The stem cell system is one of the unique systems that have evolved only in multicellular organisms. Major questions about this system include what type(s) of stem cells are involved (pluri-, multi- or uni-potent stem cells), and how the self-renewal and differentiation of stem cells are regulated. To understand the origin of the stem cell system in metazoans and to get insights into the ancestral stem cell itself, it is important to discover the molecular and cellular mechanisms of the stem cell system in sponges (Porifera), the evolutionarily oldest extant metazoans. Histological studies here provided a body of evidence that archeocytes are the stem cells in sponges, and recent molecular studies of sponges, especially the finding of the expression of Piwi homologues in archeocytes and choanocytes in a freshwater sponge, Ephydatia fluviatilis, have provided critical insights into the stem cell system in demosponges. Here I introduce archeocytes and discuss (i) modes of archeocyte differentiation, (ii) our current model of the stem cell system in sponges composed of both archeocytes and choanocytes based on our molecular analysis and previous microscopic studies suggesting the maintenance of pluripotency in choanocytes, (iii) the inference that the Piwi and piRNA function in maintaining stem cells (which also give rise to gametes) may have already been achieved in the ancestral metazoan, and (iv) possible hypotheses about how the migrating stem cells arose in the urmetazoan (protometazoan) and about the evolutionary origin of germline cells in the urbilaterian (protobilaterian).
How stem cells evolved and how “the stem cell system” (the regulatory mechanisms to control their differentiation and self-renewal) developed during evolution are two of the fundamental questions of developmental biology. The porifera are the oldest living metazoans, and thus they can yield insights into the acquisition of the primordial stem cells in the last common ancestor of metazoans (the urmetazoan), the evolution of the stem cells, and their regulatory system. It is well known that sponges have remarkable reconstitutive and regenerative abilities (Wilson 1907; reviewed in Simpson 1984). Archeocytes, thought to be stem cells in sponges, are essential for the reconstitution of dissociated sponge cells (De Sutter & Buscema 1977; De Sutter & Van De Vyver 1977). After dissociation, some populations of the dissociated cells die, and some are phagocytosed by archeocytes, while others migrate and adhere to each other to form cell aggregates. Certain-sized aggregates attach to the substratum and eventually reconstruct a sponge body. When the archeocyte-rich fraction was eliminated from dissociated cells, the remaining cells formed aggregates but these aggregates did not adhere to the substratum and eventually underwent cytolysis. In contrast, cells of the archeocyte-rich fraction aggregated and formed functional sponges (De Sutter & Van De Vyver 1977, Funayama et al. unpubl. data). These results revealed that the function of archeocytes producing the requisite types of cells was indispensable for Wilson’s reconstitution experiments. In other words, sponges possess a well-developed system to regulate stem cell differentiation and probably also self-renewal.
Based mostly on histological studies, it is generally agreed that the archeocytes are likely to be pluripotent stem cells in sponges, and that other types of cells differentiate from archeocytes (Borojevic 1966 reviewed in Borojevic 1970; Simpson 1984; Müller 2006; Funayama 2008). The presumably pluripotent archeocytes to support both sexual and asexual reproduction systems in sponges. In sexual reproduction, gametes are derived from archeocytes (which mostly give rise to oocytes) or choanocytes (food-entrapping cells; they mostly give rise to sperm) (about oocytes: Saller 1988; about sperm: Tuzet et al. 1970a; Gaino et al. 1984; Paulus & Weissenfels 1986; Saller 1988; reviewed in Simpson 1984; Leys & Ereskovsky 2006). In one of the sponge asexual reproduction systems, gemmule hatching, thousands of resting archeocytes (which are all packed inside the gemmule coat) are activated and then proliferate and differentiate to form functional juvenile sponges as described below (Höhr 1977; Tanaka & Watanabe 1984; reviewed in Simpson 1984). However, thus far, archeocytes have only been defined based essentially on their morphological features (reviewed in Simpson 1984).
Most of the key studies of the stem cell system in sponges thus far have been microscopic studies, but in the past few years, another big step forward has been made by achieving molecular studies of this system in sponges. Several genes have been suggested to be expressed during the process of archeocyte differentiation, based mostly on molecular studies using two systems. One system is the system using cell aggregates (termed primmorphs) of the demosponge Suberites domuncula (Custodio et al. 1998; Reviewed in Müller 2006). As shown in Wilson’s study, dissociated cells of sponges gather together and form cell aggregates and then eventually the aggregates attach to the substratum, and sponge bodies are then reconstituted from them (Wilson 1907; De Sutter & Van De Vyver 1977; reviewed in Simpson 1984). The primmorph is a compact, spherical cell aggregate formed from dissociated cells. The primmorph is thought to be constituted mainly of proliferating cells, presumably archeocytes, and can be maintained for several months without further attachment to a substratum, leading to morphogenesis to form functional sponges (Custodio et al. 1998; Reviewed in Müller 2006).
The other system is gemmule hatching: an asexual reproduction system in which functional juvenile sponges are formed from thousands of archeocytes originally packed inside the gemmule coats in various sponges, including the freshwater sponge E. fluviatilis (Höhr 1977; Tanaka & Watanabe 1984; Funayama et al. 2005a; Funayama 2008; reviewed in Simpson 1984; ). Gemmules are small particles (about 0.5 mm in diameter in E. fluviatilis) formed within the sponge tissue (Fig. 1Aa). As gemmules are resistant to the unfavorable conditions, the hatching from the gemmule is thought to be a system for surviving the severe environmental conditions. Hatching is usually suppressed by unidentified inhibitors (molecular weight less than 1 kDa, inactivated at 100°C for 10 min, resistant to proteinase treatment, Rasmont 1975) from the surrounding sponge tissue. When sponges die and these inhibitors are subsequently eliminated around the gemmules, gemmule hatching starts (Fig. 1Ab,c). Functional juvenile sponges are formed within about 7 days (about 1.2–1.5 mm in diameter in E. fluviatilis in our culture conditions, Figs 1Ad, 1B). Since all types of cells constructing the sponge body are derived from archeocytes that were originally packed inside the gemmule coat, we are focusing on the gemmule hatching of E. fluviatilis as a unique and suitable system for characterizing the stem cell system in sponges (Funayama et al. 2005a,b; Funayama 2008; Mohri et al. 2008; Funayama et al. 2010 in press).
In this review, archeocytes (active stem cells in sponges) and our recently proposed stem cell system in sponges involving not only archeocytes but also choanocytes are introduced. Furthermore, insights into the origin of stem cells in the urmetazoan and the evolution of the stem cell system suggested by a comparison of stem cell systems among organisms near the evolutionary base of metazoans are discussed. The significance of the expression of Piwi in multi-/pluri-potent stem cells that can give rise to gametes in the organisms near the evolutionary base of metazoans is also discussed.
Cell types constructing the sponge body
Porifera is a large phylum that includes many species with diverse morphologies and consists of the classes Demospongiae, Calcarea, and Hexactinellida. In this review, I focus on studies using cellular sponges, especially demosponges, rather than hexactinellids, which consist largely of a single syncytial tissue (for a review of Hexactinellida, see Leys et al. 2007). More than 10 types of cells have been reported to make up the sponge body. Most of these cells have been defined only according to their morphological features (reviewed in Simpson 1984; About the cellular organization of E. fluviatilis: Weissenfels 1975). The body of a cellular sponge can be roughly divided into three parts (Fig. 1C): (i) outer epithelial cells that cover the sponge body, (ii) inner epithelial cells constructing canal systems, and mesohyl, (iii) the space between the outer and inner epithelial cells. The mesohyl includes many types of cells, collagen fibrils, and spicules. The outer epithelial cells (that are not attached to the substratum) are known as exopinacocytes (Fig. 1C: Exp), and the inner epithelial cells as endopinacocytes (Fig. 1C: Enp). Basal epithelial cells (that are attached to the substratum) are known as basopinacocytes (Fig. 1C: Bp). The stem cells (which are migrating in the mesohyl) are known as archeocytes (Fig. 1C: A). Each choanocyte has a collar and a single flagellum (Fig. 1C: Ch). Choanocytes form a single-layered, nearly spherical epithelial sac, known as a choanocyte chamber (Fig. 1B, C: ChC). Choanocyte chambers are connected with the canals (Fig. 1C: Ca), and by moving their flagella, choanocytes create water currents that help them to take up nutrients from the outside. Sclerocytes are the spicule-making cells (Fig. 1C: Sc). Additionally, spongocytes, collencytes and lophocytes (Fig. 1C: Spo, Co, L, respectively) have been reported to be cells that are active in collagen biosynthesis. Resting state archeocytes that constitute the gemmule (small particles formed within the sponge body for asexual reproduction, Fig. 1a,b) are specially designated as “thesocytes”. In the process of gemmule formation, accumulated archeocytes finally undergo nuclear division, and enter a resting state to become thesocytes. They are structurally similar to archeocytes except that they lack phagosomes (Langenbruch 1981; reviewed in Simpson 1984). They contain numerous vitelline platelets (in freshwater sponges) or storage granules (in marine sponges) (Langenbruch 1981; Simpson & Fell 1974; reviewed in Simpson 1984). In spongillid (e.g. E. fluviatilis) gemmules, thesocytes have two nuclei (Langenbruch 1981). There are additional types of cells that have been reported in many but not all sponge species (e.g. myocytes as contractile cells, glycocytes as cells involved in the storage of glycogen, and spherulous cells with many inclusions) (reviewed in Simpson 1984; Boury-Esnault 2006).
The criteria for judging that cells are multi-/pluri-potent stem cells is having the ability (i) to proliferate (including self-renewal), and (ii) to differentiate into at least two types of cells. It is generally accepted that the archeocytes are likely to be the pluripotent stem cells in sponges, and other types of cells differentiate from archeocytes, based mostly on histological studies (Borojevic 1966 and 1970; Brien 1976; reviewed in Simpson 1984; Müller 2006; Funayama 2008). The recent finding that Piwi homolog mRNAs are expressed specifically in archeocytes and choanocytes in E. fluviatilis, and the precise analyses of the expression patterns of these genes (EfPiwiA and EfPiwiB paved the way for analyzing whether these cells meet the criteria described above for stem cells. The results of those analyses strongly suggested that they did meet these criteria (see details in following section “Two different modes of archeocyte differentiation”). In this section, I will introduce the cell morphological features of archeocytes and then the background knowledge about archeocytes based on rather classical observational studies.
Harrison described the archeocytes as morphologically similar to the unspecialized stem cells in other animals (i.e. they have a nucleus with a single nucleolus) (Harrison 1974). Archeocytes are defined as large amoeboid cells that have a nucleus with a single large nucleolus (Fig. 2). Archeocytes actively migrate within the mesohyl (Fig. 1C) and are highly proliferative, phagocytic cells (Rozenfeld & Rasmont 1976; Garrone & Rozenfeld 1981; reviewed in Simpson 1984).
In one of the asexual reproduction systems (Fig. 1), small particles known as gemmules are formed in the sponge tissue. Thousands of thesocytes (resting archeocytes) are packed inside the collagenous gemmule coat. In E. fluviatilis, at the initial stage of the gemmule hatching process, binucleated thesocytes exit their dormant state and undergo cytokinesis, and then become uninucleated archeocytes before they migrate out from the gemmule coat (Brien 1932; Wintermann 1951; Höhr 1977). Since thesocytes contain numerous vitelline platelets as a source of nutrition, archeocytes in early developmental stages of gemmule hatching contain many vitelline platelets (Fig. 2Aa). In order to form a functional juvenile, archeocytes undergo proliferation and differentiation (reviewed in Simpson 1984). Both in gemmule-hatched juveniles of the freshwater sponge E. fluviatilis and in the tips of the branches of the adult freshwater sponge Spongilla lacustris, it was reported that archeocytes migrate into the new growth area and then (may proliferate and) differentiate (Brien 1976). Their differentiation into specialized types of cells, including choanocytes and sclerocytes (two cell types that can be easily distinguished by their morphology) was reported (Brien 1976). Sponges routinely undergo remodeling of the tissue (i) to refine their canal systems, (ii) to reconstruct the tissues after degeneration of tissue during sexual reproduction, and (iii) to functionally connect newly grown areas with the original canal systems. In all cases, archeocytes undergo both phagocytosis of other cells and differentiation into other types of cells (Diaz 1979; reviewed in Simpson 1984).
Although archeocytes are the cells most extensively documented as likely pluripotent stem cells in sponges, the possible pluripotency of choanocytes has also been indicated by several facts shown by observational studies. In this section, a) cell morphological features of choanocytes, and observational studies suggesting the abilities of choanocytes b) to give raise gametes and c) to transform into archeocytes are introduced.
Cell morphological features of choanocytes
Choanocytes form chambers that are one of the prominent structures of the sponge body (Fig. 1B, 3A–C). Choanocytes face inward to form the choanocyte chambers (Fig. 3B–E). As described in the section “Cell types constructing the sponge body”, choanocytes forming chambers along the canals create the water currents, take up nutrients from outside the sponges, and then pass these nutrients to other cells via vesicles.
It should be noted that the morphology of choanocytes is strikingly similar to that of choanoflagellates, the protists that are situated closest to metazoans in the molecular phylogenetic tree (King & Carroll 2001; King 2004; King et al. 2008; Suga et al. 2008; Fig. 3D–F). Both sponge choanocytes (Fig. 3D, E) and choanoflagellates (Fig. 3F) have a flagellum and microvilli surrounding the flagellum like a collar (Fig. 3G). The microvilli of sponge choanocytes are laterally connected to each other, to form the cylindrical collars of the choanocytes (Fig. 3D, E, G). Furthermore, choanocytes possess the fundamental features of unicellular organisms, namely, the abilities to proliferate and to take up nutrients. Immature choanocytes proliferate within the cell cluster of immature choanocytes during the process of chamber formation (Weissenfels 1981; Tanaka & Watanabe 1984; Funayama et al. 2005a; Funayama et al. unpubl. data). Also, choanocytes proliferate to maintain the choanocyte chambers (reviewed in Simpson 1984).
The ability of choanocytes to give rise to gametes
Precise electron microscopic studies revealed that choanocytes forming chambers directly transformed to spermatogonia in situ (Tuzet et al. 1970a,b; Diaz & Connes 1980; reviewed in Simpson 1984). In this process, choanocytes lose their collars and the phagosomes at the base of the cells, and migrate into the lumen of the chambers. The flagellum is maintained during the process of spermatogenesis. Spermatogonia undergo meiosis to produce sperm. Epithelial cells gather around the cells undergoing spermatogenesis to form spermatogenic cysts. In sum, choanocyte chambers convert to spermatogenic cysts. The fact that choanocytes can transform to spermatogonia suggests that choanocytes have pluripotency, at least in the sexual reproduction process.
The possible ability of choanocytes to transform into archeocytes
Although choanocytes are often thought of as fully differentiated cells that have specialized function and morphology, there have been some conflicting opinions concerning the potential of choanocytes to differentiate (reviewed in Simpson 1984). In contrast to one well-known view of choanocytes as terminally differentiated cells, Connes et al. (1974) concluded that choanocytes can act as a stock of archeocytes and are also able to form gametes. There are several reports that choanocytes can transdifferentiate into archeocytes during gemmule formation (Connes et al. 1974) and upon disorganization of sponge tissue and subsequent remodeling (Diaz 1979).
Our current working model of the stem cell system in sponges, composed of archeocytes and choanocytes
Two different modes of archeocyte differentiation
Our recent molecular biological analysis using the gemmule hatching process of E. fluviatilis confirmed that archeocytes are stem cells by demonstrating their ability to proliferate (using BrdU pulse-incorporation experiments) and to differentiate into at least three types of cells (using in situ hybridization analysis). Furthermore, we proposed a model that the sponge stem cell system is composed of archeocytes and choanocytes (Funayama et al. 2010 in press).
EfPiwiA/B-expressing cells with archeocyte morphology were strongly suggested to be at least multi-potent stem cells by their ability to proliferate and differentiate into at least three types of cells: sclerocytes, EfLectin-expressing cells that presumably act in innate immunity, and choanocytes, based on precise analysis of the expression patterns of EfPiwiA/B and cell-lineage-specific molecular markers (EfAnnexin for the choanocyte lineage, EfSilicateinM1 for the sclerocyte-lineage and EfLectin for a lineage of cells that act in innate immunity against bacteria). The co-expression of EfPiwiA with particular cell-type-specific lineage marker genes in presumptive committed archeocytes suggests the ability of archeocytes to differentiate into these three types of cells (Funayama et al. 2010 in press, Fig. 4). Although the self-renewal of EfPiwiA-expressing archeocytes has not been directly demonstrated, it was strongly suggested by the facts that EfPiwiA-expressing archeocytes can proliferate (as indicated by BrdU incorporation, Funayama et al. 2010 in press) and that the number of EfPiwiA-expressing archeocytes did not markedly decrease during the process of hatching from the gemmule. Moreover, The video “Life of the Freshwater Sponge” has recorded the image, where cells of archeocyte morphology produced daughter cells with the same morphology, indicating the self-renewal of archeocyte during gemmule hatching of E. fluviatilis (Tokyocinema, Co., Ltd 1996 edited by Dr. Yoko Watanabe). In addition, our studies of EfPiwiA/B expression clearly revealed that there are two different modes of archeocyte differentiation: In the first mode, gene expression transition from EfPiwiA/B to cell-lineage-specific genes seemed to be a general occurrence during the process of archeocyte differentiation (Fig. 4A–C). The other, exceptional, mode could be seen during the process of archeocyte differentiation into choanocytes, in which EfPiwiA/B expression was maintained even in the mature-sized choanocyte chambers of fully functional filter-feeding sponges (Fig. 4D, E). Considering the general function of Piwi family proteins and earlier observational studies suggesting the ability of choanocytes to transform into archeocytes in specific circumstances (Connes et al. 1974; Diaz 1979), the maintenance of EfPiwiA/B expression in mature choanocytes might indicate the retention of the potential pluripotency of choanocytes (Funayama et al. 2010 in press). These data suggest that choanocytes maintain pluripotency in spite of having specialized functions and specialized cell morphology. We speculate that EfPiwiA/B acts as one of the molecular bases for the retention of pluripotency in both archeocytes and choanocytes.
Our current working model of the stem cell system in demosponges
Based on all the findings described above, we propose that the sponge stem cell system is composed of two types of potentially pluripotent cells: archeocytes and choanocytes (Fig. 5). Archeocytes are active stem cells that self-renew and differentiate into multiple types of cells, including choanocytes (Figs 4, 5). Under general conditions, choanocytes play the major role in the filter-feeding functions of sponges, as water-flow-creating and food-entrapping cells. Their expression of EfPiwiA/B suggests that choanocytes retain pluripotency, as opposed to re-gaining the pluripotency under specific circumstances. The pluripotency of choanocytes is manifested under specific conditions, in which they delaminate and then can be transformed into active stem cells (archeocytes), or undergo meiosis to produce gametes (mostly sperm). Archeocytes are also known to give rise to gametes (mostly oocytes). In the process of gemmule formation, accumulated archeocytes finally undergo nuclear division, and enter a resting state to become thesocytes.
Future studies required to assess/prove this possible model of the stem cell system in demosponges consisting of archeocytes and choanocytes
In model organisms, providing that cells meet the criteria for stem cells is performed principally by lineage-tracing and/or transplantation analyses. In sponges, gene introduction has not yet succeeded. Furthermore, the fact that most of the cells constructing the sponge body (except for epithelial cells) are highly mobile and two daughter cells derived from cell division of archeocytes immediately start moving after cytokinesis impedes the lineage-tracing of particular cells. To further investigate the validity of this hypothetical model of the stem cell system consisting of archeocytes and choanocytes, it will be necessary to overcome these problems. The establishment of the method(s) to introduce gene(s) into archeocytes or choanocytes, or to perform live labeling of these cells, combined with the time-lapse video recording will make it possible to test which cells meet the criteria for stem cells in sponges.
One would also need molecular analysis in addition to such lineage-tracing analysis to prove (i) the ability of archeocytes to undergo self-renewal, (ii) the transformation of choanocytes into archeocytes (suggested by previous observational studies) and (iii) the abilities of archeocytes and choanocytes to give rise to gametes (also suggested by previous histological studies). The identification of genes that are expressed in both archeocytes and choanocytes in addition to EfPiwiA/B, and also of genes that are specifically expressed only in archeocytes, will be necessary for such further analyses, especially for analyses of the self-renewal of archeocytes and the transformation of choanocytes. Furthermore, since there are no methods yet for gene functional analyses, the establishment of methods for introduction of genes and for gene knockdown (i.e. RNAi) will be a critical step forward in studies of the sponge biology.
Genes suggested to be involved in archeocyte differentiation or regulation
Extensive studies using primmorphs suggested that several genes are expressed during the process of archeocyte differentiation (although whether these genes are expressed in archeocytes or cells around archeocytes is still obscure; reviewed in Müller 2006): Noggin and MSCP during the process of archeocyte differentiation to sclerocytes (Müller et al. 2003; Schröder et al. 2004; respectively) induced by adding ferric ions to the culture medium for primmorphs, Iroquois during differentiation to epithelial cells (Perovic et al. 2003), and Myotrophin during differentiation to myocytes (contractile cells, Schröder et al. 2000). In adult sponges of S. domuncula, mRNA expression of EED (a component of the polycomb repressive complex) is detected in gemmules, and mRNA of a sponge-specific receptor tyrosine kinase (RTKvs) is highly expressed in embryos (Perovic-Ottstadt et al. 2004; Müller 2006, respectively).
In addition, several cell signaling genes or transcriptional factors, such as Wnt, and Tgf-beta (Adamska et al. 2007a), Hedgling (a gene encoding a hedge-domain-containing protein, Adamska et al. 2007b), Delta, and a gene that belongs to the bHLH family (AmqbHLH1) in embryos of the marine demosponge Amphimedon queenslandica; (Richards et al. 2008) were reported to be expressed and suggested to regulate differentiation of cells and morphogenesis. Further precise molecular analysis of the expression of these genes and genes that are expressed in archeocytes, functional analyses (with the establishment of methods for gene introduction, gene knockdown, and/or using drugs that affect the function of specific proteins) will reveal the stem cell regulatory mechanisms that operate in sponges to form the functional body.
Piwi and piRNA appear to have been involved in maintaining the somatic multi-/pluri-potent stem cells that have the ability to give rise to germ cells from the first appearance of metazoans
In organisms near the evolutionary base of metazoans (flatworms and hydra), several genes that have been reported to be specifically expressed in germ cells in higher organisms are expressed in both germ cells and somatic pluri/multi-potent stem cells. The Piwi gene is such a gene whose expression has been reported in multi-/pluri-potent somatic stem cells in cnidarians and flatworms. Furthermore, in planarians it was revealed that functions of Piwi and piRNAs (Piwi-interacting RNAs) are necessary for maintaining the pluripotent somatic stem cells (Mineta et al. 2003; Reddien et al. 2005; Rossi et al. 2006; Palakodeti et al. 2008; Friedlander et al. 2009; Hayashi et al. 2010; reviewed in Sanchez Alvarado et al. 2002; Sanchez Alvarado 2007; Shibata et al. 2010). Although the expression of Piwi in multipotent stem cells in hydra or nematostella has not been reported, there are several reports of its expression in jellyfish. Multipotent muscle cells and gonadal cells in medusa of Podocoryne carnea (jellyfish), and stem cells in Clytia hemisphaerica express Piwi (Seipel et al. 2004; Denker et al.2008; respectively). The expression of Piwi genes in these organisms near the evolutionary base of metazoans supports the notion that the evolutionary origin of germ cells was the somatic multi-/pluri-potent stem cells (Agata et al. 2006; Extavour 2007). Our recent findings of EfPiwiA/B expression in two types of potentially multi-/pluri-potent cells in sponges (archeocytes and choanocytes) (Funayama et al. 2010 in press) extends this generalization by indicating that Piwi is expressed in presumptive somatic cells that directly give rise to gametes in the most basal extant multicellular animals, sponges.
By what molecular mechanisms does EfPiwiA/B function to maintain the pluripotency of muti-/pluri-potent somatic stem cells in basal metazoans and flatworms? Recently, the presence of Piwi and piRNAs was reported in Nematostella vectensis (cnidaria) and Amphimedon queenslandica (marine demosponge) based on genome data (Grimson et al. 2008). Together with the facts described above that Piwi genes are commonly expressed in the multi-/pluri-potent somatic stem cells in basal metazoans and flatworms, and that Piwi and piRNAs (Piwi-interacting RNAs) are necessary for maintaining the pluripotent somatic stem cells in planarians, and these findings of piRNAs in basal metazoans suggest that the Piwi protein functions in maintaining the multi-/pluri-potent somatic stem cells in a piRNA-mediated manner in the organisms near the root of the evolutionary tree. It should be noted that in a ciliate, Tetrahymena thermophila, it was clearly demonstrated that Piwi and piRNA play an epigenetic role in protecting the somatic genome, the “macronucleus genome” (Mochizuki & Gorovsky 2004). Furthermore, the Piwi gene has often been reported to be missing in extant protists, including a choanoflagellate, Monosiga brevicollis, and even in a protozoan, Trichoplax adhaerens (Grimson et al. 2008). Altogether, these facts make it tempting to speculate that the urmetazoan already possessed a Piwi/piRNA system similar to that of Drosophila or vertebrates, in which Piwi/piRNA system used for maintaining the pluripotency of somatic stem cells that could give rise to germ cells. Later during the further evolution of metazoans, this system seems to have become restricted to use in maintaining germ cells.
Insights into the origin of stem cells and germ cells
Hypothesis about how migrating stem cells arose in evolution
In the above section “Our current working model of the stem cell system in sponges, composed of archeocytes and choanocytes”, the potential pluripotency and possible role of choanocytes as a “storage” state of archeocytes were discussed. One might next ask whether the original stem cell in the urmetazoan was more similar to choanocytes or archeocytes. The remarkable morphological similarities and several features shared by sponge choanocytes and the unicellular choanoflagellates (described in the “Cell morphological features of choanocytes” suggested to us a hypothetical model in which stem cells might have originated from flagellated cells at the surface of the body of the urmetazoan, similar to the hypothesis of Buss (Buss 1987; King 2004). Choanoflagellates have been suggested to be one of the evolutionarily closest unicellular organisms to multicellular organisms, based on phylogenetic studies (King & Carroll 2001; King et al. 2008; Suga et al. 2008; reviewed in King 2004). The facts that choanoflagellates are morphologically similar to choanocytes and that some specific species conglomerate have led some to suggest that the ancestral multicellular organism might have evolved from a conglomerate of unicellular flagellated organisms (reviewed in Buss 1987; King 2004). In this regard, it is noteworthy that choanocytes are not only morphologically similar to choanoflagellates, but also they have the ability to proliferate and to take up nutrients, like unicellular organisms (see also the above section “Choanocytes”). Furthermore, there are reports of the ability of choanocytes to transform to archeocytes when they delaminate and migrate into the mesohyl under specific circumstances (Connes et al. 1974; Diaz 1979).
The evolution of stem cells might have been related to the restriction of the mitotic ability to a subset of cells (reviewed in Buss 1987; King 2004). For example, one of the hypotheses about the evolution of stem cells is the following scenario: Sometime after the ancestral multicellular organism that was like a conglomerate of flagellated protists with all cells located in the surface had evolved, the cells came to divide into two types: proliferative cells and non-proliferative cells (with the proliferative cells giving rise to the non-proliferative cells). These can be referred to as uni-potent stem cells and non-stem cells. Multicellularity allowed the cells to share the labors of living and paved the way for the evolution of specialized cell types, including stem cells (ancestral multi/pluri-potent stem cells). When cells located inside the body evolved as the larger body size of the organism evolved, it would have become difficult to generate non-stem cells far from the stem cells located in the surface of the body. This might have led to the evolution of “migrating stem cells”, like the archeocytes of the sponge or somatic stem cells (neoblasts) in planarians: they could migrate to the interior of the body and reach places where they should differentiate to generate non-stem cells (reviewed in Sanchez Alvarado 2007Umesono & Agata 2009; Shibata et al. 2010). It is tempting to speculate that in sponges, choanocytes may represent the prototype of stem cells located in the surface of the body and archeocytes may represent a newer type of migrating stem cells. It will be interesting to clarify to what extent choanocytes and archeocytes share the molecular mechanisms for maintaining multi-/pluri-potency, not only in order to delineate the molecular mechanisms involved in these two types of cells constituting the stem cell system in demosponges but also to get further insights into how the stem cell system might have evolved in the urmetazoan.
Hypothesis about the origin of germline cells in the extant bilaterians
Further studies of the relationships among archeocytes, choanocytes, and gametes in sponges, together with the comparison of gene profiles of archeocytes, choanocytes and germline stem cells in cnidarians and flatworms, might yield insights into the evolutionary origin of germ cells. In organisms near the evolutionary base of metazoans (flatworms and hydra), several genes that have been reported to be specifically expressed in germ cells in higher organisms are expressed in both germ cells and somatic pluri/multi-potent stem cells. In addition to Piwi, these genes include Vasa, a gene encoding a DEAD-box RNA helicase (in planarians: Shibata et al. 1999; in hydra: Mochizuki et al. 2001), and Nanos, a CCHC zinc finger RNA-binding protein expressed in germline stem cells (in planarians: Sato et al. 2006, in hydra: Mochizuki et al. 2000). Based on these findings, the idea that the evolutionary origin of the germ cells in the extant bilaterians was the somatic pluri-/multi-potent stem cells in the urbilaterian and that the closest extant type of cells to the urbilaterian germline cells might be similar to the archeocytes in sponges has been proposed (Agata et al. 2006) and gained some acceptance (Extavour 2007). As the presence of Nanos and Vasa homologs in sponges has been shown by the isolation of partial sequences of these genes in E. fluviatilis (Mochizuki et al. 2000, 2001), examining whether Nanos or Vasa homologs are expressed in archeocytes and/or choanocytes will be the next important step to further understand the evolutionary origin of germ cells.
Furthermore, such analyses might yield clues for speculating about the involvement of Nanos and Vasa homologs in the specification of germline (stem) cells during evolution. For this, it will be interesting to clarify whether (i) Nanos or Vasa homologs are expressed in archeocytes and choanocytes even in the asexual season and up-regulated in the sexual season (as in the case of hydra described below) or (ii) these genes are expressed only in the sexual reproduction season in archeocytes and choanocytes to give rise to gametes. In hydra, homologs of Nanos (Cnnos1 and Cnnos2) are expressed at a low level in multi-potent somatic stem cells known as interstitial cells (I cells), but when the germline stem cells are differentiated from I cells, the expression of Cnnos1 is up-regulated in germline stem cells (Mochizuki et al. 2000), while homologs of Vasa (Cnvas1 and Cnvas2) are expressed in less strongly in I cells but strongly in germline cells (Mochizuki et al. 2001). These data suggest that up-regulation of these particular homologs is required to produce germline cells in cnidarians.
Conclusions and future prospects
Recent advances in human ES cell and iPS cell research are paving the way for regenerative medicine. The need to understand stem cells and their regulatory mechanisms is increasing not only for developmental biology, but also for the medical field, because of their therapeutic potential. Recent studies suggest that the stem cells of organisms near the evolutionary base of metazoans (planarians and hydra) and higher organisms share similar intrinsic and extrinsic molecular mechanisms for regulating the balance between self-renewal and differentiation (Reddien et al. 2007; Petersen & Reddien 2008; Molina et al. 2009; Khalturin et al. 2007; Takahashi et al. 2009: about hydra, reviewed in Bosch 2008, Bosch et al. 2010; about planarians, reviewed in Shibata et al. 2010). Sponges or cnidarians can reconstruct a functional body from dissociated cells, in addition to their regenerative abilities. This reconstructive ability has been lost in bilaterians. The stem cell systems in basal metazoans have a remarkable ability to sense what types of cells are needed and to regulate stem cell differentiation in the requisite places. Especially, gemmule hatching of sponges is a natural asexual reproduction system for constructing a functional juvenile sponge exclusively from the thousands of archeocytes derived from the gemmule. Generally, during the embryogenesis of metazoans, blastomeres differ from the early developmental stages based on differences of maternal factors, signals from the surrounding tissues or sequential cell-to-cell interactions between blastomeres. The gemmule hatching of sponges is completely different from such embryogenesis, as the body formation starts from a group of (at least morphologically) uniform archeocytes. In this process, archeocytes sense their positions and differentiate, some signaling centers may be induced to organize the body plan, and cell-to-cell interactions cause differentiation of particular types of cells or organogenesis (such as construction of canal systems, formation of choanocyte chambers, connection of choanocyte chambers and canal systems, etc.). This system is not only a suitable system for investigating the evolutionarily oldest extant regulatory mechanisms of stem cells, but can also provide clues about the fundamental mechanisms that will infrom regenerative medicine about how to construct functional tissues from a mass of ES or iPS cells.
Comparative studies of multi-/pluri-potent stem cells organisms near the evolutionary base of metazoans will illuminate the fundamental stem cell system and its evolution in metazoans. Furthermore, focusing on sponge choanocytes (which may represent the ancestral stem cells similar to choanoflagellates) will give us clues about two important aspects of the evolution of stem cells: (i) comparison of the expressed genes (transcriptome) of choanocytes and the genome data of choanoflagellates will be very important to get insights into the evolutionary acquisition of stem cells. In E. fluviatilis (estimated genome size 360 × 106 bp; Ishijima et al. 2008), a highly enriched choanocyte fraction can be isolated (Funayama et al. 2005a), and should be useful for future studies. And (ii), the regulatory mechanisms of the transition from choanocytes to archeocytes in sponges will be very informative for clarifying the evolution of internal stem cells (archeocyte-type) from the choanocyte-type ancestral stem cell (Figs 5, 6).
I would like to thank Dr. Kiyokazu Agata for his support and discussions, and Dr. Elizabeth Nakajima for discussion, intensive reading and correction of the English of this manuscript. I am grateful to Dr. Norito Shibata and Dr. Kurato Mohri for helpful discussions.
N. F. was supported in part by a Grant-in-Aid for Exploratory Research and a Sunbor Grant from the Suntory Institute for Bioorganic Research.