Immunological Mechanisms of Human Resistance To Persistent

Machine Translated by Google
Reviews
Immunological mechanisms of human

resistance to persistent
Mycobacterium tuberculosis infection
Jason D. Simmons1*, Catherine M. Stein2,3, Chetan Seshadri1, Monica Campo1,
Galit Alter4, Sarah Fortune5, Erwin Schurr6, Robert S. Wallis3,7, Gavin Churchyard7,
Harriet Mayanja-Kizza 8, W. Henry Boom3 and Thomas R. Hawn1
Abstract | Mycobacterium tuberculosis is a leading cause of mortality worldwide and establishes
a long-lived latent infection in a substantial proportion of the human population. Multiple lines of
evidence suggest that some individuals are resistant to latent M. tuberculosis infection despite
long-term and intense exposure, and we term these individuals ‘resisters’. In this Review,
we discuss the epidemiological and genetic data that support the existence of resisters and
propose criteria to optimally define and characterize the resister phenotype. We review recent
insights into the immune mechanisms of M. tuberculosis clearance, including responses
mediated by macrophages, Tÿcells and B cells. Understanding the cellular mechanisms that
underlie resistance to M. tuberculosis infection may reveal immune correlates of protection that
could be utilized for improved diagnostics, vaccine development and novel host-directed therapeutic strategie
Through a long coevolutionary history with humans, stimulation with two to three specific M. tuberculosis
Purified protein derivative
(PPD) skin reactivity test Mycobacterium tuberculosis has developed into a highly antigens and thus avoids false positive results from prior
A delayed-type successful pathogen that infects 23–32% of the world BCG vaccination8 . In the absence of a direct measure of
hypersensitivity test that population1,2 and tuberculosis (TB) is the leading infec- infection, the best surrogates for initial infection are the
measures induration at the site
tious cause of death3 . Following aerosol exposure to M. PPD skin reactivity test and/or IGRA responses that
of an intradermal injection of
peptide extract from tuberculosis, three potential clinical outcomes are possi- convert from negative to positive (that is, incident positive)
mycobacterial culture filtrate. ble, namely, resistance or early clearance of the bacillus, after M. tuberculosis exposure.
A positive result reflects a pre- asymptomatic or latent M. tuberculosis infection (LTBI) In endemic TB settings, some adults who are heavily
specified minimal diameter of
that can persist for decades, or symptomatic ‘active and repeatedly exposed to M. tuberculosis remain neg-
skin induration and suggests
the presence of latent
tuberculosis’, which includes pulmonary disease that can ative for reactivity in the PPD test and the IGRA. These
result in further transmission. Recent whole-blood individuals, whom we term resisters, can be defined clin-
Mycobacterium tuberculosis
infection but may also result transcriptomic profiling4,5 and advanced lung imaging ically as resistant to infection (Fig.ÿ1). However, as this
from non-M. tuberculosis
modalities6 have provided new insight into the transi-tion designation includes several assumptions about expo-
sensitization, including prior from subclinical to active TB7,8
vaccination with
. However, why some sure and the stability of the results of these diagnostic
Mycobacterium bovis Bacillus heavily exposed individuals never acquire, or perhaps tests, we propose the following enrolment criteria for
Calmette–Guérin. Also known immediately eliminate, infection is poorly understood. studies of resisters (Boxÿ1). First, a high level of exposure
as the Mantoux tuberculin skin No reliable test exists that directly detects the presence is crucial and should include indices of exposure inten-
test (TST).
or absence of M. tuberculosis in asymptomatic individu- sity (for example, high bacillary load of a known index case
Bacillus Calmette–Guérin
als. The purified protein derivative (PPD) skin reactivity test and close proximity of contact, such as sharing a room or
(BCG). A culture-adapted, measures delayed-type hypersensitivity to mycobac-terial bed) as well as indices of exposure duration that capture
attenuated Mycobacterium antigens and has been the gold standard for the diagnosis cumulative exposure (for example, repeated household
bovis strain that is used for
of LTBI for>100 years9,10. As PPD is enriched for protein exposures or employment in settings of docu-mented high
vaccination against
antigens11 that are not necessarily specific to M. M. tuberculosis transmission), all of which can be assessed
Mycobacterium tuberculosis.
tuberculosis, prior immunization with Mycobacterium bovis using validated exposure risk scores12,13.
Bacillus Calmette–Guérin (BCG) or exposure to non- Second, diagnosis of a resister requires a negative result
tuberculous mycobacteria can yield false positive results. for both the PPD skin reactivity test and the IGRA to avoid
*email: jasonds@uw.edu IFNÿ release assays (IGRAs) were developed as a whole- misclassification. Third, these diagnostic tests should be
https://doi.org/10.1038/ blood diagnostic test that measures IFNÿ release from M. carried out serially following documented exposures to
s41577-018-0025-3
tuberculosis-antigen-specific CD4+ Tÿcells after capture any conversions to LTBI (or TB).
NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 575

Reviews
Fig. 1 | The spectrum of human resistance to infection by Mycobacterium tuberculosis. The extent of resistance
to Mycobacterium tuberculosis infection is proportional to the duration and the intensity of exposure to M. tuberculosis.
Individuals who resist infection despite heavy exposure to M. tuberculosis are more likely to have immunogenetic
mechanisms of resistance than those who resist infection after a lower extent of exposure. Individuals with resistance to
M. tuberculosis infection after intense exposure are termed ‘resisters’ (white), whereas individuals who test positive in the
purified protein derivative (PPD) skin reactivity test and/or the IFNÿ release assay (IGRA) can either be asymptomatic
(latent M. tuberculosis infection (LTBI); pink) or be symptomatic with tuberculosis (TB; red). DC, dendritic cell.
The resister phenotype is probably heterogeneous and development of host-directed therapies or more effective
IFNÿ release assays
(IGRAs). Whole-blood immune might involve innate immunological mechanisms of early vaccines and adjuvants.
assays in which a patient’s clearance of M. tuberculosis (innate resisters).
blood is cultured in the Alternatively, protective immunity might also occur through Epidemiology
presence of select antigens
adaptive immune responses (adaptive resisters), including Despite the absence of a direct microbiological meas-
specific to Mycobacterium
tuberculosis and the secretion IFNÿ-independent responses that have been measured ure of M. tuberculosis in latent infection, historical and
of IFNÿ is quantified. Positive in preclinical models14–16, and thus a poten-tial role for rigorous modern epidemiological studies suggest that
results are independent of prior B cell or unconventional Tÿcell responses in early resisters exist, although their estimated frequency varies
Bacillus Calmette–Guérin in different studies.
clearance of M. tuberculosis should be consid-ered.
vaccination and negative results
Longitudinal, rigorous epidemiological studies of resisters,
ensure a sufficient response to
a mitogen positive control. together with data from genetic and immunolo-gical Historical case–contact studies. Strong epidemiologi-
platforms, are providing new insights into human cal evidence suggests that some individuals are natu-
I resisted resistance to M. tuberculosis infection. rally resistant to infection, even after heavy exposure to
An individual who remains
In this Review, we describe the epidemiology of the M. tuberculosis in closed environments. In 1966, all
negative for the purified protein
derivative skin reactivity test
resister phenotype and discuss several caveats about the enlisted personnel and officers aboard the destroyer
and the IFNÿ release assay definition of a resister. Next, we review the genetic deter- U.S.S. Richard E. Byrd were enrolled in a study after a 5
throughout serial testing minants of the resister phenotype and discuss potential cm pulmonary cavity was detected (by radiography) in a
despite heavy exposure to effector mechanisms in macrophages, unconventional crew member who suffered progressive respiratory
Mycobacterium tuberculosis.
Tÿcells and B cells. We examine immune factors that symptoms for the prior 6 months aboard the ship17.
pre-vent or limit M. tuberculosis infection in innate resisters Of the 308 at-risk enlisted crew members, 7 developed
or adaptive resisters, which could be harnessed for the active TB, 6 of whom shared the same berthing com-
partment as the index case. Most crew members in this
‘high-burden’ compartment showed evidence of new
Author addresses infection (either active TB or an incident positive PPD),
1Department of Medicine, University of Washington, Seattle, WA, USA. whereas seven crew members (~10%) were negative for
2Department of Population & Quantitative Health Sciences, Case Western Reserve University, the PPD skin test during the initial study and remained so
Cleveland, OH, USA. at follow-up. Other studies, including an evaluation of
3Department of Medicine, Case Western Reserve University, Cleveland, OH, USA. nursing students in the pre-antibiotic era, also showed
4Ragon Institute of MGH, MIT and Harvard, Cambridge, MA, USA. evi-dence of PPD-negative responses despite intense
5Department of Immunology and Infectious Diseases, Harvard School of Public Health, Boston, MA, expos-ure to patients with TB (Tableÿ1). A systematic
USA. review of household contact studies detected heterogeneity
6Program in Infectious Diseases and Immunity in Global Health, Research Institute of the McGill University in resistance to infection, with an average of 50% of close
Health Centre, Montreal, Québec, Canada. contacts remaining uninfected18. However, these studies
7The Aurum Institute, Parktown, South Africa. reveal substantial variability in PPD-conversion rates,
8Department of Medicine, School of Medicine, Makerere University, Kampala, Uganda. which is consistent with unmeasured exposure variables.
576 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

Reviews
Innate resisters Not surprisingly, proximity and duration of contact with the M. tuberculosis infection (that is, induration = 0 mm in the
Conceptually, resisters index case as well as the infectivity of the index case (that PPD skin test)31. In this extremely high-burden
who immediately clear is, bacillary burden) were predictors of PPD conversion. environment, any gold miners who remain persistently
Mycobacterium tuberculosis The results of most early studies in hyperendemic uninfected are likely resisters. According to mathe-matical
following exposure to the
conditions suggest that a minority of the population (5– models, the likelihood that a 40-year-old gold miner
bacterium and before priming of
50%) has the resister phenotype, and this variability without evidence of M. tuberculosis infection
any adaptive immune
responses. probably reflects exposure intensity and the duration of (induration=0mm in the PPD skin test) is a resister is
follow-up17,19,20. However, the results of some studies 93%, compared with only 22% among the general community32.
Adaptive resisters sug-gest that rates of resistance are higher (for example, Two prospective studies in South Africa are currently
Conceptually, resisters who
clear or contain
up to 70% of exposed individuals are resisters), even after being carried out to identify HIV-uninfected or HIV-infected
Mycobacterium tuberculosis periods of high exposure (that is, sleeping in the same resisters who have been mining for ÿ15 years, are 33–60
infection via Tÿcell and B cell bed or room) to the index case21–24. Importantly, most of years of age, do not have evidence of past or active TB
mechanisms but who remain
these studies either had a short duration of follow-up or or silicosis and have neither a history of treatment for
negative for the purified protein
contained data that were difficult to accurately review by LTBI nor any non-HIV immunosuppressive conditions.
derivative skin reactivity test
and the IFNÿ release assay.
modern epidemiological standards. Finally, the strain of Few longitudinal studies of conversion in the PPD skin
M. tuberculosis probably differed among the various reactivity test or the IGRA in hyperendemic set-tings
Household contact studies and thus might affect transmission rates. For have incorporated extended follow-up periods.
An individual who resides
example, increased virulence and transmission might be Uganda is a high-burden TB setting with a 3% annual risk
within the same domicile as an
one of several factors that are involved in the increased of M. tuberculosis infection, as determined by PPD skin
index case (patient with
range and incidence of the lineage 2/Beijing strains of M. test and/or IGRA conversion33. LTBI is consider-ably
tuberculosis) for a pre-specified
amount of time. tuberculosis25–27. more prevalent among household contacts than among
The results of these early studies suggest that a community members, and consequently, there is a risk
Hyperendemic conditions
resister phenotype exists, but the epidemiological methods difference of 30% among adults34. In Kampala, Uganda,
Regions in which ongoing
Mycobacterium tuberculosis
that were used could not ensure high exposure and household contacts of index cases with culture-
transmission is unusually high adequate follow-up to confirm the presence of clinical confirmed TB were evaluated for active TB and under-
owing to an elevated resistance to M. tuberculosis infection. went PPD skin testing35. The annual incidence of new
prevalence of tuberculosis.
cases of culture-positive TB in this study was 740 per
Silicosis
Contemporary cohort studies. Population studies of 100,000 individuals, compared with country-wide esti-
An occupational pulmonary individuals in hyperendemic TB settings in the past 5 mates of 159 per 100,000 individuals33. If an individual
disease that is prevalent years have provided opportunities to examine resist-ance was negative in the PPD skin reactivity test at baseline
among gold miners owing to to M. tuberculosis infection. For example, among South ( <10 mm induration for HIV-negative individuals
protracted inhalation of silicate
African gold miners, exposure to silica dust and the >5 years old,<5mm induration for HIV-negative indivi-
dust, which leads to scarring
and increased Mycobacterium
congregate working, living and social conditions that are dualsÿ5 years old or<5mm induration for all HIV-positive
tuberculosis infection. associated with mining contribute to the high level of individuals), repeated PPD skin tests were carried out at
ongoing M. tuberculosis transmission28,29. Mathematical 3, 6, 12 and 24 months after enrolment. Among the entire
modelling suggests that the annual risk of infection (that study cohort, 11.7% of adult close household con-tacts
is, conversion in the PPD skin test) among gold min-ers remained persistently negative in the PPD skin test
is as high as 20%30. The prevalence of LTBI in these throughout the two year follow-up12,36. Both the South
gold miners was estimated at 89%, and 13% (15/115) of African and Ugandan cohorts described here provide an
HIV-negative participants showed no evidence of opportunity to rigorously define the resister phenotype
and explore its underlying cellular mechanisms.
Box 1 | Proposed criteria to define the resister phenotype Complexities of the resister phenotype. Given that the
Resisters (individuals who show resistance to Mycobacterium tuberculosis infection definition of a resister relies on epidemiological data and
despite long-term, intense exposure to the bacillus) are exemplified by the following utilizes an indirect, immunological measurement of M.
characteristics: tuberculosis infection, it is likely that some resisters are
misclassified. Several alternative hypotheses may explain
Extent of exposure to Mycobacterium tuberculosis
the resister phenotype.
•Intensity: high bacillary burden of index case (for example, sputum-positive), closed
Despite meeting enrolment criteria as a household
environment (household, mine shafts and so on) and use of validated epidemiological risk
scoring12,13
contact, an individual with a negative result in the PPD
skin reactivity test may simply have had insufficient
•Duration: cumulative exposures (that is, health-care workers, miners, household
exposure, either from a low intensity of exposure (that is,
contacts or adults with community exposure in hyperendemic areas)
a low bacillary burden in the index case, minimal contact
Result of diagnostic tests with the index case or good ventilation within the
•Negative result in both the purified protein derivative (PPD) skin reactivity test and household) or a short duration of contact with an index
the IFNÿ release assay (IGRA) case. In the Ugandan study, a published modified
epidemiological exposure risk score13, consisting of char-
Durability of responses
acteristics of the TB index case and the intensity of the
•Multiple negative results for the PPD skin reactivity test and the IGRA in the first year following contact with the index case, did not differ between resist-
exposure
ers and individuals with LTBI12. This result suggested
•Additional longitudinal testing, when available and if exposures persist, to minimize false
that all household contacts were highly exposed to M.
negative test results and enrich for conversions after cumulative exposure
tuberculosis. Participants in the Ugandan household

Reviews
owing to the inherent limitations of an immunolog-ical

Table 1 | Prevalence of the resister phenotype
definition of a clinical phenotype. Both PPD skin test and
Exposed population Length of follow-up Cumulative Refs
IGRA responses are incomplete measures of anti-
resister prevalence
mycobacterial immunity for a number of reasons.
(percentage)a
As discussed later, these LTBI diagnostic tests fail to
Historical purified protein derivative studies
detect the response of Tÿcells that are specific for non-
Sailors sharing berthing >6 months 9/65 (13.8) 17
protein mycobacterial antigens via MHC-independent
compartment with an pathways, and IGRAs do not detect the response of
individual with active TB
antigen-specific Tÿcells with alternative (that is, IFNÿ-
Sailors with frequent 6–12 months 76/226 (33.7) 173
independent) cytokine-producing profiles40–42. Tÿcell
contact with two individuals with
anergy could also cause false negative results in the
incident TB aboard ship
PPD skin test. In the Uganda study, HIV-negative indi-
Young nursing students 3 years 0/1,453 44
viduals were not more likely to be PPD-negative than
working in heavy-burden TB
HIV-positive individuals, for whom a positive PPD test
hospital in Norway
was defined as induration>5mm (ref. 12). Furthermore,
Nursing students at 3 years 0/637 45
in recent follow-up studies, all resisters demonstrated

Philadelphia General
Hospital (USA) in 1935–1939 reac-tivity to the mitogen positive control that was
174
included in the IGRA methodology (W.H.B. and C.M.S.,
Nurses training at Boston 3 years 6/126 (4.8)
City Hospital (USA) in
unpub-lished observation), suggesting that generalized CD4+
1932–1936 Tÿcell anergy or depletion alone does not explain the
175
resister phenotype. This does not rule out a contribution
University of Minnesota 3 years 16/184 (8.7)
nursing classes graduating in of pathogen-specific anergy, which has been observed
1932–1935 (before in human lepromatous leprosy43.
removal of TB ward and Resistance to M. tuberculosis probably has a thresh-
installation of infection
old, above which relative resistance can be overcome
control)
by a sufficiently high exposure. For example, in two
Professional students of RNs: 3 years RNs: 30/112 (26.8) 176
land-mark studies in the 1930s and 1940s in Norway
Wisconsin General Hospital and
Wisconsin State MDs: 4 years MDs: 50/283 (17.7) and at the Philadelphia General Hospital, USA, the
Sanatorium in 1934–1944 prevalence of LTBI among nursing students at
(RNs) and 1934–1943 (MDs) matriculation was 45.3% (668/1,453)44 and 57%
Close family contacts of 5 years following 110/851 (12.9) 22 (366/643)45, respectively. At 3-year follow-up, however,
index pulmonary TB cases in diagnosis of index case 100% of students that were initially negative for the PPD
India skin test in both studies converted, and the majority of
Close family contacts of 6 months following 103/292 (35.3) 21 these negative-to-positive conversions occurred during
individuals with pulmonary diagnosis of index case the early stages of training. Considering these data,
TB in Pakistan together with those from the contemporary studies
Household contacts of index 12 months following 83/253 (32.8) 23
discussed earlier, it is possible that repeated testing (for
cases with sputum positive for diagnosis of index case example, serial testing every 4 months with 2 doses45)
the AFB-smear test
and the use of tuberculin of non-standardized purity (for
Contemporary studies example, the two early studies were carried out before
Gold miners in South Africa Variable 15/115 (13.0) 31,32 large-scale standardized preparations of tuberculin, such
as PPD-S, became available in 1941) effectively
Close household contacts of 2 years 142/1,210 (11.7) 12,177
sensitized some individuals in the absence of natural

individuals with pulmonary TB
in Uganda infection. Non-tuberculous mycobacterial infec-tions are
AFB, acid fast bacilli; MDs, medical students; RNs, nursing students; TB, tuberculosis. also known to produce false positive results in the PPD
a The number of contacts persistently negative for reactivity in the purified protein derivative skin test in some populations, a response that could
(PPD) skin test (and the IFNÿ release assay, if available) as a proportion of all contacts enrolled potentially be boosted by serial PPD testing, as was
in the study (that is, PPD-positive contacts, PPD-negative contacts and PPD converters).
seen with M. tuberculosis infection46. Furthermore, the
induration thresholds that were used for each method
contact study were re-contacted 8–10 years after the (Mantoux or von Pirquet) were not described in these
original study and were tested multiple times with IGRAs studies. Nonetheless, the results of these pre-antibiotic
(3 assays over 3 years) and once with a repeat PPD era studies suggest that absolute M. tuberculosis resistance occurred
skin reactivity test. Although they were not necessarily re- Relative resistance to M. tuberculosis disease might
exposed within the same household in the interval since also be a function of the infectious dose. The classical
their initial enrolment as a household contact, all these work of Lange and Chaussé showed that small droplets
Relative resistance individuals continued to live in areas of Kampala with carrying one bacillus or a few bacilli are the effectors of
A threshold phenomenon in high TB prevalence in which community exposure to M. aerosol transmission and that the vast majority of
which mechanisms of host tuberculosis might be substantial37–39. These data will individuals with TB show only a single Ghon complex47,48.
resistance to Mycobacterium
further clarify the durability of the resister phenotype. In natural infections, the frequency of exposure to M.
tuberculosis infection may be
overcome owing to frequent or
An alternative explanation for the resister pheno- tuberculosis is the most parsimonious surrogate for
lengthy exposure to contacts type is that persistently negative results in PPD skin infectious dose, but how consistency of exposure is
with a high bacillary burden. reactivity tests or IGRAs are false negative results linked to innate resistance mechanisms is not known.

Reviews
The epidemiological studies discussed above suggest household contacts is correlated among siblings but not
that some individuals are resistant to M. tuberculosis among unrelated children of the same household61.
infection, as defined by a persistently negative result in In a genome-wide linkage analysis of the Ugandan
the PPD skin test and/or the IGRA, despite sustained cohort, suggestive evidence for a linkage of regions on
high-level exposures over extended periods of time. chromosome 2q21–2q24 and chromosome 5p13–5q22
Furthermore, these individuals fail to develop active TB to the resister phenotype was detected57. These regions
over an extended follow-up period. Future research to are distinct from the regions that are linked to protection
uncover the mechanisms that are involved in resistance from TB, suggesting that the genetic underpinnings of
to M. tuberculosis infection will require detailed assess- these two ends of the TB spectrum (that is, asympto-
ments of exposure to the bacillus to define the resister matic infection and clinical disease) are unique. In a low
phenotype (Box 1). The true prevalence of resisters endemicity setting, a candidate-gene association study
among individuals who have been exposed to M. tuber- of individuals with or without LTBI who had close contact
culosis is not known, but evidence from studies that with an index pulmonary TB case reported that poly-
assessed exposure variables and included longitudinal morphisms in Toll-interacting protein (TOLLIP) and
sampling suggest that the prevalence is<10%. Unc51-like kinase 1 (ULK1) were associated with LTBI
susceptibility62,63. The specific roles of these proteins
Genetics in the resister phenotype are discussed later.
The cellular mechanisms that mediate M. tuberculosis Another genome-wide analysis assessed PPD skin
resistance are unknown, although multiple susceptibility test reactivity in M. tuberculosis-exposed South African
loci have been linked to or associated with the resister individuals. This cross-sectional study measured rates of
phenotype in genome-wide analyses. PPD-positivity both as a binary trait (that is, PPD=0mm
versus >0 mm) and as a quantitative trait by measure-
Genetic basis of resistance to other pathogens. ment of skin induration (in millimetres)64. As a binary
Genetic studies indicate that polymorphisms and trait, a significant linkage signal that met the stringent
mutations exist that are associated with resistance to genome-wide linkage cut-off was found at the TST1
some infections. For example, a 32bp deletion in CC– locus on chromosome 11p14. Strikingly, in a later study
chemokine receptor 5 (CCR5) in CD4+ Tÿcells results in of the same families, the TST1 locus was found to be
relative resistance to infection with R5-tropic HIV-1 and genetically indistinguishable (at a resolution of 2–14Mb)
delays disease progression49. Similarly, mutations in from a major quantitative trait locus (QTL) that controls
the gene encod-ing fucosyltransferase 2 (FUT2), which tumour necrosis factor (TNF) production following
is required for the cell surface expression of histo-blood stimulation of whole blood with either BCG alone or BCG
group anti-gens (HBGAs; the receptors for norovirus), plus IFNÿ65. These findings indicate that a genetic
results in decreased HBGA levels on mucosal surfaces determinant linked to TNF secretion in response to BCG
and thus prevents virus uptake and entry into epithelial cells.and IFNÿ might also determine PPD negativity, perhaps
Consequently, individuals with these FUT2 mutations through TNF-induced microbicidal pathways. However,
are completely (100% penetrant) resistant to Norwalk the genomic resolution around the TST1 locus in these
and other norovirus genogroups50,51. Atypical chemokine studies was low (2–14Mb), and this locus includes other
receptor 1 (ACKR1; also known as DARC and CD234) genes besides TNF. When reactivity in the PPD skin test
is the receptor on erythrocytes for Plasmodium vivax. was measured in the same study as a quantitative trait
Most west African individuals are homozygous for a (that is, the size of skin induration in millimetres), sig-
promoter mutation that ablates ACKR1 expression in nificant linkage was found with a second locus, TST2,
erythrocytes and, consequently, are resistant to P. vivax on chromosome 5p15 (ref. 64). This region of
No complex infection52. Although these examples show Mendelian chromosome 5 overlaps with the region that was
The focus of primary inheritance (for example, single-gene-inactivating associated with the resister phenotype in the Ugandan
Mycobacterium tuberculosis mutations that result in highly penetrant resistance), it is cohort57. Although the TST1 locus was strictly linked to
infection in the lung resistance to M. tuberculosis infection and invokes
likely that complex polygenic inheritance patterns are
parenchyma that histologically
corresponds to a granuloma important for resistance to other pathogens, including M. mechanisms that precede Tÿcell priming, the TST2 locus
and radiographically is
tuberculosis. might modulate Tÿcell responses that govern the intensity
recognized by its associated of the response in the PPD skin test.
lymphadenopathy and Genetic studies of purified protein derivative skin HIV-infected resisters, who have an increased likeli-
calcification.
reactivity. Numerous lines of evidence, including results hood of having inadequate adaptive immune responses
Quantitative trait locus from twin53,54, Mendelian primary immunodeficiency55, owing to CD4+ Tÿcell depletion, might be particularly
(QTL). A genomic region genome-wide linkage56–58 and candidate-gene dependent on innate immune mechanisms for resistance
containing polymorphisms that studies59,60, suggest that host genetics influences to M. tuberculosis. As HIV and TB are co-prevalent in
are associated with a
susceptibility to TB. Mendelian Susceptibility to Uganda and Tanzania, this hypothesis was tested in a
quantifiable phenotype, such
as the extent of induration (in Mycobacterial Disease (MSMD) is an immunodeficiency genome-wide association study of PPD reactivity in HIV-
millimetres) in the purified disorder in young children who are susceptible to infected individuals, which found that a locus on
protein derivative skin disseminated BCG and non-tuberculous mycobacteria chromosome 5q31.1 was significantly associated with
reactivity test. When variation and who have mutations in the IFNÿ and IL-12 signalling both binary (positive or negative in the PPD skin test; P
at these loci affect the
pathways55. Several studies also suggest that host = 1.22 × 10–8) and quantitative (induration in milli-
expression levels of specific
genes, they are referred to as genetic factors modulate resistance to M. tuberculosis metres; P=1.45×10–8) responses in the PPD skin test66.
expression QTLs. Thisexposure
infection. For example, the intensity of tuberculin reactivity after region contains
to four genes, including the gene

Reviews
encoding IL-9, which is expressed and secreted by mast Inflammatory signalling pathways. Genetic regulation
cells and T helper 2 (TH2) cells and was previously of inflammatory pathways is well described85–88, and
associ-ated with bronchial hyper-responsiveness. some studies have associated susceptibility to myco-
Interestingly, an inverse relationship between the bacterial infections with polymorphisms in Toll-like
incidence of asthma and TB was reported67,68, receptor (TLR) pathway components89–96, regulators
suggesting that IL-9 and its pleiotropic effects on lung of TLR function97,98, C-type lectin receptors99,
inflammatory responses pro-tect against M. tuberculosis autophagy proteins100,101 and other pathways for pro-
infection. Although IL9 is an intriguing possible resistance inflammatory cytokine production102. However, only
gene, the other genes in the 5q31.1 region may also some of these studies define functional polymorphisms
have a role in resistance to M. tuberculosis. The study (that affect TLR expression or signalling, for example),
also replicated associations in the regions previously and the identified associations with clinical phenotypes
identified in HIV-negative resist-ers (on chromosomes 2, require replication in other populations. Interestingly,
5 and 11; discussed previously), including susceptibility loci patients
near TST1.with Mendelian primary immunodeficiency of
Together, these results suggest that the resister phe- myeloid differentiation primary response 88 (MYD88) or
notype has polygenic variance; that is, many individual IL-1 receptor-associated kinase 4 (IRAK4), which are
genetic variants each with small individual effects con- cen-tral to most TLR signalling pathways, do not develop
tribute additively to the phenotype. Although evidence to mycobacterial diseases103. Furthermore, few human
support broad epistatic effects on infectious disease genetic studies have found an association of resistance
susceptibility is lacking, gene pair interactions have been to M. tuberculosis infection (as opposed to susceptibility
described in genome-wide analyses of susceptibility to to disease) with polymorphisms in genes involved in any
TB69,70 and leprosy71. Nonetheless, the substantial of these cellular pathways. A recently discovered single
overlap in the loci that are associated with the resister nucleotide polymorphism (rs5743854) in the promoter of
pheno-type or in the results of PPD skin reactivity tests the gene encoding TOLLIP is an expression QTL for
that was detected in independent analyses argues that which the G/G allele is associated with lower TOLLIP
evolution-ary pressure has selected variants that protect expression in monocytes and with increased risk of LTBI,
against the initial M. tuberculosis infection, at least in pulmonary TB and meningeal TB63. Although the
geographical regions that are hyperendemic for TB. specific mechanisms through which high TOLLIP
expression contributes to M. tuberculosis resistance as
Macrophage-mediated resistance well as to protection from disease are not known, the
After inhalation, aerosolized M. tuberculosis must known role of TOLLIP in directing autophagy and its
navigate multiple intrinsic barriers in the upper air-ways, negative regulation of TLR2, TLR4 and IL-1 receptor
including airway mucins and ciliated epithelial cells. The (IL-1R) suggest that these pathways are mechanisms of
bacterium then encounters resident alveo-lar resistance. Although replication of the genetic associa-
macrophages, which exhibit pro-inflammatory and anti- tion findings in independent cohorts is needed, the stud-
inflammatory responses and are uniquely posi-tioned to ies discussed above suggest that resistance alleles exist
have a key role in mediating early clearance of M. and may be relevant to the resister phenotype.
tuberculosis72–74. The effector functions of alveolar
mac-rophages might be modulated by neutrophils Autophagy. Autophagy limits intracellular survival of M.
through, for example, their formation of neutrophil tuberculosis and maturation of the mycobacterial
extracellular traps; however, the role of neutrophils in phagosome104,105. An inÿvitro genome-wide screen
both microbicidal and immunopathological activities is iden-tified autophagy factors as the host factors that
complex and is reviewed elsewhere75,76. M. tuberculosis were most influential in restricting the intracellular growth
is recognized by multiple pattern recognition receptors of M. tuberculosis106. The M. tuberculosis ESX1 secre-
(PRRs) at the cell surface or within the phagosome or tion system is required for autophagy initiation through
by cytosolic PRRs after phago-somal rupture77,78. cGAS–stimulator of interferon genes (STING)-triggered
Subsequently, the outcome of infection depends on recruitment of microtubule-associated proteins 1A/1B
effective phagosome acidification, phago-some–lysosome light chain 3 (MAP1LC3) to autophagosomes82,107,108,
fusion, 6kDa early secretory antigenic target (ESAT6) but ESX1 may also be involved in the impairment of
secretion system 1 (ESX1)-dependent escape of M. autophagosome maturation109. Whether autophagy is
tuberculosis from the phagosome79, activa-tion of the involved in host protection versus virulence in the mouse
inflammasome (resulting in maturation of pro-IL-1ÿ and model is controversial. Genetic deletion of the key
pro-IL-18)78, upregulation of host genes that are either autophagy factor Atg5 in mice results in more severe
required for intracellular killing mecha-nisms (for example, disease that correlates with both uncontrolled M. tuber-
reactive oxygen species, nitric oxide (produced by culosis growth and exacerbated lung immunopathol-
inducible nitric oxide synthase), cathelici-din antimicrobial ogy110. However, this severe disease was later
peptide (CAMP)80 and autophagy) or detrimental to the suggested to result from increased neutrophil recruitment
host (that is, IFNÿ production follow-ing intracellular in the autophagy protein 5 (ATG5)-deficient animals and
sensing of M. tuberculosis DNA by cyclic GMP–AMP not from defective autophagy111. Genetic variants within
synthase (cGAS)81,82) and, ultimately, the acti-vation the autophagy pathway have been associated with TB
of host cell death pathways (for example, pyropto-sis, in candidate-gene association studies100,101, but
apoptosis and necroptosis)83,84. Resistance alleles whether similar variants control LTBI risk is poorly studied.
could modulate a number of these airway clearance functions A recent
or macrophage
candidate-gene
antimicrobial
association
mechanisms
study found
(Fig.ÿ2).
that

Reviews
Fig. 2 | Macrophage-mediated resistance to Mycobacterium tuberculosis. Macrophages can provide resistance

to Mycobacterium tuberculosis infection at multiple points in the infection process, including the initial uptake of bacilli
into phagosomes and through any of the possible fates of these phagosomes. These fates include effective (versus
delayed) phagosome maturation (resulting in the generation of microbicidal products, such as nitric oxide, reactive
oxygen species and cathelicidin), fusion with the lysosome (resulting in degradation of the phagosomal contents),
recruitment of autophagy factors (resulting ultimately in autophagosome–lysosome fusion) or phagosomal escape of
bacteria into the cytosol. Pro-inflammatory cytokine production may result from Toll-like receptor (TLR) ligation or the
recognition of signals following phagosomal rupture, including inflammasome-triggered maturation of pro-IL-1ÿ or
pro-IL-18 and recognition of cytosolic M. tuberculosis DNA by cyclic GMP–AMP synthase (cGAS) and the resultant
stimulator of interferon genes (STING)-dependent interferon regulatory factor 3 (IRF3)-driven transcriptional
response. Vitamin D3 receptor (VDR)-dependent transcription, which requires prior activation of vitamin D to
1ÿ,25(OH)2D3 through TLR or IFNÿ signalling events, results in cathelicidin expression. Genetic and cellular evidence
exists for a role for the negative TLR regulator Toll-interacting protein (TOLLIP), the autophagy factor Unc51-like
kinase 1 (ULK1) and histone deacetylases (HDACs; which are master transcriptional regulators) in resistance.
AIM2, absent in melanoma 2; cGAMP, cyclic GMP–AMP; ESAT6, 6kDa early secretory antigenic target; mTOR,
mechanistic target of rapamycin; MYD88, myeloid differentiation primary response protein MYD88; NF-ÿB, nuclear
factor-ÿB; NLRP3, NOD-, LRR- and pyrin domain-containing 3; TBK1, TANK-binding kinase 1; TFs,
transcription factors; TNF, tumour necrosis factor.

Reviews
two polymorphisms in ULK1, which encodes a compo- controversial, as its production by CD4+ Tÿcells may be
nent of an upstream protein complex that transduces sig- dispensable for control of M. tuberculosis infection in some
nals to central autophagy effectors, were associated with mouse models14–16, whereas IFNÿ production is essential
LTBI in Asian contacts of index cases62. After adjustment in other mouse models123. Furthermore, IFNÿ-producing
for clinical risk factors for LTBI, each of these ULK1 CD4+ Tÿcells have not emerged as an immunological cor-
minor alleles was estimated to confer an 80% reduction in relate of protection from infection in humans, which sug-
LTBI risk. This result might be due to an improvement in gests that this profile is important but is not sufficient for
selective autophagy, decreased M. tuberculosis repli- protection against M. tuberculosis infection124,125. In
cation or increased TNF secretion in response to TLR indivi-duals with LTBI, MHC-restricted ÿÿ Tÿcells recognize
ligands, as suggested from mechanistic studies using a broad repertoire of MHC class I-restricted and MHC
ULK1-deficient cells62. class II-restricted M. tuberculosis peptides and exemplify
a classical TH1 cell-like response. Implicit in their definition,
Monocyte gene expression and histone deacetylase resisters lack reactivity in the PPD skin test or secretion of
pathways. One hypothesis to explain the resistance to M. IFNÿ by Tÿcells following antigen stimulation, although,
tuberculosis infection in the Uganda study12 is that fol- hypothetically, these individuals might have IFNÿ-
lowing infection, the expression profiles of blood mono- independent Tÿcell responses that mediate resistance to
cytes from individuals with LTBI and resisters are distinct. M. tuberculosis or its clearance.
In a comparative transcriptional profiling study, mono-
cytes from each clinical group were infected ex vivo with IFNÿ-independent Tÿcell responses.IFNÿ-independent
M. tuberculosis112, and the microarray data were analysed Tÿcell responses to mycobacterial antigens have been
with Gene Set Enrichment Analysis (GSEA)113 to iden- described (Fig.ÿ3). For example, Cambodian patients with
tify host pathways that are associated with each clini-cal pulmonary TB who remained anergic to PPD following
phenotype (resister and LTBI) according to curated gene treatment completion displayed tuberculin antigen-
sets from transcriptional databases. The gene set that specific Tÿcell responses inÿvitro that were characterized
was most significantly associated with the resister group by the production of IL-10 rather than IFNÿ126,127. IL-10,
was from a previous study of sodium butyrate- which is known for its anti-inflammatory role in mitigat-ing
stimulated cells114. Sodium butyrate is a short chain fatty TH1 cell responses to minimize the potentially dele-
acid that inhibits histone deacetylases (HDACs), which terious effects of TNF and IFNÿ, is mostly implicated in
modify chromatin and regulate cellular function, including the immune response to TB, and no evidence exists for a
signalling pathways115,116. Interestingly, phenyl- protective role of IL-10 in the early stages of M. tubercu-
butyrate, a class I HDAC inhibitor that is known to losis infection128. However, a genotyping study in Ghana
synergize with vitamin D to induce CAMP production and found that individuals with the IL10 promoter haplotype
inhibit M. tuberculosis growth in macrophages117, has with the highest association with low circulating levels of
been tested as a potential host-directed therapy for IL-10 were more likely to have TB or be PPD-positive than
pulmonary TB118. Activated vitamin D (1ÿ,25(OH)2D3) to be PPD-negative129. Whether high IL-10 levels in these
binds to the vitamin D3 receptor (VDR), which trans- individuals truly confer resistance to infection, as opposed
locates to the nucleus to induce expression of CAMP and to PPD-specific anergy in the presence of LTBI, remains
drive other host processes, including autophagy and to be determined.
autophagosome–lysosome fusion80. CAMP has direct Additional CD4+ Tÿcell subsets, including TH17 cells,
antimicrobial activity against M. tuberculosis in the could contribute to the resister phenotype through IFNÿ-
phagosomal compartment80. Together, these data independent mechanisms. In mice, the produc-tion of
suggest that monocytes from resisters and individuals IL-17A, one of several TH17-related cytokines, recruits
with LTBI respond differently to ex vivo M. tuberculosis neutrophils and other inflammatory cells to the lung and is
infection. Targeting HDACs with inhibitors might provide a involved in granuloma maturation130.
novel therapeutic option for M. tuberculosis Adoptive transfer of BCG-specific TH17 cells har-vested
infections; however, the specific mechanisms by which from immunized IFNÿ-deficient mice into recipient V(D)J
HDAC-dependent pathways are altered in resisters, and recombination-activating protein (RAG)-deficient mice
whether epigenetic mechanisms or ‘trained immunity’ are gave these recipient mice a sur-vival advantage when
involved, are unknown. challenged with M. tuberculosis, which was comparable
to that seen with transfer of TH1 cells from IFNÿ-competent
Tÿcell-mediated resistance mice131. This IFNÿ-
Several lines of evidence from human studies and animal independent partial protection provided by TH17 cells
challenge models underscore the importance of Tÿcells in against M. tuberculosis was also seen in another adoptive
controlling M. tuberculosis infection119–121. HIV-infected transfer model132. Unlike IL-10, IL-17 is not known to
individuals with low CD4+ Tÿcell counts have increased directly interfere with IFNÿ production, and whether these
risk of progressing from LTBI to TB. In addition, patients IFNÿ-independent TH17 activities are relevant to the
with MSMD have defects in the IL-12–IFNÿ pathway and resister phenotype is unknown. However, taken together,
increased susceptibility to disseminated mycobac-terial these examples illustrate that despite having persistent
infections55. Furthermore, mouse models of acute M. negative responses in PPD skin reactivity tests and IGRAs,
tuberculosis infection have demonstrated the impor-tance it is possible that resisters have other mycobacteria-
of IFNÿ production in controlling bacterial replica-tion and specific MHC class I-restricted or MHC class II-restricted
survival122. However, the source of IFNÿ remains Tÿcell responses.

Reviews
gd
gd
Fig. 3 | Tÿcell-mediated resistance to Mycobacterium tuberculosis. After phagocytosis of Mycobacterium tuberculosis

by macrophages and dendritic cells (not shown), various Tÿcell responses are stimulated, including conventional
MHC-restricted responses (peptide-specific Tÿcells) and MHC-independent responses (ÿÿ Tÿcells, which detect
phosphorylated prenyl metabolites (also known as phosphoantigens)), CD1-restricted (lipid-specific) Tÿcells and MHC class
I-related gene protein (MR1)-restricted Tÿcells (also known as mucosal-associated invariant T (MAIT) cells); MR1 presents
small molecules, such as vitamin B2 derivatives. Production of IFNÿ by CD4+ Tÿcells is a hallmark of the Tÿcell response
in individuals with latent M. tuberculosis infection (LTBI), whereas by definition, ‘resisters’ lack this IFNÿ response.
Qualitative differences in Tÿcell responses between resisters and individuals with LTBI lead to the production of different
macrophage-activating cytokines and chemokines, resulting in either the elimination of M. tuberculosis by resisters or persistence of M. tub
in individuals with LTBI. BTN3A1, butyrophilin subfamily 3 member A1; NTM, non-tuberculous mycobacteria; TCR, Tÿcell
receptor; TLR, Toll-like receptor.
Unconventional Tÿcell responses.In addition to cross-sectional study of South African adolescents with
or without LTBI, there were low intra-donor correlations
myco-bacterial protein antigens, Tÿcells also recognize non-
protein mycobacterial antigens (Fig.ÿ3). For example, between Tÿcell responses to protein and lipid
Tÿcells are activated by mycobacterial cell wall lipids antigens135, indicating that Tÿcell responses to lipid
that are bound to CD1 proteins, which are homologous antigens are not redundant with those to protein
to MHC class I molecules but are functionally non- antigens and that the responses to the different
polymorphic40,133,134. The human CD1 locus encodes antigens could be comple-mentary. Immunity to lipid
four proteins (CD1A–CD1D) that are localized at the antigens could be primed by exposure in early life to
cell surface and present lipid antigens to Tÿcells. In a non-tuberculous mycobac-teria, which could facilitate clearance of M

Reviews
following infection, thus limiting the initiation of peptide- of LTBI in individuals who are unable to produce IFNÿ in
specific Tÿcell responses that are mostly respon-sible for response to M. tuberculosis antigens.
producing a positive result in the PPD skin test.
In addition, lipid-specific Tÿcells could also be primed by B cell-mediated resistance
vaccination with BCG. Functionally, Tÿcells specific for Antibodies are powerful mediators of protective immu-nity
mycobacterial lipids have been shown to have a TH1 cell against many infectious diseases, and stimulating their
cytokine-producing profile ex vivo, supporting their role in production is the goal of vaccination strategies147,148.
the clearance of M. tuberculosis-infected cells136. In M. tuberculosis infection and disease, however, the role
The MHC class I-related gene protein (MR1) antigen- of humoral immunity is probably complex and remains
presenting molecule presents small molecules instead of unclear149,150. Passive transfer of antibodies has not
peptides or lipids to Tÿcells137 and is important for the consistently conferred protection151–154, and immuno-
development of mucosal-associated invariant T (MAIT) globulin G (IgG)-deficient individuals do not exhibit reduced
cells138–140. MAIT cells can kill M. tuberculosis-infected control of bacterial infections155. By contrast, a post hoc
cells in an MR1-dependent manner41, and Mr1-knockout analysis of the MVA85A vaccine trial found that elevated
mice have a higher lung burden of M. bovis BCG than wild- Ag85A-specific IgG titres correlated with protection against
type mice141. Furthermore, in candidate-gene association TB156. In addition, depletion of B cells in a non-human
studies, MR1 and CD1 polymorphisms were associated primate model of M. tuberculosis
with susceptibility to TB142,143. Finally, ÿÿ Tÿcells are a infection resulted in increased lesional bacterial bur-
Tÿcell lineage that express a heterodimeric Tÿcell receptor den157. Furthermore, in patients with active TB, B cell
and are activated by phosphorylated prenyl function is initially abnormal but returns to normal after
metabolites144,145. Vÿ9Vÿ2 Tÿcells are activated by BCG- successful treatment158,159. These observations highlight
infected cells and can lyse BCG-infected target cells146. the immunoregulatory role of B cells in controlling M.
Thus, in PPD-negative individuals, the CD1-specific, tuberculosis infections, which extends beyond their direct
MR1-specific and ÿÿ Tÿcell populations might provide antibody-mediated effector functions (Fig.ÿ4).
cellular immunity against M. tuberculosis that is inde-
pendent of the IFNÿ production by PPD-specific CD4+ Antibodies and latent Mycobacterium tuber-culosis
Tÿcells that defines LTBI. Further study is required to infection. Antibodies, plasma cells and antibody-responsive
determine whether these other populations of Tÿcells innate immune cells that express crystallizable fragment
protect against M. tuberculosis infection in resisters or, (Fc) receptors are abundant in M. tuberculosis
alternatively, whether these Tÿcells prevent progression granulomas160,161, indicating that these
Fig. 4 | Antibody-mediated resistance to Mycobacterium tuberculosis. Beyond their role in clearing pathogens,
antibodies also direct the rapid destruction of infected cells via the recruitment of innate immune cells (such as
phagocytes) that express crystallizable fragment (Fc) receptors. Two modifications to the Fc domain of an antibody control
its affinity for Fc receptors, namely, changes in the antibody subclass or isotype and its glycosylation. Fab, antigen-binding
fragment; NK cell, natural killer cell; ROS, reactive oxygen species.

Reviews
innate immune cells are actively recruited and might intense exposure. However, a number of challenges
potentially contribute to antimicrobial activity162,163. have impeded the elucidation of potential resistance
Distinct Fc effector profiles were observed in indi- mechanisms, including the limitations of using an indi-
viduals with LTBI versus those with active disease. rect immunological test to define the resister phenotype
Individuals with LTBI had improved antibody func- and the need to incorporate measurements of exposure
tionality that was linked to elevated binding to the low- intensity in study designs. Despite this, contemporary
affinity immunoglobulin-ÿ Fc region receptor III-A epidemiological evidence supports the existence of
(FcÿR3A)164, which is highly expressed by natural killer resisters. Genome-wide association analysis of data
cells, neutrophils, mature macrophages and dendritic from independent studies has identified loci that are
cells163. Furthermore, these antibodies not only associ-ated with resistance to M. tuberculosis infection.
improved innate immune cell activity but also contributed The elu-cidation of immunological and genetic
to the restriction of intracellular survival of M. tuberculosis mechanisms of the resister phenotype is at an early
in previously infected primary macrophages. Interestingly, stage, but these stud-ies might reveal heterogeneous
this differential functional activity was not related to protective responses that are relevant to different types
altered subclass selection but was instead related to of resisters. For example, macrophage-dependent
sub-stantial changes in antibody glycosylation among pathways that prevent bacterial uptake or rapidly clear
the M. tuberculosis-infected individuals164. M. tuberculosis before the devel-opment of an adaptive
immune response define innate resisters. Alternatively,
Antibodies and the resister phenotype. It is unclear adaptive resisters are individuals in which Tÿcell and B
whether antibodies contribute to the resister phenotype, cell effector functions eliminate or restrict M. tuberculosis
but it is plausible that at the time of repeated M. tubercu- infections, either independently of IFNÿ production or
losis exposure that pre-existing antibodies might direct through priming by non-protein antigens. Further study
the antimicrobial activity of macrophages, neutrophils or is required to clarify whether adaptive resisters have
dendritic cells to eliminate the bacteria more effec- greater resistance to progression to active TB than
tively. Although inÿvitro studies indicate a limited role for individuals with traditionally defined LTBI; however,
antibody-mediated opsonophagocytic prevention of these studies could identify novel cor-relates of immune
macrophage infection by M. tuberculosis, the addition of protection or aid current prognostic efforts to stratify
pooled IgG derived from individuals with LTBI to previ- those individuals who are most at risk of disease progression4,5 .
ously infected healthy donor primary macrophages can Looking ahead, integrated research approaches
drive bacterial elimination by a process that probably using advanced epidemiological, genetic and
involves activation of the inflammasome164. These immunological tools will be required to dissect
results suggest that, in addition to their minor role in mechanisms of resist-ance. In future studies, features of
limiting initial infection through opsonization, antibodies an optimally defined resister should incorporate several
‘cure’ infected macrophages if they are present at the variables outlined in Boxÿ1, including, first, indices of
right time and place. Although the majority of healthy exposure (both intensity and duration) using, at a
donor macrophages restricted M. tuberculosis survival minimum, validated epidemio-logical risk scoring12,13;
in the presence of pooled IgG from individuals with second, the use of both the PPD skin reactivity test and
LTBI164, the level of restriction varied by macrophage the IGRA to avoid misclassifica-tion bias resulting from
donor, which suggests that antibody-mediated killing of discordant immunological tests; and third, longitudinal sampling for s
M. tuberculosis is also dependent on factors that are Epidemiological risk scores must include an assess-
intrinsic to innate immune cells. ment of the nature of contact between the index case or
Thus, whether natural or affinity-matured humoral cases and the exposed individual (that is, whether the
immunity exists among resisters is unknown, but humoral index case is the primary caregiver and the frequency
immunity might provide an exciting novel immunological and proximity of contact) and some assessment of
axis that links the adaptive and innate immune responses infectivity of the exposure environment (for example,
to target and eradicate M. tuberculosis whether the index case is coughing, whether their spu-
after exposure and warrants further exploration. tum is smear-positive for M. tuberculosis and whether
there are multiple index cases within a household or
Conclusions and outlook employment setting). Although not yet ready for prac-
Historically, significant breakthroughs in treating infec- tical implementation, more advanced assessments of
tious diseases have been made by studying mechanisms infectious risk, such as aerobiological environmental
of resistance rather than susceptibility to infections. The sampling where transmission risk has been correlated
most notable historical example is the observation that with the degree of ventilation and with the number of
milkmaids exposed to cowpox were resistant to small- viable bacteria sampled from cough aerosols, could be
pox165. A more contemporary example is the HIV resist- incorporated in future studies37,169,170.
ance of individuals with a variant of CCR5, CCR5ÿ32 Even in areas of high TB endemicity, some level of
(refs166,167). Both these observations led to discordance exists between the results of PPD skin
transformative therapeutic interventions — the smallpox reac-tivity tests and IGRAs171. Consequently, individuals
vaccine and CCR5 inhibitors, respectively168. with discordant results in the PPD skin reactivity test
For much of the past century, many investigators and the IGRA are not categorized as resisters, although
have postulated that some individuals are naturally retaining clinical samples from these individuals is
resistant to M. tuberculosis infection in the face of important to facilitate studies to better understand the mechanistic

Reviews
basis of this discordance. As the resister category might new advances in stem cell work (for example, induced
encompass multiple biological phenotypes, at this early pluripotent stem cells derived from peripheral blood
stage in their characterization, we favour more stringent mononuclear cells) have made possible the creation of
criteria to avoid misclassification, which justifies the permanent cell lines from individuals, which can be diff-
increased expense and logistical complexity of imple- erentiated into different cell types that are more relevant for
menting multiple testing modalities (PPD skin reac-tivity studying M. tuberculosis pathogenesis (for example,
tests and IGRAs) and serial testing. Longitudinal sampling alveolar macrophages). Finally, if M. tuberculosis
should include, at a minimum, multiple LTBI diagnostic cultures from index cases are available, comparisons of
tests in the year following initial exposure, as most but not strain type, host phenotype (resister versus LTBI versus
all conversions occur in the first 3 months following TB) and host genotype, as well as determination of whether
exposure36. In circumstances in which resist-ers remain transmission events are sympatric or allopatric172, are
in regions of high M. tuberculosis transmis-sion, follow-up exciting avenues of investigation.
testing may be warranted to ensure that any new The integration of rigorous epidemiological pheno-
conversions owing to community exposures are detected. typing with mechanistic studies that investigate global
immune pathways in macrophages, Tÿcells and B cells,
Advances in the past decade have transformed genetic and how these pathways might differ between resisters and
methods and platforms, which now include affordable and individuals with LTBI, should provide further biological
standardized genome-wide detection of polymorphisms. insight into the mechanisms of early clearance or preven-
These tools provide assessment of polymorphisms that are tion of infection by M. tuberculosis. These potential bio-
annotated in public databases. logical mechanisms should provide targets for drugs and
Whole-genome sequencing is also more affordable and can vaccines that could improve specific immune functions and
be used to examine resisters for unidentified poly- improve therapeutic options for treating TB.
morphisms that are not annotated in public databases.
To address limitations in the availability of patient samples, Published online 12 June 2018
1. Dye, C., Scheele, S., Dolin, P., Pathania, V. & by PD-1 to prevent lethal disease. PLoS Pathog. 12, 28. Churchyard, G. J. etÿal. A trial of mass isoniazid
Raviglione, M. C. Consensus statement. Global burden of e1005667 (2016). preventive therapy for tuberculosis control. N. Engl.
tuberculosis: estimated incidence, prevalence, and 16. Smith, C. M. etÿal. Tuberculosis susceptibility and J. Med. 370, 301–310 (2014).
mortality by country. WHO global surveillance and vaccine protection are independently controlled by 29. Cowie, R. L. Short course chemoprophylaxis with
monitoring project. JAMA 282, 677–686 (1999). host genotype. mBio 7, e01516 (2016). rifampicin, isoniazid and pyrazinamide for tuberculosis
2. Houben, R. M. & Dodd, P. J. The global burden of latent 17. Houk, V. N., Baker, J. H., Sorensen, K. & Kent, D. C. evaluated in gold miners with chronic silicosis: a
tuberculosis infection: a re-estimation using The epidemiology of tuberculosis infection in a closed double-blind placebo controlled trial. Tuber Lung Dis.
mathematical modelling. PLoS Med. 13, e1002152 environment. Arch. Environ. Health 16, 26–35 77, 239–243 (1996).
(2016). (1968). 30. Vynnycky, E. et al. Tuberculosis control in the South
3. World Health Organization. Global Tuberculosis 18. Morrison, J., Pai, M. & Hopewell, P. C. Tuberculosis and African gold mines: mathematical modeling of a trial of
Report 2017. (WHO, 2017). latent tuberculosis infection in close contacts of community-wide isoniazid preventive therapy.
4. Thompson, E. G. etÿal. Host blood RNA signatures people with pulmonary tuberculosis in low-income and Am. J. Epidemiol. 181, 619–632 (2015).
predict the outcome of tuberculosis treatment. middle-income countries: a systematic review and 31. Hanifa, Y. etÿal. Prevalence of latent tuberculosis
Tuberculosis 107, 48–58 (2017). meta-analysis. Lancet Infect. Dis. 8, 359–368 (2008). infection among gold miners in South Africa.
5. Zak, D.E. et al. A blood RNA signature for This is a systematic review and meta-analysis of Int. J. Tuberc Lung Dis. 13, 39–46 (2009).
tuberculosis disease risk: a prospective cohort study. historical and contemporary case–contact studies 32. Wallis, R. S. Mathematical models of tuberculosis
Lancet 387, 2312–2322 (2016). that finds that 51.4% of individuals with a TB reactivation and relapse. Front. Microbiol. 7, 669
6. Esmail, H. etÿal. Characterization of progressive HIV- household contact have LTBI, although substantial (2016).
associated tuberculosis using 2-deoxy-2-[18F]fluoro-D- study heterogeneity exists that likely reflects This study presents a Markov model that uses
glucose positron emission and computed differing exposure intensities across studies and/or published estimates of annual risk of acquiring M.
tomography. Nat. Med. 22, 1090–1093 (2016). participants. tuberculosis infection to estimate the probability
7. Barry, C. E. 3rd etÿal. The spectrum of latent 19. Amberson, J. B. & Riggins, H. Tuberculosis among that an induration of 0 mm in the PPD skin
tuberculosis: rethinking the biology and intervention student nurses: a five year study at Bellevue hospital. reactivity test indicates a true resister and finds
strategies. Nat. Rev. Microbiol. 7, 845–855 Ann. Internal. With. 10, 156–165 (1936). that this value is 22% in South African non-miners
(2009). 20. Daniels, M. Primary tuberculous infection in nurses: and 93% in gold miners.
8. Pai, M. & Behr, M. Latent Mycobacterium manifestations and prognosis. Lancet 244, 201–204 33. World Health Organization. Global tuberculosis report
tuberculosis infection and interferon-gamma release (1944). 2015, 20th ed. WHO http://apps.who.int/iris/
assays. Microbiol. Spectr. https://doi.org/10.1128/ 21. Aziz, A., Ishaq, M. & Akhwand, R. Infection risk of handle/10665/191102 (2015).
microbiolspec.TBTB2-0023-2016 (2016). sputum positive tuberculosis patients to their family 34. Whalen, C. C. etÿal. Secondary attack rate of
9. Mendel, F. The von Pirquet skin reaction and intravenous contacts with and without chemotherapy. J. Pak. Med. tuberculosis in urban households in Kampala,
tuberculin treatment. Med. Klin. 4, 402–404 (1908). Assoc. 35, 249–252 (1985). Uganda. PLoS ONE 6, e16137 (2011).
22. Devadatta, S. etÿal. Attack rate of tuberculosis in a 5- 35. Stein, C. M., Hall, N. B., Malone, L. L. & Mupere, E.
10. Seibert, F. B. History of the development of purified year period among close family contacts of The household contact study design for genetic
protein derivative tuberculin. Am. Rev. Tuberc 44, 1–8 tuberculous patients under domiciliary treatment with epidemiological studies of infectious diseases.
(1941). isoniazid plus PAS or isoniazid alone. Bull. World Front. Genet. 4, 61 (2013).
11. Yang, H., Kruh-Garcia, N. A. & Dobos, K. M. Purified Health Organ. 42, 337–351 (1970). 36. Stein, C. M. etÿal. Resistance and susceptibility to
protein derivatives of tuberculin — past, present, and 23. Lemos , AC , Matos , ED , Pedral-Sampaio , DB & Netto , Mycobacterium Tuberculosis infection and disease in
future. FEMS Immunol. Med. Microbiol. 66, 273–280 EM Risk of tuberculosis among household contacts in tuberculosis households in Kampala, Uganda.
(2012). Salvador, Bahia . Braz. J. Infect. This. 8, 424–430 Am. J. Epidemiol. https://doi.org/10.1093/aje/kwx380
12. Ma, N. etÿal. Clinical and epidemiological characteristics of (2004). (2018).
individuals resistant to M. tuberculosis infection in a 24. Hill, P. C. etÿal. Longitudinal assessment of an This study of 872 households containing
longitudinal TB household contact study in Kampala. ELISPOT test for Mycobacterium tuberculosis individuals with TB in Uganda monitors these
BMC Infect. Dis. 14, 352 (2014). infection. PLoS Med. 4, e192 (2007). individuals for ÿ2 years and finds that ~10% of
13. Mandalakas, A. M. etÿal. Well-quantified tuberculosis 25. Aguilar, D. etÿal. Mycobacterium tuberculosis strains with household contacts are resisters.
exposure is a reliable surrogate measure of the Beijing genotype demonstrate variability in virulence 37. Andrews, J. R., Morrow, C. & Wood, R. Modeling the
tuberculosis infection. Int. J. Tuberc Lung Dis. 16, associated with transmission. Tuberculosis role of public transportation in sustaining tuberculosis
1033–1039 (2012). 90, 319–325 (2010). transmission in South Africa. Am. J. Epidemiol. 177,
14. Cowley, S. C. & Elkins, K. L. CD4+ Tÿcells mediate 26. Parwati, I., van Crevel, R. & van Soolingen, D. Possible 556–561 (2013).
IFN-gamma-independent control of Mycobacterium underlying mechanisms for successful emergence of 38. Classen, C. N. etÿal. Impact of social interactions in the
tuberculosis infection both inÿvitro and inÿvivo. the Mycobacterium tuberculosis Beijing genotype community on the transmission of tuberculosis in a
J. Immunol. 171, 4689–4699 (2003). strains. Lancet Infect. Dis. 10, 103–111 (2010). high incidence area. Thorax 54, 136–140 (1999).
15. Sakai, S. etÿal. CD4 Tÿcell-derived IFN-gamma plays a 27. Merker, M. etÿal. Evolutionary history and global
minimal role in control of pulmonary Mycobacterium spread of the Mycobacterium tuberculosis Beijing 39. Verver, S. etÿal. Proportion of tuberculosis
tuberculosis infection and must be actively repressed lineage. Nat. Genet. 47, 242–249 (2015). transmission that takes place in households in a

Reviews
high-incidence area. Lancet 363, 212–214 61. Sepulveda, R. L. etÿal. Evaluation of tuberculin 83. Behar, S. M., Divangahi, M. & Remold, H. G. Evasion of
(2004). reactivity in BCG-immunized siblings. Am. J. Respir. innate immunity by Mycobacterium tuberculosis: is death an
40. Beckman, E. M. etÿal. Recognition of a lipid antigen by CD1- Crit. Which Med. 149, 620–624 (1994). exit strategy? Nat. Rev. Microbiol. 8, 668–674
restricted alpha beta+ Tÿcells. Nature 372, 691–694 62. Horne, D. J. etÿal. Human ULK1 Variation and (2010).
(1994). susceptibility to Mycobacterium tuberculosis 84. Verrall, A. J., Netea, M. G., Alisjahbana, B., Hill, P. C. & van
41. Gold, M. C. etÿal. Human mucosal associated invariant Tÿcells infection. J. Infect. Dis. 214, 1260–1267 (2016). Crevel, R. Early clearance of Mycobacterium tuberculosis:
detect bacterially infected cells. PLoS Biol. 8, e1000407 63. Shah, J. A. etÿal. A functional TOLLIP variant is a new frontier in prevention. Immunology
(2010). associated with BCG-specific immune responses and 141, 506–513 (2014).
42. Tanaka, Y. etÿal. Natural and synthetic non-peptide tuberculosis. Am. J. Respir. Crit. Care Med. 196, 502– 85. Arbour, N. C. etÿal. TLR4 mutations are associated
antigens recognized by human gamma delta Tÿcells. 511 (2017). with endotoxin hyporesponsiveness in humans. Nat.
Nature 375, 155–158 (1995). This study identifies a SNP in the promoter region of Genet. 25, 187–191 (2000).
43. Matos, D. S. etÿal. Differential interferon- gamma TOLLIP that reduces the expression of TOLLIP (a 86. Hawn, T. R. etÿal. A common dominant TLR5 stop
production characterizes the cytokine responses to negative regulator of signalling through TLR2, TLR4 codon polymorphism abolishes flagellin signaling and is
Leishmania and Mycobacterium leprae antigens in and IL-1R) and finds that this polymorphism is associated with susceptibility to legionnaires’ disease.
concomitant mucocutaneous leishmaniasis and associated with increased risk of LTBI in J. Exp. Med. 198, 1563–1572 (2003).
lepromatous leprosy. Clin. Infect. Dis. 40, e5–12 (2005). M. tuberculosis-exposed individuals. 87. Hawn, T. R. etÿal. A common human TLR1
44. Heimbeck, J. Incidence of tuberculosis in young adult women, 64. Cobat, A. etÿal. Two loci control tuberculin skin test reactivity polymorphism regulates the innate immune response to
with special reference to employment. in an area hyperendemic for tuberculosis. lipopeptides. Eur. J. Immunol. 37, 2280–2289 (2007).
Br. J. Tuberc. 32, 154–166 (1938). J. Exp. With. 206, 2583–2591 (2009).
45. Israel, H. L., Hetherington, H. W. & Ord, J. G. A study of This study carries out a genome-wide linkage 88. Hoffman, HM, Mueller, JL, Broide, DH,
tuberculosis among students of nursing. JAMA 117, 839–844 analysis in South Africans and identifies two loci that Wanderer, A. A. & Kolodner, R. D. Mutation of a new gene
(1941). are associated with a negative result in the PPD skin encoding a putative pyrin-like protein causes familial
46. Menzies, D. Interpretation of repeated tuberculin tests. reactivity test; one of these loci has been subsequently cold autoinflammatory syndrome and Muckle-Wells
Boosting, conversion, and reversion. linked to a SNP that is associated with TNF syndrome. Nat. Genet. 29, 301–305 (2001).
I have. J. Breathe. Crit. Which Med. 159, 15–21 (1999). production (see reference 65). 89. Ma, X. etÿal. Full-exon resequencing reveals toll-like receptor
47. Chaussé, P. New research on the contagion of tuberculosis 65. Cobat, A. etÿal. Identification of a major locus, TNF1, that variants contribute to human susceptibility to tuberculosis
through the air exhaled during coughing. controls BCG-triggered tumor necrosis factor production disease. PLoS ONE 2, e1318 (2007).
Ann. Inst. Pasteur 30, 613–641 (1916). by leukocytes in an area hyperendemic for tuberculosis. 90. Meyer, C. G. etÿal. TLR1 variant H305L associated with
48. Lange, B. The epidemiology of tuberculosis. Clin. Infect. Dis. 57, 963–970 (2013). protection from pulmonary tuberculosis.
Central Journal for Bacteriology, Parasitology and PLoS ONE 11, e0156046 (2016).
Infectious Diseases 127, 25–46 (1933). 66. Sobota, R. S. etÿal. A chromosome 5q31.1 locus 91. Johnson, C. M. etÿal. Cutting edge: a common
49. Dean, M. etÿal. Genetic restriction of HIV-1 infection and associates with tuberculin skin test reactivity in HIV- polymorphism impairs cell surface trafficking and
progression to AIDS by a deletion allele of the CKR5 positive individuals from tuberculosis hyper- functional responses of TLR1 but protects against
structural gene. hemophilia growth and development endemic regions in east Africa. PLoS Genet. 13, e1006710 leprosy. J. Immunol. 178, 7520–7524 (2007).
study, multicenter AIDS cohort study, multicenter (2017). 92. Velez, D. R. etÿal. Variants in toll-like receptors 2 and 9
hemophilia cohort study, San Francisco city cohort, ALIVE 67. von Hertzen, L., Klaukka, T., Mattila, H. & Haahtela, T. influence susceptibility to pulmonary tuberculosis in
Study. Science 273, 1856–1862 (1996). Mycobacterium tuberculosis infection and the Caucasians, African-Americans, and West Africans.
subsequent development of asthma and allergic Hum. Genet. 127, 65–73 (2010).
50. Lindesmith, L. etÿal. Human susceptibility and conditions. J. Allergy Clin. Immunol. 104, 1211–1214 (1999). 93. Davila, S. etÿal. Genetic association and expression studies
resistance to Norwalk virus infection. Nat. Med. 9, 548– indicate a role of toll-like receptor 8 in pulmonary
553 (2003). 68. von Mutius, E. et al. International patterns of tuberculosis. PLoS Genet. 4, e1000218 (2008).
51. Lindesmith, L. C. etÿal. Mechanisms of GII.4 norovirus tuberculosis and the prevalence of symptoms of
persistence in human populations. PLoS Med. 5, e31 asthma, rhinitis, and eczema. Thorax 55, 449–453 (2000). 94. Hawn, T. R. etÿal. A polymorphism in Toll-interleukin 1 receptor
(2008). domain containing adaptor protein is associated
52. Miller, L. H., Mason, S. J., Clyde, D. F. & 69. Daya, M., vanÿder Merwe, L., van Helden, P. D., Moller, with susceptibility to meningeal tuberculosis. J.
McGinniss, M. H. The resistance factor to M. & Hoal, E. G. Investigating the role of gene-gene Infect. Dis. 194, 1127–1134 (2006).
Plasmodium vivax in blacks. The Duffy-blood-group interactions in TB susceptibility. PLoS ONE 95. Khor, C. C. etÿal. A Mal functional variant is associated with
genotype, FyFy. N. Engl. J. Med. 295, 302–304 (1976). 10, e0123970 (2014). protection against invasive pneumococcal disease,
70. Stein, C. M. etÿal. Genomics of human pulmonary bacteremia, malaria and tuberculosis.
53. Comstock, G. W. Tuberculosis in twins: a re-analysis of tuberculosis: from genes to pathways. Curr. Genet. Nat. Genet. 39, 523–528 (2007).
the Prophit survey. Am. Rev. Respir. Dis. 117, 621– With. Rope. 5, 149–166 (2017). 96. Thuong, N. T. etÿal. A polymorphism in human TLR2 is
624 (1978). 71. Zhang, F. etÿal. Identification of two new loci at IL23R and associated with increased susceptibility to
54. Kallmann, F. & Reisner, D. Twin studies on the RAB32 that influence susceptibility to leprosy. tuberculous meningitis. Genes Immun. 8, 422–428 (2007).
significance of genetic factors in tuberculosis. Am. Rev. Nat. Genet. 43, 1247–1251 (2011).
Tuberc. 47, 549–574 (1943). 72. Hussell, T. & Bell, T. J. Alveolar macrophages: 97. Shah, J. A. etÿal. Human TOLLIP regulates TLR2 and TLR4
55. Casanova, J.-L. & Abel, L. Genetic dissection of plasticity in a tissue-specific context. Nat. Rev. signaling and its polymorphisms are associated with
immunity to mycobacteria: the human model. Immunol. 14, 81–93 (2014). susceptibility to tuberculosis. J. Immunol. 189, 1737–1746
Year. Rev. Immunol. 20, 581–620 (2002). 73. Kopf, M., Schneider, C. & Nobs, S. P. The development and (2012).
56. Hill, A. V. Aspects of genetic susceptibility to human function of lung-resident macrophages and dendritic 98. Shah, J. A. etÿal. Genetic variation in toll-interacting protein
infectious diseases. Annu. Rev. Genet. 40, 469–486 cells. Nat. Immunol. 16, 36–44 (2015). is associated with leprosy susceptibility and cutaneous
(2006). 74. Torrelles, JB & Schlesinger, LS Integrating lung expression of interleukin 1 receptor antagonist. J.
57. Stein, C. M. etÿal. Genome scan of M. tuberculosis physiology, immunology, and tuberculosis. Infect. Dis. 213, 1189–1197 (2016).
infection and disease in Ugandans. PLoS ONE 3, Trends Microbiol. 25, 688–697 (2017). [ PubMed ] 99. Barreiro, LB et al. Promote variation in the
e4094 (2008). 75. Liu, C. H., Liu, H. & Ge, B. Innate immunity in DC-SIGN-encoding gene CD209 is associated with
In this study, a genome scan identifies two loci tuberculosis: host defense versus pathogen evasion. tuberculosis. PLoS Med. 3, e20 (2006).
(chromosome 2q21–22q24 and chromosome 5p13– Cell. Mol. Immunol. 14, 963–975 (2017). 100. Intemann, C. D. etÿal. Autophagy gene variant IRGM -261T
15q22) that are linked to the resister phenotype; 76. Warren, E., Teskey, G. & Venketaraman, V. Effector contributes to protection from tuberculosis caused by
the latter locus overlaps with a region that was mechanisms of neutrophils within the innate immune system Mycobacterium tuberculosis but not by M. africanum strains.
associated with the intensity of reactivity in the PPD in response to Mycobacterium tuberculosis PLoS Pathog. 5, e1000577 (2009).
skin test (that is, induration in millimetres) in infection. J. Clin. Med. 6, 15 (2017).
an independent cohort study (reference 64). 77. Mortaz, E. etÿal. Interaction of pattern recognition receptors 101. Lu, Y. etÿal. Association of autophagy-related IRGM
with mycobacterium tuberculosis. J. Clin. polymorphisms with latent versus active tuberculosis
58. Thye, T. etÿal. Genome-wide association analyses Immunol. 35, 1–10 (2015). infection in a Chinese population. Tuberculosis 97, 47–
identifies a susceptibility locus for tuberculosis on 78. Stamm, C. E., Collins, A. C. & Shiloh, M. U. Sensing of 51 (2016).
chromosome 18q11.2. Nat. Genet. 42, 739–741 (2010). Mycobacterium tuberculosis and consequences to both 102. Tobin, D. M. etÿal. Host genotype-specific therapies can
host and bacillus. Immunol. Rev. 264, 204–219 (2015). optimize the inflammatory response to
This analysis of a large population (11,425 patients) from mycobacterial infections. Cell 148, 434–446 (2012).
two independent genome-wide association studies 79. Groschel, M. I., Sayes, F., Simeone, R., Majlessi, L. & Brosch, 103. Picard, C. etÿal. Clinical features and outcome of
comparing patients with TB and healthy individuals R. ESX secretion systems: mycobacterial evolution to patients with IRAK-4 and MyD88 deficiency. Medicine
identifies a SNP (on chromosome 18q11.2) that counter host immunity. Nat. Rev. 89, 403–425 (2010).
is associated with TB susceptibility, which Microbiol. 14, 677–691 (2016). 104. Gutierrez, M. G. etÿal. Autophagy is a defense
demonstrates that non-MHC susceptibility loci can be 80. Fabri, M. etÿal. Vitamin D is required for IFN-gamma- mechanism inhibiting BCG and Mycobacterium
identified, even in populations with high genetic mediated antimicrobial activity of human tuberculosis survival in infected macrophages. Cell
diversity, when the sample size is increased by macrophages. Sci. Transl. Med. 3, 104ra102 (2011). 119, 753–766 (2004).
combining data from multiple studies. 81. Collins, A. C. etÿal. Cyclic GMP-AMP synthase is an innate 105. Singh, S. B., Davis, A. S., Taylor, G. A. & Deretic, V.
59. Azad, A. K., Sadee, W. & Schlesinger, L. S. Innate immune DNA sensor for Mycobacterium tuberculosis. Human IRGM induces autophagy to eliminate
immune gene polymorphisms in tuberculosis. Cell Host Microbe 17, 820–828 (2015). intracellular mycobacteria. Science 313, 1438–1441 (2006).
Infect. Immun. 80, 3343–3359 (2012).
60. Tosh, K. etÿal. Variants in the SP110 gene are 82. Watson, R. O. etÿal. The cytosolic sensor cGAS detects 106. Kumar, D. etÿal. Genome-wide analysis of the host
associated with genetic susceptibility to tuberculosis in Mycobacterium tuberculosis DNA to induce type I intracellular network that regulates survival of
West Africa. Proc. Natl Acad. Sci. USA 103, 10364– interferons and activate autophagy. Cell Host Microbe Mycobacterium tuberculosis. Cell 140, 731–743 (2010).
10368 (2006). 17, 811–819 (2015).

Reviews
107. Watson , RO , Chamomile , PS & Cox , JS 126. Boussiotis, V. A. etÿal. IL-10-producing Tÿcells suppress 146. Spencer, C. T., Abate, G., Blazevic, A. & Hoft, D. F.
Extracellular M. tuberculosis DNA targets bacteria for immune responses in anergic tuberculosis patients. Only a subset of phosphoantigen-responsive 9 2T cells
autophagy by activating the host DNA-sensing J. Clin. Invest. 105, 1317–1325 (2000). mediate protective tuberculosis immunity. J. Immunol.
pathway. Cell 150, 803–815 (2012). 127. Delgado, J. C. etÿal. Antigen-specific and persistent 181, 4471–4484 (2008).
108. Majlessi, L. & Brosch, R. Mycobacterium tuberculosis tuberculin anergy in a cohort of pulmonary 147. Plotkin, S. A. Vaccines: correlates of vaccine-induced
meets the cytosol: the role of cGAS in tuberculosis patients from rural Cambodia. Proc. Natl immunity. Clin. Infect. Dis. 47, 401–409 (2008).
anti-mycobacterial immunity. Cell Host Microbe 17, 733– Acad. Sci. 99, 7576–7581 (2002). 148. Plotkin, S. A. Correlates of protection induced by
735 (2015). This investigation of a subset of patients with vaccination. Clin. Vaccine Immunol. 17, 1055–1065
109. Romagnoli, A. etÿal. ESX-1 dependent impairment of pulmonary TB indicates that persistent inÿvivo (2010).
autophagic flux by Mycobacterium tuberculosis in anergy in the PPD skin reactivity test following 149. Achkar, J. M., Chan, J. & Casadevall, A. B cells and
human dendritic cells. Autophagy 8, 1357–1370 treatment completion is not a global anergy but antibodies in the defense against Mycobacterium
(2012). instead is PPD antigen-specific and is correlated tuberculosis infection. Immunol. Rev. 264, 167–181
110. Castillo, E. F. etÿal. Autophagy protects against active with elevated IL-10 production and reduced (2015).
tuberculosis by suppressing bacterial burden and secretion of IL-2 and IFNÿ by Tÿcells following ex 150. Kozakiewicz, L., Phuah, J., Flynn, J. & Chan, J. The
inflammation. Proc. Natl Acad. Sci. USA 109, vivo stimulation with PPD antigens. role of B cells and humoral immunity in
E3168–E3176 (2012). 128. Redford, P. S., Murray, P. J. & O’Garra, A. The role Mycobacterium tuberculosis infection. Adv. Exp. Med.
111. Kimmey, J. M. etÿal. Unique role for ATG5 in of IL-10 in immune regulation during M. tuberculosis Biol. 783, 225–250 (2013).
neutrophil-mediated immunopathology during M. infection. Mucosal Immunol. 4, 261–270 (2011). 151. Balu, S. etÿal. A novel human IgA monoclonal antibody
tuberculosis infection. Nature 528, 565–569 (2015). 129. Thye, T. etÿal. IL10 haplotype associated with protects against tuberculosis. J. Immunol. 186,
tuberculin skin test response but not with pulmonary TB. 3113–3119 (2011).
112. Seshadri, C. etÿal. Transcriptional networks are PLoS ONE 4, e5420 (2009). 152. Hamasur, B. etÿal. A mycobacterial lipoarabinomannan
associated with resistance to Mycobacterium 130. Okamoto Yoshida, Y. etÿal. Essential role of IL-17A in the specific monoclonal antibody and its F(ab’) fragment
tuberculosis infection. PLoS ONE 12, e0175844 formation of a mycobacterial infection-induced prolong survival of mice infected with Mycobacterium
(2017). granuloma in the lung. J. Immunol. 184, 4414–4422 (2010). tuberculosis. Clin. Exp. Immunol. 138, 30–38 (2004).
In this study, the transcriptional responses of 153. Suter, E. Multiplication of tubercle bacilli within
monocytes from resisters and individuals with 131. Wozniak, T. M., Saunders, B. M., Ryan, A. A. & mononuclear phagocytes in tissue cultures derived
LTBI following ex vivo M. tuberculosis infection are Britton, W. J. Mycobacterium bovis BCG-specific from normal animals and animals vaccinated with
shown to be distinct, and network analysis Th17 cells confer partial protection against BCG. J. Exp. Med. 97, 235–245 (1953).
identifies HDACs as potential mediators of this Mycobacterium tuberculosis infection in the absence of 154. Teitelbaum, R. etÿal. A mAb recognizing a surface
difference. gamma interferon. Infect. Immun. 78, 4187–4194 (2010). antigen of Mycobacterium tuberculosis enhances host
113. Subramanian, A. etÿal. Gene set enrichment analysis: a survival. Proc. Natl Acad. Sci. USA 95, 15688–15693
knowledge-based approach for interpreting 132. Gallegos, A. M. etÿal. A gamma interferon independent (1998).
genome-wide expression profiles. Proc. Natl Acad. mechanism of CD4 Tÿcell mediated control of M. 155. Glatman-Freedman, A. & Casadevall, A. Serum
Sci. USA 102, 15545–15550 (2005). tuberculosis infection inÿvivo. PLoS Pathog. 7, therapy for tuberculosis revisited: reappraisal of the role
114. Joseph, J. etÿal. Expression profiling of sodium e1002052 (2011). of antibody-mediated immunity against
butyrate (NaB)-treated cells: identification of 133. Han, M. & Hannick, LI, DiBrino, M. & Robinson, MA Mycobacterium tuberculosis. Clin. Microbiol. Rev. 11, 514–
regulation of genes related to cytokine signaling and Polymorphism of human CD1 genes. Tissue Antigens 532 (1998).
cancer metastasis by NaB. Oncogene 23, 6304–6315 54, 122–127 (1999). 156. Fletcher, H. A. etÿal. Tÿcell activation is an immune
(2004). 134. Oteo, M. etÿal. Single strand conformational correlate of risk in BCG vaccinated infants. Nat.
115. Chandran, A. etÿal. Mycobacterium tuberculosis polymorphism analysis of human CD1 genes in Common. 7, 11290 (2016).
infection induces HDAC1-mediated suppression of different ethnic groups. Tissue Antigens 53, 545–550 157. Phuah, J. etÿal. Effects of B cell depletion on early
IL-12B gene expression in macrophages. Front. Cell. (1999). Mycobacterium tuberculosis infection in cynomolgus
Infect. Microbiol. 5, 90 (2015). 135. Seshadri, C. et al. T cell responses against macaques. Infect. Immun. 84, 1301–1311 (2016).
116. Wang, Y., Curry, H. M., Zwilling, B. S. & Lafuse, W. P. mycobacterial lipids and proteins are poorly 158. Abreu, M. T. etÿal. Alterations in the peripheral blood
Mycobacteria inhibition of IFN-gamma induced HLA- correlated in South African adolescents. J. Immunol. B cell subpopulations of multidrug-resistant
DR gene expression by up-regulating histone 195, 4595–4603 (2015). tuberculosis patients. Clin. Exp. Med. 14, 423–429
deacetylation at the promoter region in human THP-1 This study shows that the response of CD4+ Tÿcells (2014).
monocytic cells. J. Immunol. 174, 5687–5694 to mycobacterial lipid (CD1-restricted) antigens ex 159. Joosten, S. A. etÿal. Patients with tuberculosis have a
(2005). vivo correlates poorly with their response to dysfunctional circulating B-cell compartment, which
117. Coussens, A. K., Wilkinson, R. J. & Martineau, A. R. peptide antigens. However, as the abundance of normalizes following successful treatment. PLoS
Phenylbutyrate is bacteriostatic against lipid-specific Tÿcells does not differ between Pathog. 12, e1005687 (2016).
Mycobacterium tuberculosis and regulates the infected and uninfected individuals in this study, it is 160. Phuah, JY, Mattila, JT, Lin, PL & Flynn, JL
macrophage response to infection, synergistically with 25- plausible that individuals negative for the IGRA are Activated B cells in the granulomas of nonhuman
hydroxy-vitamin D3. PLoS Pathog. 11, e1005007 (2015). M. tuberculosis-exposed and that their Tÿcells primates infected with Mycobacterium tuberculosis.
produce a similar repertoire of cytokines as those Am. J. Pathol. 181, 508–514 (2012).
118. Mily, A. etÿal. Significant effects of oral phenylbutyrate and from IGRA-positive individuals, although they do 161. Tsai, M. C. etÿal. Characterization of the tuberculous
vitamin D3 adjunctive therapy in pulmonary not secrete IFNÿ. granuloma in murine and human lungs: cellular
tuberculosis: a randomized controlled trial. PLoS ONE 136. Van Rhijn, I., Ly, D. & Moody, DB CD1a, CD1b, and CD1c composition and relative tissue oxygen tension.
10, e0138340 (2015). in immunity against mycobacteria. Adv. Exp. Cell. Microbiol. 8, 218–232 (2006).
119. Behar, S. M., Dascher, C. C., Grusby, M. J., Wang, C. R. With. Biol. 783, 181–197 (2013). 162. Casadevall, A. & Pirofski, LA A reappraisal of
& Brenner, M. B. Susceptibility of mice deficient in 137. Kjer-Nielsen, L. et al. MR1 presents microbial vitamin B humoral immunity based on mechanisms of antibody-
CD1D or TAP1 to infection with Mycobacterium metabolites to MAIT cells. Nature 491, 717–723 (2012). mediated protection against intracellular pathogens.
tuberculosis. J. Exp. Med. 189, 1973–1980 (1999). Adv. Immunol. 91, 1–44 (2006).
120. Lin, P. L. etÿal. CD4 Tÿcell depletion exacerbates acute 138. Martin, E. etÿal. Stepwise development of mait cells in 163. Nimmerjahn, F. & Ravetch, J. V. Fcgamma receptors as
Mycobacterium tuberculosis while reactivation of latent mouse and human. PLoS Biol. 7, 0525–0536 (2009). regulators of immune responses. Nat. Rev. Immunol.
infection is dependent on severity of tissue depletion 8, 34–47 (2008).
in cynomolgus macaques. AIDS Res. Hum. 139. Porcelli, S., Yockey, C. E., Brenner, M. B. & Balk, S. P. 164. Lu, L. L. etÿal. A functional role for antibodies in
Retroviruses 28, 1693–1702 (2012). Analysis of Tÿcell antigen receptor (TCR) expression by tuberculosis. Cell 167, 433–443 e414 (2016).
121. Mogues, T., Goodrich, M. E., Ryan, L., LaCourse, R. & human peripheral blood CD4-8- alpha/beta Tÿcells In this study, antibody profiling indicates that the
North, R. J. The relative importance of Tÿcell subsets in demonstrates preferential use of several V beta genes and pattern of glycosylation in the Fc domain of
immunity and immunopathology of airborne an invariant TCR alpha chain. J. Exp. Med. 178, 1–16 antibodies from individuals with LTBI is distinct
Mycobacterium tuberculosis infection in mice. J. Exp. (1993). from that of antibodies from patients with active
With. 193, 271–280 (2001). 140. Treiner, E. etÿal. Selection of evolutionarily conserved TB and that addition of latent antibodies to
122. Flynn, J. L. etÿal. An essential role for interferon mucosal-associated invariant Tÿcells by MR1. Nature macrophages previously infected with
gamma in resistance to Mycobacterium tuberculosis 422, 164–169 (2003). M. tuberculosis results in improved phagolysosomal
infection. J. Exp. Med. 178, 2249–2254 141. Chua, W. J. etÿal. Polyclonal mucosa-associated maturation and killing of intracellular bacteria by
(1993). invariant Tÿcells have unique innate functions in these macrophages.
123. Green, A. M., Difazio, R. & Flynn, J. L. IFN-gamma bacterial infection. Infect. Immun. 80, 3256–3267 (2012). 165. Riedel, S. Edward Jenner and the history of smallpox and
from CD4 Tÿcells is essential for host survival and vaccination. Proc 18, 21–25 (2005).
enhances CD8 Tÿcell function during Mycobacterium 142. Seshadri, C. etÿal. A polymorphism in human MR1 is 166. Liu, R. etÿal. Homozygous defect in HIV-1 coreceptor
tuberculosis infection. J. Immunol. 190, 270–277 associated with mRNA expression and susceptibility to accounts for resistance of some multiply-exposed
(2013). tuberculosis. Genes Immun. 18, 8–14 (2017). individuals to HIV-1 infection. Cell 86, 367–377
124. Kagina, B. M. N. etÿal. Specific Tÿcell frequency and 143. Seshadri, C. etÿal. A polymorphism in human CD1A is (1996).
cytokine expression profile do not correlate with associated with susceptibility to tuberculosis. Genes 167. Samson, M. etÿal. Resistance to HIV-1 infection in
protection against tuberculosis after bacillus Calmette- Immun. 15, 195–198 (2014). caucasian individuals bearing mutant alleles of the
Guérin vaccination of newborns. Am. J. Respir. Crit. 144. Sandstrom, A. etÿal. The intracellular B30.2 domain of CCR-5 chemokine receptor gene. Nature 382,
Which Med. 182, 1073–1079 (2010). butyrophilin 3A1 binds phosphoantigens to mediate 722–725 (1996).
125. Tameris, M. D. etÿal. Safety and efficacy of MVA85A, a activation of human Vÿ9Vÿ2T Cells. Immunity 40, 168. Gulick, R. M. etÿal. Maraviroc for previously treated
new tuberculosis vaccine, in infants previously 490–500 (2014). patients with R5 HIV-1 infection. N. Engl. J. Med. 359, 1429–
vaccinated with BCG: a randomised, placebo- 145. Vavassori, S. et al. Butyrophilin 3A1 binds 1441 (2008).
controlled phase 2b trial. Lancet 381, 1021–1028 (2013). phosphorylated antigens and stimulates human ÿÿ 169. Jones-Lopez, E. C. etÿal. Cough Aerosols of
Tÿcells. Nat. Immunol. 14, 908–916 (2013). Mycobacterium tuberculosis in the prediction of

Reviews
incident tuberculosis disease in household contacts. 176. Dickie, H. A. Tuberculosis in student nurses and Witwatersrand, Johannesburg, South Africa, and the Advancing
Clin. Infect. Dis. 63, 10–20 (2016). medical students at the University of Wisconsin. Care and Treatment for TB/HIV, South African Medical Research
170. Patterson, B. etÿal. Detection of Mycobacterium Ann. Internal. With. 33, 941–959 (1950). Council, Johannesburg, South Africa. E.S. is also affiliated with
tuberculosis bacilli in bio-aerosols from untreated TB 177. Stein, C. M. etÿal. Resistance and susceptibility to the McGill International TB Centre, McGill University, Montreal,
patients. Gates Open Res. 1, 11 (2017). Mycobacterium tuberculosis infection and disease in Quebec, Canada.
171. Zwerling, A. etÿal. Trajectories of tuberculosis-specific tuberculosis households in Kampala, Uganda.
interferon-gamma release assay responses among Am. J. Epidemiol. https://doi.org/10.1093/aje/kwx380 Author contributions
medical and nursing students in rural India. (2018). J.D.S., C.S. and T.R.H. contributed to the drafting and editing of
J. Epidemiol. Glob. Health 3, 105–117 (2013). all sections. Focused contributions were made to the
172. Gagneux, S. etÿal. Variable host-pathogen Acknowledgements Epidemiology and Genetics sections (C.M.S., E.S. and G.C.),
compatibility in Mycobacterium tuberculosis. Proc. This work was supported by grants from the US National the B cell-mediated resistance section (G.A.) and the Macrophage-
Natl Acad. Know USA 103, 2869–2873 (2006). Institutes of Health grants R01AI124348 (to W.H.B., T.R.H., mediated resistance section (M.C.). Initial figure drafts were
173. Hardy, M. A. & Schmidek, H. H. Epidemiology of C.S., C.M.S. and H.M.-K.), U01AI115642 (to W.H.B., T.R.H., provided by J.D.S., M.C. and G.A. Detailed review and editing
tuberculosis aboard a ship. JAMA 203, 175–179 C.S., C.M.S. and H.M.-K.), R01AI124349 (to E.S.) and were additionally provided by S.F., R.S.W., W.H.B. and H.M.-K.
(1968). T32AI007044 (to J.D.S.) and contract number NO1AI70022 (to
174. Badger, T. L. & Spink, W. W. First-infection type of W.H.B., C.M.S., T.R.H. and H.M.-K.); the Bill and Melinda Gates
tuberculosis in adults — a five-year study of student Foundation grant OPP1151836 (to T.R.H., W.H.B., C.S., C.M.S., Competing interests
nurses at the Boston City Hospital. New Engl. J. Med. H.M.-K., G.C. and R.S.W.) and grant OPP1151840 (to G.A. and The authors declare no competing interests.
217, 424–431 (1937). S.F.); the Canadian Institutes of Health Research grant
175. Myers, J. A., Boynton, R. E. & Diehl, R. E. Prevention of FDN143332 (to E.S.); and the South African Medical Research Publisher’s note
tuberculosis among students of nursing. Council grant ACT4TB/HIV (to G.C. and R.S.W.). G.C. is also Springer Nature remains neutral with regard to jurisdictional
Am. J. Nurs. 47, 661–666 (1947). affiliated with the School of Public Health, University of claims in published maps and institutional affiliations.

Immunological Mechanisms of Human Resistance To Persistent

Uploaded by

Copyright:

Available Formats

Immunological Mechanisms of Human Resistance To Persistent

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Immunological Mechanisms of Human Resistance To Persistent

Uploaded by

Copyright:

Available Formats

Machine Translated by Google

Immunological mechanisms of human

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 575

576 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 577

owing to the inherent limitations of an immunolog-ical

in recent follow-up studies, all resisters demonstrated

sensitized some individuals in the absence of natural

578 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 579

580 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

Fig. 2 | Macrophage-mediated resistance to Mycobacterium tuberculosis. Macrophages can provide resistance

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 581

582 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

Fig. 3 | Tÿcell-mediated resistance to Mycobacterium tuberculosis. After phagocytosis of Mycobacterium tuberculosis

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 583

584 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 585

586 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 587

588 | SEPTEMBER 2018 | volume 18 www.nature.com/nri

NATuRe RevIewS | IMMunoLogy volume 18 | SEPTEMBER 2018 | 589

You might also like