SEARCH

SEARCH BY CITATION

Keywords:

  • regeneration;
  • planarian;
  • intercalation;
  • blastema;
  • positional cue;
  • A-P axis;
  • D-V axis;
  • Hox;
  • noggin

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

How can a planarian regenerate its entire body from a small portion of its body? Neoblasts, the totipotent stem cells of planarian, are assumed to be able to produce all missing cell types. However, we do not know how the cell fate of these cells is controlled during regeneration. Our recent studies with molecular markers suggest that intercalary regeneration is the fundamental principle in planarian regeneration. Here, we introduce the intercalation induced by ectopic grafting along the anteroposterior (A-P), dorsoventral (D-V), and left–right (L-R) axes. Blastema formation is evoked by ectopic D-V interactions after wound closure. Intercalation between the blastema and stump induces rearrangement of the positional identities along the A-P axis. Consequently, totipotent stem cells change their differentiation patterns according to the newly rearranged positional identities along the A-P, D-V, and L-R axes. According to the classic view, the blastema is regarded as the place where undifferentiated cells accumulate and regenerative events occur. Here, we propose a new interpretation, i.e., that the blastema may work as a signaling center inducing intercalary regeneration. Also, the roles of molecules and genes involved in intercalary regeneration are discussed. Developmental Dynamics 226:308–316, 2003. © 2003 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

The concept of “intercalary regeneration” was established by a series of grafting experiments in urodele and cockroach legs (Iten and Bryant, 1975; French, 1976). In these experiments, it was demonstrated that when a distal fragment of the leg was grafted onto the proximal remaining stump, the missing parts were restored with normal size and proportions (Fig. 1A; French, 1976). However, when the distal fragment was jointed to a more distal level, the original pattern of the leg was not restored. Instead an intercalary structure with opposite polarity regenerated, resulting in the formation of a longer leg (Fig. 1B). When the legs were joined in D-V reversed orientation, two extra legs were formed between the host and graft boundary (Fig. 1C). From these observations, the authors concluded that each cell may have recognized its own position along the leg in a polar-coordinate manner and that these cells carried out intercalary restoration of positional gaps when the cells were faced against originally non-neighboring cells (Bryant and Iten, 1976). However, it was not until the discovery of homeobox genes that insight was gained into the molecular basis for this intercalary regeneration. After the discovery of genes involved in pattern formation of Drosophila, the “boundary model” was proposed as a revised version of the polar coordinate model (Campbell and Tomlinson, 1995). When two different signaling molecules, wg and dpp, face each other ectopically, they induce dll expression (distalization) at the boundary, which evokes intercalary regeneration in an imaginal disc.

thumbnail image

Figure 1. Intercalary regeneration of cockroach limbs after grafting. A: When the distal fragment at level 8 is joined to the proximal stump at level 3, the missing parts can be regenerated (4–7) and legs with normal size and structures are restored. B: When the distal fragment at level 3 is joined to the proximal stump at level 8, an intercalary structure with opposite polarity is regenerated (7–4), resulting in the formation of a longer leg. C: When the legs are joined with reversed dorsoventral orientation, two extra legs are formed at the joint boundary.

Download figure to PowerPoint

Planarians are particularly well known for their high level of regeneration activity. Planarian totipotent stem cells may be present throughout in the mesenchymal space, from head to tail, and give rise to all cell types. However, we do not know how the mechanism by which the cell fate of these cells is controlled during regeneration remains unknown. Blastemas are always formed in the process of planarian regeneration, and lost tissues and organs appear in these regions, which suggests that the blastema seems to be in the site where stem cells may be committed. Thomas H. Morgan found that a new pharynx is formed in the old tissue of regenerants generated from posterior pieces, although a new head appears in the anterior blastema of all the pieces (Morgan, 1898). These observations suggest that pharynx-forming cells may be committed in the old stump. Kido observed a stream of new tissue from the anterior blastema to the old stump order to form the process of a new pharynx formation. He proposed that pharynx-forming cells might be committed in the anterior blastema and then migrate to positions in the stump (Kido, 1961). To understand the control mechanism of the regulating stem cells during planarian regeneration, it is indispensable to know the precise location where the cell fates of these cells are determined. For this purpose we have isolated a variety of cell type-specific genes, developed RNA in situ hybridization methods to detect committed cells at an early stage of differentiation, and accumulated observations at the individual cell level about the processes of both normal regeneration and morphogenesis after ectopic grafting (Agata et al., 1989; Cebrià et al., 2002a, 2002b; Kato et al., 1999, 2001; Kobayashi et al., 1999a, 1999b; Koinuma et al., 2000; Ogawa et al., 1998, 2002a; Sakai et al., 2000; Tazaki et al., 1999; Umesono et al., 1997, 1999). These observations clearly indicated that pharynx-forming cells are committed in the mesenchymal space in the old stump, whereas brain-forming cells always appear in the anterior blastema, supporting Morgan's observation at the cellular level. Here we review these observations and provide the new insights into planarian regeneration.

Morgan described pharynx formation in the old stump of the posterior fragments, but did not indicate where pharynx-forming cells appeared. At that time he had no tools to show the regeneration process at cellular level. However, our recent studies using molecular probes clearly indicated regeneration processes at the cellular level using molecular probes. Here we would like to reconsider regeneration process of planarians and purpose a new model.

PLANARIANS DO INTERCALATE

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

What is the composition of the blastema in regenerating planarians? Is it simply a miniaturized missing part or a cluster of undifferentiated cells? The ability to trace cell lineage with molecular markers was advantageous to investigate fundamental questions. By isolating cell type-specific genes and staining planarians with these gene probes by whole-mount in situ RNA hybridization (Fig. 2A–C), we found that the planarian body is divided into at least four regions: head, prepharyngeal, pharynx, and tail (Fig. 2D). In the head, two eyes and an inverted U-shaped brain are formed. No mucus-producing cells or pharynx muscles are detected in the head (Agata et al., 1998). Mucus-producing cells are found in both the ventral and dorsal sides of the prepharyngeal region (Kato et al., 1999). The pharynx, a cluster of muscle cells expressing DjMHC-A, is located in the central portion of the body (Kobayashi et al., 1998). The tail is defined by the distribution of three strips of mucus-producing cells (Kato et al., 1999).

thumbnail image

Figure 2. Structure of the planarian body. The left-most panel shows the expression of cell type-specific genes. DjPC2 (A; Agata et al., 1998), DjMHC-A (B; Kobayashi et al., 1998), and PN8 (Kato et al., 1999) stain central nervous system, pharynx muscles, and mucus-producing cells, respectively. C: Double staining for DjPC2 and PN8 clearly indicates that mucus-producing cells are not distributed in the head region. D: A schematic drawing of the planarian body. The planarian body can be classified into 4 regions: head (H), prepharyngeal (pP), pharynx (P), and tail (T) regions. Neural cells are indicated in purple. Red shows pharynx muscle-forming cells. Mucus-producing cells are indicated in blue and light blue. Blue shows the mucus-producing cells in the prepharyngeal region.

Download figure to PowerPoint

We traced cellular events during regeneration with these probes after cutting planarians into four pieces (Agata et al., 1998; Kobayashi et al., 1999a; Tanaka et al., unpublished observations). Figure 3A shows the process of regeneration from the head fragment stained for both DjPC2 (a neural marker) and PN8 (a marker for mucus-producing cells). Nine days after amputation, the head fragment regenerated a complete body with normal proportions. Although we could clearly observe a blastema during regeneration as a cluster of the DjvlgA-positive cells (Fig. 3B), the prepharyngeal and pharynx-forming cells clearly appeared in the stump, not in the blastema. This observation demonstrates that stem cells are committed in the mesenchymal space of the stump. The pharynx-forming cells appeared more proximally to the blastema than did the prepharyngeal cells (Fig. 3C), suggesting that body regionality may be rearranged during regeneration. In the intermediate stage of regeneration, the original head containing the differentiated brain overlapped with the newly formed prepharyngeal region. At later stages of regeneration, committed cells migrate from the intercalation zone into the posterior blastema and finally overlapped regions segregated to restore original body proportion. In the case of regeneration from a tail piece, the brain primodium is formed in the anterior blastema, prepharyngeal cells appear near the border of blastema, and the old tissues in the stump region. The pharynx-forming cells appear in the central portion of the tail fragment (Fig. 4A, lower panel). We did not observe migration of the committed cells from the anterior blastema into the central portion of the stump. This indicates that the blastema cannot be the compact primordium of the missing part, and so we must reconsider the role of the blastema. The posterior blastema itself may become the tail region, and the anterior blastema itself may develop into the head region.

thumbnail image

Figure 3. Regeneration process from the head fragment. A: Whole-mount samples stained for DjPC2 and PN8. No signal was detected in the posterior blastema. B,C: Transverse sections of 5-day regenerates stained for DjvlgA (a stem cell marker; Shibata et al., 1999) and DjMHC-A (a pharynx marker). Newly prepharyngeal region- and pharynx-forming cells appear in the mesenchymal space of the stump, not in the blastema. The blastema region is surrounded by red circles in A. The blastema/stump boundary is indicated by red lines in B,C.

Download figure to PowerPoint

thumbnail image

Figure 4. Schematic drawings of the regeneration process and the results of grafting experiments. A: Schematic drawings of regenerating head and tail pieces. B: The result of a grafting experiment from anterior to posterior. C: The graft of a posterior fragment into the anterior region. Comparison between the regeneration process and intercalary events suggests that the anterior and posterior blastemas may have head and tail characters, respectively. H, head; pP, prepharyngeal; P, pharynx; T, tail.

Download figure to PowerPoint

Intercalation Along the A-P Axis

Based on these observations, we speculated that planarian regeneration occurs by means of intercalation. If the posterior blastema differentiates into the tail, intercalation between the blastema and old stump can induce rearrangement of the body regionality, resulting in the formation of a new prepharyngeal and pharynx region (Fig. 4A, upper panel). That is, the blastema may stimulate the rearrangement of body regionality by producing posterior signals. Conversely, in the case of regeneration from posterior fragments, the anterior blastema may produce anterior signals, inducing intercalary events between the blastema and old tail fragments, and the pharynx is to be formed in the central portion of the tail fragment (Fig. 4A, lower panel).

To demonstrate intercalation, we performed transplantation experiments along the A-P axis and investigated cellular events with molecular probes (Kobayashi et al., 1999b). Insertion of the head piece into the tail induced intercalary regional rearrangement on both sides of the graft (Fig. 4B). Regional rearrangement was also confirmed by examination of Hox gene expression after grafting (Kobayashi et al., 1999b). Similar intercalary events were also observed when a tail piece was grafted into the prepharyngeal region (Fig. 4C). Therefore, grafting experiments mimic normal regeneration processes after blastema formation, suggesting that the blastema induces intercalation during regeneration, while no blastema formation was observed in the case of grafting experiments. Of interest, in planarians, cell proliferation is observed in the region proximal to the blastema, not in the blastema per se (Baguñà, 1975; Newmark et al., 2000; Salvetti et al., 2000). Studies addressing the nature of the planarian stem cells are described in our previous review (Agata and Watanabe, 1999) and those of others (Newmark and Sánchez Alvarado, 2002; Saló and Baguñà, 2002).

Intercalation Along the D-V Axis

Intercalation also appears to be involved in D-V repatterning. When a small piece was grafted into its original position in the host, but with reversed D-V orientation, depigmented regions were formed in the boundary between the host and the graft on the both dorsal and ventral sides within 3 days (Kato et al., 1999). The depigmented regions then grew and formed cup-like projections (Fig. 5A). When these cup-like projections were allowed to further develop for 3 weeks, they produced ectopic structures with morphologies that would be expected in each site (Fig. 5B,C). Chimeric analysis clearly indicated that the cup-like projections were formed on the boundary between the host and donor tissues, i.e., between the dorsal and ventral sides (Kato et al., 1999). Molecular analyses also indicated that the dorsal and ventral structures were newly formed in the cup-like projections (Fig. 5A). These results suggest that the blastema might be formed to restore some missing structure between the dorsal and ventral sides. We concluded that ectopic D-V interaction may evoke blastema formation and that intercalary events along the D-V axis are induced in the blastema. The cup-like projections finally became tail- or head-like structures upon long-term cultivation of the hosts, suggesting that the blastema may have the character of the most distal parts of the planarians, namely, the anterior or posterior regions.

thumbnail image

Figure 5. Morphogenesis after dorsoventral (D-V) reverse grafting. A: A whole-mount view of the D-V reversed graft developed for 30 days, stained with DjIFb (a D-V boundary maker; Kato et al., 1999). D-V boundaries are formed in the cup-like projections. B,C: Head- and tail-like structures are formed in the head and tail regions after D-V reversed grafting, respectively. The brain is formed in the cup-shaped projections in the head region. The cup-shaped projection formed in the tail region strongly expresses Plox2 (a Hox gene strongly expressed in the tail region; Orii et al., 1999). D-V reversed graft developed the structure from the grafted position to the most distal tip.

Download figure to PowerPoint

Molecular Background for Intercalation

A widely used mechanism for pattern formation is based on positional information in which cells acquire positional identities. Hox genes are well known to be involved in A-P patterning in many animals and are considered to be key genes providing positional cues (Wolpert, 1994). The expression of Hox genes should be rearranged in the process of regeneration to repattern body structures. In planarians, Hox genes are expressed in the posterior region along the A-P axis, as in other animals (Orii et al., 1999). Initially, a Spanish group reported that all of the planarian Hox genes showed a novel expression pattern in which they were non-colinearly activated in both the anterior and posterior blastemas (Bayascus et al., 1997). However, it has since been found that that study produced artifactual images, and the conclusions about the expression pattern have been accordingly revised (Orii et al., 1999; Saló and Baguñà, 2002). The expression of Hox genes is rapidly rearranged along the A-P axis during the regeneration process. Strong activation of Hox genes is always observed in the posterior blastema of any portion of the fragments but not in the anterior blastema. Posterior blastemas may produce signal molecules that rearrange the Hox gene expression pattern. We have not yet identified any such signal molecules specifically produced in the posterior blastema. Hox genes may not be involved in anterior patterning because expression of these genes is not observed in the anterior blastema.

Recently, we have succeeded in identifying a signal molecule, DjNLG, whose expression is specifically induced by D-V interaction. The DjNLG protein has a similar structure to vertebrate NOGGIN molecules (Ogawa et al., 2002b). Expression of Djnlg is rapidly induced after wound closure in the stump, after which Djnlg-positive cells become restricted to the ventral side of the region proximal to the blastema. A planarian BMP homolog (DjBMP) is constantly expressed in the dorsal-most region of the entire body (Orii et al., 1998). We expect that the BMP signal may be involved in the induction of D-V intercalation.

Djnlg-positive cells are not eliminated by X-ray irradiation (Ogawa et al., 2002b), although stem cells are completely eliminated (Shibata et al., 1999; Ogawa et al., 2002a). We previously demonstrated by a combination of X-ray irradiation and grafting experiments that the dorsal and ventral positional cues inducing the cup-like projections are retained in X-ray-irradiated tissues, suggesting that differentiated cells may provide positional cues (Kato et al., 2001). Djnlg-positive cells are detected beneath the muscle layer where fixed parenchymal cells are observed (Hori, 1991). We speculate that stroma-like cells attached to the stem cells may provide positional cues. Hox genes are expressed in both X-ray–sensitive and –insensitive cells (Orii et al., 1999). They may play essential roles in providing positional cues to control stem cell fate.

SUMMARY OF PLANARIAN REGENERATION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

We have summarized the regeneration process of planarians in Figure 6. When the body is cut, the wound is rapidly closed by muscle contraction. As a consequence of wound closure, the dorsal part adheres to the ventral side, inducing Djnlg expression at the junction with the blastema. The blastema may then start to produce signal molecules to induce intercalation, and the expression pattern of Hox genes becomes rearranged. Intercalated structures are formed in the blastema/stump region along the D-V and A-P axes. The stem cells begin to differentiate into appropriate cell types according to their newly rearranged positions along the A-P and D-V axes. Images of intercalary regeneration are shown in Figure 7. When the planarian body was divided into nine pieces, all of the fragments could rearrange their positional characteristics along the A-P axis by intercalation between the blastema and stump.

thumbnail image

Figure 6. Summary of planarian regeneration process. Ectopic dorsoventral (D-V) interaction caused by wound closure can evoke blastema formation. Ectopic interaction between the dorsal and ventral regions after wound closure is confirmed by the induction of Djnlg expression (Ogawa et al., 2002). The blastema works as a signaling center inducing intercalary regeneration. The totipotent stem cells respond to positional cues reorganized by the intercalary molecular system.

Download figure to PowerPoint

thumbnail image

Figure 7. How can planarians regenerate entire bodies from any small fragment from any region? Image of intercalary regeneration from small pieces after amputation into nine pieces.

Download figure to PowerPoint

Intercalation Along the L-R Axis

How does the planarian organize its bilateral structures? We have not detected any morphologic asymmetry to date. To investigate the L-R axis formation, we grafted a small piece into an ectopic region along the L-R axis (Saito et al., this issue). These grafting experiments clearly indicated that the L-R axis of the planarian body is organized by mediolateral (M-L) intercalation (Saito et al., 2003), suggesting that the planarian body has true symmetry. Even when the planarian body is cut along the longitudinal axis, it can form a blastema and regenerate. M-L intercalation may be sufficient to establish the L-R axis in planarians.

Planarians are the simplest animals possessing three distinct body axes. All of these three axes are formed by the intercalation system. The entire body of chicken embryos at the late gastrula stage can be regenerated from the lateral blastoderm lacking Hensen's node and the primitive streak (Yuan and Schoenwolf, 1999), suggesting that early embryos of other animals may also have an intercalary system like that of planarians along the A-P, D-V, and M-L axes.

Morphallaxis vs. Epimorphosis

The phenomenon of regeneration commonly has been classified into two categories, namely morphallaxis and epimorphosis. In the epimorphosis model, all missing parts are produced in the newly formed blastema. In the morphallaxis model, body proportion is rearranged without overt blastema formation. Morgan described planarian regeneration as requiring the proliferation of new material preceding the development of the new part (epimorphosis) as well as remodeling of the stump tissue (morphallaxis) (Morgan, 1990, 1902). From the above discussion, it appears that we have reached a point where it is necessary to revise these categories. In classic regeneration studies, there were no observations at the cellular level. Our analysis of the regeneration process using molecular markers has demonstrated that stem cells and differentiated cells show different responses and behaviors during regeneration (Fig. 8). When we observe the planarian regeneration process with respect to differentiated cells, the regeneration shows the characteristics of epimorphosis, because the differentiated brain remains in the old stump for a long time even after formation of the blastema. However, when we observe regeneration with respect to stem cells, it clearly shows the characteristics of morphallaxis. The stem cells in the old stump changed differentiation patterns immediately after formation of the blastema. Thus, planarian regeneration cannot be classified into either of the two classic categories. The literature about the terms epimorphosis and morphallaxis is a potential source of confusion. This has recently been made worse by the unfortunate misinterpretation of these phenomena by Saló and Baguñà in their recent postulation of their “Morphaollaxis-Epimorphosis” theory (Saló & Baguñà, 2002). We propose that all of the regeneration phenomena should be described more precisely at the single cellular level, and that the fundamental principles should be reconsidered based on from the results of cellular level analyses.

thumbnail image

Figure 8. Different response of differentiated cells and stem cells during regeneration. The upper and lower panels show the regeneration process at the cellular level from the head and tail pieces, respectively. The left side shows the behavior of differentiated cells. Active apoptosis of differentiated cells has not been demonstrated so far. The right side shows the behavior of the stem cells. The differentiation patterns of the stem cells change according to the region of interest.

Download figure to PowerPoint

FUTURE PROSPECTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

To better understand planarian regeneration, it will be important to identify molecules involved in determining body polarity and regeneration. The morphogenetic gradient theory is the most familiar postulation to explain both polarity and patterning in biological systems (Morgan, 1905; Wolpert, 1994). Reaction–diffusion mechanisms of local self-activation and long-range inhibition have been postulated to generate gradients of diffusible morphogens (Meinhardt, 1978). It is possible to speculate that such a molecular system may be involved in inducing intercalary events when both anterior and posterior gradients are postulated. However, recent molecular studies to understand pattern formation suggested the existence of more complex molecular systems. For example, in the case of D-V patterning in Drosophila and Xenopus, BMP signaling is modulated by the formation of complexes with neutralizing factors such as chordin, noggin, and follistatin, which are produced in the opposite side of the body (Thomsen, 1997). In addition, BMP receptors also form complicated complexes with different isoforms including a natural dominant-negative receptors (BAMBI) (Onichtchouk et al., 1999). Gradient formation of the BMP signal may be controlled by the formation of such complicated complexes.

The recent discovery of the nou-darake (ndk) gene may provide important cues to improve the understanding of planarian regeneration at the molecular level (Cebrià et al., 2002c). The ndk gene product may act to capture prospective FGF molecules, inhibiting diffusion of FGF from head to trunk. When the function of the nou-darake gene is knocked down by RNAi, FGF may diffuse to the posterior portion, stimulating brain formation in the entire all regions of the body. In the classical view, it was speculated that a brain inhibitor molecule was secreted from the brain and then inhibited brain formation in the posterior region of the body. However, following the discovery of the ndk gene, we must now reconsider this simple activator/inhibitor model by discovery of nou-darake gene. Recently a variety of neutralizing factors acting against signal molecules have been identified. It is possible to think that such molecules, by interacting with signal molecules, may play an important role in the chain of intercalary events during planarian regeneration by interacting signal molecules. Identifying signal molecules and their interacting molecules is the one of the most important fields of study in the field of regeneration biology.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES

We thank Shigeru Kuratani and Douglas Sipp for critical reading of the manuscript. We also thank Tokindo S. Okada for encouraging planarian studies at the cellular level. K.A. received Special Coordination Funds for Promoting Science and Technology, a Grant-in-Aid for Creative Basic Research, and a Grant-in-Aid for Scientific research on Priority Areas.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PLANARIANS DO INTERCALATE
  5. SUMMARY OF PLANARIAN REGENERATION
  6. FUTURE PROSPECTS
  7. Acknowledgements
  8. REFERENCES