The F. hepatica kinome

Our bioinformatic screen (S1A Fig) of available F. hepatica genomes has enabled us to generate a putative F. hepatica kinome. In total, 271 PKs were identified (S1 Table and S1 File) with 254 ePKs and 17 aPKs. This constitutes ~1.6% of the predicted F. hepatica proteome, a little smaller than the 1.8-1.9% seen in H. sapiens, S. mansoni, S. haematobium and T. solium [23,28–30]. This could be due to the large size of the F. hepatica genome, relative to other pathogens [47], or the incomplete nature of the genome assembly, as the total number of PKs identified is comparable to that of S. haematobium (269, [30]) and S. mansoni (253; updated genome assemblies have removed putative PKs since Grevelding et al. [28]). All putative F. hepatica kinase genes were BLASTed (e-value threshold = 0.01) against the S. mansoni genome with the top hit recorded. These hits were compared against the S. mansoni kinome, revealing that 208/253 predicted S. mansoni kinase genes had a matching hit in our F. hepatica kinome. Members of all nine ePK families were represented, with calcium/calmodulin-dependent protein kinase (CAMK) and CMGC families found to be the largest (Fig 1A), representing around 30% of all ePKs. These proportions are similar to those of other parasitic helminths, such as T. solium, Schistosoma spp., B. malayi and H. contortus (S1B Fig; [28–31,62,63]). This contrasts with free-living species (C. elegans, D. melanogaster and humans), where tyrosine kinases (TK) are the largest family (S1B Fig; [23,48,49]), suggesting a reduced role of TKs in parasitism.

Download:

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

Fig 1. Fasciola hepatica exhibit conservation of all major kinase families.

CLANs analysis (1M iterations) of F. hepatica (circles), Schistosoma mansoni (triangles) and Homo sapiens (stars) eukaryotic protein kinases shows presence of all nine major families in F. hepatica with conservation among CMGC kinases, mitogen-activated protein kinases (STEs), receptor guanylate cyclases (RGCs), casein kinase I (CKs) and calcium/calmodulin kinases (CAMKs), but divergence among tyrosine kinase-like kinases (TKLs).

https://doi.org/10.1371/journal.ppat.1013406.g001

The largest F. hepatica ePK family (CAMK) contains both substrate restricted and multifunctional kinases with a range of homeostatic roles [50]. Due to their critical roles in Ca²⁺ signalling pathways, they have been proposed as potential anthelmintic targets. Schistosome CAMKII increases its expression/activity following treatment with the paralysis-inducing drug praziquantel, with RNAi reducing motility/viability [51,52], suggesting that CAMKII plays a role in worm motor function/motility. The CMGC family is the second largest ePK family in F. hepatica. This highly conserved family consists of members that play critical roles in cell cycle control, such as the cyclin-dependent kinases (CDKs), mitogen-activated protein kinases (MAP kinases) and CDK-like kinases (CDKLs). This high degree of conservation is unsurprising given the crucial role these kinases play in fundamental cell biology, especially in proliferating cells. This may provide opportunities for the development of flukicides that disrupt growth via CDK and MAP kinase inhibition. Indeed, inhibitors targeting these kinases have already been developed for cancer treatment [53], and repurposing/redeveloping them to target the neoblast-like cells of F. hepatica and other parasitic flatworms could lead to their use as novel anthelmintics.

We examined the structural relatedness of F. hepatica ePKs alongside S. mansoni and H. sapiens ePKs using CLANs software, finding the ‘Other’ grouping to be the most divergent (Fig 1). This was not surprising given the diversity of this group which includes the polo-like kinases, NimA-related kinases and aurora kinases among others. In contrast to this, the AGC, CMGC, receptor guanylate cyclases (RGCs, a family with a catalytically inactive kinase domain similar in structure to tyrosine kinases) and TK families are the most closely clustered, a reflection of their conserved roles across multiple species. The CAMK and casein kinase I (CK1) families clustered closely with only a few divergent members. However, clustering was more diffuse for the tyrosine-like kinases (TKLs) (Fig 1), with F. hepatica and S. mansoni bone-morphogenic protein receptors (FhHiC23_g1262 and Smp_124450) notable outliers. These bone morphogenic protein receptors may be druggable targets, especially with only a 24–35% amino acid sequence identity when compared to human bone morphogenic protein receptors. Recent publications in F. hepatica have demonstrated the potential of targeting kinases (protein kinase C β and p21-activated kinase 4) to disrupt motility/viability [27,54].

We also identified 17 putative F. hepatica aPKs (S1 Table), functional PKs with a typical catalytic kinase fold that lack sequence similarity to ePKs [55]. As in other organisms, there are considerably fewer aPKs than ePKs, though the number identified here is comparable to that found in C. elegans [56]. Several conserved F. hepatica aPKs (S1C Fig) play vital roles in cell proliferation, potentially linking them to their neoblast-like stem cells, such as mammalian target of rapamycin (mTOR), ataxia–telangiectasia mutated kinase (ATM), ataxia–telangiectasia and Rad3 related kinase (ATR) and the phosphoinositide 3-kinases [57–59], again highlighting numerous pathways that could be exploited for juvenile fluke control.

FhPLK expression linked to developing juveniles

We used published life-stage transcriptomic data [47] to examine expression profiles of putative PKs. Of the 266 kinases for which we have life stage transcriptomic data, 116 show the highest relative expression in NEJs or immature worms (migrating three-week-juveniles) (S1D Fig). Seven of these genes with enriched expression are cdks, crucial cell cycle regulators [60]. Among them is a putative cdk1 gene (FhHiC23_g16347), a critical component of M-CDK that drives the cell into mitosis, which has also been implicated in regulating the pluripotency of embryonic stem cells [61], and can drive the cell cycle in early embryonic cells lacking all other CDKs [62]. Additionally, four putative F. hepatica fibroblast growth-factor receptors (fgfrs) and two plks are most highly expressed in these developing life stages. We also utilised a published transcriptomic dataset that compared in vitro F. hepatica juveniles with faster-growing in vivo juveniles [46] and found 81% of kinases were more highly expressed in the faster-growing in vivo juveniles (S1D Fig), again supporting the hypothesis that kinases play an integral role in growth/development.

Development-associated kinases were further interrogated using the F. hepatica in vitro/in vivo transcriptomic datasets [46] alongside a dataset which identified genes associated with F. hepatica neoblast-like cells [21]. Six kinases were downregulated following neoblast-like stem cell ablation, and 21 kinases were upregulated in faster-growing worms (S2 Table). Notably, two genes were prominent in both of these datasets, a fgfr (FhHiC23_g5821) and plk1 (FhHiC23_g6132). FGFRs are tyrosine-kinase receptors linked to proliferation [63] with drug inhibition, RNAi or CRISPR-interference in S. mansoni and E. multilocularis reducing neoblast-like cell proliferation [11,12,64,65]. Three FGFR inhibitors have been approved by the FDA for treatment of cancers and idiopathic pulmonary fibrosis [66], highlighting the druggability of these kinases.

PLK1 is a core cell cycle effector, regulating centrosome assembly, mitotic entry, the spindle, cytokinesis and more besides [39]. PLK1 has been linked to cell differentiation and survival in other parasitic flatworms [40–43], while in F. hepatica its downregulation following irradiation and upregulation in faster growing worms (S2 Table) support its link with neoblast-like stem cells. Our HMM screen and additional BLAST searches identified two further putative F. hepatica PLKs (FhPLK2 [FhHiC23_g13468] and FhPLK4 [FhHiC23_g6087]; Fig 2A). These genes are distinct from each other, sharing only 25–27% identity at the amino acid level with only FhPLK1 predicted (full sequence confirmed using transcriptomic reads) to possess all signature PLK1 domains (Fig 2C), including the ATP-binding motif [67], catalytic motifs, Mg²⁺ chelating motif [68], phosphorylation site [69] and phosphopeptide binding sites [70,71]. Published life stage transcriptomic data [47] show that fhplk1 expression peaks in eggs (Fig 2B), potentially indicating significant proliferative cell activity as the embryo develops into the unhatched miracidia, like that seen in S. mansoni [72]. Following eggs, relative expression was highest in immature worms and adults, life stages with significant neoblast-like stem cell and germ cell activities respectively [46,73,74]. Moreover, transcriptomic data from Robb et al. (Fig 2B, [46]) show greater fhplk1 and fhplk4 expression in faster-growing in vivo worms, correlating with observed increased cell proliferation. Houhou et al. [44] also found higher fhplk1 expression in immature worms and adults compared to NEJs (24 h post-excystment) via PCR, though they also observed ~6-fold increase in expression in adults compared to immature worms. The difference seen here between transcriptomic and qPCR methodologies could be due to low replicate numbers in the published transcriptomic datasets [47]. As plk1 has previously been localised to S. mansoni reproductive structures [40], the increased expression in adult F. hepatica is likely associated with germ cell activity in this life stage. Together, these transcriptomic data indicate that fhplk1 is associated with cell proliferation in F. hepatica.

Download:

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

Fig 2. Fasciola hepatica polo-like kinase 1 (PLK1) possesses all functional kinase domains.

(A) Schematic of F. hepatica PLKs, with domains as predicted by Interpro, illustrating N-terminal serine/threonine kinase domains and C-terminal polo-boxes. (B) Expression of fhplk genes from published datasets; Log₂-transformed median transcripts per mapped million reads (TPMs) of F. hepatica plk genes across various life stages (egg, metacercariae, 1h NEJ, 3h NEJ, 24h NEJ, 3-week juvenile, adult; n = 1–3 for each lifestage; [47]); Log₂-transformed median transcripts per mapped million reads (TPMs) of F. hepatica plk genes in in vitro and in vivo 3-week juveniles (n = 3 for each treatment; [46]). (C) Conservation of functional motifs in FhPLK1 but not FhPLK2/4 when compared to PLK1 sequences from Echinococcus multilocularis, Schistosoma mansoni, Drosophila melanogaster, Homo sapiens and Mus musculus; * = conserved residue,: = highly conserved amino acid properties,. = weakly conserved amino acid properties (our transcriptomic reads were used to confirm sequence).

https://doi.org/10.1371/journal.ppat.1013406.g002

RNAi-mediated gene silencing of fhplk1 disrupts juvenile growth

To test if fhplk1 is integral to maintenance of the F. hepatica neoblast-like cell population we employed RNAi to probe its function in vitro. Parasites were treated with fhplk1-specific dsRNA and monitored for signs of abnormal growth. Transcript knockdown of fhplk1 was measured after four weeks and found to be reduced by ~90% relative to control treatment (Fig 3A; mean transcript expression ±SEM: control dsRNA = 127.67 ± 9.49%, fhplk1 dsRNA = 9.33 ± 2.73%). We also found that transcript expression of another proliferative cell marker (histone 2b (fhh2b)) was reduced by 81.7 ± 9%. Though not significantly different, it suggested that fhh2b expressing cells were depleted in fhplk1-RNAi worms; expression of a muscle cell marker (myosin light-chain (fhmlc)) was reduced by 21.3 ± 30% relative to untreated worms, although again this was not significant (Fig 3A). H2B is highly expressed throughout the S-phase of the cell cycle when it participates in chromatin organisation [75], with RNAi-mediated silencing in free-living and parasitic flatworms impairing cell proliferation [12,16,21,76], therefore demonstrating how plk1-RNAi in F. hepatica has led to the depletion of a proliferative cell-associated transcript. Here, we observed that fhplk1-RNAi affected fluke development in vitro, with fhplk1-RNAi juveniles significantly smaller than control worms three weeks after the first RNAi trigger (Fig 3B; mean juvenile area ±SEM at three-weeks: untreated = 97350 ± 2559 µm², control-dsRNA = 94214 ± 4266 µm², fhplk1 dsRNA = 73964 ± 2559 µm²). This phenotype was further exaggerated after a fourth week (Fig 3B; mean juvenile area ±SEM at four-weeks: untreated = 118368 ± 4732 µm², control-dsRNA = 117781 ± 5130 µm², fhplk1 dsRNA = 75942 ± 2495 µm²). We also found that fhplk1 knockdown impaired cell proliferation in four-week-old juveniles, thereby disrupting the neoblast-like stem cell population (Fig 3C and 3D; mean # 5-ethynyl-2-deoxyuridine (EdU)⁺ nuclei ±SEM: control-dsRNA = 477.5 ± 47; fhplk1-dsRNA = 9.8 ± 5.4). RNAi of plk1 in mammalian cancer cell lines in vitro also reduced proliferation [77], and reduced tumour size in vivo in mice with non-small cell lung carcinoma [78], supporting a conserved role for plk1 in proliferative cells. Interestingly, following four weeks of plk1 silencing, juveniles cultured for a week in the absence of a RNAi-trigger showed no signs of recovery and did not resume proliferation (S2 Fig); this observation is consistent with the hypothesis that prolonged fhplk1-RNAi results in the ablation of F. hepatica neoblast-like stem cells. It is possible that if culture was continued for longer in the absence of a RNAi-trigger, proliferative cells could recover if the systemic effects of RNAi do not persist. Together, these data support a conserved function for PLK1 in F. hepatica that ensures cell-cycle progression. As PLK1s are functionally conserved in S. mansoni and E. multilocularis [40,43], targeting PLK1 activity could be an effective pan-phylum control strategy.

Download:

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

Fig 3. Knockdown of fhplk1 disrupts growth and cell proliferation in juvenile Fasciola hepatica in vitro.

(A) Mean expression ±SEM of fhplk1 (white), fhh2b (green) and fhmlc (purple) in dsRNA-treated juvenile F. hepatica relative to no dsRNA-treated juveniles following four weeks of repeated exposures to fhplk1 dsRNA in vitro shows significant knockdown of fhplk1 and reduction in expression of neoblast-like cell associated gene fhh2b relative to control dsRNA treated worms (n = 3 for each treatment, unpaired t tests). (B) Mean area ±SEM of juvenile F. hepatica significantly diminished over three/four weeks of fhplk1 dsRNA exposures (n ≥ 37 for each treatment; Kruskal-Wallis and Dunn’s posthoc tests). (C) Mean # EdU⁺ nuclei ±SEM significantly decreased after four weeks of fhplk1 dsRNA treatments in juvenile F. hepatica (n ≥ 15 for each treatment; Mann-Whitney U test). (D) Confocal images of EdU staining (green) in four-week-old F. hepatica treated with fhplk1 dsRNA in vitro shows reduction in # EdU⁺ nuclei compared to control dsRNA treated worms; scale bar = 100 µm, DAPI counterstain (magenta). **, p < 0.01; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.ppat.1013406.g003

RNAi of fhplk1 downregulates cell cycle effectors and ribosome biogenesis

We previously used irradiation to identify transcripts associated with F. hepatica neoblast-like cells [21]. However, while irradiation is effective at ablating proliferating cells, the potential for damage to differentiated cells limits its utility. Transcriptomic analysis following fhplk1 silencing would help inform the role of fhplk1 role in juvenile fluke and support identification of other genes associated with growth/development. Therefore, we carried out target-specific RNAi for three weeks before RNA extraction and transcriptome sequencing (Fig 4A), confirming that fhplk1-RNAi reduces growth and ablates neoblast-like stem cells (S3A and S3B Fig). RNA-Seq was then used to identify differentially expressed genes (DESeq2, padj < 0.001; S4 Table), with 946 downregulated genes and 1244 upregulated genes (Fig 4B). TOPGO analysis showed that genes associated with protein translation, ribosome assembly and DNA replication were overrepresented among downregulated genes (Fig 4C). Many of these GO terms are also overrepresented among genes associated with the germinative cells found in E. multilocularis [79]. The downregulation of genes associated with protein synthesis (‘translation’ and ‘protein folding’; Fig 4C) was interesting as studies in mouse cells showed that protein synthesis rates in early differentiated mouse cells are two-fold of that in embryonic stem cells [80], a result also observed in other mammalian cell cultures (reviewed in [81]). As fhplk1-RNAi depletes neoblast-like cells (i.e., stem cells) the proportion of differentiated cells might be expected to increase relative to control worms, although the production of more undifferentiated cells in the control groups could explain the higher levels of protein synthesis. Further, other GO terms (‘ribosomal small subunit assembly’, ‘structural component of ribosome’; Figs 4C and S3C) suggested that the downregulation of translation may be related to reductions in ribosomal activity. KEGG pathway analysis showed the downregulation of ribosomal subunits/biogenesis following fhplk1-RNAi (S3D Fig) which matches observations in other platforms where stem cells have lower translation rates than differentiated cells, except in the case of ribosome biogenesis, which is upregulated in proliferating cells [82]. The downregulation of ribosome biogenesis-associated genes was also observed 14 days after irradiation in S. mansoni [13], demonstrating that this occurs in related species. It is also possible that the differential expression of ribosome-associated genes/pathways is indicative of general stress, as downregulation of these genes/pathways has previously been noted in nematodes following drug treatment (with non-stem cell targeting drugs) and local environment change [83,84].

Download:

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

Fig 4. fhplk1 knockdown downregulates cell cycle and translation-associated transcripts in juvenile Fasciola hepatica in vitro.

(A) Experimental timeline showing dsRNA treatments (24 h dsRNA exposure each time) carried out on juvenile F. hepatica across three weeks to generate transcriptomes with final treatment from day 21–22. (B) Volcano plot showing differentially expressed genes (red dots; DESeq2, padj < 0.001) after fhplk1 RNAi in juvenile F. hepatica where blue shading indicates downregulated genes and red shading indicates upregulated genes. (C) Top 10 overrepresented biological process GO terms among downregulated genes following fhplk1 RNAi shows genes involved in translation and nucleotide interactions. (D) Many components of KEGG cell cycle pathway downregulated following fhplk1 RNAi (red = upregulated; grey = no change; blue = downregulated; white = unassigned KEGG ID). (E) Venn diagram showing overlap between genes downregulated following fhplk1 RNAi in juvenile F. hepatica (blue), irradiation in juvenile F. hepatica (green) or irradiation in adult S. mansoni (yellow). (F) Log₂foldchange of differentially expressed kinases following fhplk1 RNAi shows downregulation of proliferation-associated kinases such as MAPKs and FGFRs (colours correspond to log₂foldchange; red = upregulated; blue = downregulated). Experimental timeline schematic created in BioRender.

https://doi.org/10.1371/journal.ppat.1013406.g004

Transcripts associated with DNA/RNA interactions and the cell cycle were also downregulated following fhplk1 knockdown (e.g., DNA-replication initiation; Figs 4C, 4D and S3). Indeed, 4/5 KEGG pathways downregulated following irradiation of F. hepatica juveniles [21] were downregulated following fhplk1 RNAi (S3D Fig). Numerous core cell cycle genes were downregulated (Fig 4D and S3 Table), including atm and atr kinases, cell division control proteins, cyclinB, members of the mini-chromosone complex (mcm2–7) and origin recognition complex subunits (orc2 & 3). Additional core cell cycle regulators not annotated by eggNOG, but identified through BLAST searches, were also downregulated, such as a p53 homologue (FhHiC23_g6525). p53 is a tumour suppressor which can arrest the cell cycle and cause apoptosis when activated [85]. Mutations in p53 are linked to cancerous growths, with compounds developed to reverse these adverse mutation effects currently in clinical trials [86]. Furthermore, a S. mansoni p53 homologue has been linked to their neoblast-like cells [18]. While many of the cell cycle effectors discussed likely have conserved structure/function in both parasites and their hosts, inhibitors that indirectly target them [87] may be worth exploring for their anthelmintic potential. One surprising observation in the cell cycle schematic was ‘upregulation’ of fhplk1 following fhplk1 RNAi (DESeq2, fhplk1 log₂Fold Change = ↑ 3.73). Our earlier trials had confirmed fhplk1 knockdown following dsRNA treatment via qPCR (Fig 3A), so this upregulation was unexpected. However, examination of the read files via the Integrative Genomics Viewer (https://igv.org/) revealed that fhplk1 dsRNA was likely sequenced as reads were observed only in the 192 bp dsRNA amplicon region of fhplk1 dsRNA samples (S3E and S3F Fig). Control dsRNA samples on the other hand had reads across the length of the gene, with exons clearly visible (S3E Fig). The data indicate that this sequencing anomaly is related to our RNAi methodology as we ensure knockdown by repeatedly soaking worms in 100 ng/µL dsRNA) [88], including the 24 h period immediately preceding worm extraction.

We aimed to identify pan-trematode transcripts clearly linked to parasitic flatworm neoblast-like cells by cross-referencing the transcriptomic data generated here against the transcriptomes of irradiated F. hepatica and S. mansoni (orthologues were identified based on the top hit in BLAST searches against the S. mansoni genome and can be found in S3 Table) [12,21]. We found 21 genes downregulated in all three transcriptomic datasets following neoblast-like cell ablation (Fig 4E; note that the relatively small number of downregulated genes found in both irradiated datasets is likely due to the different life stages investigated as discussed in [21]). Many of the genes downregulated in all three datasets are known for their links to proliferation (S4 Table), including the cell cycle regulators p53, cyclinB and the mcm complex members. A further 34 genes were downregulated following fhplk1-RNAi in F. hepatica and irradiation in S. mansoni, including hepatic leukaemia factor, mastL (greatwall kinase) and wdhd1 (WD repeat and HMG-box DNA-binding protein/and-1) which again have been associated with cell cycle regulation [89], DNA replication [90] and stem cell population maintenance [91] in Xenopus/human cells. Finally, 46 genes were downregulated in F. hepatica following either fhplk1-RNAi or irradiation with eight transcription factors/regulators included among them as well as a fgfr, implying conserved roles related to F. hepatica neoblast-like cells. While there were these shared downregulated genes between these two F. hepatica datasets there were 879 genes only downregulated following fhplk1-RNAi (Fig 4E). We hypothesise that this difference is caused by the loss of neoblast-like cells over an extended 3-week period of time (unlike the irradiated dataset in which juveniles were sequenced 48 hours after X-ray exposure), with this having knock-on effects across tissues in the worm.

As several of the genes downregulated in both datasets were kinases we explored the expression of all F. hepatica kinases following fhplk1 RNAi; 48 kinases were differentially expressed (22 downregulated and 26 upregulated; Fig 4F). Among downregulated kinases were three fgfrs that were also upregulated in developing life stages (FhHiC23_g6174, FhHiC23_g1614 and FhHiC23_g5821), an aurora kinase and members of the MAPK signalling cascade which regulates cell cycle progression [92]. Again, these kinases are integral to activate proliferation in other species and may be appealing targets for control. AGC kinases and CAMKs were more represented among upregulated kinases, though their functions are more diverse and less clearly linked to proliferation.

One final point to note is that of 46 genes downregulated after irradiation or fhplk1-RNAi in F. hepatica, ten had no clear BLAST hit in model species, though they all have S. mansoni homologues and most have F. gigantica homologues. Furthermore, of the 50 most downregulated transcripts (ranked by fold-change) only 8 had clear BLAST hits in model species (e-value <0.001, S3 Table), though most were found to have orthologues in at least one other parasitic flatworm. While we know little of the function of these genes/proteins, their downregulation links them to putative roles in F. hepatica growth/development associated with neoblast-like stem cells, especially where also downregulated following irradiation. These unknown genes could eventually become attractive drug targets, given their lack of homology to genes in host species.

Inter-cell signalling upregulated following fhplk1 RNAi

Over 1200 juvenile fluke genes were upregulated following three weeks of fhplk1-RNAi (Fig 4B and S3 Table). TOPGO analysis of these genes showed that GO terms associated with inter-cell signalling systems such as ion transport, cell-adhesion and synapses were overrepresented (Figs 5A and S4A) among the upregulated genes. Furthermore, cellular component GO terms suggested many predicted proteins are associated with the cell membrane and extracellular space/synapse (S4A Fig), supporting evidence for the upregulation of inter-cell signalling. KEGG pathway analysis further reinforced this, highlighting upregulation of multiple neuronal-ligand receptors, including many GPCRs (Fig 5B). Many signalling pathways were also upregulated, such as those linked to the gap junction as well as GABAergic, serotonergic and cholinergic synapses (S4B Fig). Gap junctions are critical to cell-cell connectivities that help coordinate cellular functions in multicellular organisms, though invertebrates lack the associated connexin proteins and instead possess innexins [93]. Interestingly, a recent study in S. mediterranea identified two innexins that appear to play roles in stem cell response to injury [94]. While the gap junction itself is unmarked in the KEGG pathway of S4C Fig as the pathway is based on that of humans, we found five predicted innexins to be significantly upregulated following fhplk1-RNAi (FhHiC23_g648, FhHiC23_g6222, FhHiC23_g9940, FhHiC23_g15433 and FhHiC23_g15004; S3 Table). Innexins have been localised in various planarian tissues, including the nervous system [95], and their inhibition can lead to developmental abnormalities through polarity disruption during regeneration [95,96], or ablation of neoblast-cell populations [97], suggesting key roles in developmental regulation.

Download:

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

Fig 5. RNAi of fhplk1 in Fasciola hepatica results in upregulation of inter-cell signalling systems.

(A) Biological process GO terms (top 10) associated with ion transport and cell signalling are overrepresented among upregulated genes following fhplk1 RNAi. (B) KEGG pathway showing upregulation of several neuroactive-ligand receptors following fhplk1 RNAi (red = upregulated; grey = no change; blue = downregulated; white = unassigned KEGG ID; red box = unassigned KEGG ID but identified as upregulated from BLAST searches). (C) Log₂foldchange ±SEM of putative F. hepatica neuropeptides following fhplk1 RNAi shows all 22 upregulated, with 15 differentially expressed (black bars). (D) Log₂foldchange of differentially expressed putative receptors and channel subunits (identified via BLAST searches) following fhplk1 RNAi shows upregulation of 134/155 (n = 3; colours correspond to log₂ foldchange; red = upregulated; blue = downregulated).

https://doi.org/10.1371/journal.ppat.1013406.g005

Within the upregulated classical transmitter pathways there was upregulation of ligand synthesis enyzmes, receptors and downstream signalling molecules such as G proteins (serotonergic example shown in S4D Fig). Indeed, of the known F. hepatica classical transmitter pathways [98], only the histamine pathway had no evidence (based on TOPGO, KEGG and BLAST analysis) of upregulation, a poorly understood pathway in parasitic flatworms [99]. Curiously, some upregulated classical transmitter pathways have opposing effects on F. hepatica motility (e.g., acetylcholine is believed to be inhibitory [100], while serotonin is excitatory [101]), making their concurrent upregulation interesting and potentially linked to increased tonic regulation. Serotonin has been implicated in proliferation in other species [102–105] and has been shown to directly stimulate growth/proliferation in S. mansoni and E. multilocularis [106,107]. While the roles of acetylcholine and dopamine in flatworm growth are unknown, they have been linked to proliferation in mammals [108,109]. Furthermore, dopamine deficiency in C. elegans and D. melanogaster constrains development [110,111] and a dopamine antagonist reduced S. mansoni miracidial transformation [104]. C. elegans lacking choline acetyltransferase (ChAT) also display slower growth [112].

Additionally, 16/22 putative neuropeptide encoding genes identified in our dataset (S5 Table) were significantly upregulated in plk1-RNAi worms (Fig 5C), as were their processing enzymes; prohormone convertase 2 (pc2; FhHiC23_g9059), carboxypeptidase E (cpe; FhHiC23_g10997), peptidylglycine alpha-hydroxylating monooxygenase (phm; FhHiC23_g3039) and peptidyl-alpha-hydroxyglycine alpha-amidating lyase (pal; FhHiC23_g9107), along with several predicted neuropeptide receptors (Fig 5B). Neuropeptides have been shown to drive mammalian cell proliferation [113], as well as germline proliferation and regeneration in planarians [114–117]. Furthermore, two neuropeptides and their GPCRs are required for normal maturation of female reproductive structures in S. mansoni [118]. If classical transmitters and neuropeptides have conserved roles in managing growth/proliferation in F. hepatica, then the upregulation observed could indicate that fhplk1-RNAi juveniles lacking neoblast-like cells are attempting to boost cell proliferation via enhancement of selected inter-cell signalling systems.

Previously, we highlighted that signalling systems were upregulated in slower growing F. hepatica with decreased cell proliferation [46]. Here we saw the loss of neoblast-like stem cells, concomitant with nervous system upregulation, indicating potential interplay between proliferating neoblast-like stem cells and neuronal signalling systems. When we compared upregulated genes in our neoblast-like stem cell ablated juveniles with downregulated genes in neoblast-like stem cell-enriched in vivo juveniles [46], we identified 244 shared genes, including 12 neuropeptide encoding-genes (S6 Table). Further, TOPGO analysis of these genes showed that neurotransmitter and Wnt signalling were enriched (S4E Fig). Enrichment of the Wnt pathway is notable as Wnts support proliferation and polarity orientation in regenerating planarians [119], with similar roles proposed in parasitic flatworms [120–122]. RNAi of wnt genes in regenerating planarians can result in aberrant neurogenesis [123–126] and their silencing was recently shown to inhibit F. hepatica neoblast-like cell proliferation and growth/development in vitro [127].

To help us understand the scale of signalling upregulation we examined all BLAST results for the terms ‘receptor’ or ‘channel’ to identify putative intercellular signalling-related genes. Of the 155 putative receptors/channels identified through BLAST searches, 134 (86%) were significantly upregulated following plk1-RNAi (Fig 5D). Those that were downregulated were mainly kinases or nucleus-associated receptors that could be associated with worm neoblast-like stem cells. S4G Fig shows that 22% of the upregulated receptors/channels were predicted rhodopsin peptide GPCRs with voltage-gated potassium channel subunits constituting another large component. Among these genes were two putative FMRFamide-activated amiloride-sensitive Na+ channel subunits [128] that may be of interest as putative drug targets due to their unusual pharmacology and proposed absence in mammals. These data illustrate the extent of upregulation seen within signalling systems following sustained fhplk1-RNAi.

We recognise that the reduced size of plk1-dsRNA worms (in comparison to control dsRNA worms) may increase the proportionality of neuronal tissue, thereby increasing the balance of neuronal-associated transcripts. However, tissue proportionality does not appear to be a key factor in the observed upregulation for a variety of reasons. Firstly, while fhplk1-RNAi worms were indeed smaller (S3A Fig), the difference in size was modest compared to that seen in the gene-function RNAi trials (Fig 3B). Furthermore, while we found DAPI staining (calculated as a percentage of worm area) decreased following fhplk1-RNAi (S4F Fig), the decrease was ~ 20% which may be directly attributable to the absence of neoblast-like stem cells (neoblasts account for around 20–35% of all cells in free-living flatworms [129]). Second, another study observed the downregulation of signalling systems in faster growing, stem cell enriched F. hepatica [46], with much greater size differences (~15x) than those observed in our study. Despite this, genes associated with signalling had comparable fold changes in expression in most instances (S7 Table), again suggesting size/tissue proportionality is not a factor in nervous system upregulation. Thirdly, we did not observe consistent upregulation of transcripts associated with other tissue types, such as the tegument (tsp2) or the gut (18 cathepsins upregulated and 14 downregulated) which indicates that the nervous system transcript upregulation is unusual. Finally, we found this pattern of increased signalling in adult male S. mansoni datasets in which worms were subjected to transcriptomics two weeks after irradiation, or three weeks after the onset of smfgfrA or smh2b RNAi [13]. TOPGO analysis of these datasets showed again that there was an increase in transcripts associated with signalling systems, ion transport and cell communication (S8 Table). Given that these worms were adults of comparable size and that the anterior portion, including cerebral ganglia, had been removed (to exclude testes from analysis) the upregulation of signalling systems following neoblast-like stem cell ablation in diverse circumstances supports the validity of these transcriptional changes in neuronal signalling systems and warrants further investigation.

These data challenge the traditional orthodoxy of using short-term motility-based assays to assess the drug target candidature of various helminth signalling pathways. If the signalling systems highlighted in this study do play significant roles in cell proliferation, then potential long-term impacts of drugs on worm growth/development could be missed with restricted phenotype monitoring windows. Glutamate-gated chloride channels and nicotinic acetylcholine receptors are already targets of current anthelmintics, while GPCRs are thought of as highly ‘druggable’ targets [130]. Repurposing drugs to disrupt neoblast-like stem cell biology could expose novel control strategies.

No correlation between DE miRNAs and predicted DE mRNA targets

We carried out miRNA sequencing on the same RNA samples and used miRDeep2 to map reads to the 150 predicted miRNAs [131]. We found 68 of these miRNAs were expressed in at least two of the six samples, with 13 downregulated and 8 upregulated (S5A Fig; DESeq2, padj <0.001; S9 Table). Target prediction software was then used to identify putative target transcripts of these differentially expressed miRNAs (S9 Table). TOPGO analysis did not identify any enriched GO terms among potential target transcripts. We then plotted GO term frequency of predicted target transcripts against fold change (S5B Fig) but found little, to no overlap with GO terms of differentially expressed mRNAs. Finally, only 36 targets predicted to be downregulated were found to be so, while only 38 of those predicted to be upregulated were so. No clear links between fhplk1-RNAi silencing in juvenile F. hepatica and endogenous miRNA activity were uncovered.

Pharmacological inhibition of FhPLK1 phenocopies fhplk1-RNAi

We next targeted FhPLK1 using the commercially available inhibitor BI 2536. Overexpression of PLK1 in various cancers has led to the development of inhibitors, including BI 2536, a highly selective PLK1 inhibitor which can blunt cell proliferation in various cancer types, both in vitro and in vivo [132,133]. As BI 2536 is an ATP-competitive inhibitor [132], high levels of sequence similarity across human PLK1–3 kinases compromises its selectivity [132,133]. As discussed previously we identified two additional PLK genes in F. hepatica, one similar to human PLK2 (FhPLK2), with the other similar to both human PLK4 (FhPLK4) and SAK (Serum-inducible kinase Akin Kinase) in S. mansoni and E. multilocularis [43,134]. PLK4 divergence in humans renders it less sensitive to BI 2536 [132], a trait also observed for S. mansoni SmSAK [134]. Crystallisation experiments identified Leu₁₃₂ of HsPLK1 as a key residue for facilitating drug-target interaction and selectivity. This leucine was conserved in FhPLK1 (Leu₁₀₄), but not in FhPLK2 or FhPLK4 (S6A Fig), suggesting that the activity of BI 2536 would be specific to FhPLK1 in F. hepatica.

Initial experiments involved exposing NEJs to a broad range of BI 2536 concentrations (0.001 µM – 10 µM) for 18 hours in RPMI, with motility measured as an indicator of viability as motility of F. hepatica had previously been disrupted following BI 2536 exposure [45]. However, in our hands, none of the BI 2536-treated groups exhibited reduced motility (S6B Fig). We monitored rhythmic movement (length of individual parasites plotted across frames) for subtle motility changes (S6C Fig), finding that even those treated with 10 µM BI 2536 retained the classical, ‘accordion’-style movement associated with F. hepatica NEJs (rhythmical lengthening and shortening in a probing fashion) [135]. This was in stark contrast to treatment with 1 µM TCBZ which paralysed juveniles (S6C Fig). Our BI 2536 results are not directly comparable to those of Morawietz et al. [45] as they treated older worms (ex vivo immature and adults) with higher concentrations (20 µM – 100 µM).

As neoblast-like cell inhibition is associated with growth defects [20,21], worms were treated with BI 2536 for 18 h and then cultured for one week in the absence of drug. There were no significant differences between the sizes of juveniles from any treated group and the DMSO-treated controls (Fig 6A). However, incubation of a subset of juveniles with EdU showed significantly reduced proliferation in those treated with 10 µM BI 2536 (Fig 6B; # EdU⁺ nuclei ±SEM: DMSO = 465.1 ± 109.5, 10 µM BI 2536 = 55.1 ± 25.9). A small reduction in EdU⁺ nuclei was also seen in juveniles treated with 1 µM BI 2536, though this difference was not significant. These data suggest that culture following drug exposure afforded the opportunity for flukes to recover. To test this, parasites were cultured for one week before treatment with BI 2536 and subsequent EdU exposure. These worms displayed concentration-dependent reduced cell proliferation across all groups, with significance observed at both 1 µM and 10 µM BI 2536 (Fig 6C; # EdU⁺ nuclei ±SEM: DMSO = 173.8 ± 11.5, 1 µM BI 2536 = 12.7 ± 4.7, 10 µM BI 2536 = 0 ± 0). The observed phenotypic responses are consistent with BI 2536 inhibiting PLK1 activity in F. hepatica, matching observations in other parasitic flatworms [40,43].

Download:

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

Fig 6. Repeated exposure of juvenile Fasciola hepatica to the PLK1 inhibitor BI 2536 in vitro reduces growth and proliferation.

(A) Mean worm area ±SEM of 7-day-old juvenile F. hepatica after exposure to BI 2536 on day 0, followed by 7 days in culture, shows no effect on growth (n ≥ 39 for each treatment). (B) Mean # EdU⁺ nuclei ±SEM in 9-day-old F. hepatica after exposure to BI 2536 on day 0, followed by 8 days in culture, shows 10 µM significantly reduced the number of proliferating cells (n ≥ 7 for each treatment; Kruskal-Wallis and Dunn’s posthoc tests). (C) Mean # EdU⁺ nuclei ±SEM in 9-day-old F. hepatica after exposure to BI 2536 on day 7, shows both 1 µM and 10 µM significantly reduced the number of proliferating cells (n ≥ 7 for each treatment; Kruskal-Wallis and Dunn’s posthoc tests). (D) Mean area (µm²) ±SEM of F. hepatica across 13 days following BI 2536 exposures on days 0, 4 and 8 shows 10 µM significantly reduced growth rate (n ≥ 45 for each treatment at each timepoint; Kruskal-Wallis and Dunn’s posthoc tests). (E) Mean # EdU⁺ nuclei ±SEM in 14-day-old F. hepatica after repeated BI 2536 exposures on days 0, 4 and 8 (timeline from (E) applies to this graph) shows all tested concentrations significantly reduced the number of proliferating cells (n ≥ 5 for each treatment; ANOVA and Holm-Šídák’s posthoc tests). (F) Mean size of EdU⁺ nuclei ±SEM in 14-day-old F. hepatica after repeated BI 2536 exposures (1 and 10 µM) on days 0, 4 and 8 (timeline from (E) applies to this graph) shows increased nuclei area (n ≥ 33 for each treatment at each timepoint; Kruskal-Wallis and Dunn’s posthoc tests). (G) Confocal images of EdU staining (green) in 14-day-old F. hepatica exposed to various concentrations of BI 2536 on days 0, 4 and 8 shows a dose-dependent reduction in # EdU⁺ nuclei and increase in EdU⁺ nuclei area compared to DMSO-treated control; scale bar = 100 µm, DAPI counterstain (magenta). *, p < 0.05; ***, p < 0.001; ****, p < 0.0001.

https://doi.org/10.1371/journal.ppat.1013406.g006

As BI 2536 disrupted neoblast-like cell proliferation we investigated whether it could inhibit juvenile development. For these experiments, parasites were treated with BI 2536 multiple times (on days 0, 4 and 8) with measurements taken on days 1, 5, 9 and 13. After 9 days, worms treated with 10 µM were significantly smaller than control worms (Fig 6D; day 9 mean juvenile area ±SEM: DMSO = 42368 ± 1286 µm², 10 µM BI 2536 = 33861 ± 907 µm²). This phenotype was still evident following a drug-free period post treatment (Fig 6D; day 13 mean juvenile area ±SEM: DMSO = 57568 ± 2035 µm², 10 µM BI 2536 = 43799 ± 1122 µm²). EdU exposure revealed significant, concentration-dependent reductions in cell proliferation across all groups (Fig 6E and 6G; # EdU⁺ nuclei ±SEM: DMSO = 423.8 ± 72.3, 0.1 µM BI 2536 = 268.6 ± 35.5, 1 µM BI 2536 = 107 ± 10, 10 µM BI 2536 = 6 ± 1.8). These results largely phenocopy those seen following fhplk1-RNAi (Fig 3), supporting FhPLK1’s conserved regulatory role in the proliferation of F. hepatica neoblast-like stem cells. We noticed that EdU⁺ nuclei appeared larger in worms treated with higher concentrations of the drug. Therefore, we measured 10 randomly selected EdU⁺ nuclei from five juveniles in each treatment group. This revealed that the EdU⁺ nuclei were significantly larger in worms treated with 1 µM and 10 µM BI 2536 (Fig 6F and 6G; EdU⁺ nuclei area: DMSO = 63.2 ± 1.6 µm², 1 µM BI 2536 = 114.3 ± 8.2 µm², 10 µM BI 2536 = 142.1 ± 18.1 µm²). These larger nuclei could be indicative of neoblast-like cells that had passed through S phase before being delayed/arrested at either the G2/M transition or during mitosis. BI 2536 has previously been shown to increase the DNA content in cancer cells, which was attributed to an inability to undergo cytokinesis [133]. We did not observe larger EdU⁺ nuclei throughout our RNAi experiments. This could be due to the short-term effects of inhibitor exposure which could be reversed following its removal in contrast to the long-term systemic changes to the biology of the worms throughout fhplk1 RNAi [136].

While BI 2536 has previously been shown to affect F. hepatica movement in vitro [45], our plk1-RNAi and pharmacological inhibition experiments highlight how disrupting FhPLK1 activity can undermine neoblast-like cell mediated growth and development. This could weaken F. hepatica juveniles within the host environment, aiding parasite clearance by the immune system. BI 2536 has proven anthelmintic activity against both S. mansoni and E. multilocularis, disrupting gamete production and preventing metacestode formation from germinative cells, respectively [40,43]. Taken together, these data suggest that developing/repurposing drugs to target PLKs could be a promising anthelmintic strategy to control a range of parasitic flatworms, including different life stages.

Identification of putative F. hepatica protein kinases

The PK sequences of the nine Homo sapiens ePK families, along with the aPKs, were downloaded from Uniprot (https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/docs/pkinfam.txt). HMMER (v3.4; http://hmmer.org/) was used to assemble query sequences for each ePK family and the aPKs. HMM searches were carried out using HMMER against the University of Liverpool and Washington University F. hepatica predicted proteomes [47,138] downloaded from Wormbase ParaSite (WBPS19, https://parasite.wormbase.org) [139]. We also looked for additional kinases in a set of novel transcripts. These were generated by aligning all publicly available F. hepatica transcriptomic reads [21,46,47] and in-house transcriptomic datasets against the University of Liverpool genome (WBPS19, https://parasite.wormbase.org/Fasciola_hepatica_prjeb58756/Info/Index/) using HISAT2 (v.2.1.0), before transcript assembly with Stringtie (v1.3.6) to create a master reference gtf file containing 14,684 ‘novel’ transcripts that did not overlap with genes in the reference assembly. All putative hits were screened via BLAST searches against the SwissProt database, and domains annotated using InterPro (https://www.ebi.ac.uk/interpro/about/interproscan/). Hits were retained only when the top BLAST result was a PK, and when a kinase domain was predicted to be present. Duplicates/putative gene copies were identified using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) and manually inspected before removal. The remaining putative F. hepatica PKs were used to create two HMM query sequences (one ePK and one aPK) which were then used in further HMM searches against both F. hepatica predicted proteomes [47,138] to ensure no PKs were missed. Resulting hits were sorted through the same pipeline. BLAST results were used to classify which family/subfamily putative F. hepatica kinases belonged to.

Putative ePKs were clustered using CLANs analysis software (1,000,000 rounds of clustering; [140]). MEGA11 (https://megasoftware.net/, [141]) was used to align aPKs (ClustalW), identify the optimum model for phylogenetic analysis, and construct a Maximum Likelihood (ML) tree with 500 bootstraps. This ML tree was visualised using the Interactive Tree of Life (https://itol.embl.de/).

Published life-stage transcriptomic data from eggs, metacercariae, 1 h newly excysted juveniles (NEJs), 3 h NEJs, 24 h NEJs, three-week juveniles and adults [47] were used to obtain median TPM values for all predicted PKs. Z scores were generated, and heatmaps produced using Morpheus software (https://software.broadinstitute.org/morpheus).

F. hepatica in vitro excystment and culture

The Italian strain of F. hepatica (Ridgeway Research, UK) was used for all experiments. Excystments were carried out as previously described [88,142], and a detailed protocol is available at https://dx.doi.org/10.17504/protocols.io.14egn212qg5d/v1. Briefly, metacercariae were physically ‘popped’ from their outer cyst walls and treated with 10% (v/v) sodium hypochlorite (#1056142500, Sigma-Aldrich) diluted in double-distilled water (ddH₂O) for 2–3 minutes (exact time dependent on metacercariae batch). Popped and bleached metacercariae were then washed in ddH₂O at least five times to remove excess sodium hypochlorite, prior to being stimulated to excyst as detailed by McVeigh et al. [142]. Parasites were cultured in 200 µL CS50 (50% (v/v) chicken serum (#16110082, Thermo Fisher Scientific) in RPMI (#11835105, Thermo Fisher Scientific) with antibiotic/antimycotic (#A5955, Sigma Aldrich)) in 96-well, round-bottomed plates (#83.3925, Sarstedt) in a humidified incubator at 37°C with 5% CO₂ as previously described [20]. Media changes were performed three times per week. F. hepatica in culture media are referred to as juveniles.

RNAi of fhplk1 in vitro

We generated cDNA from four-week-old juvenile F. hepatica by snap freezing 20 juveniles in liquid N₂. Tissues were lysed in a Qiagen tissue lyser with a stainless-steel bead (50 oscillations/sec for 1 min) prior to mRNA extraction with the Dynabeads mRNA Direct Purification Kit (#61011, Thermo Fisher Scientific), DNase treatment (Turbo DNA-free Kit, # AM1907, Thermo Fisher Scientific) and reverse transcription to cDNA using the High-Capacity RNA-to-cDNA Kit (#4387406, Thermo Fisher Scientific). All cDNAs were diluted 1:1 in ddH₂O prior to use. Double-stranded (ds)RNA templates specific to fhplk1 (FhHiC23_g6132) were generated using 0.4 µM T7-labelled primers (5’-TAATACGACTCACTATAGGGT-3’) and the FastStart Taq DNA Polymerase, dNTPack (#4738357001, Millipore Sigma). Primers were designed using Primer3Plus [143], and amplicons were sequenced by Eurofins Genomics (https://eurofinsgenomics.eu/en/) to ensure target specificity. Primers used to generate fhplk1-specific dsRNA templates can be found in S10 Table. Negative control dsRNA targeting bacterial neomycin phosphotransferase [U55762] was generated as previously detailed [142]. Template sizes were checked on a 1–2% agarose gel prior to purification using the ChargeSwitch PCR Clean-Up Kit (#CS12000, Thermo Fisher Scientific) and dsRNA generation using the T7 RiboMAX Express RNA System (#P1700, Promega). All dsRNAs were resuspended in ddH₂O with concentrations/purities checked on a DeNovix DS-11 FX spectrophotometer prior to storage as single-use aliquots at -20°C. RNAi treatments were performed by soaking worms in 50 µL of 100 ng/µL dsRNA diluted in RPMI for 24 h in 96-well round-bottomed plates under standard conditions detailed above. NEJs were immediately exposed to target-specific dsRNA with juveniles subsequently treated twice a week across four weeks to ensure knockdown [88]. Juveniles were cultured in CS50 between RNAi treatments as standard. All RNAi experiments were carried out in triplicate and included no dsRNA-treatment controls.

Juveniles were imaged each week by capturing darkfield videos/images on an Olympus SC50 camera attached to an Olympus SZX10 microscope. Video analyses were performed using the wrMTrck plugin (http://www.phage.dk/plugins/wrmtrck.html, [144]) for ImageJ calibrated to a 1 mm scale. The wrMTrck plugin was used as default except for modifications to minSize (10 pixels), maxSize (1000 pixels), maxVelocity (1000 pixels/frame), AreaChange (200% change in area between frames), Rawdata (2) and benddetect (0). Motility was measured by calculating changes in juvenile length between frames, which provided length change (µm)/minute. Movement of individual juveniles were normalised against mean movement of control juveniles. wrMTrck also provided area (µm²) measurements for each worm.

At the conclusion of RNAi trials juveniles were frozen in liquid N₂ in batches of 20 worms/replicate before cDNA extraction as described above. Quantification of target transcript knockdown was carried out on a Qiagen RotorGene Q 5-plex HRM instrument under the following running conditions: 95°C for 10 minutes; 40 cycles @ 95°C for 10 sec, 55°C for 15 sec and 72°C for 20 sec. Melt-curve analyses were enabled to confirm product specificity and qPCR reactions were performed in triplicate using the SensiFast SBYR No-ROX kit (#BIO-98005, Bioline) with final primer concentrations of 200 nM. Glyceraldehyde phosphate dehydrogenase (fhgapdh) [AY005475] was used as a housekeeper gene. Alongside fhplk1, fhh2b [D915_007751] and fhmlc [FhHiC23_g308] were amplified as markers for proliferating cells and muscle cells respectively. The oligonucleotide primers used can be found in S10 Table. Pfaffl’s Augmented ΔΔCt method [145] was used to calculate relative gene expression. Transcript expression was plotted for both target and control-dsRNA-treated groups relative to no dsRNA control.

Visualisation of proliferative cell activity

The thymidine analogue, EdU (Thermo Fisher Scientific) was used to label nuclei of proliferating neoblast-like cells. A 10 mM stock of EdU was stored at -20°C in PBS. Stock solution was diluted in CS50 to a final working concentration of 500 µM and placed on juveniles for 24 h under standard culture conditions. Immediately following EdU exposure worms were flat-fixed under a coverslip in 4% paraformaldehyde (w/v) in PBS (PFA) for 15 minutes, prior to a further 4 h free-fixing in 4% PFA at room temperature while constantly rotating. Worms were permeabilised in 0.5% Triton X-100 in PBS for 30 min at room temperature. EdU detection was achieved using the Click-iT EdU Proliferation Kit for Imaging with the Alexa Fluor 488 dye (#C10337, Thermo Fisher Scientific). Worms were counterstained using 4′,6-diamidino-2-phenylindole (#D1306, DAPI) in PBS for 20 minutes at room temperature prior to mounting in Vectashield (#H-1000–10, Vector Laboratories). A detailed protocol is available at https://dx.doi.org/10.17504/protocols.io.eq2lyjnrrlx9/v1.

Samples were imaged on a Leica TCS SP8 inverted microscope. Z-stacks consisting of 10–15 optical sections between the dorsal and ventral surfaces were captured for each worm and maximally projected for analyses. EdU⁺ nuclei counts were performed using the cell counter plugin for ImageJ (https://imagej.nih.gov/ij/plugins/cell-counter.html). Nuclei area (µm²) measurements were also performed in ImageJ, where images were calibrated to a 50 µm scale.

RNA-seq and miRNA-seq of fhplk1 RNAi F. hepatica juveniles

NEJs were excysted and treated with target-specific dsRNA as described above (~50 worms/replicate split across two wells). Three replicates for each treatment were carried out in total. Juveniles were treated with dsRNA in RPMI for the final time on day 21 before being snap frozen in liquid N₂ the following day and stored at -80°C. Total RNA was extracted using Trizol reagent (#15596026, Thermo Fisher Scientific) before being sent to the Genomics Core Technology Unit at Queen’s University Belfast (https://www.qub.ac.uk/sites/core-technology-units/Genomics/) for RNA quantification/purity assessment on an AATI fragment analyser prior to library generation using the KAPA mRNA HyperPrep Kit (#KK8580, Roche). Libraries were sequenced (paired end, 2 x 50 bp) on an Illumina Nova seq 6000 SP100 with ~50M reads/sample. Raw read files were uploaded to the European Nucleotide Archive and are available under accession PRJEB85227. Read quality was assessed by fastqc (v.0.11.8) prior to read alignment via HISAT2 (v.2.1.0) against the University of Liverpool F. hepatica genome [47] using genome and GTF files available on Wormbase ParaSite (WBPS19, https://parasite.wormbase.org/Fasciola_hepatica_prjeb58756/Info/Index/). Transcripts were assembled in Stringtie (v1.3.6) and a master reference file, excluding isoforms, was generated by merging transcripts from different replicates with the reference genome. This master reference file was used to generate gene counts via Stringtie (v1.3.6) before export into a format readable by DESeq2. ‘Novel’ transcripts (MSTRGs) generated by Stringtie (v1.3.6) can be found in S2 File.

‘Novel’ MSTRG sequences with no BLAST hit in the NCBI nr database other than F. hepatica and genes with <10 counts across all samples were removed before differential gene expression analysis in R (v4.2.1) by DESeq2 (v.1.34.0). A p-value threshold of 0.001 false discovery rate (FDR) was set to identify differentially expressed genes (DEGs). These DEGs were annotated using the OmicsBox BLAST2GO suite [146] via BLASTx searches against the NCBI non-redundant, landmark and SwissProt databases. Additional BLASTs were carried out against the S. mansoni genome (v10, WBPS19, https://parasite.wormbase.org/Schistosoma_mansoni_prjea36577/Info/Index/). Gene ontology (GO) terms were mapped to genes based on BLAST searches and InterproScan results by BLAST2GO annotation software [146] before TOPGO (v.2.36) analysis (parameters: FDR < 0.05, method = weight01, statistic = fisher) was carried out on DEGs. KEGG IDs were assigned to transcripts using the eggNOG-mapper [147] tool in OmicsBox (https://www.biobam.com/omicsbox) before gage (v.2.44) and pathview (v.1.34) were used to identify up and downregulated pathways. TOPGO (v.2.36) analysis of S. mansoni upregulated genes from Collins et al. [13] was performed by updating all genes to their current IDs and downloading GO terms from Wormbase ParaSite (WBPS19). Parameters were as those used for F. hepatica TOPGO analysis. R scripts are available on Github (https://github.com/pmccusker09/F.hepatica_plk1-RNAi_transcriptome_R_analysis.git).

Remaining total RNA was processed by the Genomics Core Technology Unit at Queen’s University Belfast for miRNA sequencing. Briefly, total RNA was library prepped with the QIASeq miRNA prep kit (#331502, Qiagen) and sequenced on an Illumina Next Seq 2000 with ~8 M single reads per sample (100 bp). Adaptor sequences were trimmed using trimgalore (v.0.4.4) and read quality was assessed by fastqc (v.0.11.8). Raw read files are available under accession PRJEB85228 at the European Nucleotide Archive. Reads were mapped to the University of Liverpool genome downloaded from WormBase ParaSite (WBPS18, https://parasite.wormbase.org/Fasciola_hepatica_prjeb25283/Info/Index/) by bowtie2 (v.2.3.5.1) and miRDeep2 (v.0.1.3) before miRNA counts were generated by miRDeep2 (v.0.1.3), using predicted mature F. hepatica miRNAs [131] as a guide. Only counts that mapped to miRNAs previously identified by Herron et al. [131] were taken forward for differential expression analysis by DESeq2 (v.1.34.0) in R (v4.2.1). The targets of differentially expressed miRNAs were predicted by using the same thresholds as Gillan et al. [148]; miRanda (v.; total score >145, energy < -10), RNAhybrid (v.; p < 0.1, energy < -22) and PITA (v.; seed sequence of 8 bases with DDG < -10). Target prediction was based on the WBPS18 version of the F. hepatica genome, as the miRNAs were predicted from that assembly [131], and so RNA-seq analysis was also carried out against this previous version of the genome assembly using the same methodology as detailed above.

Exposure of juvenile F. hepatica to PLK1 inhibitor in vitro

The PLK inhibitor, BI 2536 (Millipore Sigma) was chosen for testing on juvenile F. hepatica in vitro. BI 2536 has a strong selectivity for PLK1 and so the amino acid sequence of FhPLK1 was examined to determine if key amino acid residues implicated in drug binding were conserved [132]. The PLK1 sequences of F. hepatica, H. sapiens [P53350], S. mansoni [Q5UES2] and E. multilocularis [U6HQM1], were aligned using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/). Other putative F. hepatica PLKs identified in our bioinformatic screen were included in the alignment to help predict BI 2536 activity. Further BLAST searches against the F. hepatica predicted proteomes (WBPS19, https://parasite.wormbase.org) with H. sapiens PLKs1-5 [P53350; Q9NYY3; Q9H4B4; O00444; Q496M5] as query sequences did not identify additional F. hepatica PLKs.

BI 2536 (#A10134, AdooQ Bioscience) and TCBZ (#1681611, Millipore Sigma) were dissolved in 1.7 mL NoStick hydrophobic microtubes (#1210-S0, SSIbio) using dimethyl sulphoxide (DMSO). Juvenile F. hepatica were washed three times in 1 mL RPMI 1640 to remove excess CS50. They were then transferred to 3.5 cm petri dishes (#82.1135.500, Sarstedt) for exposure in up to 10 µM test compound in a final volume of 3 mL RPMI 1640 for 18 h in standard culture conditions. Compounds were added such that the final DMSO concentration was 0.1% (v/v) in 3 mL RPMI. A vehicle control group (DMSO-treated) was included, with treatments carried out in triplicate. At the conclusion of drug incubations juveniles were subjected to phenotypic observation as detailed above. Where NEJs/juveniles were being further cultured post-exposure, they were washed three times in 1 mL RPMI 1640 to remove excess drug, prior to further culturing as previously described.

Statistical analyses

Unless otherwise stated graphs were generated and statistical analyses performed in GraphPad Prism (v10, La Jolla, CA USA). All datasets were tested for normality using either Kolmogorov-Smirnov or Shapiro-Wilk tests. Where data were normally distributed, parametric tests were performed, whereas nonparametric tests were employed where data were non-normally distributed. Relevant post-hoc tests (e.g. Dunn’s, Dunnett’s and Tukey’s multiple comparisons tests) were performed to identify significant differences between multiple groups. Volcano plots were produced using R (v4.2.1). Mean ±standard error mean (SEM) is displayed on graphs unless otherwise stated. Raw data used to generate figures can be found in S11 Table.