*4.2. Complexity of TB Granulomas*

Long lasting therapies are also attributable to the complex pathology of TB. In the lungs of patients with active and latent TB, a spectrum of heterogeneous granulomatous lesions coexist, ranging from well-vascularized cellular granulomas, in which a rim of lymphocytes surrounds macrophages and neutrophils, to avascular caseous granulomas, characterized by a necrotic center with a cheese-like aspect (caseum) formed by the lysis of host cells and bacteria [29,30]. In these lesions, tubercle bacilli range from actively replicating (AR) stages, particularly in cellular granulomas, to dormant, slowly-replicating or NR stages, typical of hypoxic caseous granulomas [31]. In Mtb-infected rabbits, the fraction unbound of a drug penetrates the caseum via passive diffusion, and caseum binding of a drug is proportional to hydrophobicity (cLogP) and aromatic ring count [32].

The current 4-drugs therapeutic regimen (RIF-INH-PZA-EMB) is effective against AR intracellular bacilli in cellular granulomas, while NR extracellular bacilli localized in pH-neutral, caseous granulomas are refractory to drug action [17,33–36]. The necrotic center of caseous granulomas contains NR bacilli phenotypically resistant to several drugs (drug-tolerant persisters), with the exception of rifamycins, which are known to sterilize caseum in ex-vivo assays [35,36]. Spatial and temporal differences in drug distribution and the kinetics of accumulation of drugs in specific lesion compartments may create local windows of monotherapy that increase the risk of the emergence of drug-resistance [17,37]. This is in keeping with the knowledge that genetically resistant mutants of Mtb may emerge from the persistence phase of some TB drugs, due to hydroxyl radical-mediated genome-wide random mutagenesis [38–40]. In this view, drug combinations should contain complementary drugs preferentially distributing in lesions in which their most vulnerable target population resides [17].

In the event of caseous granulomas expansion, the necrotic centers fuse with the airway structures of bronchi to form pulmonary cavities in which are found both extracellular bacilli from liquefied caseum and intracellular bacilli derived from the lysis of infected macrophages of the cavity walls. In contact with the atmospheric oxygen, these bacilli rapidly proliferate in the lumen of cavities, and later appear in the sputum of TB patients [17]. Due to high bacterial load in pulmonary cavities, genetically resistant bacilli with chromosomal mutations may be generated, playing an important role in the development of resistance [16]. Noticeably, in comparison with paired sputum isolates, additional resistances were found in Mtb isolates recovered from surgically resected cavities of the same patient [41]. A single founder Mtb strain underwent genetic mutations during treatment, leading to the acquisition of additional drug resistance in different sections of the lung of the same patient, preferentially in the cavity wall [42]. In keeping with this observation, drug-specific gradients in the walls of human pulmonary cavities were reported to be associated with the development of acquired resistance in patients with MDR-TB, due to the low level of some drugs in the cavities centers, where there is a high number of replicating bacilli [43]. In the latter study, spatial heterogeneity of drug concentrations across the pulmonary cavity resulted in the development of mutations in the Mtb genes *gyrA* (FQ resistance) and *gydB* (aminoglycoside resistance), consistent with evolution from MDR- to XDR-TB after about five months of therapy [43]. Overall, these observations indicate that acquired Mtb resistance may be related to the formation of drug-penetration gradients in TB lesions generating suboptimal drug concentrations in non-vascularized caseous granulomas and in liquefied caseum in the cavity centers [16,17,43].

#### *4.3. Intrinsic Drug-Resistance of Mtb*

During the evolution, Mtb developed mechanisms of intrinsic resistance to antibiotics involving cell envelope, efflux systems and other mechanisms (drug degradation and modification, target modification), which allowed the organism to reach high drug resistance levels. Some examples of these mechanisms are provided in the following sections.

## 4.3.1. Cell Envelope

The constituents of the mycobacterial cell envelope are: the cytoplasmic membrane, the periplasmic space (PS), a network of peptidoglycan (PG), the arabinogalactan (AG), the long-chain mycolic acids (MA) and the capsule, made of a loose matrix of glucans and secreted proteins [44]. As to the first-line TB drugs, the bactericidal agent INH inhibits MA synthesis, while the bacteriostatic EMB inhibits AG synthesis and may sensitize Mtb to other drugs [44].

It is assumed that the innermost hydrophilic layers of PG and AG hinder the penetration of hydrophobic molecules. Instead, in the external part of the envelope, the PG and AG layers are linked to the hydrophobic MA layer, formed by long-chain fatty acids that restrict the penetration of hydrophilic drugs [18,20]. In principle, more lipophilic drugs, such as rifamycins, macrolides, and some FQs, diffuse by passive transport into and through the lipid-rich cell wall [45,46]. In early studies, mutants defective in the biosynthesis of cell wall components were very useful to demonstrate the role of the cell wall in the intrinsic resistance to drugs. For instance, a mycolate defective *Mycobacterium smegmatis* mutant showed increased susceptibility to RIF, chloramphenicol (CF), novobiocin and erythromycin [47,48]. Also, insertions in genes involved in the mycolate biosynthesis of Mtb (*mymA* operon) showed enhanced chemical penetration and sensitivity to RIF, INH, PZA and ciprofloxacin [49].

Small hydrophilic drugs traverse the cell wall of bacteria via water-filled porins, without energy consumption. *M. tuberculosis* encodes at least two porin-like proteins (OmpA, encoded by *Rv0899* and *Rv1698*), but the role of porins in Mtb drug uptake and susceptibility needs to be further investigated [18,20,50,51]. Penetration of hydrophilic β-lactam antibiotics through the mycobacterial cell was about 100 times lower than in the *Escherichia coli* cell wall [20]. The β-lactamases, probably in conjunction with slow drug penetration, were shown to be major determinants of Mtb resistance to β-lactams [52,53]. In Mtb, the PG is remodeled by nonclassical l,d-transpeptidases (LDT). The structural basis and the inactivation mechanism of LDT and the active role of carbapenems were investigated, providing a basis for their potential use in inhibiting Mtb [54]. Indeed, the carbapenems IPM-CLN and MPM (both to be used with CLV, available only in formulations combined with amoxicillin) have been listed as add-on drugs in the recent WHO treatment guidelines of MDR/XDR TB [8].

Overall, it is thought that anti-TB drugs have the peculiarity of being more lipophilic than many other antimicrobial agents, likely due to improved penetration through the waxy mycobacterial cell wall [45,46]. However, the issue is perhaps more complex, since some studies showed that lipophilicity is an important but not exclusive factor of compound permeability [50,55].

#### 4.3.2. Drug Efflux

Efflux pumps (EPs) are transmembrane proteins that provide resistance by expelling the drugs from the interior to the exterior of the cell. Five EP families are known, organized on the basis of energetic and structural characteristics: the ATP-binding cassette (ABC) superfamily, the major facilitator superfamily (MFS), the multidrug and toxic compound extrusion (MATE) family, the small multidrug resistance (SMR) family and the resistance nodulation division (RND) superfamily [18,19,46,56,57]. The ATP-energized ABC members are primary transporters, while the others are secondary transporters energized by proton gradients (MFS, SMR, RND) or sodium gradients (MATE). The EP of Mtb belongs to the ABC (representing 2.5% of the entire Mtb genome), MFS and RND superfamilies, and to the SMR family.

Following exposure of Mtb to sub-inhibitory concentration of INH and EMB, EP genes are overexpressed resulting in the development of low-level resistance for a prolonged period of time. After several weeks, a high level of acquired resistance develops, caused by chromosomal mutations in the genes encoding the target proteins [58,59]. These observations indicate that inappropriate TB

treatment may generate pressure by sub-inhibitory drug concentrations that increase drug efflux, allowing a subsequent selection of mutants with high-level resistance [46,57].

Several EPs are known to be associated with resistance. For instance, Mtb exposure to INH induces the overexpression of *MmpL7* and *mmR* EP genes [60,61]. Furthermore, several EPs are involved in resistance to several drugs. Thus, the EP Tap mediates low-level resistance to tetracycline (TC) and aminoglycosides, whereas EPs encoded by the *Rv0194* gene is associated with resistance to β-lactams, SM, TC, CF and vancomycin. Mutations in the *Rv0678* gene caused an up-regulation of the transport protein, MmpL5, which caused EP-mediated cross-resistance to both BDQ and CFZ [60,62,63]. This is a potentially dangerous evolution of Mtb against antibiotics particularly in recent times, since BDQ and CFZ have just been included in the new WHO treatment guidelines of MDR/XDR-TB [8,11].

A strategy used to inhibit efflux-mediated drug resistance is efflux inhibition by non-antibiotic molecules that block the EP or inhibit the EP energy sources [57,64]. The most studied inhibitors are verapamil (VP), thioridazine (TZ), reserpine, piperine, protonophores [57,64], to be used in combination with anti-TB drugs in order to decrease or abolish the drug resistance caused by EP activity. Verapamil, an FDA-approved calcium channel blocker, decreased the MICs of BDQ, CFZ and other drugs [57]. This synergism was confirmed in various studies, but it was found that the effect of VP was not due to intra-mycobacterial drug accumulation, but on the disruption of membrane functions [65]. In Mtb-infected mice, VP increased the bioavailability and efficacy of BQ but not CFZ [66]. EP inhibitors are not presently used for the treatment of human TB, with the exception of TZ, which was administered in compassionate therapy for some XDR-TB cases [67].
