Targeting drug tolerance in mycobacteria: a perspective from mycobacterial biofilms

Abstract

Multidrug chemotherapy for 6–9-months is one of the primary treatments in effective control of tuberculosis, although the mechanisms underlying the persistence of its etiological agent, Mycobacterium tuberculosis, against antibiotics remain unclear. Ever-mounting evidence indicates that the survival of many environmental and pathogenic microbial species against antibiotics is influenced by their ability to grow as surface-associated multicellular communities called biofilms. In recent years, several mycobacterial species, including M. tuberculosis, have been found to form drug-tolerant biofilms in vitro through genetically controlled mechanisms. In this review, the authors discuss the relevance of the in vitro mycobacterial biofilms in understanding the antibiotic recalcitrance of tuberculosis infections.

Mycobacterial infections are uniquely recalcitrant to antibiotics

Treatment of a typical infection of Mycobacterium tuberculosis – the causative pathogen of tuberculosis (TB) – requires 6–9 months of chemotherapy with isoniazid, pyrazinamide, rifampicin and often ethambutol – a multidrug regimen developed nearly 40 years ago [1,2]. This extended treatment has had a devastating impact on the global burden of TB, which is estimated to be prevalent in either clinical or subclinical states in about a third of the global human population, and kills approximately 1.7 million people every year [3]. In particular, the long chemotherapy regimen can be attributed to patients’ noncompliance and poor case management in resource-limited countries with a high burden of the disease. An incomplete course of treatment is considered to be one of the primary contributors to the emergence of multidrug-resistant TB and extremely drug-resistant TB [4,5]. A long and complex chemotherapy of TB poses further challenges to the treatment of HIV–TB coinfections – a leading cause of death in AIDS patients [6,7]. The extended chemotherapy is necessary to sterilize a small subpopulation of drug-tolerant M. tuberculosis bacilli, although it remains unclear where and how these persisters survive in the tissues [8]. Overall, an improvement in case management and subsequent control of active TB require a new generation of drugs that can clear the infection efficiently in a shorter period of time. Thus, it is time to evaluate the perspectives of mycobacterial persistence against antibiotics in the context of existing broader concepts emerging from studies on other bacterial pathogens. In this review, the authors discuss how the established paradigm of microbial growth and persistence in organized multicellular communities, called biofilms, could be relevant to the understanding of drug recalcitrance in TB infections.

Mycobacteria have a strong propensity to grow in multicellular structures

Cultures of mycobacteria in vitro spontaneously produce macroscopic structures, leading to the development of pellicles on the air–media interface. To circumvent the challenges of these growth characteristics in clonal purification of the strains, Dubos and Middlebrook described a method of growing M. tuberculosis in the presence of polysorbate-80, which produced a homogeneously dispersed suspension without any signif icant impact on the virulence properties of the pathogen [9]. The detergent is now widely used in the mycobacterial growth medium, although there are potentially significant pitfalls in the interpretations of biological characteristics of the bacilli grown in such a condition. First, the interaction of a detergent with surface-exposed, noncovalently-associated lipids of the bacteria, and subsequent alteration in the structure of the envelope, can potentially have a profound impact on the functional properties of the envelope, including altered permeability to small molecules. For instance, mycobacteria grown in media with polysorbate-80 are significantly more sensitive to antibiotics than those grown without the detergent [10]. Second, bacteria in a planktonic culture are homogeneously exposed to nutrients and oxygen, unlike those in the clusters exposed to heterogeneous microenvironments, with a self-generated gradient of nutrients and oxygen from the surface to the interior core. While cells on the surface of the clusters would presumably be exposed to rich growth conditions, those at the core could possibly encounter bacteriostatic conditions. The resulting nonuniform adaptive responses in the population can foster unique phenotypic diversity, most likely leading to the emergence of a stress tolerant subpopulation.

Thus, structured growth of mycobacteria in clusters can promote phenotypic persistence of constituent bacilli through physical protection from the environmental threats and self-formed bacteriostatic microenvironments, adaptation to which could facilitate the development of stress-tolerant physiology. However, complexities of mycobacterial growth in structured aggregates and involvement of biological mechanisms in the formation of the structures is under-appreciated by the long-held simplistic perspective in which cellular aggregation is believed to be a likely consequence of hydrophobic interactions between the waxy surfaces of the bacilli.

Biofilms: a genetically programmed lifestyle of microbes

In a landmark microscopic study of cystic fibrosis lung tissues, Costerton and colleagues observed that the growth of Pseudomonas aeruginosa, the causative bacterial pathogen, occurred in microcolonies encapsulated by extracellular polysaccharides (EPS) [11]. Subsequently, a similar pattern of clustered growth of Staphylococcus aureus was observed upon microscopic examination of infected implantation devices [12]. These studies initiated a shift in the view of microbes, from unicellular, motile and planktonic organisms to surface-attached sessile biofilm communities, coexisting along with many other species [13–24]. In biofilms, the resident microbes are self-organized into 3D, matrix-encapsulated structures, with a differentiated interior consisting of water channels and cavities [14,15,18,20,25,26]. Development of microbial biofilms proceeds through genetically controlled specific stages that typically start with surface attachment and colonization of planktonic cells, followed by sessile growth, differentiation and encapsulation of structures [13–15,17,18,21,27–30] (Figure 1). Attachment of planktonic cells is the most critical step in biofilm formation and is dependent on surface properties of both the substratum and the cells. While most biotic and abiotic surfaces are suitable for microbial colonization, certain patterns such as sharklet, or diamond-shaped nanodenticles of shark, are inhibitory [31]. From the microbial perspective, surface attachment is facilitated by both the environmental cues like nutrient limitation, as well as genetic factors. Species of proteobacteria and cyanobacteria utilize a noncovalent assembly of proteinaceous structure, called pili, for surface attachment [32,33].

Figure 1. Schematic model representing distinct developmental stages of microbial biofilms.

Each stage is associated with specific sets of phenotypic switches, facilitated by tightly regulated changes in gene expression patterns.

Surface attachment triggers changes in gene expression that facilitate the growth of biofilms. For example, a master regulator (CsgD) in Escherichia coli promotes biofilm formation by repressing flagellar synthesis and motility, and inducing cyclic diGMP – a secondary messenger for EPS synthesis – through activation of adrA [34]. While characterization of CsgD is limited to E. coli and Salmonella spp., cyclic diGMP-dependent induction of EPS synthesis and biofilm formation has been found in several Gram-negative species [35]. In Vibrio cholerae, cyclic diGMP, and therefore biofilms and virulence genes, are negatively regulated, but quorum sensing – a phenomenon of high-density, specific intercellular communication in bacteria – is positively regulated by hapR [36]. The inverse correlation between quorum sensing and biofilms in V. cholerae suggests biofilm formation as a colonization strategy, while quorum sensing and subsequent suppression of virulence indicate dispersal activity [36,37]. The quorum sensing in P. aeruginosa, however, appears to positively induce maturation of biofilms [38].

Biofilm development in Gram-positive species is modeled around the extensively studied mechanisms in Bacillus subtilis. The EPS synthesis and matrix-formation genes are under negative and positive regulation of SinR and SinI regulators, respectively [39,40]. Furthermore, biofilm formation is negatively controlled by the regulators that trigger sporulation, spo0A-P, indicating that sporulation and biofilm formation are the mutually exclusive choices made by B. subtilis in response to environmental cues [41].

Reconstructing these studies reveals a fascinating overall picture of the microbial world, in which bacteria switch from a motile planktonic form to sessile, surface-attached biofilms by reprogramming the patterns of gene expression (Figure 1). While motility in the planktonic state could be imagined to facilitate the search of individual cells for new niches, development of robust architecture in biofilms and matrix-encapsulation of cells presumably contribute to the phenotypic tolerance against environmental stress, including antibiotics [17,18,20,24,42–47]. The tolerance is believed to be a cumulative effect of the physical protection from the matrix and the physiological adaptation in the resident population in response to the nutrients and oxygen gradients that generate heterogeneity in the microenvironments [13,42–44,48]. The latter process can lead to extensive phenotypic diversity in the population, as elegantly demonstrated by several transcriptomics and genetic studies of B. subtilis biofilms [23,49]. The expression patterns of genes involved in motility, matrix production and sporulation in B. subtilis biofilms are associated with nonoverlapping subpopulations, which are spatially and temporally separated from each other in the multicellular structure [23].

Biofilms: the predominant manifestation of chronic microbial infection

Biofilms are the common clinical manifestation of pathogenic colonization in host tissues and implant devices [16,20,24,25,42,50–53]. Besides the biofilms of P. aeruginosa in cystic fibrosis, other well-documented examples of pathogenic biofilms include: polymicrobial growth in dental plaques and otitis media, mono-infections of Staphylococcus and Streptococcus spp. in implants, pneumonia, osteomyelitis, infective endocarditis, meningitis, dermatitis and in nasal passages, as well as uropathogenic E. coli infection in the urinary tract [12,51,52,54,55].

Regardless of the causative pathogen, infectious biofilms display extraordinary tolerance to antibiotics and subversion to the host immune system [20,24,56,57]. For example, a matrix component of P. aeruginosa biofilms, rhamnolipid, triggers efficient lysis of neutrophils [58]. While the process directly protects the population from the antimicrobial activities of the neutrophils, it also indirectly creates a protected niche by causing necrosis of the tissues through the release of peroxides and proteases from the lysed cells. Modulation of the immune response by biofilms is further evident in the skewed Th1/Th2 response during infections of P. aeruginosa and S. aureus. The early infections of these extracellular pathogens produce a strong Th1-dependent cell-mediated adaptive immune response that is effective against the intracellular pathogen, whereas the later stages of infection produce Th2-mediated humoral immunity [59–62]. The antibodies produced at high bacterial burden are perhaps inadequate at clearing the infection. Although the precise molecular components responsible for the skewed Th1/Th2 response remain unidentified, factors influencing the development of biofilms, such as quorum sensing, have been implicated in the host response [57]. From a broader perspective, the evidence of misdirection and evasion of immune response by biofilms are generally consistent with their subclinical pathology.

Treatment of biofilm infections is extremely difficult. The planktonic cultures of clinical isolates of Staphylococci have been found to be approximately 20–50 times more sensitive to antibiotics than their biofilms [63]. Similarly, biofilms of pathogenic E. coli and P. aeruginosa are 100–1000 times more tolerant to all tested antibiotics than their planktonic counterpart [64]. The drug tolerance in biofilms, acquired through the mechanisms described in the previous section, necessitates a very aggressive prophylactic and therapeutic regimen [56]. Although several preventive measures, such as the coating of implant devices with antimicrobial agents such as metal chelators, are proposed as effective control measures against biofilms, there is no clear therapeutic strategy as yet that involves dispersal of biofilms prior to treatment [65]. However, in one of the most promising developments in dispersal strategies of biofilms, Losick and colleagues recently discovered that biofilms of Gram-positive and Gram-negative species could be efficiently dispersed by nanomolar concentrations of D-amino-acids, although their clinical efficacies in conjunction with other antibiotics remain untested [66].

In summary, biofilm formation is a highly effective and ubiquitous strategy for the pathogen to proliferate as a stress-tolerant community in protected host niches, with limited invasion from the immune system. Thus, biofilm infections can potentially pose significant diagnostic and therapeutic challenges in clinical settings.

Do infections of M. tuberculosis display characteristics of biofilms?

Antibiotic recalcitrance & chronicity of M. tuberculosis infection: the common features associated with pathogenic biofilm infections

Infections of M. tuberculosis are often asymptomatic, chronic in a clinically symptomatic state and highly recalcitrant to antibiotics. Although molecular mechanisms underlying persistence of the pathogen against the host immune system and chemotherapy remain unclear, these clinical features bear similarities with the characteristics of infections associated with microbial biofilms; thereby raising the question as to whether or not M. tuberculosis forms biofilms in vivo, and whether biofilm formation contributes to their persistence. While reminiscence of M. tuberculosis growth in large multicellular clusters has been previously described in histopathological studies of infected lungs, it is unclear if these structures represent a genetically programmed growth pattern of persistent biofilms [67]. However, preliminary evidence suggests that biofilms could perhaps be an in vivo lifestyle of M. tuberculosis, contributing to their persistence against antibiotics. In a pathological study by Lenaerts et al., the surviving population of the pathogen after drug treatment of infected guinea pigs was found to be in microcolonies of bacteria located around the acellular rim in the granulomas [68]. Furthermore, the discovery of a M. tuberculosis-encoded pilin-like protein, which is not only expressed in vivo but also binds strongly to the eukaryotic extracellular matrix, supports the notion that the pathogen’s surface could be actively engaged in surface attachment [69]. The idea of M. tuberculosis persistence in biofilms is supported by the recent findings that portray a highly complex and dynamic state of host–pathogen interface in both acute and latent TB [70–77]. Regardless of the phase of TB infection in nonhuman primates, or murine models, replication of bacilli and engagement of immune cells most likely remain active within a localized host–pathogen interface [70,74,76,77]. Furthermore, the TB lesions representing the host–pathogen interfaces are pathologically heterogeneous, distributed over a wide range of sizes and morphologies in a single individual, irrespective of the clinical symptoms of infection [71,78].

The implications of these findings are that in any given infection of M. tuberculosis, the pathogen thrives in diverse micro-environments, possibly in a wide range of morphological and physiological states that could include biofilms. The question then arises as to which of these states can give rise to drug-tolerant persisters. The drug-tolerant persisters of M. tuberculosis are widely understood to be a small subpopulation of either nonreplicating or slow-replicating variants, presumably emerging through the adaptive processes in a bacteriostatic environment of calcified or closed lesions [79,80]. Persistence of M. tuberculosis against antibiotics in a nonreplicative state has been inferred from the microscopic detection of acid-fast bacilli in culture-negative smears of drug-treated patients [81–83]. This was later strengthened by the Cornell model of TB reactivation, in which the disease relapsed at higher frequencies when drug treatments in infected mice were terminated soon after the sterilization of culturable bacilli [84,85]. Furthermore, low concentrations of oxygen were found in lesions associated with acid-fast bacilli in drug-treated infections, leading to a notion that hypoxia could likely be the primary inducer of the nonreplicating physiology in M. tuberculosis [86–88]. Wayne and colleagues demonstrated the emergence of nonreplicating persisters during depletion of oxygen from in vitro M. tuberculosis cultures, which has been extensively investigated for molecular insight into mycobacterial adaption to limiting oxygen and its influence on development of drug-tolerance physiology [89–91]. Besides physiological adaptation to hypoxia, several additional cellular and physiological factors have also been attributed to the intrinsic antibiotic tolerance in mycobacteria. These include: restricted permeability of a thick mycobacterial envelope, metabolic plasticity under nonhypoxic stress and asymmetric growth pattern [78,92–95]. Thus it is plausible that the drug-tolerant sub-population in any single TB patient could emerge through multiple mechanisms based on the growth and adaptation of the pathogen to its immediate microenvironment, including those within the core of its large multicellular aggregates.

Mycobacterial species form biofilms in vitro

Multiple mycobacterial species, most notably Mycobacterium avium, have been found to exist as multicellular communities in the environment, as well as in clinical settings [96–99]. The prevalence of these mycobacterial communities in their natural habitat can be further appreciated by evidence that the aggregates and pellicles of mycobacteria, routinely observed in detergent-free in vitro cultures, represent a genetically programmed development of organized, drug-tolerant communities – the key features of biofilms. Kolter and colleagues first reported that surface attachment, sliding motility and biofilm formation in a nonpathogenic mycobacteria, Mycobacterium smegmatis, require an acetylated derivative of glycopeptidolipids (GPL) [100]. While the mutant of GPL acetyl-transferase produced defective biofilms, it had no growth defect in planktonic form [100]. GPL biosynthesis was also induced during multicellular growth of M. avium, suggesting that the two species share the mechanisms for biofilm development [101]. Besides GPL, mycolyl-diacylglycerol (MDAG) and free mycolic acids (FM) are the other two surface molecules known so far that have an important role in biofilm development of M. smegmatis [27,102,103]. Thus, lipids could likely have critical roles in intercellular and cell-to-substratum interactions in mycobacterial biofilms. While MDAG synthesis is regulated by a nucleoid-associated protein, Lsr2, FM synthesis is induced during the maturation of biofilms through a GroEL1-dependent modulation of type II fatty acid synthases [27,102–104]. The interaction between GroEL1 and β-keto-acyl ACP synthases (KasA and KasB), the enzymes of FASII, is specifically induced during the later stages of biofilm formation, indicating that biofilm development in M. smegmatis proceeds through distinct stages not associated with planktonic growth [104]. The development of M. smegmatis biofilms also requires a greater abundance of intracellular iron, which is facilitated by induced activity of siderophore synthesis [105]. Interestingly, a strict dependence on iron availability for biofilm formation is also found in P. aeruginosa [106].

In vitro growth of M. tuberculosis in detergent-free media also proceeds through genetically controlled developmental stages, ultimately maturing into FM-rich biofilms on the air–media interface (Figure 2A & 2B) [102]. There are at least three genetic loci, pks16, helY and pks1, implicated in the process of M. tuberculosis biofilm formation [102,107]. Mutants of all three genes fail to produce matured biofilms while growing indistinguishably from the wild-type in planktonic form. Intriguingly, biofilm formation in M. tuberculosis is also sensitive to the gaseous environment on the air–media interface, consistent with the idea that a distinct gaseous composition could induce intercellular or cell-surface interactions in slow growing mycobacteria [102].

Figure 2. In vitro biofilms of Mycobacterium tuberculosis are drug tolerant.

(A) Biofilms of Mycobacterium tuberculosis on the air–media interface grown in a 12-well plate. (B) Biofilms of green fluorescent protein expressing M. tuberculosis on a polycarbonate surface, showing distinct microcolonies. (C) Presence of rifampicin-tolerant persisters in M. tuberculosis biofilms formed on air–media interface. The data from the original article are reproduced in accordance with the policy of the journal publishing the original article.

bf: Biofilm; plnk: Planktonic; Rif: Rifampicin.

Adapted from [102].

Nontuberculous mycobacteria colonize as biofilms

At least two nontuberculous mycobacterial species, Mycobacterium ulcerans and M. avium, have been reported to colonize in the host as multicellular communities [108,109]. Infection of an aquatic insect, Naucoris cimicoides, by M. ulcerans produces a multicellular structure of the pathogen encapsulated by an extracellular matrix [108]. Interestingly, the matrix of M. ulcerans multicellular structures are laden with the toxin mycolactone, which is required for the colonization and virulence of the pathogen [108,110]. Thus, formation of the extracellular matrix, the hallmark of biofilms, by M. ulcerans has a direct influence on its virulence properties. Implication of biofilms in M. avium infection is demonstrated by the inability of the biofilm-defective mutant strain to invade and translocate the bronchial epithelial cells [109].

Mycobacterial growth in biofilms: a relevant in vitro growth model

While the questions as to how, when and where M. tuberculosis forms biofilms in vivo remain open to investigation, it is clear that the pathogen, along with other mycobacteria form these structures in vitro through dedicated genetic pathways. This provides a compelling argument that the in vitro biofilms represent a spontaneous growth characteristic and therefore should be adopted as a growth model for basic mycobacteriology. Moreover, growth of mycobacteria in planktonic form, usually in the presence of detergent, grossly misrepresents their natural behavior and metabolic activities, which can only be revealed by biofilms. Besides the dedicated genetic pathways for structural development, mycobacterial biofilms also acquire unique phenotypes not associated with planktonic forms. The biofilms of M. tuberculosis and M. smegmatis harbor a subpopulation of drug tolerant persisters at several orders of magnitude higher than both defective biofilms of the mutants, as well as the planktonic cultures of the wild-type (Figure 2C) [102,103,111]. Interestingly, presence of antibiotics has been found to influence phenotypic heterogeneity in the biofilms of M. avium [112]. Growth of M. smegmatis in biofilms has been found to be critical for the conjugal transfer of DNA, suggesting that cell-to-cell contact in biofilms could provide a unique opportunity for intercellular transfer of genetic material [113].

The distinction between phenotypes of planktonic and biofilm mycobacteria is further represented by the fact that 82 genes are exclusively induced during maturation of M. smegmatis biofilms (Table 1) [105]. Based on sequence homology, these genes belong to diverse functional categories, including transport of small molecules across the envelope, DNA replication and repair, adaptation to carbon and oxidative stresses, as well as lipid biosynthesis and envelope remodeling (Table 1). The overall pattern of induction is consistent with the idea that the microenvironments of mycobacterial biofilms pose unique challenges to which the constituent cells must adapt. These adaptive processes could possibly be the origin of their drug tolerance behavior. For example, bacterial multidrug efflux pumps have been implicated in physiological export of lipids, raising a possibility that the biofilm-induced transporters could also confer drug tolerance behavior to the bacilli [114]. Induction in fatty acid biosynthesis pathways, perhaps to facilitate the synthesis of biofilm-associated waxy extracellular components such as MDAG and FM, could reduce the permeability of the envelope and induce the drug tolerance properties. Interestingly, Lsr-2, the regulator of MDAG synthesis in M. smegmatis, has also been implicated in regulating the expression of multidrug-tolerance operon, iniABC, in M. tuberculosis [115].

Table 1.

Genes exclusively induced during later developmental stages (4-day) of Mycobacterium smegmatis biofilms.

Locus tag	Annotation	Mycobacterium tuberculosis homologs
MSMEG_0011	Iron utilization protein, putative	Rv2895c
MSMEG_0013	Ferricexochelin uptake
MSMEG_0015	Ferricexochelin uptake
MSMEG_0020	FxuD protein	Rv0265c
MSMEG_0225	Membrane protein, MmpL family	Rv2339
MSMEG_0313	6-phosphogluconate dehydratase	Rv0189c
MSMEG_0316	Transcription regulator ROK family VC2007 (imported), putative
MSMEG_0422	Carboxyphosphonoenolpyruvatephosphonomutase-like protein	Rv1998c
MSMEG_0457	DNA gyrase subunit B, putative	Rv0005
MSMEG_0504	Ribokinase	Rv2436
MSMEG_0534	Multidrug-resistance protein, putative
MSMEG_0550	ABC transporter, periplasmic substrate-binding protein
MSMEG_0782	Aminotransferase, class III	Rv2598
MSMEG_0790	MUTT3	Rv0413
MSMEG_0911	Isocitratelyase	Rv0467
MSMEG_1123	Putativecobalamin synthesis protein
MSMEG_1129	D-amino acid dehydrogenase, small subunit, putative
MSMEG_1268	MutT-nudix family protein	Rv1593c
MSMEG_1269	Serine-threonine protein phosphatase
MSMEG_1332	Conserved hypothetical protein	Rv0633c
MSMEG_1336	Flavohemoprotein, putative	Rv3571
MSMEG_1419	Conserved hypothetical protein
MSMEG_1515	Sensor histidine kinase
MSMEG_1742	Oxidoreductase	Rv3554
MSMEG_1743	DESA3	Rv3229c
MSMEG_1757	Lhr	Rv3296
MSMEG_1953	Unnamed protein product	Rv3260c
MSMEG_2188	Putative integral membrane protein
MSMEG_2238	AldA	Rv0768
MSMEG_2268	Thioredoxin-like protein
MSMEG_2269	Conserved hypothetical protein
MSMEG_2293	Conserved hypothetical protein
MSMEG_2311	Predicted protein, putative
MSMEG_2350	Conserved hypothetical protein	Rv3030
MSMEG_2380	Sugar transporter family protein	Rv2994
MSMEG_2601	Protocatechuate 3,4-dioxygenase β subunit
MSMEG_2735	Diaminopimelateepimerase	Rv2726c
MSMEG_2740	LexA repressor	Rv2720
MSMEG_2743	Conserved hypothetical protein TIGR00244	Rv2718c
MSMEG_2774	GGDEF domain protein
MSMEG_2945	Holliday junction DNA helicase RuvB	Rv2592c
MSMEG_3183	Threoninedehydratase, biosynthetic
MSMEG_3194	Biotin synthase	Rv1589
MSMEG_3297	Adenylatecyclase	Rv1900c
MSMEG_3300	Oxidoreductase, Gfo-Idh-MocA family
MSMEG_3301	Oxidoreductase, putative	Rv0791c
MSMEG_3305	Integral membrane protein domain protein
MSMEG_3568	BFD-like (2Fe–2S) binding domain family
MSMEG_4038	Vanillin dehydrogenase
MSMEG_4057	Transcriptional regulator, GntR family	Rv0165
MSMEG_4383	Membrane protein, MmpL family	Rv0676c
MSMEG_4477	Hydrolase, α-β hydrolase fold family	Rv1900c
MSMEG_4512	Polyketide synthase, putative	Rv2381c
MSMEG_4532	Sulfate ABC transporter, permease protein CysT	Rv2399c
MSMEG_5004	Ser-Thr protein phosphatase family	Rv1277
MSMEG_5048	Conserved hypothetical protein	Rv1249c
MSMEG_5102	ABC transporter domain protein
MSMEG_5130	LpqW	Rv1166
MSMEG_5216	Glyoxalase family protein superfamily
MSMEG_5308	Surface antigen, putative	RV1057
MSMEG_5310	SAM-dependent methyltransferases	Rv2622
MSMEG_5312	ABC transport protein
MSMEG_5556	Major facilitator family transporter
MSMEG_5659	ABC transporter, ATP-binding-permease protein	Rv0194
MSMEG_5680	Glyoxalase family protein	Rv0887c
MSMEG_5716	Conserved hypothetical protein
MSMEG_5732	N5,N10-methylene-tetrahydromethanopterin reductase (mer), putative	Rv3093c
MSMEG_6030	P450 heme-thiolate protein, putative	Rv1777
MSMEG_6060	Predicted permease superfamily
MSMEG_6255	Conserved hypothetical protein	Rv3707c
MSMEG_6476	Putative large secreted protein
MSMEG_6487	Amidohydrolase family superfamily
MSMEG_6540	Virulence factor	Rv0589
MSMEG_6555	Transcriptional regulator, TetR family, putative
MSMEG_6567	COG2837: predicted iron-dependent peroxidase
MSMEG_6569	Uncharacterized BCR
MSMEG_6570	Membrane protein, putative
MSMEG_6618	Superfamily II DNA and RNA helicases	Rv2092c
MSMEG_6619	Conserved hypothetical protein	Rv3096
MSMEG_6742	Conserved hypothetical protein
MSMEG_6783	Putative integral membrane protein
MSMEG_6923	Conserved hypothetical protein	Rv0036c

The list was originally published by Ojha et al. in a transcriptomic analysis of planktonic and biofilm cultures of the nonpathogenic mycobacterial species using microarrays [105]. The Mycobacterium tuberculosis homologs denote the open reading frames with at least 25% identity. The data from the original article are reproduced in accordance with the policy of the journal publishing the original article.

Data taken from [105].

Adaptation to oxidative stress in biofilms, as indicated by the induction of multiple oxidoreductases, LexA and thioredoxin-like proteins, could also influence the tolerance to antibiotics that generate high concentrations of intracellular free radicals (Table 1). This is supported by a correlation between isoniazid tolerance and induced biosynthesis of mycothiol – a predominant reducing agent that maintains redox homeostasis [116].

In conclusion, in vitro biofilms represent a complex but spontaneously self-assembled multicellular architecture of mycobacteria, with a treasury of unexplored knowledge about the mechanisms that shape their stress tolerance behavior.

Expert commentary & five-year view

The in vitro growth of M. tuberculosis in biofilms, together with the evidence of in vivo biofilms of nontuberculous mycobacteria, opens up the question as to whether biofilm formation could be a survival strategy of M. tuberculosis in chronic infection. Moreover, this question is timely with the recent emerging shift in pathological and bacteriological understanding of chronic TB. In the coming years, it will most likely become clear if the multicellular clusters of M. tuberculosis in infected lungs indeed represent the drug tolerant biofilms, held together by an extracellular matrix. It is noteworthy that such in vivo studies would heavily rely upon the basic understanding of intercellular interactions and matrix synthesis of the pathogen in cultures in vitro. The fundamental mechanisms of surface attachment, intercellular signaling and matrix production, identif ied in vitro, can be argued to have reasonable predictability under in vivo conditions, regardless of the differences between the two growth environments. Because the development of drug-tolerant persisters is intricately linked to the physical integrity of the matured biofilms, factors influencing the formation or dissociation of biofilms can serve as potential targets for faster clearance of mycobacterial infections [102].

Key issues.

• Mycobacterium tuberculosis infections are uniquely recalcitrant to antibiotics and are sterilized with 6–9 months of multidrug chemotherapy.

• Although extended therapy is necessary to eliminate a minor subpopulation of drug-tolerant persisters, it is not clear how and where these persisters develop in the host.

• Most microbial species naturally persist through a sessile growth mechanism that leads to the formation of robust, multicellular communities called biofilms.

• Mycobacterial species grown in vitro produce biofilms developed through dedicated genetic pathways, held together by waxy extracellular materials, and harbor drug-tolerant persisters.

• Histopathological studies from the autopsies of tuberculosis lesions have revealed multicellular clusters of M. tuberculosis.

• This raises the possibility that the clustered growth of mycobacteria could represent biofilms harboring drug-tolerant persisters.