Organisms and culture methods

FGF stimulation and inhibitor experiments were performed with the natural E. multilocularis isolate H95 [14]. Whole mount in situ hybridization was carried out using isolate GH09 which, in contrast to H95, is still capable of producing brood capsules and protoscoleces in vitro [14]. All isolates were continuously passaged in mongolian jirds (Meriones unguiculatus) as previously described [9]. The generation of metacestode vesicles and axenic cultivation of mature vesicles was performed essentially as previously described [7,9] with media changes usually every three days. Primary cell cultures were isolated from mature vesicles of isolate H95 and propagated in vitro essentially as previously described [8–10] with media changes every three days unless indicated otherwise. For FGF stimulation assays, 10 nM or 100 nM of recombinant human acidic FGF (FGF1) or basic FGF (FGF2)(both from ImmunoTools GmbH, Friesoythe, Germany) were freshly added to parasite cultures during medium changes. In the case of primary cells, cultivation was usually performed in cMEM medium which is host hepatocyte-conditioned DMEM (prepared as described in [10]). For inhibitor studies, specific concentrations of BIBF 1120 (Selleck Chemicals LLC, Houston, TX, USA) were added to parasite cultures as indicated and as negative control DMSO (0.1%) was used. The formation of mature metacestode vesicles from primary cells and measurement of metacestode vesicles size was performed essentially as previously described [11].

Nucleic acid isolation, cloning and sequencing

RNA isolation from in vitro cultivated axenic metacestode vesicles (isolate H95), protoscoleces (isolate GH09), and primary cells (H95, GH09) was performed using a Trizol (5Prime, Hamburg, Germany)-based method as previously described [11]. For reverse transcription, 2 μg total RNA was used and cDNA synthesis was performed using oligonucleotide CD3-RT (5’-ATC TCT TGA AAG GAT CCT GCA GGT₂₆ V-3’). PCR products were cloned using the PCR cloning Kit (QIAGEN, Hilden, Germany) or the TOPO XL cloning Kit (invitrogen), and sequenced employing an ABI prism 377 DNA sequencer (Perkin-Elmer).

The full-length emfr1 cDNA was cloned using as starting material the partial sequence of a cDNA of the closely related cestode E. granulosus, which encoded a FGFR-like TK domain but which lacked the coding regions for transmembrane and extracellular parts [28]. Using primers directed against the E. granulosus sequence (5’-CTA CGC GTG CGT TTT CTG ATG-3’for first PCR; 5’-CCC TCT GAT CCA ACC TAC GAG-3’for nested PCR), the 3’ end of the corresponding E. multilocularis cDNA was subsequently PCR amplified from a metacestode (isolate H95) cDNA preparation using primers CD3 and CD3nest as previously described [29]. 5’-RACE was performed using the SMART RACE cDNA amplification kit (Clontech) according to the manufacturer’s instructions using primers 5’-ACC GTA TTT GGG TTG TGG TCG-3’ (first PCR) and 5’-GAA CAG GCA GAT CGG CAG-3’ (touchdown PCR) as previously described [30]. The presence of an in frame TAA stop codon 110 bp upstream of the emfr1 ATG start codon indicated that the correct 5’ end had been identified. In a final step, the entire emfr1 cDNA was PCR amplified from metacestode cDNA using primers 5’-GAC ACA TCT CCT TGG CCG-3’ and 5’-GCG AGT TGA TAC TTT ATG AGA G-3’ and cloned using the TOPO XL PCR cloning kit (Invitrogen). The sequence is available in the GenBank^TM, EMBL, and DDJB databases under the accession number LT599044.

For emfr2 cloning we first identified by BLAST analyses on the published E. multilocularis genome sequence [14] a reading frame encoding a FGFR-like TKD annotated as EmuJ_000196200. Transcriptome analyses [14] and 5’_RACE experiments, however, indicated that there is actually read-through transcription between gene models EmuJ_000196300 and EmuJ_000196200. We thus designed primers 5’-ATG TGT CTC CGA GCT CTC TG-3’, binding to the 5’ end regions of gene model EmuJ_000196300, and primer 5’-TTA CTC GCT CGA TCG TGG GG-3’, binding to the reading frame 3’ end of gene model EmuJ_000196200, to PCR amplify the entire reading frame from metacestode cDNA. The resulting PCR fragment was subsequently cloned using the TOPO XL cloning kit (Invitrogen) and fully sequenced. The sequence is available in the GenBank^TM, EMBL, and DDJB databases under the accession number LT599045.

For emfr3 cloning and sequencing we used primers directed against the CDS 5’ end (5’-ATG GCA CCT AAG GTT GTG TCA GGA-3’) and 3’ end (5’-GCA GAT GAG TAA GAA ACC CTC-3’) of gene model EmuJ_000893600 [14] for direct PCR amplification of the reading frame from metacestode cDNA. The resulting PCR fragment was subsequently cloned using the TOPO XL cloning kit (Invitrogen) and sequenced. The sequence is available in the GenBank^TM, EMBL, and DDJB databases under the accession number LT599046.

BrdU proliferation assays

Proliferation of E. multilocularis metacestode vesicles and primary cells was assessed by a bromodesoxyuridine (BrdU)-based method. Axenically cultivated metacestode vesicles (2–4 mm in diameter) were manually picked and incubated in 12-well plates (Greiner BioOne, Kremsmünster, Germany; 8 vesicles per well) in DMEM medium without serum for 2 days. Freshly isolated primary cells were plated on 12-well plates and grown for 2 days under axenic conditions in conditioned DMEM (cMEM) medium with serum [8]. BrdU (SigmaAldrich, taufkirchen, Germany) as well as recombinant human FGF1 and FGF2 were added at 1mM (BrdU) and 100 nM or 10 nM (FGF1, FGF2) final concentrations as indicated. Cultures were incubated for 48 h at 37°C under 5% CO₂ for metacestode vesicles or under nitrogen atmosphere [7,8] in the case of primary cells. Samples were analysed in duplicates in three independent experiments. As controls, metacestode vesicles or primary cells were incubated in either DMEM without serum or conditioned DMEM, without addition of FGFs.

Primary cells and metacestode vesicles were then isolated for genomic DNA analysis. In detail, vesicles and primary cells were first washed with 1xPBS, pelleted, and subsequently transferred to lysis buffer (100 mM NaCl, 10 mM Tris-HCl, pH 8.0; 50 mM EDTA, pH 8.0, 0,5% SDS) supplemented with 20 μg/ml RNAse A and 0,1 mg/ml proteinase K. Samples were then incubated at 50°C for 4 h under constant shaking for complete lysis. DNA was isolated by two rounds of phenol/chlorophorm extraction (1 vol of phenol/chlorophorm/isoamylalcohol 25:24:1). DNA was then precipitated with 2 vol of 96% ethanol and 0,1 vol of LiCl (pH 4,5) after overnight incubation at -20°C and centrifugation at 20.000 rcf for 30 min at 4°C and washed with 70% ethanol. The pellet was then air dried for 15 min an resuspended in 1 x TE buffer (10 mM Tris, 1 mM EDTA, pH 8,0).

The DNA was then prepared for coating onto a 96-well plate (96 well optical bottom plates, Nunc, Langenselbold, Germany). To this end, 5 μg of DNA were combined with 1 vol of Reacti-Bind DNA Coating solution (Pierce Biotechnology, Rockford, IL, USA) and mixed for 10 min. The DNA mix was then added to the microplates in duplicates and incubated overnight at room temperature with gentle agitation. The TE/Reacti-Bind DNA coating solution mix served as a negative control. Unbound DNA was removed by washing three times with 1xPBS. After blocking with 5% nonfat dry milk in 1xPBS for 1 h at room temperature and extensive washing with 1xPBS, 100 μl of anti-BrdU-POD (Cell Proliferation ELISA, BrdU; Roche Applied Science, Mannheim, Germany) was added and incubated for 90 min at room temperature. After incubation, microplates were washed three times with 1xPBS buffer before substrate solution (Cell Proliferation ELISA, BrdU; Roche Applied Science, Mannheim, Germany) was added and the wells were incubated for 60 min. Stop-solution (25 μl of 1 M H₂SO₄) was added and absorbance of the samples was measured using an ELISA reader at 450 nm.

In situ hybridization protocols

Coding sequences of FGF receptors from E. multilocularis were amplified by RT-PCR and cloned into the vectors pDrive (Qiagen) or pJET1.2 (Thermo Fisher). In the case of emfr2, the full length coding sequence was amplified using primers 5’-ATG TGT CTC CGA GCT CTC TG-3’(forward) and 5’-TTA CTC GCT CGA TCG TGG GG-3’ (reverse), whereas partial coding sequences were amplified for emfr1 (using forward primer 5’-GCA GTG GGC GTC TTC TTT CAC-3’ and reverse primer 5’-GTA AAT GTG GGC CGA CAC TCA G-3’) and for emfr3 (using forward primer 5’-TTG CCC AGT CAT CCG CTA CAA G-3’ and reverse primer 5’-GCA AGC GGT CAT GAG GCT GTA G-3’). The recombinant plasmids were used for in vitro transcription of digoxigenin-labelled RNA probes as previously described [6]. These probes were used for fluorescent WMISH of in vitro cultured E. multilocularis larvae as described in [6]. Control WMISH experiments using the corresponding sense probes were always negative.

Xenopus oocyte expression assays

For expression in the Xenopus system, the emfr1, emfr2, and emfr3 coding sequences without predicted signal peptide information were cloned into the pSecTag2/Hygro expression system (ThermoFisher Scientific, Germany) leading to an in frame fusion of the Igk leader sequence (provided by the vector system) and the E. multilocularis FGF receptor sequences under control of the T7 promoter. Capped messenger RNAs (cRNA) encoding EmFR1, EmFR2 and EmFR3 were then synthesized in vitro using the T7 mMessage mMachine Kit (Ambion, USA). Microinjection of EmFGFR cRNAs (60 ng in 60 μl) was performed in stage VI Xenopus laevis oocytes according to the procedure previously described [31]. Following 48h of receptor expression, human FGF1 or FGF2 (R & D systems, UK) were added to the extracellular medium at the final concentration of 10 nM. cRNA of Pleurodeles FGFR1 identified as homologous to human receptor [32] was a gift of Shi D.L. (CNRS UMR 722, Paris VI) and was used as a positive control.

In some experiments, BIBF1120 (stock solution 10mM in DMSO, Selleck Chemicals LLC) was added (0.1 to 20 μM final concentration) 1 h before the addition of 10 nM FGF1 or FGF2 on EmFR1, EmFR2, EmFR3 and Pleurodeles FGFR1 expressing oocytes. Following 15 h of FGF1 or FGF2 stimulation, oocytes were analyzed for their state of progression in the cell cycle. The detection of a white spot at the animal pole of the oocyte attested to G2/M transition and GVBD. Non-injected oocytes treated with progesterone (10 μM) were used as positive controls of GVBD. For each assay, sets of 20–30 oocytes removed from 3 different animals were used.

Dead kinase (TK-) receptors were obtained by site-directed mutagenesis of the EmFR1, EmFR2 and EmFR3 constructs. The active DFG sites present in EmFR1 (D₄₄₂FG), EmFR2 (D₆₄₇FG) and EmFR3 (D₇₀₁FG) were replaced by an inactivating motif (DNA) as described in [31].

For western blot analysis, oocytes were lysed in buffer A (50 mM Hepes pH 7.4, 500 mM NaCl, 0.05% SDS, 5 mM MgCl2, 1 mg ml−1 bovine serum albumin, 10 μg ml−1 leupeptin, 10 μg ml−1 aprotinin, 10 μg ml−1 soybean trypsin inhibitor, 10 μg ml−1 benzamidine, 1 mM PMSF, 1 mM sodium vanadate) and centrifuged at 4°C for 15 min at 10,000 g. Membrane pellets were resuspended and incubated for 15 min at 4°C in buffer A supplemented with 1% Triton X-100 and then centrifuged under the same conditions. Supernatants were analyzed by SDS-PAGE. Proteins were transferred to a Hybond ECL membrane (Amersham Biosciences, France). Membranes were incubated with anti-myc (1/50 000, Invitrogen France) or anti-PTyr (1/8000, BD Biosciences, France) antibodies and secondary anti-mouse antibodies (1/50 000, Biorad, France). Signals were detected by the ECL advance Western blotting detection kit (Amersham Biosciences, France)

MAPK cascade activation assays

Axenically cultivated metacestode vesicles of about 0.5 cm in diameter were incubated in DMEM medium with or without 10% FCS for 4 days. Vesicles cultivated without FCS were subsequently incubated with 10 nM FGF1 (aFGF) or 10 nM FGF2 (bFGF) for 30 sec, 60 sec or 60 min. Immediately after stimulation, vesicles were harvested, cut by a scalpel to remove cyst fluid and then subjected to protein isolation as described previously [33]. Isolated protein lysates were then separated on a 12% acrylamide gel and analysed by Western blotting using a polyclonal anti-Erk1/Erk2 antibody (ThermoFisher Scientific; #61–7400), recognizing Erk-like MAP kinases in phosphorylated and non-phosphorylated form, as well as a polyclonal antibody against phospho-Erk1/Erk2 (ThermoFisher Scientific; #44-680G), specifically directed against the double-phosphorylated (activated) form of Erk1/Erk2 (Thr-185, Tyr-187). We had previously shown that these antibodies also recognize the Erk-like MAP kinase EmMPK1 of E. multilocularis in phosphorylated and non-phosphorylated form [33]. As secondary antibody, a peroxidase-conjugated anti-mouse IgG antibody (Dianova, Hamburg, Germany) was used. In inhibitor experiments, axenically cultivated metacestode vesicles were incubated with either 5 μM or 10 μM BIBF 1120 for 30 min and then processed essentially as described above.

Quantitative reverse transcription PCR (qRT-PCR)

qRT-PCR experiments have been performed as previously described by del Puerto et al. [34] with several modifications. Briefly, RNA samples were prepared with TRI-reagent and the DirectzolTM RNA Mini Prep Kit (Zymo Research, USA). Prior to RNA-isolation, in vitro cultured metacestode vesicles were deprived of stem cells by treatment with hydroxyurea (HU) and the Polo-like kinase inhibitor BI 2536 essentially as previously described by Koziol et al. [6] and Schubert et al. [35], respectively. Reverse transcription was performed with Omniscript RT Kit (Qiagen, Germany) from 500 ng of RNA using oligo dT23 as primers according to the manufacturer’s instructions. qRT-PCRs were then performed using primers Forward F_emfr1_qPCR (5’-CCG TAT GAA GGG AAA TGG TCG TGT T-3’) and Reverse R_emfr1_qPCR (5’-TGG TGA ATC GCC AAG GCT GAA A-3’) for emfr1, Forward F_emfr2_qPCR (5’-GGG AAT TTC CAA GGT CAT CAG GGA C-3’) and Reverse R_emfr2_qPCR (5’-ATC GTG GGG GCA CAA CAT AAT TGC-3’) for emfr2, as well as Forward F_emfr3_qPCR (5’-GTC TAC CTT GAG GAA ATT GCT GTG GTC-3’) and Reverse R_emfr3_qPCR (5’-CGT GAG GAA TGA CGC AGG C-3’) for emfr3. As a control gene for normalization, the constitutively expressed gene elp was used with primers as described previously [34]. qRT-PCR conditions and primer amplification efficiency control experiments were performed essentially as previously described [34]. The relative expression of emfr genes in each sample was estimated using the delt-delta-Ct method [36], normalising with elp expression levels using the StepOne software (Thermo Fisher). All experiments were carried out as technical triplicates.

Computer analyses and statistics

Amino acid comparisons were performed using BLAST on the nr-aa and swissprot database collections available under (https://www.genome.jp/). Genomic analyses and BLAST searches against the E. multilocularis genome [14] were done using resources at (https://parasite.wormbase.org/index.html). CLUSTAL W alignments were generated using MegAlign software (DNASTAR Version 12.0.0) applying the BLOSUM62 matrix. Domain predictions were carried out using the simple modular architecture research tool (SMART) available under (http://smart.embl-heidelberg.de/) as well as PROSITE scans available under (https://prosite.expasy.org/scanprosite/). Two-tailed, unpaired student’s T-tests were performed for statistical analyses (GraphPad Prism, version 4). Error bars represent standard error of the mean. Differences were considered significant for p-values below 0.05 (indicated by *).

Ethics statement

All experiments were carried out in accordance with European and German regulations on the protection of animals (Tierschutzgesetz). Ethical approval of the study was obtained from the local ethics committee of the government of Lower Franconia (permit no. 55.2 DMS 2532-2-354).

The E. multilocularis genome encodes FGF receptor kinases but no canonical FGF ligands

By cDNA library screening and mining of the available E. multilocularis genome sequence we identified a total of three Echinococcus genes encoding members of the FGFR family of RTK. A partial cDNA for a gene encoding a TK with homology to FGFRs was previously cloned for E. granulosus [28] and by RT-PCR amplification of metacestode cDNA as well as 5’-RACE, the entire cDNA of the E. multilocularis ortholog, designated emfr1 (E. multilocularis fibroblast growth factor receptor 1), was subsequently cloned. As shown in Fig 1, the encoded protein, EmFR1, contained an N-terminal export directing signal peptide, followed by one single Ig-like domain, a transmembrane region, and an intracellular TK domain (Figs 1 and S1). In the recently released E. multilocularis genome information [14], this gene was correctly predicted on the basis of genome and transcriptome data (E. multilocularis gene designation: EmuJ_000833200). In the upstream genomic regions of emfr1, no information encoding potential Ig-like domains was identified which, together with the presence of a signal peptide sequence upstream of the single Ig-like domain, indicated that EmFR1 indeed contained only one single Ig-like domain. Amino acid sequence alignments indicated that the kinase domain of EmFR1 contains all residues critical for enzymatic activity at the corresponding positions (S2 Fig) and SWISS-PROT database searches revealed highest similarity between the EmFR1 kinase domain and that of human FGFR4 (42% identical aa; 59% similar aa).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Domain structure of E. multilocularis FGF receptors.

Schematic representation of the domain structure of the E. multilocularis receptors EmFR1, EmFR2, and EmFR3 according to SMART (Simple Modular Architecture Tool) analyses. As a comparison, the structure of the human FGFR1 (HsFGFR1) is shown. Displayed are the location and size of the following predicted domains: TK domain (TKD) in purple and IG-domain (Ig) in orange. Predicted signal peptides are depicted in red and transmembrane domains are shown as blue bars. The numbers of amino acids in full length receptors are shown to the right.

https://doi.org/10.1371/journal.pntd.0006959.g001

A second gene encoding a TK with significant homology to known FGFRs was identified on the available E. multilocularis genome sequence [14] under the annotation EmuJ_000196200. The amino acid sequence of the predicted protein only contained an intracellular TK domain, a transmembrane region, and one extracellular Ig-like domain, but no putative signal peptide. We therefore carried out 5’-RACE analyses on a cDNA preparation deriving from protoscolex RNA and identified the remaining 5’ portion of the cDNA, which contained one additional Ig-like domain and a predicted signal peptide. In the genome annotation, these remaining parts were wrongly annotated as a separate gene under the designation EmuJ_000196300. Hence, the second FGFR encoding gene of E. multilocularis, emfr2, encoding the protein EmFR2, actually comprises the gene models EmuJ_000196300 and EmuJ_000196200 of the genome sequence. EmFR2 thus comprises a signal peptide, two extracellular Ig-like domains, a transmembrane region, and an intracellular TKD (Figs 1 and S1). The TKD contained all residues critical for TK activity (S2 Fig) and, in SWISS-PROT BLASTP analyses, showed highest similarity to two FGF receptor kinases of the flatworm Dugesia japonica (45% identical, 65% similar residues), to the S. mansoni receptor FGFRB (55%, 68%) and to human FGFR3 (48%, 62%).

The third FGF receptor encoding gene of E. multilocularis, emfr3, was identified under the designation EmuJ_000893600 and was originally listed as an ortholog of the tyrosine protein kinase Fes:Fps [14]. However, unlike the Fes:Fps kinase which contains FCH and SH2 domains, the EmuJ_000893600 gene product, EmFR3, comprised an N-terminal signal peptide, two extracellular Ig-like domains, a transmembrane region, and an intracellular TKD (Figs 1 and S1), in which 22 of 30 highly conserved residues of TK are present at the corresponding position (S2 Fig). Furthermore in SWISS-PROT BLASTP analyses the EmFR3 TKD displayed highest similarity to several vertebrate FGF receptors and to human FGFR2 (32%, 47%). We thus concluded that EmuJ_000893600 actually encoded a third Echinococcus FGF receptor TK.

Apart from emfr1, emfr2, and emfr3, no further genes were identified in the E. multilocularis genome which displayed clear homology to known FGFR TKDs and which contained characteristic IG domains in the extracellular portions. In vertebrates, structural homology has been described between the TKDs of the receptor families of FGF receptors, the vascular endothelial growth factor (VEGF) receptors, and the platelet-derived growth factor (PDGF) receptors, which also contain varying number of Ig domains in the extracellular parts [37]. Furthermore, VEGF receptor-like molecules have also been described in invertebrates such as Hydra [37]. We therefore carried out additional BLASTP searches on the E. multilocularis genome using human VEGF- and PDGF-receptors as queries, but only obtained significant hits with the above mentioned TKDs of EmFR1, EmFR2, and EmFR3. These data indicated that members of the VEGF- and PDGF-receptor families are absent in Echinococcus, as has already been described for the closely related schistosomes [24].

Genes encoding canonical FGF ligands have so far neither been identified in genome projects of free-living flatworms [38], nor in those of trematodes [39] or cestodes [14]. In vertebrates [15] as well as several invertebrate phyla [40–42], however, canonical FGF ligands are clearly expressed. To investigate the situation more closely, we carried out BLASTP and TBLASTN analyses against the predicted proteins and E. multilocularis contig information, respectively, using FGF ligand sequences of insect, nematode, and cnidarian origin as queries. The product of only one E. multilocularis gene, EmuJ_000840500 (annotated as ‘conserved hypothetical protein’) showed certain similarity to these FGF ligands and according to SMART protein domain analyses could contain a FGF-ligand domain between amino acids 166 and 258, although this prediction was clearly below the prediction threshold and of low probability (E-value: 817). No export directing signal peptide was predicted for the EmuJ_000840500 protein, as would be typical for FGF ligands. Furthermore, although EmuJ_000840500 had clear orthologs in the cestodes Taenia solium (TsM_000953800) and Hymenolepis microstoma (HmN_000558500), none of these gene models had any prediction of an FGF-ligand domain in SMART analyses (nor predicted signal peptides). We thus considered it highly unlikely that EmuJ_000840500 encodes a so far not identified flatworm FGF-ligand.

Taken together, our analyses indicated that E. multilocularis contains genomic information for three members of the FGFR family of RTK with either one (EmFR1) or two (EmFR2, EmFR3) extracellular Ig-like domains, of which one, EmFR3, showed higher divergence within the TKD as it contained only 22 of otherwise 30 highly conserved amino acid residues of TK. On the other hand, no members of the VEGF- and PDGF-receptor families are encoded by the E. multilocularis genome, nor does it contain genes coding for canonical FGF-ligands.

Expression of FGF receptors in E. multilocularis larval stages

To investigate gene expression patterns of the Echinococcus FGF receptors in parasite larvae, we first inspected Illumina transcriptome data for parasite primary cells after 2 and 11 days of culture (PC2d, PC11d, respectively), metacestode vesicles without or with brood capsules (MC-, MC+, respectively), as well as protoscoleces before or after activation by low pH/pepsin treatment (PS-, PS+, respectively), that had be produced during the E. multilocularis genome project [14]. According to these data, emfr1 was moderately expressed in primary cells, metacestode vesicles, and protoscoleces (S3 Fig). Likewise, emfr2 was expressed in all stages, but very lowly in primary cells, somewhat more in metacestode vesicles, and highest in protoscoleces. emfr3, on the other hand, was low to moderately expressed in primary cells, low in metacestode vesicles, and highest in protoscoleces (S3 Fig). Since primary cell preparations are characterized by a much higher content of germinative (stem) cells than metacestode vesicles [6], these expression patterns could indicate that emfr3 is stem cell-specifically expressed. We therefore carried out quantitative RT-PCR experiments on cDNA preparations from in vitro-cultivated metacestode vesicles (MV) versus metacestode vesicles after treatment with hudroxyurea (MV-HU) or the Polo-like kinase inhibitor BI 2536 (MV-BI), in which the stem cell population had been selectively depleted [6,40]. While emfr1 expression levels were equal in MV when compared to MV-HU and MV-BI, both emfr2 and emfr3 transcripts were significantly reduced in MV-HU and MV-BI (S4 Fig). This indicated that emfr1 does not have a typical stem cell-specific expression pattern whereas emfr2 and emfr3 could be expressed in the parasite’s stem cells or their immediate progeny (since HU- and BI 2536-treatment has to be carried out for at least one week [6,34]).

To clarify the situation we carried out whole-mount in situ hybridization experiments on metacestode vesicles according to recently established protocols [6,43,44]. In these experiments, proliferating parasite stem cells were labeled by incorporation of the nucleotide analog EdU, which was combined with detection of gene transcripts by using fluorescently labeled probes. According to these experiments, emfr1 was expressed at low levels throughout the germinal layer and the signal was somewhat higher during the development of protoscoleces (Fig 2), maybe due to higher cell density in the protoscolex. The intensity of the signal was heterogenous, but no clear pattern could be discerned and it was too low to clearly determine the percentage of positive cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. WMISH analysis of emfr1 expression.

Left panel, antisense probe detecting emfr1 expression. Right panel, sense probe (negative control). WMISH signal is shown in green, nuclear DAPI staining in blue. gl, germinal layer; ps, developing protoscolex. Bar: 20 μm.

https://doi.org/10.1371/journal.pntd.0006959.g002

emfr2, on the other hand, was specifically expressed in a population of small sized cells, which comprised 1.7% to 6.3% of all cells in the germinal layer (Fig 3A). None of these emfr2⁺ cells were also EdU⁺, indicating that they were post-mitotic [6]. During initial protoscolex development, emfr2⁺ cells accumulated in the peripheral-most layer of cells, as well as in the anterior-most region (which will form the rostellum), and in three longitudinal bands of cells in the interior of the protoscolex buds (Fig 3B). Again, practically none of the emfr2 labelled cells was EdU⁺ (less than 1% of emfr2⁺ cells were EdU⁺; Fig 3C). Later during development, some emfr2⁺ cells were found in the protoscolex body, but most accumulated in the developing rostellum and the suckers (Fig 3D). Importantly, emfr2 expression was restricted to the sucker cup, where cells are differentiating into em-tpm-hmw⁺ muscle cells [6], and not the sucker base, where EdU⁺ germinative cells accumulate [6] (Fig 3D). Taken together, these data indicate that emfr2 is expressed in post-mitotic cells, of which many are likely to be differentiating or differentiated muscle cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. WMISH analysis of emfr2 expression.

Shown are WMISH analyses of emfr2 expression (A, B) and co-detection of emfr2 WMISH and EdU incorporation (C, D, E) during metacestode development. In C, D, and E, metacestodes were cultured in vitro with 50 μM EdU for 5 h and were then processed for WMISH and EdU detection. In all panels the WMISH signal is shown in green, EdU detection in red, and DAPI nuclear staining in blue. A, detail of the germinal layer. B, general view of a region of a metacestode with protoscoleces in different stages of development. C, early protoscolex development. D, late protoscolex development. E, Detail of a sucker, maximum intensity projection of a confocal Z-stack. r, rostellum; s, sucker. Bars are 20 μm for A, 50 μm for B, 10 μm for C, and 50 μm for D. Sense probe (negative control) was negative in all samples.

https://doi.org/10.1371/journal.pntd.0006959.g003

Finally, emfr3 was expressed in very few cells in the germinal layer (less than 1% of all cells, although the number is difficult to estimate since they were absent in most random microscopy fields) (Fig 4A). emfr3⁺ cells accumulated in small numbers during brood capsule and protoscolex development (Fig 4B). The emfr3⁺ cells had a large nucleus and nucleolus and, thus, had the typical morphology of germinative cells (S5 Fig). At the final stages of protoscolex development, few emfr3⁺ cells were present in the body region and some signals were also present in the rostellar region (S5 Fig). In the developing protoscolex and in the germinal layer, emfr3⁺ EdU⁺ double positive cells were found (Fig 4). These data indicated that emfr3 is expressed in a very small number of proliferating cells with the typical morphology of germinative cells.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Co-detection of emfr3 WMISH and EdU incorporation during metacestode development.

Metacestodes were cultured in vitro with 50 μM EdU for 5 h and were then processed for WMISH and EdU detection. In panels displaying multiple channels, the WMISH signal is shown in green, EdU detection in red, and nuclear DAPI staining in blue. A, germinal layer. B, developing protoscolex. Arrowheads indicate Edu⁺ emfr3⁺ double-positive cells. Bars are 25 μm for A, 50 μm for B (main panel) and 10 μm for B (enlarged inset).

https://doi.org/10.1371/journal.pntd.0006959.g004

In summary, the three E. multilocularis FGF receptor genes showed very different expression patterns in metacestode and protoscolex larval stages. While emfr1 was lowly expressed in cells that are dispersed throughout the germinal layer, emfr2 displayed an expression pattern indicative of differentiating/differentiated muscle cells. emfr3, on the other hand, appeared to be expressed in a minor sub-population of germinative cells.

Host FGF stimulates E. multilocularis larval proliferation and development

Since E. multilocularis larvae do not express canonical FGF ligands (see above), but usually develop in an environment in which FGF1 and FGF2 are abundant [17], we next tested whether host-derived FGF ligands can stimulate parasite development in vitro. To this end, we employed two different cultivation systems which we had previously established [7–10]. In the axenic metacestode vesicle cultivation system, mature metacestode vesicles of the parasite are cultivated in the absence of host cells under reducing culture conditions (e.g. low oxygen)[7,9]. In the primary cell cultivation system [8,10], axenically cultivated metacestode vesicles are digested to set up cell cultures which are highly enriched in parasite germinative (stem) cells (~ 80% [6]), but which also contain certain amounts of differentiated cells such as muscle or nerve cells. In the primary cell cultivation system, mature metacestode vesicles are typically formed within 2–3 weeks, which is critically dependent on proliferation and differentiation of the germinative cell population [6] in a manner highly reminiscent of the oncosphere-metacestode transition [3].

As shown in Fig 5A, the exogenous addition of 10 nM FGF1 to mature metacestode vesicles already resulted in an elevated incorporation of BrdU, indicating enhanced proliferation of parasite stem cells, which was even more pronounced in the presence of 100 nM FGF1. In the case of FGF2, addition of 100 nM resulted in enhanced BrdU incorporation in a statistically significant manner (Fig 5A). Likewise, metacestode vesicles cultured for 4 weeks in the presence of 10 nM FGF1 or FGF2 displayed a considerably larger volume (about two-fold) when compared to non-FGF-stimulated vesicles (Fig 5B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Effects of host FGFs on E. multilocularis proliferation and development.

A, effects of FGFs on metacestode proliferation. Axenically cultivated metacestode vesicles (eight per well) were incubated for 48 h in DMEM medium (without FCS), then BrdU (1 mM) and FGFs (10 or 100 nM) were added and incubation proceeded for another 48 h before cell lysis, DNA isolation, and BrdU detection. Bars represent percentage of BrdU incorporation with the cMEM control set to 100%. Statistical evaluation of four independent experiments (n = 4) which were conducted in duplicates is shown. Student’s t-test (two tailed): *p<0.05. B, effects of host FGFs on metacestode vesicle development. Single axenically generated metacestode vesicles were in vitro cultivated for four weeks in the presence of 10 nM FGF1 (aFGF) or 5 nM FGF2 (bFGF) with daily medium changes. Control vesicles were kept in cMEM. Growth (in ml) of vesicles was monitored. In each of three independent experiments (n = 3), four vesicles were examined for every single condition. C, effects of FGFs on E. multilocularis primary cell proliferation. Freshly isolated E. multilocularis primary cells were incubated for 48 h in 2 ml cMEM, then BrdU (1 mM) and FGFs (100 nM) were added and cells were incubated for another 48 h before DNA isolation and BrdU detection. Statistical analysis was performed as in A. D, effects of host FGFs on metacestode vesicle development. Freshly prepared E. multilocularis primary cells were cultured for 21 days in cMEM medium in the presence or absence of 100 nM FGF1 (aFGF) or FGF2 (bFGF). Half of the medium volume was renewed every second day. The number of newly formed metacestode vesicles at day 21 was analysed. The bars represent the percentage of formed vesicles with the cMEM control set to 100%. The statistical evaluation of three independent experiments (n = 3) which were conducted in duplicates is shown. Student’s t-test (two-tailed): *p<0.05.

https://doi.org/10.1371/journal.pntd.0006959.g005

In the primary cell cultivation system, 100 nM concentrations of host ligands had to be added to observe statistically significant effects. Again, the incorporation of BrdU by primary cell cultures was stimulated in the presence of host-derived FGF ligands (Fig 5C), as was the formation of mature metacestode vesicles from primary cell cultures (Fig 5D).

Taken together, these results indicated that host-derived FGF ligands, and in particular FGF1, can stimulate cell proliferation and development of E. multilocularis primary cell cultures and mature metacestode vesicles.

The E. multilocularis FGF receptors are activated by host-derived FGF ligands

Having shown that host-derived FGF ligands can stimulate parasite proliferation and development in vitro, we were interested whether these effects might be mediated by one or all three of the parasite’s FGF receptors which are expressed by the metacestode larval stage. To this end, we first made use of the Xenopus oocyte expression system in which the activity of heterologously expressed protein kinases can be measured by germinal vesicle breakdown (GVBD). This system has previously been used to measure the activities of the TKD of schistosome FGF receptors [24], as well as the host-EGF (epidermal growth factor) dependent activation of a schistosome member of the EGF receptor family [31]. We, thus, expressed Pleurodeles FGFR1 (as a positive control), which is highly similar to human FGFR1 [32], and all three parasite FGF receptors in Xenopus oocytes, which were then stimulated by the addition of 10 nM FGF1 or FGF2. As negative controls, we also expressed kinase-dead versions of all Echinococcus FGF receptors in Xenopus oocytes.

As can be deduced from Table 1, control (non-stimulated) oocytes were negative for GVBD, whereas progesterone-stimulated oocytes displayed 100% vesicle breakdown. Expression of FGFR1 did not yield GVBD but, after stimulation with 10 nM FGF1, 100% of oocytes underwent GVBD, indicating stimulation of the Pleurodeles receptor by FGF1 (as expected). Upon expression of any of the parasite receptors in Xenopus oocytes, no GVBD was observed when no ligand was added. The addition of 10 nM FGF1 to these receptors, however, resulted in 100% GVBD for EmFR1 and EmFR2, as well as to 80% GVBD in the case of EmFR3. In the case of human FGF2 (10 nM), 90% GVBD was observed for EmFR1 and EmFR3, and 85% for EmFR2. No GVBD was observed upon addition of 10 nM FGF1 or FGF2 when the kinase-dead versions of the parasite receptors were expressed (S6 Fig). These data clearly indicated that all three parasite receptors were responsive to host derived FGF ligands (albeit to somewhat different extent) and that the kinase activity of the parasite receptors was essential to transmit the signal. We also measured the phosphorylation state of the parasite FGF receptors upon addition of exogenous FGF1 and FGF2 (10 nM each) to Xenopus oocytes and obtained significantly induced levels of tyrosine phosphorylation in all three cases (S6 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Effects of host FGFs and BIBF 1120 on E. multilocularis FGF receptors.

https://doi.org/10.1371/journal.pntd.0006959.t001

Taken together, these data indicated that all three parasite FGF receptors were activated by binding of host-derived FGF1 and FGF2, which was followed by auto-phosphorylation of the intracellular kinase domain and downstream transmission of the signal to the Xenopus oocyte signaling systems which induce GVBD.

Inhibition of the E. multilocularis FGF receptors by BIBF 1120

The small molecule compound BIBF 1120 (also known as Nintedanib or Vargatef) is a well-studied and highly selective, ATP-competitive inhibitor of mammalian members of the FGF-, VEGF-, and PDGF-receptor families with very limited affinity to other RTK [45,46]. As a possible agent to selectively inhibit FGF RTK activities in the parasite, we measured the effects of BIBF 1120 on EmFR1, EmFR2, and EmFR3 upon expression in the Xenopus oocyte system. As can be deduced from Table 1, a concentration of 1 μM of exogenously added BIBF 1120 already diminished the activity (after stimulation with 10 nM FGF1) of FGFR1 to 40%, and led to a complete block of kinase activity upon addition of 10 μM BIBF 1120. In the case of EmFR1 and EmFR2, 1 μM BIBF1120 also led to a marked decrease of receptor kinase activity, although to a somewhat lower extent than in the case of FGFR1. In the presence of 10 μM BIBF 1120, on the other hand, the activities of EmFR1 and EmFR2 were completely blocked. In the case of EmFR3, exogenous addition of 1 μM BIBF 1120 had only slight effects on GVBD, whereas 10 μM BIBF 1120 reduced the activity to less than 50% and 20 μM BIBF 1120 completely blocked TK-dependent GVBD (Table 1). Upon addition of 20 μM BIBF 1120, TK activity of all receptors tested was completely inhibited.

Taken together, these data indicated that all three Echinococcus FGF RTK were affected by BIBF 1120, although in all three cases higher concentrations of the inhibitor were necessary to completely block TK activity when compared to FGFR1. BIBF 1120 treatment had the lowest effects on the activity of EmFR3.

BIBF 1120 inhibits E. multilocularis larval development

We next tested the effects of different concentrations of BIBF 1120 on parasite development and survival in the primary cell and metacestode vesicle culture systems. As shown in Fig 6A, the addition of 1–10 μM BIBF 1120 had clear concentration-dependent effects on mature metacestode vesicle survival which, after cultivation for 18 days, led to about 20% surviving vesicles in the presence of 1 μM BIBF 1120, 10% surviving vesicles in the presence of 5 μM BIBF 1120, and no survival when 10 μM BIBF 1120 was applied. To test whether the metacestode vesicles were indeed no longer capable of parasite tissue regeneration, we set up primary cell cultures from microscopically intact vesicles which had been treated with 5 μM BIBF 1120 for 5 days (90% intact vesicles) and let the cultures recover in medium without inhibitor. In these cultures, however, we never observed the formation of mature vesicles, indicating that either the parasite’s stem cell population, and/or other cell types necessary for parasite development in the primary cell culture system, were severely damaged after BIBF 1120 treatment.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Effects of BIBF 1120 on E. multilcoularis metacestode vesicles and primary cells.

A, axenically cultivated metacestode vesicles (8 per well in 2 ml volume) were incubated for 18 days in the presence of 0.1% DMSO or BIBF 1120 in different concentrations (as indicated) and the structural integrity of the vesicles was monitored. Structurally intact vesicles (B) and damaged vesicles (C) are shown to the right. The experiment was repeated three times in duplicates. D, freshly isolated E. multilocularis primary cells were cultured for 21 days in cMEM medium in the presence of DMSO (0.1%) or BIBF 1120 at different concentrations (as indicated). After 21 days, newly formed metacestode vesicles were counted. The DMSO control for each of the three independent experiments was set to 100%. Microscopic images (25x magnification) of cultures after 2 weeks with DMSO (E) or 10 μM BIBF 1120 (F) are shown to the right.

https://doi.org/10.1371/journal.pntd.0006959.g006

We then also tested the effects of BIBF 1120 on fresh primary cell cultures from previously untreated metacestode vesicles. As shown in Fig 6B, a concentration of 1 μM BIBF 1120 had no effect on the formation of metacestode vesicles from these cultures, whereas vesicle formation was completely blocked in the presence of 5 μM or 10 μM BIBF 1120.

Altogether, these results clearly indicated detrimental effects of BIBF 1120 on parasite development already at concentrations as low as 5 μM. Since the parasite does not express known alternative targets for BIBF 1120, such as VEGF- or PDGFR-receptors, we deduced that these effects are due to the inhibition of one or more of the parasite’s FGF RTK.

Host FGF ligands stimulate Erk signaling in E. multiocularis

One of the major downstream targets of FGF signaling in other organisms is the Erk-like MAPK cascade, a complete module of which we had previously identified in E. multilocularis [33,47,48]. In particular, we had previously shown that the phosphorylation status of the parasite’s Erk-like MAP kinase, EmMPK1, can be measured by using antibodies against the phosphorylated form of the human Erk-kinase [33]. To investigate whether exogenously added host FGF can affect the E. multilocularis Erk-like MAPK cascade module, we first incubated mature metacestode vesicles for 4 days in serum-free medium (which has no effect on vesicle integrity [29]) and then stimulated these vesicles for 30 sec, 60 sec, and 60 min with 10 nM FGF1 and FGF2. As shown in Fig 7A, FGF1 treatment had a clear effect of EmMPK1 phosphorylation already after 30 sec of exposure. In the case of FGF2, the effect was still measurable, but clearly less pronounced than in the case of FGF1 (Fig 7A). We then also measured the effects of BIBF 1120 treatment on EmMPK1 phosphorylation. To this end, metacestode vesicles were incubated in hepatocyte-conditioned medium and were then subjected to BIBF 1120 treatment (5 μM, 10 μM) for 30 min. As shown in Fig 7B, this led to diminished phosphorylation of EmMPK1 when 10 μM BIBF 1120 was used.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Effects of host FGFs and BIBF 1120 on EmMPK1 phosphorylation in metacestode vesicles.

A, axenically cultivated metacestode vesicles were incubated in cMEM (control) or in medium without FCS (0 s), upon which FGF1 (aFGF) or FGF2 (bFGF) were added at a concentration of 10 nM for 30 sec (30s), 60 sec (60s) or 60 min (60min). Protein lysates were subsequently separated by 12% SDS-PAGE and Western blots were analysed using polyclonal antibodies against Erk-like MAP kinases (anti-ERK) or double phosphorylated Erk-like MAP kinases (anti-p-ERK). B, axenically cultivated metacestode vesicles were incubated with DMSO (negative control), 5 mM or 10 mM BIBF1120 (30 min each) and cell lysates were subsequently analysed as described above. Both experiments were performed in triplicate with similar results.

https://doi.org/10.1371/journal.pntd.0006959.g007

Taken together, these data indicated that, like in mammals and other invertebrates, the E. multilocularis Erk-like MAPK cascade can be activated through FGF signaling, initiated by exogenously added, host-derived FGF ligands.

Conclusions

In the present work we provide clear evidence that human FGF ligands are capable of activating evolutionarily conserved TK of the FGF receptor family that are expressed by the larval stage of E. multilocularis and that the parasite’s Erk-like MAPK cascade is stimulated upon exogenous addition of human FGFs to metacestode vesicles. We also showed that human FGF1 and FGF2 are stimulating the development of metacestode vesicles from parasite primary cell cultures and that they accelerate metacestode vesicle proliferation and growth in vitro. Since FGF1 and FGF2 are expressed in considerable amounts within the host liver, the primary target organ for the establishment of the E. multilocularis metacestode, and since FGF ligands are also constantly produced during liver regeneration and fibrosis, which are consequences of parasite growth within the intermediate host, we consider the observed in vitro effects on parasite FGF signalling and metacestode development also of high relevance in vivo. Liver-specific activities of host FGF could thus support the development of metacestode vesicles from stem cells that are delivered to the liver by the oncosphere larva, and could constantly stimulate asexual proliferation of the metacestode during an infection. We finally showed that at least one compound that inhibits the activities of mammalian FGF receptors, BIBF 1120, also inhibits the parasite orthologs, leads to metacestode inactivation, and prevents parasite development of stem cell-containing primary cell cultures. This opens new ways for the development of anti-Echinococcus drugs using the parasite FGF receptors as target molecules.