All inhibitor, RNAi, and in situ hybridization experiments were performed with two isolates of E. multilocularis, H95 from a naturally infected fox (Vulpes vulpes) (Jura et al., 1996) and GH09 from a naturally infected cynomolgus monkey (Macaca fascicularis) (Tappe et al., 2007). At the time point of these experiments, GH09 was still capable of brood capsule and protoscolex formation, thus serving as the isolate used for WISH. H95 had lost the capacity to form protoscoleces, thus producing larger metacestode vesicles in culture that served for inhibitor experiments and yielding sufficient amounts of primary cells for RNA interference and primary cell culture. The isolates have been passaged in Mongolian jirds (Meriones unguiculatus) as described (Spiliotis et al., 2004; Spiliotis and Brehm, 2009). Preparation and use of conditioned media for axenic culture were carried out as previously described (Koike et al., 2022). Primary cell cultures had been set up essentially as previously described (Herz et al., 2024) and served as a cellular source for RNA interference experiments (Spiliotis et al., 2010) and for metacestode development from parasite stem cells (Herz et al., 2024).

Yeast Two-Hybrid (Y2H) Gold strain from Clontech (MATa, trp1-901, leu2-3, 112, ura3-52, his3-200, gal4Δ, gal80Δ, LYS2:: GAL1_UAS–Gal1_TATA–His3, GAL2_UAS–Gal2_TATA–Ade2, URA3:: MEL1_UAS–Mel1_TATA, AUR1-C MEL1) used in yeast two hybrid experiments was passaged once a month on conventional YPD plates and kept at 4°C. Dropout medium and plates were prepared with Difco™ Yeast Nitrogen Base without Amino Acid (Becton Dickinson) and DO supplements (Clontech) following the manufacturer’s instructions.

Afatinib, osimertinib (both from Cayman chemicals), and dacomitinib (abcr GmbH) were prepared as 10 mM stock solutions in DMSO and utilized at the indicated concentrations (100 nM – 100 µM) in inhibitor studies with DMSO alone as control. Primary cells for cell viability assay and vesicle regeneration assay were isolated from in vitro cultivated metacestode vesicles using a previously established protocol (Spiliotis et al., 2010). Primary cell unit definition and experimental procedure of cell viability and regeneration assay were described in (Koike et al., 2022). The respective experiments were performed in 3 technical replicates and 3 biological replicates. The number of newly regenerated vesicles was analyzed by Kruskal-Wallis tests followed by Dunn’s tests in Graphpad Prism 10.1.2 (Graphpad software). The procedure of the mature vesicle assay is described by Koike et al. (2022). In this experiment, 3 biological replicates were used. The percentage of the intact vesicles in the mature vesicle assay and that of cell viability were analyzed with one-way-ANOVA followed by Dunnett’s tests in Graphpad Prism 10.1.2 (Graphpad software).

RNA isolation from HEK293T cells was performed using a Trizol-based method as described (Hemer et al., 2014). From total RNA, complementary DNA (cDNA) was synthesized with oligo(dT)20 primer and Superscript IV reverse transcriptase (Invitrogen) following the manufacturer’s instruction. For constructs of HsEGF, this cDNA was used as the PCR template. For the constructs of HER1, pMK-RQ-ErbB1 was used as a PCR template. For other constructs of Echinococcus, pJG4–5 based yeast-two-hybrid library from Hubert et al. (2004) were used as PCR templates. PCR primers were purchased from Merck (Germany) and are listed in Supplementary Table S1. Purified DNA fragments were cloned into linearized plasmids and sequenced by Sanger Sequencing (Microsynth seqlab). Sequences of newly analyzed genes of this work have been deposited in the GenBank database and accession numbers are listed in Supplementary Table S1. Prior to synthesis of capped mRNA (cRNA), pSecTag2Hygro-based plasmids (EmER1, EmER2, EmER3) and pcDNA3.1-SER (Vicogne et al., 2004) were linearized by Pme I and pOBER plasmid including HER1 (Opresko and Wiley, 1990) was linearized by Not I. cRNA was synthesized from these vectors as templates using the T7 mMessage mMachine Kit (Ambion). After purification and quantification, the size of the RNA fragment and the lack of abortive transcripts were confirmed through gel electrophoresis.

Oocytes of Xenopus laevis were surgically removed from animals and prepared for the microinjection, as described (Vicogne et al., 2004). Oocyte viability was evaluated by counting GVBD+ (germinal vesicle breakdown) oocytes without microinjection, after progesterone treatment. 60ng of emer1, emer2, emer3, SER, and human HER1 cRNA were injected into oocytes. 8 or 24 hours after microinjection of cRNA, the oocytes were stimulated by 50nM recombinant human EGF or progesterone. GVBD was evaluated 15 hours after progesterone/EGF stimulation. The lysates of oocytes were purified by myc-IP and tyrosine phosphorylation of the recombinant proteins was detected by Western Blot with anti PY100 antibody. To evaluate the effect of inhibitors, oocytes with or without cRNA microinjection were treated with various concentrations of inhibitors, dacomitinib, afatinib, and osimertinib for 30 minutes before EGF/progesterone stimulation. All experiments were performed using 2–3 technical replicates. Percentages of GVBD-positive oocytes were statistically analyzed with one-way ANOVA with Dunnett’s multiple comparison tests in Graphpad Prism 10.1.2 (Graphpad software).

Fragments of the em-nrg cDNA were amplified from a pJG4-5-based yeast-two-hybrid library (Hubert et al., 2004) and cloned into pJET1.2 with CloneJET PCR Cloning Kit (Thermo Fisher Scientific). Primers used for cloning are listed in Supplementary Table S1. Digoxygenin (DIG)-labeled probes were synthesized from this pJet 1.2-based plasmid, purified, and quantified as described (Koike et al., 2022). In vitro cultivated metacestode vesicles for whole-mount in situ hybridization (WISH) were treated with 40 mM hydroxyurea (HU) for 7 days as described (Koziol et al., 2014) and recovered for 0, 3, 6, or 10 days. All vesicles were labelled with 50 μM 5-ethynyl-2’-deoxyuridine (EdU) in vitro at 37°C for 5 h prior to fixation by 4% paraformaldehyde. WISH followed by fluorescent detection of EdU with Alexa Flour 555 was carried out as previously described (Koziol et al., 2014; Herz et al., 2024). Fluorescent specimens were imaged with a Nikon Eclipse Ti2E confocal microscope and processed with Fiji/ImageJ as described (Koike et al., 2022). One way ANOVA followed by Dunnett’s multiple comparison test was used in Graphpad Prism 10.1.2 (Graphpad software) to compare the number of cells with each signal. Pulse-chase experiments were carried out essentially as recently described (Herz et al., 2024) with a 5 h EdU pulse followed by 72 h recovery prior to emer1-specific WISH and double staining.

The previously established MALAR-Y2H system (Li et al., 2016) was used for assessing ligand-receptor interactions, with plasmids pGAD_SP-WBP1_cloning_ linker_TMP_Cub_GAL4 (for ligands), and pGBKT7_SP_OST1_cloning_NubG (for receptors). cDNA fragments encoding sequences downstream of the signal peptide and upstream of the transmembrane domain of the Echinococcus ligand cDNA were amplified from pJC4–5 based plasmid library (Hubert et al., 2004), and integrated into SmaI-digested pGAD_SP-WBP1_cloning_linker_TMP_Cub_GAL4. For the human EGF construct, cDNA synthesized from HEK293T cells was used for the PCR template. Similarly, fragments between signal peptide and juxtamembrane region of the receptor cDNA were amplified and integrated into SmaI-digested pGBKT7_SP_OST1_cloning_NubG. Primers used for cloning are listed in Supplementary Table S1. The Saccharomyces cerevisiae Y2HGold strain (Clontech) was transformed by these plasmids using the protocol of Tripp et al. (2013). Transformants were initially inoculated onto Leu-/Trp- double dropout plates, and later transferred onto Leu-/Trp-/His- triple dropout plate and Leu-/Trp-/Ade-/His- quadruple dropout plates with three densities, as described (Koike et al., 2022). Growth level was quantified by Fiji/ImageJ using the protocol of Petropavlovskiy et al. (2020). The quantified level of growth on quadruple dropout plates was statistically analyzed using one-way-ANOVA followed by Tukey’s multiple comparison tests in Graphpad Prism 10.1.2 (Graphpad software).

RNA interference (RNAi) was performed using a slightly modified protocol of Spiliotis et al. (2010). siPOOLs (mixture of 30 siRNA fragments) were designed and synthesized by siTOOLs Biotech (Planegg, Germany). Other siRNAs were designed with siDirect version 2.0 (Naito et al., 2009) and purchased from Sigma-Aldrich. As negative and positive controls for siPOOLs, randomized siRNA fragments and fragments against the bcat-1 gene (Herrmann et al., 2025) were used, respectively. As negative and positive controls for siRNA (Sigma), siRNA against eGFP and bcat-1 were used, respectively. The sequences of the siRNAs are listed in Supplementary Table S1. E. multilocularis primary cells for RNAi experiments were isolated with the same procedure as those for cell viability assay and vesicle regeneration assay. 150 U of primary cells isolated from mature metacestode vesicles were dispensed into an electroporation cuvette (1 mm Gap) from BTX with 90 µl of electroporation buffer (siPort) including 200 pmol siRNA fragments and electroporated with time constant protocol at 200 V for 0.5 ms with the electroporator GenePulser Xcell (BioRad). The electroporated primary cells were incubated for 12 min at 37°C and centrifuged at room temperature. After removal of the supernatant, cells were resuspended by the A6/B4 condition medium and dispensed into a 96-well plate (150U/well). Primary cells were then cultured for 21 days, and half of the conditioned medium was exchanged twice a week. Images were taken under an optical microscope (Nikon eclipse Ts2-FL) and the number of regenerated vesicles was counted. Three technical replicates and three biological replicates were prepared for counting regenerated vesicles. The numbers of regenerated vesicles were analyzed by Kruscal-Wallis tests followed by Dunn’s tests in Graphpad Prism 10.1.2 (Graphpad software). With only three technical replicates, cell viability was measured with Cell Titer Glo (Promega) after 21 days of incubation.

Bioinformatic analysis was essentially performed as described (Koike et al., 2022). Briefly, BLASTP searches against protein databases of E. multilocularis, Schistosoma mansoni, and Schmidtea mediterranea were performed on Wormbase ParaSite (Howe et al., 2017) using the amino acid sequences of human EGF receptors and EGF ligands as queries. The KEGG database at Genomenet (https://www.genome.jp/) was used for the reciprocal BLASTP analysis. Gene predictions based on the E. multilocularis genome sequence (Tsai et al., 2013), as available in Wormbase ParaSite, were verified or corrected using the Integrated Genome Viewer (Thorvaldsdóttir et al., 2012) and previously generated transcriptome data (Tsai et al., 2013; Herz et al., 2024). For domain detection, SMART (Letunic et al., 2021) was used. Multiple sequence alignments were performed using CLUSTALW 2.1 (Thompson et al., 1994). Statistical analyses were performed using Graphpad Prism 10.1.2 throughout. Statistical comparisons were performed only where biologically meaningful reference groups were available (e.g., treated versus control conditions). For datasets describing gene expression levels across different developmental stages without a defined control condition, no statistical testing was applied, as such comparisons would not be biologically informative.

To systematically dissect EGF signalling in E. multilocularis, we first identified and molecularly characterized the parasite’s EGF receptor repertoire before assessing receptor function, pharmacological inhibition, and ligand–receptor interactions. One receptor of this family, EmER, had previously been described and shown to exhibit the typical EGFR architecture: an intracellular tyrosine kinase domain and two extracellular receptor L domains accompanied by several furin-like repeats (Spiliotis et al., 2003). This structural combination is a hallmark of the EGF and insulin receptor families (Ward et al., 2007).

Inspection of the annotated E. multilocularis genome sequence (Tsai et al., 2013) identified only five gene models encoding proteins with this composition. Two (EmuJ_000962900, EmuJ_000981300) belonged to the insulin receptor family, which we had characterized previously (Konrad et al., 2003; Hemer et al., 2014); one corresponded to EmER (EmuJ_000075800); and two others (EmuJ_000617300, EmuJ_000969600) were annotated as receptor tyrosine kinases. BLASTP analyses revealed that the latter proteins showed the highest similarity to EGF receptors from other metazoans. We therefore concluded that EmuJ_000617300 and EmuJ_000969600 represent two additional EGF receptor genes, consistent with previous analyses of EGFR evolution in metazoans (Barberan et al., 2016).

To ensure that no further genes had been overlooked, we carried out TBLASTN searches against the genome using the tyrosine kinase and receptor L domains of EmuJ_000075800, EmuJ_000617300, and EmuJ_000969600 as queries. No additional candidates were identified. Thus, in contrast to planarians, which encode six EGF receptor genes (Barberan et al., 2016), the E. multilocularis genome contains only three. Guided by the genome sequence and recent transcriptome data (Herz et al., 2024), we designed primers to clone the remaining EGF receptors in full length and deposited the cDNA sequences in GenBank (Supplementary Table S1).

Nomenclature has been inconsistent in the literature. The first cloned receptor, corresponding to EmuJ_000075800, was originally named emer (Spiliotis et al., 2003), but Barberan et al. (2016) referred to it as EGFR-2, designating EGFR-1 for EmuJ_000617300 and EGFR-3 for EmuJ_000969600. In earlier genome analyses, we named these genes emerb and emerc (Brehm and Koziol, 2017). To harmonize with established conventions for naming insulin (Hemer et al., 2014), fibroblast growth factor (Förster et al., 2019), and transforming growth factor (Kaethner et al., 2023) receptors, we now designate them emer1 (EmuJ_000075800), emer2 (EmuJ_000617300), and emer3 (EmuJ_000969600), encoding EmER1, EmER2, and EmER3, respectively. Notably, while the cDNAs for EmER1 and EmER3 were correctly predicted in genome sequencing (Tsai et al., 2013), the annotation of EmER2 was incomplete: the true cDNA encodes an additional 31 N-terminal amino acids, including a signal peptide.

As shown in Figure 1A, all three Echinococcus EGF receptors possess predicted tyrosine kinase domains in their cytoplasmic regions, as well as signal peptides, two receptor L domains, and several furin-like domains extracellularly. In EmER2, the SMART algorithm (Letunic et al., 2020) did not detect a transmembrane helix; however, Kyte–Doolittle hydropathy analysis (Kyte and Doolittle, 1982) identified a strongly hydrophobic stretch between residues 970 and 1000, which we propose constitutes the transmembrane region. Sequence alignments of the tyrosine kinase domains further confirmed that all three receptors retain the canonical residues required for enzymatic activity (Figure 1B), indicating that they function as active tyrosine kinases.

We investigated the expression profiles of emer1, emer2, and emer3 across different developmental stages of Echinococcus. To this end, we analyzed previously generated transcriptome datasets covering protoscoleces (before and after activation by low pH/pepsin treatment) and adult worms (Tsai et al., 2013), as well as metacestode vesicles (before and after hydroxyurea treatment to deplete stem cells) and parasite primary cell cultures at various developmental stages (Herz et al., 2024). As shown in Figure 2, all three genes were expressed in metacestode vesicles, with emer2 consistently exhibiting the lowest overall expression levels. A similar pattern was observed in protoscoleces, where emer2 again displayed the weakest expression. Importantly, none of the EGFR-encoding genes showed significant reductions in transcript abundance after hydroxyurea treatment of metacestode vesicles, a feature we previously identified as characteristic of stem cell-associated genes (Herz et al., 2024). This suggests that emer1, emer2, and emer3 are expressed predominantly in differentiated or differentiating cells. By contrast, expression was comparatively low in adult worms. Examination of a transcriptome dataset for oncospheres (Huang et al., 2016) revealed substantial expression only for emer1 and emer3. Overall, emer1 and emer3 are prominently expressed throughout the parasite life cycle, whereas emer2 appears to be more weakly transcribed.

To further define the expression patterns of Echinococcus EGFR genes, we performed WISH on in vitro cultivated metacestode vesicles, in combination with EdU staining (5 h pulse) to label S-phase stem cells in metacestode vesicles without brood capsules. As shown in Figure 2, all three genes produced numerous signals distributed across the germinal layer. In the case of emer1, we counted 18.1% (± 0.8%) of all cells positive in WISH with 7.1% (± 0,4) EdU+ cells and 1.7% (± 0.1) cells that stained both positive for EdU and for WISH (n = 8405). For emer2, we counted 27.5% (± 4,0) as WISH+, 7.4% (± 0.5) as EdU+ and less than 0.5% positive for both (n = 8374). For emer3 we counted 22.0% (± 3.5) WISH+, 7.8% (± 0.7) EdU+ and, again, less than 0.5% positive for both signals (n = 8395). Together, these data indicate that all three genes are predominantly expressed in post-mitotic cells that are either differentiated or differentiating and that emer1 is the only gene that, to a certain extent, is also expressed in S-phase stem cells. It is interesting to note that emer2, albeit showing the lowest expression level of all three genes in transcriptomic analyses (Herz et al., 2024), stains positive for overall more cells within the germinal layer. This indicates that emer1 and emer3 are expressed to a relatively high level within the WISH+ cells.

We then also carried out WISH in vesicles that produced brood capsules and protoscoleces. As depicted in Figure 2, we obtained strong signals for emer1 in young and developing brood capsules (Figures 2C–E), whereas no signal was detected in the non-activated protoscolex within vesicles (Figure 2E). Likewise, we found emer2 expressed in young and developing brood capsules (Figures 2H, I) but no signal in the fully developed, dormant protoscolex (Figure 2J). For emer3, we observed a strong signal at the base of developing brood capsules (Figure 2M) and, later, a striking expression pattern in lateral cells of the protoscolex, (Figures 2N,O) suggesting a role in protoscolex pattern formation.

In summary, our analyses demonstrate that all three Echinococcus EGFR genes are highly expressed in the metacestode stage, predominantly in post-mitotic cells, and are likely to play roles in protoscolex development.

We previously established a method for silencing Echinococcus genes in primary cell cultures using RNA interference (RNAi) (Spiliotis et al., 2010). In this system, primary cell cultures that are enriched in parasite stem cells (Koziol et al., 2014) are set up from in vitro cultivated metacestode vesicles and typically yield mature metacestode vesicles due to stem cell proliferation and differentiation after about 2–3 weeks (Koziol et al., 2014; Brehm and Koziol, 2017; Herz et al., 2024). The primary cell culture system is currently also the only system for functional, RNAi-based studies on metacestode-expressed genes in Echinococcus (Spiliotis et al., 2010; Herrmann et al., 2025). Since all three EGFR encoding genes are well expressed in parasite primary cells (Figure 2), we examined their roles in metacestode development by combining our RNAi protocol with siPOOL technology (Hannus et al., 2014) to minimize off-target effects. Using this strategy, transcript levels were reduced by approximately 60% for each gene three days after RNAi application (Supplementary Figure S1). We then monitored vesicle formation over 21 days. Notably, emer1(RNAi) cultures failed to generate mature metacestode vesicles, whereas emer2(RNAi) had no detectable effect and emer3(RNAi) even enhanced vesicle formation (Figure 3A). To investigate the cellular basis of the emer1(RNAi) phenotype, we assessed primary cell viability (Koike et al., 2022) following knockdown. These experiments revealed significantly reduced viability in emer1(RNAi) cultures, while emer2(RNAi) and emer3(RNAi) cultures remained unaffected (Figure 3B). Collectively, these findings demonstrate that emer1 is indispensable for metacestode vesicle formation and that its loss compromises parasite cell viability.

Cheng et al. (2020) reported that treatment with afatinib (BIBW2992) at concentrations around 5 µM disrupts the structural integrity of metacestode vesicles in vitro, induces apoptosis in germinative cells, and reduces parasite burden in infected mice. Thus, afatinib is currently the most promising EGFR inhibitor with potential for repurposing against AE. However, the precise Echinococcus EGFR targeted by afatinib has remained unknown. To address this, and to further evaluate the effects of afatinib, we conducted comparative analyses including several related EGFR inhibitors.

We first tested the effects of inhibitors on isolated parasite primary cells for cell viability and metacestode vesicle regeneration capacity as well as on mature metacestode vesicles, according to a recently conducted study on anti-parasitic activities of PIM kinase inhibitors (Koike et al., 2022). As shown in Supplementary Figure S2, multiple EGFR inhibitors showed strong effects against primary cells in the screening. For further analysis, we therefore chose afatinib and dacomitinib (second-generation EGFR receptors) together with osimertinib as a representative of the third generation.

We then tested the compounds on isolated parasite primary cells, assessing both cell viability and vesicle regeneration capacity, as well as on mature vesicles, following protocols previously applied to PIM kinase inhibitors (Koike et al., 2022). As shown in Figure 4, afatinib, dacomitinib, and osimertinib (1–10 µM) all reduced the viability of primary cells, with afatinib exerting the strongest effects. At 10–30 µM, afatinib and dacomitinib completely blocked vesicle regeneration, while osimertinib achieved this only at 30 µM. Interestingly, dacomitinib at 3 µM even stimulated vesicle regeneration (Figure 4B). All three inhibitors also compromised vesicle integrity in vitro, again with afatinib being the most potent (Figure 4C). These results corroborate the observations of Cheng et al. (2020) and additionally identify dacomitinib and osimertinib as EGFR inhibitors with anti-Echinococcus activity, although less effective than afatinib.

To identify the molecular target of afatinib, we employed the Xenopus oocyte system, previously used to assess flatworm EGFR activity (Vicogne et al., 2004). As ligand stimulation is required in this assay, we used human EGF (HsEGF) based on prior work (Cheng et al., 2017) and included the S. mansoni receptor SER as a positive control. As shown in Figure 5A, EmER1 expression combined with HsEGF stimulation induced 30% GVBD, comparable to SER (50%), confirming that EmER1 is an active kinase responsive to HsEGF. By contrast, EmER2 and EmER3 elicited little to no GVBD, suggesting they are either inactive kinases or do not interact with HsEGF. Expression and phosphorylation assays in oocytes further supported this conclusion: all receptors were expressed, but phosphorylation was only observed for EmER1 upon HsEGF stimulation (Supplementary Figure S3).

We next tested afatinib and osimertinib on EmER1 activity in Xenopus oocytes. Importantly, progesterone-induced GVBD remained intact in the presence of both inhibitors, excluding general toxicity (Figure 5C). Dacomitinib was not pursued due to non-specific GVBD induction (Figure 5B). Both afatinib and osimertinib inhibited EmER1 in a dose-dependent manner, with ~50% inhibition already at 1 µM (Figure 5D). Further titration from 100nM to 100 μM revealed that afatinib was highly effective even at 100nM against both HER1 and EmER1, whereas osimertinib consistently showed weaker effects against HER1 than against EmER1 (Figures 5E, F).

Taken together, our data confirm afatinib’s profound activity against Echinococcus cells and vesicles, extend its efficacy to germinative cell-enriched cultures (~80%; Koziol et al., 2014), and identify EmER1 as its principal molecular target. These findings strongly suggest that the anti-parasitic effects of afatinib observed in vitro and in vivo (Cheng et al., 2020) are primarily mediated through inhibition of EmER1.

Having identified EmER1 as a key regulator of metacestode development and the principal target of EGFR inhibitors, we next asked whether E. multilocularis encodes endogenous EGF-like ligands capable of activating this receptor. The defining feature of EGF-like ligands is the presence of an EGF domain with six characteristic cysteine residues forming disulfide bridges. However, such domains also occur in the extracellular regions of many unrelated transmembrane proteins (Wouters et al., 2005). In a previous bioinformatic survey, Barberan et al. (2016) applied stringent criteria for identifying bona fide EGF-like ligands in metazoans and specifically, the presence of a single EGF domain in combination with a transmembrane domain and, in the case of neuregulin subtypes, an immunoglobulin (IG) domain. Based on these criteria, they identified two potential Echinococcus EGF-like ligands. One, represented by gene model EmuJ_000753300 (EmEGF), was categorized as an EGF-type ligand, while the second (EmuJ_000090400; EmNRG) showed structural characteristics of the neuregulin family, which are recognized as cognate ligands of EGFRs (Marchionni, 2014). Guided by the E. multilocularis genome sequence (Tsai et al., 2013) and recently generated transcriptome data (Herz et al., 2024), we designed primers to clone the full-length cDNAs of both genes, which we designated em-egf (EmuJ_000753300) and em-nrg (EmuJ_000090400), encoding the proteins EmEGF and EmNRG, respectively.

As illustrated in Figure 6A, both proteins possess N-terminal signal peptides and a single EGF domain. EmNRG also contains an immunoglobulin C-2 domain, a hallmark of the neuregulin subfamily of EGF-like ligands (Marchionni, 2014). Moreover, EmNRG harbors a transmembrane domain, suggesting either direct involvement in cell-cell communication or the requirement for proteolytic cleavage of its extracellular portion to generate a soluble ligand. In both proteins, the EGF domains feature six conserved cysteine residues responsible for disulfide bond formation (Figure 6B). We next examined previously published (Tsai et al., 2013) and more recent (Herz et al., 2024) transcriptome datasets derived from E. multilocularis primary cell cultures and various developmental stages. As shown in Figure 6C, em-egf expression was predominantly detected in protoscoleces and adult worms, with only minor expression in metacestodes. In contrast, em-nrg displayed stronger expression in the metacestode stage, and its transcripts were enriched following depletion of metacestode vesicles of stem cells, indicating preferential expression in post-mitotic cells (Figure 6D).

In summary, our analyses demonstrate that both identified Echinococcus EGF-like ligands exhibit structural hallmarks characteristic of this family of cytokine ligands, and that em-nrg represents a strong candidate for regulating stem cell dynamics in the metacestode stage.

We next investigated whether the identified Echinococcus EGF-like ligands interact with the three parasite EGF receptors. For this purpose, we employed the MALAR yeast two-hybrid system, originally developed to detect extracellular interactions between peptide ligands and their cognate receptors (Li et al., 2016). Although this system was first designed for CXC-like chemokines and G-protein–coupled receptors (GPCRs), it has since been successfully adapted to study interactions between neuropeptides and GPCRs (Weth et al., 2020). Here, we applied it to examine potential interactions between EGF-like ligands and EGF receptors.

To generate ligand-bearing plasmids, we cloned cDNA regions encoding the complete extracellular domains of EmEGF, EmNRG, and human EGF, replacing the N-terminal signal peptide and transmembrane domains with vector-derived sequences. For receptor-bearing plasmids, we used the extracellular regions of EmER1, EmER2, EmER3, and the human EGFR HER1. Unfortunately, all ligand-bearing constructs for Em-EGF triggered yeast growth on quadruple dropout plates even with empty receptor vectors, rendering this system unsuitable for assessing EmEGF-receptor interactions. In contrast, ligand-bearing constructs for EmNRG showed only weak background growth in the absence of receptor baits.

Since our Xenopus oocyte expression experiments had already demonstrated that human EGF can activate EmER1 and, to a lesser extent, EmER2, we first tested these combinations. As shown in Figure 7, human EGF interacted with both parasite receptors in a manner comparable to its interaction with the cognate human receptor HER1. For EmER3, only very weak growth was detected with human EGF. Testing Em-NRG revealed robust interactions with EmER1 and EmER2, and even with human HER1 (Figure 7), but none with EmER3. Positive and negative controls included previously validated plasmids (Li et al., 2016) encoding the murine chemokine CXCL12 and its cognate receptor CXCR4, as well as CXCL12 with an empty vector. As expected, these controls showed either strong or negligible yeast growth, respectively.

Taken together, these findings indicate that Em-NRG functions as a ligand for EmER1 and EmER2 in Echinococcus.

Under steady-state conditions, planarian stem cells are thought to divide asymmetrically, producing one self-renewing stem cell and one differentiating progeny (Wagner et al., 2012). To assess em-nrg expression under comparable conditions, we performed em-nrg–specific WISH on in vitro cultivated metacestode vesicles, combined with EdU incorporation to label S-phase germinative cells. As shown in Figure 8A, only very few (<1%) cells in the germinal layer stained positive for em-nrg, and none had incorporated EdU. Occasionally, we observed foci containing ten or more em-nrg+/EdU- cells (Figure 8B), but these were rare (2–3 per vesicle) and not present in every vesicle. Early brood capsules lacked em-nrg expressing cells, whereas strong signals were detected in more advanced brood capsules (Figures 8C–F). Toward the end of protoscolex formation, when germinative cell proliferation typically ceases (Koziol et al., 2014), no em-nrg+ cells were detected (Figures 8E, F). These observations indicate that under steady-state conditions em-nrg is expressed in a small subset of post-mitotic cells in the germinal layer, consistent with prior transcriptomic analyses of stem cell depleted metacestode vesicles (Herz et al., 2024). In those analyses, em-nrg expression increased significantly following stem cell depletion (Herz et al., 2024), further suggesting that the gene is primarily expressed in post-mitotic cells.

In metacestode vesicles, a 7 day hydroxyurea treatment markedly reduces the germinative cell population, after which the remaining cells undergo several rounds of clonal expansion to reconstitute the stem cell pool (Koziol et al., 2014). Under these conditions, germinative cells are expected to divide predominantly symmetrically. We therefore examined em-nrg expression after stem cell depletion. Vesicles were treated for 7 days until only few stem cells remained, then allowed to recover for 10 days. During recovery, we quantified EdU+ germinative cells and assessed em-nrg expression by WISH. As shown in Figure 9, immediately after depletion only few EdU+ cells were present and em-nrg expression remained relatively low. After 3 days of recovery, EdU+ cells increased modestly, consistent with previous reports that most metacestode stem cells cycle within ~3 days (Herz et al., 2024; Koziol et al., 2014), whereas em-nrg+ cells rose sharply from near zero to ~2,250 cells per mm² of metacestode tissue. After 6 days of recovery, EdU+ cells increased steeply to ~500 per mm², while em-nrg+ cells declined slightly (Figure 9). At 10 days, ~400 EdU+ cells per mm² were still observed, and em-nrg+ cells decreased further (Figure 9). Collectively, these data indicate that em-nrg is particularly highly expressed in metacestode tissue during stem-cell clonal expansion. Notably, even under these conditions, em-nrg+ cells were only rarely EdU+, again pointing to predominant expression in post-mitotic cells.

We next investigated whether RNAi-mediated knockdown of em-nrg in E. multilocularis primary cell cultures affects regeneration into metacestode vesicles. To this end, we applied a previously established RNAi protocol for Echinococcus primary cells (Spiliotis et al., 2010). As shown in Supplementary Figure S4, em-nrg transcript levels were reduced to approximately 70% compared with GFP and random controls. Notably, both metacestode vesicle formation and the viability of primary cell cultures were significantly impaired following em-nrg knockdown (Figure 10), indicating that expression of this EGF-like ligand gene is essential for proper stem cell function and parasite development.

Previous analyses by Lei et al. (2016) showed that the majority (~85%) of Smed-egfr3⁺ cells in Schmidtea mediterranea are neoblasts (i.e. also positive for the stem cell marker smedwi1), consistent with the role of the encoded receptor in regulating symmetric versus asymmetric neoblast divisions. In Echinococcus, however, our own data did not reveal a significant decrease in emer1 transcript abundance in metacestode vesicles following stem cell depletion by hydroxyurea treatment (Herz et al., 2024). Moreover, as shown above, only ~9% of emer1⁺ cells co-stained with EdU after a 5 h EdU pulse, marking S-phase stem cells. To determine whether emer1 might be expressed in stem cell progeny or in late phases of the germinative cell cycle, we performed pulse-chase experiments consisting of a 5 h EdU pulse followed by emer1-specific WISH after 72 h. This timing was chosen because previous work (Herz et al., 2024; Koziol et al., 2014; Cheng et al., 2017) indicated that most germinative cells complete a division cycle within approximately three days.

As shown in Figure 11, immediately after the 5 h EdU pulse followed by WISH, we detected 6.2 ± 0.2% (n = 8244) EdU⁺ cells, 18.4 ± 0.6% emer1⁺ cells, and 1.5 ± 0.1% double-positive cells. This confirms our above results and indicates that roughly 8.2 ± 0.5% of all emer1⁺ cells correspond to S-phase stem cells. After a 72 h chase, we counted 11.9 ± 0.3% (n = 8086) EdU⁺ cells, indicating that most labeled stem cells had undergone mitosis. At this time point, 18.1 ± 1.3% of all cells were emer1⁺, with 6.1 ± 0.3% positive for both signals. These findings suggest that approximately one third (32.6 ± 2%) of all emer1⁺ cells represent stem cell progeny or germinative cells in the terminal phase of the cell cycle.