Lysosomal Disorders Drive Susceptibility to Tuberculosis by Compromising Macrophage Migration

1Tuberculosis (TB) involves a series of interactions between macrophages and the infecting mycobacterium with this proposed sequence of events: inhaled mycobacteria are engulfed by lung alveolar macrophages and, if not cleared during this initial interaction, are transported deeper into the lung. Here, newly recruited myeloid and other immune cells aggregate around the infected cells to form organized granulomas.

2The study of zebrafish infected with M. marinum has enabled the dissection of these steps of TB pathogenesis, aided by the genetic tractability of this model organism and its optical transparency during its first few weeks of life. Newly infecting bacteria can be transported across epithelial barriers by permissive macrophages. Additional macrophages are recruited to the initial infected macrophage to form the tuberculous granuloma. Cellular expansion of the granuloma, and intracellular bacterial growth within it, proceeds through apoptosis of the infected macrophages and their phagocytosis by newly arriving uninfected macrophages. On the one hand, bacterially mediated granuloma expansion can promote infection through bacterial spread into newly recruited macrophages. On the other hand, if the supply of uninfected macrophages is limiting, apoptotic infected cells in the granuloma undergo secondary necrosis, causing granuloma breakdown and the release of bacteria into the extracellular space, which enables their accelerated growth.

3In this work, we characterize a zebrafish mutant identified in a forward genetic screen to reveal how, during genetic lysosomal storage disorders, the accumulation of undegraded products in the macrophage lysosome impairs the migration of these phagocytic cells. The disruption of macrophage migration contributes to the pathogenesis of the lysosomal storage disease in the uninfected state and causes granuloma breakdown during tuberculous infection, which underlies hypersusceptibility. The mutation maps to snapc1b, a transcriptional co-regulator that causes lysosomal storage through reduced expression of lysosomal cysteine cathepsins B and L. Using zebrafish models of human lysosomal storage diseases, we generalize our findings to show that the accumulation of diverse biological substrates, as well as inert particles, compromises macrophage migration through the derangement of endocytic recycling. We then show that lysosomal storage in macrophages inhibits their migration to engulf newly infecting bacteria. Because the resident alveolar macrophages of human cigarette smokers have been reported to accumulate particulate material, we asked whether smokers’ macrophages are similarly compromised in their response to mycobacterial infection. We find that the majority of smokers’ alveolar macrophages have enlarged lysosomes filled with opaque material and are impaired in their migration to M. tuberculosis. The compromised migration of these highly microbicidal first responding macrophages suggests a mechanistic explanation for the reported association between smoking and TB.

4The zebrafish mutant fh111, identified in a forward genetic screen, was hypersusceptible to M. marinum, displaying increased bacterial growth relative to wild-type siblings after intravenous infection (Figures 1A and 1B). fh111 infection was characterized by the breakdown of forming granulomas accompanied by bacterial cording, a characteristic morphology acquired by rapidly growing extracellular bacteria after release from necrotic macrophages (Figure 1C). We used bacterial cording as a sensitive and specific phenotype to map fh111 (Figure 1D). fh111 maps to a splice acceptor site mutation in the exon 1–2 junction of the zebrafish snapc1b gene on chromosome 13 (Figure S1A), one of two orthologs of human SNAPC1 (Small Nuclear RNA Activating Complex Polypeptide 1) that encodes a component of the basal transcriptional machinery for RNA Pol II and III-dependent transcription. Zebrafish snapc1b has higher amino acid identity to human SNAPC1 than its paralog snapc1a, located on chromosome 20 (39% versus 35%, respectively). RNA sequencing (RNA-seq) analysis of wild-type (WT) animals at 6 days post-fertilization (dpf) showed that snapc1b RNAs were ∼35-fold more abundant than snapc1a RNAs (GEO: GSE74196). We confirmed the location and transcriptional consequence of fh111 by RNA-seq and qRT-PCR (Figures S1A and S1B). Causality of the fh111 mutation was confirmed by a splice-blocking antisense oligonucleotide (morpholino) that targeted the same exon 1–2 splice junction of snapc1b (Table S1) that phenocopied fh111 susceptibility (Figures S1C and S1D) and by non-complementation with an independent retroviral insertion allele that disrupts exon 1 of snapc1b (la010158) (Figures S1E and S1F).

5In sum, our findings suggest that Snapc1b deficiency causes hypersusceptibility to mycobacterial infection through early granuloma breakdown, which releases mycobacteria into the extracellular milieu that is more growth permissive than the intracellular environment, culminating in bacterial cording morphology.

6Granuloma breakdown can result from a global reduction in macrophage numbers available to replenish the granuloma. We were surprised to find that, even in uninfected snapc1b mutants, macrophage numbers were increased as revealed by increased numbers of fluorescent macrophages in transgenic animals and by staining with neutral red, a vital dye that accumulates in macrophages (Figures 1E–1G). The increased abundance of microglia, tissue resident macrophages of the brain derived from a primitive hematopoietic lineage, suggested a derangement in multiple waves of myelopoiesis (Figures 1H and 1I).

7Most of the macrophages in mutants were enlarged and discoid in shape (Figures 1J, 1K, and S2A–S2C). The enlarged macrophages contained prominent vacuoles, which were revealed to be fused lysosomes by staining with LysoTracker (Figure 1L). The mutant lysosomes were larger than wild-type and occupied a greater proportion of total macrophage volume (Figure 1M). This aberrant macrophage morphology is similar to what has been observed in human lysosomal storage disorders. Our observation of an increased abundance of tissue resident macrophages (histiocytosis) accompanied by increased expression of myeloid growth factors (Table S2) also mirrors findings in several human lysosomal storage disorders. Neutrophils, the other myeloid cells present at this developmental stage, were not increased in snapc1b mutants and displayed normal morphology, consistent with their lack of involvement in homeostatic scavenger functions (Figure S2D; data not shown).

8To understand the development of vacuolated morphology, we compared microglial morphology and dynamics in uninfected animals during physiological efferocytosis of apoptotic neurons. In wild-type animals, macrophages assumed a vacuolated morphology upon phagocytosis of particulate material (presumably cell debris) but reverted to normal within 4 hr (Figure 2A; Movie S1). Vacuolated morphology was accompanied by a transient reduction in speed of homeostatic migration; movement resumed upon reversion to normal morphology (Figures 2A and 2B; Movie S1). In snapc1b mutants, macrophages became irreversibly vacuolated after only a few phagocytic events, accompanied by sustained immotility and consequently reduced displacement—36 μm versus 131 μm for wild-type—over a 13-hr observation period (Figures 2A and 2B; Movie S1). These vacuolated macrophages did not have obvious phagocytic defects and continued to extend pseudopods in apparent phagocytic events (Figure 2A; Movie S1). Vacuolated macrophages were also unable to migrate in response to the chemotactic molecule CCL2 injected into the hindbrain ventricle (HBV), indicating a broad migratory defect toward both cell debris and chemotactic factors (Figure 2C). These findings are consistent with observations that macrophages from patients with Gaucher’s disease, the most common human lysosomal storage disorder, are defective for migration but competent for phagocytosis.

9Acridine orange staining confirmed that the lysosomal accumulations in mutant macrophages were phagocytosed apoptotic bodies (Figure 2D). Together, these findings suggested that lysosomal accumulation of undigested cell debris in mutant macrophages impairs migration irreversibly and leads to progressive macrophage incapacitation. If so, then we hypothesized that mutants would have an accumulation of extracellular apoptotic bodies in the brain as seen in several human lysosomal storage disorders. They did (Figures 2E and 2F). Furthermore, a global reduction in apoptosis induced by the pancaspase inhibitor Q-VD-OPh reduced extracellular cell debris (Figure S3A) and decreased the number of macrophages with lysosomal accumulations of the debris and with enlarged lysosomes (Figures S3B and S3C).

10Together, these findings show that snapc1b mutant macrophages become irreversibly vacuolated due to their inability to degrade phagocytosed cell debris. As more and more macrophages lose their migratory capability, a deficit of functioning macrophages may develop.

11We recently showed that reduction in the macrophage supply reduces granuloma macrophage replenishment to the point at which apoptotic infected macrophages, failing to be engulfed, undergo secondary necrosis. Having observed that the vacuolated macrophages of snapc1b mutants failed to migrate to dying cells in the uninfected state and failed to migrate to newly infecting bacteria, we wondered whether they were also unable to migrate to dying cells in the tuberculous granuloma. If so, a functional macrophage deficiency could develop in the context of global macrophage excess, providing an explanation for our observation of early granuloma breakdown.

12We performed detailed time-lapse confocal imaging of forming granulomas in wild-type and snapc1b mutant larvae over ∼18 hr. Wild-type granulomas retained cellularity over this period through continuous influx of macrophages (Movie S2; Figures 3A and 3B). In contrast, Snapc1b-deficient granulomas broke down soon after formation (Movie S2). Mutant granulomas were surrounded by mostly vacuolated macrophages that failed to migrate into the granuloma (Figures 3A and 3B; Movie S2). The migration deficit in snapc1b mutants was specific to the vacuolated macrophages; the morphologically normal macrophages in the mutants maintained displacements and speeds similar to those of wild-type animals (Figures 3C and 3D; Movie S2). Thus, macrophage lysosomal storage leads to granuloma breakdown by preventing migration to the forming structure and is functionally equivalent to a global macrophage deficiency. Our findings suggest that, once the snapc1b mutants have exhausted their migration-competent macrophages, the tuberculous granuloma breaks down resulting in bacterial cording.

13Human SNAPC1 is involved in global RNA polymerase II-dependent transcription. RNA-seq analysis of snapc1b mutants and wild-type siblings revealed more than 1,000 differentially expressed genes in the mutant (Table S3), suggesting zebrafish Snapc1b functions in a similar manner. How might a broadly acting transcriptional regulator produce such specific phenotypes? Guided by the snapc1b mutant phenotype, we analyzed the RNA-seq dataset for lysosomal genes including those associated with human lysosomal storage disorders (Table S2). Only two, the myeloid cell-specific lysosomal cysteine cathepsins B and L1 (ctsbb and ctsl1), were underrepresented in the mutant, at 9% and 13% of wild-type levels, respectively (Table S2), and we confirmed their commensurate reduction by qRT-PCR analysis (85% and 83%, respectively) (Figure 4A; data not shown). We were able to test lysosomal cathepsin activity in situ using MagicRed (MR)-Cathepsin L, a modified cathepsin L target sequence, which fluoresces only when cleaved. In wild-type animals, brain macrophages quickly cleaved injected MR-cathepsin L; this number was reduced as expected by administration of the irreversible pan-cysteine cathepsin inhibitor, E64d (42.7% of macrophages in control versus 3.63% in E64d-treated larvae, p < 0.0001) (Figure 4B). snapc1b mutants displayed reduced MR-cathepsin-L cleavage compared to wild-type siblings, indicative of reduced lysosomal cathepsin L activity (47.2% of macrophages in WT animals versus 5.3% in mutants, p < 0.0001) (Figure 4C).

14We next tested whether cysteine cathepsin deficiency underlies all of the snapc1b mutant phenotypes. Inhibition of cysteine cathepsins by E64d recapitulated both baseline and infected snapc1b mutant phenotypes—macrophage lysosomal storage with accompanying migratory defects at baseline, and hypersusceptibility to infection with bacterial cording (Figures 4D–4F). We attempted morpholino knockdown of ctsbb and ctsl1 to probe their individual culpabilities in the snapc1b mutant phenotypes (Table S1). As the ctsbb morpholino was highly toxic, we could only pursue ctsl1 further. ctsl1 morphants recapitulated the snapc1b phenotypes (Figures 4G–4I). Transient overexpression of ctsl1 mRNA in snapc1b mutant larvae restored normal macrophage morphology in uninfected snapc1b mutants and rescued cording (Figures 4J and 4K). Together these experiments implicate cysteine cathepsins in snapc1b hypersusceptibility resulting from macrophage incapacitation. Our data ascribe a substantial portion of snapc1b phenotypes to cathepsin L1 deficiency though we cannot rule out a minor role for cathepsin B deficiency. Prior findings that cathepsin L knockout mice are not hypersusceptible to M. tuberculosis may reflect functional redundancies present in the mouse but not the zebrafish.

15Our finding that cathepsin L1 deficiency mediated hypersusceptibility prompted us to ask whether this lysosomal hydrolase might play a role in macrophage microbicidal activity for two reasons. First, a deficit in macrophage microbicidal activity (e.g., through TNF deficiency) has been shown to result in granuloma breakdown with bacterial cording. Second, in vitro, cathepsin L has been reported to indirectly facilitate mycobacterial killing by cleaving ubiquitin into microbicidal peptides. However, we found that macrophages of both snapc1b mutants and ctsl1 morphants restricted bacterial growth normally (Figure S4). These findings suggest that cathepsin L-mediated macrophage microbicidal capacity is dispensable in vivo and confirms that its deficiency induces susceptibility by compromising macrophage migration.

16While cathepsin deficiency causes protein accumulation in lysosomes, many human genetic lysosomal storage disorders result from the accumulation of diverse lipid species. Patients with Gaucher’s disease, the most common lysosomal storage disease, have macrophages with migration defects in vitro and are susceptible to a variety of pathogens including mycobacteria, though this may be due to concomitant immune defects including pancytopenias. We asked whether the mechanism of susceptibility uncovered for ctsl1 deficiency extended to lysosomal storage disorders characterized by lipid accumulation. Knockdown of the zebrafish orthologs of the genes responsible for Gaucher’s disease, Tay-Sachs disease, and metachromatic leukodystrophy produced increased numbers of vacuolated macrophages with enlarged lysosomes and migratory defects (Tables S1 and S2; Figures 5A–5C). Upon infection, all three exhibited early granuloma breakdown and bacterial cording (Figures 5D–5F). Thus, etiologically diverse lysosomal storage disorders can increase susceptibility to tuberculous infection, regardless of the nature of the accumulated material.

17Our work so far had linked macrophage lysosomal storage to impaired migration to increased susceptibility to mycobacteria. Having understood the cellular basis of the link between impaired macrophage migration and susceptibility, we sought to understand how macrophage lysosomal storage might impair migration. Recycling between the endosomal and plasma membranes is known to be required for cell migration. This recycling delivers membrane lipids and proteins required for movement to the plasma membrane and facilitates adjustments in cell-surface area that are critical for cell motility. Lysosomes share contents with endosomes, and recent evidence suggests that, like endosomes, they participate in recycling to the plasma membrane. Accordingly, embryonic fibroblasts isolated from two mouse models of severe lysosomal storage disorders display broad dysregulation of the entire endocytic pathway.

18We asked whether endocytic recycling was disrupted in the zebrafish macrophages with lysosomal storage by monitoring the fate of fluorescently labeled high-molecular-weight dextran (10,000 MW) in normal animals and those with macrophage lysosomal storage. Following endocytosis, high-molecular-weight dextran is trafficked to lysosomes but not readily degraded, and its loss from lysosomes strictly reflects trafficking from them. In pulse-chase experiments, dextran-labeled lysosomes have been shown to fuse rapidly with endosomes, and several hours later the dextran is released into the extracellular medium suggesting subsequent fusion events that involve trafficking to the plasma membrane.

19We injected fluorescent dextran into the brains of 3-dpf zebrafish larvae—wild-type, cathepsin deficient by E64d treatment, and gba-deficient morphants. In all groups, 74%–82% of the macrophages had taken up the dye within 5 hr (Figure 6). After 30 hr, only 33% and 39% of the macrophages in the wild-type fish retained the dextran, whereas 77% and 79% did in the cathepsin and gba-deficient animals, respectively (Figure 6), suggesting that stalling of the entire endocytic system is a common feature of lysosomal storage diseases and underlies the defective migration displayed by vacuolated macrophages.

20Our finding that the lysosomal accumulation of diverse biomolecules compromises endocytic recycling, and thus cell motility, suggested a common mechanism independent of the specific lysosomal substrate. If so, then lysosomal storage induced by non-biological particles should produce the same phenotypes. We injected beads into the HBV, which were phagocytosed by brain resident macrophages (Figure S5A). Bead-laden macrophages were compromised for homeostatic migration and exhibited disruption of endocytic recycling (Figures S5B–S5E).

21In addition to their role in forming and maintaining the granuloma, resident macrophages are the first cells to migrate to mycobacteria at the initial site of infection. This first macrophage-mycobacterium interaction can be visualized in the zebrafish hindbrain ventricle (HBV), a cavity into which phagocytes migrate in response to mycobacteria. In the snapc1b mutant, only the subset of brain-resident macrophages that still had normal morphology migrated to the bacteria and phagocytosed them, while the vacuolated macrophages, failing to migrate from the adjacent brain parenchyma, remained uninfected (data not shown). We could not directly test the migration of bead-laden brain resident macrophages, as only a minority of them engulfed sufficient numbers of beads injected into the HBV. So we injected either beads or the nuclear stain Hoechst 33342 into the caudal vein followed by bacteria into the HBV (Figure S5F). As observed previously, the Hoechst-stained macrophages could be discerned by their blue nuclei and were morphologically normal (Figure S5G). After confirming that similar numbers of circulating macrophages were labeled blue by either dye or beads, we injected bacteria into the HBV (Figure S5F). Multiple Hoechst-positive macrophages migrated to the HBV in response to the bacteria, as expected, but hardly any bead-filled ones did (Figures S5G–S5I). Thus, the accumulation of indigestible inert particles in macrophage lysosomes compromises their migration so as to preclude their ability to phagocytose infecting mycobacteria.

22Human TB is thought to begin when mycobacteria are phagocytosed by pulmonary alveolar macrophages, the resident macrophages at the air-lung interface. Consistent with their role in primary defense against diverse inhaled bacteria, the ability of a substantial number of individuals to clear M. tuberculosis early after infection has been ascribed to the microbicidal activity of the alveolar macrophage. Despite their central defensive role, many alveoli are normally devoid of macrophages because their numbers are limiting. Therefore, efficient and complete phagocytosis of inhaled particulates is predicated on the rapid migration of alveolar macrophages from nearby alveoli. This migration should be particularly relevant to TB, the outcome of which depends upon the fate of the 1-3 bacteria deposited in a distal alveolus, which might not contain a macrophage. In light of our findings that bead-laden macrophages were compromised for migration to newly infecting bacteria in the zebrafish, we wondered whether the accumulation of tobacco smoke particulates in the alveolar macrophages of cigarette smokers might be similarly compromised, accounting for the poorly understood association between smoking and the acquisition of new TB infection. If an infecting mycobacterium were to be deposited in a macrophage-deficient alveolus and not rapidly phagocytosed by nearby alveolar macrophages rendered immotile secondary to lysosomal engorgement, it would have an extended period of extracellular growth before engulfment by alveolar macrophages or other myeloid cells recruited from afar.

23We examined alveolar macrophages obtained from smokers, nonsmokers, and ex-smokers by bronchoalveolar lavage (Table S4; Supplemental Experimental Procedures). The majority of smokers’ alveolar macrophages exhibited vacuolated morphology and had accumulated opaque material in large lysosomal inclusions as evidenced by staining with neutral red, a vital dye that concentrates in lysosomes (Figures 7A and 7B). The abnormal cells were readily identified by their autofluorescence, consistent with previous findings. These cells were present at a lower frequency in ex-smokers and virtually absent in nonsmokers (Figures 7A and 7B).

24Using a transwell assay, we confirmed prior reports that alveolar macrophages from nonsmokers and ex-smokers migrate to zymosan-activated serum, a rich source of the chemoattractant C5a (Figure S6A). In this assay, nonsmokers’ and ex-smokers’ alveolar macrophages also migrated to M. tuberculosis within 2 hr (Figures 7C and S6B). Migration of smokers’ macrophages to M. tuberculosis was impaired (Figure 7D). Our hypothesis predicts that this overall migration impairment is due to a selective inability of the vacuolated subset to migrate. Indeed, by calculating the fraction of the smokers’ normal versus vacuolated macrophages that migrated, we found that the migration impairment was specific to the vacuolated subset (Figure 7E). In sum, we show that the majority of smokers’ macrophages fail to migrate toward M. tuberculosis due to lysosomal accumulation of particulates, and their non-participation may contribute to the susceptibility of these individuals to TB.

25We have described a zebrafish mutant in the snapc1b basal transcription factor component that displays the hallmark characteristics of human lysosomal storage disorders and is hypersusceptible to M. marinum infection. RNA-seq of snapc1b mutants revealed reduced expression of the lysosomal degradative cathepsins L and B, and pharmacological inhibition of cathepsin activity or knockdown of cathepsin L recapitulates the key mutant phenotypes of vacuolated macrophage morphology and susceptibility to infection.

26Though cathepsin L is involved in the lysosomal degradation of phagocytosed material, its deficiency here mediates susceptibility to mycobacteria not by reducing macrophage microbicidal capacity but rather by causing lysosomal accumulation of undigested cell debris. This disrupts endocytic membrane recycling and thereby compromises macrophage migration in a variety of contexts. By modeling human lysosomal storage diseases in the zebrafish, we find that the accumulation of diverse substrates causes susceptibility to infection through this same mechanism. Our studies provide insights into the fundamental and common role played by macrophages as scavengers of dying cells during homeostasis and during tuberculous granuloma maintenance. These insights shed light on the protective role of tissue macrophages in early tuberculous infection and how lysosomal accumulation of tobacco smoke products may compromise this role.

27Sequential live visualization of the developing snapc1b mutant highlights the continuous scavenging role of macrophages under homeostatic conditions. Our work suggests that the accumulation of undigested cell debris in macrophage lysosomes may itself contribute substantially to the pathogenesis of human lysosomal storage diseases. We find that defects in macrophage degradative function render the cell vacuolated, immotile, and unable to further perform a critical scavenging function, which depends on directed migration to the dying cell. This may contribute to the accumulation of unphagocytosed debris from cells undergoing apoptosis in the course of homeostatic tissue remodeling and repair, and the pathological consequences of their secondary necrosis.

28The increased number of apoptotic bodies observed in human lysosomal storage disorders has been attributed to increased cell death triggered by the accumulation of lysosomal substrates. However, macrophages also accumulate lysosomal substrates in a variety of human lysosomal storage diseases; our findings suggest that the resultant immotility of an increasing proportion of macrophages may contribute to the accumulation of dead cells. Because tissue turnover is high in the developing brain, the macrophage scavenging deficit we propose may be particularly relevant for the pathogenesis of the neurological manifestations of lysosomal storage disorders hitherto attributed to neuronal dysfunction. Hematopoietic stem cell transplants in humans and mice improve clinical manifestations of lysosomal storage disorders, including neurological ones that are recalcitrant to enzyme replacement therapy. In light of our findings, we speculate that hematopoietic stem cell transplantation alleviates disease pathology by restoring macrophage degradative function and consequently migration to engulf cell debris.

29Our studies of the snapc1b mutant, in which vacuolated macrophages fail to migrate into the tuberculous granuloma, reveal the inextricable link between macrophage homeostatic and immune function. Like the brain, the forming tuberculous granuloma is an environment with high cell turnover and the maintenance of its cellularity depends on the continuous migration of new macrophages that engulf dying infected macrophages. In the context of tissue remodeling, the clearance of dying cells prevents their secondary necrosis and release of inflammatory material into the extracellular space. Likewise, in the TB granuloma, timely engulfment of dying infected macrophages prevents their secondary necrosis and release of bacteria into the extracellular milieu. Granuloma breakdown is clinically significant because it increases both disease severity and risk of transmission.

30Human lysosomal storage disorders are rare and often lethal within the first year of life, and thus unlikely to be significant contributors to the global burden of TB. Likewise, snapc1b mutant zebrafish fail to reach adulthood, and SNAPC1-null mutations in humans are likely embryonic lethal. However, even relatively small reductions in the macrophage supply to the granuloma can accelerate its breakdown. It is possible that subtle alterations in macrophage degradative function, caused by altered expression of SNAPC1, or lysosomal cathepsins or other hydrolases, could create local macrophage deficits and increased susceptibility to TB. Thus, macrophage lysosomal accumulation from diverse genetic etiologies may together be not insignificant contributors to the global TB burden.

31Finally, our findings that lysosomal storage also compromises the migration of lung resident alveolar macrophages to mycobacteria suggests a mechanism for the observed susceptibility of smokers to new TB infection. There is accumulating evidence for a role for alveolar macrophages being first-responding protective cells in TB. In mice, aerosolized M. tuberculosis is found almost exclusively in alveolar macrophages for the first 7 days, after which infection moves into other myeloid cells such as monocytes and dendritic cells recruited from the lung interstitium or circulation. Their greater microbicidal capacity is mirrored in humans whose alveolar macrophages inhibit M. tuberculosis growth in contrast to peripheral blood monocytes, which are growth permissive. Our findings suggest that migration defects resulting from macrophage lysosomal engorgement impede the rapid engulfment (and therefore eradication) of infecting microbes at points of entry and may therefore facilitate bacterial entry into growth-permissive cells.

32Cigarette smoking increases not only the risk of progression to active pulmonary TB disease, but also the risk of new TB infection, suggesting defective early response mechanisms in smokers. Smokers’ alveolar macrophages phagocytose bacteria and yeast normally and have normal bactericidal activity against M. tuberculosis. The incapacitation of alveolar macrophages by tobacco smoke particulates may contribute to increased risk of infection in two ways: (1) delayed time to phagocytosis by the alveolar macrophage, allowing for a longer extracellular growth period by the bacteria, and (2) increased chance of initial phagocytosis by a recruited, growth-permissive macrophage.

33In addition to providing an explanation for the increased susceptibility of individuals with genetic lysosomal storage disorders to respiratory, skin, and mucosal infections, our findings may constitute a basis for the susceptibility of smokers to other respiratory infections, Finally, this mechanism may also contribute to the poorly understood association between indoor air pollution and TB.

34Smokers’ increased susceptibility to infection may be reversible. A longitudinal study of alveolar macrophages after transplant of a smoker’s lung into a nonsmoker revealed a progressive decrease in “smokers alveolar macrophages” from >90% to 3% in 3 years. In our cohort, not only did ex-smokers have significantly fewer alveolar macrophages with lysosomal storage than smokers, but overall migration to M. tuberculosis was restored. These findings provide an additional rationale for smoking cessation as a prescription for TB prevention.

35Detailed methods and bacterial and zebrafish strains associated with all procedures below are available in Supplemental Experimental Procedures.

36Zebrafish husbandry and experiments were conducted according to guidelines from the UK Home Office, and the US NIH (approved by the University of Washington Institutional Animal Care and Use Committee). The wild-type AB strain was used for experiments except those in which the snapc1b(fh111) line or transgenic lines were used. Unless noted, crosses using snapc1b(fh111) were performed as heterozygote incrosses, which were genotyped at the completion of the experiment, to ensure blinded scoring of phenotypes. Except where noted, “WT” refers to snapc1bfh111/+ and snapc1b+/+. Bacteria, beads, and dye were injected into the caudal vein and/or hindbrain ventricle.

37Alveolar macrophages (AM) were retrieved at bronchoscopy after informed consent and as approved by the Research Ethics Committee of St. James’ Hospital, using a protocol that preserves viability of macrophages from both smokers and nonsmokers. Macrophage migration and microscopical visualization procedures are detailed in Supplemental Experimental Procedures.

38Wild-type M. marinum (Mm) (strain M - ATCC #BAA-535) expressing tdTomato under the constitutive promoter msp12 was used for fluorescence microscopy and quantification of intracellular bacterial burdens. The attenuated Δerp mutant Mm was used to enumerate intracellular bacteria, and WT Mm was used for all other assays. Bacterial were cultures and prepared for injection as described.

39M. tuberculosis H37Ra (ATCC 25177) was used for the human alveolar macrophage studies and prepared as described in Supplemental Experimental Procedures.

40Statistical analyses were performed using Prism 6 (GraphPad). Not significant, p ≥ 0.05, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001.

41R.D.B., S.L., C.J.C., J.C., K.T., C.B.M., D.M.T., and L.R. conceived, designed, and analyzed zebrafish experiments; R.D.B., S.L., C.J.C., J.C., K.T., and D.M.T. performed these experiments; S.L., M.P.O., S.M.O., J.K., and L.R. conceived, designed, and analyzed human experiments; M.P.O. and S.M.O. performed these experiments; S.L., R.D.B., and L.R. wrote the paper with input from J.K., D.M.T., M.P.O., S.M.O., and C.B.M.; K.T., and S.L. prepared the figures.