The End of the Binary Era: Revisiting the

Spectrum of Tuberculosis
Philana Ling Lin and JoAnne L. Flynn
This information is current as J Immunol 2018; 201:2541-2548; ;
of October 22, 2018. doi: 10.4049/jimmunol.1800993
http://www.jimmunol.org/content/201/9/2541

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

References This article cites 54 articles, 11 of which you can access for free at:
http://www.jimmunol.org/content/201/9/2541.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision
• No Triage! Every submission reviewed by practicing scientists
• Fast Publication! 4 weeks from acceptance to publication
*average

Subscription Information about subscribing to The Journal of Immunology is online at:

http://jimmunol.org/subscription
Permissions Submit copyright permission requests at:
http://www.aai.org/About/Publications/JI/copyright.html
Email Alerts Receive free email-alerts when new articles cite this article. Sign up at:
http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by

The American Association of Immunologists, Inc.,
1451 Rockville Pike, Suite 650, Rockville, MD 20852
Copyright © 2018 by The American Association of
Immunologists, Inc. All rights reserved.
Print ISSN: 0022-1767 Online ISSN: 1550-6606.
The End of the Binary Era: Revisiting the Spectrum of
Tuberculosis
Philana Ling Lin* and JoAnne L. Flynn†
Human Mycobacterium tuberculosis infection was vaccine, or even correlates of protective immunity, has
thought to result in either active symptomatic tuber- proved difficult.
culosis (TB) or latent asymptomatic infection. It is Mycobacterium tuberculosis, the bacterium that causes TB, is
now clear that this binary classification is insufficient an intracellular pathogen transmitted in aerosolized droplets,
to describe the myriad of infection outcomes. In active typically by coughing from a person with active TB disease. In
TB, symptomatic disease can be mild to severe, with a the airways, the bacillus primarily infects alveolar macro-
range of lung and thoracic lymph node involvement or phages but can infect other cell types. It then transits to the
extrapulmonary manifestations. Most humans control lung parenchyma where it can infect resident macrophages

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

the infection and develop latent TB infection, with or other phagocytes, including neutrophils. The ability to
differential risks of reactivation to active TB. However, replicate in macrophages is critical to the pathogenesis of
some frequently exposed persons appear to be resistant M. tuberculosis. Immune signals produced by infected mac-
to infection, whereas others may initially become in- rophages and bacterial products recruit other immune cells,
fected yet subsequently eliminate all bacilli. The im- including monocytes, which differentiate into macrophages
and provide additional targets for infection. Simultaneously,
munologic factors influencing these varied outcomes
the bacilli can be taken up by dendritic cells and transported
are still not clear, but likely involve a range of different
to the thoracic lymph nodes where T cells are primed against
responses. In this article, we review the data supporting
a broad range of M. tuberculosis Ags. These T cells return to
the spectrum of M. tuberculosis infection in humans as the site of infection in the lung and organize around infected
well as data in nonhuman primates that allow dissec- macrophages to form granulomas, the pathologic structure
tion of the immune responses leading to the varied most closely associated with TB (Fig. 1). Granulomas also
outcomes of infection. The Journal of Immunology, form in thoracic lymph nodes where they serve as a reservoir
2018, 201: 2541–2548. of infection and dissemination for M. tuberculosis.
The course of M. tuberculosis infection is extremely variable
in humans. Once infection is established, most people control

T
uberculosis (TB) remains a major cause of morbidity
and mortality worldwide and is now the most com- but do not eliminate the initial infection. Infected but
mon cause of death from an infectious disease, sur- asymptomatic individuals are classified as having latent
passing HIV/AIDS (1). Even with improved measures to M. tuberculosis infection (LTBI). Persons with LTBI are
diagnose and treat TB, the disease continues to ravage de- characterized as having evidence of infection by an immu-
veloping countries, causing 1.5 million deaths in 2016 (2). nologic test [positive tuberculin skin test (TST) or IFN-g
Standard therapy for TB has not changed in almost 50 years release assay (IGRA)], with no signs or symptoms of TB
and involves 6 mo of treatment (four drugs for 2 mo, fol- and are presumed to be noncontagious. A small percentage
lowed by two drugs for an additional 4 mo). This long course (5–15%) of these people progress to active, symptomatic, and
of treatment can result in poor compliance and the potential transmissible TB within 2 y of infection, likely representing a
for developing drug resistance. Drug-resistant TB is compli- lack of initial control of infection; this is termed primary TB.
cated to manage and requires longer treatment with less ef- Symptoms of disease evolve slowly, possibly due to the slow
fective drugs and greater risk of adverse effects. Although there growth of the M. tuberculosis. Not surprisingly, active TB can
are newer drugs being used for drug-resistant TB, an ef- present in a variety of ways, most commonly as pulmonary
fective vaccine that prevents infection or disease is essen- TB, but the bacillus can infect any organ in the body.
tial for ending the scourge of TB. However, identifying Pulmonary TB can present with mild, moderate, or severe
the protective mechanisms necessary for an effective TB respiratory and systemic symptoms, with involvement of

*Department of Pediatrics, Children’s Hospital of Pittsburgh of the University of Address correspondence and reprint requests to Dr. JoAnne L. Flynn, Department of
Pittsburgh Medical Center, University of Pittsburgh School of Medicine, Pittsburgh, Microbiology and Molecular Genetics, University of Pittsburgh School of Medicine,
PA 15224; and †Department of Microbiology and Molecular Genetics, University of Room 5058 Biomedical Science Tower 3, 3501 Fifth Avenue, Pittsburgh, PA 15261.
Pittsburgh School of Medicine, Pittsburgh, PA 15261 E-mail address: joanne@pitt.edu
ORCIDs: 0000-0001-5169-3714 (P.L.L.); 0000-0001-7874-8981 (J.L.F.). Abbreviations used in this article: BCG, bacille Calmette–Guérin; CT, computed to-
mography; FDG, 2-deoxy-2-[18F]-fluoro-D glucose; IGRA, IFN-g release assay; LTBI,
Received for publication July 17, 2018. Accepted for publication July 28, 2018.
latent M. tuberculosis infection; PET, positron emission tomography; TB, tuberculosis;
This work was supported by the Bill and Melinda Gates Foundation (to J.L.F. and P.L.L.), TST, tuberculin skin test.
Aeras (to J.L.F.), and National Institutes of Health Grants AI111871 and AI134195 (to
P.L.L.) and HL110811, AI114674, AI123093, and AI094745 (to J.L.F.). Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$37.50

www.jimmunol.org/cgi/doi/10.4049/jimmunol.1800993
2542 BRIEF REVIEWS: SPECTRUM OF M. TUBERCULOSIS INFECTION

appreciated over the last decade owing to technologic ad-

vances in transcriptional profiling, in vivo imaging, and more
innovative and comprehensive clinical research that has en-
abled us to gain more thorough insights into TB pathogenesis.
The expanding definition of LTBI: resisters and reverters. A growing
appreciation for the complexities of M. tuberculosis exposure
and infection has developed from household contact studies
in which certain individuals, despite multiple exposures to an
index case, have persistently negative TST or IGRA results
and appear to be resistant to M. tuberculosis infection. These
individuals are termed “resisters.” This phenomenon is de-
scribed in some of the older literature and was nicely dem-
onstrated in a detailed and extended contact-tracing study
involving a single TB index case aboard a naval vessel
FIGURE 1. Caseous granuloma from the lung of an M. tuberculosis–infected (known as the Byrd study) (8). In that study, despite
macaque. Original magnification 34. similar and prolonged exposure to the index case, several
subjects remained TST-negative, whereas many others
single or multiple lung lobes. Cavitary TB is seen in some, but became TST-positive (8). Resisters were initially thought to

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

not all, individuals with pulmonary TB. Active TB can also be subjects who were not as frequently or severely exposed to
occur by reactivation of LTBI, resulting in a similar range of the index case. However, household contact investigation
severity as primary TB. It is estimated that there is a 10% studies have since identified subjects with no evidence of
lifetime risk of reactivation in those with LTBI (3). However, infection (9) as long as 2 y after index case exposure
the more recent paradigm of LTBI (4) as a spectrum of in- (10, 11). This is a fascinating cohort of people from which
fection suggests that this risk is not the same for each person clues about resistance to infection can be gleaned; this resister
with LTBI. The factors defining this reactivation risk are phenotype has been better defined recently (reviewed in Ref.
unclear and likely involve both microbiologic and immuno- 12). It should be noted that the current diagnostics for
logic components. M. tuberculosis infection are relatively poor at identifying this
Diagnosis of M. tuberculosis infection is indirect and is resister population. The TST can be cross-reactive to other
based on detecting cellular immune responses against myco- Mycobacterial spp., and the IGRA depends on detecting only
bacterial proteins, either by a positive TST or IGRA. The one immune response to M. tuberculosis infection (IFN-g
TST measures a delayed-type hypersensitivity reaction to the produced in response to two Ags). Thus, it is possible
injection of a poorly defined mixture of mycobacterial Ags, that some IGRA-negative subjects may, in fact, have been
known as purified protein derivative, but can yield false infected but are controlling (or eliminating) the infection by
positive results from cross-reactivity with nontuberculous IFN-g–independent mechanisms or have T cells that recognize
Mycobacterium spp. as well as M. bovis. More recently, IGRA Ags other than ESAT-6 and CFP-10. Functional Ab
blood tests have been used, which rely on measuring IFN-g responses have been shown recently to distinguish LTBI from
produced by T cells ex vivo in response to the M. tuberculosis– active TB cases (13), but there are no data yet on Abs in resister
specific Ags ESAT-6 and CFP-10. These Ags are absent from cohorts. Other immune mechanisms, such as nonclassical
the commonly used TB vaccine bacille Calmette–Guérin T cells (e.g., gd T cells or mucosal-associated invariant
(BCG), an attenuated strain of M. bovis, and thus IGRAs can T cells) or innate immune functions of alveolar macrophages,
distinguish reactions to M. tuberculosis from those to BCG. could be responsible for the apparent lack of infection observed
However, these Ags are present in certain other mycobacteria, in this population (14). In any event, this special group of
such as M. africanum. Thus, although this is an improvement people provide a unique opportunity for more detailed
over purified protein derivative skin testing, IGRAs remain an immunologic and transcriptomic studies to determine the
imperfect test for M. tuberculosis infection. Importantly, nei- mechanisms of protection against infection or disease.
ther IGRA or TST differentiates active TB versus LTBI, and In addition to resisters, another group of interesting subjects
active TB must be diagnosed by the presence of clinical signs, has been recently identified: household contacts of an index
radiography, sputum smear for acid-fast bacilli, or culture for case or those who live in a high endemic area who develop
M. tuberculosis. a positive IGRA but then revert to negative within 2 y
(“reverters”) (15–17). Some cases of reversion likely represent
The spectrum of TB in humans the imprecise nature of the cut-off between positive and
The dogma of the binary nature of M. tuberculosis infection negative IGRA results, given the marginally positive IFN-g
(active TB versus LTBI) is an oversimplified and now out- levels often observed. However, a substantial number of pa-
dated concept. M. tuberculosis infection has a spectrum of tients with high IFN-g responses truly reverted to IGRA-
manifestations (5–7) that encompass a broad range of out- negative within 2 y. This unique subset of patients may
comes, including resisters (no evidence of infection despite represent initial infection with subsequent self-cure or eradi-
repeated exposure to M. tuberculosis); infected initially but cation. A recent vaccine trial in humans found that revacci-
able to eradicate M. tuberculosis; infected but asymptomatic nation with BCG or with the H4 subunit vaccine of
and stable; latently infected but at high risk of reactivation; mycobacterial proteins did not significantly reduce initial
active TB with chronic symptoms; and fulminant, severe tu- infection and IGRA conversion but instead reduced the rate
berculous disease (Fig. 2). This concept has become better of sustained IGRA conversion over the 2-y follow-up (18).
The Journal of Immunology 2543

FIGURE 2. The outcome of M. tuberculosis infection includes a host of outcomes and clinical manifestations. Those with severe outcomes are often highly
symptomatic (e.g., weight loss, chills, night sweats, fevers, cough) with a positive skin test or IGRA, chest x-ray (CXR) with substantial TB disease, and growth of
M. tuberculosis on sputum culture. In contrast, some individuals may have been infected in the past (with or without a positive IGRA/skin test) and have cleared
the infection resulting in the absence of symptoms, negative sputum culture, and normal chest x-ray. Although the greatest proportion of infected individuals are

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

able to control infection (LTBI), the exact distribution of outcomes depicted in this article is not known. PPD, purified protein derivative TST.

BCG revaccination of previously BCG vaccinated subjects more disseminated disease (23). The quality of chest x-rays
resulted in reversion of an initially positive IGRA with an has greatly improved from the time they were first used more
efficacy of 45%, suggesting these vaccines had a robust re- than five decades ago and this has led to improved diagnosis
duction, or possibly elimination, of an initial M. tuberculosis of TB. More sophisticated imaging technologies such as
infection. As with resisters, the biologic mechanisms leading computed tomography (CT) have been used in research and
to this outcome are of interest for development of preventive specific clinical settings, resulting in even more sensitive
strategies and provide new outcome measures for vaccine detection of disease. Positron emission tomography (PET),
trials. Additional studies to identify the true infection status using 2-deoxy-2-[18F]-fluoro-D glucose (FDG, a PET probe
and more detailed analysis of immunologic responses of these measuring metabolic activity), combined with CT has
reverters are critical to more fully understand the spectrum of enabled researchers and clinicians to find metabolically active
LTBI. It is possible that there is a link between resisters and granulomas in otherwise asymptomatic or LTBI patients. In
reverters, with common immunologic mechanisms responsi- fact, a spectrum of image patterns has been observed in subjects
ble for preventing or eliminating infection, respectively. with clinically defined LTBI (24) and in patients with
Subclinical disease. Subclinical TB, often described as having symptomatic active disease (reviewed in Ref. 25). Esmail
detectable M. tuberculosis in sputum but with a normal chest et al. (26) reported on 35 HIV-infected, antiretroviral naive
x-ray and no symptoms of TB, has been documented patients without signs or symptoms of active TB and identified
in immune competent individuals. However, it is likely 10 patients with PET/CT characteristics of active or subclinical
underreported, because sputum examinations are not disease (e.g., infiltrates or metabolically active nodules) and
typically performed on patients without symptoms or with who also exhibited a higher risk of progression to active TB.
negative chest radiographs (8, 19). In the aforementioned In this same study, there were 195 HIV+ patients who were
Byrd study (8), one of the exposed subjects was found to be screened for TB but who were excluded from the study for
sputum positive with a normal chest radiograph and no overt various reasons. Of these 195 patients, 10 were found to be
symptoms. Only after intensive questioning did this sputum positive, of whom five were asymptomatic with normal
individual admit to having 6 wk of cough and weight loss, chest x-rays. Clearly, there is discordance between the
which resolved (8). Surprisingly, recent studies identified traditional classification of LTBI and active TB.
some patients who would have been defined as having LTBI Transcriptional studies of human blood have shown over-
but had a positive sputum sample, performed serendipitously lapping signatures between clinically defined active TB and
for research purposes (19). Similarly, features of active TB LBTI (19, 27) as well as between LTBI and uninfected sub-
have been historically observed at autopsy of patients who jects (28). These studies support the existence of a full spec-
died of other non-TB causes, suggesting that these patients trum of M. tuberculosis infection beyond the dogmatic,
were otherwise asymptomatic of TB (20). With the clinically defined, binary outcome. Recent blood transcrip-
resurgence of TB during the HIV epidemic, clinicians have tional studies identified a signature that could predict active
recognized that immunodeficiency is often associated with TB as early as 18 mo prior to clinical diagnosis (29). Although
atypical clinical and radiographic manifestations of TB. these studies do not provide biological mechanisms, the pre-
HIV-infected patients, especially those with low CD4 dictive signatures reflect ongoing disease evolution that cannot
counts, do not generally have cavity formation and can have otherwise be detected by clinical signs or symptoms. To ac-
atypical or even normal chest radiographs despite having count for this broader classification of M. tuberculosis infec-
confirmed disease (reviewed in Refs. 21 and 22). Patients tion, some have coined the term “incipient TB.” Although
on TNF inhibitors, a major risk factor for reactivation of this term has been disputed since the early 1900s (30), it is
LTBI, also can have atypical manifestations of TB with now defined by the World Health Organization as “prolonged
2544 BRIEF REVIEWS: SPECTRUM OF M. TUBERCULOSIS INFECTION

asymptomatic phase of early disease during which pathology epithelial cells to express DEFB4 and other antimicrobial
evolves, prior to clinical presentation as active disease” (31). effectors that can kill M. tuberculosis directly (36). These very
Tests that can identify this phase of infection will facilitate early events are consistent with our findings in which serial
more strategic treatment of large populations, preferably by PET/CT imaging was performed during M. tuberculosis in-
identifying and treating those at greatest risk of active TB fection in macaques to determine whether the pattern of lung
before they become overtly contagious, thereby reducing granulomas could predict outcome. Animals that would later
transmission. To more fully understand the spectrum of develop active TB had more granulomas as early as 3 wk
M. tuberculosis infection and the underlying mechanisms that postinfection (before the adaptive immune response is estab-
drive outcomes (especially with regard to incipient infection), lished) with significantly more new granulomas developing
more in-depth studies of basic TB pathogenesis and biology because of dissemination from existing granulomas between
of the host and pathogen are required. 3–6 wk compared with animals that would later present with
LTBI (38). These data suggest that innate and early adaptive
Animal models for understanding the spectrum responses are critical to outcome.
of M. tuberculosis infection Bacterial burden within individual granulomas peaks at
The human data that support the idea of a spectrum of ∼4 wk postinfection when there is minimal bacterial killing
M. tuberculosis infection outcomes are compelling. However, detected (39). Substantial bacterial killing occurs within
for practical reasons, much of this data comes from blood, granulomas by 10–11 wk postinfection, coinciding with the
whereas TB is a disease primarily of the lungs. Animal models development of an adaptive immune response. RNA tran-

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

have contributed greatly to our understanding of the key scriptional profiling of granulomas at 4 wk shows a pre-
events during infection that dictate outcome and allow one dominantly proinflammatory profile that is later reduced by
to control the infection dose and timing, to perform inten- 12 wk postinfection (40). Transcriptional signatures in blood
sive monitoring and serial sampling, and to harvest tissue from macaques after M. tuberculosis infection appear to be
samples, including granulomas, lung, and lymph nodes, at similar to human signatures (41–43). Interestingly, some of
various timepoints postinfection. Although murine models the signatures (e.g., IFN-related signatures) that predicted
are generally used to study pathogenesis and immunology of active TB during the period of clinical stability in hu-
TB, mice do not replicate many key aspects of human mans (so-called incipient TB) (29) were observed prior to
M. tuberculosis infection, including granuloma formation and M. tuberculosis infection in our macaque model (41). In a
LTBI. For these reasons, the available data on LBTI and in- rhesus study of a CMV vector-based vaccine, Hansen et al.
dividual granulomas are primarily from nonhuman primate (42) observed upregulation of a transcriptional module
models. Macaques are unique models for TB in that they related to neutrophil and innate immune function associ-
recapitulate the entire human range of infection outcomes and ated with vaccine-induced protection that was present be-
exhibit pathology, including granulomas, that is identical to fore infection. These data reinforce the notion that early
human M. tuberculosis infection. In particular, cynomolgus immune factors are critical determinants of outcome.
macaques infected with a low dose of virulent M. tuberculosis
develop the full spectrum of M. tuberculosis infections seen in The success and failure of the granuloma
humans, from LTBI to severe disease, over the course of The histopathologic hallmark of TB is the granuloma, an
several months (32, 33). Rhesus macaques are more suscep- organized spherical structure composed of a variety of cell
tible, with most developing active TB within 4 mo; however, types, primarily macrophages and lymphocytes. The classic
even in rhesus macaques, there is a spectrum of active TB granuloma has a necrotic (caseous) core, surrounded by epi-
from mild to severe (34). Thus, these macaque models pro- thelioid macrophages, with an outer cuff of lymphocytes and
vide unique opportunities to investigate M. tuberculosis host– macrophages. However, there is a range of granuloma types,
pathogen interactions with remarkable similarities to humans. including nonnecrotic, suppurative (neutrophilic), fibrotic,
and mineralized. These diverse granuloma types were recog-
Early events in M. tuberculosis infection nized early in studies on human TB, and we observe the full
In humans, it is extremely difficult to study early events of range of granuloma types in macaques (44). The granuloma
infection, because the time of infection is rarely known and the contains and prevents dissemination of the infection, and it
diagnostics rely on development of an adaptive immune re- provides an immune microenvironment where macrophages
sponse, which usually takes 6–8 wk to be detected by IGRA or are activated to kill or restrain the growth of M. tuberculosis
TST. However, the available data indicate that early immune bacilli. However, the bacterium has evolved to persist in some
control of M. tuberculosis infection plays a critical role in granulomas, often for years or even the lifetime of the host.
outcome (35). As an intracellular respiratory pathogen, the The actual metabolic state of the bacilli in granulomas over
bacilli first encounter the immune response in the airways, the long term is not known, although there likely is replica-
which includes cellular (e.g., alveolar macrophages, T cells, tion occurring, even if quite slowly or intermittently. What is
innate lymphocytes) and noncellular (e.g., antimicrobial becoming clear is that there is substantial heterogeneity of
peptides) (36) components. Alveolar macrophages are capable granulomas, even within the same host, in terms of host
of killing M. tuberculosis soon postinfection, but these cells are response, bacterial growth and killing, dissemination, and
quite heterogeneous in both humans and macaques, with reactivation. This is exemplified by serial PET/CT imaging of
likely variable bactericidal capacity. Specific subsets of alveolar individual granulomas in which individual granulomas appear
macrophages observed before M. tuberculosis infection were dynamic and independent, increasing or decreasing in size
associated with differences in infection outcome (active ver- (measured by CT) and metabolic activity (as measured by
sus LTBI) (37). Infected macrophages can induce bronchial FDG avidity) within the same macaque and even within the
The Journal of Immunology 2545

same lung lobe (38, 45) (Fig. 3). Even in humans, reso-
lution of some granulomas and progression of others dur-
ing active TB can be observed by PET/CT imaging over as
little as 2 mo (46).
We have shown that individual bacilli that enter the lung
give rise to individual granulomas, but these granulomas can
have different trajectories (39). The heterogeneity among
individual granulomas within the same host can be observed
in terms of bacterial burden and killing, immune responses,
risk of dissemination, and granuloma type (39, 44, 47, 48).
Using a library of “bar-coded” M. tuberculosis, in which in-
dividual bacteria can be distinguished by a DNA tag, in
conjunction with serial PET/CT imaging, we demonstrated
that only a subset of granulomas within a single host dis-
seminate to form new granulomas and advance disease (48).
Thus, despite the global host immune response to the path-
ogen, it is the local host–pathogen response in each granu-
loma that determines overall control of the infection. During

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

the first 4 wk of M. tuberculosis infection when adaptive im-
munity is just being initiated, bacterial burden in granulomas
is at its greatest with little to no bacterial killing (39). During FIGURE 3. Granulomas are independent and dynamic by PET/CT during
the course of infection, granulomas from animals with active M. tuberculosis infection in macaques. Top row; between 6 and 16 wk
postinfection, one granuloma increases in size but maintains the same FDG
or latent disease can sterilize the infection, although the
avidity (yellow arrow), whereas the other granuloma increases in size and
proportion of sterile granulomas is greater in those with LTBI FDG avidity (orange arrow) in the accessory lobe. Bottom row; this same
(39). Thus, unlike most infections, the clinical outcome of animal has a granuloma in the right lower lobe that has decreased in size and
infection is more accurately represented by the cumulative FDG avidity (blue arrow) at the same timepoints. Black line = 1 cm.
host–pathogen interactions of all sites of involvement. That is,
all granulomas within a host need to either kill or control
M. tuberculosis bacilli to prevent development of disease; one infection resulted in 80% of animals with worsening TB pa-
poorly functioning granuloma can lead to loss of containment thology, greater bacterial burden, and dissemination compared
and TB progression. with controls (50). In contrast, CD4+ T cell depletion during
The success or failure of each granuloma is dependent on the LTBI resulted in only 50% of animals reactivating (50). Similar
immune response within that granuloma. The granuloma findings have been observed with TNF neutralization, which
functions to immunologically contain M. tuberculosis. Yet, consistently exacerbates acute infection, but only results in
granulomas are heterogeneous in terms of cell types present reactivation in ∼50% of LTBI macaques (4, 51). The greater
(T cells, B cells, macrophages, neutrophils), cellular immune susceptibility to disseminated disease during early infection is
responses, and bacterial burden (47), and this heterogeneity likely due to higher bacterial loads that require more of a robust
likely contributes to the variable range of infection outcomes immune response to control within any individual granuloma
(Fig. 4). Surprisingly, ,10% of T cells within the granuloma and the very low frequency of sterile granulomas at early
produce Th1 or Th17 cytokines (47). Our recent data suggest timepoints. For example, depletion of CD20+ B cells during
that this is not due to exhaustion of T cells (49), as might be early infection did not result in overt clinical deterioration but
expected in a chronic infection. Interestingly, the combina- did significantly increase variability of bacterial burden and
tion of IL-10–, TNF-, or IL-17–producing T cells, rather influenced T cell responses within individual granulomas (52).
than a preponderance of Th1 cytokines, within individual Thus, B cells may be important in stabilizing the granuloma
granulomas was best associated with low bacterial burden and immune responses early in infection.
sterile granulomas (47). Thus, the “success” (bacterial killing Outcomes in individual granulomas are likely associated with the TB
or prevention of dissemination) of an individual granuloma spectrum of risk. Given the heterogeneity of lesion types
appears to be the result of both pro- and anti-inflammatory seen among the various outcomes of M. tuberculosis infec-
processes rather than being depending on any single cytokine, tion, it is not surprising that the PET/CT patterns during
chemokine, or cell type. Although the host–pathogen inter- treatment would be extremely variable. Malherbe et al. (53)
action within an individual granuloma is relatively complex, characterized PET/CT patterns of active TB cases treated with
even greater complexity is observed at the organismal level the standard four-drug regimen and correlated the findings
because of the number of heterogeneous granulomas present with clinical and microbiological outcomes. PET/CTs were
within a single host. done at baseline, 1 and 6 mo after the start of treatment.
This concept is important when examining the impact of Six months after treatment, PET/CT patterns were classified
systemic immunologic interventions on individual granulomas as resolved (i.e., minimal to no increase in FDG from
and on the overall outcome of the host. For example, the pretreatment baseline, regardless of CT abnormalities),
granulomas during early infection are more homogeneous with improved (i.e., decreased FDG intensity of all lesions but
high bacterial burden and later become much more hetero- with $1 lesion with increased FDG uptake), and mixed
geneous and likely less susceptible to immune suppression. For (i.e., $1 lesion with higher FDG intensity compared with
example, depletion of CD4+ T cells during early M. tuberculosis baseline). Heterogeneous PET/CT scans were observed
2546 BRIEF REVIEWS: SPECTRUM OF M. TUBERCULOSIS INFECTION

FIGURE 4. The hypothetical spec-

trum of granuloma types is seen in any
infected individual with each gran-
uloma associated with a range of
immune control and bacterial burden.
A granuloma that has lost its overall
structure and function (progressive and
disseminating granulomas) is associated
with poor immune control and high
bacterial burden. In contrast, fibrotic
granulomas are often sterile reflecting a
robust immune response.

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

during treatment. Patients who failed treatment had a single granuloma) and the presence of an extrapulmonary
exclusively mixed patterns; however, mixed patterns could lesion (spleen or liver) predicted reactivation with 92% sen-
also be seen among cured and recurrent patients. Moreover, sitivity and specificity. To determine the importance of these
M. tuberculosis mRNA was identified in sputum 6 mo after metrics, we predicted reactivation risk in a large number of
initiation of drug treatment in cured, failed, and recurrent LTBI macaques who did not undergo TNF neutralization and
TB cohorts. The significance of this is yet unclear. The characterized their granulomas by PET/CT. The macaques
spectrum of patterns seen during treatment likely reflects predicted to be at high risk using our PET/CT metrics had a
the heterogeneous effect of drug treatment on individual single granuloma with higher bacterial burden than those
granulomas due to variable drug penetration into various predicted to be at low risk, supporting the concept that a
granuloma types (54, 55) and the efficacy of the individual single granuloma can put a host at higher risk of reactivation
drugs against bacilli in different metabolic states depending and that successful immune responses in all granulomas are
on the microenvironment of each granuloma (56) [reviewed essential for long-term prevention of disease. However, the
in (57)]. Clearly, more comprehensive studies are required to immunologic microenvironment within granulomas is com-
better understand outcome and the appropriate treatment of plex and each granuloma likely takes its own path to success
disease. or failure, making it unlikely that a single set of immune
Reactivation risk in LTBI may well depend on a single responses uniformly leads to success (35). This paradigm has
granuloma that is slightly impaired for bacterial control. In major implications for vaccines and host-directed therapies,
LTBI, it traditionally was assumed that all granulomas are as protective responses likely involve a variety of immune
successfully controlling or eliminating the infection. However, mechanisms from different cell types.
the new paradigm states that LTBI exists as a spectrum of
infection outcomes. What defines the spectrum immunolog- Conclusions
ically is not clear, but it is unlikely that this will be fully Given the heterogeneous nature of bacterial dynamics and host
understood simply from blood signatures. Data from our immune responses in individual granulomas, it is no wonder
macaque models indicate that immune responses in blood do that outcomes cannot be classified by binary outcomes of active
not reliably reflect immune responses in either individual or LTBI. Moreover, these factors significantly complicate ef-
granulomas or the average of all granulomas (47). We per- forts to improve TB treatment and develop protective vaccines.
formed a large study of reactivation risk in LTBI cynomolgus Perhaps it will be more accurate to classify infection outcomes
macaques, in which TNF was neutralized (a major risk factor as an individual’s risk of primary disease or reactivation. This
for reactivation TB in humans). PET/CT scans were per- requires more precise diagnostics with greater specificity and
formed on macaques with LTBI over the 8-wk course of TNF sensitivity, especially for immune-compromised hosts, such as
neutralization (4). Fifty percent of the macaques experienced those with HIV who have the greatest risk of disease. It is
reactivated TB during this time. We assessed PET/CT fea- possible that risk assessments may require a series of tests
tures prior to TNF neutralization that were associated with and rely on a battery of immune responses that could be
reactivation and found that the number of granulomas in the difficult to implement in the field instead of simply relying on
lung was not a factor. Instead, lung inflammation (often from TST or IGRA results. Non–culture-based assays to identify
The Journal of Immunology 2547

M. tuberculosis components, such as the urinary lipo- 11. Stein, C. M., S. Zalwango, L. L. Malone, B. Thiel, E. Mupere, M. Nsereko,
B. Okware, H. Kisingo, C. L. Lancioni, C. M. Bark, et al. 2018. Resistance and
arabinomannan assay (58) or trehalose assay (59), show susceptibility to Mycobacterium tuberculosis infection and disease in tuberculosis
promise as a point-of-care diagnostic. Host-specific risks, households in Kampala, Uganda. Am. J. Epidemiol. 187: 1477–1489.
12. Simmons, J. D., C. M. Stein, C. Seshadri, M. Campo, G. Alter, S. Fortune,
such as immune competence, age, genetic background, and E. Schurr, R. S. Wallis, G. Churchyard, H. Mayanja-Kizza, et al. 2018. Immu-
comorbidities, will need to be accounted for, not just once, nological mechanisms of human resistance to persistent Mycobacterium tuberculosis
but over time as the risk can change over the life of the host. infection. Nat. Rev. Immunol. 18: 575–589.
13. Lu, L. L., A. W. Chung, T. R. Rosebrock, M. Ghebremichael, W. H. Yu,
That said, even with improved diagnostic methods and a P. S. Grace, M. K. Schoen, F. Tafesse, C. Martin, V. Leung, et al. 2016. A func-
more thorough understanding of the disease and risk, we need tional role for antibodies in tuberculosis. Cell 167: 433–443.e14.
14. Seshadri, C., N. Sedaghat, M. Campo, G. Peterson, R. D. Wells, G. S. Olson,
to ask ourselves some key questions. Given the complex D. R. Sherman, C. M. Stein, H. Mayanja-Kizza, A. Shojaie, et al; Tuberculosis
spectrum of M. tuberculosis infection, what will the optimal Research Unit (TBRU). 2017. Transcriptional networks are associated with resis-
tance to Mycobacterium tuberculosis infection. PLoS One 12: e0175844.
treatment regimen look like? For example, what is the optimal 15. Ewer, K., K. A. Millington, J. J. Deeks, L. Alvarez, G. Bryant, and A. Lalvani. 2006.
treatment for incipient disease? Is a four-drug therapy neces- Dynamic antigen-specific T-cell responses after point-source exposure to Mycobac-
terium tuberculosis. Am. J. Respir. Crit. Care Med. 174: 831–839.
sary for a stage that may have a lower bacterial burden 16. Pai, M., R. Joshi, S. Dogra, A. A. Zwerling, D. Gajalakshmi, K. Goswami,
than symptomatically overt, active TB? Will individualized M. V. Reddy, A. Kalantri, P. C. Hill, D. Menzies, and P. C. Hopewell. 2009. T-cell
host-directed therapies, in conjunction with antimicrobial assay conversions and reversions among household contacts of tuberculosis patients
in rural India. Int. J. Tuberc. Lung Dis. 13: 84–92.
treatment, shorten treatment duration? Because the disease 17. Andrews, J. R., M. Hatherill, H. Mahomed, W. A. Hanekom, M. Campo,
progression can change over time and the time of infection is T. R. Hawn, R. Wood, and T. J. Scriba. 2015. The dynamics of QuantiFERON-
TB gold in-tube conversion and reversion in a cohort of South African adolescents.
unknown, do patients require repeated testing for incipient Am. J. Respir. Crit. Care Med. 191: 584–591.

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

and overtly active TB? Will there be an accurate and cost- 18. Nemes, E., H. Geldenhuys, V. Rozot, K. T. Rutkowski, F. Ratangee, N. Bilek,
S. Mabwe, L. Makhethe, M. Erasmus, A. Toefy, et al; C-040-404 Study Team.
effective method for identifying those at low-to-no risk of 2018. Prevention of M. tuberculosis infection with H4:IC31 vaccine or BCG
reactivation? Does that risk change over time? Can tests of revaccination. N. Engl. J. Med. 379: 138–149.
incipient TB that all currently rely on transcriptional signa- 19. Berry, M. P., C. M. Graham, F. W. McNab, Z. Xu, S. A. Bloch, T. Oni,
K. A. Wilkinson, R. Banchereau, J. Skinner, R. J. Wilkinson, et al. 2010. An
tures be converted to point-of-care assays that are practical in interferon-inducible neutrophil-driven blood transcriptional signature in human
resource-limited settings? Clearly more research is required to tuberculosis. Nature 466: 973–977.
20. Opie, E. L. 1917. The relation of apical tuberculosis of adults to the focal tuber-
better understand the pathogenesis and immune responses culosis of children. J. Exp. Med. 26: 263–277.
that categorize status of infection and risk of disease on both a 21. Burman, W. J., and B. E. Jones. 2003. Clinical and radiographic features of HIV-
related tuberculosis. Semin. Respir. Infect. 18: 263–271.
microscopic (e.g., within the granuloma) and macroscopic 22. Greenberg, S. D., D. Frager, B. Suster, S. Walker, C. Stavropoulos, and
(clinical outcomes) level to drive better diagnostic and ther- A. Rothpearl. 1994. Active pulmonary tuberculosis in patients with AIDS: spectrum
of radiographic findings (including a normal appearance). Radiology 193: 115–119.
apeutic modalities for the future. 23. Keane, J., S. Gershon, R. P. Wise, E. Mirabile-Levens, J. Kasznica,
W. D. Schwieterman, J. N. Siegel, and M. M. Braun. 2001. Tuberculosis associated
Acknowledgments with infliximab, a tumor necrosis factor alpha-neutralizing agent. N. Engl. J. Med.
345: 1098–1104.
We thank all the members of the Flynn and Lin laboratories and our colleagues 24. Goo, J. M., J. G. Im, K. H. Do, J. S. Yeo, J. B. Seo, H. Y. Kim, and J. K. Chung.
Drs. Sarah Fortune, Denise Kirschner, Charles Scanga, Edwin Klein, Hannah 2000. Pulmonary tuberculoma evaluated by means of FDG PET: findings in 10
Gideon, and Joshua Mattila for helpful discussions that contributed to this cases. Radiology 216: 117–121.
25. Ankrah, A. O., A. W. J. M. Glaudemans, A. Maes, C. Van de Wiele,
review. Special thanks to Elise Y. Chu and Pauline Maiello for artistic and R. A. J. O. Dierckx, M. Vorster, and M. M. Sathekge. 2018. Tuberculosis. Semin.
graphic contributions. Nucl. Med. 48: 108–130.
26. Esmail, H., R. P. Lai, M. Lesosky, K. A. Wilkinson, C. M. Graham,
A. K. Coussens, T. Oni, J. M. Warwick, Q. Said-Hartley, C. F. Koegelenberg, et al.
Disclosures 2016. Characterization of progressive HIV-associated tuberculosis using 2-deoxy-2-
[18F]fluoro-D-glucose positron emission and computed tomography. Nat. Med. 22:
The authors have no financial conflicts of interest.
1090–1093.
27. Sweeney, T. E., L. Braviak, C. M. Tato, and P. Khatri. 2016. Genome-wide ex-
pression for diagnosis of pulmonary tuberculosis: a multicohort analysis. Lancet
References Respir. Med. 4: 213–224.
1. World Health Organization. 2017. Global Tuberculosis Report 2017. World Health 28. Burel, J. G., C. S. Lindestam Arlehamn, N. Khan, G. Seumois, J. A. Greenbaum,
Organization, Geneva, Switzerland. R. Taplitz, R. H. Gilman, M. Saito, P. Vijayanand, A. Sette, and B. Peters. 2018.
2. World Health Organization. 2018. Tuberculosis (Fact Sheet). World Health Orga- Transcriptomic analysis of CD4+ T cells reveals novel immune signatures of latent
nization, Geneva, Switzerland. tuberculosis. J. Immunol. 200: 3283–3290.
3. Zumla, A., M. Raviglione, R. Hafner, and C. F. von Reyn. 2013. Tuberculosis. 29. Zak, D. E., A. Penn-Nicholson, T. J. Scriba, E. Thompson, S. Suliman,
N. Engl. J. Med. 368: 745–755. L. M. Amon, H. Mahomed, M. Erasmus, W. Whatney, G. D. Hussey, et al; ACS
4. Lin, P. L., P. Maiello, H. P. Gideon, M. T. Coleman, A. M. Cadena, and GC6-74 cohort study groups. 2016. A blood RNA signature for tuberculosis
M. A. Rodgers, R. Gregg, M. O’Malley, J. Tomko, D. Fillmore, et al. 2016. PET disease risk: a prospective cohort study. Lancet 387: 2312–2322.
CT identifies reactivation risk in cynomolgus macaques with latent M. tuberculosis. 30. Ritter, J. 1916. So-called “incipient” tuberculosis. J. Am. Med. Assoc. LXVI:
PLoS Pathog. 12: e1005739. 592.
5. Barry III, C. E., H. I. Boshoff, V. Dartois, T. Dick, S. Ehrt, J. Flynn, D. Schnappinger, 31. World Health Organization. 2017. Development of a Target Product Profile (TPP)
R. J. Wilkinson, and D. Young. 2009. The spectrum of latent tuberculosis: rethinking the and A Framework for Evaluation for a Test for Predicting Progression from Tuberculosis
biology and intervention strategies. Nat. Rev. Microbiol. 7: 845–855. Infection to Active Disease. World Health Organization, Geneva, Switzerland.
6. Gideon, H. P., and J. L. Flynn. 2011. Latent tuberculosis: what the host “sees”? 32. Capuano, S. V. , III, D. A. Croix, S. Pawar, A. Zinovik, A. Myers, P. L. Lin,
Immunol. Res. 50: 202–212. S. Bissel, C. Fuhrman, E. Klein, and J. L. Flynn. 2003. Experimental Mycobacterium
7. Lin, P. L., and J. L. Flynn. 2010. Understanding latent tuberculosis: a moving tuberculosis infection of cynomolgus macaques closely resembles the various mani-
target. J. Immunol. 185: 15–22. festations of human M. tuberculosis infection. Infect. Immun. 71: 5831–5844.
8. Houk, V. N. 1980. Spread of tuberculosis via recirculated air in a naval vessel: the 33. Lin, P. L., M. Rodgers, L. Smith, M. Bigbee, A. Myers, C. Bigbee, I. Chiosea,
Byrd study. Ann. N. Y. Acad. Sci. 353: 10–24. S. V. Capuano, C. Fuhrman, E. Klein, and J. L. Flynn. 2009. Quantitative com-
9. Morrison, J., M. Pai, and P. C. Hopewell. 2008. Tuberculosis and latent tuberculosis parison of active and latent tuberculosis in the cynomolgus macaque model. Infect.
infection in close contacts of people with pulmonary tuberculosis in low-income and Immun. 77: 4631–4642.
middle-income countries: a systematic review and meta-analysis. Lancet Infect. Dis. 8: 34. Maiello, P., R. M. DiFazio, A. M. Cadena, M. A. Rodgers, P. L. Lin, C. A. Scanga,
359–368. and J. L. Flynn. 2018. Rhesus macaques are more susceptible to progressive tu-
10. Ma, N., S. Zalwango, L. L. Malone, M. Nsereko, E. M. Wampande, B. A. Thiel, berculosis than cynomolgus macaques: a quantitative comparison. Infect. Immun.
B. Okware, R. P. Igo, Jr., M. L. Joloba, E. Mupere, et al; Tuberculosis Research 86: e00505–17.
Unit (TBRU). 2014. Clinical and epidemiological characteristics of individuals 35. Cadena, A. M., J. L. Flynn, and S. M. Fortune. 2016. The importance of first
resistant to M. tuberculosis infection in a longitudinal TB household contact study in impressions: early events in Mycobacterium tuberculosis infection influence outcome.
Kampala, Uganda. BMC Infect. Dis. 14: 352. MBio 7: e00342–e16.
2548 BRIEF REVIEWS: SPECTRUM OF M. TUBERCULOSIS INFECTION

36. Reuschl, A. K., M. R. Edwards, R. Parker, D. W. Connell, L. Hoang, A. Halliday, balance of pro- and anti-inflammatory cytokines is associated with steriliza-
H. Jarvis, N. Siddiqui, C. Wright, S. Bremang, et al. 2017. Innate activation of tion. PLoS Pathog. 11: e1004603.
human primary epithelial cells broadens the host response to Mycobacterium tu- 48. Martin, C. J., A. M. Cadena, V. W. Leung, P. L. Lin, P. Maiello, N. Hicks,
berculosis in the airways. PLoS Pathog. 13: e1006577. M. R. Chase, J. L. Flynn, and S. M. Fortune. 2017. Digitally barcoding Myco-
37. Silver, R. F., A. J. Myers, J. Jarvela, J. Flynn, T. Rutledge, T. Bonfield, and bacterium tuberculosis reveals in vivo infection dynamics in the macaque model of
P. L. Lin. 2016. Diversity of human and macaque airway immune cells at tuberculosis. MBio 8: e00312–17.
baseline and during tuberculosis infection. Am. J. Respir. Cell Mol. Biol. 55: 49. Wong, E. A., L. Joslyn, N. L. Grant, E. Klein, P. L. Lin, D. E. Kirschner, and
899–908. J. L. Flynn. 2018. Low levels of T cell exhaustion in tuberculous lung granulomas.
38. Coleman, M. T., P. Maiello, J. Tomko, L. J. Frye, D. Fillmore, C. Janssen, Infect. Immun. 86: e00426–18.
E. Klein, and P. L. Lin. 2014. Early Changes by (18)Fluorodeoxyglucose positron 50. Lin, P. L., T. Rutledge, A. M. Green, M. Bigbee, C. Fuhrman, E. Klein, and
emission tomography coregistered with computed tomography predict outcome J. L. Flynn. 2012. CD4 T cell depletion exacerbates acute Mycobacterium tubercu-
after Mycobacterium tuberculosis infection in cynomolgus macaques. Infect. Immun. losis while reactivation of latent infection is dependent on severity of tissue depletion
82: 2400–2404. in cynomolgus macaques. AIDS Res. Hum. Retroviruses 28: 1693–1702.
39. Lin, P. L., C. B. Ford, M. T. Coleman, A. J. Myers, R. Gawande, T. Ioerger, 51. Lin, P. L., A. Myers, L. Smith, C. Bigbee, M. Bigbee, C. Fuhrman, H. Grieser,
J. Sacchettini, S. M. Fortune, and J. L. Flynn. 2014. Sterilization of granulomas is I. Chiosea, N. N. Voitenek, S. V. Capuano, et al. 2010. Tumor necrosis factor
common in active and latent tuberculosis despite within-host variability in bacterial neutralization results in disseminated disease in acute and latent Mycobacterium
killing. Nat. Med. 20: 75–79. tuberculosis infection with normal granuloma structure in a cynomolgus macaque
40. Mehra, S., B. Pahar, N. K. Dutta, C. N. Conerly, K. Philippi-Falkenstein, model. Arthritis Rheum. 62: 340–350.
X. Alvarez, and D. Kaushal. 2010. Transcriptional reprogramming in nonhuman 52. Phuah, J., E. A. Wong, H. P. Gideon, P. Maiello, M. T. Coleman,
primate (rhesus macaque) tuberculosis granulomas. PLoS One 5: e12266. M. R. Hendricks, R. Ruden, L. R. Cirrincione, J. Chan, P. L. Lin, and J. L. Flynn.
41. Gideon, H. P., J. A. Skinner, N. Baldwin, J. L. Flynn, and P. L. Lin. 2016. Early 2016. Effects of B cell depletion on early Mycobacterium tuberculosis infection in
whole blood transcriptional signatures are associated with severity of lung inflam- cynomolgus macaques. Infect. Immun. 84: 1301–1311.
mation in cynomolgus macaques with Mycobacterium tuberculosis infection. 53. Malherbe, S. T., S. Shenai, K. Ronacher, A. G. Loxton, G. Dolganov, M. Kriel,
J. Immunol. 197: 4817–4828. T. Van, R. Y. Chen, J. Warwick, L. E. Via, et al; Catalysis TB–Biomarker Con-
42. Hansen, S. G., D. E. Zak, G. Xu, J. C. Ford, E. E. Marshall, D. Malouli, sortium. 2016. Persisting positron emission tomography lesion activity and Myco-
R. M. Gilbride, C. M. Hughes, A. B. Ventura, E. Ainslie, et al. 2018. Prevention of bacterium tuberculosis mRNA after tuberculosis cure. [Published erratum appears in

Downloaded from http://www.jimmunol.org/ by guest on October 22, 2018

tuberculosis in rhesus macaques by a cytomegalovirus-based vaccine. Nat. Med. 24: 2017 Nat. Med. 23: 1499; 23: 526.] Nat. Med. 22: 1094–1100.
130–143. 54. Zimmerman, M., L. Blanc, P. Y. Chen, V. Dartois, and B. Prideaux. 2018. Spatial
43. Thompson, E. G., S. Shankar, H. P. Gideon, J. Braun, J. Valvo, J. A. Skinner, quantification of drugs in pulmonary tuberculosis lesions by laser capture micro-
A. Aderem, J. L. Flynn, P. L. Lin, and D. E. Zak. 2018. Prospective discrimination dissection liquid chromatography mass spectrometry (LCM-LC/MS). J. Vis. Exp.
of controllers from progressors early after low-dose Mycobacterium tuberculosis in- (134): e57402.
fection of cynomolgus macaques using blood RNA signatures. J. Infect. Dis. 217: 55. Prideaux, B., L. E. Via, M. D. Zimmerman, S. Eum, J. Sarathy, P. O’Brien,
1318–1322. C. Chen, F. Kaya, D. M. Weiner, P. Y. Chen, et al. 2015. The association between
44. Flynn, J. L., and E. Klein. 2010. Pulmonary tuberculosis in monkeys. In A Color sterilizing activity and drug distribution into tuberculosis lesions. Nat. Med. 21:
Atlas of Comparative Pathology of Pulmonary Tuberculosis. CRC Press, Boca Raton, 1223–1227.
FL, p. 83–105. 56. Sarathy, J. P., L. E. Via, D. Weiner, L. Blanc, H. Boshoff, E. A. Eugenin,
45. Lin, P. L., T. Coleman, J. P. Carney, B. J. Lopresti, J. Tomko, D. Fillmore, C. E. Barry, III, and V. A. Dartois. 2018. Extreme drug tolerance of Mycobacterium
V. Dartois, C. Scanga, L. J. Frye, C. Janssen, et al. 2013. Radiologic responses in tuberculosis in Caseum. Antimicrob. Agents Chemother. 62: e02266–17.
cynomolgus macaques for assessing tuberculosis chemotherapy regimens. Anti- 57. Lenaerts, A., C. E. Barry, III, and V. Dartois. 2015. Heterogeneity in tuberculosis
microb. Agents Chemother. 57: 4237–4244. pathology, microenvironments and therapeutic responses. Immunol. Rev. 264: 288–
46. Coleman, M. T., R. Y. Chen, M. Lee, P. L. Lin, L. E. Dodd, P. Maiello, L. E. Via, 307.
Y. Kim, G. Marriner, V. Dartois, et al. 2014. PET/CT imaging reveals a therapeutic 58. Kerkhoff, A. D., and S. D. Lawn. 2016. A breakthrough urine-based diagnostic test
response to oxazolidinones in macaques and humans with tuberculosis. Sci. Transl. for HIV-associated tuberculosis. Lancet 387: 1139–1141.
Med. 6: 265ra167. 59. Kamariza, M., P. Shieh, C. S. Ealand, J. S. Peters, B. Chu, F. P. Rodriguez-Rivera,
47. Gideon, H. P., J. Phuah, A. J. Myers, B. D. Bryson, M. A. Rodgers, M. R. Babu Sait, W. V. Treuren, N. Martinson, R. Kalscheuer, et al. 2018. Rapid
M. T. Coleman, P. Maiello, T. Rutledge, S. Marino, S. M. Fortune, et al. detection of Mycobacterium tuberculosis in sputum with a solvatochromic trehalose
2015. Variability in tuberculosis granuloma T cell responses exists, but a probe. Sci. Transl. Med. 10: eaam6310.