Medical Microbiology and Immunology

, Volume 199, Issue 3, pp 247–259

The role of evolutionarily conserved signalling systems in Echinococcus multilocularis development and host–parasite interaction

Authors

    • Institute of Hygiene and MicrobiologyUniversity of Würzburg
Review

DOI: 10.1007/s00430-010-0154-1

Cite this article as:
Brehm, K. Med Microbiol Immunol (2010) 199: 247. doi:10.1007/s00430-010-0154-1

Abstract

Alveolar echinococcosis, one of the most serious and life-threatening zoonoses in the world, is caused by the metacestode larval stage of the fox-tapeworm Echinococcus multilocularis. Mostly due to its accessibility to in vitro cultivation, this parasite has recently evolved into an experimental model system to study larval cestode development and associated host–parasite interaction mechanisms. Respective advances include the establishment of axenic in vitro cultivation systems for parasite larvae as well as culture systems by which the early development of metacestode vesicles from totipotent parasite stem cells can be reconstituted under controlled laboratory conditions. A series of evolutionarily conserved signalling molecules of the insulin, epidermal growth factor and transforming growth factor-β pathways that are able to functionally interact with corresponding host cytokines have been described in E. multilocularis and most likely play a crucial role in parasite development within the liver of the intermediate host. Furthermore, a whole genome sequencing project has been initiated by which a comprehensive picture on E. multilocularis cell–cell communication systems will be available in due time, including information on parasite cytokines that are secreted towards host tissue and thus might affect the immune response. In this article, an overview of our current picture on Echinococcus signalling systems will be given, and the potential to exploit these pathways as targets for anti-parasitic chemotherapy will be discussed.

Keywords

SignallingChemotherapyHost–parasite interactionStem cellsNeoblasts

Introduction

Alveolar echinococcosis (AE), cystic echinococcosis (CE) and neurocysticercosis (NCC) are important diseases that are caused by the metacestode larval stages of the three related tapeworms Echinococcus multilocularis, E. granulosus and Taenia solium, respectively. In all three cases, infection of the intermediate host (domesticated animals, rodents and humans) is initiated by the oral uptake of so-called ‘infectious eggs’ that contain the first larval stage, the oncosphere. Once activated within the stomach and intestine of the intermediate host, the oncosphere hatches from the egg, actively penetrates the intestinal barrier and subsequently gains access to the inner organs where it undergoes a metamorphosis towards the second larval stage, the bladder-like metacestode [13]. In the case of T. solium, the metacestode stage (also called ‘cysticercus’) primarily infects muscle and brain tissue and consists of defined vesicles with a cellular, germinal layer that gives rise to one single scolex per vesicle in a later stage of the infection. In E. granulosus, the metacestode vesicles primarily affect the host’s liver and lung and can develop into massive ‘hydatid cysts’ that contain an outer, acellular ‘laminated layer’ composed of a carbohydrate network to which parasite and host proteins are attached, as well as inner, germinal layer that is morphologically similar to that of T. solium and that produces numerous protoscoleces in later time points of the infection. In contrast to T. solium and E. granulosus, the E. multilocularis metacestode displays a very strong organ-tropism towards the liver of the intermediate host. Its vesicles are structurally similar to those of E. granulosus but with a much thinner laminated layer. E. multilocularis metacestode vesicles can grow by size but are also able to produce daughter vesicles through budding, which results in a complex, multi-vesicular parasite tissue that grows infiltratively, like a malignant tumour, into the surrounding host tissue. At later time points of the infection, protoscoleces are produced from the germinal layer and are passed onto the definitive host when it takes the prey. Furthermore, and in contrast to E. granulosus and T. solium, the E. multilocularis metacestode has the capacity to metastasize, yielding proliferating parasite tissue at secondary organs such as brain, heart or kidney during prolonged infections [13].

Despite the distinct morphological features of the metacestodes of E. multilocularis, E. granulosus and T. solium, it is expected that the initial phase of parasite establishment within the different target organs of the host is largely identical, at least from the view-point of developmental biology. Central to the biology of all flatworms is a population of totipotent somatic stem cells, called ‘neoblasts’, which account for the amazing regenerative capabilities and morphological plasticity in free-living turbellaria [47]. In parasitic flatworms (trematodes and cestodes), the functional equivalents of neoblasts have been named ‘germinal cells’, ‘germinative cells’,’regenerative cells’ or sometimes just ‘stem cells’, depending on the system studied [4, 5]. Nevertheless, despite this different nomenclature, it is expected that the basic biological functions, including regulatory mechanisms of stem cell maintenance and differentiation, are largely shared between neoblasts of free-living and parasitic flatworms and that this cell type played a master role in the evolution of the complex life cycles of trematodes and cestodes [4]. In cestodes, the somatic stem cells are present in the neck region of the adult where they constantly give rise to new proglottides. Within the proglottides, the somatic stem cells then initiate the formation of the genital anlagen and, after sexual reproduction, are newly formed within the oncosphere. Upon infection of the intermediate host, the oncosphere’s stem cells (between 6 and 12, depending on the species) are the only cells that contribute to the formation of the metacestode, and it is expected that stem cells attached to the germinal layer are also crucial for protoscolex production in the late phase of an infection [4, 5]. If we want to understand how the complex cestode life cycles have evolved, more molecular and cellular knowledge concerning the function and regulation of cestode stem cells is required. Furthermore, by elucidating the specific differences between stem cells from different cestode species, fundamental knowledge concerning organ-tropism and general aspects of molecular host–parasite interactions can be gained.

Although cestode species such as T. solium or E. granulosus are clinically more relevant, based on the number of infections they cause worldwide, E. multilocularis has the big advantage that it can be cultured much easier under laboratory conditions, which is a necessary prerequisite to undertake molecular studies concerning host–parasite interaction. Unlike in the case of the other two species, larval material of E. multilocularis can be routinely maintained in the peritoneal cavity of laboratory animals and has already been very successfully used to set up in vitro cultivation systems [4, 8]. In the present article, I will give a brief overview of the different cultivation systems that have been developed recently. I will further introduce evolutionarily conserved signalling systems that have been characterized in E. multilocularis and that are highly relevant to understand mechanisms of stem cell differentiation. Finally, I will discuss how molecular and biochemical knowledge on these signalling systems might be exploited to design novel anti-parasitic drugs and will describe some first steps that have been gone in this direction.

Recent advances in the in vitro cultivation of E. multilocularis larvae and stem cells

The first successful attempts to cultivate E. multilocularis metacestode material in vitro date back to the mid-1990s when Hemphill and Gottstein [9] as well as Jura et al. [10] introduced so-called co-cultivation systems that relied on the co-incubation of parasite tissue, isolated from the peritoneum of laboratory mice, with host feeder cells such as rat hepatocytes or CACO-2 colon carcinoma cells (reviewed in [8]). In both systems, the presence of serum and feeder cells was absolutely necessary to support parasite development, which strongly indicated that host cells secrete soluble growth factors for the metacestode. Although metacestode vesicles that had been produced by either of these methods proved to be highly useful in drug screening assays [11] or as a resource for nucleic acids to establish cDNA libraries [12, 13], they were only of limited use for the identification of defined host factors that support parasite growth due to the continuous presence of feeder cells. To overcome these obstacles, my group developed an axenic cultivation system [14] by which the proliferation of metacestode vesicles, and even their differentiation towards the protoscolex stage, can be investigated in vitro in the complete absence of host cells. In this system, it proved to be crucial that the parasite vesicles were kept under reducing conditions and under a nitrogen atmosphere, indicating that E. multilocularis larval material is highly sensitive to reactive oxygen intermediates that are produced during culture in an oxygen-containing atmosphere [8, 14]. Furthermore, although reducing conditions and serum allowed survival of metacestode vesicles for several weeks, parasite growth and development depended on the addition of feeder-cell-conditioned medium, again indicating that those host cells secrete soluble growth factors that support parasite development. Since not only supernatant of hepatoma or CACO-2 cells, but also medium that had been conditioned with a number of other cell lines readily promoted parasite growth [14], the growth factors in question are either widely expressed by mammalian tissue or are products typically secreted by immortal cell lines. Interestingly, epidermal growth factor (EGF) and EGF-like cytokines such as transforming growth factor-α (TGF-α) or amphiregulin have been shown to be frequently secreted by transformed cell lines [15], including rat hepatoma and colon carcinoma cells [16, 17], and thus might directly stimulate metacestode growth through binding to a parasite receptor tyrosine kinase of the EGF receptor family, EmER, as will be discussed below.

One of the decisive advantages of the axenic cultivation system was that it yielded metacestode vesicles that were essentially free of contaminating host cells. In earlier attempts to establish germinal cell cultivation systems for E. multilocularis, parasite material isolated from the peritoneal cavity of mice had routinely been used, which eventually resulted in an overgrowth of parasite cells through host fibroblasts [4, 8]. In our system, particularly after having applied some modifications to yield high numbers of axenic metacestode vesicles [18], a sufficient quantity of primary cells could be isolated that now allowed long-term incubation without host fibroblasts [19]. On the basis of electron microscopic investigations and flow cytometry analyses, at least 30% of the cells in freshly established primary cultures were shown to represented germinal cells. Interestingly, when these parasite cells were incubated under favourable growth conditions, they first underwent proliferation that resulted in the production of cell aggregates after 1–2 weeks of cultivation. Within these aggregates, small internal cavities arose that continually enlarged and resulted, after 4–5 weeks, in young vesicles that were surrounded by a growing, syncythial germinal layer to which germinal cells were attached [19]. After around 6 weeks of incubation, the formation of the laminated layer occurred, upon which the mature vesicles detached from the aggregates and could be further incubated until protoscolex formation was initiated [19]. When injected into the peritoneum of laboratory mice, these in vitro regenerated metacestode vesicles yielded high loads of parasite material, including numerous fully developed protoscoleces, thus confirming totipotency of the germinal cells that had been used to set up the original culture. Furthermore, in vitro regenerated vesicles could, again, be used as a source for stem cells to initiate new rounds of vesicle formation, indicating that the regenerative capacity of E. multilocularis metacestode tissue even exceeds that of free-living flatworms [19].

Morphologically, the in vitro metacestode vesicle regeneration process closely resembled the generation of metastatic foci, as previously observed by Mehlhorn et al. [20]. The E. multilocularis primary cell culture system should thus serve as a highly useful tool to study the molecular host–parasite interaction mechanisms associated with metastasis formation as it occurs during prolonged infections in humans [4, 8]. Furthermore, clear similarities between in vitro regenerating E. multilocularis vesicles and the early development of metacestode tissue from the oncosphere in experimentally infected mice were observed [4, 8, 18]. Hence, despite the different origin of Echinococcus stem cells in the regeneration system (metacestode-derived) and in natural infections (oncosphere-derived), there seems to be no qualitative difference in how they react to comparable environmental conditions. The in vitro regeneration system should therefore also be a highly relevant tool to study the molecular developmental processes during the early establishment of the parasitic oncosphere within the intermediate host’s liver. This is supported by the fact that regenerating parasite vesicles, already after 1 week of in vitro culture, express antigens such as EM95 [21], which have previously been described to be oncosphere-specific surface factors (K. Brehm, unpublished observation). Finally, since the early developing metacestode (prior to the establishment of the laminated layer) should be highly vulnerable to the attacking host immune system [22], the in vitro regeneration system can now also be used to study the excretory/secretory parasite products that are relevant for immuno-modulatory processes in the establishment phase of the disease.

Functional characterization of evolutionarily conserved signalling systems in E. multilocularis

As multicellular organisms, parasitic helminths rely on developmental mechanisms that have arisen very early in metazoan evolution and are, therefore, also present in their various vertebrate and invertebrate hosts [1, 23]. The basic toolkit of cell–cell communication systems that are used by metazoans to properly control body pattern formation was already established in the most basal phyla such as sponges or cnidarians and comprises, among others, peptide hormones and cytokines such as insulin and insulin-like growth factors (IGF), the epidermal growth factor (EGF) and related molecules, fibroblast growth factors (FGF) and the dimeric cytokines of the transforming growth factor-β (TGF-β)/bone morphogenetic protein (BMP) family, including their cognate receptors of the receptor tyrosine kinase (RTK) and the receptor serine/threonine kinase (RSTK) families [1, 23]. Interestingly, the respective hormones/cytokines and receptors from different animal phyla are not only structurally related, but can even functionally replace each other. Early studies on signalling systems of the invertebrate model systems Drosophila melanogaster and Caenorhabditis elegans have, for example, demonstrated that mammalian insulin- and BMP-like molecules can stimulate respective RTKs and RSTKs of invertebrates [24, 25]. In the case of tissue- and blood-dwelling helminth parasites, this has led to the interesting concept of ‘hormonal host–parasite cross-communication’, which proposes that host cytokines of the insulin, EGF, FGF, and TGF-β/BMP families might bind to corresponding RTKs and RSTKs at the surface of helminths, thus directing parasite development at suitable sites within the host, while parasite cytokines of these families could bind to corresponding host receptors, thus affecting the immune response and host physiology in their favour [1, 23, 2628].

In the case of E. multilocularis, respective interactions could be particularly relevant at the primary site of infection, the liver, since a number of evolutionarily conserved cytokines and hormones are either expressed in relatively high concentrations by the liver parenchyma or are otherwise produced during the immune response. EGF- and FGF-like cytokines are, for example, readily expressed during liver regeneration [29], which might be induced once the oncosphere or the immune response has inflicted damage. TGF-β as well as activins and inhibins are secreted in significant quantities by host cells during the immune response [30] and cytokines of the BMP family regulate extracellular matrix production, regeneration and iron homoeostasis within the liver [3134]. Finally, the liver is also the host organ where the highest concentrations of insulin can be found [35]. Several studies during recent years therefore concentrated on the identification of evolutionarily conserved E. multilocularis signalling systems and on questions whether these can interact with corresponding cytokines of the host.

Receptor tyrosine kinase signalling in Echinococcus

Peptide hormones and cytokines of the insulin/IGF, the EGF and the FGF families signal through surface-associated RTKs in both vertebrates and invertebrates. Upon cytokine/hormone binding, dimeric receptor complexes usually undergo auto-phosphorylation at tyrosine residues within the intracellular domain to which downstream signalling proteins bind, mostly via phospho-tyrosine-specific SH2 domains [36, 37]. In E. multilocularis, my group could so far functionally characterize two RTKs of the insulin (EmIR) and the EGF receptor families (EmER) [38, 39] (Table 1). Predicted in the first assembly version of the E. multilocularis genome (see below) are one additional insulin-receptor-like RTK (EmIR2) and one member of the FGF receptor family (EmFR). The situation in E. multilocularis therefore closely resembles that in the related trematode Schistosoma mansoni, the causative agent of schistosomiasis, which also expresses two members of the insulin receptor family, SmIR-1 and SmIR-2 [40], and one member of the EGF receptor family, SER [41], all with considerable amino acid sequence homologies to EmIR, EmIR2 and EmER, respectively. The overall sequence homologies between the flatworm EGF and insulin receptors to those of mammalian origin range between 25 and 28% but are considerably higher when only the ligand-binding domains (30–45%) or the tyrosine kinase domains (48–56%) are considered. A biochemical peculiarity of the cestode and trematode insulin receptors is the presence of inserts of up to 172 amino acids within the otherwise highly conserved tyrosine kinase domains that cannot be found in insulin receptor RTKs of other organisms. Although the precise function of these inserts is unknown at present, they are most probably involved in the regulation of signalling processes such as receptor conformational adaptation or cross-talk with other signalling systems [38, 40].
Table 1

Published E. multilocularis signalling factors and their location in the current assembly version of the genome

Protein

Gene

Function

Contig#

Position

Size (g)a

Size (p)b

Ref.

EmIR

emir

RTK, insulin receptor family

3363

431848–448063

16.2

190

[38]

EmER

emer

RTK, EGF receptor family

3085

107210–62240

44.9

175

[39]

EmRas

emras

Small GTPase; Ras-family

3020

17764–17166

0.6

21

[43]

EmRal

emral

Small GTPase; Ras-family

3450

178651–182761

4.1

23

[46]

EmRaf

emraf

MAPKKK; Raf-family

2891

15109–27423

12.3

87

[43]

EmMKK1

emmkk1

MAPKK; MKK3/6-family

2775

68417–57536

10.9

38

[45]

EmMKK2

emmkk2

MAPKK; MEK1/2-family

1752

61701–67031

5.3

57

[45]

EmMPK1

emmpk1

MAPK; Erk-family

3450

517953–524269

6.3

42

[44]

EmMPK2

emmpk2

MAPK; p38-family

1841

22335–15545

6.8

42

[47]

EmTR1

emtr1

RSTK; TGF-β-type I family

2863

314321–312182

2.1

61

[55]

EmSmadA

emsmadA

AR-Smad; activin signalling

1896

24674–18200

6.5

35

[56]

EmSmadB

emsmadB

BR-Smad; BMP signalling

2104

27820–25171

2.6

48

[56]

EmSmadC

emsmadC

AR-Smad; activin signalling

4678

90486–94261

3.8

37

[57]

EmSmadD

emsmadD

Co-Smad; TGF-β/BMP sign.

3408

19628–4395

15.2

77

[57]

EmSkip

emskip

SNW/SKIP fam. transcr. cor.

3396

104098–102401

1.7

60

[58]

EmPDZ-1

empdz1

PDZ-domain scaffolding fact.

2733

109195–107121

2.1

23

[49]

Em14-3-3.2

em1433.2

14-3-3 Family scaffold. Fact.

5645

28659–29489

0.8

30

[48]

aGene size (kb)

bProtein size (kDa)

In addition to RTKs, a number of downstream-acting intracellular signalling molecules could be identified (Table 1). EGF- and FGF-like signals are typically transmitted from the cell surface to the nucleus via Erk-like mitogen-activated protein kinase (MAPK) modules that, in mammals, comprise small GTPases of the Ras-family, Raf-like MAPK kinase kinases (MAPKKK), the two MAPK kinases (MAPKK) MEK1 and MEK2, and the MAPKs Erk1 and Erk2 [42]. In the E. multilocularis genome, one member of each of these signalling molecule families is encoded, and RT–PCR experiments as well as recent transcriptome analyses indicated that the respective genes are expressed throughout the E. multilocularis life cycle [4345]. Furthermore, by biochemical analyses, we could show that the GTPase EmRas, the Raf-like MAPKKK EmRaf, the MEK1/2-like MAPKK EmMKK2 and the Erk1/2-like MAPK EmMPK1 form a functional MAPK module [43, 45] that acts downstream of EmER [44, 45] and EmFR (Lorenz et al., unpublished data) (Fig. 1). Further, MAPK cascade signalling components that we have identified are the small GTPase EmRal [46], the MKK3/6-like MAPKK EmMKK1 [45] and the p38-like MAPK EmMPK2 [47]. Although it became clear from biochemical analyses than none of these components forms part of the parasite’s Erk-like MAPK pathway [4547], further investigations are necessary to define the MAPK cascade pathways in which they participate. Finally, members of the 14-3-3 and the PDZ-1-domain containing protein families were described in E. multilocularis [48, 49] that could both be involved in RTK signalling events. 14-3-3 proteins usually act as scaffolds in regulating protein–protein interaction and are well-established interaction partners of Raf-like MAPKKKs in mammals [50]. Interestingly, we could recently demonstrate that one of the E. multilocularis 14-3-3 isoforms, Em14-3-3.2, interacts with EmMKK2, which forms part of the Erk-like MAPK cascade [45]. Since PDZ-domain-containing proteins can act as linkers between RTKs such as the EGF receptor and cytoskeleton elements [51], the so far only known member of this family in E. multilocularis, EmPDZ-1, could fulfil a similar role between EmER and the ezrin-radixin-moesin family component Elp [49].
https://static-content.springer.com/image/art%3A10.1007%2Fs00430-010-0154-1/MediaObjects/430_2010_154_Fig1_HTML.gif
Fig. 1

Schematic representation of the E. multilocularis Erk-like MAPK module and its activation through host-derived EGF. Direct interactions between EmRas and EmRaf, EmRaf and EmMKK2, as well as EmMKK2 and EmMPK1 have been demonstrated by yeast two-hybrid studies [43, 45]. Inhibition of EmRaf through BAY 43-9006 or EmMKK2 through PD 184352 leads to de-phosphorylation of EmMPK1 and growth arrest [47]. Addition of exogenous host EGF to metacestode vesicles leads to an induction the parasite’s Erk-like MAPK cascade and supports parasite development [44]. E. multilocularis orthologs to Grb (growth factor receptor bound protein) and SOS (son of sevenless) have not yet been biochemically characterized but are encoded by the parasite’s genome [4] and most likely form a link between EmER and EmRas

Direct evidence that Echinococcus RTKs can interact with corresponding host cytokines has been obtained for EmIR. Using the yeast two-hybrid system, we could show that the extracellular ligand-binding domain of EmIR is capable of interacting with mammalian insulin, but not with IGF [38]. Similar results could be obtained for the insulin receptors from S. mansoni [40] and, very recently, also for those of S. japonicum (You et al., submitted for publication). Since we have already observed positive effects of insulin on parasite development in the E. multilocularis in vitro systems for primary cells and mature metacestode vesicles (C. Konrad et al., unpublished observations), EmIR could indeed play a crucial role for the initiation of parasite development within the host’s liver. Furthermore, since schistosome maturation from schistosomules to adults also requires passage through the host’s liver, a similar role could be fulfilled by the trematode counterparts of EmIR [40].

At least indirect evidence that EmER can interact with host-derived EGF has been obtained by us in in vitro cultivation experiments. When exogenously added to metacestode vesicles, mammalian EGF resulted in a significant induction of the phosphorylation of EmMPK1, the central component of the parasite Erk-like MAPK cascade [44]. This could be prevented in the presence of inhibitors specific for Raf- and MEK-like kinases, leading to growth arrest of the vesicles [45]. Hence, host EGF might directly stimulate the parasite’s EmRas/EmRaf/EmMKK2/EmMPK1 MAPK module via binding to EmER, thus supporting parasite growth and development (Fig. 1). That flatworm EGF receptors are principally able to functionally interact with mammalian EGF has been demonstrated by Vicogne et al. [41] who obtained direct stimulation of S. mansoni SER by human EGF upon expression in the Xenopus oocyte system. Like in the case of E. multilocularis, it appeared to be the classical Erk-like MAPK cascade that was activated in S. mansoni in response to EGF [41].

Receptor serine/threonine kinase signalling in Echinococcus

Cytokines of the TGF-β/BMP family signal through surface RSTKs of which two subfamilies exist, the type I and the type II receptors. Upon formation of the receptor ligand complex, the type II receptor phosphorylates, and thereby activates, the type I receptor at the so-called GS domain that is located in the type I receptor’s intracellular domain. Once activated, the type I receptor recruits, phosphorylates and activates downstream Smad signalling components [5254]. On the basis of the sequence of the L45-loop of the type I receptors, which is important for the interaction with downstream factors, several families of type I receptors can be distinguished. Of these, the ALK5 receptor group is typically activated by the TGF-β/Activin subfamily of ligands, the ALK3 group exclusively by BMP family ligands and the ALK1 group by ligands of both subfamilies [5254]. In E. multilocularis, we have previously characterized the type I receptor EmTR1, which is a member of the ALK1 group on the basis of its L45-loop structure [55]. Similar to the situation in EmIR, EmTR1 displays significant homologies to known TGF-β/BMP receptors in the tyrosine kinase and ligand-binding domains but contains extensive inserts in the tyrosine kinase domain, the biochemical function of which is unknown [55]. Encoded on the parasite’s genome are another two receptors of the type I family, of which one presumably transmits TGF-β/Activin signals, whereas the other is, like EmTR1, a member of the ALK1 receptor family, and a single type II receptor [4]. According to our own transcriptional analyses, all four receptors are expressed throughout the parasite’s larval development from primary cells to protoscoleces ([55] and unpublished results).

As already mentioned, the most important downstream signalling components of TGF-β/BMP family receptors are the Smad transcription factors, of which three sub-families exist. The so-called receptor-regulated Smads (R-Smads) are directly phosphorylated by type I receptors at a highly conserved, C-terminal pSXpS motif [54]. In mammals, five R-Smads are expressed, of which two, Smad2 and Smad3, are exclusively involved in the transmission of signals of the TGF-β/activin family (they are therefore called AR-Smads), whereas Smad1, Smad5 and Smad8 exclusively transmit BMP signals (BR-Smads). In E. multilocularis, three R-Smads could so far be identified and were closer analysed by biochemical and cell biological methods. On the basis of sequence homologies, two of these R-Smads, EmSmadA and EmSmadC, belong to the AR-Smad family, whereas EmSmadB is a BR-Smad [56, 57]. Interestingly, the two Echinococcus AR-Smads display some biochemical peculiarities. While R-Smads typically contain two relatively well-conserved sub-domains called MH1 (N-terminal) and MH2 (C-terminal), EmSmadA and EmSmadC lack any discernable MH1-domain [56, 57]. Since the MH1-domain is usually conferring DNA binding specificity, it has already been speculated that EmSmadA and EmSmadC could have a broadly overlapping spectrum of target promoters [57]. This does not, however, indicate that they are functionally redundant, since remarkable differences have been identified concerning the upstream activation process of both Smads. While EmSmadC is exclusively activated by type I receptors of the TGF-β/activin subfamily when heterologously expressed in HEK 293 cells [57], as one would expect from its structure, EmSmadA is activated by both type I receptors of the TGF-β/activin and by those of the BMP-subfamily [56]. Hence, the strict separation between the TGF-β/activin and the BMP pathways that is typical for vertebrates seems to be relaxed in E. multilocularis (and maybe also other invertebrates). Unlike EmSmadA and EmSmadC, the E. multilocularis BR-Smad EmSmadB contains MH1 and MH2 domains and, upon heterologous expression in HEK 293 cells, was exclusively activated through BMP receptors [57], indicating that it functionally resembles all other BR-Smad members that have been identified so far. Upon release of the first assembly version of the E. multilocularis genome [4], we could recently identify another BR-Smad, EmSmadE, which is expressed throughout the parasite’s life cycle. Like in the case of EmSmadB, EmSmadE contains MH1 and MH2 domains and is exclusively activated by BMP receptors (K. Epping et al., unpublished data). Hence, the biochemical peculiarities that have been observed for the parasite’s AR-Smads seem not to apply to its BR-Smad system.

Two further Smad sub-families are the common-mediator Smads (Co-Smads), such as vertebrate Smad4, and the inhibitory Smads (I-Smads), such as Smad6 and Smad7. I-Smads lack the N-terminal MH1 domain as well as the C-terminal pSXpS motif and primarily act as antagonists of TGF-β/BMP signalling by counteracting the functions of the R-Smads. So far, no I-Smad has been biochemically characterized in E. multilocularis, although evidence for the expression of at least one member of this protein family has been obtained through genomic and transcriptomic analyses (Epping and Brehm, unpublished data). Although Co-Smads contain the conserved MH1 and MH2 domains, they lack the C-terminal pSXpS motif and are not phosphorylated by upstream receptors. Instead, they form heteromeric complexes with both AR- and BR-Smads once these are activated. The heteromeric R-Smad/Co-Smad complexes are then translocated into the nucleus and regulate the transcription of target genes [54]. In E. multilocularis, one single Co-Smad, EmSmadD, is expressed, and biochemical analyses showed that EmSmadD is able to undergo heteromer formation with the three R-Smads, EmSmadA-C [57]. Interestingly, in the linker region between the conserved MH1 and MH2 domains, EmSmadD contains putative target phosphorylation sites for members of the MAPK family, and direct phosphorylation of EmSmadD in this region through EmMPK1 was demonstrated [57]. Hence, EmSmadD is most likely a key factor for cross-regulation between the Erk-like MAPK cascade and TGF-β/BMP signalling pathways in E. multilocularis. Further evidence for cross-regulation between E. multilocularis TGF-β/BMP signalling and other pathways has been obtained through yeast two-hybrid analyses showing that EmSmadC (but no other E. multilocularis Smad) can interact with EmSKIP [57, 58], a member of the SNW/SKIP family of transcriptional co-regulators [58]. Since the SNW/SKIP family comprises important transcriptional co-regulators for members of the nuclear hormone receptor (NHR) family, which are typically involved in the signal transduction of lipophilic hormones [59], the formation and action of heteromeric EmSmadB/EmSmadD-complexes might be one Echinococcus signalling event on which information input from several different pathways converges.

Like in the case of Echinococcus RTKs and cognate ligands of the host, we have already obtained direct evidence that the parasite’s TGF-β/BMP receptors can functionally interact with host cytokines. Upon heterologous expression of EmTR1 in HEK 293 cells, together with a human type II receptor of the BMP receptor family, specific phosphorylation of EmSmadB (but not EmSmadA and EmSmadC) at the conserved pCXpC motif was obtained in presence of exogenously added human BMP2 [55, 57]. We meanwhile also obtained evidence that this was not dependent on the human type II receptor component, since EmTR1 is also able to form a functional BMP2-responsive receptor complex with the Echinococcus type II receptor (Bernthaler et al., unpublished observations). However, although the parasite’s TGF-β/BMP receptors are clearly responsive to host cytokines when expressed in HEK 293 cells, we have so far never observed significant phenotypic effects on the development of Echinococcus primary cells or metacestode vesicles when host BMP2 was added in vitro (unpublished observations). This somehow reflects the situation in S. mansoni, in which a TGF-β type I/type II receptor complex is expressed at the surface of adult worms [60]. Although it could be demonstrated that this complex interacts with human TGF-β in activation assays [60, 61], no phenotypic changes have yet been reported when schistosomes were exposed to host-derived TGF-β in vitro, and the expression of a TGF-β-regulated gene in S. mansoni, SmGCP, was only slightly induced in response to exogenously added host cytokine [60]. Taken together, these data from E. multilocularis and S. mansoni demonstrate that the flatworm TGF-β/BMP receptor systems are principally able to interact with host-derived TGF-β/BMP cytokines, but it is presently unclear whether respective ligand–receptor complexes are formed during an infection or whether the host cytokines affect parasite development.

The fact that the TGF-β/BMP ligand–receptor systems of cestodes, trematodes and vertebrates are compatible is not only relevant for questions concerning the influence of the host on parasite development, but also for influences of flatworm parasites on the immune response and the physiology of the host. As already mentioned, ligands of the TGF-β/activin families are important regulators of the immune response, and BMP cytokines are involved in the regulation of extracellular matrix production, iron homoeostasis and regeneration within the liver [3134]. Hence, through the secretion of such cytokines towards host tissue during an infection, flatworm parasites could affect host cell reactions in their favour. Such a situation has already been reported for the nematode Brugia malayi, the adult stage of which secretes a member of the TGF-β cytokine family, Tgh-2, that is able to activate TGF-β-receptors of the host, thus possibly playing an important role in immuno-suppressive activities of the parasite [62]. Due to the availability of extensive genome information for S. mansoni [63] and E. multilocularis [4], TGF-β/BMP-like ligands could recently also be identified in flatworm parasites. In S. mansoni, the expression of one member of the BMP-subfamily, SmBMP [64], and one of the TGF-β/activin family, SmInACT [27], was reported, and in the genome of E. multilocularis, two BMP-like cytokines and one activin-like factor are encoded [4]. Although these cytokines are surely involved in the regulation of parasite development, as already demonstrated for SmInACT concerning S. mansoni embryo formation [65], they might also be released to affect host cells via binding to TGF-β/BMP receptors. Investigations whether this is indeed the case are currently underway.

It should be mentioned in this context that evolutionarily conserved protein–protein interaction patterns might also play important roles in host–helminth interactions that are not directly associated with signalling and cell–cell communication. In E. multilocularis, we have recently identified a protein, EmABP, which displayed considerable amino acid sequence homologies (60% identical residues) to mammalian AI-BP (apolipoprotein A-I (apoA-I) binding protein; [66]). Although the parasite itself does not express apolipoprotein-like molecules, EmABP was able to biochemically interact with host-derived apoA-I, which is an important component of high-density lipoprotein particles that are involved in reverse cholesterol transport [66]. Furthermore, we could show that EmABP is released into the extra-parasitic environment by metacestode vesicles and that it can also have access to host-derived apoA-I within hydatid fluid [66]. Since E. multilocularis, like other parasitic flatworms, is not able to synthesize cholesterol de novo, secreted EmABP could therefore have an important role in cholesterol uptake mechanisms of the parasite [66].

The E. multilocularis whole genome sequencing project

In cooperation with the Parasite Sequencing Unit of the Wellcome Trust Sanger Institute (Hinxton, Cambridge, UK), led by Matt Berriman, my group has recently initiated a whole genome sequencing project for E. multilocularis that is currently in a very advanced stage (introduced in more detail in [4]). By combining classical capillary sequencing on BAC (bacterial artificial chromosome) libraries with modern techniques such as 454- and Solexa-sequencing, ~140-fold sequence coverage of the genome has so far been obtained and, according to the latest assembly version, the sequence information is now present within 1841 contigs and 597 supercontigs, with 50% of the parasite’s genome assembled into 17 scaffolds of more than 1.6 Mb. The current assembly, including BLAST analysis tools, is publicly available under http://www.sanger.ac.uk/Projects/Echinococcus/. In addition to the E. multilocularis project, efforts are presently ongoing to also sequence the genomes of other cestodes such as T. solium [67], E. granulosus and Hymenolepis microstoma (http://www.sanger.ac.uk/Projects/Helminths/) so that comprehensive information on cestode genome and gene structures should soon be available. Annotation is supported by extensive EST data from both E. multilocularis and E. granulosus [12, 13, 68]. Furthermore, the entire transcriptome of several key stages of the E. multilocularis life cycle is currently being determined using an Illumina sequencing platform (Zheng et al., manuscript submitted). So far, extensive transcriptome data have been obtained for regenerating parasite vesicles (1 week primary cell cultures), for mature metacestode vesicles and for dormant as well as activated (low pH; pepsin) protoscoleces. This will soon be complemented by transcriptome data of primary cell cultures at later stages, protoscolex-forming metacestode vesicles (with brood-capsules), the adult stage and the oncosphere. Taken together, these investigations will provide valuable information on stage-specific gene expression patterns, covering the entire E. multilocularis life cycle, in due time.

By using the available genome sequence information, a detailed characterization of evolutionarily conserved core-signalling pathways in E. multilocularis is now possible. On the one hand, components of RTK-, RSTK- and MAPK-signalling pathways that have so far not been addressed (e.g. the JNK branch of MAPKs) are now easier accessible. Furthermore, investigations into the cross-communication of the parasite’s MAPKs and Smad transcription factors with other evolutionarily conserved pathways will be facilitated. Particularly important in this context would be the Delta/Notch-, Hedgehog (Hh)- and Wingless-related (Wnt) signalling systems since these are known to significantly modify the processing of host-signals that are transmitted through RTKs and RSTKs [69]. Components of each of these pathways, such as several frizzled-, Notch- and patched-like receptors as well as a number of Wnt-, jagged- and Hh-like ligands, have already been identified on the parasite’s genome [4], and at least from transcriptome data, Delta/Notch and Hh-signalling appear to be important for metacestode development, whereas Wnt-signalling is probably particularly relevant for the differentiation towards the protoscolex stage (Brehm, unpublished data). Finally, since the RTK, RSTK, Wnt, Hh and Notch pathways are primarily expected to play a role in mechanisms of stem cell differentiation [4], it will also be interesting to identify pathways that are important for stem cell maintenance and self-renewal. In free-living flatworms, which have emerged as experimental models for stem cell research during recent years, a number of factors that control neoblast maintenance have already been characterized and frequently form part of the cellular machinery that controls micro-RNA-dependent translation and RNA decay such as DEAD-box family RNA helicases or members of the PIWI/Argonaute family of proteins [4]. Supporting the notion that the germinal cells of cestodes are the functional equivalents of the neoblasts of free-living flatworms, close orthologs of these proteins are encoded by the E. multilocularis genome and are particularly well expressed in the primary culture system, which contains the highest proportion of stem cells when compared to mature metacestode vesicles or protoscoleces [4].

Echinococcus signalling systems as targets for anti-parasitic chemotherapy

Due to the tumour-like, infiltrative growth of the E. multilocularis metacestode and the relatively close phylogenetic relationship between parasite and host, there are currently only very limited options for adequate treatment of AE. Besides surgical removal of parasite tissue, which is only possible in ~20% of cases, treatment of patients with anthelminthic benzimidazole derivatives is the most important form of AE therapy [70]. Benzimidazoles have a high affinity for helminth-specific β-tubulin isoforms, thus inhibiting microtubule polymerization that eventually leads to parasite death [11]. However, in AE, this treatment is mostly parasitostatic rather than parasitocidal and, as a consequence, chemotherapy has to be given life long. Furthermore, long-term benzimidazole chemotherapy is often associated with severe side effects [70]. This is due to the interaction of the drugs with β-tubulin of the host, which, on amino acid sequence level, is more than 90% identical to that of Echinococcus [71]. Although there have been several recent reports on the activity of other drugs such as amphothericin B, isoflavones or nitazoxanide on E. multilocularis larvae cultivated in vitro [11], no reliable chemotherapeutic alternative can presently be given to patients who do not tolerate or do not respond to benzimidazoles. Thus, it has become evident that new chemotherapeutic strategies against AE are urgently needed.

On the one hand, the E. multilocularis genome project would now offer a unique opportunity to identify parasite-specific proteins as targets for anti-parasitic chemotherapy that could minimize the occurrence of side effects. However, there is generally no knowledge on the precise role of such parasite-specific proteins, whether they are necessary for parasite survival and development, how they function biochemically or how their activities can be inhibited by small-molecule compounds. Hence, from the identification of parasite-specific target proteins to the development of effective chemotherapeutic compounds, there is usually a very long way to go, involving considerable investments of time and money, and an uncertain outcome.

Evolutionarily conserved signalling pathways, on the other hand, although present in both host and parasite, are decisively involved in cellular survival and development and, particularly in the case of RTK, RSTK and MAPK signalling, are extremely well studied from the biochemical point of view. In humans, the majority of these pathways are involved in cancer formation, and in the course of cancer research, a plethora of small-molecule compounds to inhibit or modify their functions has already been developed. The RTK and RSTK signalling pathways of helminth parasites have therefore already been suggested as promising targets for chemotherapeutic treatment [1, 8, 37], and the challenge of the coming years will be to identify helminth signalling components that display sufficient functional or structural differences to their mammalian counterparts so that strictly parasite-specific compounds can be designed.

Towards this end, we have recently targeted components of the Erk-like MAPK module of E. multilocularis using the ATP-competitive small-molecule compounds BAY 43-9006 (also called Sorafenib) and PD184352 [45] (Fig. 2). BAY 43-9006 is known as a potent inhibitor of Raf-like MAPKKKs [72] and, when added to in vitro cultivated metacestode vesicles at a concentration of 100 μM, effective de-phosphorylation of the Echinococcus Erk-like MAPK EmMPK1 was obtained [45], indicating that EmRaf, the parasite’s only member of the Raf MKKK family [4, 43], is acting upstream of EmMPK1. Likewise, incubation of metacestode vesicles with PD184352, a potent and highly specific inhibitor of MEK1/2-like MAPKKs [73], resulted in EmMPK1 de-phosphorylation, which was mediated through inhibition of EmMKK2 [45]. Interestingly, however, although the parasite’s Erk-like MAPK module could be effectively inhibited by this treatment, no vesicle killing occurred, even after prolonged periods of treatment. The vesicles merely ceased growth and proliferation [45], indicating that BAY 43-9006 and PD184352 could be used as parasitostatics but are only of limited use as parasitocidal compounds against E. multilocularis. Due to the decisive role of Erk-like MAPK modules in the differentiation, but not in self-renewal, of stem cells [74], it is highly likely that the Raf and MEK inhibitors merely inhibited the differentiation of E. multilocularis stem cells towards cell types that fuse with the germinal layer and thus contribute to vesicle proliferation, whereas the parasite stem cells themselves were not affected. In the primary cell culture system, BAY 43-9006 and PD184352 could therefore potentially be used as compounds to generate relatively pure cultures of constantly self-renewing germinal cells that do not differentiate to other Echinococcus cell types [45], a possibility that is currently tested in my laboratory.
https://static-content.springer.com/image/art%3A10.1007%2Fs00430-010-0154-1/MediaObjects/430_2010_154_Fig2_HTML.gif
Fig. 2

Small-molecule compounds affecting E. multilocularis development in vitro. The pyridinyl imidazole compounds ML3403 and SB202190 inhibit the activity of EmMPK2 in vitro and lead to effective killing of parasite vesicles after 4 days of incubation [47]. Incubation of metacestode vesicles with the compounds BAY 43-9006 (Sorafenib; inhibitor of Raf-family MAPKKK) or PD184352 (specific inhibitor of MEK-like MAPKKs) leads to an effective de-phosphorylation of EmMPK1 and to growth arrest [45]

Much more promising than the Erk-like MAPK module for the development of novel anti-parasitic drugs proved to be the E. multilocularis p38-like MAPK EmMPK2 [47]. EmMPK2 displays a significant overall sequence homology to human p38 MAPKs (~53% identical residues) and contains a TGY-motif within the activation loop that is characteristic for p38 MAPKs. Interestingly, however, when EmMPK2 was heterologously expressed in E. coli, it displayed much higher enzyme activities than human p38 MAPK expressed under the same conditions. Furthermore, EmMPK2 lacked the so-called ‘common docking domain’, which, in other p38 MAPKs, is necessary for the activation through upstream MAPKKs [47] and, in yeast two-hybrid studies, did not interact with any of the parasite’s MAPKKs [45]. Taken together, these data indicated that EmMPK2 is a de-regulated, constitutively active p38 MAPK, which is supported by the fact that it displays several distinct differences to other p38 MAPK, particularly in amino acids known to be involved in the regulation of enzyme activity [47]. Most interestingly, EmMPK2 could be effectively inhibited in in vitro activity assays and within parasite vesicles by pyridinyl imidazole compounds (Fig. 2) that have originally been developed against the human p38 MAPK isoforms, and one of these molecules, ML3403, effectively killed parasite vesicles in vitro at concentrations (5 μM) that did not affect cultured mammalian cells [47]. Hence, pyridinyl imidazoles are a novel and potent class of anti-Echinococcus compounds that act against the parasite’s p38 MAPK. Since several hundred of compounds related to ML3403 have already been synthesized [75], we are currently trying to identify those that display an even higher specificity for the EmMPK2 when compared to human p38 MAPK isoforms. Once identified, these small-molecule compounds can then be tested in experimentally infected mice, either alone or in combination with benzimidazoles, for their potential as an anti-AE therapeutic.

Conclusions

The past decade has witnessed a great increase in our methodological repertoire to undertake molecular studies on E. multilocularis development and host–parasite interaction mechanisms. The parasite cell population that decisively drives larval and adult development within the host, the totipotent germinal cells, can now be directly addressed from the cell biological point of view using the primary cell in vitro culture system. Extensive genomic and transcriptomic data, covering large parts of the parasite’s life cycle, are already available and will greatly facilitate investigations concerning signalling pathways that regulate stem cell differentiation and self-renewal. This is supported by recent success in the development of methods to genetically manipulate germinal cells through RNA interference, virus-based transduction systems and intracellular bacteria as gene delivery systems [4, 8, 19]. Given the urgent need for novel, improved chemotherapeutic treatment options against AE and the decisive role of evolutionarily conserved signalling systems in parasite development, concentrated efforts to target these pathways by small-molecule compounds should pay out since the resulting drugs can either be directly used in antiparasitic chemotherapy or at least form promising lead structures for the design of parasite-specific chemotherapeutics. Pyridinyl imidazoles are only one of these promising compound families. Our own preliminary studies already showed that also the parasite’s RTKs can be effectively targeted by ATP-competitive drugs, and by exploiting their biochemical and structural differences to mammalian RTKs (e.g. tyrosine kinase insert regions), the design of parasite-specific RTK inhibitors should be possible. Since the genomic analyses undertaken so far indicate that the signalling pathways of E. multilocularis, E. granulosus, T. solium and even schistosomes are very well conserved, this could even result in the development of drugs with broad-spectrum activity against flatworm parasites. The respective compounds are already produced and available, we just have to find them.

Acknowledgments

I wish to thank all those colleagues and students working on their theses in my laboratory at the Institute of Hygiene and Microbiology (University of Würzburg) who, over the last 7 years, have contributed to the results mentioned in this review. Katja Klöpper, Dirk Radloff and Monika Bergmann are thanked for excellent technical assistance during this time. Work of the authors was supported by the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich 479.

Copyright information

© Springer-Verlag 2010