The global prevalence of mycobacterial diseases of all types has increased considerably. The most significant mycobacterial disease is tuberculosis (TB). In 2018, 1.4 million deaths were attributed to Mtb infection (World Health Organisation, 2020), meaning TB is one of the top 10 causes of death and the leading cause of worldwide death from a single infectious agent. One-quarter of the world’s population is infected with Mtb. Approximately 5 to 10% of infections progress to the active disease at some point in their host lives. Mtb is a successful pathogen due to its capacity to evade the host immune systems and utilize phagocytes as a replication niche. The bacteria can significantly inhibit the phagolysosome’s acidification and limit phagosome maturation, thereby forcing the infected cell to undergo programmed cell death (Russell et al., 2002). Several of the seminal observations regarding the antimicrobial role of autophagy have been made using Mtb (Gutierrez et al., 2004; Castillo et al., 2012) and such observations were followed by a gradual increase in studies on autophagy as a cell-autonomous, pharmacologically, physiologically and immunologically inducible anti-mycobacterial process (Gutierrez et al., 2004; Seto et al., 2012; Watson et al., 2012; Sakowski et al., 2015). These studies revealed the colocalization of Mtb with autophagosomes and increased bacterial clearance during autophagy induction (Gutierrez et al., 2004; Seto et al., 2012; Watson et al., 2012; Sakowski et al., 2015). Most of these models have been explored in the context of Atg5 and its effect on autophagy. Examination of other important autophagy markers such as Ulk1 and Atg4b in vivo has uncovered that Atg5 may play a role independent of canonical autophagy in Mtb control (Kimmey et al., 2015).

Infection with mycobacteria induces significant levels of pro-inflammatory cytokines, known to be inducers of autophagy. However, little colocalization of Mtb and autophagosomes has been observed without non-mycobacterial stimulation of autophagy. An example of immunological induction of mycobactericidal autophagy includes stimulation of infected macrophages with Th1 cytokines such as IFN-γ and TNF-α in a process that can be antagonized by Th2 cytokines including IL-4 and IL-13 (Harris et al., 2007; Ghadimi et al., 2010). Physiological induction of autophagy by IFN-γ generated significant autophagosome formation in mycobacteria-infected macrophages and dendritic cells. Although Mtb infection causes a robust IFN-γ response, the autophagy induction by virulent Mtb is limited, probably because mycobacteria inhibit IFN-γ mediated autophagy induction (Zullo and Lee, 2012a; Zullo et al., 2014). IFN-γ also plays a critical role in the nitric oxide (NO) response to Mtb infection. The deletion of IFN-γ significantly impedes NO production and leads to uncontrolled replication of bacilli in vivo (Cooper et al., 1993; Flynn et al., 1993). The inhibition of NO has previously been shown to induce autophagy substrates’ clearance, highlighting the complex role of cytokine signaling in autophagy pathways (Sarkar et al., 2011). Virulent Mtb infection induces TNF-α, which is inactivated by the increased release of TNFR2 and results in inhibition of apoptosis (Balcewicz-Sablinska et al., 1998; Keane et al., 2000). This observation underscores the potential of virulent Mtb to inhibit autophagy by modulating cytokines’ bioreactivity known to induce autophagy.

Several bacterial effector proteins are known to modulate autophagy. Many of these effectors are secreted through the type I to type VII and type IX secretion systems (Jiao and Sun, 2019). Mycobacteria has numerous Type VII secretion systems (Esx1 – Esx5). Mtb ESX-1 is responsible for the puncture of the phagosome, allowing for mycobacterial escape (Conrad et al., 2017) (Figure 1B). Mtb cytosolic DNA is recognized by the cytosolic DNA sensor, cyclic GMP-AMP synthase (cGAS), resulting in the release of cyclic guanosine monophosphate (cGAMP). cGAMP is recognized by the stimulator of interferon genes (STING), leading to type I IFN release and the recruitment of autophagy receptors p62, NDP52, and optineurin (Watson et al., 2012, 2015). These receptors are recruited to the ubiquitinated pathogen, thereby allowing for specific targeting by the autophagosome. The receptors contain an LC3 interaction region (LIR) to bind the LC3 autophagy protein (Thurston et al., 2009; Zheng et al., 2009; Wild et al., 2011).

EspB is a part of the Esx1 secretory apparatus responsible for the secretion of early secretory antigenic target-6 (ESAT-6). Treatment of macrophages with EspB protein demonstrates downregulation of the IFN-γ receptor IFN-γR1, resulting in the inhibition of STAT-1 activation even in the presence of IFN-γ (Huang and Bao, 2016). EspB and ESAT-6 are not the only Mtb proteins linked to the inhibition of autophagy. The “enhanced intracellular survival” (eis) gene of Mtb can confer enhanced survival of Mycobacterium smegmatis in macrophages. However, it is not required for the persistence of Mtb in these cells (Wei et al., 2000; Shin et al., 2010). During Mtb infection, eis significantly inhibits the activation of JNK, which prevents the induction of non-canonical autophagy through Atg7. JNK activation also induced reactive oxygen species (ROS) generation and significantly increased type 2 macrophage cell death by Mtb eis deletion mutant (Shin et al., 2010). Eis was also found to substantially inhibit the production of TNF-α, IL-4, and IL-6, while simultaneously stimulating INF-γ and IL-10 secretion (Lella and Sharma, 2007; Samuel et al., 2007; Shin et al., 2010).

Mtb inhibits autophagy to protect against bacterial clearance and host cell death, which also impedes antigen presentation. The Mtb PE_PGRS47 protein inhibits autophagy and limits MHC class II antigen presentation (Saini et al., 2016). Several other Mtb PE/PPE proteins are also known to inhibit autophagy. For example, Mtb PE_PGRS41 (Deng et al., 2017), Mycobacterium marinum MMAR_0242 (Singh et al., 2016), and Mtb PE_PGRS29 (Chai et al., 2019) interact with autophagy machinery. Mtb also secrets a probable ligase (CpsA) to inhibit the non-canonical autophagy pathway designated as LC3-associated phagocytosis (LAP) and NADPH oxidase (Köster et al., 2017) (Figure 1A). In contrast to canonical autophagosomes, LAP does not result in double-membrane structures and instead promotes rapid phagosome maturation (Fazeli and Wehman, 2017). This cellular process limits the phagocytosed pathogen’s ability to replicate by expediting phagosome maturation while regulating the IFN pathway and antigen presentation.

Additionally, studies have found that virulent Mtb, but not avirulent Mtb, can inhibit autophagy flux in macrophages and dendritic cells in an ESAT-6 and PhoP dependent manner (Chandra et al., 2015). Autophagy flux is an important cellular mechanism that degrades autophagosome cargo, which allows nutrient recycling or antigen presentation. Increased autophagic flux was found to improve bacterial clearance from macrophages and dendritic cells (Romagnoli et al., 2012; Chandra et al., 2015). The maturation of Mtb-containing autophagosomes into autolysosomes was inhibited by blocking recruitment of the late endosome marker Rab7 (Chandra et al., 2015). Inhibition of Rab5 conversion to Rab7 in endosomes is a well-established method in which mycobacteria inhibit lysosomal fusion (Via et al., 1997; Rink et al., 2005).

Mtb inhibits canonical and non-canonical autophagy by several means and it is apparent that the role of infection-induced autophagy is complicated. While overcoming autophagy inhibition by Mtb could lead to better treatment options, further consideration should be given to evidence suggesting that Mtb can inhibit autolysosome formation (Chandra et al., 2015). Additionally, there may be value in examining host-directed therapies targeting mTOR-independent autophagy pathways (Schiebler et al., 2015), since Mtb infection markedly activates mTOR. Exploring alternative autophagy-inducing pathways may lead to more efficacious drugs and may prove more useful in patients presenting with co-disease such as TB/Diabetes.

NTM is the broad term for diseases caused by over 170 mycobacteria. The most commonly isolated specie is the Mycobacterium avium Complex (MAC), accounting for 71.1% of Australian cases and 31.3% of NTM cases in South America. M. avium is the most common species isolated in Europe, Asia, South America, and North America. At the same time, M. intracellulare is prevalent in South Africa and Australia (Johnson and Odell, 2014; Griffith, 2019; Gopalaswamy et al., 2020). M. avium, MAC, and M. intracellulare most commonly present as a pulmonary disease similar to TB. In 2013, a nearly six-fold increase in NTM cases were reported in America compared to the 1980s (Donohue and Wymer, 2016), with similar trends in the United Kingdom, Denmark, and Germany (Ratnatunga et al., 2020). Some studies have attributed this rise of NTM infections to a vaccination policy change from a blanket BCG vaccination to a limited vaccination only for specific groups (SAGE Working Group of BCG Vaccines and WHO Secretariat, 2017; Kontturi et al., 2018).

Other common NTMs include Mycobacterium fortuitum, Mycobacterium kansasii, Mycobacterium abscessus, M. marinum, and Mycobacterium ulcerans. NTMs are opportunistic environmental pathogens that are typically found in soil and water. Although NTM disease presentation is most commonly pulmonary, observation of lymphatic, skin/soft tissue and disseminated disease have been reported (Griffith et al., 2007; Bodle et al., 2008; Tortoli, 2009). Many species of mycobacteria can also cause ulcerative disease. Four main presentations of the mycobacterial ulcerative disease have been designated: (i) cutaneous Mtb infection, (ii) leprosy (the second most common mycobacterial disease; caused by M. leprae or Mycobacterium lepromatosis), (iii) Buruli Ulcer, the third most common mycobacterial disease (caused by M. ulcerans), and (iv) opportunistic infections caused by other non-tuberculosis mycobacteria such as M. marinum. Franco-Paredes et al. (2018) eloquently summarized the disease presentation of cutaneous mycobacterial infections in their 2018 review.

Besides Mtb, M. leprae and M. ulcerans account for the next highest mycobacterial disease burdens. Buruli Ulcer, caused by M. ulcerans, primarily occurs in the West and Central Africa, Asia, South America, the western Pacific, and Australasia (Simpson et al., 2019). Unlike the well-studied Mtb and M. leprae, the mode of transmission of M. ulcerans remains unknown (Röltgen and Pluschke, 2015). As with other mycobacteria, treatment of Buruli Ulcer is costly and takes a long time. Traditional antimycobacterial antibiotics are used for treatment, including rifampicin, streptomycin, clarithromycin, and moxifloxacin. However, wound interventions, such as lymphedema management and surgery, are commonly used to speed up healing (Yotsu et al., 2018; Converse et al., 2019; World Health Organisation, 2019). Though leprosy transmission remains on the decline with less than 200,000 cases in 2017 (World Health Organization, 2016), improved treatment options are a vital resource for continued disease decline (Fischer, 2017; Maymone et al., 2020; Scollard, 2020). Leprosy broadly presents two different clinical manifestations; paucibacillary tuberculoid, which is characterized by negative smears for acid-fast bacilli, and multibacillary lepromatous, which is characterized by positive smears for acid-fast bacilli (Nath, 2016).

Mycobacteria’s unique cell wall and some species’ ability to form biofilms, spread by aerosolization, slow growth, and intrinsic antibiotic resistance, also contribute to their ability to survive in unique and low nutrient environments (De Groote and Huitt, 2006). Their lipid-rich cell wall influences the bacteria’s ability to modulate autophagy (Zullo and Lee, 2012a). The ability to form biofilms and survive in low nutrient environments indicates that these bacteria can form unique replication niches within the hosts’ cells that traditional mycobacterial drugs cannot penetrate to be effective (Islam et al., 2012).

The induction of autophagy by mycobacteria is species-dependent. Although all mycobacteria elicit strong mTOR activation, most non-pathogenic mycobacteria simultaneously induce significant autophagy, unlike their pathogenic relatives (Zullo and Lee, 2012a). M. smegmatis is often utilized as a model organism to study pathogenic mycobacteria due to its short culture time and BSL2 classification (Deng et al., 2017). While a low concentration of mTOR-inhibiting drugs like Rapamycin and Torin are able to inhibit mTOR activation and induce autophagy during mycobacterial infection (Zullo et al., 2014), clearance of M. smegmatis requires up to 10 times higher quantity of those drugs than needed to inhibit mTOR activation. Interestingly, this killing was observed to be independent of LC3B or Atg5, indicating a non-canonical autophagy pathway is involved in the clearance of M. smegmatis from macrophages (Zullo et al., 2014). This interesting observation suggests that targeting a non-canonical autophagy pathway for mycobacterial treatment may be useful. It has previously been shown that treatment of Mtb infected macrophages with potent autophagy inducers such as M. smegmatis can clear bacteria (Singh et al., 2017).

The role of autophagy during NTM has not been studied extensively. However, evidence exists that genetic variants in the autophagy-related genes, nucleotide-binding oligomerization domain-containing 2 (NOD2), E3 ubiquitin-protein ligase parkin (PARK2), IRGM, and autophagy-related proteins 16-1 (ATG16L1), are associated with susceptibility to mycobacterial disease (Yang et al., 2014; Capela et al., 2016; Uaska Sartori Priscila et al., 2020). A single nucleotide polymorphism (SNP) in PARK2 correlates significantly with increased susceptibility to M. ulcerans infection, while an SNP in NOD2 is associated with increased disease progression. Conversely, an SNP in ATG16L1 protects against severe disease during M. ulcerans infection (Capela et al., 2016; Manry et al., 2020). Although not directly associated with autophagy, other SNPs in iNOS and IFN-γ have been associated with increased susceptibility to Buruli Ulcer, leprosy, and TB (Bibert et al., 2017).

The major virulence factor of M. ulcerans is mycolactone, a cytotoxic, immunosuppressive polyketide-derived macrolide. Mycolactone alone induces autophagy, although it impairs autophagy flux (Gama et al., 2014). The induction of autophagy is further evidenced by mycolactone’s ability to inhibit mTOR, thereby resulting in the upregulation of apoptosis activating protein, Bim (Bieri et al., 2017). This pathway signals through the inactivation of Akt by an alternative mTOR pathway. As such, activation of mTOR could lead to inhibition of Bim and, subsequently, apoptosis, resulting in control of bacterial infection.

Two variants of M. abscessus and M. fortuitum are frequently observed: rough (R) and smooth (S) (Byrd and Lyons, 1999; Catherinot et al., 2007; Lee et al., 2016). It is widely accepted that the R variant is hypervirulent compared to its S counterpart. It is known that the loss of glycopeptidolipid (GPL) is the cause of the S-variant of M. abscessus in several animal models (Byrd and Lyons, 1999; Catherinot et al., 2007). A highly virulent clinical isolate of M. abscessus-R significantly inhibited autophagic flux than the S variant of M. abscessus. The R variant’s intracellular survival is enhanced considerably by blocking the autophagosome-lysosome fusion in macrophages compared to the S variant (Kim et al., 2017c). These immunological effects of NTMs have been mostly studied from the perspective of respiratory illness and genome comparison studies focusing on traditional virulence factors for related opportunistic pathogens (N’Goma et al., 2015) indicates we have only a minimal understanding of their impact during ulcerative infection.

Whereas M. abscessus S utilizes phosphatidyl-myoinositol to mask TLR2 activation, M. fortuitum R does not induce the anti-inflammatory molecule TNFAIP3 (Lee et al., 2016). TNFAIP3 is an anti-apoptotic molecule that inhibits NF-κB and TNF-induced cell death (Lee et al., 2000). TNF- α and the TLR2 signaling pathway appear to play an essential role in M. fortuitum infection. Some lipids of RGMs have differential terminal modifications compared to those from pathogenic slow-growing mycobacteria. Specifically, lipoarabinomannan (LAM) in RGM is capped with phosphomyo-inositol (PI) caps compared to mannose (Man) caps in pathogenic mycobacteria. Purified PI-LAM induces significantly more apoptosis than purified Man-LAM in a TLR2 dependent manner (Bohsali et al., 2010). Similarly, PILAM caused significant autophagy induction, unlike ManLAM, which did not induce autophagy (Shui et al., 2011; Singh et al., 2019). Although terminal modifications of LAM appear to play a role in the modulation of apoptosis, total lipid from both pathogenic and non-pathogenic mycobacteria can induce autophagy (Zullo et al., 2014; Kim et al., 2017c; Mishra et al., 2019). Interestingly, while total lipids from M. abscessus-R induce a significant autophagy level, live M. marinum induces autophagy and simultaneously inhibits autophagy flux, which leads to increased intracellular survival (Lerena and Colombo, 2011; Kim et al., 2017c; Oliveira et al., 2020; Pohl et al., 2020).

It has been known for many years that autophagy is an efficient mechanism to clear M. leprae from macrophages (Evans and Levy, 1972). However, it has been recently described that autophagy may be a major modulating factor in leprosy disease presentation. In patients presenting with multibacillary leprosy, there is significantly less autophagic control in macrophages taken from patient lesions than patients presenting with paucibacillary tuberculoid leprosy (Silva et al., 2017a). This supports previous studies which found that the autophagy inhibiting cytokine IL-10 is predominant in multibacillary leprosy compared to high levels of IL-26, IFN-γ, and TNF-α, autophagy inducing cytokines, found during paucibacillary tuberculoid leprosy (Yamamura, 1992; Sieling and Modlin, 1994; Nath, 2016; Dang et al., 2019). Multibacillary leprosy patients who developed type 1 reaction (T1R) episodes demonstrated dysregulation of autophagy genes and significantly increased expression of the mTOR complex leading to overexpression of the NLRP3-inflammasome-IL-1B pathway. These data demonstrate that leprosy treatment with pro-autophagic drugs may improve treatment outcomes by reducing reversal reaction risk (de Mattos Barbosa et al., 2018).

The establishment of uncontrolled mycobacterial infection in an extracellular bacterial milieu or biofilm presents significant complications for treatment (Greendyke and Byrd, 2008). Many mycobacterial species causing ulcerative diseases are widely considered to have significantly reduced sensitivity to antibiotics and a natural ability to acquire antibiotic resistance, making it very hard to treat and leading to high failure rates (Moore and Frerichs, 1953; Jarlier and Nikaido, 1990; Sanguinetti et al., 2001; Nessar et al., 2012). Utilizing host-directed therapies, such as those inducing autophagy, to inhibit bacterial release from the cell and form biofilms or bacterial milieus may enhance the efficacy of currently available antibiotics.

Rapamycin has been the most frequently used autophagy-inducing drug for host-directed therapies. While Rapamycin appears to improve pathology during Mtb infection, there is evidence that it is directly antimycobacterial in vitro at the high concentration used for the reported studies. Rapamycin does not seem to have a direct effect on M. smegmatis or BCG for short periods. Still, it was found to significantly inhibit BCG, M. kansasii, M. avium, and multiple virulent Mtb strains over 7–8 days incubation (Greenstein et al., 2008; Zullo et al., 2014). This direct antimycobacterial activity is somewhat unsurprising as Rapamycin was initially discovered as a novel antifungal antibiotic (Singh et al., 1979). Rapamycin is not the only drug evaluated as a host-directed therapy for the treatment of tuberculosis. Some medications, such as azithromycin and metformin, have been found to decrease mycobacterial infections in patients with cystic fibrosis and diabetes due to their ability to increase the autophagic clearance of bacteria (Renna et al., 2011; Tseng, 2018). Table 1 summarizes the drugs and compounds that have been tested for their ability to induce autophagy and treat mycobacterial diseases.

Of importance, the use of Rapamycin to treat infectious diseases is not practical due to its immunosuppressant actions. Gupta et al. (2014) have attempted to address this issue with the administration of Rapamycin by microparticles directly to the airway. Highlighting the delicate balance needed for host-directed therapies, the study found that the induction of autophagy in the lung macrophages was inverse to the dosing interval. In vitro and in vivo rapamycin microparticles induce autophagolysosomal formation in macrophages infected with Mtb in an mTOR-dependent manner (Gupta et al., 2014; Gupta et al., 2016). Rapamycin alone significantly improved pathology during Mtb infection in a mouse model but did not clear bacteria. Co-administration of isoniazid and rifabutin with Rapamycin microparticles considerably improved bacterial clearance (Gupta et al., 2014, 2016). As the administration of microparticles with autophagy-inducing drugs may improve traditional antimycobacterial chemotherapies, an alternative strategy is utilizing microparticles that directly induce autophagy. Poly (lactic-co-glycolic acid) microparticles were found to be antimycobacterial in human macrophages. For example, NFκB activity was increased during microparticle treatment, and antimycobacterial effects were reversed by autophagy inhibitors (Lawlor et al., 2016).

An antiprotozoal drug, nitazoxanide, has been extensively tested to treat Mtb and NTMs alone and in conjunction with traditional antimycobacterial medicines. These conventional antibiotics were not found to inhibit autophagosome formation stimulated by nitazoxanide (Lam et al., 2012). Nitazoxanide has been previously explored as an autophagy agonist for treating multiple disease states such as Alzheimer’s and cancer (Di Santo and Ehrisman, 2014; Li et al., 2020). It has also been examined as a potential treatment option against several mycobacteria including M. leprae (Bailey et al., 2017) and MAC (Rossignol, 1999). Nitazoxanide is metabolized into hydroxylamine by mycobacterial nitroreductase NfnB (Buchieri et al., 2017). Interestingly, nitazoxanide can kill replicating and non-replicating mycobacteria, emphasizing its potential role in combating latent mycobacterial infection (de Carvalho et al., 2009; Iacobino et al., 2019). Like Rapamycin and nitazoxanide, metformin also increases bacterial clearance during traditional anti-mycobacterial treatment, while inducing autophagy (Singhal et al., 2014; Lachmandas et al., 2019).

Two of the most promising experimental host-directed therapies against M. tuberculosis are Vitamin D3 and Metformin (Naicker et al., 2020). Metformin was shown to increase mitochondrial reactive oxygen species production, acting through AMPK, leading to control of drug-resistant Mtb and facilitation of phagolysosome fusion (Singhal et al., 2014; Yew et al., 2020). There also appears to be a correlation between metformin treatment for diabetes mellitus type II and delayed smear and culture conversion and reduced unfavorable outcomes (Singhal et al., 2014; Degner et al., 2017; Marupuru et al., 2017; Lee et al., 2018; Padmapriyadarsini et al., 2019). While metformin shows promise in preventing TB in type II diabetes patients, Vitamin D supplementation showed no improvement in TB treatment outcomes in patients with vitamin D sufficiency during drug sensitive Mtb infection. However, vitamin D deficiency is associated with an increased risk of Mtb infection (Ustianowski et al., 2005; Chun et al., 2011). Vitamin D supplementation did reduce the time to sputum culture conversion in patients with Taql vitamin D receptor gene polymorphism, indicating that Vitamin D does play an important role in TB treatment outcomes. Vitamin D supplementation also improved the MDR-TB sputum culture conversion rate (Zhang et al., 2019). In vitro treatment with vitamin D during HIV and Mtb co-infection or Mtb infection alone concluded that autophagy induction was responsible for the better control of both HIV and Mtb in macrophages (Yuk et al., 2009; Fabri et al., 2011; Campbell and Spector, 2012a).

Many host pathways may constitute viable targets for host-directed therapies (HDTs). Apoptosis and autophagy have been the most explored HDT targets of Mtb and NTMs. Even though apoptosis may be a possible target, there is mounting evidence that NTMs can escape apoptotic bodies to ensure survival and disease progression (Early et al., 2011; Bento et al., 2020). Autophagy presents an exciting target as the induction of autophagy promotes bacterial clearance and antigen presentation (Castillo et al., 2012; Saini et al., 2016). The current recommended treatment for NTM infection is clarithromycin or azithromycin, ethambutol, and rifamycin (Griffith, 2018; Griffith, 2019; Daley et al., 2020; Gopalaswamy et al., 2020). Azithromycin was shown to inhibit autophagosome maturation resulting in an increased risk of M. abscessus infection (Renna et al., 2011; Torfs et al., 2019). It has also been found that many traditional anti-mycobacterials, though directly antimycobacterial, also have off-target effects that promote autophagy (Kim et al., 2012; Zullo and Lee, 2012b). Unfortunately, the development of new drugs targeting particular host pathways is often slow and expensive. One potential strategy to expedite this drug discovery is studying and assessing previous medications known to increase autophagy and their effect on mycobacteria (Williams et al., 2008; Sundaramurthy et al., 2013; Stanley et al., 2014; Kim et al., 2019). Potentially repurposed autophagy targeting host-directed therapies are summarized in Table 2. Many of these drugs are of interest because of the modulation of the host immune response. Accordingly, they should also be effective against a broad range of mycobacteria and other intracellular pathogens.

As a better understanding of the role of infection-induced autophagy transpires, more targeted host-directed therapy approaches can be developed and exploited. With C₄T₄ (a TLR4 agonist), autophagy was induced in guinea pigs infected with Mtb in a CLEC4E-dependent manner through MYD88 and PrdIns3K activation, leading to reduced mycobacterial burden (Pahari et al., 2020). Along with targeting cell receptors to activate autophagy, there is increasing evidence that many microRNAs (miRNAs) can be targeted to activate autophagy during mycobacterial infection (Wang et al., 2013; Kim et al., 2015, 2017a,b; Kumar et al., 2016; Etna et al., 2018; Liu et al., 2018; Li et al., 2019). These miRNAs modulate autophagy through different upstream pathways of mTOR. Mtb infection induces miRNA-144, which targets the DNA damage regulated autophagy modulator 2 (DRAM2), resulting in autophagy inhibition through AMPK (Kim et al., 2017b). Similar to targeting CLEC4E through TLR4 agonists, TLR2 and MYD88 are required to induce miRNA-125a during Mtb infection. Mtb induces expression of miR-125a in macrophages, which results in the inhibition of autophagy by targeting UV radiation resistance-associated gene (UVRAG) in the AMPK dependent manner (Kim et al., 2015).

BCG is widely used as a vaccine against tuberculosis. BCG evades phagosome maturation, autophagy, MHC-II expression of antigen-presenting cells (APCs), and T-cell activation (Deretic et al., 1997; Singh et al., 2006; Münz, 2016; Saini et al., 2016; Khan et al., 2019). Clinical isolates of Mtb that could not inhibit autophagy showed increased TB disease outcomes and the extent of disease (Li et al., 2016), strongly indicating that autophagy is a crucial host pathway for the control of TB. The ability to utilize this essential host pathway could prove a viable avenue for improving mycobacterial vaccines (Yuk and Jo, 2014; Flores-Valdez et al., 2018; Tao and Drexler, 2020). Notably, the yellow fever vaccine, YF-17D, one of the most successful vaccines, has been found to enhance autophagy-dependent antigen presentation. The mechanisms of the vaccine efficacy by YF-17D were not well understood until its role in autophagy modulation was deciphered (Ravindran et al., 2014).

Developing new vaccines or improving the BCG by harnessing autophagy is an area of interest and has garnered examination. BCG, like Mtb, expresses a wide array of bacterial effectors that modulate autophagy, but co-immunization of mice with BCG and rapamycin-treated dendritic cells enhanced Th1-mediated protection against Mtb infection (Jagannath et al., 2009). Similarly, it was demonstrated that a recombinant BCG expressing 85C5 (BCG^85C5) induced a robust MHC-II-dependent antigen presentation to CD4⁺ T cells in vitro. The 85C5 peptide contains the TLR-2 activating C5 peptide from Mtb CFP-10 protein. The vaccine also elicited stronger Th1 cytokines from APCs of C57Bl/6 mice and enhanced MHC-II surface expression on macrophages by inhibiting the membrane associated RING-CH 1 (MARCH1) E3 ligase that degrades MHC-II. BCG^85C5 infected APCs presented antigens in a MyD88 or a TLR-2 dependent manner (Khan et al., 2019). Additionally, activation of TLR3 or TLR4 by LPS cleared mycobacteria in vitro in an autophagy-dependent way (Xu et al., 2007, 2013).

BCG was genetically modified to improve its immunogenicity by replacing the urease C encoding gene with the listeriolysin encoding gene from Listeria monocytogenes (Nieuwenhuizen et al., 2017). Listeriolysin perturbates the phagosomal membrane at acidic pH and Urease C neutralize the phagosome harboring BCG. Deletion of ureC leads to rapid phagosome acidification and promotes phagolysosome fusion. Subsequently, BCGΔureC:hly elevates apoptosis and autophagy and accelerates release of mycobacterial antigens into the cytosol. The BCGΔureC:hly vaccine completed phase I and IIa clinical trials. Upon deleting the anti-apoptotic gene nuoG to enhance cross protection, BCGΔureC:hlyΔnuoG vaccine showed reduced Mtb burden in the lungs of mice leading to less pathology and, most importantly, enhanced immune responses. It was found that the nuoG deletion leads to significant induction of autophagy and an improved safety profile (Gengenbacher et al., 2016). M. indicus pranii is another potential immunotherapy and vaccine candidate under clinical trials (Gupta et al., 2012; Saqib et al., 2016; Sharma et al., 2017). Boosting BCG vaccination with M. indicus pranii resulted in improved protection in a murine model of Mtb infection. Increased IFN-γ, IL-12, and IL-17 were observed along with increased polyfunctional T cells (Saqib et al., 2016). This increased protection and immune response were subsequently due to increased autophagy induced by M. indicus pranii, potentially due to its PILAM (Singh et al., 2017, 2019).

While improving the BCG vaccine appears a viable short-term solution to improve vaccine efficacy against Mtb, the efforts to develop new vaccines must be continued. BCG is a live attenuated vaccine, but it is not suitable for use in all cases. The development of other vaccine options, such as DNA vaccines or subunit vaccines, would significantly advance the vaccination strategy against Mtb and NTMs. A DNA vaccine encoding for the potent Mtb antigen 85B (Ag85B) significantly enhanced autophagy activation and vaccine efficacy when delivered with a plasmid encoding a kinase defective (mTOR-KD) (Meerak et al., 2013). This mTOR-KD DNA vaccine-elicited considerably higher Ag85B-specific antibodies, increased secretion of IFN-γ and IL-2 levels, and enhanced proliferation of CD4⁺ T cells. Similar to this approach, it was found that an LC3-LpqH-Ag85B DNA vaccination reduced mycobacterial burden, increased IFN-γ and IL-2, and enhanced the Th1 immune response (Hu et al., 2014).

Adjuvating or boosting BCG with autophagy-inducing substrates has also been examined as a potential way to increase the efficacy of current vaccines. Curcumin-coated nanoparticles have been found to enhance autophagy, leading to increased Th1 and Th17 central memory T cells (Ahmad et al., 2019). Other studies have identified autophagy as having an essential role in forming and surviving memory T cells (Xu et al., 2014). Like curcumin nanoparticles, boosting BCG vaccine with nanofibers acting through the autophagy pathway improved BCG efficacy (Rudra et al., 2017; Chesson et al., 2018). While some DNA vaccines directly target autophagy, the development of an adjuvant system targeting autophagy may improve the efficacy of potential subunit vaccines. Utilizing the lactic acid bacteria (LAB) as an adjuvant for Mtb antigens showed improved IFN-γ and NO responses, polarizing a Th1 response and increasing autophagosome formation. Although LAB’s protective efficacy was not tested, these improved immunological responses compared to Mtb antigen alone are promising (Ghadimi et al., 2010).