Monkey Models of Tuberculosis: Lessons Learned

Abstract

The use of animal models has been invaluable for studying the pathogenesis of Mycobacterium tuberculosis infection, as well as for testing the efficacy of vaccines and drug regimens for tuberculosis. Among the applied animal models, nonhuman primates, particularly macaques, share the greatest anatomical and physiological similarities with humans. As such, macaque models have been used for investigating tuberculosis pathogenesis and preclinical testing of drugs and vaccines. This review focuses on published major studies which illustrate how the rhesus and cynomolgus macaques have enriched and may continue to advance the field of global tuberculosis research.

INTRODUCTION

About one-third of the population worldwide is infected with Mycobacterium tuberculosis, while nearly 9 million cases of active tuberculosis (TB) are reported annually (1). Hence, the need for improved preventative and treatment strategies against TB is ever-increasing. In an attempt to globally control TB, researchers have employed multiple animal models (mice, guinea pigs, rabbits, cows, nonhuman primates, and others) for testing novel experimental vaccines and therapies for TB (2, 3). The use of nonhuman primates (NHP), especially cynomolgus macaques (CM; Macaca fascicularis, also called long-tailed macaques, a species of Old World monkeys native to Southeast Asia) and rhesus macaques (RM; Macaca mulatta, a species of Old World monkeys native to Asia; most experimental models are from India or China), has led to significant advances in TB research due to their inherent commonalities with humans, as illustrated in previous reviews in this subject area (4–6). By using the macaque model of TB, we can gain even greater insights into ways to prevent M. tuberculosis infection and disease progression.

JUSTIFICATION FOR MACAQUES IN TB RESEARCH

Macaques exhibit remarkable similarities to humans in virtually every aspect of their anatomy and physiology (7–9). As such, macaques respond similarly to many human immunological, pathological, and drug agents, providing a tremendous advantage over other animal models (6, 10). The literature shows that macaques and humans share extensive clinical manifestations of TB, including pulmonary and extrapulmonary signs and symptoms (6, 10). Clinicians and researchers can monitor the disease course in macaques by measuring nearly identical parameters tested in humans, ranging from skin and blood tests to radiographic imaging and body fluid samples (Table 1). In addition, multidrug chemotherapy for TB provides effective treatment in both humans and macaques (11, 12). Furthermore, as in humans, Mycobacterium bovis bacillus Calmette-Guérin (BCG) vaccination exhibits variable efficacy in macaques of even the same species (13–15). Table 1 highlights the similarities and differences in M. tuberculosis infection between humans and the rhesus macaques and cynomolgus macaques.

TABLE 1.

Comparison of TB in humans versus rhesus and cynomolgus macaques^a

Parameter	Humans	Rhesus macaques	Cynomolgus macaques
Clinical manifestations
Rate of disease progressionb	Acute ≪ latent (10% vs 90%)	Acute ≫ latent (90% vs 10%)	Acute > latent (60% vs 40%)
Presence of active/chronic infection symptom
Cough	+	+	+
Bloody sputum	+	+	+
Increased body temp	+	+	+
Wt loss	+	+	+
Presence of latent infection symptoms
No clinical signs	+	+	+
Activated by coinfection (HIV or SIV)	+	+	+
Clinical tests
Skin tests
PPD	+	+	+
Old tuberculin test	+	+	+
Blood tests
ELISA, ELISpot	+	+	+
Quantiferon-TB Gold	+	+	+
Primagram	−	+	+
CBC, ESR, CRP, LT	+	+	+
Imaging (chest X-Ray, MRI, PET/CT)	+	+	+
Fluid sampling (BAL, gastric aspirate)	+	+	+
Pathology
Caseous granulomas	+	+	+
Calcification	+/−	+/−	+/−
Fibrous capsule	+/−	+/−	+/−
Pulmonary cavities	+	+	+
Disseminated lesions	+/−	+/−	+/−

^aThe majority of macaques, particularly the rhesus species, develop acute or active TB after artificial infection, whereas 90% of infected humans have latent TB. Chronic infection is defined as persistent signs of active disease, radiographic involvement, or culture positivity. Although the PPD and old tuberculin skin tests are used in both humans and macaques, these diagnostic exams are less reliable in macaques than in humans (69). Also, macaques exhibit a more random than apical distribution of pulmonary cavities. Abbreviations: BAL, bronchoalveolar lavage; CBC, complete blood count; CRP, C-reactive protein; LT, lymphocyte transformation; rBCG, recombinant bacillus Calmette-Guérin, +, positive for the indicated finding or functional modality; −, absent finding or unused modality; +/−, a variable finding.

^bAcute disease lasts weeks to months, and latent disease lasts months to years. Percentages (in parentheses) indicate the global percentage of the species infected with M. tuberculosis.

HISTORICAL OUTLOOK ON MACAQUE MODELS

The use of NHP models to study M. tuberculosis infection traces back to published literature from the 1960s. This “Golden Age” of TB research using NHP, performed through the 1970s, generated valuable data on the evolving BCG vaccine (16), as well as one report on the TB drug efficacies of ethambutol and isoniazid (12). The majority of the studies focused on the BCG-induced immune reactions and vaccine efficacy (17–23). All pertinent studies published during this era used Indian RM models, which underwent intrabronchial (i.b.), intratracheal (i.t.), or aerosol infection with M. tuberculosis. The estimated mycobacterial retention rate in mammalian lungs after aerosol exposure was derived from prior animal models, including macaques exposed to anthrax spores, as well as guinea pigs and mice infected with M. tuberculosis (24). Figure 1 recaps the major published articles of NHP models with experimental M. tuberculosis infection during this time period (12, 17–23).

FIG 1.

The “Golden Age” of TB research using rhesus macaques. The timeline illustrates major studies of experimental M. tuberculosis infection in Indian rhesus macaques during the so-called Golden Age from the 1960s to 1970s (12, 17–23). Events are organized chronologically by year of publication. Thus, the 10-year study by Good (17) actually began before the first reported investigation in 1966. Locations of the experiments (red) and the corresponding author (blue) are indicated. Abbreviations: ic, intracutaneous; im, intramuscular; iv, intravenous; sc, subcutaneous; M.tb, M. tuberculosis; NBL, Naval Biological Laboratories; UC, University of California.

Not until about 20 years after the Golden Age was another study using experimentally infected macaques with M. tuberculosis published (25). This large research gap may be attributed to several factors, including the high maintenance costs and necessary space/equipment for biocontainment to properly conduct such experiments (6). Additionally, the limited animal availability, handling difficulties, and adverse public opinions discouraged TB research with NHP. However, with greater research funding and the emergence of compatible reagents for macaques, further investigations into M. tuberculosis infection using NHP models regained momentum (10). Especially amid the expansion of the National Primate Research Centers (NPRCs), TB investigators revisited the macaque model of experimental M. tuberculosis inoculation with renewed enthusiasm. A series of subsequent investigations were dedicated to macaque research to assess novel TB vaccines and drugs, as well as to gain an understanding of the pathogenesis of M. tuberculosis infection and reactivation. Figures 2 and 3 abridge the RM (13, 26–43) and CM (10, 15, 25, 42–59) models of experimental M. tuberculosis infection after the Golden Age, from 2001 to 2014.

FIG 2.

Twenty-first century TB research using rhesus macaques. The timeline shows important studies of experimental M. tuberculosis infection in rhesus macaques during the 21st century, from 2001 to 2014 (13, 26–34, 36–43). Events are organized chronologically by year of publication. Locations of the experiments (red) and the corresponding author (blue) are specified accordingly. Boxed events refer to comparative TB studies using both cynomolgus and rhesus macaques. All studies, except for those reported in references 28 to 30 and 34, used Indian rhesus macaques. Abbreviations: BPRC, Biomedical Primate Research Center; rBCG, recombinant BCG.

FIG 3.

TB research using cynomolgus macaques. The timeline includes significant studies on cynomolgus macaques experimentally infected with M. tuberculosis from 1996 to 2014 (10, 15, 25, 42–59). Events are organized chronologically by year of publication. Locations of experiments (red) and the corresponding author (blue) are listed. Boxed events refer to comparative TB studies using both cynomolgus and rhesus macaques, as shown in Fig. 2. Abbreviations: BPRC, Biomedical Primate Research Center; IFN-γ, gamma interferon; TNF-α, tumor necrosis factor alpha; Treg, regulatory T cells; U Pitt, University of Pittsburgh.

The majority of NHP models of M. tuberculosis infection have involved either CM or Indian RM, except for at least five reported research projects which used Chinese RM in Wuhan, China (34, 35), Solna, Sweden (28, 29), and Rijswijk, the Netherlands (30). So far within the scientific literature illustrating macaque models of M. tuberculosis infection, the routes of inoculation have included i.b. instillation as well as i.t., aerosol, and intranasal (i.n.) infection. Although zoonotic TB outbreaks have revealed horizontal transmission of M. tuberculosis, an experimental macaque model of natural M. tuberculosis infection has not been reported. Tables 2 to 4 summarize by route the dose and M. tuberculosis strain of infection in each study design, along with the major findings of the listed papers.

TABLE 2.

TB studies using both rhesus and cynomolgus macaques^a

Exptl design		Major findings	Reference
No. of animals	M. tuberculosis strainb (inoculation route), dose(s) (CFU)
6 RM, 6 CM	Erdman (i.t.), 3,000	BCG provides greater protective efficacy against TB in CM than in RM	42c
8 RM, 15 CM	Erdman (i.t.), 1,000–3,000	In vitro IFN-γ assays provide reproducible, reliable results while causing less stress than the PPD skin test	61
5 RM, 4 CM	Erdman (i.t.), 100	Synthetic peptides may be used in lieu of full-length ESAT-6 protein in TB diagnostic antibody detection assays	62d
26 RM	Erdman (i.t.), 30–1,000	Highest rate of TB detection achieved when skin test is combined with PrimaTB STAT-PAK immunoassay	63e
4 RM	H37Rv (i.t.), 210
3 RM	Beijing (i.t.), 1,000
16 CM	Erdman (i.t.), 100–1,000
4 RM	H37Rv (i.b.), 1,000	The multiplex microbead immunoassay profiles M. tuberculosis antibodies at multiple stages of infection/disease	64
6 CM	Erdman (i.b.), 25
9 RM, 21 CM	Erdman K01 (aerosol), 30–500	MRI and stereology provides most accurate, quantifiable measurements of TB disease burden; RM exhibit higher susceptibility to M. tuberculosis than CM	60
6 RM	Erdman (i.b.), 500	M. tuberculosis antibody profiles depend on the NHP species and infecting M. tuberculosis strain but do not significantly change with TB disease progression	43
4 RM	H37Rv (i.b.), 1,000
14 CM	Erdman (i.b.), 25

^aAll studies used Indian RM models.

^bThe Erdman strain is a virulent set of M. tuberculosis isolates existing in two forms, including the laboratory ATCC 35801 isolate and the clinically isolated K01; this strain is most commonly used to study acute tuberculosis. Here, Erdman strain ATCC 35801 was used unless K01 is indicated. H37RV is an attenuated laboratory strain of M. tuberculosis typically used to study latent infection.

^cA TB vaccine-related study (42).

^dThe study reported in reference 62 also used African green monkeys.

^eIn the study reported in reference 63, additional RM, CM, and African green monkey groups were inoculated with the mycobacteria M. kansaii and M. avium.

RHESUS VERSUS CYNOMOLGUS MACAQUE MODELS OF TB

Both RM and CM models have served to evaluate the efficacy of TB vaccines/drugs, as well as improve our understanding of the immunopathogenesis of M. tuberculosis infection and reactivation. However, there are differences between the two species. For example, one investigation revealed that BCG provides greater protective efficacy in CM than in RM against M. tuberculosis infection (42). Later investigations exploring stereological techniques for measuring the bacterial burden showed that RM are more susceptible to M. tuberculosis infection than are CM (60). As a result, RM are more often used for the study of active TB, whereas CM provide better models of latent or chronic TB (10). Depending on the route and dose of infection, as well as the strain for the inoculum, either CM or RM can develop acute, chronic, or latent TB. Therefore, several research endeavors have employed both species for developing and determining the efficacy of TB screening immunoassays (61–64) (Table 2).

RHESUS MACAQUES

RM have been used extensively to study TB (Table 3). The early investigations during the Golden Age (17–21) specifically showed that M. tuberculosis infection in RM progresses rapidly, within 8 to 9 weeks after aerosol inhalation of the H37Rv strain administered at low doses of up to 62 CFU. In 2004, the Tulane NPRC established a model of asymptomatic, or latent, M. tuberculosis infection of RM (26). The investigators there revealed that RM are a good model for not only active TB but also asymptomatic TB when the investigators used lower doses (30 CFU) of the H37Rv strain. In addition to RM of Indian origin, Chinese RM can also be used as a viable model of acute TB (34). Chinese RM are highly susceptible to M. tuberculosis infection and develop active TB regardless of the dose of strain H37Rv used (34). In an attempt to improve the methods to monitor the progression of TB disease, Helke et al. (27) from the Oregon NPRC showed that use of high-resolution radiographic and fine immunologic studies helped define the disease status in RM as in humans. Namely, computed tomography (CT) evaluation and M. tuberculosis-specific T cell frequencies measured by enzyme-linked immunosorbent spot (ELISPOT) assays correlated well with the bacterial burden and severity of disease (27).

TABLE 3.

TB studies using rhesus macaques^a

Exptl design		Major findings	Reference(s)
No. of RM	M. tuberculosis strainb (inoculation route), dose(s) (CFU)
8	Erdman (i.b.), 10–150	RM are a good model for latent TB, with use of low doses of H37Rv	26
12	H37Rv (i.b.), 30–6,000,000
4	H37Rv (i.b.), 1,000	High-resolution radiographic and fine immunologic studies provide definition of TB disease progression	27
18	Erdman (i.t.), 500	Recombinant BCG (AFRO-1) induces strong antigen-specific T cell responses with TB vaccine vector (rAD35)	28, 29c
24	Erdman (i.t.), 1,000	MVA.85 boosting of BCG and an attenuated, phoP-deficient TB vaccine show protective efficacy against TB	30c
16	Erdman K01 (aerosol), 40–65	RM may be used as models of M. tuberculosis aerosol challenge; IFN-γ (ELISpot, ELISA) does not correlate with protection against TB; only MRI offers a reliable correlate	31
NP	NP	Early TB lesions have a highly proinflammatory environment, expressing IFN-γ, TNF-α, JAK, STAT, and C-C/C-X-C chemokines; in contrast, late TB lesions have a silenced inflammatory response	32
12	326 CDC1551 Himar 1 mutants (i.n.), 100,000	Virulence mechanisms of M. tuberculosis include transport of lipid virulence factors, biosynthesis of cell wall arabinan and peptidoglycan, DNA repair, sterol metabolism, and lung cell entry	33
9	H37Rv (i.b.), 50–3,000	Chinese RM are highly susceptible to M. tuberculosis infection and develop active TB regardless of the dose of strain H37Rv or Erdman used	34, 35
24	Erdman (i.b.), 25–500
16	CDC1551 (i.n.), 50	RM are an excellent model of TB/HIV coinfection and can be used to study TB latency and reactivation	36
6	Erdman K01 (i.b.), 500	Stereological analysis quantitative data show a strong correlation between bacterial load and lung granulomas	37
13	CDC1551 (i.n.), 5,000	The M. tuberculosis stress response factor sigH is required for M. tuberculosis growth and replication in mammalian lungs	38
3	Erdman (i.b./i.n.), 5–50	Newborn macaques infected with aerosolized M. tuberculosis develop human-like immunologic responses and are a good model for pediatric TB/HIV	40
32	Erdman K01 (i.b.), 275	RM aerosol vaccination with AERAS-402 elicits transient cellular immune responses in blood and robust, sustained immune responses in BAL fluid but does not protect against high-dose M. tuberculosis infection	13c
17	CDC1551 (aerosol), 100	Clinical profiles vary considerably among RM infected with M. tuberculosis but can help identify predictive biomarkers for TB susceptibility along with gene expression profiles	41

^aAll studies, except for those reported in references 28 to 30 and 33, used Indian rhesus macaques. Abbreviations: i.d., intradermal; i.n., intranasal; NP, not provided.

^bThe Erdman strain is most commonly used to study acute TB. It is a virulent subset of M. tuberculosis and exists in two forms, the laboratory isolate ATCC 35801 and the clinical isolate, K01; the Erdman ATCC 35801 strain was used in most studies, except those for which the K01 strain is indicated. H37RV is an attenuated laboratory strain of M. tuberculosis typically used to study latent TB infection. CDC1551 is a clinical isolate of M. tuberculosis and exhibits a similar degree of virulence as the Erdman strain.

^cTB vaccine-related study.

CYNOMOLGUS MACAQUES

In addition to RM, CM have aided researchers similarly in the study of TB (Table 4), particularly after the so-called Golden Age (16). The first published CM investigation of controlled M. tuberculosis infection stemmed from Walsh et al.'s (25) work in the Philippines (Fig. 3), in collaboration with the University of California Los Angeles School of Medicine. In this project, the Philippine CM were found to make an excellent model of not only acute TB but also chronic TB. This research pioneered the intratracheal infection of Philippine CM in a TB model. Within 7 years, Capuano et al. (10) from The University of Pittsburgh developed a CM model representing the full spectrum of human M. tuberculosis infection by infecting the macaques after intrabronchial inoculation with a low dose (25 CFU) of the Erdman strain. Interestingly, this low dose of inoculum precipitated various reactions in the CM, ranging from latent TB to active-chronic and even rapidly progressive TB. The CM's pulmonary and even extrapulmonary manifestations of disease within the different stages of TB infection closely resembled the pathological and clinical findings in human TB, as confirmed by laboratory assessments, including purified protein derivative (PPD) tests and erythrocyte sedimentation rate (ESR) profiling (10) (Table 1). A substantial number of the proceeding publications on CM models of M. tuberculosis infection were similarly published by researchers at The University of Pittsburgh (47–50, 52, 56, 57). Therefore, CM models have now been used for evaluating TB drugs and vaccines (42, 46, 52).

TABLE 4.

TB studies using cynomolgus macaques^a

Exptl design		Major findings	Reference(s)
No. of CM	M. tuberculosis strain (inoculation), dose(s) (CFU)
28	Erdman (i.t.), 10–100,000	Philippine CM provide an excellent model of chronic TB	25
17	Erdman (i.b.), 25	Low-dose infection of CM represents the full spectrum of human M. tuberculosis infection and provides a model to study latent as well as active-chronic and rapidly progressive TB	10
16	Erdman (i.t.), 500	CM vaccination with the 72f rBCG vaccine provides better protective efficacy than with BCG	44b
44	Erdman (i.t.), 500	CM vaccination with the HSP65 plus IL-12/HVJ vaccine provides better protective efficacy than BCG	44, 45b
15	Erdman (i.t.), dose not reported	CM vaccination with Mtb72F/AS02A provides greater protective efficacy than BCG alone	46b
24	Erdman (i.b.), 1,000	CM vaccination with mc26020 or mc26030 provides less protection than with BCG	15b
25	Erdman (i.b.), 25	At necropsy, CM with active TB have more lung T cells and more IFN-γ from PBMC, BAL fluid, and mediastinal lymph nodes than CM with latent TB	47
24	Erdman (i.b.), 25	Neutralization of TNF results in disseminated disease in acute and latent TB infection with normal granuloma structure in a CM model	48
41	Erdman (i.b.), 25	Increased regulatory T cells in active TB occur in response to increased inflammation, not as a causal factor of disease progression	49
15	Erdman (i.b.), 25	Reactivation of latent TB with SIV is associated with early T cell depletion and not virus load	50
7	Erdman (i.b.), 25	M. tuberculosis-specific multifunctional T cells are better correlates of antigen load and disease status than of protection	51c
5	Erdman (i.b.), 200
33	Erdman (i.t.), 25–500	The multistage vaccine H56 boosts effects of BCG to protect CM against active TB and reactivation of latent TB	52
14	Erdman (i.b.), 25–200	The CM model of M. tuberculosis infection mimics human TB, particularly in granuloma type and structure	53
8	Erdman (i.t.), 250	M. tuberculosis may modulate protective immune responses via the use of indoleamine 2,3-dioxygenase (an immunosuppressant) found in nonlymphocytic regions of TB granulomas	14b
9	Erdman (i.b.), 25	Experimental and epidemiologic estimates of the M. tuberculosis mutation rate are comparable	54
27	Erdman (i.b.), 500	Early expansion/differentiation of Vγ2Vδ2 T effector cells during M. tuberculosis infection increases resistance to TB	55
26	Erdman (i.b.), 25–400	TB granulomas evolve and resolve independently within a single host; individual lesions respond differently to different drugs; overall PET and CT signals can predict successful TB drug treatment	56
12	Erdman (i.b.), 1,000	Compared to nonvaccinated CM, BCG-vaccinated CM exhibit higher expression levels of TNF-α, IL-10, IL-1b, TLR4, IL-17, IL-6, IL-12, and iNOS in lungs	58b
39	Erdman (i.b.), 25	Sterilization of TB granulomas occurs in both active and latent TB amid the differential killing of M. tuberculosis within a single host	57
2	SNP strains (i.b.), 34
8	Erdman (i.b.), 240–500	CM vaccination with BCG transiently increases levels of macrophages and lymphocytes in blood, with later recruitment in the lungs; however, M. tuberculosis continues to replicate in lungs	59b

^aAbbreviations: SNP strains, strains with a single-nucleotide polymorphism mutation; rBCG, recombinant BCG; BAL, bronchoalveolar lavage; PBMC, peripheral blood mononuclear cells; iNOS, inducible nitric oxygen synthase.

^bTB vaccine-related study.

^cAnimals were coinfected with M. tuberculosis and SIV.

SPECIFIC MACAQUE MODELS OF TB

Macaque models for TB vaccine evaluation.

Using the RM model, researchers from the Swedish Institute of Infectious Disease Control strived to augment, broaden, and prolong immune protection against M. tuberculosis. They showed that a recombinant BCG (AFRO-1) could induce strong antigen-specific T cell responses when combined with the TB vaccine vector rAD35 (28, 29). Two other investigations using RM models of M. tuberculosis infection also focused on evaluating TB vaccines (30, 31). The Biomedical Primate Research Center in Rijswijk, the Netherlands, aimed to develop a model to test TB vaccines before progressing to human clinical trials (30). By using an RM model employing intratracheal inoculation with the Erdman strain of M. tuberculosis, the study revealed that prior immunization with the MVA.85-boosted BCG and an attenuated, phoP-deficient TB vaccine provided effective protection against M. tuberculosis infection. A novel aerosol challenge model effected by Sharpe et al. (31) helped to assess the endpoints for testing the BCG/MVA.85 vaccine. This NHP model used a three-jet collision nebulizer in addition to a modified Henderson apparatus (a device for studying the infectivity and virulence of microorganisms in small air droplets; the 3 components include a continuous aerosol-generating unit (collision spray), exposure unit, and sampling unit), as described by Barclay et al. (18), in a head-out plethysmography chamber. The results of these studies indicated that the gamma interferon (IFN-γ) indices—from ELISPOT assays and enzyme-linked immunosorbent assays (ELISAs)—do not relate to protection against TB; only the magnetic resonance imaging (MRI) readouts offered a reliable correlate, using ex vivo lung samples removed at necropsy (18, 31). Another way to objectively measure the efficacy of TB vaccines and drugs was later developed by Luciw et al. (37) at the California NPRC. The research findings from these studies determined that stereological analysis (i.e., three-dimensional interpretation of planar sections of materials or tissues) from MRI could provide quantitative data, showing a significant correlation between bacterial load and lung granulomas.

Researchers did not use CM for testing TB vaccines until much later than for the early RM models of TB. Between 2005 and 2012, three reported studies using CM models evaluated the efficacy of TB vaccines that had been tested previously with smaller animal models, including guinea pigs and mice (44, 45). The first project evaluated the efficacy of the 72f recombinant BCG vaccine and HASP65 plus interleukin-12 (IL-12)/HVJ vaccine (44), which another team of investigators also tested later in Osaka, Japan (45). Both research investigations showed that the recombinant BCG vaccines were more effective than BCG alone. In subsequent years, the multistage vaccine H56 was found to boost the effects of BCG to protect CM against active TB and the reactivation of latent TB (52). Ultimately, the macaque models have exhibited the potential to help evaluate preventative methods and interventions before reaching human clinical trials, particularly for TB vaccines.

Macaque models for TB drug evaluation.

Finally, another recently reported investigation of experimental M. tuberculosis infection in a CM model was also published from the University of Pittsburgh School of Medicine (56, 57). The purpose of this research endeavor was to determine potential alternative markers for evaluating the efficacy of TB drugs. Specifically, the studies compared the overall metabolic and radiographic changes, as well as the alterations within individual granulomas, in CM infected with M. tuberculosis. Of note, the results revealed that TB granulomas evolve and resolve independently within a single host and that individual lesions respond variably to different drugs (56). Furthermore, the clinical findings concluded that the overall positron emission tomography (PET) and CT signals could be used as prognostic markers to predict successful TB drug therapies. However, the overall metabolic and radiographic changes reported by this study were nonspecific indicators of metabolic activity, measured from [¹⁸F]FDG-radiolabeled glucose in PET/CT imaging data (56).

Macaque models for the study of TB pathogenesis.

By gaining a better understanding of the pathogenesis of M. tuberculosis infection and reactivation of latent TB, researchers can further develop and improve treatment strategies. As such, investigators at the Tulane NPRC profiled the TB granuloma transcriptome in an RM model to identify key immune signaling pathways that are activated during M. tuberculosis infection (32). Previously, scientists from the Chicago Center for Biomedical Research characterized gene networks in RM after only BCG vaccination/infection (65). Mechanistic studies of M. tuberculosis delineated more specific factors employed by M. tuberculosis to successfully infect and persist in mammalian lungs (33). The identification of a potential therapeutic target sparked from the discovery of the M. tuberculosis stress response factor SigH as an important player in the growth/replication of M. tuberculosis (38, 39).

Lin et al. (47) characterized the clinical manifestations of TB disease within the three stages: latent, active-chronic, and rapidly progressive TB. The results of these studies showed that at necropsy, CM with active TB had more CD4⁺ and CD8⁺ T cells in the lungs and more gamma interferon from peripheral blood mononuclear cells, bronchoalveolar lavage fluid, and mediastinal lymph nodes than CM with latent TB (47). Investigators from the same group also showed that tumor necrosis factor neutralization resulted in disseminated disease in both acute and latent TB with normal granulomatous structures (48). Another CM study examined the role of CD4⁺ regulatory T cells in active TB; these cells occurred in response to increased inflammation rather than acting as a causative factor in the progression to active disease (49).

In light of the discovery that TB granulomas change uniquely within a host (56), the latest published investigation by Lin et al. (57) aimed to define how these structures vary between CM with latent and active TB. Interestingly, the sterilization of the granulomas occurred in both stages of TB, regardless of the differential killing of bacteria within a single host (57). Moreover, TB vaccine-related studies have elucidated possible genetic mechanisms of host resistance to M. tuberculosis after immunization (58), as well as immunosuppressant mechanisms of M. tuberculosis virulence (59). While adding insight into the molecular and pathological pathways of TB progression, these findings may also help to evaluate the disease status as well as spur the development of novel therapeutic targets.

Macaque models of TB/HIV coinfection.

Macaques also serve as an excellent model of TB/HIV coinfection, which is of importance to understand TB latency and reactivation (36). Upon inoculation with high-dose BCG (36, 66, 67) or low-dose Erdman strain (25 CFU, i.b.) (50), latently infected RM and CM, respectively, had reactivated TB when coinfected with the simian immunodeficiency virus (SIV). Pathogenic SIV-BCG interactions facilitated the development of TB-like disease (67), while antiretroviral agents restored M. tuberculosis-specific T cell immune responses (36). The reactivation of latent TB in CM infected with SIV was associated with early T cell depletion and not the virus load (50). Another particularly innovative RM model of TB was established at the Southwest NPRC (40). The purpose of that study was to institute an NHP model for pediatric TB/HIV coinfection. Newborn macaques were infected with M. tuberculosis Erdman strain intrabronchially or via the aerosol route, using an ultrasonic nebulizer specifically adapted for the newborn macaque nose. Investigators confirmed M. tuberculosis infection by various methods, including chest X-rays, ELISPOT, bronchoalveolar and gastric lavages, and necropsy. Because people with HIV/AIDS carry a high risk of M. tuberculosis infection and disease severity, it is clinically significant to have a model that mimics coinfection.

KEY LESSONS AND FUTURE DIRECTIONS

Collectively, the literature on macaque models of TB has shared four important lessons and opportunities for improvement. First, we recognize that the biosafety requirements, cost of equipment and maintenance, and animal availability have impeded scientists from using NHP models in TB research. To address these issues, researchers may experiment with smaller genera and species of NHPs that still recapitulate human TB, such as the common marmoset (Callithrix jacchus) model (68). Second, we found that different animal species, individually and as a whole, respond differently when exposed to M. tuberculosis in terms of immunopathogenesis of the disease (43, 64) (Table 2). Therefore, investigators must consider the purpose of the study for the appropriate selection of NHP species. Third, the strain of M. tuberculosis used for inoculation can significantly impact the disease outcome. Instead of employing merely the laboratory-adopted Erdman and H37Rv strains, study designs should include clinical isolates of M. tuberculosis. Although this approach may introduce more variability into the NHP studies of TB while making studies performed at different sites more difficult to compare, the findings would likely further mimic human TB disease and add valuable knowledge to the field. Lastly, most of the macaque studies so far have shown only acute TB, which is much less prevalent than latent TB in humans. Hence, it is necessary to establish more clinically relevant NHP models that resemble human passive airway transmission. By creating an NHP model of natural M. tuberculosis infection under experimental conditions, researchers would be able to test new and existing TB vaccines/drugs for protection against M. tuberculosis infection.

CONCLUDING REMARKS

Evidently, the macaque models have served as a valuable tool in TB research over the past several decades. As demonstrated in the literature, the CM and RM TB models have revealed clinically similar manifestations of TB disease or latency through multiple diagnostic and prognostic parameters. Even the variabilities in immune responses to M. tuberculosis infection imitate the diverse human host reactions to the pathogen (41). Although a number of studies on TB vaccine evaluation have been conducted using experimentally infected macaques, the majority of these studies merely showed the reduction of TB disease progression by BCG-based vaccines. Therefore, an urgent need still exists in order to establish macaque models that can demonstrate protection against M. tuberculosis infection besides TB disease progression. Fortunately, some investigators and funding institutions have revamped an interest in further developing NHP models for TB research. Undoubtedly, given the reemerging TB epidemic, many people worldwide should benefit from more advanced treatment and prevention strategies against M. tuberculosis infection and disease progression. Hence, the use of NHP models should be considered a highly effective means of reaching these common goals. In essence, from a global health standpoint, there is truth in the words of the literary science author Chris Roberson: “Everything is improved by the judicious application of primates.”

Biographies

Juliet C. Peña, M.D., M.P.H., dedicated 1 year (2013 to 2014) as a postdoctoral fellow in WenZhe Ho's laboratory at the Temple University School of Medicine in Philadelphia, PA, where she focused her studies on the immunopathogenesis of tuberculosis and HIV. Her interest in tuberculosis research stemmed from her medical and public health background, as well as her collaborations with Wen-Zhe Ho's laboratory using monkey models of tuberculosis. She earned her M.D. (2012) and M.P.H. (2014) from The University of Arizona College of Medicine and the Mel and Enid Zuckerman College of Public Health. Currently, Dr. Peña is completing a health policy fellowship at the Office of Disease Prevention and Health Promotion within the U.S. Department of Health and Human Services.

Wen-Zhe Ho has 30 years of research experience in understanding the interactions between host innate immunity and the pathogens HIV, hepatitis C virus, and M. tuberculosis. He received his M.D. and clinical training as a pediatrician in infectious diseases at Wuhan University, China in the 1980s. He was a postdoctoral fellow in infectious diseases at the Children's Hospital of Philadelphia and the Wistar Institute of the University of Pennsylvania, where he became a full professor (2005). He also received his M.P.H. at the University of Pennsylvania (2006), through which he gained a greater interest in public health issues of M. tuberculosis and HIV coinfection. In 2009, Dr. Ho moved to Temple University, where he now serves as a tenured professor in the Departments of Pathology and Cell Biology. Additionally, as Director/Professor at the Center for Animal Experiment/ABSL-III Laboratory of Wuhan University (2010 to present), he continues his research on M. tuberculosis infection using monkey models.