Metformin Suppresses Development of the Echinococcus multilocularis Larval Stage by Targeting the TOR Pathway

Alveolar echinococcosis (AE) is a severe disease caused by the larval stage of the tapeworm Echinococcus multilocularis. Current chemotherapeutic treatment options based on benzimidazoles are of limited effectiveness, which underlines the need to find new antiechinococcosis drugs. Metformin is an antihyperglycemic and antiproliferative agent that shows activity against the related parasite Echinococcus granulosus. Hence, we assessed the in vitro and in vivo effects of the drug on E. multilocularis.

ABSTRACT

Alveolar echinococcosis (AE) is a severe disease caused by the larval stage of the tapeworm Echinococcus multilocularis. Current chemotherapeutic treatment options based on benzimidazoles are of limited effectiveness, which underlines the need to find new antiechinococcosis drugs. Metformin is an antihyperglycemic and antiproliferative agent that shows activity against the related parasite Echinococcus granulosus. Hence, we assessed the in vitro and in vivo effects of the drug on E. multilocularis. Metformin exerted significant dose-dependent killing effects on in vitro cultured parasite stem cells and protoscoleces and significantly reduced the dedifferentiation of protoscoleces into metacestodes. Likewise, oral administration of metformin (50 mg/kg of body weight/day for 8 weeks) was effective in achieving a significant reduction of parasite weight in a secondary murine AE model. Our results revealed mitochondrial membrane depolarization, activation of Em-AMPK, suppression of Em-TOR, and overexpression of Em-Atg8 in the germinal layer of metformin-treated metacestode vesicles. The opposite effects on the level of active Em-TOR in response to exogenous insulin and rapamycin suggest that Em-TOR is part of the parasite’s insulin signaling pathway. Finally, the presence of the key lysosomal pathway components, through which metformin reportedly acts, was confirmed in the parasite by in silico assays. Taken together, these results introduce metformin as a promising candidate for AE treatment. Although our study highlights the importance of those direct mechanisms by which metformin reduces parasite viability, it does not necessarily preclude any additional systemic effects of the drug that might reduce parasite growth in vivo.

INTRODUCTION

Alveolar echinococcosis (AE) is a zoonosis caused by the cestode Echinococcus multilocularis, which is endemic in the Northern hemisphere (1, 2). This life-threatening disease is of increasing public health concern, especially in Europe, China, and Canada, where the parasite has become more prevalent (3, 4). The parasite is predominantly perpetuated in a sylvatic cycle, with wild carnivores (mainly foxes) as definitive hosts and small mammals (usually rodents) as intermediate hosts. Humans can accidentally acquire the infection through ingestion of parasite eggs shed in the feces of a definite host. Once an individual is infected, E. multilocularis forms metacestodes, which grow aggressively and infiltrate the host tissue (primarily the liver), thus causing AE (5). E. multilocularis metacestodes reproduce asexually by exogenous formation and budding of daughter vesicles. They are composed of a complex germinal layer which contains 20% to 25% of stem cells (called germinative cells), the only proliferating cell type in the parasite and the type responsible for metastasis formation and continuous parasite growth (6).

Treatment alternatives for AE are systemic chemotherapy and/or surgery. Surgery (radical or palliative) is complemented by postoperative pharmacotherapy (recommended in all patients for at least 2 years [7]). In those cases where surgery is not feasible, chemotherapy remains as the only option. Although there are alternative drugs, albendazole (ABZ) and mebendazole are the only ones licensed to date. Both exhibit a relatively good clinical efficacy but are associated with adverse side effects and lack of parasiticidal activity (8, 9). Therefore, the identification of better or alternative drugs is becoming increasingly urgent (5).

One of the most promising approaches to find new compounds against neglected infectious diseases is the repurposing of existing drugs (10). Metformin (Met, N,N-dimethylbiguanide), an antihyperglycemic agent widely used in type 2 diabetes mellitus treatment (11), has emerged as an anticancer drug, which is also effective against several pathogens, including Trichinella spiralis, Staphylococcus aureus, Pseudomonas aeruginosa, hepatitis B virus, hepatitis C virus, and human immunodeficiency virus (HIV) (12, 13). The drug has also been shown to suppress tumorsphere formation and to selectively target the cancer stem cells (CSCs) of various tumor types in both in vitro cultures and in vivo mouse experiments (14–16).

The effects of Met can be partially attributed to AMP-activated protein kinase (AMPK) activation, which depends on LKB1 (liver kinase B1) (17, 18). AMPK is a highly conserved sensor of cellular energy charge that regulates various metabolic pathways (19). It was demonstrated that Met treatment increases cellular levels of AMP through inhibition of complex I of the electron transport chain, leading to inhibition of the mitochondrial function and activation of AMPK (18). Nevertheless, recent studies showed that treatment of primary hepatocytes with clinically relevant concentrations of Met, as well as chronic administration of 50 mg/kg of body weight/day in mice, efficiently activates AMPK without disrupting the energy state (20). Thus, the mechanism by which Met activates AMPK is not yet fully understood.

The antiproliferative action of Met can be mediated by the indirect suppression of the TOR (target of rapamycin) pathway, mainly as a result of the activation of AMPK, and the inactivation of IGF1R (insulin-like growth factor type I receptor), two mechanisms that play a critical role in cell proliferation and growth (15, 21, 22). In all eukaryotes, TOR kinase is found in two functionally distinct complexes, TORC1 and TORC2 (23). TORC1 is a master regulator of anabolic pathways and a key hub mediating control of cell growth in response to nutrients and, in metazoans, to growth factors as well (insulin/IGF, which regulate the insulin-PI3K-AKT-pathway) (23, 24). Thereby, Met leads to phosphorylation and activation of AMPK (Thr¹⁷²), which in turn reduces the phosphorylation and activity of TOR (Ser²⁴⁴⁸). Consequently, expression/phosphorylation of TOR-downstream effectors such as ribosomal protein S6 kinase (S6K, Thr³⁸⁹), 4E-binding protein 1 (4EBP1, Thr^37/46), and IGF1R (Tyr^1135/6) is reduced, blocking protein synthesis and inactivating the cell proliferation (15, 25, 26). On the other hand, Met has been shown to activate AMPK and inhibit TOR by promoting the v-ATPase-Ragulator-AXIN/LKB1-AMPK complex to assemble into the lysosome surface. The shutdown of TORC1, in turn, promotes autophagy, a lysosomal process of bulk degradation of proteins and organelles (23, 27).

Recently, we observed a significant in vitro antiparasitic effect of Met on E. granulosus protoscoleces and metacestodes (28) and demonstrated that oral administration of this drug was effective against the larval stage of the parasite in the murine cystic echinococcosis (CE) infection model (29). In addition, we described the indirect activation of Eg-AMPK in response to Met treatment and showed that, in the parasite, the drug induces autophagy in the same way as rapamycin, an inhibitor of TOR (28, 30, 31). Since AE and CE are two related diseases that differ significantly regarding pathogenesis and metacestode morphology, it is of great interest to assess Met efficacy also against E. multilocularis. Therefore, in this work, we assessed the potential antiparasitic effect of Met on E. multilocularis stem cell-containing primary cell cultures as well as in a secondary mouse infection model of AE and examined the effect of this drug on mitochondrial function and the AMPK-TOR-autophagy pathway in the parasite. In addition, we raised the question of whether Met controls the development of Echinococcus through the indirect inhibition of TOR, as a consequence of the ATP synthesis inhibition, and/or through the direct inhibition of TOR, by the lysosomal pathway.

RESULTS

Pharmacological sensitivity of E. multilocularis primary stem cells and protoscoleces to metformin.

To study drug effects on parasite stem cells, we made use of the previously established primary cell culture system in which parasite cells are directly isolated from in vitro cultivated metacestode vesicles (32). These primary cell cultures are strongly enriched (up to 80%) in germinative stem cells (6). In order to investigate the in vitro effect of Met on the viability of these stem cell cultures, the percentage of living cells in response to different concentrations of Met was analyzed. Exposure to the drug for 72 h led to a dose-dependent decrease in the viability of primary cells, with significant effects at 5 mM or higher concentrations. After treatment with 10 mM Met, the viability percentage reached values below 50% (Fig. 1A). Since secondary echinococcosis can be induced by dedifferentiating protoscoleces (33, 34), drug effects were also assessed on this larval stage. As shown in Fig. S1A in the supplemental material, Met exerted a dose-dependent effect on the viability of protoscoleces after 8 days of incubation. The mortality rate of protoscoleces reached 20, 40, and 60% with 1, 5, and 10 mM Met, respectively. Furthermore, the drug partially arrested the in vitro dedifferentiation of protoscoleces into metacestode vesicles (Fig. 1B).

FIG 1.

In vitro effect of metformin on viability of primary stem cells of E. multilocularis and on the dedifferentiation process of protoscoleces to metacestodes. (A) Viability of primary cells incubated in the absence (C) or presence of different concentrations of metformin (1, 5, and 10 mM) for 72 h. Triton 1% was used as a positive control. Data are the mean ± standard deviation (SD) of three independent experiments. *, Statistically significant difference (P < 0.05) compared with control. (B) Box plot showing the number of metacestode vesicles per field of view recorded from untreated (C) and Met-treated (10 mM) protoscolex cultures at day 7. *, Statistically significant difference (P < 0.05) compared with control. n shows the number of images analyzed.

Molecular changes induced by metformin in the E. multilocularis larval stage.

To explore the possible inhibitory effect of Met on the complex 1 of the respiratory chain, we studied the mitochondrial functional status using the membrane potential indicator JC-1 in E. multilocularis metacestode vesicles (35).

Control and Met-treated metacestodes were examined by confocal microscopy for JC-1 fluorescence (Fig. 2a and b). Following 48 h of treatment with 10 mM Met, the relative values of red/green JC-1 fluorescence ratios showed low dispersion. At this point, untreated metacestodes showed a red/green fluorescence ratio with a mean value of 2.2, whereas Met-treated metacestodes showed a lower mean ratio of around 0.7 (Fig. 2c). Met treatment induced an increase in depolarized regions indicated by the disappearance of red fluorescence and an increase in green fluorescence (Fig. 2b).

FIG 2.

Molecular effects of metformin on in vitro generated E. multilocularis metacestodes. (a to c) Changes in mitochondrial membrane potential. (a and b) Representative confocal images showing JC-1 fluorescence in metacestodes incubated under control conditions (a) or treated with 10 mM metformin (b) for 48 h. Bars indicate 200 μm. (c) Box plot graph showing the values of the red/green JC-1 fluorescence ratios measured in control (C) and metformin (Met)-treated metacestodes with Image J Software. *, Statistically significant difference (P < 0.05) compared with control. (d to f) Pharmacological activation of Em-AMPKα. (d and e) Representative confocal images of in toto immunolocalization assays revealed with an antibody conjugated with Alexa 488 (green fluorescence) and counterstained with propidium iodide (red fluorescence). Control (d) and Met-treated (e) metacestodes incubated with anti-AMPKα-P antibody. Cytoplasmic expression is observed in green. Nuclear expression is observed in yellow/orange, corresponding to the merged fluorescences. Inset images correspond to transmission microscopy. Bars indicate 200 μm. (f) Graph depicts the ratio of p-AMPKα to nuclei fluorescence intensity in metacestodes treated with 10 mM Met relative to controls. Values are expressed as means ± standard error of the mean (*, P < 0.05 compared to the control). (g to i) Pharmacological induction of autophagy. (g and h) Representative confocal images of in toto immunolocalization assays revealed with an antibody conjugated with Alexa 488 green fluorescence and counterstained with propidium iodide (red fluorescence). Control (g) and Met-treated (h) metacestodes incubated with anti-LC3 antibody. Inset images correspond to transmission microscopy. Bars indicate 200 μm. (i) Graph depicts the ratio of Em-Atg8 to nuclei fluorescence intensity in metacestodes treated with 10 mM Met relative to controls. Values are expressed as means ± standard error of the mean (*, P < 0.05 compared to the control).

Since the maintenance of the mitochondrial membrane potential (ΔΨm) is required for ATP production, Met might activate Em-AMPK as a consequence of cellular energy charge depletion (increased cellular AMP:ATP ratio). Therefore, after confirming the expression of genes encoding the three subunits of AMPK in E. multilocularis (28), we studied the phosphorylation at Thr¹⁷⁶ of Em-AMPKα (AMPKα-P¹⁷⁶) as a readout of its activation state. To this end, in toto immunolocalization assays using a monoclonal antibody directed against the phosphorylated form of AMPKα were performed from protoscoleces (Fig. S1B) and in vitro generated metacestodes (Fig. 2d and e). As shown in Fig. S1C and Fig. 2f, a significant increase in the Em-AMPKα-P¹⁷⁶ level was observed after 48 h of treatment with 10 mM Met, indicating Em-AMPKα activation under these conditions. The expression of Em-AMPKα-P¹⁷⁶ was detected both in the nucleus and in the cytoplasm of the cells of Met-treated and control samples, although in Met-treated parasites, the nuclear expression was higher than in the control conditions. This is consistent with the presence of a nuclear export sequence at the C terminus of the catalytic subunit of Em-AMPK and its direct involvement in transcriptional regulation. The fluorescence pattern was not observed when the parasites were incubated with the secondary antibody alone (data not shown).

Since activation of AMPK by Met has been shown to induce autophagy (31, 36, 37), we further examined the effects of the drug on the autophagic pathway in the larval stage of E. multilocularis. With in toto immunolocalization assays, Em-Atg8 (an LC3β-homolog) was detected in a diffuse and punctate form in both control (Fig. 2g) and 10 mM Met-treated metacestodes (Fig. 2h), with the total fluorescence signal and the amount of punctuated structures being higher in the presence of the drug (Fig. 2i). It should be added that under the effect of the drug, the signal was evidenced in the nucleus of germinal layer cells (Fig. 2h).

Pharmacological activation of Em-TOR in the E. multilocularis larval stage.

In an attempt to determine the mechanism responsible for the antiparasitic effect of Met, its effect on TOR signaling was assessed. As TOR is activated by AKT phosphorylation at Ser²⁴⁴⁸ to promote protein synthesis and cell proliferation (38), we used a phospho-specific antibody to assess Em-TOR activity. As shown in Fig. 3, 10 mM Met treatment resulted in inhibition of Em-TOR, as demonstrated by decreased phosphorylation of Em-TOR (Ser³¹²²) in treated metacestodes (Fig. 3Ad) compared with untreated metacestodes (Fig. 3Aa). Additionally, while treatment with rapamycin (Rp) also caused inhibition of Em-TOR (Fig. 3Ac), treatment with insulin resulted in increased phosphorylation of Em-TOR^S3122 (Fig. 3Ab). Importantly, we showed that Ser³¹²² is highly conserved in the parasite protein, including the region around the serine (Fig. S2).

FIG 3.

Detection and immunolocalization of an activated form of Em-TOR in pharmacologically treated E. multilocularis metacestodes. (A) Representative confocal images of in toto immunolocalization assays revealed with an antibody conjugated with Alexa 488 (green fluorescence) and counterstained with propidium iodide (red fluorescence). Control (a) and insulin-treated (b), rapamycin-treated (c), and metformin-treated (d) metacestodes incubated with anti-TOR-P. Inset images correspond to transmission microscopy. Bars indicate 200 μm. (B) Graph depicts the ratio of Em-TOR-P^S3122 to nuclei fluorescence intensity in metacestodes treated with 1 U/ml insulin (Ins), 10 μM rapamycin (Rp), and 10 mM metformin (Met) relative to controls (C). Values are expressed as means ± standard error or the mean (*P < 0.05 compared to the control).

Subsequently, we analyzed the immunolocalization of Em-AKT, which was previously reported by Hemer et al. (39), in the larval stage of Echinococcus because it may be involved in the posttranslational regulation of Em-TOR (Fig. S1D). The expression of total Em-AKT was detected in the tegument of control and Met-treated protoscoleces (Fig. S1Da and Dc). However, its expression was generalized and included a spotted pattern in insulin-treated protoscoleces (Fig. S1Db). While the expression level of this kinase was unchanged in parasites treated with Met, it was increased in samples treated with insulin (Fig. S1E).

Anthelmintic efficacy of metformin on secondary alveolar echinococcosis in mice.

Given our results showing the in vitro effects of Met on primary stem cell viability and the ability of protoscoleces to dedifferentiate into metacestodes, we examined whether this drug could affect parasite growth in vivo. Each mouse was intraperitoneally infected with 200 μl of metacestode tissue, and Met was administered daily orally at 50 mg/kg/day for 60 days. All mice survived at the end of the experiment, and Met did not affect the animal weight and diet consumption. As shown in Fig. 4A, Met was effective in achieving a significant reduction of parasite weight (1.5 ± 1.1 g) compared to the untreated group (3.14 ± 1.1 g).

FIG 4.

In vivo efficacy of metformin against the E. multilocularis larval stage. Box plot showing the comparative distribution of the weight (in grams) of cysts recovered from untreated (C) and metformin-treated (Met, 50 mg/kg/d) mice. A significant cyst weight reduction (*, P < 0.05) was achieved in treated animals. (B) Representative SEM images of cysts recovered from untreated control mice (a to c) compared with Met-treated mice (d to f). Bars indicate 50 μm (a, d), 20 μm (b, e), and 10 μm (c, f).

To analyze the ultrastructural changes of parasite material recovered from each experimental group, scanning electron microscopy (SEM) studies were performed. Metacestode tissue from control mice appeared with protoscoleces and an intact germinal layer (Fig. 4Ba to Bc). In contrast, metacestodes collected from Met-treated mice displayed a marked reduction in the number of germinal cells (Fig. 4Bd to Bf).

In silico analysis of key components of the TOR pathway in Echinococcus.

Previously, sensitivity to rapamycin, induction of autophagy, and transcriptional expression of TOR in the larval stage of E. granulosus suggested the occurrence of Em-TOR in E. multilocularis (30, 40). The full-length open reading frame of Em-TOR, identified by a BLAST search, predicts a protein of 3,273 amino acids (annotated as CDS40303 and EmuJ_000787900 in the GenBank and GeneDB databases, respectively) with 28% to 37% overall identity to human TOR (GenBank accession number P42345). The Em-TOR has a conserved domain structure containing N-terminal HEAT (Huntington, EF3A, ATM, TOR) repeats followed by a FAT (FRAP, ATM, TTRAP) domain (∼600 residues), the FRB (FKBP-rapalog binding) domain (∼100 residues), the kinase domain, and a FATC domain (∼35 residues) at the C terminus (Fig. S2A). Although the N-terminal HEAT repeat and FAT region show a relatively low grade of conservation in the primary sequence (24% to 43% and 25% to 37% identity, respectively) with respect to vertebrate TOR orthologs, the C-terminal FRB domain, the kinase domain, and the FATC domain are highly conserved (46% to 70% identity) (Fig. S2). As in the human TOR, the FRB domain and the putative regulatory domain (RD) of Em-TOR are arranged on either side of the catalytic site (Fig. S2A and B). The RD is referred to as the negative regulatory domain (residues 2430 to 2492 in human TOR), since its deletion leads to an increase in TOR activity (41, 42). The mammalian amino acid sequence of this region is highly conserved, structurally disordered (PDB ID 4JSV), and contains regulatory phosphorylation sites such as Thr²⁴⁴⁶ and Ser²⁴⁴⁸ (KRSRTRTDSYSAGQSVE), which aligned with a conserved motif in Em-TOR (Fig. S2; 43, 44). Prediction of phosphorylation hot spot regions using NetPhos 3 revealed a score of 0.93 for Thr³¹¹⁹ and Ser³¹²², suggesting that these two putative phosphorylation sites match the TRT-X1/2-S consensus sequence. Finally, Em-TOR showed a considerable similarity in secondary and tertiary structures to human TOR (PDB 6bcu.1. A; 45).

Subsequently, the occurrence of the five subunits of Ragulator (LAMTOR1-5) was also analyzed in E. multilocularis. Extensive BLAST searches on the parasite genome revealed five genes coding for the different subunits of Ragulator (Fig. S3 and S4). The genes encode a 166-amino acid protein (named Em-LAMTOR1 and annotated as CDS42462 and EmuJ_001017100 in the GenBank and GeneDB databases, respectively), a 125-amino acid protein (named Em-LAMTOR2 and annotated as CDS41091 and EmuJ_000871100 in the GenBank and GeneDB databases, respectively), a 110-amino acid protein (named Em-LAMTOR3 and annotated as CDS43561 and EmuJ_001132600 in the GenBank and GeneDB databases, respectively), a 121-amino acid protein (named Em-LAMTOR4 and annotated as CDS38220 and EmuJ_000555600 in the GenBank and GeneDB databases, respectively), and a 103-amino acid protein (named Em-LAMTOR5 and annotated as CDS43359 and EmuJ_001111300 in the GenBank and GeneDB databases, respectively).

The predicted Em-LAMTOR1 sequence aligned with 23% and 30% identity with the Homo sapiens (GenBank accession number NP_060377) and Fasciola hepatica (THD24342) orthologs, respectively. The Em-LAMTOR1 subunit contains elements that support the identification of the protein product as a member of the LAMTOR family. The predicted protein is similar in size to the human and F. hepatica proteins. Likewise, it is predicted to be helix rich (42%) and has two N-terminal cysteines that may be sites for S-acylation (palmitoylation). In addition, Em-LAMTOR1 contains conserved key residues for its interaction with the other components of the complex (Fig. S4A). On the other hand, Em-LAMTOR2 aligned with 36% and 63% identity with the H. sapiens (NP_054736) and Hymenolepis microstoma (CDS27935) orthologs, respectively. The Em-LAMTOR2 subunit contains a conserved Roadblock/LC7 domain (pfam03259) located in the middle of the protein (Fig. S4B). For its part, the Em-LAMTOR3 subunit showed 24% identity with the H. sapiens (NP_068805) ortholog, and it presents the mitogen-activated protein kinase 1 interacting (pfam08923) (Fig. S4C). Additionally, the predicted Em-LAMTOR4 sequence aligned with 34% and 37% identity with the H. sapiens (NP_001008396) and Schistosoma haematobium (XP_012797450) orthologs, respectively (Fig. S4D). Finally, the Em-LAMTOR5 subunit showed 27% and 18% identity with the H. sapiens (O43504) and Drosophila busckii (XP_017851496) orthologs, respectively, and it contains the Roadblock/LC7 domain (pfam03259) like Hs-LAMTOR5 (Fig. S4E). Lastly, homology modeling results showed that the Em-LAMTOR1, Em-LAMTOR2, Em-LAMTOR3, and Em-LAMTOR4 proteins have a quaternary structure similar to that of human orthologs in the Ragulator complex (PDB ID 6ehr.1 and 5yk3.3, PDB 5x6v.1, PDB 5yk3.1 and PDB 5y39.1, respectively).

DISCUSSION

The severe infection caused by the E. multilocularis metacestode is currently treated with benzimidazoles. However, these drugs are ineffective for some patients, at least in part due to their low bioavailability, limited half-life in the host, and restricted uptake by the parasite. Moreover, they also lack parasiticidal activity and cause toxicity (9, 46). In this report, we demonstrated that Met reduces E. multilocularis primary stem cell viability in culture and parasite mass in a murine secondary AE model. The drug toxicity mechanism could be attributed to the mitochondrial membrane depolarization and the modulation of the AMPK-TOR-autophagy pathway in the parasite. Importantly, both the in vitro and in vivo models and the treatment schedule that we used for the screening of Met against AE were previously established (47–49).

Met not only significantly decreased the viability of primary stem cells and protoscoleces, but also reduced the dedifferentiation of protoscoleces into metacestode vesicles. These findings are consistent with results showing that the drug inhibited stem cell proliferation in a dose-dependent manner in preclinical cancer models (14). The concentrations of Met used in our in vitro experiments are high compared to those in plasma (5 to 18 μM) and liver (50 to 100 μM) reported in vivo (50). However, they are within the range used to examine the in vitro effects of the drug on cell metabolism and proliferation (51, 52). A likely cause for the need of these drug concentrations is the large amount of glucose and growth factors employed in culture medium (53). Echinococcus stem cell targeting using Met could lead to a breakthrough of therapeutic approaches for AE, given that it has been suggested that the stem cells are resistant to ABZ (54, 55). In fact, compared to other drugs with an in vitro effect on parasite stem cells, such as kinase inhibitors (56, 57), Met can also control the parasite development in vivo. In relation to this, it has been reported that Met can act on stem cells through the PKA-GSK3β signal pathway by suppressing the expression of KLF5 (Krüppel-like factor 5), a transcription factor involved in the expression of developmental genes (52), which has a putative ortholog in the Echinococcus genome (Em_000425100).

In our in vivo assay, treatment with Met showed a significant therapeutic effect against murine experimental AE. Under these conditions, a reduction in the weight of the recovered metacestodes as well as the destruction of their germinal layer were observed. Considering that germinative cells can be released into the environment of the parasite and affect neighboring organs, our in vitro results, together with the results showing ultrastructural changes induced by Met in vivo, suggest that the drug may not only be effective in preventing metastasis formation, but may also be useful for its administration during the peri- and postoperative periods, in which cell spread could occur. This could be reinforced by the ability of Met to weaken the induction or reverse the epithelial-mesenchymal transition (EMT) in cancer-like cells (58). It is important to note that the dose of Met employed in this experiment is the lowest used in mice (50 mg/kg/day) and is within the clinically relevant range for humans (59). Met has a bioavailability of 50% to 60%, it is not metabolized, and its mean plasma half-life is approximately 20 h (50, 60). After oral ingestion, Met is absorbed by the small intestine, distributed by the portal vein, and concentrates in the liver (60). This could be an advantage for AE treatment since the liver is the main target organ of the parasite. In fact, Met accumulation in E. granulosus has been reported in experimental CE (29). Currently, we are extending the study to evaluate the in vivo efficacy of Met using different infection and treatment schemes by increasing the parasitic inoculum and starting the treatment later.

Concomitantly, our results showed that in germinative cells of metacestodes, Met inhibited the complex I of the mitochondrial respiratory chain, a direct molecular target confirmed for the drug (53). Therefore, by altering the mitochondrial metabolism and oxidative phosphorylation, Met could affect the production of ATP and TCA cycle intermediates, preventing the proliferation of germinative cells, as has been reported for cancer stem cells (15). The strategy of interfering with the energy-generating systems in the larval stage of E. multilocularis has previously been proven to be effective using drugs such as buparvaquone and quinazoline-type compounds in vitro (10, 61).

On the other hand, as a consequence of ATP depletion, Met activates AMPK and inhibits TOR. The latter is a key mediator of the PI3K/AKT pathway, which responds to growth factors and insulin (17). Although in the E. multilocularis genome project it has already been reported that the parasite encodes a TOR ortholog and an insulin-like tyrosine kinase receptor (InsR) (62), no studies concerning the TOR activity in the parasite have been performed to date. In our experiments, Met led to the activation of Em-AMPK and the decrease in Em-TOR activity in both metacestode vesicles and protoscoleces. The inhibition of Em-TOR could justify the antiechinococcal effect of this drug on the parasite’s larval stage, as has been evidenced in cancer cells (15). In addition, the reduction in Em-TOR phosphorylation after Met treatment was accompanied by an increase in Em-Atg8 levels, indicating autophagy induction. Since it has been described that Met can activate prodeath autophagy in lymphoma and melanoma cells (36, 37), the antiechinococcal effects of Met could also be partially dependent on autophagy.

Although the main building blocks of the insulin-InsR-PI3K-AKT-FoxO-TOR pathway are generally conserved among mammals and invertebrates, Echinococcus spp. possess a single gene for InsR, FoxO, S6K, and EIF4B (similar to flies and free-living worms) and lack the insulin-like peptides, PTEN, TSC1/2, and Rheb (30, 31, 39, 62, 63). Given that host insulin stimulates the in vitro vesicular growth of E. multilocularis (39, 64) and that TOR represents a direct target of the InsR-PI3K-AKT pathway in insulin-stimulated cells, we also assessed the potential differential phosphorylation of Em-TOR in response to exogenous insulin and rapamycin (a direct inhibitor of TOR), demonstrating that Em-TOR reacts in the opposite manner under these stimuli. Thus, we confirmed that Em-TOR is part of the parasite signaling pathway downstream of insulin provided by the host. Similar to its orthologs, Em-TOR is a high-molecular-weight protein that contains all conserved structural domains. It also contains, between the catalytic and FATC domains, two conserved phosphorylation hot spots, Thr³¹¹⁹ and Ser³¹²² (corresponding to the phosphosites Thr²⁴⁴⁶ and Ser²⁴⁴⁸ in human TOR), which respond to insulin (41, 43). Phosphorylation of these sites is highly conserved among vertebrate species but is absent in free-living invertebrates (65). However, we showed that the phosphorylation site region is partially conserved in helminth parasites, demonstrating that the phosphorylation status of Em-TOR can be measured by using antibodies against the phosphorylated Hs-TOR-S²⁴⁴⁸ form. Probably as a result of the coevolution of the parasite and its vertebrate host, TOR activity in these helminths could be regulated in the same way as in vertebrates (39, 66). Moreover, in these TOR sequences, we identified additional peptides that were previously unobserved, with coil structures, surrounding the phosphorylation sites into RD. These sites may be druggable, providing new opportunities for the development of specific inhibitors with a high degree of specificity for parasite TOR kinase.

On the other hand, intrigued by the discovery that Met may directly act on the lysosomal pathway to promote AMPK activation and TOR inhibition (27, 50), we sought to establish the presence of LAMTOR orthologs in the Echinococcus genome. This finding, as well as the presence of a gene encoding v-ATPase (67) and two axin paralogs in the genome of this tapeworm (68), suggests that the formation of the v-ATPase-Ragulator-AXIN/LKB1-AMPK complex is possible in the parasite. This, added to the demonstration of the lysosomal pathway functionality within the invertebrate model Caenorhabditis elegans (69), emphasizes the relevance of studying this pathway to identify new pharmacological targets.

Although our study highlights the importance of the direct mechanisms by which Met reduces parasite viability, we consider that Met could also have systemic effects that contribute to its potential as an antiechinococcal therapeutic agent in vivo. These may include suppression of the Warburg effect, a metabolic strategy acquired by the Echinococcus larval stage under limited oxygen supply (29, 70), and control of liver chronic inflammation by reducing the proinflammatory cytokine levels (71) and increasing the cytotoxic response by blocking the PD-1/PD-L1 (programmed cell death-1/programmed death ligand-1) axis (72).

By targeting germinative cells, Met holds great promise for the treatment of AE. Therefore, our results provide a rational basis for testing the combination of Met and ABZ, given that drugs with different mechanisms of action could improve treatment efficacy. On the other hand, the drug presents the advantages of being commercially available, approved by the FDA, and extensively characterized in terms of bioavailability and pharmacokinetics, with a good long-term safety profile and controllable side effects (13, 16).

MATERIALS AND METHODS

Ethics statement.

Female CF-1 mice (8 weeks of age) were supplied by the National Health Service and Food Quality (SENASA), Mar del Plata, and housed in specific-pathogen-free (SPF) facilities at the bioterium of the National University of Mar del Plata (UNMdP). Experimental protocols for using mice were evaluated and approved by the Animal Experimental Committee at the Faculty of Exact and Natural Sciences, UNMdP (permit number 2555-08-15). Experiments for the continuous passage of E. multilocularis larval material in Mongolian jirds (Meriones unguiculatus) were carried out in accordance with European and German regulations on the protection of animals. Ethics approval of these studies was obtained from the local ethics committee of the government of Lower Franconia (permit number 55.2 DMS 2532-2-354).

Maintenance, culture, and collection of parasites.

E. multilocularis (isolates J2012 and 8065) was maintained by serial intraperitoneal passage in M. unguiculatus or CF-1 mice as described by Spiliotis and Brehm (47). Homogenized metacestode material obtained from M. unguiculatus or CF-1 mice was cultured in vitro with rat Reuber hepatoma or Huh7 cells as previously described (47). In addition, protoscoleces obtained from CF-1 mice were cultured in vitro as in Wang et al. (33). Once metacestode tissue or protoscoleces developed metacestode vesicles (typically after 1 to 3 months of in vitro culture), they were collected and fixed for in toto immunolocalization assays with 4% paraformaldehyde prepared in 0.1 M phosphate-buffered saline (PBS) (pH 7.4) as described by Loos et al. (30). In vitro staining of metacestodes with JC-1 was performed by incubating the metacestode vesicles for 30 min in culture medium in the presence of 10 mg/ml JC-1 (28). JC-1 is a positively charged fluorescent compound that can penetrate mitochondria and change its color as a function of the mitochondrial membrane potential (ΔΨm). It accumulates as aggregates with intense red fluorescence within the mitochondria when the ΔΨm is high or remains as green monomers in the cytoplasm and the mitochondria when the ΔΨm is low (35). Isolation of protoscoleces and primary stem cell cultures was carried out as described by Spiliotis et al. (32).

Drug treatment and viability assays.

Metformin (Sigma-Aldrich) was used dissolved in water at final concentrations of 1, 5, and 10 mM. Primary stem cells were cultivated in hepatocyte-conditioned medium (32) supplemented with Met for 72 h, and the cell vitality was assessed with an alamarBlue assay (73). Protoscoleces were incubated in medium 199 supplemented with Met for 8 days, and viability was subsequently assessed using a methylene blue staining assay (30). In all culture systems, medium was changed every third day, including fresh addition of Met. All experiments were carried out at least three times independently.

Metacestode vesicle development in the presence of metformin.

E. multilocularis protoscoleces (50 μl for each condition) were cultured in RPMI 1640 medium (Gibco) containing 25% (vol/vol) fetal bovine serum (FBS) (Gibco), 0.45% (wt/vol) yeast extract, 0.4% (wt/vol) glucose, and 100 μg/ml penicillin, streptomycin, and gentamicin in a 25-cm² culture flask at 37°C in the presence of 5% CO₂. The experimental conditions evaluated were control and 10 mM Met. The medium was changed every 7 days, including fresh addition of Met. At the same time, each culture was monitored under an optical microscope in order to record the total number of larval vesicles. The experiment was repeated three times.

In toto immunohistochemistry.

For in toto immunohistochemistry, control and treated protoscoleces and metacestode vesicles were processed for analysis of total and phosphorylated (Thr¹⁷⁴) Em-AMPKα, Em-Atg8, phosphorylated (Ser³¹²²) Em-TOR, and Em-AKT as previously described (28, 30, 31). The samples were incubated with primary monoclonal antibodies directed against phosphorylated and total human AMPKα [phospho-AMPKα-Thr¹⁷²-(40H9) rabbit monoclonal antibody (MAb) and AMPKα (D63G4) rabbit MAb (Cell Signaling Technology, USA, catalog numbers 2535 and 5832, respectively) (1:1000 dilution)], primary polyclonal antibody directed against the N terminus of human LC3 (MAP LC3β clone H-50, catalog number sc-28266; Santa Cruz Biotechnology, USA) (1:1,000 dilution), primary monoclonal antibody directed against phosphorylated human mTOR (phospho-mTOR-Ser²⁴⁴⁸-[D9C2] rabbit MAb, catalog number 5536; Cell Signaling Technology) (1:1,000 dilution), and primary polyclonal antibody directed against total mouse AKT (also known as PKB) (AKT antibody, catalog number. 9272; Cell Signaling Technology) (1:1,000 dilution). The anti-TOR antibody used in these assays is directed against an epitope which showed 30% amino acid identity with the possible orthologs of E. multilocularis (Em-TOR) (GenBank accession number CDS40303; see alignment of the conserved motif in Fig. S2B, red box). Negative controls consisted of omission of primary antibody.

Immunofluorescence images were acquired using an inverted confocal laser scanning microscope (confocal microscope C1; Nikon) with an excitation/emission wavelength of 494/517 nm for Alexa Fluor 488-conjugated antibodies and 536/617 nm for propidium iodide-stained nuclei. Fluorescence intensity was measured using ImageJ software (NIH, https://imagej.nih.gov/ij/; 74) in randomly chosen sections of control and pharmacologically treated samples. The surrounding background was subtracted before different regions of interest (ROIs) were analyzed to obtain the mean intensity values. A total of 20 images per condition of 3 independent sets of experiments were acquired and analyzed. Image files were loaded as separate image stacks. Ratios of Em-AMPKα-P, Em-Atg8, or Em-TOR-P to nuclei fluorescence intensity were calculated and displayed as bar plots. Negative controls consisted of omission of primary antibody.

Experimental animals and determination of efficacy of in vivo treatment.

Healthy CF-1 mice (30 ± 5 g) were acclimatized for 1 week before initiation of the experiment. Mice were infected by intraperitoneal infection with 200 μl of homogenized metacestode material (strain 8065) to produce experimental secondary AE (48). The animals were maintained in standard polyethylene cages (three mice per cage) under controlled laboratory conditions (temperature, 20 ± 2°C; 12 h photoperiod with lights off at 8.00 p.m.; 50% ± 5% humidity). Food and water were provided ad libitum. Every 3 to 4 days, animals were placed into a clean cage with fresh wood shavings. The pharmacological treatment was performed by intragastric administration of a drug aqueous suspension (0.3 ml/animal). At the end of the experiments, mice were euthanized by cervical dislocation after anesthesia with ketamine-xylazine (50 mg/kg/mouse ketamine and 5 mg/kg/mouse xylazine). All efforts were made to minimize suffering. The minimum number of animals was used in each experiment. At necropsy, the peritoneal cavity was opened, and the parasite tissues were carefully recovered and weighed. The efficacy of treatment was calculated using the following formula: 100 × {(mean parasite weight of control group) – (mean parasite weight of treated group)}/(mean parasite weight of control group). In addition, samples were processed for scanning electron microscopy (SEM) with a JEOL JSM-6460LV electron microscope as previously described (30).

Experimental treatment of E. multilocularis-infected mice with metformin.

At the time of infection, 12 CF-1 mice were allocated into 2 experimental groups (6 animals/group) as the untreated control group (water) and the Met-treated group (50 mg/kg/day). The drug was applied by oral gavage daily for 60 days. At the end of the treatment period, animals were euthanized and necropsied.

Sequence analysis of Echinococcus TOR and Ragulator complex.

A BLAST search for homologs of TOR and the different components of the Ragulator complex in the E. multilocularis and E. granulosus genome databases (https://www.sanger.ac.uk/Projects/Echinococcus; 62) was performed using orthologs from H. sapiens, F. hepatica, H. microstoma, Drosophila hydei, S. haematobium, and D. busckii as queries. This search allowed the identification of the putative orthologous genes encoding TOR and the Ragulator subunits (Em-lamtor1, Em-lamtor2, Em-lamtor3, Em-lamtor4, and Em-lamtor5), whose predicted open reading frames were analyzed. Orthologs were selected based on reciprocal best BLAST hits (75, 76) on an E value cutoff of 1e⁻²⁵ and on the presence of the characteristic domains in the deduced amino acid sequences. Sequence alignments were generated with the CLUSTALX software program (https://www.ebi.ac.uk/Tools/msa/clustalo/), and the modeling of secondary structures of the putative proteins was obtained from the deduced primary structures using Gen-THREADER (http://bioinf.cs.ucl.ac.uk/psipred/psiform.html). The SWISS-MODEL server (https://swissmodel.expasy.org/interactive) was used to generate alignments and homology models for Em-TOR and Em-Ragulator proteins, selecting template protein structures in PDB with a high coverage (>60% of target aligned to template) and sequence identity of >30%. Also, phosphorylation sites were predicted for Em-TOR by submitting the sequence to Web-based tools, namely, NetPhos 3.1 server (http://www.cbs.dtu.dk/services/NetPhos).

Statistics.

Data within experiments were compared using Student’s t test or the nonparametric Mann-Whitney test, and differences among groups were considered statistically significant with a P value below 0.05. Statistical analyses were performed using R software (https://www.R-project.org).