|Home | About | Journals | Submit | Contact Us | Français|
The bacterial virulence factors Shiga toxins (Stxs) are expressed by Shigella dysenteriae serotype 1 and certain Escherichia coli strains. Stxs are protein synthesis inhibitors and induce apoptosis in many cell types. Stxs induce apoptosis via prolonged ER stress signaling to activate both extrinsic and intrinsic pathways in human myeloid cells. Studies have shown that autophagy, a lysosome-dependent catabolic process, may be associated with activation of pro-survival or death processes. It is currently unknown if autophagy contributes to apoptosis or protects cells from Stxs. To study cellular responses to Stxs, we intoxicated toxin-sensitive cells (THP-1 and HK-2 cells), and toxin-resistant cells (primary human monocyte-derived macrophages) and examined toxin intracellular trafficking and autophagosome formation. Stxs translocated to different cell compartments in toxin-resistant versus toxin-sensitive cells. Confocal microscopy revealed autophagosome formation in both toxin-resistant and toxin-sensitive cells. Proteolytic cleavage of Atg5 and Beclin-1 play pivotal roles in switching non-cytotoxic autophagy to cell death signaling. We detected cleaved forms of Atg5 and Beclin-1 in Stx-treated toxin-sensitive cells, while cleaved caspases, calpains, Atg5 and Beclin-1 were not detected in toxin-resistant primary human monocytes and macrophages. These findings suggest that toxin sensitivity correlates with caspase and calpain activation, leading to Atg5 and Beclin-1 cleavage.
Despite efforts to improve hygienic conditions and regulate food and drinking water safety, the enteric pathogens, Shiga toxin (Stx)-producing Escherichia coli (STEC) and Shigella dysenteriae serotype 1 remain major public health concerns due to widespread outbreaks and the severity of diarrheal and extra-intestinal diseases they cause. The estimated incidence of food- and water-borne STEC infections in the U.S. is approximately 110,000 cases per year (Mead et al., 1999). Shigellosis remains a major public health problem in developing countries with an estimated incidence of 1.1 million fatal cases per year (Kotloff et al., 1999). Typical symptoms include nausea, vomiting, and bloody diarrhea, and these patients are at risk for the development of life-threatening complications involving acute renal failure (hemolytic uremic syndrome) and neurological abnormalities such as seizures, paralysis, blindness and death (Proulx and Tesh, 2007, Bale et al., 1980). There currently are no vaccines or effective interventional therapies to prevent or treat the diseases caused by Stx-producing bacteria. An improved understanding of host cell responses to Stxs will be necessary to devise effective means of treatment and vaccination.
The Stx family is comprised of Shiga toxin (Stx) expressed Shigella dysenteriae serotype 1, and closely related toxins designated Shiga toxin type 1 (Stx1) and Shiga toxin type 2 (Stx2) expressed by STEC. Stxs contain six protein subunits in an AB5 molecular configuration. Toxin monomeric A-subunits are potent protein synthesis inhibitors and the B-subunit proteins form homopentamers capable of binding to the neutral glycolipid globotriaosylceramide (Gb3) (Fraser et al., 2004, Lingwood, 2003, Johannes et al., 2010). Stxs appear capable of crossing the intact intestinal epithelial barrier via transcytotic or paracellular mechanisms, damage the colonic microvasculature, and then may associate with blood monocytes and neutrophils to circulate in the bloodstream, subsequently binding to susceptible cells in target organs via the toxin-binding glycolipid receptor (Hurley et al., 1999, te Loo et al., 2000, Brigotti et al., 2008, Malyukova et al., 2009). After endocytosis to an early endosome, Stxs travel via retrograde transport through the trans-Golgi network and Golgi apparatus to the endoplasmic reticulum (ER) (Sandvig et al., 1992a, Sandvig et al., 2010). The ER is the ultimate cellular compartment in the intracellular delivery of Stxs leading to cell death. The toxins act to block protein synthesis by acting on the 28S rRNA of ribosomes (Endo et al., 1988, Saxena et al., 1989). In addition to mediating protein synthesis inhibition, Stxs induce apoptosis via intrinsic and extrinsic pathways in many cell types including epithelial, endothelial and neuronal cells (reviewed in Tesh, 2010). Apoptosis is not the sole death process activated by tightly regulated mechanisms. In addition to apoptosis (type 1 cell death), an alternative cell death program called autophagic cell death (type 2 cell death) may facilitate or antagonize apoptosis, depending on the stimulus and cell type (Debnath et al., 2005, Gozuacik et al., 2007, Maiuri et al., 2007b).
The term autophagy describes the process of “self-eating”. In this process, a double layered membrane envelopes a volume of intracellular space, and the resultant autophagosome fuses with lysosomes. Originally, autophagy was characterized as a process allowing cells to survive in a nutrient deficient environment by consuming and recycling portions of the host cell intracellular machinery (Levine et al., 2004). More recently, substantive roles were defined for autophagy in cell death processes such as the induction of necrosis or apoptosis. Autophagy may coincide with apoptosis and promote cell death, or antagonize apoptosis to promote cell survival. Autophagy may also be induced by ER-stress following activation of the unfolded protein response (UPR) leading to cell death or survival. Thus, autophagy plays a critical role in the decision of cellular fate (Hoyer-Hansen et al., 2007, Moretti et al., 2007). While we have reported that Stxs activate apoptotic signaling pathways through the induction of ER stress and signaling through TRAIL and the apoptosis-inducing receptor DR5 (TRAIL-R2) in macrophage-like THP-1 cells, (Lee et al., 2008, Lee et al., 2010), few studies have been carried out to examine the capacity of the toxins to activate autophagy, and it is currently unknown if autophagy contributes to, or protects cells from, Stx-induced apoptosis. It is interesting to note, however, that signaling through DR5 and caspase-8 is also known to activate autophagy contributing to cell death (Park et al., 2009, Norman et al., 2010). The regulation of autophagy is complex, occurring through a variety of mechanisms including signaling through the PI3K/Akt/mTOR pathway, interaction of autophagy gene (Atg) products Atg5, Atg6, and Atg12 with the autophagosomal membrane, and activation of Bcl-2 proteins, p53, ARF, DAPk, and E2F1 (Periyasamy-Thandavan et al., 2009). Atg8/LC3 is considered a positive marker for autophagosome formation. The LC3B protein is lipidated and binds to phosphatidylethanolamine present in the inner and outer membranes of autophagosomes (Tanida et al., 2004). Unlike the Atg5/Atg12/Atg16L complex which is recycled following enclosure of the autophagosome, Atg8/LC3 remains associated with the autophagosome and is degraded along with the inner membrane (Tanida et al., 2004). While the precise mechanisms by which autophagy facilitates or inhibits apoptosis is not known, we show here that Stxs are routed to different intracellular compartments in toxin-sensitive and toxin-resistant human cells. Although autophagy is activated in both cell types, the activation of calpains and caspases in toxin-sensitive cells is associated with the cleavage of Beclin-1 and Atg5, thereby converting a pro-survival autophagic response to a cell death signal.
Previous studies revealed that human monocyte/macrophages and a variety of toxin-sensitive cell lines such as THP-1, HeLa and Vero cells, express the toxin-specific glycolipid receptor Gb3 on their surface (Ramegowda et al., 1996, van Setten et al., 1996, Falguières et al., 2001). Toxin binding to Gb3 is a requisite initial step leading to internalization and retrograde intracellular transport to reach the lumen of the ER through the Golgi apparatus. The cytotoxic effects of purified Stx2 for freshly isolated human peripheral blood monocytes (hMono), human monocyte-derived macrophages (hMDM), and monocytic, undifferentiated (UD-) and macrophage-like, differentiated (D-)THP-1 myeloid cells were assessed by incubation of cells with varying doses of the toxin for 24 h. As shown in Figures 1A and B, hMono and hMDM were highly refractory to cytotoxicity following exposure to Stx2, while UD- and D-THP-1 cells were relatively sensitive to the cytotoxic effects of Stx2. As we have previously reported using Stx1 (Harrison et al., 2005), differentiation of THP-1 cells to the adherent macrophage-like state is associated with increased resistance to Stx2, with CD50 values of < 1.0 ng/ml in UD-THP-1 and ~400 ng/ml in D-THP-1. We previously showed that Stx1 activates caspase-8 in THP-1 cells (Lee et al., 2007). We readily detected procaspase-8 cleavage in Stx2 treated D-THP-1 cells beginning 6 h after intoxication (Figure 1C). Caspase-8 activation required Stx enzymatic activity as incubation of D-THP-1 cells with Stx2A− toxoid or Stx2 B-subunits did not result in cleavage of procaspase-8 (Figure 1C). Resistance of hMono and hMDM to Stx1 or Stx2 was associated with failure to cleave procaspase-8 up to 24 h after intoxication (Figure 1C). Activation (cleavage) of the terminal executioner caspase, caspase-3, was also not observed in Stx1- or Stx2-treated hMono or hMDM (data not shown). Lipopolysaccharides (LPS) are potent activators of macrophage function, and treatment of THP-1 cells with Stxs and LPS resulted in increased cytotoxicity compared to treatment with Stxs alone (Harrison et al., 2005). However, treatment of primary hMDM with LPS + Stx1 for 24 h did not result in cell death (data not shown) or caspase-8 activation (Figure 1C). These data suggest that in contrast to the human myeloid leukemia cell line THP-1, primary hMono and hMDM are resistant to Stxs, and resistance correlates with failure to activate caspases-8 and -3.
The data presented above suggest that differential intracellular signaling induced by Stxs may contribute to different cell fates in toxin-sensitive versus toxin-resistant cells. We hypothesized that the induction of autophagy, a catabolic process involving the sequestration and routing of mis-folded proteins or damaged subcellular organelles to the lysosome-dependent degradation machinery, may play a critical role in altering intracellular toxin routing leading to proteolytic degradation of Stxs in toxin-resistant primary hMDM. In this case, autophagy would contribute to cell survival by eliminating the capacity of the toxins to stimulate apoptotic signaling. In contrast to this hypothesis, Sandvig et al. (1992b) used inhibitors to show that autophagy may be necessary for Stxs to induce cell lysis in toxin-sensitive Vero and MDCK cells. Therefore, we examined autophagy induction by Stxs in toxin-sensitive D-THP-1 cells and toxin-resistant hMono/hMDM by measuring two well-characterized indicators of autophagosome formation: lipidation of LC3B (LC3B-I → LC3B-II conversion) and formation of fluorescent punctate bodies of GFP-LC3. Levels of LC3B-I → LC3B-II conversion were increased in D-THP-1 cells treated with Stx1 over 0–16 h in serum-containing complete growth media (Figure 2A). Lipidated LC3B (LC3B-II) was detected 1 h after toxin exposure and remained elevated over the course of the experiment. We noted an increase in total expression of LC3B (LC3B-I + LC3B-II) following toxin exposure. Treatment of D-THP-1 cells with Stx1 B-subunits also induced autophagy with LC3B-II levels elevated 1 h after treatment and then gradually declining until 10 h after toxin treatment (Figure 2A). Total levels of LC3B were also increased in Stx1 B-subunit treated cells. Since autophagy induction occurs in response to nutrient starvation or growth factor withdrawal, as a positive control we compared levels of LC3B-I → LC3B-II conversion in response to the absence or presence of serum (Figure 2B, lanes labeled “starvation” and “serum +” respectively). LC3B-II lipidation was triggered by starvation only in the absence of serum in D-THP-1 cells, suggesting that starvation conditions do not contribute to autophagy induction by Stxs in the presence serum. As was the case with Stx1, LC3B-I → LC3B-II conversion was observed when the D-THP-1 cells were exposed to Stx2, Stx2A− or Stx2 B-subunits in the presence of serum, suggesting that toxin enzymatic activity is not required (Figure 2B). Purified Stx2 B-subunits reproducibly increased total levels of LC3B protein expression and activated LC3B-II lipidation to a significantly greater degree compared to A-subunit containing toxin preparations (Figure 2B, bar graph). UD-THP-1 cells are also highly sensitive to Stxs. Therefore, we compared autophagy induction in UD- and D-THP-1 cells. Compared to untreated control cells, elevated levels of LC3B-II were evident 2 h after Stx1 treatment in both UD-THP-1 and D-THP-1 cells (Figure 2C). Autophagy induction in Stx1 treated D-THP-1 cells was confirmed by confocal microscopy showing increased numbers of GFP-LC3 punctate aggregates as an indicator of autophagosome formation (Figure 2D). Finally, we examined autophagy induction in toxin-resistant hMono and hMDM. We did not detect LC3B-II lipidation in Stx1 or Stx2 treated hMono (Figure 2E). However, we detected LC3B-II lipidation in toxin-sensitive D-THP-1 cells after 4 h of toxin treatment, and in toxin-resistant hMDM treated with Stx1 for 24 h. LC3-GFP aggregates formed in hMDM treated with Stx1 (Figure 2F), but not in hMono (data not shown). Thus, Stxs induce autophagy in toxin-sensitive THP-1 cells, and do or do not induce autophagy in toxin-resistant cells hMDM and hMono, respectively. Autophagy induction by Stxs appears to be correlated with maturation of primary human blood monocytes to the adherent macrophage-like state.
In humans and experimental animals, renal tubules appear to be targets for destruction in acute renal failure caused by Stxs (Takeda et al., 1993, Tesh et al., 1993). Treatment of the human renal epithelial cell line HK-2 with Stx1 or Stx2 significantly diminished cell viability and induced apoptosis by triggering caspases and the release of pro-apoptotic factors (Sood et al., 2001, Nestoridi et al., 2005, Wilson et al., 2005). To extend our findings to non-myeloid cell types that may model in vivo pathogenesis, we examined the capacity of Stxs to induce autophagy in HK-2 cells. In the presence of serum, Stx2 or Stx2 B-subunit treatment activated autophagy with elevated levels of LC3B-II detected by Western blot analysis (Figure 3A). Levels of LC3B-II remained significantly elevated 3- to 4-fold over a 12 h treatment period (p < 0.05). Although Stx2 B-subunits failed to induce cell death, LC3B-I → LC3B-II conversion was increased 3-fold (p < 0.05) 3 h after Stx2 B-subunit treatment (Figure 3A, bar graph). To confirm autophagy induction, HK-2 cells incubated with serum-free media as a positive control also showed increased levels of LC3B-II by Western blotting analyses (data not shown). In the presence of serum, punctate LC3-GFP aggregates were counted and showed significant increases in the number of LC3B-II dots (> 15 dots per cell; p<0.001) in HK-2 cells treated for 6 h with Stx2 (100 pg/ml) or chloroquine diphosphate (autophagy positive control). Numbers of LC3B-II dots in Stx1-treated cells were modestly increased at 6 h, although we did not detect a significant difference compared to numbers of LC3B-II dots in untreated cells (Figure 3B). The number of aggregates formed in Stx1- and Stx2-treated HK-2 cells transfected with GFP-LC3B was significantly greater (P < 0.001) than the number of dots formed in HK-2 cells transfected with GFP-LC3BΔ(G120A), a point mutation in LC3B which prevents autophagosome formation and shows a diffuse cytosolic pattern of fluorescence (Figure 3B). Finally, we tested the role of autophagy in the cytotoxic effect of Stx2 by treatment of HK-2 cells with or without the autophagy inhibitor 3-methyladenine (3-MA) (Seglen and Gordon, 1982). Exposure of HK-2 cells to Stx2 showed a dose-dependent decrease in percentage of cell viability (p < 0.001) compared to cells treated with the enzymatic mutant Stx2A− or Stx2 B-subunits (Figure 3C), indicating that Stx2 enzymatic activity is required to induce cell death. Treatment of HK-2 cells with 3-MA alone had no effect on cell viability. However, the addition of 3-MA 40 min prior to toxin challenge significantly protected HK-2 cells against cell death caused by Stx2 (p < 0.01). These data suggest that Stx2 induces autophagy in toxin-sensitive HK-2 cells. Treatment of cells with an autophagy inhibitor protects cells from Stx2 cytotoxicity, suggesting that autophagy contributes to cell death in this cell type. Finally, Stx2 enzymatic activity, while required for cytotoxicity, is not necessary for autophagosome formation.
Fluorescently labeled Stx1 B-subunits have been extensively employed to characterize Stx intracellular routing to different cellular compartments (Falguières et al., 2001; Haicheur et al., 2003). To determine whether intracellular transport of Stx1 or Stx2 is required to activate autophagy, we investigated autophagosome formation by measuring the number of GFP-LC3B dots (green) formed during trafficking of Stx1 B-subunit- or Stx2 B-subunit-Alexa594 (red) into the cells at various time points 24 h after the transduction of GFP-LC3B or mutant GFP-LC3BΔ(G120A). In the absence of toxin, the formation of GFP-LC3B-labeled structures (GFP-LC3B dots) was rarely observed in toxin-resistant hMDM, while GFP-LC3B fluorescence was significantly increased in hMDM as Stx1 B-subunit-Alexa594 was internalized over 240 min (Figure 4A). Over time, GFP-LC3B dots appeared to migrate from discrete peripheral locations to a perinuclear region which co-localized with Stx2 B-subunits (yellow fluorescence in merged images). While GFP-LC3B dots were detected in toxin-sensitive D-THP-1 and HK-2 cells during Stx2 B subunit-Alexa594 trafficking, we failed to detect extensive co-localization of toxin and autophagosomal fluorescence, or a perinuclear pattern of fluorescence, in toxin-sensitive cells which was apparent in toxin-resistant hMDM (Figure 4B). In GFP-LC3BΔ(G120A)-transduced cells, fluorescence was diffuse rather than forming punctate dots during trafficking of the toxin (Figure 4B). These experiments indicate that autophagy is induced coincident with intracellular trafficking of Stxs. However, the patterns of toxin- and autophagosome-associated fluorescence differ in toxin-sensitive and toxin-resistant cells, so that toxin-containing autophagosomes appeared to selectively form in toxin-resistant cells.
In many toxin-sensitive cell types, Stxs undergo retrograde transport through the Golgi apparatus to reach the ER, where a fragment of the toxin A-subunit retrotranslocates into the cytoplasm (Sandvig et al., 2010) . Although human macrophages and dendritic cells express Gb3, Stx B-subunits were routed to lysosomes to undergo proteolytic degradation (Falguières et al., 2001, Haicheur et al., 2003). Given that hMono, hMDM and THP-1 cells showed different susceptibilities to Stx1 or Stx2 (Figure 1), we used confocal microscopy to examine the intracellular trafficking of Stx1 or Stx2 B-subunits conjugated to Alexa 488 dye (green) with lysosome (red)- or ER (blue)-specific fluorescent markers. In D-THP-1 cells, we observed retrograde trafficking of Stx1 and Stx2 B-subunits to the ER as indicated by the light blue fluorescence in merged images with optimal fluorescence detected 90 min after intoxication (Figure 5A). We did not detect toxin routing to lysosomes as indicated by the maintenance of punctate red fluorescence and lack of yellow fluorescence in merged images at all time points. In primary hMDM, within 30 min of intoxication, we noted Stx2B subunit-Alexa488 translocation to lysosomes (Figure 5B) and B-subunits remained associated with lysosomes up to 90 min after intoxication. Routing of Stx2B subunit-Alexa488 into the ER was not observed in hMDM (data not shown). We also examined toxin trafficking in another toxin-sensitive cell type, HK-2 cells. Stx2 B-subunits translocate to the ER as shown by the light blue fluorescence in the merged image (Figure 5B, right panel). Stx2B subunit-Alexa 488 (green) was not trafficked to lysosomes (red) in toxin-sensitive HK-2 cells (Figure 5B, left panels). Thus, using fluorescent Stx B-subunits, we show that the toxin co-localizes with autophagosomes and lysosomes in toxin-resistant cells, but co-localizes with the ER in toxin-sensitive cells.
Atg5 and Beclin-1 are essential components in the initiation of autophagosome formation. In addition, Atg5 and Beclin-1 may be cleaved by calpains and caspases, respectively, to generate proteins that amplify apoptosis by external stimuli in susceptible cells (Yousefi et al., 2006, Wirawan et al., 2010). We showed that Stxs activate calpains and caspase-8 and -3 in toxin-sensitive THP-1 cells (Lee et al., 2008). To determine if Stxs induce Atg5 and Beclin-1 cleavage, D-THP-1 cells were treated with Stx1 or Stx1 B-subunits for various time points and whole cell lysates were examined by SDS-PAGE and western immunoblotting with anti-Atg5 or anti–Beclin-1 antibodies. A significant increase in the amount of the 24 kDa Atg5 cleaved fragment was detected up to 12 h after Stx1 treatment of D-THP-1 cells, after which levels of full-length and cleaved Atg5 were decreased (Figure 6A). In contrast, treatment of toxin-resistant hMDM with Stx1 or Stx2 failed to trigger Atg5 cleavage (Figure 6B). Cleavage of Beclin-1 was also clearly evident in Stx1-treated D-THP-1 cells 1 and 4 h after intoxication. However, Beclin-1 fragments were not detected in Stx1- or Stx2-treated toxin-resistant hMDM 0 – 24 h after intoxication (Figure 6C). To examine the role of enzymatic activity in Atg5 and Beclin-1 cleavage, levels of the cleaved proteins in Stx1 B-subunit-treated D-THP-1 cells were examined over a time course of 16 h. No Atg5 or Beclin-1 cleavage products were detected in Stx1 B-subunit-treated cells (Figure 6D, upper panel). Stxs also induced Atg5 and Beclin-1 cleavage in a toxin enzymatic activity-dependent manner in UD-THP-1 cells (Figure 6D, lower panel).
We employed another toxin-sensitive cell line, HK-2 cells, to verify these results. We previously showed that Stxs induced calpain activation (cleavage) in D-THP-1 cells and calpains contributed to apoptosis (Lee et al., 2008, Lee et al., 2010). Treatment of HK-2 cells with Stx-2 led to a dose-dependent increase in levels of 28 kDa cleaved calpain fragments 24 h after intoxication (Figure 7A). In contrast, using toxin-resistant primary hMDM, calpain cleavage products were not detected following treatment with Stx1 or Stx2 for 24 h. Thus, neither calpains nor Atg5 (Figure 6B) were activated in toxin-resistant primary hMDM. We compared the ability of various doses of Stx1 and Stx2 to cleave calpains and Atg5 in HK-2 cells. Both toxins activated calpains with cleavage products detected 16 h after treatment with toxin doses as low as 0.01 ng/ml (Figure 7B, upper panel). At 16 h, the vast majority of immunoreactive Atg5 was associated with Atg12, suggesting that autophagosome formation had occurred. Low levels of free cleaved Atg5 were detected, with Stx2 appearing to be slightly more effective in inducing Atg5 cleavage (Figure 7B, upper panel). Treatment of HK-2 cells with Stx2A− toxoid failed to generate Atg5 cleavage fragments (Figure 7B, lower panel). Since optimal Atg5 cleavage reproducibly occurred using 0.1 ng/ml of Stx2 (Figure 7B), we used this toxin dose in a 24 h time course experiment to examine Atg5 and Beclin-1 cleavage in HK-2 cells. As shown in Figure 7C, Atg5 and Beclin-1 cleavage was induced by Stx2 treatment of HK-2 cells. Maximal Beclin-1 cleavage (4 h) slightly preceded maximal Atg5 cleavage (6 h). It should be noted that at longer time points, a significant fraction of HK-2 cells treated with Stx2 have undergone apoptosis. Neither Stx1 nor Stx2 treatment of hMDM for 24 h resulted in cleavage of Beclin-1, while Stx2, but not Stx1, appeared to induce modest Beclin-1 cleavage in HK-2 cells (Figure 7D). Beclin-1 was only modestly processed to generate the cleaved forms in HK-2 cells under 24 h serum starvation conditions without treatment with Stx1 or Stx2 (Figure 7D, lane C*). In summary, Stxs differentially signal the autophagic pathway in toxin-sensitive versus toxin-resistant cells. The data suggest that calpains and caspases may process Atg5 and Beclin-1 in toxin-sensitive cells to convert a pro-survival autophagic response to an apoptotic response. Failure of Stxs to activate caspase-8 and -3, calpains, Atg5 and Beclin-1 correlates with the toxin-resistant phenotype.
Stxs are a family of proteins capable of binding to host cells expressing the membrane glycolipid Gb3. The toxins are internalized and routed via the retrograde transport process to the lumen of the ER. During transport, the A-subunit undergoes furin-dependent cleavage to produce the A1-fragment, and this fragment is translocated across the ER membrane utilizing the ER-associated degradation (ERAD) machinery. Once in the cytoplasm, the A1-fragment escapes routing to the proteasome and inhibits protein synthesis by the catalytic removal of a single adenine residue from the 28S rRNA (reviewed in Johannes and Römer, 2010; Sandvig et al., 2010). In addition to mediating protein synthesis inhibition, Stxs activate multiple signaling pathways including: i) signaling through src kinases, PI3K, Akt and mTOR (Katagiri et al., 1999, Zanchi et al., 2008, Cherla et al., 2009); ii) activation of MAPKs (Foster et al., 2002, Smith et al., 2003) and their upstream kinases PKR, MK-2 and ZAK (Gray et al., 2008, Jandhyala et al., 2008, Saenz et al., 2009); iii) activation of NF-κB and AP-1 transcription factors (Sakiri et al., 1998, Zoja et al., 2002); and iv) prolonged activation of the ER stress response (Lee et al., 2009, Lee et al., 2010). Signaling through these pathways has been shown to be important in toxin internalization and routing, induction of cytokine/chemokine expression, and the induction of apoptosis. Primary human monocytes and the human myeloid cell line THP-1 express Gb3. However, primary human monocytes/macrophages are relatively resistant to the cytotoxic action of Stxs (Ramegowda et al., 1996, Falguières et al., 2001). Internalized Stx B-subunits expressing the KDEL ER-retrieval sequence were not glycosylated by ER-resident enzymes and did not interact with the ER resident chaperone BiP in human macrophages (Falguières et al., 2001, Falguières and Johannes, 2006). Rather, immunofluorescence microscopy showed that Stx B-subunits sequentially associated with EEA1-expressing early endosomes and LAMP2-expressing lysosomes. Within 2 h of internalization, 55% of macrophage-associated radiolabeled Stx B-subunits were degraded and detected in TCA-soluble fractions.
Data reported here confirm and extend these earlier findings. THP-1 cells are sensitive, while primary human monocytes/macrophages are relatively resistant, to the cytotoxic action of purified Stx2. THP-1 cell sensitivity to Stxs is associated with the rapid activation of caspase-8 and caspase-3, while we failed to detect caspase activation in primary human monocytes/macrophages. Confocal immunofluorescence microscopy showed that toxin B-subunits co-localized with an ER-specific marker, while B-subunits rapidly routed to lysosomes in primary cells. Thus, the induction of apoptosis by Stxs correlates with retrograde transport of the toxins to the ER. The precise mechanisms contributing to differences in intracellular routing of Shiga toxins remain to be characterized. It is known that Gb3 structure is heterogeneous, with differences characterized in fatty acid chain length, and degree of carbon-carbon bond saturation and hydroxylation (Distler et al., 2009). Expression of long-chain Gb3 molecules with a single unsaturated bond (C22:1) was most efficient in interacting with Stx B-subunits to mediate localized negative membrane curvature and internalization of B-subunits into membrane invaginations (Römer et al., 2007). Finally, the association of Gb3 with detergent-resistant membrane microdomains or “lipid rafts” appears to be crucial for internalization and routing of Stxs to the ER. In primary human monocytes, Gb3 does not coalesce into lipid rafts (Falguières et al., 2001). Whether differences in Gb3 isoforms or Gb3 sequestration in lipid rafts is associated with toxin retrograde transport in THP-1 myeloid leukemia cells will require additional study.
Autophagy (“self-eating”) is a catabolic process in which macromolecules within the cytoplasm, or damaged cell organelles, may be engulfed within a vacuole containing a characteristic double membrane called the autophagosome. Autophagosomal contents are subsequently degraded by fusion with lysosomes (Rabinowitz et al., 2010, Mehrpour et al., 2010). Autophagy was initially characterized as a process important in maintenance of cellular homeostasis and survival. For example, in the face of nutrient starvation, autophagy may generate metabolic substrates necessary for cell survival. However, autophagy may also contribute to cell death (autophagic or type II cell death) in response to certain stressors or under conditions in which apoptosis is inhibited (Maiuri et al., 2007b, Moretti et al., 2007). The precise relationship among signaling pathways leading to: i) apoptosis and cell death; ii) autophagy and cell survival; or iii) autophagy and cell death remain to be characterized. Sandvig and van Deurs (1992b) showed that treatment of MDCK or Vero cells with autophagy inhibitors, such as 3-methyladenine, protected cells from apoptosis induced by the protein synthesis inhibitors Stx and ricin. This protective effect occurred even in the face of protein synthesis inhibition, suggesting that the two phenomena are independent. Here we show that autophagy is induced in cells that survive intoxication (primary human monocyte-derived macrophages) or in cells induced to undergo apoptosis (THP-1 cells and HK-2 cells) when challenged with Stxs. Thus, the induction of autophagy by Stxs is not, in and of itself, sufficient to trigger cell death in primary human macrophages. Treatment of primary human monocytes with Stxs did not induce autophagy or cell death, suggesting that primary monocytes may become competent for autophagy induction upon maturation to the adherent macrophage state. In toxin-sensitive THP-1 and HK-2 cells, toxin enzymatic activity was necessary for apoptosis induction, but not for induction of autophagy, suggesting that protein synthesis inhibition and apoptosis inducing activities of the toxins may be dissociated from signaling for autophagosome formation. Our fluorescence confocal microscopy studies suggest that internalization and intracellular routing of toxin B-subunits to the ER is associated with autophagy. Additional experiments will be required to determine whether autophagy may be induced by B-subunit trafficking to intermediate intracellular compartments such as endosomes, trans-Golgi network, or Golgi apparatus. However, the capacity of Stx B-subunits to trigger autophagy in primary macrophages in which the toxin is routed to lysosomes suggests that signals for autophagy may be activated by toxin binding or from early/recycling endosomes prior to entry of the toxins into the degradative environment of lysosomes. Finally, the kinetics of autophagosome formation as measured by Atg8/LC3B lipidation were rapid and transient in cells treated with purified Stx2 B-subunits, but prolonged in cells treated with the holotoxin, suggesting that the extent and duration of signaling for autophagy may contribute to the cellular response to Stxs. Interestingly, we noted modest increases in LC3B expression in cells treated with Stxs. In this regard, Norman et al. (2010) have shown that staurosporine treatment of HeLa cells increased both LC3B-I → LC3B-II conversion and LC3B protein expression, suggesting that autophagic signaling leading to apoptosis may involve up-regulation of atg8 gene expression
ER stress is a major activator of autophagy, mediating the degradation of mis-folded proteins and damaged ER membranes. However, signaling mechanisms leading to autophagy may be cell type- and stimulus-dependent. For example, Kouroku et al. (2007) showed that ER stress signaling triggered by thapsigargin or expanded polyglutamine repeats (polyQ72) led to PERK and eIF2α activation, up-regulated atg12 expression, and increased LC3B-I→LC3B-II conversion. Ogata et al. (2006) showed that tunicamycin and thapsigargin induced autophagosome formation in neuroblastoma cells via an IRE1/JNK-dependent signaling mechanism. Autophagy was not induced in PERK-deficient or ATF6 knockdown cells, suggesting that ER stress signaling through IRE1 is primarily responsible for autophagy induction in this cell type. We have shown that Stxs activate the ER stress response in toxin-sensitive THP-1 cells, although we detected cell maturation-dependent differences in the activation profiles. Treatment of undifferentiated THP-1 cells led to activation of all proximal ER-membrane localized sensors of ER stress (PERK, IRE-1 and ATF6) and the rapid induction of apoptosis. Differentiated THP-1 cells responded to the toxins with the delayed onset of apoptosis in which ER stress signaling was mediated via activation of PERK and IRE-1, but not ATF6 (Lee et al., 2009). Interestingly, Stxs failed to activate IRE1 and PERK in toxin-resistant hMDM (unpublished data) perhaps because the toxins are not routed to the ER and ER stress is not activated in these cells. Additional cell signaling mechanisms activated in Stx-treated THP-1 cells include the release of Ca2+ from intracellular stores and the activation of the calpain proteases. Calpain inhibitors partially blocked apoptosis induced by Stx1 (Lee et al., 2008). We have also reported that Stx1 treatment of differentiated THP-1 cells led to the increased expression of the apoptosis-inducing factor TRAIL and its receptor DR5 (TRAIL-R2). Signaling through TRAIL-DR5 engagement and the activation of calpains may contribute to the cleavage of procaspase-8 that we detected in Stx-treated THP-1 cells (Lee et al., 2010).
Autophagosome formation is regulated by the autophagy-related genes (atg). Key regulators in autophagy induction are Atg5 and Atg6/Beclin-1 (Bcl2-interacting protein-1). Atg5 is a critical component in the formation of a ubiquitin-like conjugation system necessary for autophagosome formation (Mizushima et al., 1998). In spontaneous neutrophil apoptosis and apoptosis in HL-60 cells induced by staurosporine, Atg5 is cleaved from a 33 kDa form to a 24 kDa fragment (Yousefi et al., 2006). Atg5 is cleaved at residue Thr193 by calpain-1 and -2. Cleaved Atg5 translocates to the mitochondria, followed by cytochrome c release and procaspase-3 and PARP cleavage. Beclin-1 serves as an activation platform for the assembly of a multi-protein phosphatidylinositol-3-kinase class 3 (PI3KC3) complex required for autophagosome formation (Kihara et al. 2001; Wirawan et al. 2010). Beclin-1 possesses a BH3 domain capable of binding the pro-survival Bcl-2 family members Bcl-2 or Bcl-XL. Autophagy is inhibited when Beclin-1 is in the Bcl-2-bound state (Pattingre et al., 2005, Maiuri et al., 2007a). Recently, it was shown that prolonged growth factor deprivation can lead to apoptosis which is preceded by activation of autophagy. Apoptosis was associated with cleavage of Beclin-1 and the PI3KC3 complex, which coincided with caspase activation. Beclin-1 was shown to be a substrate of caspase-3, -7, and -8, which act at two conserved caspase cleavage sites to produce 37 kDa and 35 kDa Beclin-1 fragments. Not only did cleaved Beclin-1 fail to induce autophagy, a Beclin-1 fragment was shown to associate with mitochondrial membranes where it induced the release of cytochrome c and other proapoptotic factors into the cytoplasm (Wirawan et al. 2010). Thus cleavage of Atg5 and Beclin-1 by calpains and caspases, respectively, may represent a critical switch in the determination of induction of autophagy versus apoptosis in Stx-intoxicated cells. We show here that in toxin-resistant primary human macrophages, Stxs are translocated to lysosomes and autophagy is induced in the absence of calpain and caspase activation, and Atg5 and Beclin-1 cleavage (Figure 8, left panel). In toxin-sensitive THP-1 and HK-2 cells, Stxs are translocated to the ER, the ER stress response is activated, and autophagy is induced in association with the activation of calpains and caspase-8 and -3, and the cleavage of Atg5 and Beclin-1 (Figure 8, right panel). Thus, sensitivity to the cytotoxic effects of Stxs correlates with differential intracellular routing and activation of signaling pathways that convert autophagy from a cell survival response to a programmed cell death response.
Peripheral blood Leukopaks from normal healthy donors were obtained from the Gulf Coast Regional Blood Center (Houston, TX, USA). Subsequently, mononuclear cells were immediately diluted (1:1) in 2% fetal bovine serum (FBS) containing ice-cold phosphate buffered saline (PBS). Twenty-five milliliters of samples were layered onto tubes containing 15 ml Ficoll-Paque Plus, density 1.077 g/ml (GE Health Care, Piscataway, NJ) and centrifuged at 200 × g 30 min at 4°C. Supernatants were removed and mononuclear cells at the interfaces were collected, washed once, and resuspended in suitable volumes of PBS + 2% FBS to make 1:100 dilutions. Red blood cells were hypotonically lysed and cells were counted using an automated cell counter (Cellometer, Nexcellom Biosciences, Ltd., Lawrence, MA). Further purification was performed using human monocyte enrichment kits with magnetic negative selection in an automated Robosep device (Stemcell Technology, Vancouver, Canada). Purity of CD14+ monocyte fractions were >97% as assessed by flow cytometry. Purified fresh monocytes were cultured immediately.
Freshly isolated human monocytes were cultured in 6- or 96-well tissue culture plates (Falcon, Becton Dickinson, Franklin Lakes, NJ) in RPMI-1640 medium (Gibco-BRL, Grand Island, NY) containing 1% (vol/vol) penicillin/streptomycin, and 10% (vol/vol) human serum (Invitrogen, Carlsbad, CA) The cells were maintained at 37°C in a 5% CO2/95% air atmosphere in a humidified incubator. The medium was replaced every 3 days and adherent cells were used for experiments within 11 days of plating.
The human myelogenous leukemia cell line THP-1 (TIBS-202; American Type Culture Collection, Manassas, VA) was cultured in RPMI-1640 medium containing 10% fetal bovine serum (FBS; Hyclone Laboratories, Logan, UT), penicillin (100 U/ml), and streptomycin (100 µg/ml) at 37°C in 5% CO2 in a humidified incubator. Cells maintained under these conditions were considered undifferentiated, monocytic cells. Monocytic THP-1 cells (1.0 × 106 cells/ml) were differentiated to the adherent macrophage-like state with phorbol 12-myristate 13-acetate (PMA; Sigma Chemical Co., St. Louis, MO) at a concentration of 50 ng/ml for 48 h. Plastic-adherent cells were washed three times with cold, sterile Dulbecco’s phosphate-buffered saline (Sigma) and then incubated with fresh medium lacking PMA but containing 10% FBS, penicillin (100 U/ml), and streptomycin (100 µg/ml). The medium was changed every 24 hours for the next 3 days. Experiments were performed on the fourth day after PMA removal. The human kidney proximal tubule epithelial cell line HK-2 (CRL-2190) was purchased from the American Type Culture Collection. HK-2 cells were maintained in Keratinocyte-Serum Free Media (K-SFM) (Invitrogen, Carlsbad, CA) supplemented with bovine pituitary extract (BPE), human recombinant epidermal growth factor (EGF), penicillin (100 U/ml) and streptomycin (100 µg/ml) at 37°C in humidified 5% CO2.
Stx1 was expressed from Escherichia coli DH5α(pCKS112), a recombinant strain harboring plasmid pCKS112 encoding the stx1 operon under control of a thermoinducible promoter (Tesh et al., 1993). Cells were lysed and periplasmic extracts were subjected to sequential ion-exchange and chromatofocusing chromatography. Purity of toxins was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting analysis using anti-Stx1 specific antibodies. Prior to use, Stx1 preparations were shown to contain < 0.1 ng of endotoxin/ml by use of Limulus amoebocyte lysate assay (Associates of Cape Cod, East Falmouth, MA). Purified Stx1A− toxoid, a holotoxin with double point mutations (E167Q/R170L) in the gene encoding the A-subunit, was a kind gift from Dr. Shinji Yamasaki, Osaka Prefecture University, Osaka, Japan. These site-directed mutations reduce toxin N-glycosidase activity by approximately 5-logs (Ohmura et al., 1993). Purified pentameric Stx1 B-subunits were the kind gift of Dr. Cheleste Thorpe, Tufts University School of Medicine, Boston, MA. Recombinant Stx2 holotoxin, Stx2A− toxoid (Y77S/E167Q/R170L), and purified Stx2 B-subunits were obtained from the NIAID, NIH Biodefense and Emerging Infections Research Resource Repository (BEI Resources, Manassas, VA). Lipopolysaccharides (LPS) derived from Escherichia coli O111:B4 were purchased from Sigma Chemicals (St. Louis, MO).
Antibodies used in this study were directed against human calpain1/2 (Calbiochem, San Diego, CA), caspase-8, LC3B-I and LC3B-II, pan-actin, Atg5 and Beclin-1 (Cell Signaling Technologies, Beverly, MA). Horseradish peroxidase conjugated goat anti-rabbit IgG and rabbit anti-mouse IgG secondary antibodies were also obtained from Cell Signaling Technologies.
Intracellular trafficking of Stxs into human monocyte-derived macrophages, THP-1, or HK-2 cells was determined using purified Stx1 or Stx2 B-subunits conjugated to fluorescent tags (Alexa488, Alexa594). Thirty micrograms of purified Stx1 or Stx2 B-subunits were labeled with Alexa Fluor-488 or -594 dyes (Molecular Probes, Inc., Invitrogen, Eugene, OR) as described in the manufacturer’s protocol. Briefly, monocyte-derived human macrophages, differentiated THP-1, or HK-2 cells (1.0 × 105 cells/well) were seeded overnight into four-well Lab-Tek chambered borosilicate coverglass slides (Nalge-Nunc International, Rochester, NY). The cells were washed twice in complete RPMI-1640 growth media or K-SFM supplemented with BPE, EGF and penicillin (100 U/ml) before further staining for 30 min at 37°C with cell-permeant lysosome- or endoplasmic reticulum-specific dyes (100 nM Lyso-Tracker or 60 nM ER-Tracker Live Cell Staining dyes; Molecular Probes, Inc.) Complete growth media containing 100 ng/ml or 400 ng/ml Stx1B - or Stx2B-Alexa 488 or 594 were added to the cell monolayers. Cells were washed extensively and then imaged over the next 5 to 90 min. Single confocal optical sections through the middle of the majority of cells in a field of view were simultaneously taken for red, blue or green emission channels using a Stallion Digital Imaging Station (Carl Zeiss Microscopes, Gottingen, Germany) and SlideBook 4.2 image software (Olympus America, Inc., Center Valley, PA). Images shown are representative of at least two independent experiments. All data within each experiment were collected at identical settings.
Autophagy was detected by measuring the aggregation of LC3B protein coupled to green fluorescence protein (GFP) using Premo Autophagy Sensor kits (Invitrogen, Carlsbad, CA). The LC3B protein plays a critical role in autophagosome formation and is considered a positive biomarker for autophagy. Briefly, human monocyte-derived macrophages, differentiated THP-1 and HK-2 cells were plated at 5.0 × 104 cells/well. Cells were then transduced with non-replicating baculoviral vectors expressing LC3B-GFP or LC3BΔ(G120A)-GFP, a point mutation in the gene encoding LC3B which prevents cleavage and subsequent lipidation during normal autophagosome formation resulting in GFP expression that remains cytosolic and diffuse. Twenty-four hours after transduction, the cells were untreated or incubated Stx1 or Stx2 (100 pg/ml or 400 ng/ml), or with 30 µM chloroquine diphosphate as a positive control. The appearance of punctate LC3B-GFP aggregates (“LC3B-dots”) was observed and quantified as LC3B dots per cell. Autophagy was also detected by Western blotting using anti-human LC3B reactive primary antibodies as described below.
Human monocytes, human monocyte-derived macrophages, undifferentiated (monocytic) and differentiated (macrophage-like) THP-1 cells (5.0 × 104 cells/well) were seeded in 96-well microtiter plates prior to treatment with Stx1 (100 pg/ml or 400 ng/ml), Stx2 (0–500 ng/ml), Stx2A− (100 pg/ml), or Stx2 B-subunits (100 pg/ml). Adherent HK-2 cells (5.0 × 104 cells per well) were plated in 96-well plates and grown to 80% confluency at 37°C in a humidified incubator. HK-2 cells were then stimulated with various concentrations of Stx2 ranging from 10 pg/ml to 1.0 ng/ml. In some experiments, the autophagy inhibitor 3-methyladenine (3-MA; Sigma) was added to the cells for 40 min prior to stimulation with Stx1 and incubation at 37°C in humidified 5% CO2. In some experiments, cells were treated with Stx1 or Stx1A− for 24 h in RPMI-1640 medium containing 0.5% FBS (serum starvation conditions) in the presence or absence of 3-MA. Control experiments with THP-1 cells treated with 0.5% FBS or 3-MA alone were included with each experiment. Cytotoxicity was determined by colorimetric assay (Cory et al., 1991) using the tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium inert salt; (MTS, Promega, Madison, WI). MTS (50 µl/5000 cells) was added to each well and incubation continued for 2 h at 37°C in 5% CO2. Optical density was recorded with an automated microtiter plate reader at absorbance of 490 nm (Dynatech MR5000; Molecular Dynamics, Chantilly, VA). The percentage of cell death was determined using the following equation: percentage of cell death = [(average OD490 of treated cells − average OD490 of control cells) ÷ average OD490 of control cells] × 100. Background absorbance at 630 nm measured with untreated cells was subtracted from each sample reading. The reference wavelength 630 nm was used to subtract background contributed by excess cell debris and other nonspecific absorbance.
Eighteen hours prior to stimulation, differentiated THP-1 cells (5 × 106 cells/well) were washed twice in cold Dulbecco’s PBS, and RPMI-1640 containing 0.5% FBS. Cells were stimulated with Stx1 (400 ng/ml), Stx1A− (400 ng/ml), Stx1 B-subunit (800 ng/ml), or thapsigargin (10 µm) for 0–24 h. Cells were harvested and lysed with modified radioimmunoprecipitation assay (RIPA) buffer as previously described (Foster et al., 2002). Protein concentrations of each extract were determined using the Micro BCA protein Assay Kit (Pierce, Rockford, IL). Equal amounts of proteins (70–100 µg/lane) were separated by 8% or 12% Tris-glycine SDS–PAGE and transferred to nitrocellulose membranes. Membranes were blocked with 5% non-fat milk prepared with TBST (20 mM Tris [pH 7.6], 137 mM NaCl, 0.1% Tween 20). Membranes were incubated with primary antibodies at 4°C for 24 h. After washing, the membranes were incubated with horseradish peroxidase-labeled secondary antibodies for 2 h at room temperature. Bands were visualized using the Western Lightning Chemiluminescence System (NEN-Perkins Elmer, Boston, MA). Data shown are from at least three independent experiments. Relative protein expression levels were measured using NIH image J software.
Data are reported as means ± standard errors of the mean (SEM) for at least three independent experiments. Statistical analyses of data were performed with Graphpad Prism software (Graphpad, San Diego, CA). Depending on the assay, P values of < 0.05 or < 0.001 were considered significant.
This work was supported by U.S. Public Health Service grant RO1 AI34530 from the National Institute of Allergy and Infectious Diseases, National Institutes of Health (to V.L.T.). We thank Drs. Cheleste Thorpe and Shinji Yamasaki for gifts of reagents necessary to carry out these experiments, Dr. Rola Barhoumi for assistance with fluorescence microscopy, Mr. Jonathan Lei for assistance with primary human blood cell isolation, and Ms. Gay Pridgeon for artwork.