Single-cell transcriptomics of the human parasite Schistosoma mansoni first intra-molluscan stage reveals tentative ...

Abstract

Schistosomiasis is a major Neglected Tropical Disease, caused by the infection with blood flukes in the genus Schistosoma. To complete the life cycle, the parasite undergoes asexual and sexual reproduction within an intermediate snail host and a definitive mammalian host, respectively. The intra-molluscan phase provides a critical amplification step that ensures a successful transmission. However, the cellular and molecular mechanisms underlying the development of the intra-molluscan stages remain poorly understood. Here, single cell suspensions from S. mansoni mother sporocysts were produced and sequenced using the droplet-based 10X Genomics Chromium platform. Six cell clusters comprising two tegument, muscle, neuron, parenchyma and stem/germinal cell clusters were identified and validated by in situ hybridisation. Gene Ontology term analysis predicted key biological processes for each of the clusters, including three stem/germinal sub-clusters. Furthermore, putative transcription factors predicted for stem/germinal and tegument clusters may play key roles during parasite development and interaction with the intermediate host.

Single-cell atlas of the first intra-mammalian developmental stage of the human parasite Schistosoma mansoni

Article Open access 18 December 2020

The stage- and sex-specific transcriptome of the human parasite Schistosoma mansoni

Article Open access 07 November 2023

Single-cell transcriptomic analysis of bloodstream Trypanosoma brucei reconstructs cell cycle progression and developmental quorum sensing

Article Open access 06 September 2021

Introduction

Schistosomiasis, infection with parasitic flatworms in the genus Schistosoma, remains a major neglected tropical disease with more than 240 million people affected worldwide and more than 700 million at risk of infection in endemic areas^1,2. To date, a single drug (Praziquantel) is in use; however, this drug is effective only against adult worms, it does not prevent reinfection and drug resistance may be emerging in the field³. One approach that may lead to novel control strategies is to gain a better understanding of the mechanisms underlying life cycle progression, including the cells and their transcriptomic signatures across developmental stages^4,5.

Schistosoma mansoni eggs laid by adult worm pairs dwelling in the portal system of the mammalian host traverse the intestinal wall and pass with the faeces into the environment. In contact with fresh water, from the eggs hatch free-swimming larvae (miracidia) that seek, infect a suitable snail, and transform into mother sporocysts, the first intra-molluscan developmental stage. Within the mother sporocyst, groups of stem cells (historically termed 'germinal cells') start to proliferate and differentiate to develop into daughter sporocysts⁶. By 5 days post infection (5 dpi), developing daughter sporocysts, initially spherical in shape, start to grow within the brood chamber, elongate and become surrounded by a primitive epithelium derived from the mother sporocyst tegument⁶. By ~ 15 dpi they have acquired the definitive vermiform shape containing germinal cells closely packed in the medial part of the body⁶. The mature daughter sporocysts escape from the mother sporocyst, migrate through the snail tissue to the digestive gland area and start to produce whole cercariae from single germinal cells following a second round of embryogenesis⁷. Therefore, from a single miracidium, hundreds to thousands of clonal human-infective cercariae are produced. Extensive knowledge gathered for several decades through detailed histological, and electron microscopy-based studies paved the way towards an understanding of the parasite progression within the snail^8,9,10,11. This knowledge can now be scrutinised using current molecular and 'omics' technologies¹². Shining new light on the cellular and molecular basis of this parasite expansion strategy is critical not only to discover novel aspects of trematode developmental biology, but also to reveal targets for control⁴.

Single-cell transcriptome sequencing (scRNA-seq) has been employed to define cellular subtypes by revealing their specific transcriptional signatures. Compared with so-called 'bulk RNA-seq' studies of whole organisms or tissues, scRNA-seq has exceptional resolving power, being able to detect genes expressed in just a few cells or with low expression levels, but also reveals the stochastic nature of gene expression in individual cells¹³. Single cell transcriptomics have been used in several systems to understand diverse biological processes, such as cell differentiation, tissue specification and development as well as to generate cell "atlases" based on the scRNA-seq profiles across different tissues¹⁴. Studies have employed scRNA-seq in S. mansoni^4,5 to generate cell atlases for male and female adult worms¹⁵, mixed-sex schistosomula¹⁶, the first intra-mammalian developmental stage, and more recently mixed-sex miracidia¹⁷. In addition, scRNA-seq in schistosomes has been employed to define and functionally characterise stem cell populations driving the development of both intra-molluscan¹⁸ and intra-mammalian stages¹⁹.

Important contributions to our current understanding of the developmental biology of schistosome intra-molluscan stages have been made using 'bulk RNA-seq'²⁰. In particular, two germinal cell lineages with distinct proliferation properties had previously been identified and functionally characterised by 'bulk transcriptomics' and RNAi, respectively²⁰. However, there is a critical lack of transcriptomic data and knowledge of gene regulatory networks at the single cell level. Wang and collaborators pioneered work in this area by sequencing 35 sporocyst individual cells, focusing primarily on proliferating stem/germinal cells¹⁸ but most other cell types of this lifecycle stage remain uncharacterised. In the present study, we have followed an untargeted approach to characterise the individual transcriptomic signatures of more than 600 cells isolated from cultured mother sporocysts. Each of the tissue types were spatially validated by fluorescence in situ hybridisation (FISH). Furthermore, we have explored aspects of gene expression regulation, including the prediction of promoter motifs as tentative binding sites for transcription factors, in the stem/germinal and tegumental cell populations. This study contributes to the expansion of the currently scarce number single cell datasets for schistosomes⁵ and reveals key candidate genes involved in the intra-snail developmental phase of this neglected tropical disease pathogen.

Results

Six cell populations identified in the Schistosoma mansoni mother sporocyst

Freshly collected S. mansoni miracidia were transferred into sporocyst media to induce transformation into mother sporocysts. Within the first ~ 16 h most parasites have shed the cilia plates, and their tegument has been remodelled⁸; however, parasite in-vitro development is not synchronous. Therefore, we decided to culture the mother sporocysts for 5 days (named 'D5 sporocysts') to facilitate the complete transformation of > 95% of the parasites^10,21 (Fig. 1A and Supplementary Fig. S1). The D5 sporocysts were collected and processed following the dissociation protocol previously used for schistosomula¹⁶, and live cells enriched and quantified using Fluorescence-activated Cell Sorting (FACS). The droplet-based 10X Genomics Chromium platform was used to generate transcriptome-sequencing data from a total of 601 cells after applying quality-control filters. With the assistance of a high-resolution nuclei quantification protocol, based on a machine-learning imaging platform²², we estimated that a D5 sporocyst contains an average of 169 nuclei (n = 5; range: 112–254) (Supplementary Fig. S2). Therefore, the number of quality-controlled cells theoretically represents > 3.5 × coverage of all cells in a single D5 mother sporocyst.

Based on top markers identified using Seurat, annotation from schistosomula single cell data and genes curated from the literature, we identified six discrete cell populations (Supplementary Table S3); Tegument-1 (138 cells), Tegument-2 (66 cells), Muscle (189 cells), Stem/germinal (119 cells), Parenchyma (66 cells) and Neuron (23 cells) (Fig. 1B). To further explore the biological processes in which the cells of each cluster are involved, we examined over-represented Gene Ontology terms using TopGO (Fig. 1C). Within each cluster, there were clear examples of single genes with annotated roles that are statistically enriched due to their rarity in the genome (Supplementary Tables S4, S5). Although many of these conformed to expectations—for instance, 'neuropeptide signalling' is enriched in the Neuron cluster due to the expression of the known marker gene 7B2—we focussed on over-represented annotation supported by multiple genes. In the case of the Neuron cluster, this highlighted 'neurotransmitter transport', as well as less expected terms 'nucleosome assembly' and 'proton transmembrane transport. In the Muscle cluster, 'cytoskeletal organisation' was enriched, and previously characterised serotonin receptors in schistosomes²³ showed muscle-specific expression (Supplementary Fig. S3). The Parenchyma cluster showed enrichment for several aspects of metabolism including, 'iron homeostasis', 'amino acid metabolism' and 'glycolysis'; whereas the Stem/germinal cell cluster, unsurprisingly, was highly enriched for genes involved in DNA replication, ribogenesis and translation.

In both Tegument clusters, 'purine ribonucleotide/side salvage' was highly enriched, as was 'microtubule-based process' due to the expression of 6–8 dyneins and the dynein-domain protein SmTAL2 (Fig. 1C and Supplementary Tables S4, S5). To further explore differences between the two Tegument subclusters, we first identified differentially expressed genes between Tegument 1 and 2 (Supplementary Tables S6, S7), and then performed a GO term enrichment analysis for each of the sub-clusters. 'Proteolysis', 'metalloendopeptidase activity', and 'metal ion binding' were GO terms supported by at least 2 genes found enriched in Tegument 1 (Supplementary Figure S4A and Table S8). On the other hand, Tegument 2 showed a significant enrichment in the GO terms 'translation', 'structural constituent of ribosome', 'ribosome' and 'large ribosomal subunit', i.e., biological processes and cellular components associated with protein synthesis. Remarkably, four out of the top 5 upregulated genes in Tegument 2 compared to Tegument 1 were ribosomal proteins (Supplementary Table S7); however, the proportion of cells in each of the subclusters expressing these genes was similar, e.g., the 40S ribosomal protein S24 is expressed in more than 90% of the cells in both Tegument subclusters (Supplementary Table S7). Most of the genes highly expressed in Tegument 2 were also expressed in Tegument 1, with few exceptions including an arrestin_C domain-containing protein (Smp_121950) expressed in ~ 54 and ~ 74% of the cells in Tegument 1 and 2, respectively (Supplementary Fig. S4B). On the contrary, several genes were enriched in Tegument 1, e.g. genes expressed in > 90% of Tegument 1 cells and < 50% of Tegument 2 cells, including a Plexin domain-containing protein (Smp_348500 -Supplementary Fig. S4B). Remarkably, 12 genes out of the top 15 upregulated genes in Tegument 1 are uncharacterised proteins (Supplementary Table S6).

The unexpected overrepresentation of biological processes associated with 'nucleotide/side metabolism' in the sporocyst tegument prompted us to compare among the most significantly represented GO terms and top marker genes in the tegument of other developmental stages for which single-cell transcriptomic data are available^15,16,17. The comparative analysis among the miracidium¹⁷, schistosomulum¹⁶, adult¹⁵ and sporocyst showed not only an expected GO term present across all stages (i.e., 'microtubule-based process'), but also biological processes found only in the sporocyst tegument, including purine 'ribonucleoside salvage', 'purine ribonucleotide salvage', 'nucleotide salvage', biological processes linked to nucleic acid metabolism as indicated above (Supplementary Fig. S5 and Table S9). The comparison between top marker genes in the tegument cluster(s) across different developmental stages, revealed genes that may be specific to the sporocyst tegument with tentative functions associated with the absorption of molecules, e.g., Smp_329690 described as nose resistant to fluoxetine protein 6; Smp_169090, a major facilitator superfamily (MFS) domain-containing protein (Supplementary Fig. S6 and Table S10).

Following the comparative analysis across developmental stages performed for the tegument clusters and to further understand the similarities between life stages and the consistency of marker genes, we extended this analysis to other tissues. This revealed conserved and tentative developmental stage-specific GO terms and top marker genes for stem cells (Supplementary Figs. S7, S8 and Tables S9, S11), parenchyma (Supplementary Figs. S9, S10 and Tables S9, S12), muscle cells (Supplementary Figs. S11, S12 and Tables S9, S13) and neurons (Supplementary Figs. S13, S14 and Tables S9, S14). It is important to highlight the limitations of this approach. The data obtained from each developmental stage were generated by different research teams, different versions of the genomes were employed for mapping, and different approaches were followed for the analyses (see Methods). Notwithstanding these caveats, several GO terms specific to the sporocyst tissues were identified, providing indications of tentative functional differences in this intra-molluscan stage compared to other developmental stages. For instance, the GO term 'cellular protein complex disassembly' is only found in neurons of sporocysts, but not in other developmental stages (Supplementary Table S9 and Fig. S13). Similarly, genes expressed in the sporocyst Parenchyma cluster were involved in carbohydrate, amino acid, lipid and iron metabolism, i.e., genes with tentative roles in the catabolism of nutrient molecules derived from the snail tissues. The enzymes UDP- glucose 4—epimerase (Smp_070780) and Ornithine aminotransferase (Smp_000660), the 14 kDa fatty acid-binding protein (Smp_095360), and Ferritin (Smp_063530, Smp_311630 and Smp_311640) were all within the top 30 marker gene list in the Parenchyma cluster of sporocysts, but not in the Parenchyma cluster(s) of the other developmental stages (Supplementary Figs. S9, S10 and Tables S9, S12).

Overall, these findings suggest that cluster-specific gene products display distinct functions in different molecular pathways, and that the identified cell populations may be involved in different biological processes in the mother sporocyst (Fig. 1C and Supplementary Tables S4 and S5). To spatially validate the predicted cell clusters, we defined highly specific cluster-defining marker genes (Fig. 2A and Supplementary Table S15), for which Fluorescence in situ Hybridization (FISH) probes were generated (Supplementary Table S2). We identified cells expressing the Muscle-specific marker myosin heavy chain, location of which correlated with actin filaments following the anterior–posterior axis of the sporocyst (Fig. 2B and Supplementary Fig. S15A, B). Whilst a Neuron-specific marker was expressed across a handful scattered cells in the mid region of the parasite (Fig. 2B and Supplementary Figu. S15C, D), cells expressing the Stem/germinal cluster-specific marker histone H2A were mainly located in clusters towards one pole (Fig. 2C). As stem cells are located in the posterior half of the miracidium larva¹⁷, we hypothesise this is the same in the mother sporocyst, so the Stem/germinal cell markers are thus likely to be highlighting the posterior end. In addition, a few individual Stem/germinal cells were located in the medial region towards the surface of the animal (Supplementary Fig. S15E, F). The two Tegument cell clusters that highly expressed the micro-exon gene 6 (MEG-6), were spatially validated by a strong FISH signal from cells lining the surface of the parasites (Fig. 2D, Supplementary Fig. S16 and Supplementary Video S1). The Parenchyma cluster cells were identified by the cluster specific expression of Smp_318890 (encoding a hypothetical protein). Even though this marker was expressed in < 50% of the cells (Fig. 2A), it showed a high and specific expression in parenchyma cells. These cells seem to be distributed throughout the whole parasite body with a tendency towards the posterior pole (Fig. 2E, Supplementary Fig. S15G, H, and Supplementary Video S2). The parenchymal cells showed clear anterior–posterior cytoplasmic projections containing Smp_318890 transcripts (Fig. 2E, yellow arrowheads).

Stem cell heterogeneity revealed by self-assembling manifold algorithm

Using the self-assembling manifold (SAM) algorithm²⁴, the 119 stem/germinal cells were further analysed. Three discrete subclusters (clusters 0, 1 and 2) with distinct transcriptional profiles were identified (Figs. 3A, 3B and Supplementary Tables S16, S17). We used Scanpy²⁵ to rank genes that characterise each of the stem cell subclusters (Supplementary Table S18) and the five top-ranked genes for each cluster were used for visualisation. STRING²⁶ was used to predict protein–protein interactions among the top 50 ranked genes within each subcluster (Supplementary Table S19). The analysis of subcluster 0 showed a network with strong connectivity of inferred protein–protein interactions among the top marker genes (Figs. 3C). The top STRING terms for subcluster 0, across all categories, are related to ribosomes and translation (Supplementary Table S19). In contrast, weak connectivity was found between genes expressed in subclusters 1 and 2, with no network, far fewer STRING terms, and weaker statistical support (Supplementary Table S19).

Using TopGO we further explored the biological roles of cells within each stem/germinal subcluster by examining over-represented Gene Ontology terms. Related annotations from multiple genes within each cluster were evident; for example, and consistent with the STRING findings, subcluster 0 was highly enriched for biological processes 'translation' and 'ribosome biogenesis' (Figs. 3D and Supplementary Table S20). Statistically significant biological processes, supported by a minimum of two genes, were identified for the other two subclusters. In subcluster 1, 'deoxyribonucleotide biosynthesis' and 'chromosome segregation' suggest active DNA synthesis and cell division (Supplementary Table S20). In subcluster 2, 'actin filament organisation' and 'microtubule-based process', supported by three and six genes, respectively, as well as 'purine ribonucleoside salvage', 'purine ribonucleotide salvage' and 'purine ribonucleoside monophosphate biosynthetic process' indicate considerable similarity to the tegumental gene clusters described above (Supplementary Table S20).

Previously, three stem cell populations in mother sporocysts¹⁸ were identified as kappa, delta and phi based on the differential expression of seven marker genes (klf, nanos-2, fgfrA, fgfrB, p53, zfp-1, and hesl). Although these markers were expressed in our dataset, we were not able to use them to unambiguously assign the Stem/germinal cell subclusters identified by SAM to the kappa, delta and phi cell populations (Supplementary Figs. S17–S19).

Promoter motifs and transcription factor binding sites enriched in two major cell populations

To further investigate gene expression regulation in the Stem/germinal cell population, we searched for putative Transcription Factor Binding Sites (TFBSs). The Tegument-1 cell cluster was included as a control because it represents a somatic/differentiated cell population with a similar number of cells as the Stem/germinal cell population. Motif analysis was performed for marker genes specific to the Stem/germinal cluster (12 genes; Fig. 4A) and the Tegument-1 cluster (49 genes; Supplementary Fig. S20A). We identified five enriched motifs in the promoter region (i.e., 1 kb upstream the Transcription Start Site—TSS) of Stem/germinal cell marker genes, and ten in those of tegumental cells (Fig. 4B, Supplementary Fig. S20B and Table S21). Interestingly, nearly all enriched motifs had ≥ 1 significant match(s) to known TFBSs in the model worm C. elegans based on JASPAR 2022 database, and no overlapping motifs were identified between the two analysed cell clusters (Supplementary Table S22). Within both the Stem/germinal (Fig. 4C) and Tegument-1 cell clusters (Supplementary Fig. S20C), most marker genes have binding sites for multiple transcription factors (TFs). Notably, from the 12 Stem/germinal cell marker genes, 11 contain the motif S-STREME-1, that is similar to the binding site for ceh-22, a homeobox gene and orthologue of the human NKX2-2 gene (Fig. 4D). A further eight genes share the binding motif for pha-4 (S-STREME-5), a forkhead/winged helix factor, and eight genes share a motif (S-STREME-4) similar to the TFBSs for both ceh-22 and vab-7 homeobox genes (Fig. 4C). Six genes share S-STREME-2, which is similar to the unc-30 homeobox binding motif, and three genes share S-STREME-3, similar to the binding motif for the transcription factor sma-4 (SMAD/NF-1 DNA-binding domain factors; Supplementary Table S22). Detailed information of motif locations within the promoter region of Stem/germinal and Tegument 1 marker genes is provided in Supplementary Table S23.

Using the above TFs from C. elegans, we identified putative S. mansoni orthologs. The putative S. mansoni TFs were examined by KEGG ortholog search, and the binding profiles were further confirmed for some candidates using the Jaspar Profile Inference tool²⁷. This screening produced a list of 12 candidate S. mansoni TF genes (Supplementary Table S24): one ortholog of sma-4 (Smp_033950); two orthologs of ceh-22 (Smp_027990 and Smp_186930), pha-4 (FoxA) (Smp_331700 and Smp_158750) and unc-30 (Smp_124010 and Smp_163140); and five orthologs of vab-7 (Smp_147640, Smp_138140, Smp_347890, Smp_308310, Smp_134690). Two of the 12 candidates (Smp_186930 and Smp_033950) were not annotated as TFs in the KEGG search but indicated as orthologues to C. elegans counterparts. Based on a meta-analysis of gene expression for S. mansoni developmental stages²⁸, the Stem/germinal cell cluster specific markers (Fig. 4A) with promoter motifs and TFBSs identified above, tend to show a high expression in the ovary of mature females, miracidia and sporocysts (Supplementary Fig. S22A). Interestingly, the sma-4 ortholog Smp_033950 (Mothers against decapentaplegic or MAD homolog) also showed a high expression across these developmental stages, particularly in sporocysts, relative to the other stages (Supplementary Figure S22A). At the single-cell level, this gene was expressed across all cell clusters, with the highest expression and percentage of cells in the Neuron cluster (Supplementary Figure S22B). Even though the TF Smp_033950 was expressed in < 5% of the stem/germinal cells (Supplementary Fig. S22B), it was the only one among the 12 putative TFs identified in this study (Supplementary Table S22) that showed a Stem/germinal cluster-specific expression and almost exclusively in subclusters 1 and 2 (Supplementary Fig. S22C).

To explore the evolutionary conservation of the predicted of TFBSs within the promoter region of Stem/germinal and Tegument cell cluster marker genes, we searched for enriched common regulatory binding sites across seven Schistosoma species in addition to S. mansoni (Fig. 5A and Supplementary Fig. S21A). Several conserved motifs and combinations of motifs were significantly enriched within 1 kb upstream of the TSS of S. mansoni marker gene orthologues across all the analysed Schistosoma species (Fig. 5B and C, Supplementary Fig. S21B and Supplementary Tables S25–S27). Moreover, the closer the species in the phylogenetic tree, the more conserved the sets of promoter motifs for each marker gene.

In addition to the predicted TFs and TFBSs expressed in Stem/germinal cluster and Tegument clusters, we investigated in our dataset the expression of well-characterised TFs in other developmental stages²⁹. The expression of the flatworm-specific zinc finger proteins genes zfp-1 and zfp-1-1, involved in the tegument specification in adult worms, was detected in few sporocyst cells (< 20) in the Stem/Germinal cluster (in the three subclusters—Supplementary Table S18) and in the Tegument (in particular in Tegument 2—Supplementary Table S3), respectively (Supplementary Figure S23).

Discussion

In this study, we employed single cell transcriptomics, followed by spatial validation, to unveil the cellular components of the mother sporocyst, the life cycle stage that marks the start of rapid asexual proliferation within the intermediate host. Due to the experimental challenges of studying parasites within snail tissues, we have used mother sporocysts cultured in vitro for 5 days, an approach that has been shown to mirror many aspects of in-vivo parasite development in snails^30,31,32,33. For instance, the first scRNA-seq study in schistosomes described the transcriptomic signatures of proliferating cells derived from mother sporocysts transformed in vitro¹⁸. The authors validated their findings in parasites developing in vivo, following parasite progeny through genesis of cercariae and intra-mammalian stages¹⁸. Molecular and cellular differences cannot be completely ruled out; however, in-vivo validation of observations across several stages of development¹⁸ provides strong evidence that the in-vitro miracidium–sporocyst transformation, and culture of early sporocyst stages, is a reliable model to study the parasite development.

Striking changes have been reported during the miracidium–sporocyst transition, including shedding of the ciliary plates, tegumental remodelling and an overall tissue reorganisation²¹ that transforms the free-living miracidium with ~ 365 cells¹⁷ into a simple 'sac-like structure' enriched with stem/germinal cells¹⁰. With the assistance of a novel machine learning approach²², we estimated an average of 169 nuclei in a single D5 mother sporocyst (ranging from 112 to 254 nuclei) and have revealed previously uncharacterised variability in cell number among the parasites at ...

Naegleria Fowleri Symptoms