, M.S.J., Chen, B., Kelley, C.L., Collins, F.M. & Jacobs, W.R. Jr. Vaccine effi-
cacy of a lysine auxotroph of Mycobacterium tuberculosis. Infect. Immun. 71,
4190–4192 (2003).
38. Smith, D.A., Parish, T., Stoker, N.G. & Bancroft, G.J. Characterization of auxotrophic
mutants of Mycobacterium tuberculosis and their potential as vaccine candidates.
Infect. Immun. 69, 1142–1150 (2001).
39. Berthet, F.-X. et al. Attenuation of virulence by disruption of the Mycobacterium
tuberculosis erp gene. Science 282, 759–762 (1998).
40. de Mendonca-Lima, L. et al. Erp, an extracellular protein family specific to mycobac-
teria. Microbiology 147, 2315–2320 (2001).
41. Tullius, M.V., Harth, G. & Horwitz, M.A. Glutamine synthetase GlnA1 is essential for
growth of Mycobacterium tuberculosis in human THP-1 macrophages and guinea
pigs. Infect. Immun. 71, 3927–3936 (2003).
42. De Voss, J.J. et al. The salicylate-derived mycobactin siderophores of Mycobacterium
tuberculosis are essential for growth in macrophages. Proc. Natl. Acad. Sci. USA 97,
1252–1257 (2000).
43. Hensel, M., Shea, J.E., Gleeson, C., Jones, M.D., Dalton, E. & Holden, D.W.
Simultaneous identification of bacterial virulence genes by negative selection.
Science 269, 400–403 (1995).
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology

Page 7

REVIEW
44. Camacho, L.R., Ensergueix, D., Perez, E., Gicquel, B. & Guilhot, C. Identification of
a virulence gene cluster of Mycobacterium tuberculosis by signature-tagged transpo-
son mutagenesis. Mol. Micro. 34, 257–267 (1999).
45. Cox, J.S., Chen, B., McNeil, M. & Jacobs, W.R. Jr. Complex lipid determines tissue-
specific replication of Mycobacterium tuberculosis in mice. Nature 402, 79–83 (1999).
46. Perez, E. et al. An essential role for phoP in Mycobacterium tuberculosis virulence.
Mol. Micro. 41, 179–187 (2001).
47. Nigou, J., Gilleron, M. & Puzo, G. Lipoarabinomannans: from structure to biosynthe-
sis. Biochimie 85, 153–166 (2003).
48. Braunstein, M., Espinosa, B.J., Chan, J., Belisle, J.T. & Jacobs, W.R.J. SecA2 func-
tions in the secretion of superoxide dismutase A and in the virulence of
Mycobacterium tuberculosis. Mol. Micro. 48, 453–464 (2003).
49. Hernandez-Pando, R. et al. Persistence of DNA from Mycobacterium tuberculosis in
superficially normal lung tissue during latent infection. Lancet 356, 2113–2114
(2000).
50. McCune, R. & Tompsett, R. Fate of Mycobacterium tuberculosis in mouse tissues as
determined by the microbial enumeration technique I. The persistence of drug-sus-
ceptible tubercle bacilli in the tissues despite prolonged antimicrobial therapy.
J. Exp. Med. 104, 737–762 (1957).
51. McCune, R., Tompsett, R. & McDermott, W. Fate of Mycobacterium tuberculosis in
mouse tissues as determined by the microbial enumeration technique II. The conver-
sion of tuberculous infection to the latent state by the administration of pyrazinamide
and a companion drug. J. Exp. Med. 104, 763–804 (1957).
52. Koch, R. The Aetiology of tuberculosis (ed. Pinner, D.a.M.M.) (National Tuberculosis
Association, New York, 1882).
53. Glickman, M.S., Cox, J.S. & Jacobs, W.R. Jr. A novel mycolic acid cyclopropane syn-
thetase is required for coding, persistence, and virulence of Mycobacterium tubercu-
losis. Mol. Cell 5, 717–727 (2000).
54. McKinney, J.D. et al. Persistence of Mycobacterium tuberculosis in macrophages
and mice requires the glyoxylate shunt enzyme isocitrate lyase. Nature 406,
735–738 (2000).
55. Raynaud, C. et al. Phospholipases C are involved in the virulence of Mycobacterium
tuberculosis. Mol. Micro. 45, 203–217 (2002).
56. Johansen, K.A., Gill, R.E. & Vasil, M.L. Biochemical and molecular analysis of phos-
pholipase C and phospholipase D activity in mycobacteria. Infect. Immun. 64,
3259–3266 (1996).
57. Dahl, J.L. et al. The role of RelMtb-mediated adaptation to stationary phase in long-
term persistence of Mycobacterium tuberculosis in mice. Proc. Natl. Acad. Sci. USA
100, 10026–10031 (2003).
58. Avarbock, D., Avarbock, A. & Rubin, H. Differential regulation of opposing RelMtb
activities by the aminoacylation state of a tRNA.ribosome.mRNA.RelMtb complex.
Biochem. 39, 11640–11648 (2000).
59. Primm, T.P. et al. The stringent response of Mycobacterium tuberculosis is required
for long-term survival. J. Bacteriol. 182, 4889–4898 (2000).
60. Boshoff, H.I.M., Reed, M.B., Barry, C.E. 3rd & Mizrahi, V. DnaE2 polymerase con-
tributes to in vivo survival and the emergence of drug resistance in Mycobacterium
tuberculosis. Cell 113, 183–193 (2003).
61. Stewart, G.R. et al. Overexpression of heat-shock proteins reduces the survival of
Mycobacterium tuberculosis in the chronic phase of infection. Nat. Med. 7, 732–737
(2001).
62. North, R.J., Ryan, L., LaCource, R., Mogues, T. & Goodrich, M.E. Growth rate of
mycobacteria in mice as an unreliable indicator of mycobacterial virulence. Infect.
Immun. 67, 5483–5485 (1999).
63. Kaushal, D. et al. Reduced immunopathology and mortality despite tissue persis-
tence in a Mycobacterium tuberculosis mutant lacking alternative sigma factor, SigH.
Proc. Natl. Acad. Sci. USA 99, 8330–8335 (2002).
64. Collins, D.M. et al. Mutation of the principal sigma factor causes loss of virulence in
a strain of the Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA 92,
8036–8040 (1995).
65. Steyn, A.J. et al. Mycobacterium tuberculosis WhiB3 interacts with RpoV to affect
host survival but is dispensable for in vivo growth. Proc. Natl. Acad. Sci. USA 99,
3147–3152 (2002).
66. Pethe, K. et al. The heparin-binding haemagglutinin of M. tuberculosis is required for
extrapulmonary dissemination. Nature 412, 190–194. (2001).
67. Talbot, E.A., Perkins, M.D., Silva, S.F. & Frothingham, R. Disseminated bacille
Calmette-Guerin disease after vaccination: case report and review. Clin. Infect. Dis.
24, 1139–1146 (1997).
68. Rosenfeldt, V., Paerregaard, A. & Valerius, N.H. Disseminated infection with Bacillus
Calmette-Guerin in a child with advanced HIV disease. Scand. J. Infect. Dis. 29,
526–527 (1997).
69. Raton, J.A. et al. Disseminated bacillus Calmette-Guerin infection in an HIV-infected
child: a case with cutaneous lesions. Med. J. Austr. 156, 286–287 (1992).
70. Reynes, J.P.C., Lamaury, I., Janbon, F. & Bertrand, A. Bacille Calmette-Guerin adenitis
30 years after immunization in a patient with AIDS. Lancet 356, 2133–2138 (2000).
71. Smith, E., Thybo, S. & Bennedsen, J. Infection with Mycobacterium bovis in a patient
with AIDS: a late complication of BCG vaccination. Scand J. Infect. Dis. 24,
109–110 (1992).
72. McDonough, K.A. & Kress, Y. Cytotoxicity for lung epithelial cells is a virulence-asso-
ciated phenotype of Mycobacterium tuberculosis. Infect. Immun. 63, 4802–4811
(1995).
73. Mueller-Ortiz, S.L. et al. Decreased infectivity despite unaltered C3 binding by a ∆hbha
mutant of Mycobacterium tuberculosis. Infect. Immun. 70, 6751–6760 (2002).
NATURE IMMUNOLOGY
VOLUME 4 NUMBER 10 OCTOBER 2003
955
© 2003 Nature Publishing Group http://www.nature.com/natureimmunology