Mycobacterial Biofilms: Revisiting Tuberculosis Bacilli in Extracellular Necrotizing Lesions

ABSTRACT

Under detergent-free in vitro conditions, Mycobacterium tuberculosis, the etiological agent of tuberculosis in humans, spontaneously forms organized multicellular structures called biofilms. Moreover, in vitro biofilms of M. tuberculosis are more persistent against antibiotics than their single-cell planktonic counterparts, thereby raising questions about the occurrence of biofilms in the host tissues and their significance in persistence during chemotherapy of tuberculosis. In this article, we present arguments that extracellular M. tuberculosis in necrotizing lesions likely grows as biofilms.

INTRODUCTION

The ongoing emergence of multidrug-resistant (MDR) and extensively drug-resistant (XDR) strains of Mycobacterium tuberculosis not only underscores the limitations of our current tuberculosis (TB) control strategies but is also escalating the TB epidemic to a new level. Realizing the imminent threats of MDR- and XDR-TB and the urgency for new TB control measures, the World Health Organization has maintained TB control as high priority and set an ambitious goal of eradicating the disease by 2030 (1). What remains an urgent need is the development of a shorter-duration combination of antimicrobial drug treatments that is more effective at eradicating drug-susceptible and drug-resistant strains of M. tuberculosis. However, progress toward this goal is hampered by a lack of understanding of factors that contribute to the expression of in vivo drug tolerance by M. tuberculosis, which contributes significantly to the need to treat patients from 6 to 9 months with antimicrobial drug combinations that have toxic side effects. In this review, we discuss the current state of our understanding of the host and pathogen factors that contribute to M. tuberculosis drug tolerance. Moreover, we highlight potential strategies that can be used to improve the efficacy of existing drugs against drug-tolerant M. tuberculosis. These strategies are based on our current knowledge of how and where drug-tolerant bacilli persist and on features of the complex host response that likely limit the penetration of antibiotics. A better understanding of the factors that contribute to the expression of drug tolerance reveals the potential value of adjunctive therapies that can be used to potentiate the effectiveness of existing and future anti-TB drugs.

NECROTIZING LESIONS: THE CHARACTERISTIC PATHOLOGY OF ACTIVE PULMONARY TB

In general, the host response to M. tuberculosis infection is best characterized as mixed inflammation, composed primarily of macrophages and lymphocytes that accumulate at the site of primary infection in the lung or in extrapulmonary tissues. In humans and some animal model species, mixed inflammatory cells are organized into a nodular mass referred to as a granuloma. In other species, like most strains of mice, the inflammatory and immune cell types, although similar, fail to organize into discrete granulomas. The morphological features of TB granulomas are dynamic and variable, being influenced by combinations of host, pathogen, and environmental factors. Besides the aforementioned species-specific differences, granuloma composition and structure can also be influenced by the relative susceptibility of the host, stage of infection, virulence of the M. tuberculosis strain(s), presence/absence of comorbidities, and whether individuals are treated and responding appropriately to antimicrobial drug treatment. Despite the different granuloma morphotypes, the prototypical TB granuloma is composed of centrally located macrophages surrounded by an ill-defined rim of different lymphocyte subsets, fewer plasma cells, multinucleated giant cells, and granulocytes. In addition, granuloma morphology can be altered by the presence or absence of central necrosis, dystrophic calcification, and fibrous encapsulation. Calcification and fibrosis are indicative of wound or lesion healing that occurs concurrently with active inflammation, especially in patients with chronic TB. The development of granuloma necrosis and calcification is significant in that they represent, at the least, a localized loss of normal tissue structure and function and associated irreversible tissue damage that can persist for the life of the patient.

While the granuloma is recognized as the response to primary infection in humans and some animals, the clinical signs of active TB in humans are often the result of a post-primary manifestation that can occur years or decades following the initial exposure (2). The pathogenesis of postprimary tuberculosis is complex and poorly understood, and is not easily reproduced in animal models including non-human primates. An emerging hypothesis is that postprimary TB is associated with airway obstruction that alters the lung microenvironment to favor rapid M. tuberculosis proliferation, which stimulates an aggressive and destructive proinflammatory response that predisposes to cavitary disease and thus large numbers of extracellular bacilli (3, 4). Cavitary TB is the most severe manifestation of active TB disease, in which an unregulated immune response degrades and replaces normal lung parenchyma. The formation of an open cavity is determined in part by the location of the destructive inflammatory response and whether lesion necrosis develops adjacent to and communicates with conducting airways. The progression of TB disease to necrosis or lesion cavitation is not only detrimental to the host, but also represents an important transition from a predominantly intracellular infection to now include extracellular bacilli released from infected cells. In addition, the relatively normal oxygen concentration in lesions connected with airways supports the proliferation of high numbers of extracellular bacilli, which further contributes to inflammation and necrosis. This transition alone contributes to the complexity of the host microenvironment as well as the physiological state of different M. tuberculosis populations (5–7).

EXTRACELLULAR M. TUBERCULOSIS IN NECROTIZING LESIONS: A PROTECTED NICHE FOR THE PATHOGEN

It is generally accepted that granuloma formation in response to M. tuberculosis infection functions as a protective barrier that contains bacilli from spreading from the site of primary infection within a single host or between hosts. The accumulation of mixed immune cells that make up the granulomatous response acts not only as a mechanical barrier, but also as a functional barrier given that cell-mediated immunity is critical to controlling M. tuberculosis infection. As mentioned above, the development of lesion necrosis and cavitation further contributes to the functional diversity of M. tuberculosis populations within individual lesions even within a single host. As infection of macrophages allows bacilli to evade innate and adaptive immune surveillance, extracellular bacilli also gain a survival advantage through physical separation from circulating immune cells and resistance to phagocytosis. By virtue of being associated with cellular and tissue necrosis, extracellular bacilli are sequestered within a microenvironment that has little or no blood supply that not only limits oxygen delivery, but also the ability of effector immune cells to circulate within lesions, and limits the penetration and accumulation of antimicrobial drugs (8–10). Moreover, even though M. tuberculosis requires oxygen to effectively replicate in vitro and in vivo, bacilli can survive or persist in a nonreplicative state for a long period of time under hypoxic or anoxic conditions (11–13). Studies have shown that, in humans and animal model species that typically form necrotic TB lesions and thus harbor viable, extracellular bacilli, lesions are measurably hypoxic (12) and drug penetration is limited or absent (10, 14). Besides host factors, the changing lesion microenvironment contributes to changes in bacilli physiology. In vitro studies have clearly demonstrated that hypoxia is among the most important inducers of the dosRS-dependent dormancy regulon in M. tuberculosis (15).

Since the discovery that M. tuberculosis infects and survives within macrophages in vitro, much has been learned about how M. tuberculosis circumvents intracellular killing (16, 17). However, very little is known about the importance of extracellular bacilli in the pathogenesis of active TB disease or the clinical manifestations of in vivo drug tolerance in humans and animals. Even in the pre-antibiotic era, investigators recognized the importance of extracellular M. tuberculosis in human TB lesions, especially those with central necrosis or cavitation (18–21). Typically, the distribution of both intracellular and extracellular bacilli in the context of naturally occurring TB in humans and experimental infections in animals has been studied through the use of acid-fast staining of histological sections postmortem or from surgical biopsies. Studies in both humans and animals using different staining techniques suggest that acid-fast positive bacilli represent a fraction of the organisms that make up the total bacterial burden in TB lesions (22). In humans, Nyka et al. demonstrated in multiple studies that acid-fast negative, extracellular bacilli in human TB lesions form large microbial communities morphologically resembling biofilms formed by other pathogenic bacteria that cause extracellular infection (23–26). Biofilms are defined as matrix-encapsulated microbial communities attached to biotic or abiotic surfaces that are self-assembled through a genetically programmed developmental process (27). The high cell density, cell-cell contacts, and different nutrient and oxygen gradients within the interiors of biofilms facilitate expression of several unique phenotypes, including antimicrobial tolerance, that are not expressed by the same organisms grown as unattached, free-living organisms referred to as planktonic growth (28, 29). The most striking and clinically significant feature of biofilm-forming bacteria is their extraordinary recalcitrance to antibiotics (30, 31). Several mycobacterial species including M. tuberculosis spontaneously form biofilms in vitro (32–36) (Fig. 1), although the architectural and functional properties of the extracellular M. tuberculosis aggregates in necrotizing lesions remain unknown.

FIGURE 1.

Visualization of M. smegmatis growth in a microfluidic device by time-lapse microscopy. The numbers at the bottom of the snapshots denote the time in minutes at which the snaps were taken. Note the distinct foci of multicellular communities from growth of individual cells. (Data collected by Jacob Richards in the laboratory of Anil Ojha).

Orme and, more recently, Wong et al. refer to these extracellular bacilli in vivo as necrosis-associated extracellular clusters or NECs (37, 38). As mentioned previously, a large proportion of these extracellular bacilli are acid-fast negative. In an effort to visualize heterogeneous populations of extracellular bacilli in animal models, Lenaerts et al. developed a fluorescent DNA-staining protocol that shows both acid-fast negative and positive bacilli. As a consequence, they confirmed that M. tuberculosis persists as large clusters of both intracellular and extracellular bacilli in the C3HFeJ strain of mice that develop necrotic lesions following aerosol exposure (22). These data show that the propensity of M. tuberculosis to form in vivo microbial communities is not limited to extracellular bacilli but can accumulate within an intracellular compartment as well. The concept of intracellular biofilm formation was first suggested in the study of uropathogenic Escherichia coli in the urinary bladder of patients with recurring urinary tract infections (39), and later shown to persist against antibiotics within transitional epithelium (40–42). The possibility that intracellular and extracellular communities of M. tuberculosis contribute to the expression of drug tolerance in vivo is in need of further investigation.

EXTRACELLULAR M. TUBERCULOSIS PERSISTS AGAINST ANTIBIOTICS TREATMENT

Recent animal studies have found that extracellular bacilli are among the populations of M. tuberculosis that persist and express in vivo antimicrobial drug tolerance, especially in model species that develop necrotic granulomas (7, 43, 44). Using the guinea pig model, which develops well-organized TB granulomas similar to those in humans, Lenaerts et al. reported that the majority of the acid-fast M. tuberculosis bacilli that recovered from a truncated exposure of antibiotics were localized in acellular rims of necrotizing lesions (44). This was later verified by comparing the host response to M. tuberculosis infection in different strains of mice that do or do not form necrotic lung lesions (6). Drug treatment of M. tuberculosis-infected mice produced a more uniform decline in acid-fast bacilli across the lesions in strains that fail to develop lesion necrosis, whereas antimicrobial drug treatment of mice strains that do develop necrotic lesions, and of guinea pigs, resulted in disproportionately higher levels of extracellular, drug-tolerant bacilli (6). Although the visualization methods in these studies fail to clarify the live/dead status of the recovered acid-fast bacilli from drug-treated tissues, the findings nevertheless provide clues as to the host factors that contribute to the in vivo expression of antimicrobial drug by M. tuberculosis.

BIOFILMS: A NEW PERSPECTIVE OF EXTRACELLULAR M. TUBERCULOSIS IN NECROTIZING LESIONS

In the animal studies discussed above, the investigators made an intriguing observation of diffused rhodamine staining pattern around the acellular rims of necrotizing lesions. In light of the fact that rhodamine readily stains mycolic acids, the authors speculated that the diffused material could likely be M. tuberculosis-derived mycolic acids, either actively secreted by viable bacilli or accumulated after bacterial death and degradation (6). The former possibility assumes greater significance from the fact that mycolic acids are abundantly produced and secreted as free acids by in vitro cultures of mycobacteria in detergent-free medium, in which the bacilli typically grow as self-organized, surface-associated, multicellular communities, leading to development of pellicles on air-medium interface (33) or colonies on solid substratum (Fig. 1). Moreover, a direct role of free-mycolic acids (FM) in formation of pellicle is suggested by a groEL1 mutant of Mycobacterium smegmatis, in which the defects of the mutant in forming pellicle are linked to instability of mycolic acid modulating enzyme KasA and KasB that results in lower abundance of FM (32, 45).

The phenotypic and functional characteristics of mycobacteria grown as pellicles in vitro adhere to the general definition of biofilms in a way that not only are the pellicles more resistant to antibiotics, but also that their development proceeds through genetically distinct stages (32, 33). Taken together, colocalization of diffused (secreted) mycolic acids and drug-tolerant persisters in the acellular rim of necrotizing lesions support the hypothesis that mycobacteria could likely grow as biofilms in such host niches. Interestingly, Wong and Jacobs suggested that the multicellular growth of mycobacteria inside the host could be an active process regulated by the pathogen (37). The authors found that M. tuberculosis actively produces signals through the ESX-1 pathway to induce lysis of macrophages. Subsequent release of host DNA in the extracellular compartment, also called extracellular traps, appears to facilitate aggregated growth of the pathogen (37). Given that DNA is a key component of extracellular matrix in biofilms produced by many pathogenic bacteria, it is reasonable to argue that host-derived DNA can also contribute significantly to the extracellular matrix of M. tuberculosis biofilms, especially in necrotic lesions (46).

Nick et al. showed that neutrophil-derived eDNA contributed significantly to biofilm formation by Pseudomonas aeruginosa in vitro, and that targeting DNA and the host cytoskeletal protein actin enzymatically dispersed microbial communities, which restored antimicrobial drug susceptibility (47, 48). Ackart et al. developed a similar in vitro assay in which lysed human neutrophils served as an attachment matrix for extracellular M. tuberculosis. They went on to show that M. tuberculosis formed complex microbial communities similar to those described for other known biofilm-forming bacterial species (49). Moreover, bacilli attached to host-derived macromolecules expressed a nonreplicating phenotype and extreme tolerance to first-line anti-TB drugs alone or in combination. This in vitro model system was used as a platform to screen a library of 2-aminoimidazole (2-AI)-based small molecules that have been shown to have biofilm-inhibiting and -dispersing activity against a wide variety of Gram-positive and Gram-negative bacteria (50). These data showed that second-generation 2-AI small molecules were even more effective at restoring susceptibility of drug-tolerant bacilli to isoniazid and rifampin by directly targeting attached communities of M. tuberculosis (51). In more recent unpublished studies, these investigators have shown that 2-AI compounds are also effective at reversing the inherent resistance of M. smegmatis and M. tuberculosis to beta-lactam antibiotics. These data demonstrate the potential use of small molecules as adjunctive therapy to restore antimicrobial susceptibility of M. tuberculosis expressing drug tolerance through extracellular biofilm formation.

Besides an extraordinary recalcitrance to antibiotics, pathogenic bacterial biofilms also successfully subvert the host immunity to establish chronic infections. For example, biofilm formation by Streptococcus pneumoniae evades recognition by the immune system by inhibiting complement binding and phagocytosis (52). In an in vitro culture of E. coli, the phagocytosis of planktonic bacteria by macrophages was significantly more efficient than that of bacteria maintained as a biofilm (53). Biofilm formation can also impair antimicrobial killing by neutrophils. Neutrophils are often the first cells to encounter bacteria during the early stages of infection and have multiple antimicrobial strategies for killing both intracellular and extracellular bacteria. Recent studies have shown that microbial communities are not completely resistant to killing by neutrophils but the impairment of antimicrobial defenses is somewhat dependent on a combination of host and pathogen factors (54). In the case of M. tuberculosis, Lenaerts et al. showed that growth of bacilli under hypoxic conditions in vitro resulted in the secretion of extracellular DNA (22, 55), which may, in combination with host DNA and other macromolecules, impair phagocytosis and extracellular killing (56, 57). This raises a possible linkage between the characteristic chronic infection by M. tuberculosis and its ability to form aggregated community in extracellular niches.

SUMMARY AND OUTLOOK

Taken together, in vivo aggregates of extracellular M. tuberculosis represent an interesting and perhaps physiologically distinct entity that could influence the clinical characteristics of TB. For further investigation into the significance of these aggregates in M. tuberculosis persistence in vivo, in vitro studies on growth characteristics of the pathogen are crucial. Using in vitro growth models, addressing fundamental questions such as how do individual bacteria attach and aggregate to biotic and abiotic surfaces, what facilitates their adaptation to interior microenvironment of the aggregates, and how these processes impact their fitness against host defense and drug treatment stresses would provide molecular tools for characterization of in vivo aggregates. Mutational analysis is a powerful approach to address these questions. However, a straightforward genetic correlation between in vitro and in vivo phenotypes of a mutant is often difficult to infer because of multiple possible effects of a mutation on the pathogen. For example, mmaA4, a methyltransferase involved in synthesis of oxygenated mycolic acids, influences cell wall permeability of M. tuberculosis as well as host-pathogen interaction (58, 59). Because of such pleiotropic effects of mmaA4, its requirement in biofilm formation in vitro and growth in vivo offers limited correlation between the two phenotypes (60, 61). These limitations could be circumvented by an integrative approach that combines advanced microscopy with high-throughput genetics (Tn-seq), transcriptomics (RNAseq), and bioinformatics techniques to determine the specific biomarkers associated with M. tuberculosis biofilms in vitro. Such biomarkers could serve as valuable reagents for understanding the role of extracellular aggregate in M. tuberculosis persistence and pathogenesis.