Cathepsin B modulates lysosomal biogenesis and host defense against Francisella novicida infection

Lysosomes are cytoplasmic membrane–enclosed organelles filled with numerous acidic hydrolases. These organelles are dynamic and can fuse with a variety of vesicles to mediate degradation of extracellular materials that have been endocytosed or intracellular components that have been sequestered by autophagy. Furthermore, lysosomes are recycled and reformed after fusion with cargo-containing vacuoles (Luzio et al., 2007; Yu et al., 2010). Fundamental cellular processes orchestrated by lysosomes require cathepsin proteases; dysregulation of these activities contributes to the development of malignancy and Alzheimer’s disease (Gocheva et al., 2006; Mueller-Steiner et al., 2006). Secretion of cathepsins from lysosomes is involved in lysosomal membrane permeabilization–mediated cell death via caspase-dependent and caspase-independent pathways (Kreuzaler et al., 2011; Aits and Jäättelä, 2013). In addition, lysosomal rupture and cytoplasmic release of cathepsin B triggered by phagocytosis of crystals lead to activation of the NLRP3 inflammasome and maturation of the proinflammatory cytokines IL-1β and IL-18, which are key events linked to the development of atherosclerosis and Alzheimer’s disease (Halle et al., 2008; Duewell et al., 2010).

Certain lysosomal proteases also provide antimicrobial host defense against infection. Lysosomal acidic hydrolases, especially cathepsins, are executioner proteases used for degradation of intracellular bacteria within lysosomes. Macrophages lacking cathepsin D have a reduced capacity to kill bacteria during pneumococcal infection (Bewley et al., 2011). Similarly, mice with a genomic deletion of cathepsin E have increased susceptibility to infection by Staphylococcus aureus and Porphyromonas gingivalis (Tsukuba et al., 2006). Furthermore, a protective role for cathepsin L has been reported for Mycoplasma pulmonis infection (Xu et al., 2013). However, the relative contribution of different members within this large protease family in antibacterial host defense is largely underappreciated.

Here, we identified an important role for cathepsin B in driving increased susceptibility to the cytosolic pathogen Francisella novicida but not to the vacuolar-restricted pathogen Salmonella enterica serovar Typhimurium (S. Typhimurium). Genetic deletion or pharmacological inhibition of cathepsin B triggered signaling via down-regulation of mechanistic target of rapamycin (mTOR) and the lysosomal calcium channel TRPML1 to activate transcription factor EB (TFEB), which in turn results in increased lysosomal biogenesis and autophagy to mediate clearance of the bacteria. This phenomenon was not observed with genetic deletion of cathepsin G, elastase, or elastase and neutrophil proteinase 3. Our results, therefore, identified a fundamental negative regulatory function of cathepsin B in the host cell against F. novicida infection.

Intracellularly replicating bacterial pathogens have evolved strategies to evade degradation by lysosomes to survive and replicate in the hostile environment of the host cell (Ray et al., 2009; Huang and Brumell, 2014). F. novicida escapes the vacuole to undergo replication in the cytoplasm of infected cells, whereas S. Typhimurium replicates inside S. enterica–containing vacuoles. Despite decades of research on cathepsin biology, the role of lysosomal cathepsins in host defense against intracellularly replicating bacterial pathogens with distinct intracellular lifestyles has remained enigmatic. To address this, WT mice and mice deficient in the lysosomal cathepsin B, cathepsin G, or the extralysosomal proteases elastase and neutrophil proteinase 3 were infected with F. novicida or S. Typhimurium, and survival was monitored over time. Mice of all genotypes succumbed with similar kinetics to S. Typhimurium infection with most animals dying within 8 d after infection (Fig. 1 A). In contrast, when infected with F. novicida, cathepsin B–deficient mice (Ctsb^−/− mice) were significantly protected from lethality, with 65% of animals from this group surviving beyond day 10, compared with only 25% of WT mice (Fig. 1 B). Enhanced resistance to F. novicida infection was specific to Ctsb^−/− mice because mice lacking cathepsin G, elastase, or elastase and neutrophil proteinase 3 died with similar kinetics as WT mice (Fig. 1 B).

Reduced mortality in Ctsb^−/− mice could be caused by reduced bacterial burden. Therefore, we determined the number of bacteria in the spleen and liver of WT and Ctsb^−/− mice after 1 and 3 d of infection with F. novicida. Bacterial loads in the liver and spleen of Ctsb^−/− mice were significantly lower than those in WT mice on day 3 but not on day 1 after infection (Fig. 1, C and D). In agreement, less granulomas were formed in the liver of infected Ctsb^−/− mice relative to WT controls (Fig. 1 E). Reduced bacterial burden of Ctsb^−/− mice on day 3 after infection was also associated with reduced production of proinflammatory cytokines and chemokines in the liver (Fig. 1 F). Moreover, cell death and associated activation of the apoptotic caspases 3, 7, and 8 was diminished in the liver tissue of Ctsb^−/− mice 3 d after infection (Fig. 1, G and H). These results suggested cathepsin B plays negative roles in host defense against F. novicida infection.

Recruitment of immune cells and production of proinflammatory cytokines are hallmarks of effective innate immune responses to infection (Cowley and Elkins, 2011). One possibility is that the improved early innate immune response in Ctsb^−/− mice might restrict bacterial replication or increase bacterial killing and clearance. Therefore, we investigated the number of circulating total white blood cells, red blood cells, neutrophils, and monocytes in the blood of uninfected WT and Ctsb^−/− mice and WT and Ctsb^−/− mice infected with F. novicida. Relative to their uninfected controls, both WT and Ctsb^−/− mice exhibited a global reduction in the numbers of circulating white blood cells (lymphocytes and myeloid cells) and monocytes and an increase in the number of neutrophils 24 h after F. novicida infection, whereas total red blood cell counts remained unchanged (Fig. 2 A). These results suggested that certain populations of circulating immune cells might have migrated from the BM to the sites of infection. However, no significant differences were noted in the prevalence of specific immune cell populations in circulation between infected WT and Ctsb^−/− mice (Fig. 2 A). We also did not observe any difference in the number of CD11b⁺ Ly6g⁺ neutrophils and CD11b⁺ Ly6g⁻ monocyte populations in the liver and spleen of WT and Ctsb^−/− mice 24 h after F. novicida infection when WT and Ctsb^−/− mice have comparable bacterial burden in the liver and spleen (Fig. 2 B). Finally, the levels of the proinflammatory cytokines TNF, IL-1β, and IL-6 and the keratinocyte-derived chemokine (CXCL1) in liver and spleen of infected WT and Ctsb^−/− mice were similar after 24 h of infection (Fig. 2 C), suggesting that cathepsin B controlled host defense against F. novicida independently of immune cell recruitment and cytokine or chemokine production.

Macrophages play a key role in controlling intracellular replication and dissemination of F. novicida (Hall et al., 2008). To investigate the role of cathepsin B in the control of F. novicida replication in macrophages, we infected primary BMDMs from WT and Ctsb^−/− mice with F. novicida and determined the number of intracellular bacteria over time. Notably, Ctsb^−/− BMDMs cleared bacteria more efficiently relative to WT BMDMs at 8 and 24 h after infection (Fig. 2 D). In agreement, inhibition of the cathepsin B protease activity with CA-074 methyl ester (Me) also enhanced the bactericidal activity of WT BMDMs (Fig. 2 E).

F. novicida infection in macrophages is detected by the DNA-sensing AIM2 inflammasome (Fernandes-Alnemri et al., 2010; Jones et al., 2010; Rathinam et al., 2010; Man et al., 2016), but cathepsin B was not required for AIM2-dependent activation of caspase 1, the subsequent release of the caspase 1 substrate IL-1β, and pyroptotic cell death in BMDMs infected with F. novicida (Fig. 3, A–C). Other members of the caspase family also contribute to innate immunity during infection (Man and Kanneganti, 2016). However, infected Ctsb^−/− BMDMs displayed similar levels of caspases 3, 7, and 8 activation compared with infected WT BMDMs, in agreement with results from CA-074 Me–pretreated WT BMDMs (Fig. 3 D). Differential activation of apoptotic caspases was also not observed in liver tissues from WT and Ctsb^−/− mice after 24 h of infection (Fig. 3 E).

Apart from inflammasome signaling, type I interferons, inducible nitric oxide synthase (iNOS), and antimicrobial peptides regulate host protection against F. novicida infection (Lindgren et al., 2005; Henry et al., 2007). However, WT and Ctsb^−/− BMDMs that have been infected with F. novicida expressed similar transcript levels of Ifnb, iNos, and the antimicrobial glycoprotein lipocalin 2 (Lcn2; Fig. 3 F). Also, the levels of activated NF-κB and extracellular signal–regulated kinase (ERK) as well as production of IL-6 and TNF were not differentially regulated in F. novicida–infected WT and Ctsb^−/− BMDMs (Fig. 3, G and H). These data collectively suggested that the role of cathepsin B in the regulation of F. novicida replication is uncoupled from macrophage cell death, inflammasome activation, and inflammatory cytokine production.

Fusion of lysosomes with pathogen-occupied vacuoles is a major bactericidal mechanism in macrophages (Roy et al., 2006). In response, bacterial pathogens have evolved mechanisms to evade the lysosomal degradation pathway. During F. novicida infection, bacteria escape the vacuoles of human macrophages after 3–4 h of infection (Santic et al., 2005; Clemens et al., 2009). To investigate whether improved control of F. novicida replication in the absence of cathepsin B was caused by increased fusion of F. novicida–containing vacuoles with lysosomes, we examined colocalization of a GFP-expressing strain of F. novicida with LysoTracker in infected WT and Ctsb^−/− BMDMs. We found a significantly elevated frequency of F. novicida colocalizing with lysosomes in Ctsb^−/− BMDMs compared with the numbers observed in WT BMDMs at 5 h (67.2 ± 1.8% vs. 51.1 ± 2.2%) and 12 h (94.8 ± 0.4% vs. 86.8 ± 0.9%) after infection (Fig. 4, A–D). These results suggested that the prevalence of lysosome-associated F. novicida increased in Ctsb^−/− BMDMs compared with WT BMDMs. In agreement, treatment of WT BMDMs with a cathepsin B protease inhibitor CA-074 Me significantly increased the prevalence of LysoTracker-positive F. novicida (Fig. 4 E). Notably, we observed that Ctsb^−/− BMDMs contained significantly more LysoTracker-positive vacuoles before infection, and the number of LysoTracker puncta in uninfected Ctsb^−/− BMDMs was nearly twofold that of WT BMDMs (15.9 ± 0.5 vs. 8.3 ± 0.3 per cell; Fig. 4, F and G). The increased prevalence of LysoTracker-positive vacuoles and increased expression of the lysosomal membrane protein LAMP1 in Ctsb^−/− BMDMs was further confirmed by flow cytometry and immunoblotting analysis, respectively (Fig. 4, H–K). Similar to Ctsb^−/− BMDMs, the cathepsin B inhibitor CA-074 Me increased the number of LysoTracker puncta and expression of LAMP1 in WT BMDMs (Fig. 4, L–N).

Transmission electron microscopy reveled that uninfected Ctsb^−/− BMDMs contained larger single-membrane lysosomes and double-membrane autophagosomes compared with that in uninfected WT BMDMs (Fig. 5, A and C). Remarkably, higher resolution analysis of these micrographs revealed increased accumulation of partially digested vesicles in autolysosomes of Ctsb^−/− BMDMs (Fig. 5 B), suggesting impaired nutrient recycling by Ctsb^−/− lysosomes. Collectively, these results indicated that the enhanced antimicrobial activity of Ctsb^−/− BMDMs to F. novicida might be attributed to their increased lysosomal biogenesis, which may result in increased efficiency in delivering the bacteria to lysosomes for killing.

We performed genome-wide expression analysis to profile differentially expressed genes in uninfected WT and Ctsb^−/− BMDMs to examine whether the increased number of lysosomes in Ctsb^−/− BMDMs is caused by enhanced lysosomal biogenesis. We identified increased expression of the gene encoding the TFEB and differentially regulated genes involved in the Coordinated Lysosomal Expression and Regulation (CLEAR) network as well as genes encoding lysosomal hydrolases, lysosomal membrane proteins, lysosomal acidification proteins, and autophagy proteins (Fig. 6 A and Fig. S1; Palmieri et al., 2011). Importantly, autophagy and lysosomal biogenesis in response to starvation signals are up-regulated by the CLEAR gene network and its regulatory transcription factor TFEB (Sardiello et al., 2009; Settembre et al., 2011). Consistent with the increased expression of TFEB mRNA, we found a substantial increase in the protein expression level of TFEB in uninfected Ctsb^−/− BMDMs relative to those of WT controls (Fig. 6 B). Activated TFEB translocates into the nucleus to drive transcription of genes encoding lysosomal proteins containing the CLEAR element (Sardiello et al., 2009). We indeed observed increased nuclear localization of TFEB in Ctsb^−/− BMDMs by immunofluorescence microscopy and immunoblotting analysis of TFEB (Fig. 6, C–E). These findings indicated that Ctsb^−/− BMDMs have a superior capacity to drive lysosomal biogenesis by unleashing the activity of TFEB.

Activation and nuclear localization of TFEB are tightly regulated by signaling via mTOR (Settembre et al., 2012). Activated mTOR retains TFEB in the cytoplasm by inducing phosphorylation of TFEB, whereas inhibition of mTOR results in activation and nuclear translocation of TFEB. We investigated whether the mechanism of increased TFEB activation in the absence of cathepsin B was caused by down-regulated mTOR activation. Remarkably, phosphorylation of both mTOR and its direct target 4E-BP were significantly diminished in uninfected Ctsb^−/− BMDMs compared with uninfected WT BMDMs (Fig. 6 F). We further confirmed this observation and found that WT BMDMs treated with CA-074 Me recapitulated the results obtained in Ctsb^−/− BMDMs (Fig. 6 F).

Inhibition of mTOR is mediated by the conserved sensor of cellular energy and nutrient status, known as adenosine monophosphate–activated protein kinase (AMPK; Alers et al., 2012). AMPK also directly activates ULK1 to induce the autophagy pathway under stressed conditions (Kim et al., 2011). Notably, phosphorylation of both AMPK and ULK1 were substantially elevated in Ctsb^−/− BMDMs (Fig. 6 G). Moreover, WT BMDMs treated with CA-074 Me phenocopied Ctsb^−/− BMDMs (Fig. 6 G). To further confirm alterations in autophagy, we examined conversion of the autophagy marker LC3-I to the lipidated form of LC3-II. We indeed observed increased level of LC3-II in untreated Ctsb^−/− BMDMs or in WT BMDMs treated with the cathepsin B inhibitor compared with untreated WT BMDMs (Fig. 6 G).

Next, we examined which pathway governed by mTOR, AMPK, and TFEB was the apical signal initiating the increase in lysosomal biogenesis and autophagy. A time course experiment assessing WT BMDMs treated with CA-074 Me showed that LC3-II accumulation and reduction of phosphorylated mTOR occurred within 30 min of cathepsin B inhibition (Fig. 6 H). In line with reduced phosphorylation of mTOR, the expression of TFEB was induced after treatment with the inhibitor for 30 min (Fig. 6 H). Phosphorylation of AMPK and ULK1 increased after 2 h of treatment with the cathepsin B inhibitor (Fig. 6 H), suggesting that AMPK and ULK1 activation succeeded LC3-II accumulation, reduction of phosphorylated mTOR, and activation of TFEB. These results suggested that impaired lysosomal degradation and recycling contributed to lysosomal biogenesis, whereas signaling via AMPK preferentially regulated the autophagy pathway and is not the apical signal driving increased lysosomal biogenesis in Ctsb^−/− BMDMs. Importantly, we did not observe substantial differences in the number of lysosomes, phosphorylation of AMPK and ULK1, and protein expression of LC3-II, TFEB, and LAMP1 in Cathepsin G^−/− (Ctsg^−/−), Elastase^−/−, or Nepr3^−/− BMDMs compared with WT BMDMs (Fig. 6, I and J), suggesting that enhanced lysosomal biogenesis and autophagy are specific to Ctsb^−/− BMDMs. Collectively, these results provided evidence to support that cathepsin B negatively regulated lysosomal biogenesis and autophagy at both the transcriptional and posttranslational levels through a mechanism that involved TFEB activation and ULK1 phosphorylation.

Lysosomal ion channels and transporters play essential roles in regulating signals from the lysosome (Xu and Ren, 2015). For instance, the vacuolar ATPase, a key player in lysosome nutrient sensing machinery, can sense lysosomal nutrient levels and regulate mTORC1 activity (Settembre et al., 2013). Calcium signaling mediated by the lysosomal calcium channel TRPML1 (MCOLN1) regulates TFEB activity from the lysosome through the phosphatase calcineurin. Local calcineurin activation via TRPML1 is essential for nuclear translocation and activation of TFEB. The pathway regulating lysosomal biogenesis and autophagy seems to operate independently of mTOR (Medina et al., 2015). Previous studies have shown that cathepsin B mediates proteolytic cleavage of the lysosomal calcium flux channel TRPML1 (Kiselyov et al., 2005) and that a loss of TRPML1 induces leakage of cathepsin B into the cytoplasm (Colletti et al., 2012). It is possible that cathepsin B might mediate TRPML1 channel activity to regulate TFEB activity. Indeed, the protein level of the full-length TRPML1 was increased in Ctsb^−/− BMDMs, whereas differential levels of the transcription of the gene encoding TRPML1 was not observed between untreated WT and Ctsb^−/− BMDMs (Fig. 7, A and B), suggesting decreased cleavage of TRPML1 in Ctsb^−/− BMDMs. We enriched lysosome fractions from BMDMs to investigate the level of TRPML1 cleavage. Cleavage of TRPML1 in the lysosome was substantially impaired in untreated Ctsb^−/− BMDMs compared with untreated WT BMDMs (Fig. 7 C). We further examined whether an increased level of the full-length TRPML1 protein in Ctsb^−/− BMDMs resulted in altered Ca²⁺ efflux from the lysosome. WT and Ctsb^−/− BMDMs were treated with the TRPML1 agonist ML-SA1 (Shen et al., 2012). We observed increased Ca²⁺ release from the lysosome in Ctsb^−/− BMDMs stimulated with ML-SA1 compared with treated WT BMDMs (Fig. 7 D). To further confirm whether lysosomal calcium signaling supported the increased TFEB activity in Ctsb^−/− BMDMs, we treated Ctsb^−/− BMDMs with the lysosomal TRPML1 inhibitor PI(4,5)P₂ (Zhang et al., 2012) and found that, over 8 h of inhibitor treatment, the expression of TFEB was substantially reduced in Ctsb^−/− BMDMs (Fig. 7, E and F). Similar effects were observed in Ctsb^−/− BMDMs treated with the calcineurin inhibitor cyclosporine A (Fig. 7, G and H). Collectively, these results suggested that cathepsin B negatively regulated TFEB activity, in part through lysosomal calcium signaling and degradation of TRPML1.

Autophagy has been previously shown to mediate reentry of cytosolic-dwelling F. novicida to the endocytic pathway and fusion with lysosomes (Checroun et al., 2006). To investigate how cathepsin B modulates lysosomal biogenesis and autophagy during F. novicida infection, WT and Ctsb^−/− BMDMs were infected with F. novicida and analyzed for LC3⁺ puncta formation and TFEB expression. We found that F. novicida infection induced autophagy formation and significantly elevated the number of LC3⁺ puncta per cell in Ctsb^−/− BMDMs compared with infected WT BMDMs (Fig. 8, A and B). Moreover, the size of LC3⁺ puncta in infected Ctsb^−/− BMDMs was markedly increased relative to those observed in infected WT BMDMs (Fig. 8 C). Notably, the frequency of autophagosome-associated F. novicida was significantly higher in Ctsb^−/− BMDMs than in WT BMDMs (Fig. 8 D). Furthermore, we found an increased frequency of nuclear-localized TFEB in Ctsb^−/− BMDMs compared with WT BMDMs after F. novicida infection (Fig. 8, E and F). Accordingly, both mRNA and protein levels of TFEB in infected Ctsb^−/− BMDMs were higher than that of infected WT BMDMs (Fig. 8, G and H). We observed elevated phosphorylation of AMPK and ULK1 in Ctsb^−/− BMDMs compared with their WT counterparts after F. novicida infection (Fig. 8, H and I). In line with the observation of increased phosphorylation of AMPK, the ratio of ADP to ATP in Ctsb^−/− BMDMs was significantly higher than that of WT BMDMs after F. novicida infection (Fig. 8 J). Phosphorylated mTOR was further reduced in Ctsb^−/− BMDMs infected with F. novicida (Fig. 8, H and I).

To further confirm that autophagy could play a protective role in the host defense against F. novicida infection, BMDMs from WT and autophagy-defective Atg7^−/− mice (Martinez et al., 2011) were infected with F. novicida and analyzed for LC3 conversion and intracellular bacterial burden. Reduced LC3-II was observed in infected Atg7^−/− BMDMs relative to infected WT BMDMs (Fig. 9, A and B), which was associated with increased numbers of F. novicida bacteria recovered from infected Atg7^−/− BMDMs compared with the number recovered from infected WT BMDMs (Fig. 9 C).

To investigate whether the molecular mechanism orchestrated by cathepsin B to regulate autophagy and lysosomal biogenesis is recapitulated in a physiological setting, we analyzed signaling pathways in liver tissue homogenates from WT and Ctsb^−/− mice. In agreement with our cellular analyses, uninfected Ctsb^−/− mice or those infected with F. novicida exhibited a higher level of LC3-II compared with the corresponding WT controls (Fig. 9, D and E). Remarkably, LAMP1 expression was dramatically up-regulated in the liver of uninfected Ctsb^−/− mice compared with uninfected WT mice, and both infected WT and Ctsb^−/− mice maintained elevated levels of LAMP1 protein expression (Fig. 9 D). Moreover, expression of TFEB was significantly increased in Ctsb^−/− mice (Fig. 9 F). In contrast, expression of Lcn2, a bacteriostatic factor produced during the innate immune response to bacterial infection, was similar in WT and Ctsb^−/− mice after F. novicida infection (Fig. 9 F). Collectively, we identified a central role for cathepsin B in governing activities of TFEB and the kinase ULK1, resulting in suppression of lysosomal biogenesis and autophagy, respectively. Thus, the coordination of autophagy and lysosomal biogenesis induced by inhibition of cathepsin B facilitates host defense against F. novicida infection (Fig. 10).

Lysosomal cathepsins have been conventionally recognized as executive proteases for cargo degradation and cell death. Our study identified a central role for cathepsin B in governing activities of TFEB and the kinase ULK1, resulting in suppression of lysosomal biogenesis and autophagy, respectively. The involvement of cathepsin B in the lysosomal cell death pathway and activation of the NLRP3 inflammasome under certain physiological settings have attracted scientific interests (Tschopp and Schroder, 2010; Man and Kanneganti, 2015). Our results, however, indicated that genetic deletion of cathepsin B is dispensable for activation of the AIM2 inflammasome, cytokine production, and cell death typically associated with F. novicida infection (Henry et al., 2007; Fernandes-Alnemri et al., 2010; Jones et al., 2010; Rathinam et al., 2010; Man et al., 2015; Meunier et al., 2015), suggesting that there were additional roles of cathepsin B in host defense beyond inflammasome activation and cell death.

In this study, we found a role for cathepsin B in negatively regulating lysosomal biogenesis and autophagy, such that cells lacking cathepsin B possessed a superior ability to induce fusion of their lysosomes with F. novicida. It is also interesting to speculate that F. novicida might be engulfed by autophagosomes and require cathepsin B to mediate entry to the cytoplasm for further replication. In addition, it is possible that the localization of cathepsin B proteins to a particular location within the host cell might contribute to their ability to influence the number of bacteria in the host cell.

We found that genomic deletion of cathepsin B in mice did not confer protection against the vacuolar-adapted pathogen S. Typhimurium, which could be because of differences in the mechanisms used by this bacterium to evade the lysosomal degradation pathway or to the functional redundancy among cathepsin proteases that are capable of inducing degradation of S. Typhimurium (Diacovich and Gorvel, 2010; Huang and Brumell, 2014). Further studies investigating the role of cytosolic pathogens, including both bacterial and viral pathogens, would yield insight into the biology of cathepsin B in infectious diseases. In particular, examining the effect of a range of doses of F. novicida infection in vivo would provide information on the strength of the impact of cathepsin B deletion (Nano et al., 2004; Kanistanon et al., 2012; Chu et al., 2014).

The role of autophagy in the host defense against members of the Francisella genus is controversial. The ability of different species or strains of the Francisella genus to evade the autophagy pathway might determine their intracellular survival (Chiu et al., 2009; Cremer et al., 2009; Chong et al., 2012; Steele et al., 2013). Results from the ATG7-deficient BMDMs from our study suggested that canonical autophagy is protective in the host defense against F. novicida infection. Lysosomes play critical roles in nutrient sensing and energy homeostasis (Settembre et al., 2013). During infection, neutralization of lysosomal pH by uropathogenic Escherichia coli in infected bladder epithelial cells can be sensed by the lysosomal cation channel TRPML3, subsequently triggering expulsion of exosome-encased bacteria (Miao et al., 2015). Our study now reveals a critical role of cathepsin B in controlling activity of the key lysosomal Ca²⁺ channel TRPML1. A lack of cathepsin B would leave the activity of TRPML1 unchecked, and a previous study has demonstrated that Ca²⁺ release by lysosomes via TRPML1 activates the phosphatase calcineurin, which, in turn, drives nuclear translocation of TFEB (Medina et al., 2015). Stressed signals from the lysosome itself might also inhibit mTOR activity and reduce phosphorylation of TFEB, leading to TFEB-dependent transcription of genes encoding molecules involved in lysosomal biogenesis and autophagy. The down-regulation of mTOR activity in Ctsb^−/− BMDMs is consistent with a recent study showing that the mTORC1 pathway is activated by lysosomal degradation (Palm et al., 2015). It is not clear which specific signal elicited in the absence of cathepsin B down-regulated the phosphorylation of mTOR. However, enhanced AMPK phosphorylation in the absence of cathepsin B might provide a positive feedback reflecting increased lysosomal biogenesis, as this event occurred subsequent to TFEB and mTOR activation.

Our results unveiled a fundamental biological function of cathepsin B in providing a checkpoint for homeostatic maintenance of the lysosome population and basic recycling functions in the cell. Importantly, deficiency in cathepsin proteases causes lysosome storage diseases in humans owing to accumulation of partially digested macromolecules and enlarged lysosomes (Alroy and Lyons, 2014). TFEB-mediated lysosomal biogenesis and autophagy are crucial in exogenous antigen presentation and cancer metabolism (Perera et al., 2015; Samie and Cresswell, 2015). Cathepsin B can also mediate trypsinogen activation to promote the development of acute pancreatitis (van Acker et al., 2002). Our findings suggested that cathepsin B is a negative regulator of lysosomal biogenesis and autophagy in the intracellular milieu during a microbial infection, which broadens the biological repertoire of this protease in the context of health and disease. Our study, therefore, raises the intriguing possibility that modulating cathepsin B activity may hold therapeutic potential as a mean to increase host immunity against lethal F. novicida infection and to offer treatment options to those with debilitating lysosome storage diseases.

WT C57BL/6J mice were purchased from The Jackson Laboratory and bred in house. Ctsb^−/− mice were provided by T. Reinheckel (Albert-Ludwigs-University Freiburg, Freiburg, Germany), generated as described previously (Halangk et al., 2000), and backcrossed to the C57BL/6J background. Ctsg^−/− (MacIvor et al., 1999) and Nepr3^−/− (Kessenbrock et al., 2008) mice were provided by C. Pham (Washington University School of Medicine, St. Louis, MO). Aim2^−/− (Jones et al., 2010) mice were provided by V.M. Dixit (Genentech, San Francisco, CA). LysM-Cre-Atg7^f/f mice were described previously (Martinez et al., 2011). Elastase^−/− mice were purchased from The Jackson Laboratory (stock no. 006112; Belaaouaj et al., 1998). All mutant mice were crossed with the C57BL/6J mice for >10 generations. All mice were kept in specific pathogen–free conditions within the Animal Resource Center at St. Jude Children’s Research Hospital. Animal studies were conducted under protocols approved by the Institutional Animal Care and Use Committee of St. Jude Children’s Research Hospital.

The bacterial strains used in this study were F. novicida strain U112, GFP-expressing F. novicida (from D.M. Monack, Stanford University, Stanford, CA), and S. Typhimurium strain SL1344. F. novicida and S. Typhimurium were grown overnight with shaking in tryptic soy broth (TSB) and Luria-Bertani medium, respectively, subcultured for an additional 3–4 h, and resuspended in PBS. 8–10-wk-old mice were injected subcutaneously with F. novicida U112 (1.5 × 10⁵ CFU per mouse) as described previously (Man et al., 2015). Mice were weighted and monitored daily over a period of 3 wk. Mice were euthanized at days 1 and 3 after infection, and liver and spleen were harvested to determine the bacterial burden. Liver and spleen were homogenized, serially diluted, plated onto TSB agar plates, and incubated overnight at 37°C as described previously (Man et al., 2015). For S. Typhimurium infection, mice were infected i.p. with 5 × 10³ CFU per mouse. For immune cell infiltration analysis, whole blood from WT and Ctsb^−/− mice was analyzed on a hematology analyzer (Forcyte; Oxford Sciences) by a hematologist blinded to the experimental groups.

To generate BMDMs, BM cells were cultured in L929 cell–conditioned IMDM supplemented with 10% FBS, 1% nonessential amino acids, and 1% penicillin-streptomycin for 5 d. Immortalized BM-derived cells from Trpml1^−/− mice used for controls were from H. Fares (University of Arizona, Tucson, AZ; Miller et al., 2015). BMDMs were seeded in 12-well plates (1 million cells per well) and cultured overnight. The next day, cells were washed and supplied with fresh media without antibiotics. BMDMs were infected with F. novicida for the indicated times. For inhibitor treatment, BMDMs were treated with the cathepsin B inhibitor CA-074 Me (sc-214647; Santa Cruz Biotechnology, Inc.), AMPK inhibitor compound C (171260; EMD Millipore), calcineurin inhibitor cyclosporine A (PHR1092; Sigma-Aldrich), or lysosomal TRPML1 inhibitor PI(4,5)P₂ (R4504; Echelon Biosciences; Zhang et al., 2012). BMDMs were lysed in radioimmunoprecipitation assay buffer with protease and phosphatase inhibitors (Roche) for immunoblot analysis.

BMDMs were infected with F. novicida with a multiplicity of infection (MOI) of 10 for 2 h and washed, and 50 µg/ml gentamicin was added to kill extracellular bacteria. After 1 h, cells were washed twice and cultured in fresh media. BMDMs were lysed in sterile water at the indicated times after infection, serially diluted, plated onto TSB agar plates, and incubated overnight for CFU enumeration.

Liver tissues were fixed in 10% formalin and embedded in paraffin for immunochemical staining. 4-µm–thick liver sections were stained with anti–cleaved caspase 3 or anti-myeloperoxidase (MPO) followed by incubation for 30 min with Rabbit-on-Rodent HRP polymer (RMR622; BioCare Medical). TUNEL (terminal deoxynucleotidyl transferase deoxyuridine triphosphate nick end labeling) staining was performed with the Dead End kit (G7130) according to the manufacturer’s instructions (Promega). Liver tissues were homogenized in radioimmunoprecipitation assay buffer with protease and phosphatase inhibitors for immunoblot analysis. Protein concentrations were determined using a bicinchoninic acid assay kit (Thermo Fisher Scientific) according to manufacturer’s instructions.

Samples were separated by 12% SDS-PAGE and followed by electrophoretic transfer onto polyvinylidene fluoride membranes as described previously (Karki et al., 2015). Membranes were blocked with 5% nonfat milk and further incubated overnight in primary antibody at 4°C. The following primary antibodies were used: mouse IκBα, phosphorylated IκBα (p-IκBα; S32), ERK, p-ERK (T202Y204 [Cell signaling Technology], 9242L, 2859L, 9101L, and 9102L), anti–caspase 1 (AG-20B-0042; Adipogen), anti-LC3B (NB600-1384; Novus Biologicals), anti-LAMP1 (14-1071-85; eBioscience), antibodies for caspase 3, 7, and 8 (9661S, 9491S, and 8592S; Cell Signaling Technology), anti–cathepsin B (AF965; R&D Systems), antibodies for p-ULK1 (S555), total mTOR, p-mTOR (S2448), p–4E-BP1 (T37/46), total AMPK, p-AMPK (T172 [Cell Signaling Technology], 5869S, 2972S, 5536S, 2855P, 2793S, and 2535P), anti-ULK1 (A7481; Sigma-Aldrich), anti-TFEB (A303-673A; Bethyl Laboratories, Inc.), anti-TRPML1 (M1571; Sigma-Aldrich), anti-ATP6V1B2 (ab73404; Abcam), anti-MARCKS (myristoylated alanine-rich C kinase substrate; ab51100; Abcam), anti–lamin B (SC-6216; Santa Cruz Biotechnology, Inc.), and anti-GAPDH (5174S; Cell Signaling Technology). The secondary antibodies used were HRP-labeled anti–rabbit, anti–mouse, or anti–goat antibodies (Jackson ImmunoResearch Laboratories, Inc.). Quantification of the relative protein expression was processed using ImageJ (National Institutes of Health).

For LysoTracker staining and colocalization with F. novicida, WT and Ctsb^−/− BMDMs were infected with GFP-expressing F. novicida and followed by labeling with 100 nM LysoTracker (L-7528; Molecular Probes) for the last 30 min. Cells were washed and fixed with 4% paraformaldehyde for 15 min at room temperature. Cells were washed three times and mounted using mounting medium (H-1200; Vector Laboratories). For LC3B, LAMP1, and TFEB immunostaining, uninfected and infected BMDMs were fixed in 4% paraformaldehyde for 15 min at room temperature. Cells were then permeabilized with cold methanol for 5 min, washed with PBS, and blocked in 1× ELISA buffer with 0.1% saponin for 1 h. Cells were stained with anti-LC3B (NB600-1384; Novus Biologicals), anti-LAMP1 (14-1071-85; eBioscience), or anti-TFEB (A303-673A; Bethyl Laboratories, Inc.), all at 1:500 dilution, overnight at 4°C. Cells were washed, stained with fluorescence-conjugated secondary antibody for 1 h, and mounted using mounting medium (H-1200; Vector Laboratories). Cells were observed on a confocal microscope (Axio Observer; Z1; SlideBook 6 software; ZEISS) for image acquisition and data analysis.

BMDMs were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 1 h at 37°C. Cells were embedded and sectioned for transmission electron microscopy by the Cell and Tissue Imaging Core Facility of St. Jude Children’s Research Hospital. Quantification of lysosome and autophagosome size was performed with ImageJ.

Lysosomes from BMDMs were isolated using a lysosome enrichment kit with cultured cell sonication and density gradient centrifugation, followed by lysosome precipitation according to the manufacturer’s manual (89839; Thermo Fisher Scientific). The purity of isolated lysosomes was determined by detecting the lysosomal membrane protein LAMP1, luminal lysosomal protease cathepsin B, nuclear protein lamin B, and GAPDH using immunoblotting analysis.

WT and Ctsb^−/− BMDMs were labeled with a calcium indicator (O-6807; Thermo Fisher Scientific) in calcium-free media for 15 min. Calcium release from the lysosome induced by 20 µM ML-SA1 and 200 µM glycylphenylalanine 2-naphthylamide was analyzed according to fluorescence intensity at 488 nm recorded with a confocal imaging system as described previously (Shen et al., 2012).

Both infected and uninfected WT and Ctsb^−/− BMDMs were lysed in assay buffer, and ATP and ADP content were measured using an ADP/ATP ratio assay kit according to the manufacturer’s protocols (MAK135; Sigma-Aldrich).

Total RNA was extracted from WT and Ctsb^−/− BMDMs, reverse transcribed into biotin-labeled cRNA with an Ambion WT Expression kit (Thermo Fisher Scientific), and hybridized to a Mouse Gene 2.0 ST GeneChip (Affymetrix). Gene expression data were normalized and transformed into log₂ transcript expression values with the robust multiarray average algorithm (version 6.6; Partek Genomics Suite; Irizarry et al., 2003). Differential expression was defined by application of a difference in expression of 0.5-fold (log₂ signal) between conditions. The dataset has been deposited in the GEO database under accession no. GSE79508.

Cell culture supernatants were collected at the indicated times, and lactate dehydrogenase activity was measured by using the Promega cytotoxicity kit according to the manufacturer’s protocols.

Cell culture supernatant or sera from mice were analyzed for cytokine and chemokine release using a 22-multiplex assay (EMD Millipore) following the manufacturer’s instructions.

Total RNA was isolated from BMDMs using TRIzol (Invitrogen). cDNA was reverse transcribed by using Superscript III reverse transcriptase (Invitrogen). Real-time quantitative PCR was performed on the ABI Prism 7500 sequence detection system (Applied Biosystems). Primer sequences are listed in Table S1.

Data are given as means ± SEM. Statistical analyses were performed using one-way ANOVA with multiple comparisons, two-tailed Student t tests, and log-rank tests. P-values ≤0.05 were considered significant.

Fig. S1 shows the expression of genes encoding TFEB, and its targets are differentially regulated in the absence of cathepsin B. Table S1 shows real-time quantitative PCR primer sequences.

We thank T. Reinheckel for the Ctsb^−/− mice, C. Pham for Ctsg^−/− and Nepr3^−/− mice, V.M. Dixit for the Aim2^−/− mice, D.M. Monack for GFP-expressing F. novicida, H. Fares for the immortalized Trpml1^−/− cells, and M. Barr, R. Johnson, A. Burton, and S. Frase (St. Jude Children’s Research Hospital) for technical assistance.

This work was supported by the US National Institutes of Health (AI101935, AI124346, AR056296, and CA163507 to T.-D. Kanneganti), the American Lebanese Syrian Associated Charities (grant to T.-D. Kanneganti), the European Research Council (281600 to M. Lamkanfi), the Fund for Scientific Research-Flanders (G030212N to M. Lamkanfi), and the National Health and Medical Research Council of Australia (R.G. Menzies Early Career Fellowship APP1091544 to S.M. Man).

The authors declare no competing financial interests.

Author contributions: X. Qi, S.M. Man, and T.-D. Kanneganti designed the study; X. Qi, S.M. Man, R.K.S. Malireddi, R. Karki, C. Lupfer, P. Gurung, C.S. Guy, and G. Neale performed experiments and analyzed the data; and X. Qi, S.M. Man, M. Lamkanfi, and T.-D. Kanneganti wrote the manuscript.