Stem cell systems and regeneration in planaria

Planarians are members of the Platyhelminthes (flatworms). These animals have evolved a remarkable stem cell system. A single pluripotent adult stem cell type (“neoblast”) gives rise to the entire range of cell types and organs in the planarian body plan, including a brain, digestive-, excretory-, sensory- and reproductive systems. Neoblasts are abundantly present throughout the mesenchyme and divide continuously. The resulting stream of progenitors and turnover of differentiated cells drive the rapid self-renewal of the entire animal within a matter of weeks. Planarians grow and literally de-grow (“shrink”) by the food supply-dependent adjustment of organismal turnover rates, scaling body plan proportions over as much as a 50-fold size range. Their dynamic body architecture further allows astonishing regenerative abilities, including the regeneration of complete and perfectly proportioned animals even from tiny tissue remnants. Planarians as an experimental system, therefore, provide unique opportunities for addressing a spectrum of current problems in stem cell research, including the evolutionary conservation of pluripotency, the dynamic organization of differentiation lineages and the mechanisms underlying organismal stem cell homeostasis. The first part of this review focuses on the molecular biology of neoblasts as pluripotent stem cells. The second part examines the fascinating mechanistic and conceptual challenges posed by a stem cell system that epitomizes a universal design principle of biological systems: the dynamic steady state.

Planarians are flat, soft-bodied worms with a triploblastic body plan and lacking visible segmentation. Many hundred species are known, including fresh- and seawater dwellers and even terrestrial forms. They are members of the phylum Platyhelminthes (platy = flat; helminth = worm), which additionally includes the parasitic clades Cestoda (tapeworms), Trematoda (flukes) and Monogenea (fish gill parasites). “Planaria” is a colloquial term that generally refers to free-living members of the order Tricladida. Planarians have long since attracted the attention of biologists due to their astonishing regenerative abilities, food supply dependent scaling of body size and their great abundance of adult stem cells. The fresh water species are easy and cheap to maintain in the laboratory and several species are studied. The two “workhorses” are Schmidtea mediterranea (Smed) and Dugesia japonica (Dj). Both have excellent regenerative abilities and clonal strains originating from single animals are used. Whether a particular research group works on S. mediterranea or D. japonica is mainly a question of habit and results are so far assumed to be comparable. Other planarian model species include Schmidtea polychroa (Spol; used as embryogenesis model due to consistent fecundity) and Dugesia ryukyuensis (Dr; model for switching between sexual and asexual reproduction modes) (Ishizuka et al. 2007). Gene names in the planarian literature carry a prefix designating the species (Reddien et al. 2008) (e.g., Smed-actin for Schmidtea mediterranea actin). The genome of S. mediterranea has been sequenced (Robb et al. 2008), and a genome project for D. japonica is underway. The current planarian tool kit further includes organism-wide RNAi (Sánchez Alvarado and Newmark 1999; Reddien et al. 2005a), BrdU-labeling (Newmark and Sánchez Alvarado 2000), in situ hybridization (Pearson et al.

2009; Umesono et al. 1997), FACS fractionation of stem cell populations (Hayashi and Agata 2012; Hayashi et al. 2006) and next generation sequencing techniques (Friedländer et al. 2009; Palakodeti et al. 2008). Beyond the Tricladidans, the flatworm species Macrostomum lignano is increasingly studied (Morris et al. 2006). Flatworms are amongst the first model systems within the so far scarcely investigated superphylum Lophotrochozoa and therefore also provide interesting evolutionary perspectives.

Superficially, fresh water planarians may seem rather boring — flattened, mostly drab-colored worms without visible appendages (Fig. 1). A closer look reveals a set of organ systems similar to other triploblastic animals: A brain comprising diverse neurotransmitter systems (Umesono and Agata 2009), a highly branched gastrovascular cavity tasked with both the digestion and distribution of nutrients (Forsthoefel et al. 2011), a protonephridial excretory system with interesting evolutionary homologies to the vertebrate kidney (Rink et al. 2011; Scimone et al. 2011), diverse suites of sensory organs and a hermaphroditic reproductive system (Newmark et al. 2008). However, planarians are truly astonishing in terms of their biology. Akin to mythological beasts, they have the ability to regenerate in their entirety even from tiny injury remnants and the asexual strains appear to be exempt from the mortal’s plight of physiological ageing (Mouton et al. 2011; Pearson and Sánchez Alvarado 2008; Tan et al. 2012).Fig. 1Examples of European planarian species. From left to right: Polycelis sp., Planaria torva (recently fed), Dendrocoelum lacteum, Schmidtea polychroa, Dugesia gonocephala (recently fed), Schmidtea mediterranea

Both traits originate from an abundance of adult stem cells. Collectively referred to as “Neoblasts”, these cells have been investigated for more than a century (Baguñà 2012). Recently, Wagner and colleagues showed that transplantation of a single Neoblast into a lethally irradiated (=stem cell depleted) worm rescued the recipient and gave rise to a perfectly healthy animal of the donor genotype (Wagner et al. 2011). This experiment conclusively demonstrated the pluripotency of Neoblasts (defined here as the ability to differentiate into all zygotic cell types). Neoblasts are likely even totipotent (differentiation into all zygotic cell types + extraembryonic tissues), but the use of donor Neoblasts from asexual animals in the above experiments precluded the reconstitution of sexual reproduction and thus the formal demonstration of totipotency. Neoblasts as naturally occurring pluripotent adult stem cells are remarkable, because adult stem cells in other model systems are lineage-restricted and somatic pluripotency exists only transiently during early embryonic development. A further unusual feature of planarian Neoblasts is their high basal mitotic activity. The resulting stream of progenitors drives the continuous turnover of all planarian tissues, which are likely devoid of any long-lived cell types. Dynamic turnover characterizes also the vertebrate intestine, for example (van der Flier and Clevers 2009). Yet the complete turnover of an entire triploblastic animal in a matter of weeks is surely a further fascinating feature of planarians.

The above points raise a number of intriguing questions with respect to planarian biology and stem cell systems in general: What makes Neoblasts pluripotent and is pluripotency evolutionarily conserved? Which principles and mechanisms orchestrate the orderly generation of all organismal cell types from one pluripotent stem cell population? What maintains organismal homeostasis in face of changing flux rates between stem- and differentiated cell compartments? The revival of planarians as molecular model system is starting to provide insights into the above questions. This text takes a deliberately stem cell focused approach. Several recent reviews examine in detail regeneration, pattern formation or the unusual embryonic development of planarians (Tanaka and Reddien 2011; Forsthoefel and Newmark 2009; Aboobaker 2011; Adell et al. 2010; Martín-Durán et al. 2012; Reddien 2011).

The long history of Neoblast research has been recently reviewed (Baguñà 2012). Nevertheless, the lasting legacy of the pre-molecular era requires a brief historical excursion also at this point. The term “Neoblast” came into use to describe the small, roundish cells found abundantly throughout the planarian mesenchyme, except for the area in front of the photoreceptors and the pharynx (which are the only areas incapable of regeneration) (Reddien and Alvarado 2004). In transmission electron microscope (TEM) images, Neoblasts appear as 5- to 10-μm diameter cells with a thin rim of cytosol, lots of free ribosomes, few discernible organelles, prominent chromatoid bodies (CBs) (see below) and a large nucleus with little heterochromatin (Pedersen 1959; Hori 1982; Hay and Coward 1975; Coward 1974). The accumulation of Neoblasts at regenerating wounds and their rapid loss after regeneration-inhibiting doses of irradiation linked Neoblasts to regeneration (Wolff and Dubois 1948). The observation that all cell divisions in planarians occurred exclusively in cells meeting the above morphological criteria culminated in the statement “in planarians, Neoblasts are the only cells that divide” (Morita and Best 1984). This made the term “Neoblasts” practically synonymous with “dividing cells”. Consequently, generic cell division markers such as phospho-Histone H3 Ser10 (H3P), BrdU incorporation, or expression of cell division machinery components such as PCNA (Newmark and Sánchez Alvarado 2000; Orii et al. 2005; Salvetti et al. 2000) have been and are commonly used as Neoblast markers.

PCNA and H3P as stem cell markers? In other systems, true stem cells capable of self-renewal and the production of multiple progeny are generally rare amongst the cycling cells within a tissue (Weissman 2000). Transit amplifying cells, that is dividing stem cell progeny with limited self-renewal potential, tend to constitute the greatest proportion of mitotic cells and even differentiated cell types like fibroblasts or endothelial cells may divide occasionally in vertebrates. Generic cell division markers such as PCNA are consequently of little use in singling out stem cells from other dividing cell types. In planarians, however, cell divisions never occur within the confines of differentiated tissues. Not even high turnover tissues like gut or epidermis harbor dividing cells (Forsthoefel et al. 2011; Baguñà 1976a; Newmark and Sánchez Alvarado 2000), contrasting sharply with the high rates of cell divisions and organ-specific stem cell populations in the corresponding tissues of vertebrates or the Drosophila gut (van der Flier and Clevers 2009; Blanpain and Fuchs 2009; Micchelli and Perrimon 2006). Instead, planarian cell divisions are strictly limited to the loosely organized mesenchyme surrounding all organs and, more specifically, to cells meeting the morphological definitions of Neoblasts detailed above. Stating that “Neoblasts are the only dividing cells in planarians” therefore accurately captures the restriction of mitotic activity to one morphologically homogenous cell population and the absence of organ-specific stem cells or autonomously self-renewing cells outside the planarian mesenchyme.

On the other hand, the term “Neoblasts” and general statements in the planarian literature about what Neoblasts do or do not do often imply functional homogeneity in absence of hard evidence. For example, it is simply not known whether all PCNA- or H3P-positive “Neoblasts” are bonafide stem cells or whether planarians also harbor transient amplifying cell types that would consequently contribute to the PCNA-positive population. Moreover, neither morphological criteria nor the current molecular markers can draw an exact boundary between dividing cells and early committed progeny (see below). It is therefore important to stress that the term “Neoblasts” in the planarian literature and explicitly also in this text needs to be understood as a general reference to the planarian stem cell system, inclusive of pluripotent stem cells, earliest postmitotic progeny and any possible intermediate stages. The source of such ambiguity is not the use of a historical term, but the current lack of knowledge regarding the planarian stem cell system. Instead of introducing new terminology, I will therefore adapt the term “cNeoblasts” (“clonogenic Neoblasts”) from Wagner et al. for explicit references to pluripotent stem cells (Wagner et al. 2011) and “Neoblast progeny” when referring to early postmitotic differentiation stages. Neoblast heterogeneity will be further discussed below.

Pluripotent stem cells as potentially unlimited in vitro source of each and every cell type are of the greatest interest to regenerative medicine. Adult vertebrates do not harbor pluripotent stem cells, but pluripotent cells can be derived either from early embryos (embryonic stem [ES] cells; Weissman 2000) or via the recently discovered “reprogramming” of somatic cells back into a quasi-embryonic state (induced pluripotent stem [IPS] cells; Takahashi and Yamanaka 2006). Fundamental questions remain regarding the mechanistic basis of cellular pluripotency and the orderly transition to progenitor differentiation. Their continuous activity, abundance in adult tissues and evolutionary distance to vertebrates predestine planarian Neoblasts as model system for stem cell pluripotency (Fig. 2).Fig. 2Organization of the planarian stem cell system. a Top: stem cell distribution as visualized by whole mount in situ hybridization with the neoblast marker smedwi-1. The image is a maximum intensity projection spanning the D/V axis (anterior is to the right). The stem cell devoid pharynx occupies the dark central area. Scale bar: 200 μm. Bottom: cartoon illustration of the stem cell distribution in a transverse section at the level of the pharynx. Stem cells (red), Ep epithelium, Bm basement membrane, Phx pharynx, VNC ventral nerve cords, Gut lateral branch of the gastrovascular system. b Magnified head region. smedwi-1 RNA expressing Neoblasts (red) are largely absent from the area in front of the photoreceptors (asterisks), but SMEDWI-1 protein persists in anteriorly migrating stem cell progeny (green, antibody staining). Scale bar: 200 μm. c Only smedwi-1 RNA expressing cells (red) divide (blue, H3P antibody staining). Scale bar: 25 μm

Understanding the mechanistic basis of Neoblast pluripotency consequently has been and continues to be a major focus of planarian research. The operational definition of Neoblasts as only dividing cells and the rapid elimination of dividing cells by X-irradiation have strongly shaped the experimental approaches. Comparisons between irradiated and non-irradiated samples yielded riboprobes for the localization of Neoblasts (Shibata et al. 1999; Salvetti et al. 2000; Reddien et al. 2005b), gating criteria for isolating Neoblast-containing cell populations by FACS (Hayashi et al. 2006) and, more recently, a basis for large-scale surveys of radiation-sensitive gene expression profiles (Rossi et al. 2007; Eisenhoffer et al. 2008; Wagner et al. 2012; Solana et al. 2012; Friedländer et al. 2009; Blythe et al. 2010). Further, the stereotypical progression of head regression, ventral curling and eventual lysis following lethal irradiation (Bardeen and Baetjer 1904) established a benchmark for Neoblast ablation RNAi phenotypes (Reddien et al. 2005a, b). Recently, Neoblast depletion by RNAi against a Neoblast specific Histone variant H2B has been proposed as more specific alternative to irradiation, circumventing the substantial transcriptional responses to organism-wide DNA damage (Solana et al. 2012).

Amongst the long list of genes now known to be expressed in and/or required for Neoblast function, three groups clearly stand out (Solana et al. 2012; Wagner et al. 2012; Labbé et al. 2012; Eisenhoffer et al. 2008; Rossi et al. 2007; Shibata et al. 2012; Onal et al. 2012; Rouhana et al. 2010): First, and hardly surprising given the experimental premises, components of the cell division machinery. Telomerase is found in this category and its regulation entails intriguing responsiveness to regeneration and reproduction modes (Tan et al. 2012). Second, a striking number of evolutionarily conserved proteins involved in posttranscriptional regulation of gene expression. Third and barely characterized, a growing list of chromatin modifiers.

The planarian piwi homologues have arguably received the greatest attention amongst Neoblast markers. S. mediterranea piwi-1 (smedwi-1) is highly expressed in an abundant cell population between the gut branches (Fig. 2a). smedwi-1 + cells are absent from the pharynx and the area in front of the photoreceptors (Fig. 2b), become almost undetectable within 24 h of irradiation (Reddien et al. 2005b) and only smedwi-1 expressing cells divide (Fig. 2c) (Eisenhoffer et al. 2008; Guo et al. 2006; Yoshida-Kashikawa et al. 2007). The additional S. mediterranea piwi homologues -2 and -3 are likewise expressed in Neoblasts (Reddien et al. 2005b; Palakodeti et al. 2008; Nakagawa et al. 2012b). Further, RNAi-mediated knock down of smedwi-2 and -3 or the respective Dugesia ryukyuensis homologues result in lethal Neoblast depletion, recapitulating the stereotypical deterioration of irradiated animals (Reddien et al. 2005b; Palakodeti et al. 2008; Nakagawa et al. 2012b). Overall, the planarian piwi genes therefore extend the morphological definition of Neoblasts to the molecular level.

Piwi and the closely related Argonautes are highly conserved mediators of gene regulation (Farazi et al. 2008). Argonautes are widely expressed and mediate gene silencing via small RNA like miRNA and siRNAs. In contrast, piwi homologues in established model systems are mostly germ line specific. The planarian piwi gene family has undergone a drastic expansion, generating at least seven piwi genes in S. mediterranea as opposed to three family members in fly or mouse (Juliano et al. 2011; Palakodeti et al. 2008; Friedländer et al. 2009). One essential function of piwi proteins is transposon silencing via piwi-associated small RNAs (piRNA) (Thomson and Lin 2009), sparking the honorific “guardians of the genome” (O’Donnell and Boeke 2007). Transposons have been estimated to account for 31% of the S. mediterranea genome (Friedländer et al. 2009). Since transposon-mediated mutations and genome rearrangements are likely as important a concern in indefinitely self-renewing pluripotent Neoblasts as in the germ line, could the essential role of piwi genes in Neoblasts reflect a similar role in the maintenance of genome integrity?

At least indirect evidence for a role of planarian PIWI proteins in transposon silencing comes from a recent study investigating the evolutionarily conserved arginine methyl-transferase Smed-PRMT-5 (Rouhana et al. 2012). PRMT-5 methylates RG containing motifs in RNA-binding proteins such as PIWI, SmB and Vasa, which subsequently serve as docking sites for proteins containing methyl-binding TUDOR domains. This interaction network provides an evolutionarily conserved assembly scaffold for ribonucleoprotein (RNP) complexes such as germ granules. Germ granules are a near-universal feature of germ cells and play essential roles in their formation and maintenance (see below) (Voronina et al. 2011; Eddy 1975). The conspicuous CBs of planarian Neoblasts are also RNPs (Auladell et al. 1993; Hori 1982; Coward 1974) and Rouhana et al. found that CBs carry SMED-PRMT-5 methylations. Importantly, smedwi-3 was identified as PRMT-5 substrate and PRMT-5(RNAi) resulted in gradual Neoblast depletion and eventual death of the animals. The concomitant and massive upregulation of transposon transcription and the enrichment of piRNA-like RNA in CBs demonstrated a function of CBs and smedwi-3 in transposon silencing (Rouhana et al. 2012). Consistently, abundant and diverse populations of likely piRNAs have been identified in Neoblasts, of which >30% map to annotated transposons (Palakodeti et al. 2008; Friedländer et al. 2009). However, the rapid decline of Neoblast numbers after smedwi-3(RNAi) and especially smedwi-2(RNAi) in comparison to the gradual decline after PRMT-5(RNAi) strongly suggests that the planarian PIWI proteins are required for more than transposon silencing, as are their germ line counterparts in other organisms (Klenov et al. 2011; Juliano et al. 2011).

In fact, the smedwi-2 and -3(RNAi) phenotypes suggest a role in Neoblast differentiation as well as self-renewal (Reddien et al. 2005b; Wagner et al. 2012). The future characterization of PIWI-associated small RNAs might identify the relevant targets and thus provide mechanistic insights into maintenance and exit from pluripotency.

Besides smedwi-3 and PRMT-5, a surprising number of known Neoblast genes have links to CBs. The TUDOR homologue SPOL-TUD localizes to CBs and Spol-TUD-1(RNAi) depletes Neoblasts with very similar kinetics to Smed-PRMT-5(RNAi), consistent with a scaffold function of methyl-RG/TUDOR interactions in CBs (Solana et al. 2009; Rouhana et al. 2012). However, neither RNAi treatment completely disassembled CBs, which could indicate redundant mechanisms or molecular heterogeneity amongst CBs. The DEAD-box RNA helicase DJ-CBC-1 is required for Neoblast differentiation (Rouhana et al. 2010) and localizes to a subset of CBs, consistent with the latter possibility (Yoshida-Kashikawa et al. 2007). SMED-SmB, a further likely substrate of PRMT-5 (Rouhana et al. 2012), localizes to CBs and SmB(RNAi) causes rapid chromatoid body disassembly and Neoblast loss within a matter of days (Fernandéz-Taboada et al. 2010). RNAi against the RNA binding proteins Smed-bruli (a Bruno homologue) (Guo et al. 2006), Dj-pumilio (Salvetti et al. 2005) and Smed-vasa-1 (Wagner et al. 2012) all result in the loss of Neoblasts and/or differentiation defects.

The above list of molecules is intriguing also because it reads like a veritable “who is who” in germ line specification. Homologues of the above neoblast constituents are generally necessary for germ granule and/or germ cell function, revealing molecular parallels between CBs and germ granules and between Neoblasts and germ cells in general (Shibata et al. 2010; Wang et al. 2010; Wagner et al. 2012; Solana et al. 2012; Voronina et al. 2011; Ewen-Campen et al. 2010).

Parallels between germ cells and Neoblasts are surprising, because Neoblast are not germ cells. Sexually reproducing planarians harbor a bonafide germ line as part of their hermaphroditic reproductive system (Newmark et al. 2008). Both germ line and the somatic gonads can regenerate de novo, likely involving a population of putative germ line stem cells. These cells are very similar to “ordinary” Neoblasts in terms of morphology and radiation sensitivity, even though they may cycle at a slower rate (Wang et al. 2007; Sato et al. 2006; Handberg-Thorsager and Saló 2007). Interestingly, the factor that distinguishes somatic Neoblasts from germ line Neoblasts is expression of nanos, so far the only highly conserved germ line determinant that is not constitutively expressed in all Neoblasts (Wang et al. 2007; Sato et al. 2006; Handberg-Thorsager and Saló 2007). Consistently, nanos(RNAi) does not affect Neoblast maintenance, but prevents gonad regeneration (Wang et al. 2007; Nakagawa et al. 2012a). Neoblasts therefore appear poised fascinatingly close to the soma/germ line divide, requiring only a small push involving nanos expression to transition into germ line fate. The resemblance between Neoblasts and germ cells could reflect a peculiarity of planarian biology. However, the pluripotency of Neoblasts provides a conceptual link between the two cell types. Even though germ cells only give rise to gametes, gamete fusion during fertilization initiates development of the complete organism, thus necessitating pluripotency within the germ line. The ectopic differentiation of somatic cell types in germ line tumors (Teratomas) dramatically emphasizes the cryptic pluripotency of germ line cells (Juliano et al. 2010; Ciosk 2006; Seydoux and Braun 2006).

Could the striking component sharing between germ cells and Neoblasts at the level of their RNPs reflect an ancient role of these structures in pluripotency (Juliano et al. 2010; Seydoux and Braun 2006)?

“Germ granules” are likely a heterogenous collection of germ line RNPs, which share components with RNPs in somatic cells (Anderson and Kedersha 2009). CB-like RNPs also occur in planarian neurons (Yoshida-Kashikawa et al. 2007) and the commonly observed expression of “Neoblast genes” in neurons might reflect similar component sharing (Solana et al. 2009, 2012; Rouhana et al. 2010; Rossi et al. 2012; Wagner et al. 2012). Germ granule disassembly frequently results in germ line loss and induction of excess granules can induce ectopic germ cells (reviewed by Ewen-Campen et al. 2010). Such importance of RNPs arises from their role as veritable control centers of posttranscriptional gene regulation (PTGR), including the transcriptional repression and storage of mRNA, transcript degradation via piRNA and miRNA, translational activation of specific RNAs and even epigenetic chromatin modification (reviewed by Voronina et al. 2011). Germ granules therefore likely exert a considerable influence on the translatable pool of mRNAs and thus ultimately the proteome of the cell. This role might be of particular importance to pluripotent cells: Their ability to differentiate into all cell types necessitates globally accessible chromatin, yet at the same time the emergence of differentiation promoting transcriptional circuits must be suppressed (reviewed by Koh et al. 2010). Pluripotency therefore might generally come at the cost of reduced transcriptional control options, consistent with widespread low-level transcription of the genome in pluripotent ES cells (Efroni et al. 2008; Gaspar-Maia et al. 2011). PTGR via germ granules and CBs could supply additional layers of gene expression control important for stabilizing the pluripotent state (Seydoux and Braun 2006; Juliano et al. 2010).

Further, the predominance of PTGR in the orchestration of germ cell development has been interpreted in favor of reduced transcriptional control options (Ewen-Campen et al. 2010). PTGR components are also strikingly enriched in Neoblasts (Rouhana et al. 2010; Solana et al. 2012; Onal et al. 2012; Labbé et al. 2012), including the practically Neoblast-specific expression of what common knowledge would consider generic translation initiation factors (Rouhana et al. 2010). PTGR can therefore be expected to play similarly important roles during Neoblast maintenance and differentiation as during germ line differentiation, possibly reflecting common pluripotency related constraints on transcriptional control mechanisms.

However, the discovery that overexpressing a cocktail of transcription factors can convert terminally differentiated cells back into a pluripotent state has placed dramatic emphasis on the importance of transcriptional mechanisms in the establishment and maintenance of pluripotency (Takahashi and Yamanaka 2006). IPS cell formation usually requires expression of the POU-domain transcription factor Oct4 and Sox2, which via mutual binding sites in their promoters establish self-maintaining transcriptional networks incorporating the pluripotency factor Nanog (Ng and Surani 2011). However, only a small fraction of cells overexpressing all reprogramming factors ever become pluripotent (Takahashi and Yamanaka 2006). Chromatin rearrangements are thought to constitute a rate limiting step in pluripotency establishment (Rouhana et al. 2012; Koh et al. 2010). Epigenetic chromatin modifications as well as the mechanistic contributions of Oct4/Sox2/Nanog targets to the pluripotent state are therefore an intense focus of current stem cell research.

Which, if any, of the above mechanisms participate in Neoblast pluripotency? Besides the much-reduced content of heterochromatin in TEM images (Hay and Coward 1975), practically nothing is known about chromatin regulation in planarian Neoblasts. The striking enrichment of chromatin modifiers in Neoblast gene expression profiles (Solana et al. 2012; Wagner et al. 2012; Onal et al. 2012; Labbé et al. 2012) emphasize the urgent need for detailed studies. A planarian Nanog-homologue has not been reported yet, but Neoblasts specifically express Sox- and POU-domain transcription factors (Wagner et al. 2012; Onal et al. 2012) and Smed-soxP1 is required for Neoblast maintenance (Wagner et al. 2012). Further research needs to address whether these factors maintain a molecularly conserved pluripotency network or whether they carry out general “housekeeping” functions in planarian Neoblasts. Two recent studies directly compared Neoblasts to vertebrate stem cells. Onal et al. examined statistically the relative homologue abundance of vertebrate pluripotency factors and known Oct4/Nanog target genes amongst Neoblast transcriptomes and proteomes (Onal et al. 2012). In both cases, the authors found statistically significant enrichments, consistent with broad evolutionary conservation of pluripotency. Similarly, Labbé and colleagues reported particularly high degrees of overlap between a vertebrate pluripotency associated gene set and genes specifically upregulated in Neoblasts (Labbé et al. 2012). Together, these findings provide an exciting starting point for the comparative dissection of pluripotency across large evolutionary distances (planaria: protostomes/lophotrochozoa; vertebrates: deuterostomes).

A first requirement are data addressing the function of planarian Sox and POU factors in molecular detail, including DNA binding sites, regulatory targets and cross-species rescue experiments in vertebrates. Should such studies reveal the existence of an ancient pluripotency core network, comparisons between Neoblasts and ES/IPS cells could identify conserved components/targets and thus aide in the mechanistic understanding of pluripotency. The alternative outcome, namely pluripotency as an epiphenomenon not relying on explicitly conserved molecular circuits, would be equally interesting. In this case, comparisons could help in revealing the general cellular processes that collectively give rise to pluripotency as an emergent systems property (e.g., “open chromatin”, “PTGR”, particular metabolic states). Further, the mechanistic contribution of CBs and germ granules to pluripotency remain an important area for future exploration, despite the fact that neither ES nor IPS cells seem to harbor prominent RNPs. It is important to stress that both ES and IPS cells are artificial laboratory products. In vertebrates at least, somatic pluripotency in vivo exists only transiently during the earliest stages of embryonic development. Neoblasts and germ cells on the other hand have evolved as permanently pluripotent cells. Their joint reliance on RNPs could therefore reflect critical accessory functions required for the long-term maintenance of pluripotency. The established role of germ granules and CBs in safeguarding genome integrity (see above) fits this picture. The erasure of epigenetic marks, which remains incomplete in reprogrammed cells (Lister et al. 2011; Ohi et al. 2011), could be a second such possibility worth exploring.

Overall, planarian cNeoblasts as naturally occurring pluripotent stem cells ideally complement vertebrate in vitro systems towards the far goal of medically exploiting pluripotency.

Beyond molecular mechanisms, the distribution of pluripotency within the neoblast population raises the next set of urgent questions. Are all cells expressing smedwi-1, PRMT-5 or any of the other Neoblast markers pluripotent cNeoblasts? Or are cNeoblasts a minority amongst a majority of transient amplifying progenitors? Expression patterns of Neoblast genes have so far been of little help in this respect. Many Neoblast genes display the typical expression pattern epitomized by smedwi-1 and, importantly, are coexpressed at the single-cell level (Guo et al. 2006; Wagner et al. 2012; Eisenhoffer et al. 2008; Reddien et al. 2005b). Others are likely expressed only in subpopulations of smedwi-1 + cells (Sakurai et al. 2012; Rossi et al. 2006; Scimone et al. 2010). Further evidence for heterogeneity in terms of morphology and gene expression comes from FACS-sorted radiation sensitive cell populations (Higuchi et al. 2007; Morris et al. 2006; Hayashi et al. 2010; Shibata et al. 2012). Dynamic expression pattern changes during the recovery from near-lethal irradiation have further been interpreted in favor of a more radiation-resistant Neoblast subpopulation in association with the ventral nerve cords (Salvetti et al. 2009). However, the relevance of these observations to Neoblast potency remains tentative, because a conclusive differentiation between stem cells and transient amplifying cells necessitates assessment of cell lineages.

The single Neoblast transplantations by Wagner and colleagues (2011) are landmark achievements also because they introduce this dimension to Neoblast research. Starting from a cell population sorted according to size and complexity ((X1)FS), the authors report seven whole-animal reconstitutions out of 130 single cell transplant attempts (Wagner et al. 2011). The starting population therefore must have contained at least 5% pluripotent cells, whereas the actual proportion might be significantly higher due to the technical challenges involved.

More than 5% pluripotent cells in a label-free sorted fraction lead to the important conclusion that cNeoblasts cannot be as exceedingly rare as, for example, vertebrate haematopoietic stem cells (Lensch and Daley 2004). However, the proportion of cNeoblasts amongst smedwi-1 + cells and consequently the possible existence of transient amplifying or lineage-restricted populations remain unknown. A further important milestone will be the localization of cNeoblasts within the planarian tissues. Addressing these issues will require a new generation of experiments linking molecular markers with assays of differentiation potential. The recently reported sub-fractionation of living Neoblasts with surface antibodies represents a further step into this direction (Moritz et al. 2012).

The maintenance of appropriate stem cell numbers is a central challenge in all stem cell systems. Excessive stem cell divisions can lead to cancer, a loss of stem cells to a halt of tissue turnover and premature ageing (Arwert et al. 2012). Stem cell homoeostasis has to balance the two foundations of stemness, self renewal and differentiation into multiple progeny (Weissman 2000). Every single stem cell division challenges homeostasis: Symmetric divisions resulting in two stem cells increase stem cell numbers. Symmetric divisions resulting in two differentiating daughters signify a net loss of one stem cell. Only asymmetric divisions (i.e., the production of one stem cell and one differentiating daughter) maintain the status quo. Drosophila neuroblasts regulate the fate choice cell intrinsically via the asymmetric segregation of differentiation determinants into their daughter cells (Prehoda 2009). The mammalian gut is an example of a stem cell system operating without intrinsic division asymmetries (van der Flier and Clevers 2009). Such systems can maintain homeostasis at the population level despite stochastic outcomes of individual divisions (González-Estévez et al. 2012b; Simons and Clevers 2011). Practically nothing is known regarding the regulation of cNeoblast divisions at the single cell level. However, the fact that cNeoblasts self-amplify (i.e., divide symmetrically) in addition to producing progeny (i.e., asymmetric divisions) during the recovery from partial irradiation strongly suggests population-level control mechanisms (Wagner et al. 2011, 2012). Further progress along these lines, including questions regarding the niche of cNeoblasts, will have to await their localization within the tissue.