experimental studies of the regeneration of planaria maculata

• DOI: 10.1007/BF02161491
• Corpus ID: 5724633

Experimental studies of the regeneration of Planaria maculata

• Published in Roux's archives of… 9 August 2015
• Roux's archives of developmental biology

241 Citations

Wnt/notum spatial feedback inhibition controls neoblast differentiation to regulate reversible growth of the planarian brain.

• Highly Influenced

The expanding epigenetic landscape of non-model organisms

Fundamentals of planarian regeneration., characterising demixed polymer surfaces to stimulate mesenchymal stromal cell activity and influence tissue development during osteochrondral regeneration, regenerating the bra in : lessons from the red spotted, ethics of biohybrid robotic jellyfish modification and invertebrate research, the cellular and molecular basis for planarian regeneration, neoblast-enriched zinc finger protein fir1 triggers local proliferation during planarian regeneration, positional information specifies the site of organ regeneration and not tissue maintenance in planarians, pbx is required for pole and eye regeneration in planarians, 9 references, observations and experiments on regeneration in planarians, experimental studies on hydra, regeneration in allolobophora foetida, studien über das regulationsvermögen der organismen, ueber heteromorphose bei planarien, a study of metamerism, notes on regeneration and heteromorphosis of tubularian hydroids, ueber regenerationsvorgänge bei lumbriciden, untersuchungen zur physiologischen morphologie der thiere, related papers.

Showing 1 through 3 of 0 Related Papers

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• HHS Author Manuscripts

THE HISTORY AND ENDURING CONTRIBUTIONS OF PLANARIANS TO THE STUDY OF ANIMAL REGENERATION

Sarah a. elliott.

1 Stowers Institute for Medical Research; Kansas City, MO USA

2 Department of Neurobiology and Anatomy, University of Utah; Salt Lake City, UT USA

Alejandro Sánchez Alvarado

3 Howard Hughes Medical Institute

Having an almost unlimited capacity to regenerate tissues lost to age and injury, planarians have long fascinated naturalists. In the Western hemisphere alone, their documented history spans more than 200 years. Planarians were described in the early 19 th century as being “immortal under the edge of the knife,” and initial investigation of these remarkable animals was significantly influenced by studies of regeneration in other organisms and from the flourishing field of experimental embryology in the late 19 th and early 20 th centuries. This review strives to place the study of planarian regeneration into a broader historical context by focusing on the significance and evolution of knowledge in this field. It also synthesizes our current molecular understanding of the mechanisms of planarian regeneration uncovered since this animal’s relatively recent entrance into the molecular-genetic age.

PLANARIANS AND THEIR HISTORICAL CONTEXT

The study of regeneration has a rich, intertwined history with experimental embryology. In the 17th century, naturalists contemplated two ancient paradigms for thinking about embryology: preformationism versus epigenesis. Preformationism contended that animals were already formed in miniature at the time of conception and simply expanded in size over the course of development. In contrast, epigenesis stated that animals were built piece by piece during development, guided by some intrinsic information housed in the undifferentiated embryonic cells. The rediscovery of animal regeneration at the end of the 1600s cast serious doubt upon the validity of preformationism. As naturalists began gathering more insights into both regeneration and embryonic development, epigenesis eventually took center stage as one of the most important principles of biology. 1

The earliest known description of animal regeneration came from Aristotle around 350 B.C.E. Among other things, he described that the tails of lizards regenerate. 2 In 1686, Thévenot, Perrault, and Duverney revived this finding. 3 This rediscovery of regeneration created a wave of excitement, and 18 th century naturalists began experimenting on any animals they could find to determine if this was a common phenomenon across the tree of life. de Réaumur showed that arthropods could lose appendages such as limbs and subsequently regenerate them. 4 Trembley systematically demonstrated that hydra, a member of the cnidarian phylum, regenerates after transection. 5 Bonnet proved that annelid worms regenerate, 6 and Spallanzani described the regenerative abilities in a variety of invertebrates such as snails and vertebrates like salamanders and frog tadpoles. 7

The birth of the study of planarians is most frequently associated with Pallas, who encountered them while exploring the Ural mountains in the late 18 th century. There, he observed that these animals regenerate missing body parts after fissioning. 8 However, other early reports of planarians exist. Trembley described feeding pieces of planarians to hydra in his monograph in 1774. 5 Müller described a number of planarian species in 1773, but erroneously grouped them with the trematode genus Fasciola . 9 Woodcut prints of land planarians can be found in Japanese encyclopedias dating as far back as the 17 th century ( FIG 1 ), 10 , 11 while written descriptions of the animals long precede that. 12 , 13 In fact, the oldest known reference to planarians comes from the Chinese text Yu-Yang Tsa-Tsu written around 860 AD by T’uan. He describes the animal “T’u-K’u” (likely the land planarian Bipalium ), and hints at its regenerative abilities by saying that it can “easily separate into several pieces” when touched. 14 , 12 , 13

1A: Wood-print adapted from the Japanese encyclopedia Kinmô-Zui by Tekisai Nakamura, 1666. 10 Image depicts a striped land planarian (likely Bipalium ). The caption indicates that “it is very poisonous and similar to another soil insect (nematode) previously described.” (Translation assistance provided by Dr. Tamaki Suganuma, Nobuo Ueda, and Shigeki Watanabe.)

1B: Wood-print adapted from the illustrated Japanese encyclopedia Wakan Sansai-Zue by Ryōan Terajima, 1713. 11 Image depicts a striped land planarian (likely Bipalium ) in the left column. The vertical text is translated to read:

“‘Doko’ or ‘Toku’. The animal has the shape of a Japanese belt in general appearance and is without legs. It measures up to 12 to 15 cm in length; a large specimen attains about 30 cm. The body is flattish in shape as a leaf of leek. There are yellow and black folds on the dorsal surface. The animal has a head shaped like a Japanese forceps… If the animal is touched, fission may occur… ” 13

Over the course of the 19 th century, more than a dozen different European 15 , 16 , 17 , 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 and American 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 biologists—including Darwin himself—continued to study these animals and demonstrated that the robustness of regeneration was common across planarian species. Indeed, in Dalyell’s eloquent words, planarians appear be “almost immortal under the edge of the knife,” making them tantalizing animals for study. 16

Work continued on planarians for the rest of the 20 th century, as it did in other animal models of regeneration. 34 However, in all fields, progress towards a mechanistic understanding of regeneration was hampered by an incomplete understanding of basic cell biology and genetics, in addition to a lack of tools for experimentation. Only recently have significant advances in molecular biology, genetics, and sequencing technologies reignited interest in planarians and other regenerative organisms. Today we are poised to tease apart the molecular mechanisms of regeneration, and unlock the mysteries of this biological phenomenon that have fascinated so many for so long.

WHY STUDY REGENERATION IN PLANARIANS?

Planarians are masters of regeneration.

Today’s popular model organisms were selected for a simple reason: they are biological exaggerations. T.H. Morgan selected the fruit fly and Sydney Brenner, the nematode worm because their exaggerated reproductive biology made them ideal for performing forward genetic screens. Likewise, planarians are ideal for regeneration studies because they undergo amazing feats of restorative and physiological regeneration. 35

Planarians undergo restorative regeneration in response to almost any type of injury. An upregulation of cell proliferation forms a mass of unpigmented newly-differentiating cells, called a blastema. From this blastema emerges many of the tissues lost to injury, producing a fully restored worm in as little as 1–2 weeks ( FIG 3A–B , dpa: days post amputation ). 28 , 31 This restorative response involves rebuilding anatomy de novo —a process Morgan called “epimorphosis.” It also involves remodeling the pre-existing tissues and integrating them with the newly made anatomy so that the animal regains its proper proportions and restores function to its organs. Morgan termed this type of remodeling “morphallaxis” ( FIG 3C ). 35

3A: T.H. Morgan amputated an adult planarian (red dashed line) and observed it regenerate missing anatomy (“epimorphosis”) and re-establish proper body proportions (“morphallaxis”). (A dapted from Morgan, 1900 33 )

3B: A live intact planarian was amputated (white dashed line), and the regenerating tail fragment is shown at 1, 4, and 7 dpa. Scale bar 200 um. dpa: days post amputation .

3C: The cephalic ganglia (arrowheads), pharynx (yellow asterisk), and anterior gut branch (arrow) regenerate by 7 dpa. An intact planarian (left) was amputated (white dashed line) and regenerating tail fragments were stained at timepoints indicated (right) for nervous system, pharynx (green, anti-α-tubulin antibody) and gut ( Smed-porcn-1) . Scale bars 200 um. ( Adapted from Gurley, et al., 2010 131 )

3D: The A/P decision is made by 1 dpa, preceding tissue regeneration and anatomical remodeling. Regenerating tail fragments stained at timepoints indicated for marker of anterior cell identity ( Smed-sfrp-1) . Scale bars 200 um. hpa= hours post amputation ( Adapted from Gurley, et al., 2010 131 )

In addition to restorative regeneration, planarians display physiological regeneration, repairing anatomy as it naturally ages. In the absence of an injury, these animals constantly undergo impressive levels of cell proliferation to replace old tissues. Like many cnidarians, annelids, echinoderms, and ascidians, planarians can maintain physiological regeneration for decades without losing the ability to regenerate or developing cancer. 36 , 37 This not only makes planarians useful for asking questions about regeneration. It also makes these long-lived animals tantalizing subjects for aging research, which has so far included studies of the mechanisms of planarian telomere maintenance 38 and the function of genes that affect longevity in other organisms. 39 , 40 , 41

Planarians are sufficiently complex in anatomy and behavior

Planarians are protostomes and members of the Lophotrochozoan clade. They are triploblastic and thus have tissues derived from all three germ layers (ecto-, meso-, and endoderm). While they have simpler body plans than vertebrate model systems, they have long been recognized to have discrete organ systems and behaviors amenable for regeneration studies ( FIG 2A ).

2A: Two depictions of planarian anatomy adapted from Leuckart’s zoological wall chart series entitled “Vermes,” circa 1890. 235 (Image obtained from MBLWHOI Library, Rare Books Archive)

2B: A live planarian extruding its pharynx ( arrowhead ). Scale bar 200 um.

(All animals depicted in Figures 2– 7 are the asexual strain of Schmidtea mediterranea unless otherwise noted.)

2C : Overlay of gut (blue, Smed-porcn-1 ), neurons (yellow, Smed-PC-2 ), axons, and pharynx (magenta, anti-α-tubulin antibody). Scale bar 200 um.

2D : Left panel : Head of a live planarian. Photoreceptors are darkly pigmented. Right panels : A different specimen showing neurons of the cephalic ganglia (blue, Smed-PC-2 ), photoreceptors, and commissural visual axons (red, anti-arrestin antibody; a kind gift of Dr. Kiyokazu Agata). Scale bars 200 um.

2E: Tufts of ventral cilia (yellow, anti-acetylated-tubulin antibody) projecting from epithelial cells (nuclei: magenta, TOPRO-3) facilitate swimming. Image focused around opening to the pharynx cavity (M, mouth). Scale bar 50 um.

2F: Left panel : Protonephridia, which compose the excretory system ( Smed-innexin-10) . Scale bar 200 um. Right panel : Close up of tail tip of a different specimen. Confocal maximum projection of protonephridial system, including flame cells (blue, anti-α-tubulin antibody), proximal tubules (magenta, Smed-innexin-10 ), and distal tubules (green, Smed-CAVII-1) . Scale bar 50 um. (Images provided by Hanh Thi-Kim Vu.)

2G: Markers labeling distinct body regions. Left to right: anterior cells and distal tip of pharynx ( Smed-sfrp-1) , posterior cells ( Smed-wnt11-2) , body periphery ( Smed-wnt5), and midline ( Smed-slit-1) .

Planarians eat ( VIDEO 1 ) and defecate ( VIDEO 2 ) through a muscular feeding tube, or pharynx, which connects to the gut ( FIG 2B–C ). 42 , 43 , 44 , 45 A blind gut with one anterior branch and two posterior branches occupies much of the body cavity ( FIG 2C ). 46 , 47 , 48 The animals possess an organized nervous system composed of two anterior cephalic ganglia and two parallel nerve cords that run ventrally along the length of the body ( FIG 2C ). A pair of dorsal photoreceptors is connected to the nervous system by axons that make up the optic chiasm ( FIG 2D ). 49 , 50 They possess motile cilia on their ventral epithelium that enables them to glide across surfaces ( FIG 2E ). 51 , 52 , 53 , 54 , 55 Their body plan is peppered with protonephridia, organs that facilitate osmoregulation and may ultimately prove to be homologous rather than analogous to the vertebrate kidney ( FIG 2F ). 56 , 57 , 58 While much work has focused on regeneration in asexual planarians, sexual strains also exist as cross-fertilizing hermaphrodites, regenerating ovaries and testes after amputation or starvation. 59 , 60 , 61 , 62 , 63 Planarian tissues also have intricate domains of molecularly discrete cell populations, yielding a plethora of markers to assess wound responses and general organization of the body plan ( FIG 2G ). 64 Finally, these animals display complex behaviors including negative phototaxis, fissioning in response to stimuli like changes in population density, and even cannibalism ( VIDEO 3 ). 16 , 17 , 18 , 65 , 66 , 67 , 68 Thus, planarian anatomy and behavior provide a sufficiently complex palette for studying regeneration.

An expanding molecular toolkit

Planarian studies prior to the end of the 20 th century were plagued by a lack of cellular resolution. Investigators relied principally upon basic histology, electron microscopy, and the visualization of gross anatomy under a transmitted light microscope to assess the regenerative response. They had little means of distinguishing cellular identity or tracing the lineage and movement of cells over time. They also had few ways to perturb the animals, and frequently resorted to treating planarians with pharmacological agents or toxins which had unknown mechanistic effects. Ultimately, the lack of experimental tools hampered the understanding of planarian biology.

Within the last two decades, our ability to visualize planarian tissues has improved drastically. Unlike Morgan and his inability to visualize the cellular and molecular events underpinning regeneration ( FIG 3A–B , dpa: days post amputation ), we can now assess each step of the regenerative response with extensive panels of markers. We can detect changes in gene expression and protein function, yielding a much sharper picture of the unfolding morphological and cellular dynamics of regeneration. 69 , 70 , 71 For instance, it is possible to visualize the regeneration of endodermally and ectodermally derived organ systems like the gut and brain, respectively ( FIG 3C ). Cellular activities that are not necessarily associated with organogenesis can also be assessed, such as the reestablishment of anterior domain identities after amputation ( FIG 3D , hpa: hours post amputation ).

In addition, the genome of the species Schmidtea mediterranea has been sequenced, 72 to which EST, transcriptome, proteome, and small RNA datasets can be mapped. 73 , 74 , 75 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 , 84 , 85 Genome microarrays have been generated to identify genes important for regeneration. 86 , 87 , 88 , 89 , 58 High-throughput RNAi screens can be performed to characterize gene function. 90 , 91 , 92 Fluorescence Activated Cell Sorting (FACS) is used to identify and isolate discrete cell populations, which subsequently can be used for such purposes as single-cell gene profiling or functional transplantation studies. 93 , 94 , 95 , 96 , 97 , 98 , 99 All of these tools, coupled with the ability to perform lineage tracing using BrdU, have opened the door for rigorous study of the molecular mechanisms underlying planarian regeneration. 100

PATTERNING THE PLANARIAN BODY AXIS

Polarity is maintained during regeneration.

One of the earliest uses of the term “polarity” in reference to body plan regeneration can be found in the work of Allman to describe the tubularian’s propensity to regrow a head from anterior-facing wounds and not posterior ones. 101 The mechanisms that establish and maintain polarity are fundamental questions shared by the fields of regeneration and embryology. How do cells know they are different from other cells? How are these differences translated into proper specification of the body axes and subsequent organogenesis? How is polarity re-established in the face of unexpected perturbations—whether that perturbation is a blastomere ablation in an embryo or an appendage amputation in an adult?

During regeneration, adult planarians maintain the polarity of their body axes. A small piece of tissue removed from the flank of the animal conserves the original orientation of the anterior-posterior (A/P), dorsal-ventral (D/V), and medial-lateral (M/L) axes ( FIG 4A ). 28 Additional experiments from the earlier part of the 20 th century demonstrated that the juxtaposition of tissues from different regions of the animal—such as the transplantation of anterior tissue to posterior regions—triggers abnormal regeneration, including the formation of an ectopic body axis ( FIG 4B ). 102 , 103 These results suggest that planarian tissues possess some type of intrinsic positional and polarity information. During regeneration, this starting information must be read and interpreted correctly such that the proper structures are made in the right location. Determining how polarity is re-specified and maintained is critical for understanding the mechanisms of animal regeneration, and we are now discovering that some of the genetic toolkit used to establish polarity in an embryo may also play similar roles in maintaining polarity during regeneration.

4A: Randolph showed that a small piece of tissue amputated from the flank of the body (left, red box) maintains axial polarity during regeneration (right). ( Adapted from Randolph, 1897 28 )

4B: Transplanting tissue from the anterior region of one planarian to a posterior region of another (left) results in outgrowth of a new body axis (right). ( Adapted from Santos, 1929 102 )

4C: Thin transverse amputations (left, red dashed lines) cause heteromorphic regeneration, resulting in double-headed (top) or double-tailed (bottom) regenerates. (Adapted from Morgan, 1904 120 )

4D: RNAi strategy employed for Figures 4– 6 . Animals were 1) fed dsRNA to knockdown a gene of interest, 2) amputated, and 3) allowed to regenerate.

4E: Wnt/β-catenin signaling controls A/P polarity. Live images and fixed animals stained for the nervous system ( Smed-PC-2) , anterior cell identity ( Smed-sfrp-1), and posterior cell identities ( Smed-fz-4) . Controls regenerate normally. Smed- β -catenin-1(RNAi) causes a head to regenerate from posterior blastemas. Smed-APC-1(RNAi) causes a tail to regenerate from anterior blastemas. Scale bars 200 um. ( Live images provided by Dr. Kyle A. Gurley and Dr. Jochen C. Rink.)

Historical views of regeneration polarity

Over the centuries, investigators have proposed diverse mechanisms to explain the phenomenon of polarity. Some of these models had distinctly preformationist undertones. Bonnet hypothesized that “germs” exist in the earthworm that contain a fully-formed miniature head or tail. Upon amputation, fluid flow transports “head germs” anteriorly and “tail germs” posteriorly so that a head and tail sprout at the proper locations. 6 , 104 Weismann, a declared preformationist, extended his theory of the germ-plasm to explain regeneration. He proposed that preformed cells containing an “idioplasm” facilitate the reconstitution of the limbs of salamanders and newts. “Idioplasm” is organic material that predetermines the reconstitution of the limb, regardless of cues from the environment or the regenerating appendage. As cells divide, portions of the nuclear “idioplasm” are lost, and the division progeny are left with only enough “ids” to produce the next most distal cells in the limb. Thus, this pre-determined regeneration program ensures that distal structures are never regenerated before more proximal ones. 105

Other hypotheses regarding regeneration polarity were more grounded in the ideals of epigenesis, proposing that polarity came not from preformed germs or “ids,” but instead developed progressively out of some instructive cues intrinsic to the cells and tissues. Bardeen argued that the pre-existing anatomy of a planarian exerts mechanical forces that constrain the location in which new anatomy can physically fit. He also argued that the nervous system was key in dictating polarity. 65 , 106 Pflüger proposed that the chemical composition of the pre-existing tissue’s cut surface establishes polarity. Each tissue laid down during regeneration provides a chemical signal that instructs the fate of the next layer laid down on top of it, and regeneration thus proceeds in a proximal to distal direction. 107 Child thought that gradients of metabolic activity guide regeneration. He believed that anterior tissues have higher metabolic rates and, thus, display “physiological dominance” over more posterior tissues, establishing anterior-posterior polarity early on in regeneration. 108 Brøndsted, heavily influenced by Child, proposed that unknown effectors establish A/P polarity through a time-graded regeneration field. This field exposes “high points” in a planarian blastema where regeneration of the head occurs faster and more vigorously than in other regions, subsequently releasing factors that inhibit head formation in more posterior areas. 109 , 110 ( For additional examples of regeneration polarity theories, see: 4 , 111 , 112 , 113 , 114 , 115 )

While most of these hypotheses have been proven insufficient to fully explain regeneration polarity or the defects resulting from experimental manipulation, Morgan’s theory has best withstood the test of time. Based upon meticulously-documented regeneration experiments performed in a wide variety of animals, Morgan observed that “something in the piece itself determines that a head shall develop at the anterior cut surface and a tail at the posterior cut surface. This ‘something’ is what we call polarity.” 116 He hypothesized that polarity results from some type of physical and/or chemical gradient along the body axes. 117 , 118 , 119

Anterior-posterior polarity in planarians

Early attempts to better understand axial polarity in planarians centered around perturbing regeneration through surgical means. The abnormal, surgically-produced regenerates were referred to as heteromorphoses. Before the use of chemicals, irradiation, electric fields or RNAi, heteromorphoses provided key insights from which hypotheses could be made about the mechanisms underpinning regeneration. Heteromorphoses of the A/P axis were described early on in studies of planarian regeneration, since the head was an easily recognized structure. Most notably, Morgan observed that transverse amputations producing short cross-pieces frequently regenerated bipolar heads or tails ( FIG 4C ). 31 , 33 , 120 Coupled with similar regeneration defects from experiments on earthworms and tubularians, Morgan suggested that a regenerate might interpret a gradient of chemical or physical information along the body axis to maintain proper axial polarity. Very thin slices of tissue could have too shallow of a gradient to be deciphered, causing the production of a head from posterior wounds by default.

RNAi screens have uncovered phenotypes recapitulating Morgan’s double head and double tail defects, lending support to his gradient hypothesis ( FIG 4D ). The Wnt/β-catenin pathway, which is involved in many developmental processes across metazoa including establishing polarity along the primary axis, 121 , 122 is required for A/P polarity in planarians. 123 , 124 Knockdown of the pathway’s core transcription factor Smed-β-catenin-1 results in anteriorization of the body axis, causing a head to regenerate from a posterior wound instead of a tail ( FIG 4E ). This anteriorized phenotype is also produced by knockdown of upstream ligands Dj/Smed-wnt1 and Smed-wnt11-5, receptor-associated agonists Smed-dvl-1 and Smed-dvl-2 , and the transmembrane protein required for secretion of WNTs Smed-evi/wntless. In contrast, upregulation of β-CATENIN-1 activity by knockdown of such inhibitors as Smed-APC-1 , Smed-notum, Smed-axinA, and Smed-axinB elicits the opposite phenotype in which tails regenerate from anterior wounds ( FIG 4E ). 123 , 124 , 125 , 126 , 127 , 128 , 129 , 130 These results suggest that during regeneration, graded levels of β-CATENIN activity along the body plan regulate the anterior-versus-posterior fate choice. β-CATENIN activity must be sufficiently high in posterior blastemas to facilitate tail regeneration and sufficiently low in anterior blastemas to produce a head. Likewise, this signaling system must be acting during physiological regeneration, as knockdown of Smed-β-catenin-1 anteriorizes uninjured animals too. 123 , 124 , 125 Whether or not there is a posterior-to-anterior gradient of β-CATENIN nuclear localization is unknown. However, numerous posteriorly-expressed Wnt ligands and anteriorly-expressed Wnt inhibitors suggest that there may indeed be such an activity gradient. 123 , 124 , 126 , 131 , 129 , 127

The Hedgehog pathway, well characterized during the development of many animals, 132 is also important for A/P polarity in planarians. RNAi of pathway activators Dj/Smed-hh , Dj/ Smed-gli-1 , and Smed-smo decreases Hh signaling and results in loss of posterior regeneration. In contrast to this “tailless” phenotype, increased Hh signaling through RNAi of pathway inhibitors Dj/Smed-ptc and Dj/Smed-sufu causes defects in anterior regeneration. In these animals, a tail regenerates instead of a head at anterior wounds. Thus, high levels of Hh signaling are required to properly specify posterior tissues, while lower levels are required for specifying anterior tissues. Furthermore, the Hh pathway may act upstream of the Wnt/β-catenin pathway by modulating the expression of Dj/Smed-wnt1 , which likely signals through β-CATENIN to specify posterior fates. 126 , 133 , 51 , 134 , 128

Additional parallel or convergent pathways are known to participate in the establishment and maintenance of A/P polarity. Simultaneous knockdown of putative gap junctions Smed-innexin-5 , -12 , and -13 produces double heads. 135 RNAi of the LIM homeobox transcription factor Djislet causes a tailless phenotype. 129 Knockdown of the TALE class homeobox transcription factor Smed-prep causes cyclopic and headless phenotypes. 136 Graded membrane voltage, based at least in part on high intracellular calcium levels in anterior wounds, also plays a role in establishing A/P polarity in planarians. 137 , 138 , 139 , 140 , 141 At this point, however, it is unclear how all these collective signals are integrated to properly reestablish the A/P axis.

Dorsal-ventral polarity in planarians

As in other regenerative animals, an amputation in a planarian brings dorsal and ventral tissues into close contact at the wound site. 142 Grafting experiments in animals including newts, 143 , 144 arthropods, 145 and annelids 146 have suggested that signaling between tissues from different regions of the dorsoventral (D/V) axis—such as the interaction induced by wound closure—might be an early trigger for regeneration. Classical planarian experiments have also supported this idea. In particular, Santos grafted plugs of planarian tissue into a host in either normal D/V orientation or inverted orientation ( FIG 5A ). In the former case, the tissue healed and the animal appeared normal. In the latter case, a blastema formed at the interface between the graft and host tissues and large cup-shaped protrusions emerged at the graft site. In at least one case, Santos even observed an ectopic planarian developing from the graft with inverted D/V orientation to the host’s body axis ( FIG 5A ). This suggested that the graft not only retained its original D/V polarity after transplantation, but the juxtaposition of dorsal and ventral tissues somehow triggered the formation of a new body axis. 102 , 103

5A: Santos showed that flipping the D/V orientation of a tissue plug without altering the A/P orientation results in the outgrowth of an ectopic body axis. This ectopic growth has inverted D/V polarity compared to the main body’s D/V axis. Only the tail region is pictured in Santos’ sketch; the main body’s head is to the left, out of view. (Adapted from Santos, 1931 103 )

5B: BMP signaling controls D/V polarity. Control tail fragments form a blastema (bracket) and regenerate nerve cords localized ventrally only ( Smed-PC-2, red arrowheads; compare ventral vs. dorsal views). Smed-smad4(RNAi) causes a loss of blastema formation, regeneration of the cephalic ganglia (compare brackets), and growth of ectopic dorsal nerve cords (black arrowheads; compare ventral vs. dorsal views). Scale bars 200 um.

5C: BMP signaling is required for blastema formation and organization of the midline. Control animals form anterior and posterior blastemas (white arrowheads) and regenerate a midline ( Smed-slit-1) . Smed-bmp4(RNAi) causes midline indentations in the blastemas, dorsal ruffling (red arrow), and ectopic expression of a midline marker (black arrowheads). Smed-smad4(RNAi) causes a loss of blastema formation (white arrowheads), photoreceptor regeneration in old tissue (white arrows), and ectopic expression of a midline marker (black arrowheads, compare insets). Scale bars 200 um.

Seventy years after Santos’ initial observations, investigators are revisiting his experiments with modern tools. Histological analyses and expression studies show that Santos’ inverted transplants indeed maintain their original D/V polarity after grafting. In addition, the boundary between the host and graft tissues of inverted transplants ectopically expresses a body edge marker, while non-inverted control transplants do not. 147 While future studies are needed to determine whether an ectopic body axis is truly forming from these protrusions and how this is accomplished, these results do suggest that the closer positioning of dorsal and ventral tissues after an injury and wound healing might be an important aspect of the regenerative response of planarians, helping promote blastema formation and specification of a new body edge. 148

Recent RNAi screens have uncovered numerous genes important in establishing and maintaining D/V polarity in planarians. So far, all genes identified in these screens are components of the BMP pathway, a branch of the TGF-β signal transduction cascade, which has a conserved role in organizing the D/V axis in diverse metazoans. 122 Reduction of BMP signaling by RNAi of DjBMP/Smed-bmp4 , Smed-smad1 , Smed-smad4 , Smed-admp , and Smed-noggin-like-8 ventralizes animals during both restorative and physiological regeneration ( FIG 5B ). Collectively, these ventralized defects include a dorsal duplication of the brain and ventral nerve cords, ectopic dorsal expression of ventral markers, and growth of dorsal cilia that enable the animals to swim in an inverted fashion on their dorsal side. 92 , 149 , 150 , 151 , 152 , 153 Likewise, increasing BMP signaling through knockdown of putative inhibitors Smed-noggin-1 and Smed-noggin-2 causes the opposite dorsalized phenotype in which animals ectopically express dorsal markers on their ventral side. 152 , 153

In addition to identifying a molecular foothold for studying the regulation of the D/V axis during planarian regeneration, these defects hint that an interaction between dorsal and ventral tissues juxtaposed during wound closure may indeed be important for regeneration, as Santos proposed from his grafting experiments ( FIG 5A ). 102 , 103 All ventralized RNAi phenotypes examined thus far display reduced or absent blastemas ( FIG 5B–C ) and a loss of expression of body edge markers at the wound site. Perhaps critical signaling events between properly-specified dorsal and ventral tissues organizes or permits downstream events in regeneration. While such an interaction could explain the BMP pathway blastema phenotype and the formation of a second body axis observed by Santos, these spatially-introduced signaling events have yet to be confirmed.

Medial-lateral polarity in planarians

Randolph and Morgan both described a variety of perturbations in medial-lateral (M/L) regeneration. 28 , 31 , 33 , 59 One of the most striking experiments involved a simple midline incision in either the anterior or posterior region of the animal. This incision did not fully cut the animals in half, and the wounds were allowed to heal back together. While some animals simply healed, this midline incision triggered a duplication of the M/L axis in others. These animals became wider and sprouted ectopic pharynges and photoreceptors lateral to the pre-existing ones ( FIG 6A–B ). While this experiment has yet to be revisited with molecular markers, their results suggest that the M/L axis is tightly regulated during regeneration. Simple wounding may cause the animal to reassess the integrity of the M/L axis, and trigger cells to take on different positional identities as if the animal had been cut through and through.

6A: Randolph showed that a midline incision (red dashed line) that is allowed to heal together causes, with some frequency, a duplication of midline structures. (Adapted from Randolph, 1897 28 )

6B : Live images 14 days after the midline incision depicted in 6A. Although the tissue was allowed to heal back together, the head is duplicated at the site of the incision (white arrowhead). The tail is forked (white arrow), even though the posterior was never injured. Species Dugesia sanchezi . Scale bar 200 um.

6C: Smed-slit-1 maintains the M/L axis. Compared to control(RNAi), Smed-slit-1(RNAi) causes the cephalic ganglia, nerve cords (magenta, Smed-PC-2; blue, anti-α-tubulin antibody), photoreceptors (black arrowhead), and markers for the body periphery ( Smed-wnt5; compare insets) to collapse towards the midline. Scale bar 200 um. (Adapted from Gurley, et al., 2010 131 )

6D: Smed-wnt5 maintains the M/L axis. Compared to control(RNAi), Smed-wnt5(RNAi) causes the lateral expansion of the axon tracts, and the formation of an ectopic lateral pharynx (blue, anti-α-tubulin antibody; white arrowhead) flanked by gut branches (magenta, Smed-porcn-1) . Expression of a midline marker expands laterally ( Smed-slit-1; compare insets ) . Scale bar 200 um. (Adapted from Gurley, et al., 2010 131 )

RNAi screens have identified numerous genes important for regulating anatomical patterning with respect to the M/L axis in planarians. This includes the Slit/Netrin repulsion-attraction signaling system. Among other developmental processes, SLIT and NETRIN ligands cooperatively regulate the migration of axonal projections across the midline, with SLIT repulsing axons as NETRIN attracts them. 154 , 155 Similarly, Dj/Smed-slit-1 is required to maintain the planarian midline during restorative and physiological regeneration. RNAi knockdown causes a collapse of lateral tissues towards the midline, including the cephalic ganglia, nerve cords, photoreceptors, optic chiasm, and posterior gut branches ( FIG 6C ). 156 , 157 While it is curious that knockdown of the SLIT receptor, Dj/Smed-roboA , does not fully recapitulate these midline defects, it does cause aberrant crossing and fasciculation of the axons of the optic chiasm, in addition to a lateral displacement of the ganglia and reduction of the anterior commissure. This suggests a defect in M/L patterning. 158 , 157 In contrast to the slit-1(RNAi) phenotype, the most penetrant defects associated with knockdown of Netrin signaling via Smed-netR(RNAi) and Smed-netrin2(RNAi) are a lateral expansion of the cephalic ganglia, reduction in anterior commissure, and disorganization of the axons of the ventral nerve cords. 159 These contrasting defects in M/L patterning, coupled with simultaneous double RNAi for Dugesia japonica Slit-Netrin signaling components, 157 suggest a Slit-Netrin signaling synergy may help direct various events in M/L patterning.

The Wnt pathway is also involved in midline patterning in planarians. Smed-wnt11-2(RNAi) causes a failed extension of Smed-slit-1 cells at the tip of the regenerated tail, resulting in abnormal looping of the ventral nerve cords and inappropriate midline crossing of the posterior gut branches. 126 , 131 In contrast, knockdown of Smed-wnt5 yields a phenotype very similar to the defects observed by Randolph and Morgan after midline incisions: a lateral expansion of medial structures towards the body periphery, in addition to the growth of lateral, ectopic pharynges ( FIG 6D ). 126 , 131 , 55 In light of these phenotypes, the M/L axis must be as tightly regulated during planarian regeneration and homeostasis as the A/P and D/V axes are.

Finally, the TGF-β pathway is important not just for D/V axis organization, but also M/L organization. Smed-bmp4 (which is expressed along the dorsal midline of the adult planarian) is upregulated at the wound edge during lateral regeneration. RNAi of some components of this pathway ( Smed-bmp4, Smedolloid-1, and Smed-smad4 ) completely abolishes regeneration from amputations that bisect the animals down the midline. 151 RNAi of TGF-β pathway members that cause ventralization results in conspicuous midline indentations in anterior and posterior blastemas, suggesting that the midline does not regenerate properly ( FIG 5C ). 92 , 149 , 150 , 151 , 152 , 153 Furthermore, collapse of the nervous system and regeneration of supernumerary photoreceptors at the midline, as well as lateral ectopic pharynges are among other midline abnormalities reported thus far. 150 The perturbation of the M/L axis was molecularly verified in Smed-admp(RNAi) animals, which ectopically express the midline marker Smed-slit-1 at the body periphery, 152 as well as in Smed-smad4(RNAi) regenerates ( FIG 5C ). (See TABLE 1 for summary of all polarity phenotypes described.)

Summary of body axis patterning phenotypes described in text. Data is summarized for A/P, D/V, and M/L patterning defects. Dorsalization and ventralization refers to ectopic expression of axis markers or regeneration of ectopic anatomy. Midline defects refer to any ectopic or missing anatomy at the midline. These include misguidance of the visual axons that cross the midline; expansion or collapse of midline structures like the brain, nerve cords, photoreceptors, or pharynx; loss of neural connectivity at the midline; ectopic expression of midline or body periphery markers; reduction in lateral regeneration, or midline indentations in the blastema. (See text for references and additional details for each phenotype.)

These results leave us with many questions. Does regeneration require establishment of one body axis before another can be specified? Or can an axis be specified independently of the other two, as in the case of zebrafish development? 160 Which cells provide polarity information and which cells interpret these cues? Are there organizing centers for body axis polarity analogous to those identified during embryogenesis? And how are all three axes integrated during regeneration? In order to understand how these animals regenerate in three dimensions, studies must focus on the timing of axis specification and the manner in which these axis decisions affect signaling cascades required for subsequent organogenesis. 161 , 63 , 48 , 162 , 57 , 58

NEOBLASTS: CELLULAR AGENTS OF PLANARIAN REGENERATION

A century-long debate regarding the cellular agents of regeneration.

After the amazing regenerative abilities of planarians were discovered, the search for the cellular source of this phenomenon ensued. Surprisingly, many key insights predate the use of the molecular-genetic tools recently applied to study planarians.

The histological analysis of planarian tissues under the light microscope in the late 19 th century allowed biologists to identify a subset of parenchymal cells that undergo cell division, as evidenced by the presence of mitotic figures. This proliferating population of 6–12 micron ovoid-shaped cells possesses large decondensed nuclei and scant, basophilic cytoplasm ( FIG 7A ). These cells were correctly identified as the main source of new tissues. 23 , 24 , 25 , 30 Over the years, many names were ascribed to these cells, including verästelten bindegewebszellen (branching connective tissue cells), 163 bildungszellen (forming cells), 23 stammzellen (stem cells), 24 stoffträger (support material), 164 ersatzzellen (replacement cells), 30 cellules libres du parenchyme (free cells of the parenchyma), 165 regenerationszellen (regeneration cells), 166 and wanderzellen (migratory cells). 167 Eventually, the term neoblast permanently designated these cells, a name applied by Randolph to describe the cells responsible for regeneration in the earthworm Lumbriculus . 168 , 169 , 170

7A: Classical depiction of neoblasts by histology. (Adapted from Stéphan-Dubois, 1965 236 )

7B: Neoblasts (green, Smed-piwi-1 ) distributed throughout the parenchyma in between gut branches (blue, Smed-porcn-1 ) and in proximity to the nerve cords (magenta, anti-α-tubulin antibody). Images are confocal maximum projections. Scale bars 50 um.

7C: Irradiation disrupts physiological regeneration. Representative intact planarians at specified dpi, exposed to 10,000 rads from a cesium source. Head regression is observed by 10 dpi, followed by ventral curling around 20 dpi. Lysis generally occurs after 20 dpi. Scale bar 200 um. dpi: days post irradiation

7D: Irradiation eliminates neoblasts and proliferation. Left panels : Neoblasts (green, Smed-piwi-1) are the only mitotic cells (magenta, anti-H3P antibody). White arrowheads indicate examples of colocalization in the tail of a different animal. Right panels : Neoblasts and mitotic cells are eliminated after irradiation by 3dpi. Images are confocal maximum projections. Scale bars 200 um. (Images provided by Dr. Kyle A. Gurley)

7E: Irradiation disrupts restorative regeneration. Top panels : Representative control trunk fragments displaying unpigmented regeneration blastemas by 5 dpa (white arrowheads). Bottom panels : Representative irradiated trunk fragments do not form blastemas (white arrowheads) or regenerate new tissues. Fragments curl ventrally and eventually lyse around 13 dpi. Scale bars 200 um.

7F: Amputation induces two waves of cell proliferation. Mitotic cells (white, anti-H3P antibody) are visualized in regenerating head fragments. A global burst in proliferation is observed within 6 hpa. By 2 dpa, a second proliferative burst occurs at the wound site (yellow arrowheads). Scale bar 200 um.

Biologists initially tried to integrate what they knew about embryogenesis with what they were learning about neoblast-based regeneration. At first, these cells were referred to as an “embryonic stock,” likened to blastomeres that persisted into adulthood to replenish injured or aging tissues. 171 , 113 Keller even suggested that neoblasts comprise a previously-unidentified fourth germ layer. 24 Soon, however, some biologists challenged the idea that neoblasts were a persistent undifferentiated pool of cells, and suggested instead that they were derived from differentiated tissues that had dedifferentiated or transdifferentiated—a phenomenon termed metaplasia. 172 , 173 , 174 , 175 Still others thought that planarian regeneration and tissue homeostasis might involve a combination of these two phenomena. 176 , 106 , 177 , 178 , 179

For much of the 20 th century, the source of a planarian’s regenerative abilities created a heated debate, especially as new scientific tools facilitated more sophisticated analyses. Some groups tried to specifically label differentiated cells versus neoblasts to determine what tissues they contributed to. Attempts were also made to visualize regenerated tissues using improved histochemical techniques. These experiments led a few researchers to believe they had observed various cell types dedifferentiate into neoblasts. 180 , 181 Others took advantage of differences in ploidy between the somatic and germline tissues, and reported that the germline could dedifferentiate or transdifferentiate into tissues such as muscle. 182 Some even argued that since dedifferentiation was the mechanism of regeneration identified in most other animals studied thus far, planarian biology must work the same way. 183

However, one tool proved key to properly addressing this question. In the first half of the 20 th century, it was demonstrated that ionizing radiation primarily kills neoblasts, causing the animals to lose the ability to undergo physiological and restorative regeneration ( FIG 7C–E , dpi: days post irradiation ). 184 , 185 This simple manipulation suggested that metaplasia might not play a major role in regeneration, as the animals died even though the differentiated tissues appeared to be intact. To further test this, Wolff and Dubois used a lead block to shield portions of planarians at different positions along the A/P axis, resulting in the destruction of all neoblasts not covered by the shield. Subsequently, they amputated the animals and demonstrated that the length of time required for blastema formation was proportional to the distance of the lead shield from the wound. 185 , 170 These results suggested that the surviving neoblasts migrated to the wound to facilitate regeneration, as opposed to the dedifferentiation of local tissues.

Fueled by improvements in cell labeling methods, cell culture, grafting techniques and microscopy, additional evidence mounted in support of neoblasts being collectively totipotent migratory stem cells. 186 , 187 , 188 , 189 , 190 , 191 , 192 Of note, Baguñà and colleagues took advantage of two different strains of Schmidtea mediterranea to identify the source of regenerated tissues. One of these strains is sexual while the other is asexual, and they can be distinguished at the cellular level by a distinct chromosomal translocation. Cell fractions enriched for either neoblasts or differentiated cells were isolated by serial filtration from sexual animals and injected into irradiated asexuals and vice versa. In both cases, only the neoblast-enriched fraction rescued irradiated animals. Furthermore, the host took on the sexuality and karyotype of the animal from which the cell fractions were isolated. 193 Coupled with the observation that dividing cells migrate out of unirradiated tissue grafts into irradiated host tissues, 194 , 195 , 196 these results strongly suggested that neoblasts are a collectively totipotent, migratory stem cell population.

Modern tools demonstrate that neoblasts are collectively totipotent stem cells

In an effort to pinpoint the cellular source of regenerated tissues using modern molecular tools, early efforts focused on identifying genetic markers for neoblasts ( FIG 7B ). These efforts have included cloning candidate stem cell and proliferation-dependent genes, generating EST libraries of regenerating animals, and testing antibodies against conserved proliferation-dependent histone modifications and cell cycle regulators. 197 , 198 , 44 , 199 , 70 , 200 , 94 , 201 , 202 , 203 Adaptation of FACS protocols enabled profiling and isolation of two side populations of cells mainly composed of neoblasts (termed “X1s”) and a mixture of neoblasts and their recent division progeny (termed “X2s”). 95 Also, after 35 years of attempts to incorporate modified thymidine analogs into proliferating neoblasts, BrdU was successfully optimized for use in planarians, facilitating the tracing of neoblasts and their division progeny. 204 , 205 , 100 Furthermore, the sequencing of the planarian genome 72 led to the development of microarrays to examine global gene expression changes after ablation of neoblasts by irradiation. These microarrays not only identified genes that define a molecular “signature” for neoblasts. 87 They also revealed a number of categories of genes that disappear at different timepoints after irradiation and have distinct distributions of expression in the planarian body plan. By performing BrdU tracing and co-localization studies with neoblast markers, it was shown that these genes are actually markers for lineages of differentiating neoblasts. 88

The ability to sort, label, and trace neoblasts and their division progeny made possible an impressive series of experiments that have put the debate about neoblasts’ contribution to regeneration to rest. A single transplanted neoblast—termed a clonogenic neoblast (cNeoblast) for its ability to generate colonies of cells—has been shown sufficient to rescue a lethally irradiated planarian. cNeoblasts display extensive pluripotency, and can differentiate into all of the cell types in the animal, except possibly the germline. Furthermore, BrdU labeling and strain-specific SNPs used to discern tissues derived from the transplanted cNeoblast versus those of the host suggest that dedifferentiation is unlikely to contribute significantly to planarian regeneration. 99

Neoblasts and their division progeny are a heterogeneous population

Historically, neoblasts were characterized by their morphology alone ( FIG 7A ). As a result, all neoblasts seemed roughly equivalent, with the exception that slight differences in cell shape could be observed. 206 However, modern studies demonstrate significant molecular heterogeneity amongst neoblasts and neoblast progeny. This suggests that while some neoblasts may be pluripotent, there could be subsets of neoblasts that are lineage-restricted and able to differentiate into only certain cell types.

Dj/Smed-nanos provided the first molecular hints of neoblast heterogeneity. In asexual planarians, it is expressed in only a subset of neoblasts. 61 , 62 , 207 , 63 , 98 Subsequently, single-cell PCR, immuno-EM, and in situ hybridization studies showed that neoblasts express various combinations of the canonical neoblast markers, in addition to genes normally associated with tissue-specific differentiation. 207 , 98 , 162 , 58 , 99 , 208 Even a number of the neoblast progeny markers identified by microarray do not colocalize extensively, suggesting a diversity of progeny lineages. 99

Finally, the classical observation that neoblasts display subtle diversity in morphology has been confirmed with improved means of isolating these cells. 209 Recent functional data suggest that neoblast morphology might indeed be indicative of heterogeneity in the population. Single cNeoblasts possessing a distinct membrane protrusion produced 75% of all rescue events. However, since the rate of rescue was quite low, with only 7 out of 130 injections successfully grafting, it is possible that many cells expressing the pan-neoblast marker Smed-piwi-1 might actually be a diverse population of multipotent cells. 99 It is currently unclear whether this molecular and morphological heterogeneity simply results from lineage restriction as pluripotent neoblasts differentiate, or whether there are permanent subpopulations of neoblasts that are restricted in potential.

Neoblasts cycle rapidly, migrate, and proliferate in response to injury

Classical descriptions of the behavior of neoblasts as a cell population are being re-examined with improved resolution and accuracy. For many decades, analysis of cell division in planarians was based on scoring mitotic figures in serial histological sections. 210 With the demonstration that BrdU could be incorporated by neoblasts in 2000, 100 the door opened for detailed analyses of the planarian cell cycle. We now know from continuous labeling with BrdU that around 20% of all planarian cells are cycling neoblasts, coming in at the lower end of classical estimates based on cell macerations. 211 , 97 Nearly all neoblasts enter S-phase and can be labeled with BrdU within 2–3 days of continuous exposure, suggesting that a large population of slow-cycling or G2-arrested neoblasts is unlikely to exist, as originally proposed. 212 , 100 , 97 The length of G2 has been estimated at approximately 6 hours and the average cell cycle length is around 21 hours. 100 , 97 In addition, changes in proliferation due to nutritional state were described classically, and it has been confirmed that a large proliferative burst 12–72 hours after feeding indeed corresponds with animal growth. 212 , 213 , 97 The basis of degrowth during starvation, however, is still debated. It may result from a decrease in neoblast proliferation, an increase in cell death of neoblast division progeny, or some combination of the two. 213 , 214 , 215

Recent work has confirmed and expanded upon older descriptions of the temporospatial dynamics of neoblast proliferation. 216 It now seems that the regenerative response can be divided into two distinct mitotic phases. During the first phase, neoblasts initiate a global burst in proliferation within 6 hours of any type of amputation or wound. The second burst requires the removal of tissue and is concentrated near the blastema, peaking around 48–72 hpa ( FIG 7F ). 217 Neoblasts can migrate as they differentiate and, in accordance with classical observations, they stop dividing before entering the blastema. 218 , 100 , 88 , 217 Finally, there is evidence supporting classical hypotheses that a signal emanating from the wound may trigger proliferation, as the initial mitotic increase seems to progress away from the wound in a wave-like fashion. 219 , 217

With this expanding toolkit and an improved understanding of the dynamics of neoblasts, we can now begin to identify the genes that regulate neoblast self-renewal and differentiation. While it is possible that conserved cell cycle regulators and pluripotency genes play similar roles in neoblasts as they do in other stem cell systems, it is equally likely that novel mechanisms for regulation of neoblasts may be discovered. Already, we are learning that even the most fundamental aspects of planarian cell division are surprisingly unique. Planarians, for instance, are the first animals identified that do not seem to require a centrosome for cell division at any point in their life history. 220 It is possible that such fundamental differences in regulation of cell division might be key to understanding why planarians have exceptionally robust regenerative abilities.

Many genes required for neoblast function have been identified

RNAi of genes from microarray experiments, expression profiling, EST libraries, and candidate ortholog searches of the planarian genome have already identified close to 200 genes whose phenotypes suggest defects in neoblast self-renewal and/or differentiation ( see Table 2 for references ). These phenotypes include reduced or absent blastemas, ventral curling, tissue regression, and lysis. Around thirty of these genes have been characterized in detail ( TABLE 2 ).

Summary of neoblast dysfunction phenotypes. All phenotypes characterized that cause a loss or reduction in blastema formation (not associated with a known polarity defect) and/or ventral curling and lysis are listed in alphabetical order for each species. ( 201 , 92 , 94 , 203 , 56 , 207 , 221 , 222 , 88 , 237 , 238 , 223 , 239 , 224 , 240 , 241 , 226 , 242 , 225 , 243 , 215 , 244 , 39 , 245 , 40 , 246 , 247 , 41 , 248 , 249 , 82 , 250 , 85 ) Reported data regarding numbers of neoblasts, cell proliferation based on Phosphor-Histone H3 (Ser10) staining, expression of early and late neoblast progeny markers, 88 results from clonogenic expansion assays, 248 and miscellaneous phenotype data are included.

While most of the phenotypes examined ultimately abolish proliferation and deplete neoblasts, an examination of the earlier stages of the phenotype progression reveals that neoblasts can be perturbed in numerous ways. First, neoblast self-renewal can be abrogated, as evidenced by a variety of phenotypes that display a decrease in proliferation and number of neoblasts soon after RNAi administration. Second, neoblasts can be disrupted at the level of differentiation. For instance, Smed-piwi-2(RNAi) does not affect neoblast numbers, their migratory ability, or their proliferative response after injury. Instead, differentiation is perturbed, as evidenced by the lack of regenerated tissues and the abnormal morphology of neoblast progeny that incorporate into the epithelium during homeostasis. 94 , 221 Additional examples of differentiation defects include Smed-p53(RNAi) , Smed-CHD4(RNAi) , and Smed-PTEN-1/2(RNAi). Among other abnormalities, RNAi of these genes seems to stall differentiation, causing an accumulation of neoblasts at the expense of postmitotic progeny. 222 , 223 , 224 Lastly, the spatial distribution of proliferating neoblasts can be disrupted, as it is after RNAi of Smed-egfr-3 or administration of a putative ERK inhibitor. 225 , 226

THE EMERGING ROLE OF DIFFERENTIATED TISSUES IN REGENERATION

The regeneration field has focused much of its efforts on the study of stem cells proper. The role of differentiated tissues has been appreciated mostly within the context of a cellular microenvironment known as a niche, which protects and maintains stem cells. 227 Considering that planarian regeneration requires not only local restoration of missing tissues, but also a simultaneous reproportioning of the entire body plan, it stands to reason that differentiated tissues may play important roles in regeneration on scales larger than what has been previously described for a stem cell niche.

Historically, the function of differentiated tissues during planarian regeneration has been largely dismissed as secondary to the action of neoblasts. Brøndsted, for instance, argued that neoblasts and the blastema they generate provide inductive cues to establish axial polarity in the rest of the pre-existing tissues. 228 Likewise, Betchaku’s cell culture experiments led him to view the fixed parenchymal cells as merely a vehicle for transporting neoblasts to the wound site so they can mount a regenerative response. 229 In the 1980s, the importance of differentiated tissues in directing neoblast differentiation was proposed. 195 However, only recently have the molecular tools become available to test this idea in a background completely devoid of neoblasts. These experiments have revealed a striking plasticity of the differentiated tissues during regeneration that occurs on a body-wide scale.

After irradiation ablates neoblasts and depletes their recent division progeny, the animals cannot regenerate new tissues. However, they can still undergo normal transcriptional responses after amputation. For example, in the complete absence of neoblasts, the differentiated tissues upregulate expression of early wound-response genes, in addition to re-specifying the A/P axis within 1 dpa. 230 , 151 , 133 , 131 , 226 Further examination of cell death dynamics reveals that two distinct waves of apoptosis occur within 4 hpa and 3 dpa. Surprisingly, the amount of apoptosis measured by TUNEL positive nuclei is normal in the absence of neoblasts, meaning that the cell death required for proper tissue remodeling occurs independently of stem cells. 231 Finally, it has been shown that the differentiated tissues in irradiated animals can dynamically modify body-wide transcriptional output for at least four days after an amputation, as evidenced by the oscillation of Smed-wnt11-5 expression across the A/P axis in irradiated tail fragments. It is only after 4 dpa that obvious defects in the expression of this gene become apparent, suggesting that the differentiated tissues may eventually need to integrate their new positional identity with the regenerated anatomy or neoblasts after a certain point in time. 133 , 131 While it is still unknown whether a niche for neoblasts exists, neoblasts and their local microenvironment are likely not the only elements required to understand planarian regeneration. Something about the nature of the global pre-existing differentiated tissues could be an important factor in determining to what extent an animal—whether it be a planarian or a human—can regenerate.

LOOKING TO THE FUTURE

After centuries of fascination with planarians, their regenerative abilities have transformed from a curiosity ultimately deemed intractable for detailed study by Morgan to an established animal model of regeneration. If classical biologists could have peered into the future, they would probably have been impressed by the amount of knowledge generated in just the past 15 years. Topping this list of accomplishments, it was confirmed that a pool of pluripotent neoblasts act as stem cells to replenish missing tissues. Numerous genes important for neoblast self-renewal and differentiation have been identified, and the first markers of neoblast lineages have been described. The molecular principles underlying axial polarity and organogenesis are already being teased out. In addition, the surprisingly dynamic behavior of whole tissues devoid of neoblasts is challenging us to reassess a stem cell-centric philosophy of regeneration. These results indicate that instead of acting only as a local niche, differentiated tissues may provide a macroenvironment capable of initiating wound responses, specifying axial polarity, and integrating global positional information that direct the subsequent differentiation of neoblasts. Of course, much still remains to be learned about these fascinating organisms.

From the top down, there are many facets of planarian regeneration biology to be elucidated. At the highest level, the rapid changes in transcriptional output observed after amputation or wounding suggests that the chromatin landscape and access to genomic promoters must be tightly coordinated. Chromatin dynamics will likely play key roles in these global transcriptional responses as the cells of the animal reassess positional information and facilitate the proper differentiation of neoblast progeny. Transcription factors and the targets they regulate in response to injury should also be studied so that gene regulatory networks for regeneration of specific tissues can be made to integrate the large body of functional data that will undoubtedly be generated. Additionally, increased cellular resolution will be required to study the effects of genes that undoubtedly have diverse temporospatial roles during physiological and restorative regeneration. Tools must be developed for indelibly labeling cells in vivo for fate mapping and live imaging studies. Single-cell transcriptional profiling, in addition to analysis of post-transcriptional and post-translational modifications occurring shortly after injury will also be key to teasing apart the molecular tapestry underlying regeneration.

By using this knowledge base, we can begin exploring how the mechanisms of regeneration formally compare to planarian embryogenesis. Such a comparison would begin to address the long-standing question of whether regeneration is simply a recapitulation of development or whether it is made possible by independent mechanistic innovations. How, for instance, are embryonic stem cells functionally different from neoblasts? How and when are neoblasts specified during embryogenesis? Is the same genetic toolkit required during embryogenesis to organize the body axes and facilitate organogenesis as it is during regeneration? This comparison may provide vital insights into a particularly curious phenomenon. Specifically, a variety of organisms display an impressive ability to undergo regulative embryonic development. Animals like the mouse, fruit fly, and frog can recover from ablation of numerous blastomeres or substantial injury to embryonic organs. However, these animals display limited regenerative capacities as adults. Understanding whether regulative development happens in planarian embryos and how it might differ from these other organisms may help us identify key differences crucial to preserving regenerative abilities into adulthood.

Finally, one of the ultimate goals of studying planarian regeneration is to understand why some animals regenerate robustly while others—such as humans—do not. Comparing the mechanisms of adult regeneration in diverse animals that have varying abilities to regenerate may be another way of pinpointing the core requirements for regeneration. Such an approach may also provide insights into the permutations that evolution has enacted upon this biological process over time. The first step in this endeavor should be to compare different planarian species, some of which regenerate robustly, while others display more limited abilities depending upon the plane of amputation. Elegant irradiation and grafting experiments performed on Procotyla fluviatilis , which does not regenerate robustly after post-pharyngeal amputation, suggest that variation in regenerative ability may result from the signals provided by the differentiated tissues, 232 , 233 , 187 , 234 and may not be explained simply by total numbers of neoblasts, as first thought. 206 Revisiting these classical studies with modern tools may help us identify a core set of molecular and physical principles guiding regeneration, which could then be examined in more distantly-related animal species.

With a rich history and giant leaps forward in recent years, the future of the planarian field is bright. Our mechanistic view of regeneration will undoubtedly come into much greater focus as more techniques are developed and more investigators pursue questions in this classical model system. It will be exciting to see what biological insights these animals will reveal to us next.

ACKNOWLEDGEMENTS

The authors acknowledge Dr. Kiyokazu Agata, Dr. Kyle Gurley, Dr. Jochen Rink, Dr. Tamaki Suganuma, Nobuo Ueda, Hanh Thi-Kim Vu, and Shigeki Watanabe for contributions indicated in the figure legends, in addition to Longhua Guo for translation assistance with T’uan’s Yu-Yang Tsa-Tsu . A.S.A is supported by NIH grant R37GM057260. A.S.A is a Howard Hughes Medical Institute and a Stowers Institute for Medical Research Investigator.

FURTHER READING

Schmidtea mediterranea Genome Database (SmedGD): http://smedgd.neuro.utah.edu/

Sánchez Alvarado lab website: http://planaria.stowers.org

HathiTrust, full text of Yu-Yang Tsa-Tsu by T’uan, 860: 14 http://catalog.hathitrust.org/Record/003326204

HathiTrust, full text of Kinmô-Zui by Tekisai Nakamura, 1666: 10 http://catalog.hathitrust.org/Record/002269510

HathiTrust, full text of Wakan Sansai-Zue by Ryōan Terajima, 1713: 11 http://catalog.hathitrust.org/Record/002269488

VIDEO 1: Planarian feeding. Schmidtea mediterranea extrudes its pharynx to feed on calf liver. Sped up to approximately twice normal speed.

VIDEO 2: Planarian defecation. Through its newly-regenerated pharynx, an 8 dpa Schmidtea mediterranea head fragment defecates in response to light eposure. Sped up to approximately twice normal speed.

VIDEO 3: Planarian cannibalism. A planarian consumes a decapitated head from another planarian of the same species. Sped up to approximately twice normal speed.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 22 August 2024

SoxC and MmpReg promote blastema formation in whole-body regeneration of fragmenting potworms Enchytraeus japonensis

• Toshiyuki Fujita 1 ,
• Naoya Aoki   ORCID: orcid.org/0000-0003-0138-2496 2 ,
• Chihiro Mori 2 ,
• Koichi J. Homma   ORCID: orcid.org/0000-0002-5165-419X 2 &
• Shinji Yamaguchi   ORCID: orcid.org/0000-0003-2184-431X 1  

Nature Communications volume  15 , Article number:  6659 ( 2024 ) Cite this article

71 Accesses

27 Altmetric

Metrics details

• Regeneration

Regeneration in many animals involves the formation of a blastema, which differentiates and organizes into the appropriate missing body parts. Although the mechanisms underlying blastema formation are often fundamental to regeneration biology, information on the cellular and molecular basis of blastema formation remains limited. Here, we focus on a fragmenting potworm ( Enchytraeus japonensis ), which can regenerate its whole body from small fragments. We find soxC and mmpReg as upregulated genes in the blastema. RNAi of soxC and mmpReg reduce the number of blastema cells, indicating that soxC and mmpReg promote blastema formation. Expression analyses show that soxC -expressing cells appear to gradually accumulate in blastema and constitute a large part of the blastema. Additionally, similar expression dynamics of SoxC orthologue genes in frog ( Xenopus laevis ) are found in the regeneration blastema of tadpole tail. Our findings provide insights into the cellular and molecular mechanisms underlying blastema formation across species.

Similar content being viewed by others

TIM29 is required for enhanced stem cell activity during regeneration in the flatworm Macrostomum lignano

Spatiotemporal transcriptomic atlas reveals the dynamic characteristics and key regulators of planarian regeneration

The brittle star genome illuminates the genetic basis of animal appendage regeneration

Introduction.

Regeneration is a prominent feature in many animals 1 . The degree of restoration of missing body parts or organs varies across species 1 , 2 , 3 . In highly regenerative animals, regeneration involves the formation of a blastema, which includes an undifferentiated mass of proliferative cells and provides the source material for lost body parts 1 , 2 , 4 , 5 , 6 , 7 , 8 . Hence, the mechanisms underlying blastema formation are fundamental for regeneration.

In vertebrates, the popular model systems for studying blastema formation are limb regeneration in salamanders and tail fin regeneration in zebrafish 9 . After wound healing, studies using lineage-tracing tools have identified that the connective tissue progenitors migrate to the amputation site and form regeneration blastema in both these model systems 10 , 11 . Among invertebrates, the highly regenerative animals are broadly distributed in many groups, including Platyhelminths 12 , Annelids 13 , Xenacoelomorphs 14 , Cnidarians 15 , and Sponges 16 . Among them, some animals can regenerate whole bodies from small fragments of tissues 7 . Planaria is the most intensively studied model for whole-body regeneration, which includes blastema formation 12 . Several studies on regeneration in planaria have indicated that cell migration is crucial for blastema formation in invertebrates 17 , 18 , 19 , 20 . To clarify whether the cellular mechanisms underlying blastema formation can be generalized among broad animal species, they need to be verified in other model systems.

Annelids have long been used as model animals for studying regeneration because of their impressive regenerative abilities 4 , 13 , 21 , 22 . Some annelids exhibit regenerative abilities comparable to planaria, which can regenerate the entire body from small fragments 7 , 13 . In oligochaetes (a group of annelids that includes earthworms), a population of cells was speculated to migrate to blastema along the ventral nerve cord (VNC) from a distant location 22 , 23 , 24 , which has not yet been determined. Putative multipotent stem cells, called neoblasts in oligochaetes (in contrast, in planaria, neoblasts were shown to be multipotent by an elegant transplantation experiment 25 ), are thought to be involved in this cell migration 4 ; however, no solid evidence has been reported to date.

Three hypotheses have been put forth to explain oligochaete regeneration, and these theories have been believed for approximately 130 years despite the absence of definitive evidence: 1) the existence of multipotent neoblasts, 2) cell migration to the amputation site to form blastema, and 3) migration of cells to the blastema along the VNC 4 , 13 , 22 , 24 , 26 , 27 . Nevertheless, the germ cell precursor population has been reported to accumulate along the VNC towards the anterior blastema 28 . To date, some progress has been made in identifying the molecules involved in blastema formation in oligochaete 29 , 30 , 31 , 32 , 33 . However, these hypotheses are yet to be clearly addressed.

The potworm Enchytraeus japonensis Nakamura, 1993 34 is an enchytraeid (an oligochaete) reported in Japan. It reproduces asexually by dividing its body into several fragments, which then regenerate into complete individuals within 4–5 days 34 (Fig.  1a ). The amputated fragments can similarly regenerate into whole bodies 35 . Enchytraeus japonensis is a small whitish animal that is approximately 10 mm in length and is easy to culture and handle. Additionally, blastema formation in E. japonensis is completed in approximately 24 h postamputation (hpa) 36 . All these features make E. japonensis an excellent laboratory animal for studying molecular and cellular basis underlying blastema formation.

a Enchytraeus japonensis grows to about 10 mm in length and reproduces asexually by fragmentation approximately every 2 weeks. b Workflow of transcriptomic analyses during regeneration. Intact worms were amputated into 3 fragments (blastema-poor group) or 8 fragments (blastema-rich group). c Venn diagrams depicting selectively upregulated contigs in regenerating animals (blastema-poor and blastema-rich groups) compared with that of intact animals. d Sequential filters to select upregulated contigs in regenerating worms. We selected the genes that showed higher upregulation in the blastema-rich than in the blastema-poor group based on the FPKM value (>10). e Expression level of soxC during regeneration. soxC was upregulated at 2, 6, and 24 hpa. n  = 5 for intact animals and for each time point. Error bars indicate the standard error of the mean (S.E.M.). f, g soxC expression during blastema formation at 0, 1, 3, 5, 8, and 24 hpa in horizontal sections ( f ) and sagittal sections ( g ). The sense controls and their diagrams are shown at the top. in, intestine; bw, body wall; se, septa; vn, ventral nerve cord. Dashed lines indicate amputation sites. f’ f” f”’ Magnified view of the boxed areas in ( f ) (1 hpa, arrowhead ( f’ ); 3 hpa, arrowhead ( f” ); 5 hpa, arrowhead ( f”’ )). h Quantification of soxC -expressing areas in the anterior blastema of regenerating worms at 0 hpa ( n  = 11), 3 hpa ( n  = 10), 8 hpa ( n  = 12) and 24 hpa ( n  = 12). Data was combined from three independent experiments. The central lines and the error bars indicate the mean and standard deviation (SD), respectively. *p  < 0.05, **p  < 0.01, ***p  < 0.001 (Dunnett’s test) ( e, h ). Scale bars represent 100 µm ( f, g ) and 20 µm ( f’, f”, f”’ ). Source data are provided as a Source Data file. A: anterior, P: posterior, D: dorsal, V: ventral ( f, g ).

In this work, using large-scale RNA-sequencing (RNA-seq) and the newly established functional RNA interference (RNAi) system in E. japonensis , we try to understand the molecular and cellular basis underlying blastema formation. We find soxC and mmpReg as upregulated genes in the blastema. RNAi of soxC and mmpReg reduce the number of blastema cells, indicating that soxC and mmpReg promote blastema formation. Expression analyses show that soxC -expressing cells appear to gradually accumulate in blastema and constitute a large part of the blastema. Furthermore, to examine whether the findings obtained using E. japonensis can be generalized across species, we conduct a comparative study using a tail regeneration model of a frog ( Xenopus laevis ) tadpole. We show similarity in the expression dynamics of SoxC genes in the blastema formation in E. japonensis and X. laevis , raising the possibility of common cellular mechanisms underlying blastema formation in both species.

soxC is upregulated in blastema formation in E. japonensis

To understand the cellular and molecular basis of blastema formation in E. japonensis , we tried to identify the genes that were selectively expressed in the blastema. To this end, we explored the differentially expressed genes between blastema and other body parts in E. japonensis using RNA-seq. We devised a dose-dependent strategy to overcome the challenges associated with the surgical isolation of blastema. We increased the amount of blastema obtained by cutting the worms into multiple fragments. These fragments were sorted into two groups of blastema-containing fragments: worms cut into 3 fragments (blastema-poor group) and worms cut into 8 fragments (blastema-rich group). Additionally, the third group contained intact worms (Fig.  1b ). In E. japonensis , blastema formation is completed in approximately 24 hpa 36 . We thus cut and cultured the worms for 24 h and used them for RNA-seq analysis (Fig.  1b ). We generated a large-scale RNA-seq data and de novo assembled it to generate 291,660 contigs with N50 value of 534 base pairs from all reads derived from the three libraries: blastema-rich group, blastema-poor group, and intact group. Among the 291,660 contigs, 86,094 common non-redundant contigs were detected in all three libraries (Fig.  1b ). We identified 1177 differentially expressed contigs between the blastema-poor and intact groups (Fig.  1c left and d) and 1593 differentially expressed contigs between the blastema-rich and intact groups (Fig.  1c right and d). We found 465 annotated genes that were differentially expressed (Fig.  1d ). Among these, we identified 278 genes that were more upregulated in the blastema-rich group than in the blastema-poor group at 10 fragments per kilobase of exon per million reads mapped (FPKM) values or higher (Fig.  1c, d , Supplementary Data  1 ). Among 465 annotated genes, we found 398 genes in the blastema-poor group shared with blastema-rich group at 10 FPKM values or higher. We identified 124 genes that were more downregulated in the blastema-rich group than in the blastema-poor group at 10 FPKM values or higher (Supplementary Fig.  1a , Supplementary Data  2 ).

Among the 278 upregulated genes, we performed in situ hybridization (ISH) using 4 genes ( soxC, mmpReg, lrp8, nas36 ) and found that all 4 were expressed in the blastema ( lrp8, nas36 ; Supplementary Fig.  2a, b ), suggesting that our strategy to identify genes associated with blastema was successful.

As transcription factors that act early in regeneration often establish and maintain cell identities 37 , we focused on transcription factors (Supplementary Data  3 ) among the 278 genes. The most upregulated gene among the transcription factors (7.9-fold upregulation in the blastema-rich group, Supplementary Data  3 ) belonged to the Sry-related high-mobility group (HMG) box (SOX)C transcription factor family. It contains a highly conserved HMG box DNA-binding domain (Supplementary Fig.  1b, c ), which is distinct from other families, namely, SOXB and SOXD (Supplementary Fig.  1d ). We named the gene Ej-soxC (hereafter referred to as soxC ). As the SOX transcription factor family generally contributes to determining cell fate and identity in many lineages 38 , we further analyzed soxC by performing expression analysis using quantitative reverse-transcription polymerase chain reaction (qRT-PCR). We found that soxC was upregulated during blastema formation until 24 hpa (Fig.  1e ) and leveled off at later stages (Supplementary Fig.  1e ). This suggests that soxC is involved in the regeneration of E. japonensis , especially blastema formation.

To understand the role of soxC in blastema formation, we first examined the soxC expression pattern using ISH in horizontal and sagittal sections of the regenerating fragments (Fig.  1f, g ). We did not detect soxC -positive cells just after amputation (0 hpa) and rarely detected them in intact worms (Supplementary Fig.  3 ), showing that soxC is not highly expressed in the homeostatic state. soxC -expressing cells were found close to the body wall and septa only near the posterior amputation site at 1 hpa (Fig.  1f, f’, g ). At 3 hpa, soxC- expressing cells were found close to the body wall near both the anterior and posterior amputation sites (Fig.  1f, f”, g ). At 5, 8, and 24 hpa, we detected the soxC- expressing cells at both anterior and posterior amputation sites (Fig.  1f, f”’, g ). The area of soxC -expressing cells in the amputation site increased during blastema formation (Fig.  1h ). At 24 hpa, the area of soxC- expressing cells encompassed a large part of the developing blastema (Fig.  1f, g , Supplementary Fig.  4a, b ). The soxC- expressing cells observed close to the body wall (at 3 and 5 hpa) were spread out and flat (Fig.  1f”, f”’ ). Phalloidin staining showed that the shape of the cells close to the body wall were also spread out and flat at 5 hpa (Supplementary Fig.  4c–e ). These shapes are typical of that of migrating cells 39 , 40 .

soxC promotes blastema formation

To test whether soxC regulates blastema formation, we silenced soxC using RNAi by soaking or feeding the worms with double-stranded RNA (dsRNA) (Fig.  2a, b , Supplementary Fig.  4f ). The soxC transcript level reduced by 40% (24 hpa) after soaking and feeding RNAi (Fig.  2c , Supplementary Fig.  4g ). Interference efficiency was approximately the same in both methods. However, soaking RNAi takes less time than feeding RNAi; therefore, subsequent RNAi experiments were performed using the soaking RNAi method. soxC silencing significantly decreased the number of cells around the amputation site compared with that of the control at 6 hpa (anterior: by 29%, p  = 0.0127; posterior: by 38%, p  = 6.65 × 10 −5 ; Fig.  2d–g ) and 24 hpa (anterior: by 33%, 4.2 × 10 −4 ; posterior: by 27%, p  = 0.00571; Fig.  2h–k ). As mentioned earlier, as the area of soxC- expressing cells encompassed a large part of the developing blastema (Fig.  1f, g , Supplementary Fig.  4a, b ), we analyzed the size of blastema by visualizing soxC expression. soxC RNAi decreased the volume of the blastema compared with that in the control at 6 hpa (by 59%, p  = 5.6 × 10 −10 ) and 24 hpa (by 52%, p  = 7.4 × 10 −8 ) (Fig.  2l–o ). These data showed that soxC is required for functions in blastema formation. At 2 days post amputation (dpa), soxC RNAi-treated animals did not show a significant decrease in the number of cells in the blastema (Supplementary Fig.  5a–e ). At 4 dpa, soxC RNAi-treated animals did not show significant difference in blastema size, morphology, or the anti-acetylated tubulin antibody-labelled central and peripheral nervous system compared with that in the control RNAi-treated animals (Supplementary Fig.  5j–l ). Proliferating cells were detected in the blastema of soxC RNAi- and control RNAi-treated animals (Supplementary Fig.  5m ). These data suggested that soxC silencing did not affect regeneration at 2 and 4 dpa. Taken together, soxC is required for functions in blastema formation for at least up to 24 hpa, indicating that soxC promotes blastema formation.

a Illustration of RNAi by soaking. b Schematic drawing of knock-down. Animals were amputated after 24 h of soaking in dsRNA and cultured for 1 day. c qPCR for the RNAi-mediated downregulation of soxC transcripts at 24 hpa ( n  = 4, biologically independent sample pools, p  = 0.0289). Error bars indicate S.E.M. d, e Representative images of RNAi animals at 6 hpa. Bright-field images ( d ). The nuclei were stained with DAPI ( e ). f, g RNAi of soxC reduced the number of cells in both the anterior blastema ( GFP RNAi ( n  = 20), soxC RNAi ( n  = 42), p  = 0.0127) ( f ) and posterior blastema ( GFP RNAi ( n  = 20), soxC RNAi ( n  = 43), p  = 6.6 × 10 −5 ) ( g ). Representative of two independent experiments (f, g ). h, i Representative images of RNAi animals at 24 hpa. Bright-field images ( h ). The nuclei were stained with DAPI ( i ). j, k RNAi of soxC reduced the number of cells in both the anterior blastema ( GFP RNAi ( n  = 25), soxC RNAi (n = 36), p  = 4.2 × 10 −4 ) ( j ) and posterior blastema ( GFP RNAi ( n  = 25), soxC RNAi (n = 38), p  = 0.0057) ( k ). Representative of two independent experiments ( j, k ). l – o soxC RNAi resulted in a decrease in the blastema volume at 6 hpa ( GFP RNAi ( n  = 22), soxC RNAi ( n  = 20), p  = 5.6 × 10 −10 ) ( m ) or 24 hpa ( GFP RNAi ( n  = 21), soxC RNAi (n = 23), p  = 7.8 × 10 −8 ) ( o ). Data was combined from two independent experiments ( m, o ). Typical images are shown (RNAi 6 hpa ( l ) or 24 hpa ( n )). t-test, * p  < 0.05, ** p  < 0.01, *** p  < 0.001 ( c, f, g, j, k, m, o ). Error bars: mean ± SD ( f, g, j, k, m, o ). Scale bars represent 250 µm ( e, i, l, n ). Dashed lines indicate the position of amputation sites ( d, e, h, i, l, n ). Dotted lines indicate the outline of regenerating worms ( e, i ). Ventral views ( d, e, h, i ). Sagittal sections ( l, n ). Source data are provided as a Source Data file.

mmpReg promotes blastema formation

To further delineate the molecular basis of blastema formation in E. japonensis , we performed a Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis 41 . We mapped the 465 differentially expressed genes between regenerating and intact animals to functional categories (Fig.  3a ). Among the 465 genes, we mapped 110 and frequently detected the extracellular matrix (ECM)-receptor interaction pathway category (Fig.  3b ). This pathway analysis raised the possibility that ECM remodeling might modulate blastema formation. The cleavage of ECM components is one of the main processes during ECM remodeling 42 . We focused on matrix metalloprotease (MMP) because MMPs are the primary enzymes involved in ECM degradation 42 . Among the 278 genes that showed higher upregulation in the blastema-rich group than in the blastema-poor group, we found that only one gene encoded a protein belonging to the MMP family of extracellular proteases (35.8-fold upregulated in the blastema-rich group, Fig.  1d , Supplementary Fig.  6a and Supplementary Data  1 ). It was characterized by a pro-domain and a zinc-binding motif with potential catalytic properties and hemopexin-like motifs (Fig.  3c ). Phylogenetic analysis showed that this mmp was somehow close to Japanese scallop Mizuhopecten yessoensis MMP2-like and sea mussels Mytilus edulis MMP-17 (Supplementary Fig.  6b ). qRT-PCR analysis showed that this mmp was upregulated during blastema formation until 24 hpa (Fig.  3d ) and leveled off from 1 to 4 dpa (Supplementary Fig.  6c ). This result suggests that this mmp may be involved in regeneration in E. japonensis , particularly in blastema formation. ISH showed that this mmp was expressed in the blastema at 5, 8, and 24 hpa (Fig.  3e, f ) and in the chloragocytes, which are specialized peritoneal cells of unknown function (Supplementary Fig.  6d ). These results also suggested that this mmp might be involved in blastema formation in E. japonensis .

a KEGG pathway categories of frequently detected genes differentially expressed between regenerating and intact animals. b The KEGG pathway subcategories of frequently detected genes are categorized as environmental information processing in the upper panel. c Schematic of the predicted MmpReg structure of E. japonensis . d Expression level of mmpReg during regeneration. mmpReg was upregulated after 2, 6, and 24 hpa. ** * p  < 0.001 (Dunnett’s test). Error bars indicate S.E.M. n  = 5 for intact animals and for other time points. e, f mmpReg expression during blastema formation at 1, 5, 8, and 24 hpa using horizontal sections ( e ) and sagittal sections ( f ). Dashed lines indicate amputation sites. Scale bars represent 100 µm ( e, f ). A: anterior, P: posterior, D: dorsal, V: ventral ( e, f ). Source data are provided as a Source Data file.

To determine whether this mmp regulates blastema formation, we examined whether silencing this mmp using RNAi affected blastema formation. Soaking worms in this mmp -specific dsRNA reduced mmp expression by 40% at 24 hpa (Fig.  4a ), which significantly decreased the number of cells around the amputation site compared with that in the control at 6 hpa (anterior: by 41%, p  = 0.00109; posterior: by 38%, p  = 2.06 × 10 −4 ; Fig.  4b–e ) and 24 hpa (anterior: by 40%, p  = 9.1 × 10 −5 ; posterior: by 37%, p  = 0.00165; Fig.  4f–i ). We analyzed blastema size by visualizing soxC expression. The mmp RNAi decreased blastema volume compared with that in control at 6 hpa (by 32%, p  = 4.8 × 10 −6 ) and 24 hpa (by 38%, p  = 2.7 × 10 −4 ) (Fig.  4j–m ). Similarly, treatment with an MMP inhibitor (MMP2/MMP9 inhibitor I, N-([1,1′-biphenyl]-4-ylsulfonyl)-D-phenylalanine) decreased blastema volume (by 40%, p  = 1.4 × 10 −5 ) (Fig.  4n,o ). These data showed this mmp is required for functions in blastema formation. Therefore, we named this mmp Ej-mmpReg (Reg referring to regeneration, hereafter referred to as mmpReg ). Similar to that of soxC RNAi-induced phenotype at 2 dpa, mmpReg RNAi-treated animals did not show a significant decrease in the number of blastema cells (Supplementary Fig.  5a, f–i ). At 4 dpa, no significant difference was found in blastema size, morphology, or the nervous system (Supplementary Fig.  5j, n, o ). Proliferating cells were detected in the blastema of mmpReg RNAi- and control RNAi-treated animals (Supplementary Fig.  5p ). These data suggested that mmpReg silencing did not affect regeneration on 2 and 4 dpa. Taken together, mmpReg is required for functions in blastema formation for at least up to 24 hpa, indicating that mmpReg promotes blastema formation.

a qPCR for the RNAi-mediated downregulation of mmpReg transcripts at 24 hpa ( n  = 4, biologically independent sample pools, p  = 0.00732). Error bars indicate S.E.M. b, c Representative images of animals at 6 hpa. Bright-field images ( b ). The nuclei were stained with DAPI ( c ). d, e RNAi of mmpReg reduced the number of cells in both the anterior blastema ( GFP RNAi ( n  = 22), mmpReg RNAi (n  = 30), p  = 0.00109) ( d ) and posterior blastema ( GFP RNAi ( n  = 22), mmpReg RNAi ( n  = 29), p  = 2.06 × 10 −4 ) ( e ). Representative of two independent experiments ( d, e ). f, g Representative images of animals at 24 hpa. Bright-field images ( f ). The nuclei were stained with DAPI ( g ). h, i RNAi of mmpReg reduced the number of cells in the anterior blastema ( GFP RNAi (n = 20), mmpReg RNAi ( n  = 32), p  = 9.1 × 10 −5 ) ( h ) and posterior blastema ( GFP RNAi (n = 20), mmpReg RNAi ( n  = 31), p  = 0.00165) ( i ). Representative of two independent experiments ( h, i ). j – m mmpReg RNAi resulted in a decrease in the blastema volume at 6 hpa ( GFP RNAi ( n  = 22), mmpReg RNAi ( n  = 20), p  = 4.8 × 10 −6 ) ( k ) or 24 hpa ( GFP RNAi (n = 24), mmpReg RNAi (n = 29), p  = 2.7 × 10 −4 ) ( m ). Representative images are shown (RNAi 6 hpa ( j ) or 24 hpa ( l )). n, o MMP inhibitor (MMP2/MMP9 inhibitor I, N-([1,1’-biphenyl]−4-ylsulfonyl)-D-phenylalanine, CAS#:193807−58-8) resulted in decreased blastema volume at 24 hpa (DMSO (n = 22), inhibitor ( n  = 24), p  = 1.4 × 10 −5 ) ( o ). Representative images are shown (MMP inhibitor ( n )). Data was combined from two independent experiments ( k, m, o ). t-test, *p  < 0.05, ** p  < 0.01, *** p  < 0.001( a, d, e, h, i, k, m, o ). Error bars: mean ± SD ( d, e, h, i, k, m, o ). Scale bars represent 250 µm ( c, g, j, l, n ). Dashed lines indicate amputation sites ( b, c, f, g, j, l, n ). Dotted lines indicate the outline of regenerating worms ( c, g ). Ventral views ( b, c, f, g ). Sagittal sections ( j, l, n ). Source data are provided as a Source Data file.

A limitation of our study is that the RNAi-induced phenotypes for soxC and mmpReg appeared to be relatively mild. Whether this phenomenon reflects gene function or the technical limits of our gene silencing system (e.g., quick recovery from RNAi-induced silencing by 2 dpa) is unclear. In our normal protocol, intact worms were soaked with dsRNA for 24 h and amputated worms were allowed to undergo regeneration on filter paper kept moist with Milli-Q water. To continuously treat amputated worms with dsRNA, we cultured them on filter paper kept moist with dsRNA solution until 2 dpa. At 2 dpa, animals continuously treated with soxC and mmpReg dsRNA did not exhibit a significantly reduced cell number in the blastema (Supplementary Fig.  5q-u ), suggesting a similar phenotype to those treated following our normal protocol (Supplementary Fig.  5a–I, q–u ).

Cell proliferation and blastema formation

To clarify the mechanisms underlying soxC and mmpReg promoted blastema formation, we analyzed cell proliferation, which is a hallmark of blastema formation 1 . We investigated the cellular proliferation dynamics in the blastema using 5′-ethynyl-2′-deoxyuridine (EdU) labelling analysis to identify the time point at which the proliferating cells appeared in the blastema (Fig.  5a ). The number of DAPI-labelled cells increased in the anterior (3 vs. 5 hpa, p  < 0.05; 3 vs 8 hpa, p  < 0.001; 3 vs. 24 hpa, p  < 0.001; Fig.  5b, c, e ) and posterior (3 vs 5 hpa, p  = 0.0634; 3 vs 8 hpa, p  < 0.001; 3 vs. 24 hpa, p  < 0.001; Fig.  5b, c, g ) blastema regions, whereas the number of EdU-incorporated cells did not increased in the anterior (0–3 hpa vs 3–5 hpa, p  = 1; 0–3 hpa vs 5–8 hpa, p  = 0.464; Fig.  5b, d, f ) and posterior (0–3 hpa vs 3–5 hpa, p  = 1; 0–3 hpa vs 5–8 hpa, p  = 0.153; Fig.  5b, d, h ) blastema regions, but increased during 21–24 hpa in the blastema (vs. 0–3 hpa, anterior, p  < 0.001; posterior, p  < 0.001; Fig.  5b, d, f, h ). These data showed that the number of cells increased, whereas that of proliferating cells rarely increased at 0–8 hpa. This suggests that the cells are supplied from elsewhere to the amputation site rather than proliferate at the amputation site up to 8 hpa.

a Schematic drawing of EdU labelling. Small arrows indicate the fixation points with PFA–PBS. b, c, d Representative images of regenerating animals at 3, 5, 8 and 24 hpa. Bright-field images ( b ). The nuclei were stained with DAPI ( c ). EdU-incorporated cells ( d ). e, f, g, h The number of cells increased in both anterior blastema ( e , 3 hpa ( n  = 18), 5 hpa ( n  = 32), 8 hpa ( n  = 18), 24 hpa ( n  = 25)) and posterior blastema ( g , 3 hpa ( n  = 19), 5 hpa ( n  = 32), 8 hpa ( n  = 18), 24 hpa (n = 25)). The number of proliferating cells did not significantly increase (0–3 vs 3–5, 0–3 vs 5–8 hpa), but increased at 21–24 hpa (0–3 vs 21–24 hpa) in anterior and posterior blastema ( d, f, h ). *p  < 0.05, **p  < 0.01, *** p  < 0.001. NS indicates not significant (Dunnett’s test). Representative of two independent experiments (e, f, g, h ). Error bars: mean ± SD ( e, f, g, h ). Scale bars represent 250 µm ( c, g, j, l, n ). Dashed lines indicate amputation sites ( b, c, d ). Dotted lines indicate the outline of regenerating worms ( c, d ). Ventral views ( b, c, d ). Source data are provided as a Source Data file.

Next, to explore whether soxC and mmpReg are required for functions in cell proliferation in the blastema, we examined the effect of soxC and mmpReg RNAi on cell proliferation (Fig.  6a, i ). At 3 and 5 hpa, both soxC and mmpReg RNAi did not affect the number of anterior and posterior blastema cells ( soxC , p  > 0.05 [Fig.  6b, c, e, g ]; mmpReg , p  > 0.05 [Fig.  6j, k, m, o ], except for mmpReg RNAi posterior at 5 hpa p  = 0.0030; Tukey– Kramer’s test) or proliferating cells ( soxC , p  > 0.05 [Fig.  6b, d, f, h ]; mmpReg , p  > 0.05 [Fig.  6j, l, n, p ]; Tukey–Kramer’s test). At 8 hpa, soxC RNAi significantly decreased the number of cells around the amputation site compared with that in the control (anterior, p  < 0.01 [Fig.  6b, c, e ]; posterior, p  < 0.05 [Fig.  6b, c, g ]; Tukey–Kramer’s test), whereas the number of proliferating cells did not significantly differ at 5–8 hpa in the blastema (anterior, p  = 0.91, [Fig.  6b, d, f ]; posterior, p  = 0.33, [Fig.  6b,d,h ]; Tukey–Kramer’s test). At 24 hpa, soxC RNAi reduced both the number of DAPI-stained cells (anterior, p  < 0.01, [Fig.  6b, c, e ]; posterior, p  < 0.0001, [Fig.  6b, c, g ]; Tukey–Kramer’s test) and EdU-incorporated proliferating cells (anterior, p  < 0.001, [Fig.  6b, d, f ]; posterior, p  < 0.05, [Fig.  6b, d, h ]; Tukey–Kramer’s test). Moreover, mmpReg RNAi elicited a similar effect on the number of cells and cellular proliferation dynamics in blastema at 8 hpa (DAPI: anterior, p  < 0.01; posterior, p  < 0.01, [Fig.  6j,k,m,o ]; EdU: anterior, p  = 0.57; posterior, p  = 0.44, [Fig.  6j, l, n, p ]; Tukey–Kramer’s test) and 24 hpa (DAPI: anterior, p  < 0.0001; posterior, p  < 0.0001, [Fig.  6j, k, m, o ]; EdU: anterior, p  < 0.0001; posterior, p  < 0.01, [Fig.  6j, l, n, p ]; Tukey–Kramer’s test). These data showed that for at least up to 5–8 hpa, both soxC and mmpReg are not required for cell proliferation, but functions to supply cells to the amputation site; they are required subsequently for cell proliferation.

a Schematic drawing of EdU labelling in soxC RNAi-subjected worms. Small arrows indicate the fixation points with PFA–PBS. b, c, d Representative images of regenerating animals at 8 and 24 hpa. Bright-field images ( b ). The nuclei were stained with DAPI ( c ). EdU-incorporated cells ( d ). e, f, g, h Quantification of DAPI-stained cells. 3 hpa ( GFP ( n  = 41), soxC ( n  = 43)), 5 hpa ( GFP ( n  = 47) , soxC (n  = 44), 8 hpa ( GFP ( n  = 43), soxC ( n  = 44)), 24 hpa ( GFP ( n  = 40), soxC (n = 42)) in the anterior blastema ( e ) and 3hpa ( GFP (n = 40), soxC ( n  = 42)), 5 hpa ( GFP ( n  = 46), soxC ( n  = 41)), 8 hpa ( GFP ( n  = 43)), soxC ( n  = 44)), 24 hpa ( GFP ( n  = 38), soxC ( n  = 38)) in the posterior blastema ( g ). Quantification of EdU-incorporated cells in the anterior blastema ( f ) and posterior blastema ( h ). i Schematic drawing of EdU labelling in mmpReg RNAi worms. Small arrows indicate the fixation points with PFA–PBS. j, k, l Representative images of regenerating animals at 8 and 24 hpa. Bright-field images ( j ). The nuclei were stained with DAPI ( k ). EdU-incorporated cells ( l ). m, n, o, p Quantification of DAPI-stained cells. 3 hpa ( GFP (n = 40), soxC (n  = 31)), 5 hpa ( GFP ( n  = 35) , soxC (n  = 34), 8 hpa ( GFP ( n  = 45), soxC ( n  = 52)), 24 hpa ( GFP ( n  = 44), soxC ( n  = 44)) in the anterior blastema ( m ) and 3hpa ( GFP ( n  = 39), soxC ( n  = 31)), 5 hpa ( GFP ( n  = 35), soxC ( n  = 34)), 8 hpa ( GFP ( n  = 44)), soxC (n  = 51)), 24 hpa ( GFP ( n  = 43), soxC ( n  = 44)) in the posterior blastema ( o ). Quantification of EdU-incorporated cells in the anterior blastema ( n ) and posterior blastema ( p ). *p  < 0.05, **p  < 0.01, ***p  < 0.001. NS indicates not significant (Tukey-Kramer’s test). Error bars: mean ± SD ( e, f, g, h, m, n, o, p ). Dashed lines indicate amputation sites ( b, c, d, j, k, l ). Dotted lines indicate the outline of regenerating worms ( c, d, k, l ). Ventral views ( b, c, d, j, k, l ). Source data are provided as a Source Data file.

Expression patterns of stem cell marker genes at homeostatic state

To elucidate the relationship between soxC -expressing cells and putative stem cells at the homeostatic state, we examined the expression patterns of known stem cell marker genes in E. japonensis (0 hpa or intact worm). We used the germline multipotency program (GMP) genes, including piwi, vasa, nanos , and pl10 , as stem cell marker genes as they are broadly expressed in germline and multipotent stem cells in animals 43 . We analyzed the 86,094 assembled contigs obtained from our transcriptomic analysis and identified two vasa genes ( Ej-vasa1 and Ej-vasa2) , Ej-nanos, Ej-pl10 , and Ej-piwi (Supplementary Fig.  8a,b,c ). All five GMP genes were not differentially expressed between the blastema-rich and intact groups (>1.0- and <1.7-fold). The amino acid sequence of Ej-vasa1 was similar to that of Ej-vasa-like gene ( vlg ) -1 (89.9% identity, Supplementary Fig.  9a ), and the amino acid sequence of Ej-vasa2 was similar to that of Ej-vlg-2 (86.6% identity, Supplementary Fig.  9b ), as previously reported for E. japonensis 44 . Some previous studies have partially examined the expression patterns of some GMP genes in E. japonensis 28 , 44 , 45 . Our ISH analysis showed the presence of Ej-vasa1- positive cells close to the body wall just after amputation (0 hpa) (Fig.  7a , Supplementary Fig.  11a ) and in intact worms (Supplementary Fig.  12a, b ). The expression pattern of Ej-vasa1 at 0 hpa was highly similar to those in intact worms, suggesting that 0 hpa time point can be considered as the homeostatic state. Consistent with a previous study 28 , Ej-piwi- expressing cells were found on the VNC (Fig.  7d ). These cells act as germ cell precursors during gonad regeneration in E. japonensis 28 . Ej-nanos and Ej-pl10 were expressed in the cells close to the body wall after amputation at 0 hpa (Fig.  7b, c , Supplementary Fig.  11b,c ) but Ej-piwi and Ej-vasa2 were not detected in the cells close to the body wall after amputation at 0 hpa (Fig.  7d, e , Supplementary Fig.  11d, e ). Additionally, Ej-vasa2 was expressed in cells in the body cavity at 0 hpa (Fig.  7e ). The expression patterns of all GMP genes that were tested differed from that of soxC in a homeostatic state.

a–e GMP gene ( Ej-vasa1 ( a ), Ej-nanos ( b ), Ej-pl10 ( c ), Ej-piwi ( d ), Ej-vasa2 ( e )) expression during blastema formation at 0, 1, 5, and 24 hpa in horizontal sections (upper eight panels) and sagittal sections (lower eight panels). Dashed lines indicate amputation sites. f GMP gene expression in gonad and PGZ at 0 hpa. Expression level varies depending on the GMP gene, but all GMP genes were expressed in the gonad and PGZ, except for Ej-piwi in PGZ at 0 hpa. Arrowheads in the left panels indicate the position of the gonad, and arrowheads in the right panels indicate the position of PGZ. g GMP gene expression in neoblast at 0 hpa. Ej-vasa1 was expressed in neoblasts, whereas the other GMP genes were not detected in neoblasts at 0 hpa. The regions around the septa are shown in each panel. Dotted lines indicate the outline of neoblasts. Scale bars represent 250 µm ( a – f ) and 20 µm ( g ). A: anterior, P: posterior, D: dorsal, V: ventral ( a–f ). All images are representative of at least two independent experiments ( a – g ).

Next, we examined the expression pattern of the GMP genes in the body parts where putative stem cells are considered to be present: the posterior growth zone (PGZ), gonad (recently referred to as multipotent progenitor cell (MPC) clusters in a marine annelid Capitella teleta 46 ), and neoblasts 13 . The gonad expressed all GMP genes that were examined ( Ej-vasa1, Ej-vasa2, Ej-nanos, Ej-pl10, Ej-piwi ) (Fig.  7f ), whereas PGZ expressed all GMP genes except Ej-piwi ( Ej-vasa1, Ej-vasa2, Ej-nanos, Ej-pl10 ) (Fig.  7f , Supplementary Fig.  11f ). In contrast, the neoblasts only expressed Ej-vasa1 (Fig.  7g , Supplementary Fig.  10 ). Furthermore, soxC was expressed in PGZ but not in the gonad and neoblasts (Supplementary Figs.  3a , b, 10b ). Taken together, these data showed that the expression patterns of the tested GMP genes were different from that of soxC except for PGZ.

Expression dynamics of soxC and GMP genes are different

To understand the relationship between soxC -expressing cells and putative stem cells during blastema formation, ISH for the GMP genes was performed. At 5 hpa, soxC -expressing cells were present (Fig.  1f,g ) but Ej-vasa1 -expressing cells were absent in the blastema (Fig.  7a ). In contrast, at 24 hpa, Ej-vasa1 was expressed in a small portion of the blastema (Fig.  7a ), which is reminiscent of the results of previous studies in other annelids 47 , 48 that GMP genes were expressed de novo in the early blastema. To clarify the relationship between the Ej-vasa1 -expressing and soxC -expressing cells, we examined whether Ej-vasa1 and soxC were co-expressed in blastema. A double ISH analysis revealed that the soxC and Ej-vasa1 expression patterns did not fully overlap in the blastema (18 out of 18 fragments) (Fig.  8 ). Ej-piwi- expressing cells were not found in the blastema even at 0–24 hpa (Fig.  7d ). Similarly, Ej-vasa1, Ej-nanos , Ej-pl10 , and Ej-vasa2 were not detected in the cells near the anterior or posterior amputation sites at 0–5 hpa (Fig.  7a, b, c, e ) but were expressed in a small portion of the blastema at 24 hpa (Fig.  7a,b,c,e ). The neoblasts only expressed Ej-vasa1 at 0–24 hpa (Fig.  7g , Supplementary Fig.  10a ). The expression pattern of all the tested GMP genes varied considerably from that of soxC during blastema formation, particularly at 0–5 hpa.

DIG-labeled soxC and fluorescence-labeled Ej-vasa1 RNA antisense probes were used for double ISH in horizontal sections of regenerating worms at 24 hpa. Arrowheads indicate the position of Ej-vasa1 -expressing cells. Scale bars represent 25 µm. Representative of two independent experiments.

Expression dynamics of soxC genes during tadpole tail blastema formation

We consider the possibility that animals other than annelid may display similar expression dynamics of soxC during blastema formation because SOXC transcription factors are widely conserved across animals 49 . In many vertebrates, SOX4, SOX11, and SOX12 constitute the SOXC family 50 . We examined this possibility using blastema in regenerating the African claw frog ( X. laevis , amphibian) tadpole tail. Blastema formed within 48 hpa in the tadpole tail 51 (Fig.  9a ). ISH was performed using serial sections of regenerating tadpole tail. It showed that Xl-sox4 and Xl-sox11 were not detected just after amputation at 0 hpa (Fig.  9b,e ); only Xl-sox11 was expressed around amputation sites at 6 hpa (Fig.  9b,e ). Then, both Xl-sox4 and Xl-sox11 were expressed around amputation sites at 24 hpa (Fig.  9b,c ), particularly underneath the outer layer that covers the wounds (Fig.  9c,f ). At 48 hpa, Xl-sox4 and Xl-sox11 were expressed abundantly in the blastema (Fig.  9b,e ). Time–course ISH analysis showed that Xl-soxC -expressing cells appeared to accumulate gradually in the blastema. qRT-PCR showed that the Xl-sox4 and Xl-sox11 expression levels increased in the regenerating tadpole tail, including the blastema (Fig.  9d,g ). Overall, these data indicate that the regenerating frog tadpole tail displays similar expression dynamics as that of soxC in E. japonensis in blastema formation.

a Schematic drawing of tadpole tail blastema formation. A: anterior, P: posterior, D: dorsal, V: ventral. b Xl-sox4 expression during blastema formation at 0, 6, 24 and 48 hpa. The sense controls are shown on the lower panels. c Magnified view of the boxed areas in ( b , 24 hpa). d Expression level of Xl-sox4 during blastema formation. sox4 was upregulated after amputation. n  = 4 for time points. e Xl-sox11 expression during blastema formation at 0, 6, 24 and 48 hpa. The sense controls are shown on the lower panels. f Magnified view of the boxed areas in ( e , 24 hpa). g Expression level of Xl-sox11 during blastema formation. Xl-sox11 was upregulated after amputation. n  = 4 for time points. ( d, g ) Error bars indicate S.E.M. *p  < 0.05, **p  < 0.01, ***p  < 0.001 (Dunnett’s test). Dashed lines indicate amputation sites ( a, b, e ). Source data are provided as a Source Data file.

In this study, we identified 278 upregulated (Fig.  1d , Supplementary Data  1 ) and 124 downregulated (Supplementary Fig.  1a , Supplementary Data  2 ) genes involved in blastema formation in E. japonensis using large-scale RNA-seq. Additionally, we developed a system for silencing genes in the potworms, which involves soaking or feeding the worms with dsRNA. This system is an easily applicable, reproducible, and minimally invasive strategy for gene silencing (Figs.  2 a–o, 4 a–m, 6a–p , Supplementary Fig.  5a–p ) (the first RNAi study using injection in E. japonensis was published by Takeo, et al. 33 ). We expect that analyzing the functions of the genes on our lists using the functional RNAi tool would significantly contribute to enhancing our understanding of the molecular and cellular basis of blastema formation in E. japonensis . Here, we present the significance of two genes ( soxC and mmpReg ) that are upregulated in blastema formation.

We found that soxC was rarely expressed in the homeostatic state (Supplementary Fig.  3 ) and was upregulated during regeneration (Fig.  1f,g ). After amputation, soxC was expressed selectively close to the body wall and septa near the amputation site at 1 hpa (posterior) and 3 hpa (anterior and posterior) (Fig.  1f,g ). The soxC -expressing cells near the body wall were spread out and flat (Fig.  1f”,f”’ ), which is typical of migrating cells 39 , 40 . Subsequently, the area in which soxC was expressed gradually increased (Fig.  1f,g ). The area of soxC- expressing cells encompassed to a large part of the developing blastema at 24 hpa (Fig.  1f,g , Supplementary Fig.  4a,b ). Furthermore, soxC silencing led to a significant decrease in the number of cells around the amputation site at 6 and 24 hpa (Fig.  2a–o ), showing that soxC is required for functions in blastema formation at least up to 24 hpa. Taken together, these results suggest that soxC can be used as a representative marker for blastema formation (Fig.  10a ).

a GMP gene(s)-expressing cells (ASCs, shown in blue) are present in intact worms. soxC -expressing cells (shown in red) might be the cells supplied to the amputation site. The relationship between soxC- expressing cells and ASCs remains unclear. soxC- expressing cells were distinct from the population of cells on the VNC (shown in orange), which are Ej-piwi -positive cells that serve as germ cell precursors. b Similarity in expression dynamics of SoxC genes in E. japonensis and X. laevis raised the possibility of the common cellular mechanisms underlying blastema formation in both species. In the cnidarian polyp Hydra vulgaris , HvSoxC was expressed in a transition state between i-cells and differentiated cells 63 (*).

The multipotent stem cells maintained in the adults (hereafter referred to as adult stem cells (ASCs)) of other whole-body regeneration animals generally express GMP genes 7 , 52 . For instance, neoblasts in planaria are molecularly defined by the expression of piwi-1 53 . Multipotent interstitial stem cells (i-cells) in Hydra expressed piwi , Vasa, and PL10 54 , 55 . We found that in E. japonesis , Ej - vasa1 was expressed in all putative ASC niches, which we have described subsequently. Furthermore, Ej - vasa1 was expressed abundantly in the homeostatic state (Supplementary Fig.  12a ). These results suggest that Ej - vasa1 can be used as a representative marker for ASCs in E. japonensis . Additionally, GMP genes were not found in blastema at 0–5 hpa (Fig.  7a–e ). In contrast, soxC was not detected at 0 hpa (Fig.  1f,g ) and was expressed in blastema at 3–5 hpa (Fig.  1f,g ), showing that soxC and Ej-vasa1 distribution patterns differed in blastema. Moreover, their distribution patterns did not fully overlap in blastema (Fig.  1f, g , Fig.  7a ), as confirmed by the double ISH analysis results at 24 hpa (Fig.  8 ). Taken together, these data suggest that a part of the ASC population differed from that of the soxC- expressing cells. Currently, the relationship between soxC- expressing cells and ASCs remains unclear, and further information may be obtained by enumerating the relationship between soxC- expressing cells and ASCs in annelid blastema formation (Fig.  10a ).

In annelids, putative ASCs found in the PGZ, gonad (recently referred to as MPC clusters in a marine annelid, C. teleta 46 ), and neoblasts 13 . The PGZ of annelids is a highly proliferative region at the posterior tip of the body that produces new segments during normal growth 48 , 56 , 57 . In E. japonensis , the PGZ expressed all the tested GMP genes with the exception of Ej-piwi, showing that PGZ contains putative ASCs (Fig.  7f , Supplementary Fig.  11f ). Additionally, soxC was expressed in the PGZ at the homeostatic state (Supplementary Figs.  3 , 10b ), suggesting that soxC is also involved in normal growth. The gonad was identified in the 7 th and 8 th segments in E. japonensis 28 . A pair of neoblasts are located adjacent to the septa in each E. japonensis segment 58 . Neoblasts of annelids have been hypothesized to correspond to ASCs 4 , although this is yet to be proven. In contrast, in planaria, neoblasts were proven to be multipotent through an elegant transplantation experiment 25 . In E. japonensis , neoblasts only expressed Ej-vasa1 (Fig.  7g ), whereas the cells close to the body wall expressed Ej-vasa1, Ej-nanos , and Ej-pl10 (Fig.  7a–c ). The GMP gene-expressing cells close to the body wall appeared to be distinct from neoblasts. These results suggest the heterogeneity of vasa1 -expressing ASCs.

Historically, many studies have speculated that a population of cells might migrate to blastema along the VNC 4 , 13 , 23 , 24 instead of through the body wall in oligochaete; however, limited evidence is available. Recently, time-lapse imaging of a freshwater annelid Pristina leidyi showed wound-directed migrating cells 40 . However, these migrating cells were detected and classified using purely morphological features, with no molecular markers available. Owing to the lack of molecular markers, the types of cells that contribute to blastema formation remain unknown. To date, ISH analysis in E. japonensis using GMP genes as markers has shown that Ej-vasa1 -expressing cells and Ej-piwi -expressing cells are attached to the dorsal side of the VNC 28 , 44 (consistent with the data shown in Fig.  7a , d) and that Ej-vasa1 and Ej-piwi are co-expressed 44 . Ej-piwi- expressing germ cell precursors attached to the VNC were suggested to migrate to blastema 28 . However, Ej-piwi- expressing cells were not detected in the blastema at 0–24 hpa 28 (consistent with our data in Fig.  7d ). Hence, no evidence exists that Ej-piwi- expressing cells are related to blastema formation at least 0–24 hpa. Another study claimed that Ej-vasa1 -expressing cells found close to the body wall could migrate to blastema based on the results of an ISH analysis at 48 hpa 45 . However, we found that Ej-vasa1 -expressing cells were not detected in the blastema at 0–5 hpa (Fig.  7a ), whereas a certain number of DAPI-stained cells were detected in blastema up to 5 hpa (Fig.  5c,e,g ). Hence, no evidence exists that Ej-vasa1 -expressing cells are related to blastema formation at least at 0–5 hpa. In contrast, soxC -expressing cells were found close to the body wall and septa near the amputation site at 1 hpa (posterior) and 3 hpa (anterior and posterior) (Fig.  1f,g ). Time–course ISH analysis showed that soxC -expressing cells appeared to gradually accumulate in the blastema (Fig.  1f,g ). The area of soxC -expressing cells encompassed to a large part of the developing blastema (Fig.  1f,g , Supplementary Fig.  4a,b ). Expression dynamics of soxC tended to coincide with the increase in the number of DAPI-stained cells during blastema formation (Figs.  1 h, 5e,g ). Additionally, RNAi experiments showed that soxC is required for functions in blastema formation at 6 and 24 hpa (Fig.  2a–o ). All these data suggest that soxC- expressing cells contribute to blastema formation. We further investigated the cellular proliferation dynamics in the blastema using EdU labelling analysis and found that the number of cells in the blastema gradually increased before they started proliferating at 8 hpa (Fig.  5a–h ). This suggests that the cells were supplied from elsewhere to the amputation site up to 8 hpa rather than proliferated. Before the onset of cell proliferation, soxC RNAi led to a significant decrease in the number of cells in blastema (Fig.  6a,b,c,e,g ). This result suggests that soxC is required for functions to supply cells from elsewhere to the amputation site up to 8 hpa. Taken together, possibly, soxC -expressing cells may be the cells supplied to the amputation site (Fig.  10a ).

We identified mmpReg as another upregulated gene in blastema formation. Recently, a single-cell RNA-seq (scRNA-seq) study in salamander limb regeneration showed that Mmp8 and Mmp13 were upregulated during blastema formation, indicating extensive ECM remodeling by MMPs 59 . Similar to the results obtained using soxC RNAi, mmpReg silencing decreased the number of cells in blastema and delayed blastema formation at 6 and 24 hpa. Before onset of cell proliferation (at 8 hpa), mmpReg RNAi significantly decreased the number of cells in the blastema (Fig.  6a,b,c,e,g ). We analyzed the size of blastema by visualizing soxC expression. At 6 and 24 hpa, mmpReg RNAi and MMP inhibitor decreased the volume of the blastema (Fig.  4j–o ). Our data indicate that mmpReg promotes blastema formation. Notably, the role of MmpReg in regenerating E. japonensis needs to be clarified by determining the substrate ECM proteins and the stimulation factors that release MmpReg to better understand the whole picture of blastema formation at the molecular level. Currently, we did not detect an epistatic relationship between soxC and mmpReg in terms of level of expression by knock-down (Supplementary Fig.  7a,b ). However, further investigation of the relationship between soxC -expressing cells and MmpReg could provide valuable information.

SOXC transcription factors are widely conserved across animals 49 . SOX4 and SOX11 (in mammals, SOX4, SOX11, and SOX12 constitute the SOXC family) were highly expressed in embryonic epidermal cells and re-expressed in postnatal skin wounds in mice 60 . SOX4 and SOX11 regulate downstream cytoskeleton genes, one of which was required for cell migration 60 . In zebrafish, the result of a knock-down analysis suggested that sox11b regulates the migration of Muller glia-derived progenitor cells and their fate determination during retina regeneration 61 . During cerebral development in mice, Sox11 inhibition suppressed the radial migration of cortical neurons 62 . Overall, these results in addition to those inferred from the present study suggest that SOXC gene(s)-expression may be a characteristic of migrating cells across different contexts and animal species.

If SOXC gene(s)-expression reflects a characteristic of blastema cell mass across different animal species, similar expression dynamics in blastema formation would be found in other animals. We found Xl-sox4 and Xl-sox11 -expressing cells appeared to accumulate in the blastema in the regenerating tadpole tail. Our data point to a similarity in the expression dynamics of SoxC genes in the blastema formation in E. japonensis and X. laevis , raising the possibility of common cellular mechanisms underlying blastema formation in both species (Fig.  10b ). Recently, in the cnidarian polyp Hydra vulgaris , HvSoxC was found to be expressed in a transition state between i-cells and differentiated cells, which suggests that HvSoxC is a marker for cells in a transition state 63 . At present, whether HvSoxC -expressing cells migrate is unknown, and further studies are warranted to determine the migration capabilities of the HvSoxC -expressing cells. In planaria, scRNA-seq suggested that soxP-1 is expressed in multiple lineage progenitors 64 . In the acoel worm Hofstenia miamia , scRNA-seq suggested that sox-1 was expressed in the likely neural specialized neoblast population 65 . However, the relationship of soxP-1 in planaria or sox-1 in acoel worm and SOXC family needs to be verified. Notably, examining the expression dynamics of soxP-1 in planaria and sox-1 in acoal worm in blastema formation may provide important information. Further studies on the expression dynamics and behavior of SOXC gene(s)-expressing cells in various regeneration animal models would help understand whether the cellular mechanisms underlying blastema formation can be generalized among across species.

Enchytraeus japonensis 34 worms were reared in 1.0% agar in Holtfreter’s solution (15 mM NaCl, 0.2 mM KCl, 0.6 mM NaHCO 3 , 0.2 mM MgSO 4 7H 2 O, and 0.2 mM CaCl 2 ) in 150-mm disposable Petri dishes at 24 °C and fed with rolled oats (Quaker Oats). Under these conditions, E. japonensis grows continuously to approximately 10 mm in length and reproduces asexually by fragmentation approximately every 2 weeks. Artificial amputation of E. japonensis was performed using needle-sharp tweezers (Feather, K-715). The amputees were cultured on filter paper (Whatman No 3030917) and kept moist at 24 °C with Milli-Q water in 60 × 10 mm disposable Petri dishes. E. japonensis was provided by Dr. Yoshida-Noro, C. (Nihon University, Tokyo, Japan).

Niewkoop and Faber stage 66 51–53 Tadpoles ( X. laevis ) were purchased from a Japanese company (Watanabe Zoushoku). The tadpoles were maintained in tap water, which was left in a bucket overnight and filtered with canister filters (EHEIM Classic Model 2213). The tadpoles were fed with dumplings made of barely grass powder (Yamamoto Kampo Pharmaceutical Co.) and Ceramicron (Cera Japan Co.) mixed with Milli-Q water. The water and dumplings were replaced every day. Tail amputation was performed using fine razor blades after the tadpoles were anesthetized with ice. The amputated tadpoles were kept in the water at 20 °C and allowed to regenerate. The regenerating tail tissues including the blastema (about 5 mm long from the amputation site) were collected. These experiments were performed by following the recommendations of the Guidelines for Proper Conduct of Animal Experiments of the Science Council of Japan. The protocol was approved by the Committee on the Ethics of Animal Experiments of Teikyo University.

cDNA cloning and RNA probe preparation

Worms were cut into several fragments (normally 6–8) and allowed to regenerate for 24 h. Total RNA was extracted from approximately 100 regenerating fragments using TRIzol Reagent (Invitrogen) and reverse-transcribed using the SuperScript III kit (Invitrogen) using an oligo (dT) primer, by following the manufacturer’s protocol. RT-PCR was performed using the following gene-specific primer pairs:

soxC : (Forward) 5′-ATGATACTAAGTTCTAAAAT-3′ and (Reverse) 5′-TTAGTTCACCAATCTCTTA-3′;

mmpReg : (Forward) 5′-AACCAGTAACCAAGGCAACG-3′ and (Reverse) 5′-CCCTCATGATTTACGCCACT-3′;

Ej-vasa1 : (Forward) 5′-GCCCGTTGATCCTGATAAGA-3′ and (Reverse) 5′-TGTTACCACATCGCCCTGTA-3′;

Ej-vasa2 : (Forward) 5′- CTGGTAAAACGGCATCATTTCTC -3′ and (Reverse)5′-TTCTCCAGCCACTCCGGCAC-3′;

Ej-nanos : (Forward) 5′-GCTCGTTGGAATCGATTAGTG-3′ and (Reverse) 5′-CCCACACTGACTTGTGGTTG-3′;

Ej-pl10 : (Forward) 5′-TTCTGGCTGTGGGAAGAGTT-3′ and (Reverse) 5′-CTGCTCCTCGAGCCATTTAG-3′;

Ej-piwi : (Forward) 5′-GATCAAACAGCACACGGATG-3′ and (Reverse) 5′-CTTGGTCCCATCTTCTCTCG-3′.

The PCR products were subcloned into the pGEM-T easy vector (Promega), and the sequences were confirmed using Sanger sequencing. The plasmids containing the cDNA fragments for soxC, mmpReg, Ej-vasa1, Ej-vasa2, Ej-nanos, Ej-pl10 , and Ej-piwi were amplified using PCR with the M13 primer pair. The amplicons containing the T7 and SP6 promoter sites were purified using a PCR purification kit (Qiagen). Digoxigenin (DIG)-labeled sense and antisense RNA probes were prepared via in vitro transcription using the DIG RNA labeling kit (Roche). For double ISH analysis, Ej-vasa1 fluorescein-labeled sense and antisense RNA probes were prepared through in vitro transcription using the fluorescein RNA labeling kit (Roche). For the preparation of Xl-sox4S and Xl - sox11S RNA probes, template DNA corresponding to full-length Xl-sox4S (GenBank accession number: AF186844) or Xl-sox11S (GenBank accession number: D83650) cDNA with T7 and SP6 promoter sites were synthesized and purchased from Twist Biosciences (South San Francisco, CA, USA).

The worms were fixed overnight in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (pH 7.5, PFA–PBS) at 4 °C. For cryoprotection, the fixed samples were placed overnight in an 18% sucrose/PFA–PBS solution at 4 °C. Subsequently, the worms with sucrose substitution were embedded in Tissue-Tek OCT compound (Sakura Finetechnica) using a Cryomold (Sakura Finetechnical), frozen immediately on dry ice, and stored at −80 °C until sectioning was performed. The frozen worm blocks were cut into 12 µm-thick sections using a cryostat (CM3050S; Leica Biosystems). Ten serial sections were mounted on a glass slide, and the sectioned specimens were fixed in PFA–PBS. After incubation in 10 μg/ml of proteinase K in 10 mM Tris/HCl and 1 mM EDTA, the specimens were post-fixed for 10 min in PFA–PBS, treated with 0.2 M HCl for 10 min, washed in PBS, and treated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 7.5) for 1 h. Hybridization was performed at 70 °C for 16 h in ULTRAhybTM Hybridization Buffer (Ambion) containing 500 ng/ml of the probe. The hybridized specimens were washed twice in wash buffer (50% formamide, 2× saline-sodium citrate (SSC) buffer) at 70 °C for 10 min, subjected to RNase treatment (20 mg/ml) in TNE (10 mM Tris, 1 mM EDTA, and 0.5 M NaCl [pH 7.5]) for 10 min, and washed in TNE for 10 min. After stringent washes with a series of SSC buffers, the specimens were incubated at 25 °C for 1 h in 1.5% blocking reagent (Roche, cat. # 11096176001) in DIG buffer I (100 mM Tris, 150 mM NaCl, pH 7.5), followed by incubation at 25 °C for 30 min with the alkaline phosphatase (AP)-conjugated anti-DIG antibody (Roche, cat. # 11093274910, 1:1000 dilution) in 1.5 % blocking reagents. After incubation, the specimens were washed twice at 25 °C with DIG buffer I for 20 min each time. For signal visualization, a chromogenic reaction was performed with a nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP, Roche) at 25 °C for 18 h until the signal pattern levelled off. The sections were counterstained with nuclear fast red solution (ScyTek Laboratorie). Bright-field images of whole sections on each slide glass were semi-automatically captured using the NanoZoomer 2.0HT or NanoZoomer XR systems (Hamamatsu Photonics). The microscopic fields of interest were cropped using the NDP.view2 software (ver. 2.7.25; Hamamatsu Photonics). The cropped images were converted to 8-bit images, and their brightness and contrast were adjusted using ImageJ (ver. 1.52a; National Institutes of Health). Sense probes were used as negative controls.

ISH was performed in the regenerating tadpole tail tissues using the same protocol as that used for worms with minor modifications. The frozen blocks were cut into 10 µm-thick sections using a cryostat. Six serial sections were mounted alternatively on two glass slides. One slide glass was used for the antisense probe and the other was used for the sense probe.

The process from the re-fixation of sections to the chromogenic reaction with NBT/BCIP was performed as described earlier, except for hybridization. DIG-labeled RNA probes and fluorescein-labeled RNA probes were mixed and mounted simultaneously during hybridization. After the first chromogenic reaction with NBT/BCIP, the AP-conjugated anti-DIG antibody was detached using 100 mM glycine (pH 2.2). After washing in PBS, the fluorescein-labeled probes were immunohistochemically detected using the AP-conjugated anti-fluorescein antibody (Roche, cat # 11426338910, 1:1000 dilution). To visualize the signals, the second color chromogenic reactions were performed for 4 days at room temperature with SIGMAFAST Fast Red TR/Naphthol AS-MX tablets (Sigma-Aldrich, cat. # F4523). Sense probes were used as negative controls.

Transcriptome seq, assembly, and annotation

Total RNA was extracted from the intact group (20 worms), blastema-poor group (about 100 fragments, 24 hpa) and blastema-rich group (about 100 fragments, 24 hpa) using TRIzol. Total RNA was further purified using RNeasy Plus Mini Kit (Qiagen) and submitted to Macrogen Corp. ( https://www.macrogen-japan.co.jp/ ). The RNAs were purified by ribo-zero RNA removal kit (Illumina). Sequencing libraries were generated using the TruSeq Stranded mRNA LT Sample Prep Kit (Illumina). The libraries were sequenced on the NovaSeq 6000 platform, and 150 bp paired-end reads were generated. Raw sequences were deposited in the NCBI Short Read Archive (SRA) database ( http://www.ncbi.nlm.nih.gov/Traces/sra/ ) under the accession numbers: SRR21413950, SRR21413951, and SRR21413952. Trimmomatic 0.38 ( http://www.usadellab.org/cms/?page=trimmomatic ) was used to quality trim the reads using the default settings, and the remaining clean reads were assembled using Trinity version trinityrnaseq_r20140717, bowtie 1.1.2 ( https://github.com/trinityrnaseq/trinityrnaseq/wiki ) for transcriptome assembly without a reference genome. The longest transcript for each gene was selected as the unigene. For annotation analysis, the unigenes were searched against the KEGG, NCBI Nucleotide (NT), Pfam, Gene ontology (GO), NCBI non-redundant Protein (NR), UniProt, and EggNOG using BLASTN of NCBI BLAST version 2.9.0 ( https://blast.ncbi.nlm.nih.gov/Blast.cgi ) and BLASTX of DIAMOND version 0.9.21 ( https://github.com/bbuchfink/diamond ) software with an E-value default cut-off of 1.0E-5. The differentially expressed genes (DEGs) between the control (intact group) and regeneration groups (blastema-rich group or blastema-poor group) were identified using DEGseq analysis from the adjusted read count data. Statistical analyses were performed using fold change and Fisher’s exact test using edgeR per comparison pair. The significant DEGs were determined by setting the threshold of fold change to >2 and Fisher’s exact test at a raw p-value to <0.05. The unigenes were annotated based on BLASTX results, and the best alignments were used for downstream analyses. The GO and KEGG databases were used to predict the functions of the unigenes. Pathway analysis was performed using the KEGG Mapper tool ( http://www.genome.jp/kegg/tool/map_pathway2.html ). NCBI-gi numbers were applied as queries, and the acquired pathway search results were grouped under the KEGG pathway maps.

5′-rapid amplification of cDNA end (RACE) for mmpReg

5′-RACE was performed using the SMARTerRACE5′/3′ kit (Takara Bio) according to the manufacturer’s protocol. Briefly, after first-strand cDNA synthesis, PCR was performed using the gene-specific primer 5′- TCACCTCCTCGACCTTCTCCTGG -3′ and the universal primer included in the kit. Nested PCR was performed using the gene-specific primer 5′-CGGTACCCGGGGATCGCCATCCTGATGATAGCCTGATGAG -3′ and the universal primer short included in the kit. The gene-specific primer for nested PCR contained a sequence at its 5′ ends for cloning of RACE products (CGGTACCCGGGGATC). Cloning was performed using the In-Fusion Snap Assembly Master Mix (Takara Bio) according to the manufacturer’s protocol.

Phylogenetic analysis

Multiple sequence analysis was performed using Clustal X 2.1 [ https://clustalx.software.informer.com/2.1/ ] using complete sequences from the selected eukaryotic species. A phylogenetic tree was constructed using the Draw tree option in Clustal X 67 using the Neighbor-Joining algorithm with default settings (Gap opening:10, Gap extension:0.2, bootstrap number:1000), and the output file was analyzed using NJplot software 68 [ http://pbil.univ-lyon1.fr/software/njplot.html ]. The names of the genes and their accession numbers used for the phylogenetic analysis are listed in Supplementary Data  4 . The E. japonensis sequences have been submitted to the DNA Data Bank of Japan (DDBJ) nucleotide database (LC727633, Ej-mmpReg ; LC727634, Ej-nanos ; LC727635, Ej-piwi ; LC727636, Ej-pl10 ; LC727637, Ej-soxC : LC727638, Ej-vasa1 ; LC727639, Ej-vasa2 ; LC727632, Ej-gapdh ).

RNAi and inhibitor experiments

dsRNA was synthesized by following the protocol for RNAi in planarians with some modifications 69 . Briefly, DNA templates used for subsequent dsRNA synthesis were generated using PCR, and primers were designed to amplify limited regions of the cDNA sequence for each gene. Each primer contained a T7 promoter sequence at its 5′-end (TAATACGACTCACTATAGGGAGACCAC). The following PCR primers were used:

soxC : (Forward) 5′-TAAGCTGTGATGAGCTGAATCA-3′ and (Reverse) 5′-CACCAATCTCTTAATATTCCCA-3′;

GFP : (Forward) 5′-GTGCCCATCCTGGTCGAGCT-3′ and (Reverse) 5′-ACTTGTACAGCTCGTCCATGCC-3′.

dsRNA was synthesized using the MegaScript TM T7 kit (Invitrogen) by following the manufacturer’s instructions. The typical yield was approximately 50 μg of dsRNA per reaction. dsRNA (50 μg/100 μl) was diluted with 900 μl of Milli-Q water (5 μg/μl). Each worm was soaked in 5 μl of dsRNA in a 96-well plate at 24 °C. In all cases, the animals were transferred to a large volume of Milli-Q water and amputated 1 days after RNAi soaking. The amputated worms were allowed to undergo regeneration on filter paper that was kept moist with Milli-Q water. For the inhibitor experiment, MMP-2/MMP-9 inhibitor I (Cayman, CAS#193807-58-8) was used. The animals were soaked in 10 μM of inhibitor in 0.3% DMSO/PBS for 24 h at 24 °C, followed by amputation. The amputees were cultured on filter paper kept moist with 10 μM of inhibitor in 0.3% DMSO/PBS in Petri dishes at 24 °C.

Determination of amputation site

The amputation site was determined on the image at a site where the curvature of the body wall suddenly changes. Care was taken to avoid chloragocytes and intestines as much as possible from the amputation site to the tip (Supplementary Fig.  13 ).

Image analysis for evaluation of RNAi and inhibitor phenotype

The worms were fixed after 6 h or 24 h of amputation, and serial sections were subjected to ISH for soxC . We selected adjacent serial sections (three consecutive sections) in which signals were detected. The images, including blastema, were cropped and converted to 8-bit images using ImageJ software. The threshold for binarization was determined to appropriately measure the area of positive signals detected. The total area was calculated from the sum of the values obtained from three consecutive sections.

Fluorescent image acquisition and quantification

For EdU labeling, the Click-iT EdU Alexa Fluor 488 Imaging Kit (Life Technologies, cat. # C10337) was used by following the manufacturer’s protocol with minor modifications. Briefly, the amputated worms were allowed to regenerate for 3, 5, 8 or 24 h. The worms were exposed to EdU at a final concentration of 250 μM for the last 2 or 3 h. Then, the animals were fixed with PFA–PBS and incubated for 30 min in a freshly prepared EdU click-iT reaction cocktail. The nuclei were stained with DAPI (1/1000). The specimens were rinsed with PBS containing 0.1% Tween 20 several times and mounted on slide glasses with 90% glycerol/PBS. Imaging was performed using CellVoyager CQ1 (Yokogawa). Z-stack images were acquired at approximately 5-µm intervals (15 sections). The area containing the blastema was enclosed manually. The total number of EdU-incorporated cells or DAPI-labelled cells was measured using the Cell Pathfinder software (Yokogawa) according to the manufacturer’s protocol.

RT and qRT-PCR analysis

The worms were cut into several fragments (normally 6–8) and allowed to regenerate for different times for experiments. Total RNA was extracted from each pool of fragments (approximately 200 fragments) using TRIzol, and cDNA was synthesized from 1 μg of total RNA using a Reverse-Transcription Kit (Thermofisher). Quantitation of each gene product was performed using a LightCycler 96 System (Roche). Each reaction contained TB Green® Premix Ex Taq TM II (Qiagen), gene-specific primers, and the cDNA template and was subjected to PCR as follows: 95 °C for 30 s; 40 cycles of 95 °C for 5 s, 60 °C for 20 s; and 1 cycle at 65 °C for 15 s. The measurements were normalized to the expression level of Ej-gapdh . The mean of three replicate qRT-PCR assays was used for quantification. The following PCR primer pairs were used:

soxC : (Forward) 5′-TGCCAGCTTCTACCCTGAAGA-3′ and (Reverse) 5′-GGTGCATGGTGCCAAAACTT-3′;

mmpReg : (Forward) 5′-CGATGGGCCAGGTATGGTACT-3′ and (Reverse) 5′-ATGTCACCTCCTCGACCTTCTC-3′;

Ej-gapdh : (Forward) 5′-AGGATTGGAGAGGAGGCAGAA -3′ and (Reverse) 5′-CCGGTGGAGGATGGAATG -3′.

The tips of the tails were amputated from the tadpoles at st. 51–53 and allowed to regenerate for 0, 6, 24, and 48 h. The regenerating tail tissues including the blastema (about 5 mm long from the amputation site) were collected and pooled. Total RNA was extracted from each pool consisting of approximately 50 regenerating tail tissues. The measurements were normalized to the expression level of Xl-elf1α . The mean of four replicates was used for quantification. The following PCR primer pairs were used:

Xl-sox4 : (Forward) 5′- CATCAAGCGGCCAATGAAC -3′ and (Reverse) 5′- TGATTTTCCTCCGCTCGATCT -3′;

Xl-sox11 : (Forward) 5′- GGGCTCCCACTTCGACTTC -3′ and (Reverse) 5′- TGCGATCATCTCGCTGAGTT -3′;

Xl-elf1α : (Forward) 5′- GGAACGGTGACAACATGC -3′ and (Reverse) 5′- AGGCAGACGGAGAGGCTTA -3′.

Statistical analysis

Sample size (n) is indicated in the figure legends. Statistical analyses were performed using EZR software based on R (version 2.7.1) or R (version 4.1.2). The methods used for each analysis are presented in the corresponding sections. All tests performed were two-sided.

Reporting summary

Further information on research design is available in the  Nature Portfolio Reporting Summary linked to this article.

Data availability

Source data for graphs are provided with this paper. The raw sequences generated in this study were deposited in the NCBI Short Read Archive (SRA) database [ http://www.ncbi.nlm.nih.gov/Traces/sra/ ] under the accession numbers: SRR21413950 , SRR21413951 , and SRR21413952 . DNA sequences used for PCR in this study were deposited in the DNA Data Bank of Japan (DDBJ) nucleotide database under the accession numbers: Ej-mmpReg , LC727633 ; Ej-nanos , LC727634 ; Ej-piwi , LC727635 ; Ej-pl10 , LC727636 ; Ej-soxC , LC727637 ; Ej-vasa1 , LC727638 ; Ej-vasa2 , LC727639 ; Ej-gapdh , LC727632 .  Source data are provided with this paper.

Barresi, M. & Gilbert, S. Developmental Biology . (Oxford University Press 2019).

Morgan, T. H. Regeneration and Liability to Injury. Science 14 , 235–248 (1901).

Article   ADS   CAS   PubMed   Google Scholar  

Sanchez Alvarado, A. Regeneration in the metazoans: why does it happen? Bioessays 22 , 578–590 (2000).

Article   CAS   PubMed   Google Scholar  

Randolph, H. The regeneration of the tail in Lumbriculus. J. Morphol. 7 , 317–344 (1892).

Tanaka, E. M. & Reddien, P. W. The cellular basis for animal regeneration. Dev. Cell 21 , 172–185 (2011).

Article   CAS   PubMed   PubMed Central   Google Scholar  

Tsonis, P. A. Regeneration in vertebrates. Dev. Biol. 221 , 273–284 (2000).

Srivastava, M. Beyond Casual Resemblance: Rigorous Frameworks for Comparing Regeneration Across Species. Annu Rev. Cell Dev. Biol. 37 , 415–440 (2021).

Bely, A. E. & Nyberg, K. G. Evolution of animal regeneration: re-emergence of a field. Trends Ecol. Evol. 25 , 161–170 (2010).

Article   PubMed   Google Scholar  

Tornini, V. A. & Poss, K. D. Keeping at arm’s length during regeneration. Dev. Cell 29 , 139–145 (2014).

Currie, J. D. et al. Live Imaging of Axolotl Digit Regeneration Reveals Spatiotemporal Choreography of Diverse Connective Tissue Progenitor Pools. Dev. Cell 39 , 411–423 (2016).

Bassat, E. & Tanaka, E. M. The cellular and signaling dynamics of salamander limb regeneration. Curr. Opin. Cell Biol. 73 , 117–123 (2021).

Reddien, P. W. The Cellular and Molecular Basis for Planarian Regeneration. Cell 175 , 327–345 (2018).

Ozpolat, B. D. & Bely, A. E. Developmental and molecular biology of annelid regeneration: a comparative review of recent studies. Curr. Opin. Genet Dev. 40 , 144–153 (2016).

Srivastava, M., Mazza-Curll, K. L., van Wolfswinkel, J. C. & Reddien, P. W. Whole-body acoel regeneration is controlled by Wnt and Bmp-Admp signaling. Curr. Biol. 24 , 1107–1113 (2014).

Vogg, M. C., Galliot, B. & Tsiairis, C. D. Model systems for regeneration: Hydra. Development 146 , dev177212 (2019).

Funayama, N. The cellular and molecular bases of the sponge stem cell systems underlying reproduction, homeostasis and regeneration. Int J. Dev. Biol. 62 , 513–525 (2018).

Guedelhoefer, O. C. & Sanchez Alvarado, A. Planarian immobilization, partial irradiation, and tissue transplantation. J. Vis. Exp. 6 , 4015 (2012).

Google Scholar  

Salo, E. & Baguna, J. Cell movement in intact and regenerating planarians. Quantitation using chromosomal, nuclear and cytoplasmic markers. J. Embryol. Exp. Morphol. 89 , 57–70 (1985).

CAS   PubMed   Google Scholar  

Lapan, S. W. & Reddien, P. W. dlx and sp6-9 Control optic cup regeneration in a prototypic eye. PLoS Genet 7 , e1002226 (2011).

Abnave, P. et al. Epithelial-mesenchymal transition transcription factors control pluripotent adult stem cell migration in vivo in planarians. Development 144 , 3440–3453 (2017).

CAS   PubMed   PubMed Central   Google Scholar  

Seaver, E. C. Annelid models I: Capitella teleta. Curr. Opin. Genet Dev. 39 , 35–41 (2016).

Herlant-Meewis, H. Regeneration in annelids. Adv. Morphog. 4 , 155–215 (1964).

Cornec, J.-P., Cresp, J., Delye, P., Hoarau, F. & Reynaud, G. Tissue responses and organogenesis during regeneration in the oligochete Limnodrilus hoffmeisteri (Clap.). Can. J. Zool. 65 , 403–414 (1986).

Article   Google Scholar  

Stephan-Dubois, F. Migration and Differentiation of Neoblasts in Anterior Regeneration of Lumbriculus variegatus (Annelida, Oligochaeta). C. R. Seances Soc. Biol. Fil. 150 , 1239–1242 (1956).

Wagner, D. E., Wang, I. E. & Reddien, P. W. Clonogenic neoblasts are pluripotent adult stem cells that underlie planarian regeneration. Science 332 , 811–816 (2011).

Article   ADS   CAS   PubMed   PubMed Central   Google Scholar  

Hill, S. D. Origin of the regeneration blastema in polychaete annelids. Am. Zool. 10 , 101–112 (1970).

Martinez Acosta, V. G., Arellano-Carbajal, F., Gillen, K., Tweeten, K. A. & Zattara, E. E. It Cuts Both Ways: An Annelid Model System for the Study of Regeneration in the Laboratory and in the Classroom. Front Cell Dev. Biol. 9 , 780422 (2021).

Article   PubMed   PubMed Central   Google Scholar  

Tadokoro, R., Sugio, M., Kutsuna, J., Tochinai, S. & Takahashi, Y. Early segregation of germ and somatic lineages during gonadal regeneration in the annelid Enchytraeus japonensis. Curr. Biol. 16 , 1012–1017 (2006).

Myohara, M., Niva, C. C. & Lee, J. M. Molecular approach to annelid regeneration: cDNA subtraction cloning reveals various novel genes that are upregulated during the large-scale regeneration of the oligochaete, Enchytraeus japonensis. Dev. Dyn. 235 , 2051–2070 (2006).

Bhambri, A. et al. Large scale changes in the transcriptome of Eisenia fetida during regeneration. PLoS One 13 , e0204234 (2018).

Shao, Y. et al. Genome and single-cell RNA-sequencing of the earthworm Eisenia andrei identifies cellular mechanisms underlying regeneration. Nat. Commun. 11 , 2656 (2020).

Cho, S. J. et al. Gene expression profile in the anterior regeneration of the earthworm using expressed sequence tags. Biosci. Biotechnol. Biochem 73 , 29–34 (2009).

Takeo, M., Yoshida-Noro, C. & Tochinai, S. Functional analysis of grimp, a novel gene required for mesodermal cell proliferation at an initial stage of regeneration in Enchytraeus japonensis (Enchytraeidae, Oligochaete). Int J. Dev. Biol. 54 , 151–160 (2010).

Nakamura, Y. A new fragmenting enchytraeid species, Enchytraeus japonensis from a cropped Kuroboku soil in Fukushima, Northern Japan (enchytraeids in Japan 5). Edaphologia 50 , 37–39 (1993).

Myohara, M., Yoshida-Noro, C., Kobari, F. & Tochinai, S. Fragmenting oligochaete Enchytraeus japonensis: a new material for regeneration study. Dev. Growth Differ. 41 , 549–555 (1999).

Yoshida-Noro, C. & Tochinai, S. Stem cell system in asexual and sexual reproduction of Enchytraeus japonensis (Oligochaeta, Annelida). Dev. Growth Differ. 52 , 43–55 (2010).

Goldman, J. A. & Poss, K. D. Gene regulatory programmes of tissue regeneration. Nat. Rev. Genet 21 , 511–525 (2020).

Sarkar, A. & Hochedlinger, K. The sox family of transcription factors: versatile regulators of stem and progenitor cell fate. Cell Stem Cell 12 , 15–30 (2013).

Bretscher, M. S. On the shape of migrating cells-a ‘front-to-back’ model. J. Cell Sci. 121 , 2625–2628 (2008).

Zattara, E. E., Turlington, K. W. & Bely, A. E. Long-term time-lapse live imaging reveals extensive cell migration during annelid regeneration. BMC Dev. Biol. 16 , 6 (2016).

Okuda, S. et al. KEGG Atlas mapping for global analysis of metabolic pathways. Nucleic Acids Res 36 , W423–W426 (2008).

Bonnans, C., Chou, J. & Werb, Z. Remodelling the extracellular matrix in development and disease. Nat. Rev. Mol. Cell Biol. 15 , 786–801 (2014).

Juliano, C. E., Swartz, S. Z. & Wessel, G. M. A conserved germline multipotency program. Development 137 , 4113–4126 (2010).

Sugio, M. et al. Exploration of embryonic origins of germline stem cells and neoblasts in Enchytraeus japonensis (Oligochaeta, Annelida). Gene Expr. Patterns 8 , 227–236 (2008).

Sugio, M., Yoshida-Noro, C., Ozawa, K. & Tochinai, S. Stem cells in asexual reproduction of Enchytraeus japonensis (Oligochaeta, Annelid): proliferation and migration of neoblasts. Dev. Growth Differ. 54 , 439–450 (2012).

de Jong, D. M. & Seaver, E. C. Investigation into the cellular origins of posterior regeneration in the annelid Capitella teleta. Regenerat. (Oxf.) 5 , 61–77 (2018).

Kozin, V. V. & Kostyuchenko, R. P. Vasa, PL10, and Piwi gene expression during caudal regeneration of the polychaete annelid Alitta virens. Dev. Genes Evol. 225 , 129–138 (2015).

Ozpolat, B. D. & Bely, A. E. Gonad establishment during asexual reproduction in the annelid Pristina leidyi. Dev. Biol. 405 , 123–136 (2015).

Heenan, P., Zondag, L. & Wilson, M. J. Evolution of the Sox gene family within the chordate phylum. Gene 575 , 385–392 (2016).

Angelozzi, M. & Lefebvre, V. SOXopathies: Growing Family of Developmental Disorders Due to SOX Mutations. Trends Genet 35 , 658–671 (2019).

Tsujioka, H., Kunieda, T., Katou, Y., Shirahige, K. & Kubo, T. Unique gene expression profile of the proliferating Xenopus tadpole tail blastema cells deciphered by RNA-sequencing analysis. PLoS One 10 , e0111655 (2015).

Lai, A. G. & Aboobaker, A. A. EvoRegen in animals: Time to uncover deep conservation or convergence of adult stem cell evolution and regenerative processes. Dev. Biol. 433 , 118–131 (2018).

Reddien, P. W. et al. SMEDWI-2 is a PIWI-like protein that regulates planarian stem cells. Science 310 , 1327–1330 (2005).

Plickert, G., Frank, U. & Muller, W. A. Hydractinia, a pioneering model for stem cell biology and reprogramming somatic cells to pluripotency. Int J. Dev. Biol. 56 , 519–534 (2012).

Lim, R. S., Anand, A., Nishimiya-Fujisawa, C., Kobayashi, S. & Kai, T. Analysis of Hydra PIWI proteins and piRNAs uncover early evolutionary origins of the piRNA pathway. Dev. Biol. 386 , 237–251 (2014).

de Rosa, R., Prud’homme, B. & Balavoine, G. Caudal and even-skipped in the annelid Platynereis dumerilii and the ancestry of posterior growth. Evol. Dev. 7 , 574–587 (2005).

Zattara, E. E. & Bely, A. E. Evolution of a novel developmental trajectory: fission is distinct from regeneration in the annelid Pristina leidyi. Evol. Dev. 13 , 80–95 (2011).

Myohara, M. What role do annelid neoblasts play? A comparison of the regeneration patterns in a neoblast-bearing and a neoblast-lacking enchytraeid oligochaete. PLoS One 7 , e37319 (2012).

Gerber, T. et al. Single-cell analysis uncovers convergence of cell identities during axolotl limb regeneration. Science 362 , eaaq0681 (2018).

Article   ADS   PubMed   PubMed Central   Google Scholar  

Miao, Q. et al. SOX11 and SOX4 drive the reactivation of an embryonic gene program during murine wound repair. Nat. Commun. 10 , 4042 (2019).

Song, K. et al. Sox11b regulates the migration and fate determination of Muller glia-derived progenitors during retina regeneration in zebrafish. Neural Regen. Res 18 , 445–450 (2023).

Hoshiba, Y. et al. Sox11 Balances Dendritic Morphogenesis with Neuronal Migration in the Developing Cerebral Cortex. J. Neurosci. 36 , 5775–5784 (2016).

Siebert, S. et al. Stem cell differentiation trajectories in Hydra resolved at single-cell resolution. Science 365 , eaav9314 (2019).

Zeng, A. et al. Prospectively Isolated Tetraspanin(+) Neoblasts Are Adult Pluripotent Stem Cells Underlying Planaria Regeneration. Cell 173 , 1593–1608 e1520 (2018).

Hulett, R. E. et al. Acoel single-cell atlas reveals expression dynamics and heterogeneity of adult pluripotent stem cells. Nat. Commun. 14 , 2612 (2023).

Nieuwkoop D. P., Faber, J. Normal Table of Xenopus laevis (Daudin): A Systematical and Chronological Survey of the Development from the Fertilized Egg Till the End of Metamorphosis . Cold Spring Harbor Laboratory Press (1967).

Larkin, M. A. et al. Clustal W and Clustal X version 2.0. Bioinformatics 23 , 2947–2948 (2007).

Perriere, G. & Gouy, M. WWW-query: an on-line retrieval system for biological sequence banks. Biochimie 78 , 364–369 (1996).

Shibata, N. & Agata, K. RNA Interference in Planarians: Feeding and Injection of Synthetic dsRNA. Methods Mol. Biol. 1774 , 455–466 (2018).

Download references

Acknowledgements

We thank Dr. Takahiro Ito, Dr. Shinji Takada, and Dr. Yusuke Mii for critically reading the manuscript. We thank Dr. Mii for providing protocols and apparatus for raising the tadpoles. We would like to thank Editage ( www.editage.com ) for English language editing. This study was supported by a Teikyo University Research Encouragement Grant (T.F.), a Fund for the Promotion of Joint International Research (Fostering Joint International Research [B]) (T. F., K.J.H. 19KK0211), Grants-in-Aid for Scientific Research from the Japan Society for the Promotion of Science (S.Y., 24590096, 15K07945, 18K06667; N.A., 24790089, 20K06915; C.M., 20K16472; and K.J.H., 26440182, 17K07492, 20K06747), the Uehara Memorial Foundation (S.Y.), the Sagawa Foundation for Promotion of Cancer Research (S.Y.), and a Grant-in-Aid for Scientific Research on Innovative Areas Memory dynamism (26115522), Adaptive circuit shift (15H01449), and Evolinguistics (20H05012) from the Ministry of Education, Culture, Sports, Science and Technology (K.J.H.), the Naito Foundation (K.J.H.), and the Japan Foundation for Applied Enzymology (K.J.H.).

Author information

Authors and affiliations.

Department of Biological Sciences, Faculty of Pharmaceutical Sciences, Teikyo University, 2-11-1 Kaga, Itabashi-ku, Tokyo, Japan

Toshiyuki Fujita & Shinji Yamaguchi

Department of Molecular Biology, Faculty of Pharmaceutical Sciences, Teikyo University, 2-11-1 Kaga, Itabashi-ku, Tokyo, Japan

Naoya Aoki, Chihiro Mori & Koichi J. Homma

You can also search for this author in PubMed   Google Scholar

Contributions

T.F. and S.Y. designed the study and performed the experiments; T.F., N.A., C.M., K.J.H., and S.Y. analyzed the data and wrote the paper. All authors have reviewed the manuscript.

Corresponding author

Correspondence to Shinji Yamaguchi .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature Communications thanks José Martín-Durán and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information, peer review file, description of additional supplementary files, supplementary data 1, supplementary data 2, supplementary data 3, supplementary data 4, reporting summary, source data, source data, rights and permissions.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

Reprints and permissions

About this article

Cite this article.

Fujita, T., Aoki, N., Mori, C. et al. SoxC and MmpReg promote blastema formation in whole-body regeneration of fragmenting potworms Enchytraeus japonensis . Nat Commun 15 , 6659 (2024). https://doi.org/10.1038/s41467-024-50865-1

Download citation

Received : 03 October 2022

Accepted : 22 July 2024

Published : 22 August 2024

DOI : https://doi.org/10.1038/s41467-024-50865-1

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

By submitting a comment you agree to abide by our Terms and Community Guidelines . If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

• Developmental Biology
• Stem Cells and Regenerative Medicine

Single-cell RNA sequencing of the holothurian regenerating intestine reveals the pluripotency of the coelomic epithelium

Joshua g medina-feliciano, griselle valentín-tirado, kiara luna-martínez, yamil miranda-negrón.

• José E García-Arrarás author has email address
• Department of Biology, University of Puerto Rico, San Juan, Puerto Rico
• https://doi.org/ 10.7554/eLife.100796.1
• Open access
• Copyright information

In holothurians, the regenerative process following evisceration involves the development of a “rudiment” or “anlage” at the injured end of the mesentery. This regenerating anlage plays a pivotal role in the formation of a new intestine. Despite its significance, our understanding of the molecular characteristics inherent to the constituent cells of this structure has remained limited. To address this gap, we employed state-of-the-art scRNA-seq and HCR-FISH analyses to discern the distinct cellular populations associated with the regeneration anlage. Through this approach, we successfully identified thirteen distinct cell clusters. Among these, two clusters exhibit characteristics consistent with putative mesenchymal cells, while another four show features akin to coelomocyte cell populations. The remaining seven cell clusters collectively form a large group encompassing the coelomic epithelium of the regenerating anlage and mesentery. Within this large group of clusters, we recognized previously documented cell populations such as muscle precursors, neuroepithelial cells and actively proliferating cells. Strikingly, our analysis provides data for identifying at least four other cellular populations that we define as the precursor cells of the growing anlage. Consequently, our findings strengthen the hypothesis that the coelomic epithelium of the anlage is a pluripotent tissue that gives rise to diverse cell types of the regenerating intestinal organ. Moreover, our results provide the initial view into the transcriptomic analysis of cell populations responsible for the amazing regenerative capabilities of echinoderms.

eLife assessment

This study presents a dataset obtained through a single cell RNA-Sequencing of sea cucumber regenerating intestine 9 days post evisceration. The data were collected and analyzed using standard single cells analysis from n=2 adult sea cucumbers captured from the wild, which represents a useful resource for future studies. Although cell type validation is attempted, it is performed on samples from the same 2 animals (and not independent samples), rendering the validation incomplete . Further, the RNA localization images provided in the paper could benefit from improved spatial context, and many strong statements in the discussion should be better justified and supported by the presented data. With the validation part strengthened, this paper would be of interest to development and regeneration fields.

• https://doi.org/ 10.7554/eLife.100796.1.sa4
• Read the peer reviews
• About eLife assessments

Introduction

Animals exhibit a remarkable diversity in their responses to injury, ranging from basic wound healing to the complete regeneration of lost structures. At one end of the spectrum are species that heal wounds without regenerating the missing part, while at the other end are those capable of recreating structures identical to the original. Despite these differences, all animals possess some level of regenerative ability. This capacity for regeneration can vary not only between different species but also within the same species, depending on the tissue or organ involved.

For decades, scientists have tried to understand these vast differences in regeneration capacities across the animal kingdom. For this they have focused on those species that show amazing regenerative abilities ( Tanaka and Reddien, 2011 ), including coelenterates (hydra) ( Vogg et al., 2019 ), flatworms (planaria) ( Reddien, 2018 ), fish (zebrafish) ( Gemberling et al., 2013 ) and amphibians (axolotl) ( Roy and Gatien, 2008 ) among others. These studies have uncovered various mechanisms that regeneration-competent species exhibit to regenerate tissues, entire organs, body parts, and in some cases complete bodies. Key findings include the discovery of essential processes, such as the formation of a blastema – a mass of proliferating cells that plays a crucial role in regenerating the lost structure ( Min and Whited, 2023 ; Poleo et al., 2001 ; Santos-Ruiz et al., 2002 ; Seifert and Muneoka, 2018 ; Stocum, 2004 ; Zenjari et al., 1996 ).

Among deuterostomes, echinoderms are considered prime exponents of regenerative capability ( Candia-Carnevali et al., 2024 ). Within this group, holothurians, commonly known as sea cucumbers, exhibit an extraordinary form of regeneration. They can regenerate their internal organs following evisceration, a process in which they expel their viscera in response to stress or predation ( Byrne, 2023 ). This extraordinary ability makes them a valuable model for studying regeneration in complex organisms. The regeneration of the digestive system in holothurians, in particular, has garnered significant interest ( Dolmatov, 2021 ; Mashanov et al., 2014 ; Medina-Feliciano and García-Arrarás, 2021 ; Quispe-Parra et al., 2021 ; Su et al., 2022 ). Upon evisceration, the holothurian intestine, which constitutes nearly their entire digestive tract, begins to regenerate from the mesentery, a supportive tissue layer where the original intestine was attached ( García-Arrarás et al., 1998 ). A thickening at the injured end of the mesentery, known as an “anlage” or “rudiment,” initiates this process ( García-Arrarás et al., 1998 ). This structure is analogous to a blastema but differs in that the cell proliferation mainly occurs in the surrounding epithelium, rather than in the mesenchymal cells, as observed in classical blastemas ( Carlson, 2007 ; García-Arrarás et al., 1998 ).

Histologically, the holothurian intestinal anlage forms from dedifferentiated cells within the mesentery, which revert to a more stem-cell-like state before proliferating and migrating to form a new intestinal structure ( Candelaria et al., 2006 ; Mashanov et al., 2005 ). This dedifferentiation process is crucial for regeneration, involving a spatial and temporal gradient starting at the injury site and extending along the mesenteric border ( García-Arrarás et al., 2011 ). The new epithelial layer that forms around the anlage is distinct from the original mesothelium, shows significant morphological and molecular changes compared to the mesenteric tissue ( Candelaria et al., 2006 ; García-Arrarás et al., 1998 ; Mashanov et al., 2005 ). Further examination of holothurian regeneration reveals that most cellular division occurs within the anlage’s coelomic epithelium ( García-Arrarás et al., 2011 , 1998 ). These proliferating cells are hypothesized to differentiate into various cell types, including myocytes and neurons, as well as mesenchymal cells, through an epithelial to mesenchymal transition (EMT) ( García-Arrarás et al., 2011 , 1998 ). The gene expression profiles during the formation and growth of the anlage suggest extensive reprogramming, leading to a more plastic cell phenotype ( Ortiz-Pineda et al., 2009 ; Quispe-Parra et al., 2021 ; Rojas-Cartagena et al., 2007 ).

The process of intestinal regeneration in holothurians raises fundamental questions of regenerative phenomena particularly concerning the identity, origin, and fate of progenitor cells, involved in the process ( Alvarado and Tsonis, 2006 ; Candia-Carnevali et al., 2024 ). In this context, the role of the anlage, and specially the mesothelium (also named celothelium) in echinoderms, deserved particular attention ( Candia-Carnevali et al., 2024 ; Smiley, 1994 ). This tissue, mainly composed of coelomic epithelia and myocytes, exhibit significant morphological and gene expression changes that are associated with the dedifferentiation process. These dedifferentiated cells form the coelomic epithelium of the anlage and appear to be the principal source of cells for the new intestine. Despite these findings, little is known about the cell composition and dynamics of the anlage nor of its coelomic epithelium.

Single-cell omic studies offer tremendous promise to dissect the cellular contributions of the holothurian intestinal anlage and to identify the specific cells involved in generating a new intestine. Single-cell RNA sequencing (scRNA-seq) a powerful tool for dissecting cellular composition and dynamics has been used in related species to explore regenerating or developing tissues. In mice, for instance, scRNA-seq has been employed to investigate the response to injury. However, these studies have primarily focused on tissue healing rather than the full regenerative process ( Ayyaz et al., 2019 ; Beumer and Clevers, 2021 ; Capdevila et al., 2021 ; Parigi et al., 2022 ). Beyond healing, scRNA-seq has been pivotal in studying regenerating systems such as the axolotl limb blastema and planarian regeneration providing a detailed insights into the cellular processes at play ( Gerber et al., 2018 ; King et al., 2024 ; Leigh et al., 2018 ; Rodgers et al., 2020 ). Furthermore, considering the close relationship between regeneration and embryogenesis, scRNA-seq, has also been used to understand the cellular events in developing echinoderm embryos ( Cocurullo et al., 2023 ; Paganos et al., 2022a , 2022b , 2021 ; Satoh et al., 2022 ; Tominaga et al., 2023 ). These applications highlight the groundbreaking role of scRNA-seq in advancing our knowledge of the cellular mechanisms in both regenerative and developmental contexts.

In this study, we employ scRNA-seq to analyze the regenerating intestinal anlage of the sea cucumber Holothuria glaberrima , aiming to delineate its constituent cellular populations ( Fig 1A ). We corroborate our findings using hybridization chain reaction fluorescent in situ hybridization (HCR-FISH) to verify the presence and location of specific cell types ( Choi et al., 2018 ). The results provides transcriptomic evidence on the pluripotency of the anlage coelomic epithelia and the first description of the holothurian regenerating precursor cells. This research not only advances our understanding of the unique regenerative capabilities of holothurians but also contributes to the broader field of regenerative biology, highlighting the diverse strategies employed by different organisms to restore lost tissues.

Overview of single cell RNA sequencing of regenerating intestinal tissue of H. glaberrima .

(A) Diagram of 9-day regenerating intestine depicting mesentery and anlage components (B) UMAP plot of population identities determined through unsupervised clustering of 9-day regenerating mesentery and anlage tissues. (C) Representation of cells per tissue of both samples. (C) UMAP projections of cluster cells separated by tissue of origin. (D) Percentage of cells per cluster based on their tissue of origin.

Previous work from our laboratory has shown that the rudiment or anlage that forms at the tip of the mesentery plays a pivotal role in the formation of the new intestine ( García-Arrarás et al., 2019 ). This transient mass of cells is thought to give rise to most intestinal cell types, the sole exception being the luminal cells. Therefore, to maximize the characterization of the cells in the regenerative anlage, we chose to perform scRNA-seq in the tissues of 9-days post evisceration (dpe) regenerating animals ( García-Arrarás et al., 2011 , 1998 ). At this stage a well-formed anlage consists of epithelial cells surrounding a large area of connective tissue populated with mesenchymal cells ( Fig 1A ). More importantly, different cell populations undergoing proliferation or differentiation can be found at this stage. Since some of the cellular processes during regeneration occur in a spatio-temporal gradient along the length of the mesentery, we separately surveyed the anlage and the mesentery tissues accordingly. Thus, for each animal, the anlage was separated from the mesentery, and both tissues were processed independently, for a total of four scRNA-seq runs.

Tissues were treated with 0.05% trypsin for 15 min and cells were dissociated manually. The dissociated cells were counted and used for the scRNA-seq analyses. Additionally, samples from the cell dissociation were fixed in poly-lysine treated slides and analyzed with immuno- and cyto-chemical techniques. Therefore, these cells correspond to the same batch whose transcriptomes were sequenced.

Cell Heterogeneity: Immuno- and cyto-chemical analyses

The strength of the scRNA-seq data depends mainly on the dissociation and isolation of the cell populations from the dissected tissue. Since our focus was on the cells of the regenerating anlage, we devised a dissociation protocol that favored the isolation of the cells within this structure. To determine, at least partially, the cell types in our original dissociation, we performed immuno- and cyto-chemical analyses on the enzyme-dissociated cell suspension that was used for the scRNA-seq. The labeling obtained for each marker is shown in S1 Fig .

We found that, in the dissociated anlage, 5% of the cells were labeled with a spherulocyte (immune cell) marker, 2-7% with phalloidin (a muscle marker), 40% with a mesenchymal marker (KL14-antibody), and a large number of cells (32-60%) with a mesothelial marker (MESO-antibody). Similar populations were found in the mesentery, although in this tissue, 5-10% of cells expressed the neuronal markers, heptapeptide GFSKLYFamide ( Diaz-Miranda et al., 1995 ) and RN1 ( Diaz-Balzac et al., 2007 ). This analysis suggested that most of the cells originate from the regenerate coelomic epithlium.

Different populations were also observed with three different tubulin antibodies ( S2 Fig ). In the 9-dpe anlage, anti-acetylated alpha tubulin labeled about 7% of the cells, anti-beta tubulin labeled about 70% of the cells, while an anti-alpha tubulin labeled about 80% of the cells. Except for the acetylated alpha tubulin, which labeled about 1% of the cells dissociated from the mesentery versus 7% of those from the anlage, anti-alpha and beta labeling percentages were similar in both mesentery and anlage dissociated cells.

Finally, a cell population labeled with fluorescently-labeled phalloidin accounted for about 7% of the cells in both mesentery and anlage. These cells, however, did not correspond to the elongated muscle cells of the mesentery. Instead, they were rounded cells with labeling found in the cytoplasm surrounding one side of the nuclei.

Two cell populations from the mesentery are not represented or are greatly underrepresented in the scRNA-seq. Firstly, the muscle cells of the mesentery, due to their non-dissociation by the protocol used, were absent in the dissociated cell suspension. Their elongated morphology would have further complicated their passage via cell separation system for sequencing. Secondly, the majority of neurons from the neuronal network associated with the mesentery could not be dissociated ( Nieves-Ríos et al., 2020 ). We did, however, observe structures that resembled a tangled mass of cells immunoreactive to some of our neuronal markers. This suggests that the mesentery nervous component, being unable to be isolated as single cells, was not sequenced.

In summary, the immuno- and cytochemical results show that the dissociated cell populations sequenced correspond to cellular phenotypes that have been previously described within the regenerating anlage. The abundance of these cells in the sequenced samples corresponds to the ease of their dissociation by trypsin. Thus, dedifferentiated cells of the mesentery and anlage epithelium (which are loosely connected to each other) and those of the connective tissue are probably over-represented compared to differentiated cell types.

Cell populations defined by scRNA-seq

Analysis of scRNA-seq data resulted in a total of 3,844 cells, with 2,392 originating from the two anlage samples and 1,452 from the two mesentery samples. Upon dataset integration and graph-based clustering, we identified 13 clusters, each thought to represent singular cell types or cell states in the regenerating intestine ( Fig 1A-B ). The percentage of cells that form each cluster differs from cluster to cluster, ranging from 21% (cluster 0) to 1% (cluster 12) of the total cells. Nonetheless, each cluster consists of cells from both the mesentery and anlage samples ( Fig. 1C & D ). The number of clusters did not change dramatically under various parameters (resolution, dimensionality, and number of variable features), constantly around 12-15 clusters. Moreover, except for C3 and C4, the clusters are supported by the clustering significance analysis performed with scSHC ( Grabski et al., 2023 ), a model-based hypothesis testing method for scRNA-seq that evaluates the probability of each individual cluster being unique ( S3 Fig ). The uniqueness of C3 and C4 is suggested by additional analyses as will be addressed below.

Each identified cluster exhibits a distinctive gene expression profile relative to cells in other clusters ( Fig 2 ). The top expressed gene serves as a marker for each cell population to highlight the uniqueness of each cluster. In Fig 2 , we show the relative expression of the top gene for each cluster based on two factors: (1) difference in percentage of representation and (2) log 2 fold-change (log 2 FC) against all other clusters. Interestingly, each cluster shows dramatic differential expression values (> 2 log 2 FC) and differences in representation percentages over 50%, except for C0 through C3, with differences in representation around 30%.

UMAP visualization of clusters highlighting the expression of their top genes.

Each gene corresponds to the top gene of each independent cluster based on the percentage of representation to other clusters.

Prior to characterizing each of the 13 cell clusters, we sought to understand what, in a broad view, appears to be a segregation of ∼90% of the cells into two distinct supra-clusters. One of them encompassing 7 clusters (C0, C1, C3, C4, C5, C8, C9) that corresponded to 69.6% of all cells and the other encompassing two clusters (C2 and C7) that corresponded to 19.1% of cells. The remaining 11.3% of cells were distributed in 4 distinct isolated clusters (C6, C10, C11, and C12). As stated earlier, all clusters have representation from mesentery and anlage tissues ( Fig 1C - D ), thus excluding the possibility that the two supra-clusters represented mesentery versus anlage cells.

Mesenchymal versus Epithelial Clusters

These two supra-clusters are of interest as they appear to represent the two main cell types found in the regenerating intestine: coelomic epithelial cells and mesenchymal cells ( Fig 1B ). Understanding the cell types that correspond to each cluster is essential as cellular processes crucial for regeneration are thought to be localized to the coelomic epithelia of the regenerating anlage. Comparison between these two supra-clusters showed distinct expression profiles that allowed us to characterize their cell types. For instance, the supra-cluster composed of C2 and C7, showed ERG (transcriptional regulator ETS-related gene) as the top expressed gene, an oncogene that is related to embryonic development, differentiation, angiogenesis, and apoptosis ( Dhordain et al., 1995 ; Iwamoto et al., 2001 ; Vlaeminck-Guillem et al., 2000 ) ( Fig 2 ). More importantly, the expression of ERG has been associated with mesenchymal identity in other echinoderms ( Meyer et al., 2023 ; Tominaga et al., 2023 ). In these studies, ERG has been reported as the marker gene of mesenchymal cells of sea urchin and sea star larva by scRNA-seq analyses and the localization of ERG on embryonic precursor mesenchymal cells of the sea urchin was further confirmed by in situ analyses ( Meyer et al., 2023 ). In addition, this latter group also reported these mesenchymal cells to express a GATA transcription factor ( GATA3 ) and ETS1 , which in our dataset are also being expressed only by populations within this supra-cluster ( GATA2 and ETS1 – Fig 3B ). The gene ETS1 has been associated with the differentiating sea urchin larvae mesenchymal cells and has a vital role in EMT and cellular invasion of mesenchymal cancer cells ( Gluck et al., 2019 ; Koga et al., 2010 ; Meyer et al., 2023 ). The expression of PRG4 (proteoglycan-4) reinforces the mesenchymal identity of this supra-cluster in the regenerating intestine ( S4 Fig ), which correlates with previous studies that have identified a proteoglycan-like molecule in the mesenchyme of 7-dpe regenerating intestine ( Vázquez-Vélez et al., 2016 ).

Cluster characterization by gene expression.

(A) Top-expressed genes or (B) genes corresponding to transcription factors and intercellular signaling molecules are identified in the 13 cell clusters. Clusters are classified by their corresponding cell type, where blue corresponds to cells of the coelomic epithelium, red to those in the mesenchyme and green to coelomocytes. Color intensity shows the expression level of each gene in log2fold-change (log2FC) values. Dot size corresponds to percentage of representation of the gene in the respective cluster compared to all others. Gene identifiers starting with “g” correspond to uncharacterized gene models of H. glaberrima. (C) UMAP plot of clusters colored by cell type.

In contrast, the top marker gene in the other supra-cluster is AHNK , known as neuroblast differentiation-associated protein AHNAK. Reports have shown AHNK to have a role in calcium regulation, cellular migration, and carcinogenic transformation of colon epithelial cells ( Dumitru et al., 2013 ). Furthermore, this gene is overexpressed in regenerating rat muscle compared to normal muscle ( Huang et al., 2007 ). However, the localized expression of WNT9 in the C1, C3, C5, and C8 of this supra-cluster more clearly favors its classification as a marker for coelomic epithelial cell types ( Fig 3B and S4 Fig ). A previous study from our laboratory using in situ hybridization, details the expression of WNT9 during intestinal regeneration in H. glaberrima, where it was shown to be localized to the coelomic epithelium of the anlage and adjacent mesentery ( Mashanov et al., 2012 ). In addition, correlating with what was shown by the in situ hybridization results, the population of cells that differentially expresses WNT9 makes 15% of those in the mesentery. However, it is close to 30% of the cells in the anlage. Additional analyses, discussed below, further strengthens the coelomic epithelium identity of cells in this supra-cluster.

In summary, results show 13 individual cell clusters with distinct expression profiles in the regenerative intestinal tissue. Moreover, our results suggest a principal separation of clusters based on whether they correspond to those of the mesenchyme or to cells from the coelomic epithelia layer. The rest of the populations show top expressed genes that are immune-related suggesting that these must be coelomocyte populations ( Fig 2 ).

Cluster Identities

Rather than considering single genes, the uniqueness of each cluster can be assessed in terms of their transcriptomic profile ( Fig 3A ) and by the expression of transcription factors and intercellular signaling molecules ( Fig 3B ). As shown in Fig 3A-B , each cluster has a unique transcriptomic profile. Nonetheless, some clusters show their close relationship (epithelial vs mesenchymal) based on shared gene expression or representation. A notable example is clusters C0, C1, C3, and C4 of the epithelial layer populations, which share many top genes, albeit at different expression levels. Similarly, C8 and C9 have low expression of genes expressed by C0, C1, C3 and C4 but have expression of other genes that are not expressed by any other cluster. This is also true for the mesenchymal clusters, where these two clusters have overlap of some genes, that are not expressed by other clusters. Therefore, these results show the interaction of these clusters and their transcriptomic relationship depending on where they are localized within the regenerating intestinal tissue.

Identifying the top genes expressed by each cluster also provides essential, notwithstanding limited, information for their complete identification. To further characterize each cluster, we have used other analyses or performed additional experiments, including (1) multiple gene expression patterns, (2) enriched ontology, (3) HCR-FISH and (4) pseudo-trajectory. Initially, we describe in depth the potential identity of these clusters based on their expression pattern and enriched ontology, and we will end with their potential interactions.

An in-depth analysis of the various clusters and their possible relation to cell populations

Coelomocyte populations.

Coelomocytes are specialized cells found in the coelomic fluid and within organs of echinoderms. These cells have been associated with immunological roles including pathogen recognition, encapsulation, phagocytosis, debris removal, cytokine production and secretion among others (Barela Hudgell et al., 2022; Courtney Smith et al., 2018; García-Arrarás et al., 2006 ; Smith et al., 1995 ), In holothurians, these cells have been shown to be present at injury sites and in the regeneration anlage. Coelomocytes can be subdivided into different populations by using morphological, physiological, and molecular characteristics ( Ramírez-Gómez et al., 2010 ).

Many of the coelomocyte characteristics correlate with the top differentially expressed genes of cell populations within our data. These immune-like clusters (C6, C10, C11, and C12) have high expression of genes related to the immune system that are not shared with any other cluster. For example, C6 embodies a distinct cell population that represents a substantial number of cells (6% of all cells, 8% of the mesentery and 4% of the cells in the anlage). These cells are the only population expressing FBCD1 (fibrinogen C domain-containing protein 1). Cells in C6 also express other genes such as various tyrosine protein phosphate receptors, integrin alpha-8 ( ITA8 ), platelet glycoprotein V ( GPV ), leucine rich repeats and immunoglobulin-like domains protein 2 ( LRIG2 ) ( Fig 4A ). The immune identity of C6 is also supported by the resulting gene set enrichment of gene ontology (gseGO) terms related to immune responses, such as ubiquitin-dependent ERAD pathway , innate immune response activating cell surface receptor signaling pathway , respond to endoplasmic reticulum stress , phagocytosis and many more related to defense mechanisms ( Fig 4B ). Similarly, C10 and C12 transcriptomic profiles suggest they correspond to immune-like populations ( Fig 4C ). Specifically, C10 shows various uncharacterized genes along with HCK (tyrosine-protein kinase HCK), DMBT1 (deleted in malignant brain tumors 1), ALS (insulin-like growth factor binding protein complex acid labile subunit), GPV , and IRF8 (interferon regulatory factor 8), FER (tyrosine-protein kinase Fer). The top GO enriched terms of this cluster support its involvement in immune process, some of these being: lipase activity regulation , Fc receptor mediated stimulatory signaling pathway , cellular pigmentation , B cell activation involved in immune response , and Fc receptor signaling pathway . The transcripts expressed by C12 show a more complex profile, where most of the top genes are uncharacterized. Nonetheless, among the annotated genes are BAR3 (Balbiani ring protein 3), LYS1 (lysozyme 1), TRPM3 (transient receptor cation channel subfamily M member 3), PGCA (aggrecan core protein) and MRC1 (macrophage mannose receptor 1). The top genes of C11 include a great number of immune genes such as MUC5A (mucin-5AC), FCGBP (IgGFc-binding protein), SSPO (SCO-spondin), FCN1A (ficolin-1-A), MUC5B (mucin-5B), and TIE1 (tyrosine-protein kinase receptor Tie-1). However, the top genes of this cluster also include several genes involved in neuronal activity, which is evident in the top enriched GO terms of this cluster that include regulation of postsynaptic membrane potential , excitatory postsynaptic potential , chemical synaptic transmission , endoplasmic reticulum to Golgi vesicle-mediated transport , adult behavior , and regulation of neurotransmitter levels ( Fig 4A-B ).

Expression profile of coelomocyte cell types.

(A) Dot plot showing the top expressed genes of C6, C10, C11, and C12. Color intensity reflects the expression levels in log2FC values. Dot size correspond to the percentage of representation of each gene in the respective cluster compared to all others. (B) Top enriched GO biological processes terms of C6, C10, C11, and C12. Dot size corresponds to the enrichment score while the color reflects the adjusted p-value. (C) UMAP highlighting the coelomocyte populations. (D-E) HCR-FISH for FBCD1 in holothurian intestinal tissues. Cells expressing FBCD1 mRNA in the connective tissue layer of (D) regenerating mesentery and (E) normal intestine of H. glaberrima are shown. Red = DAPI stain, green = HCR-FISH. Bar= 10 μm.

To partially confirm the prediction that cells from C6, C10, C11 and C12 corresponded to coelomocyte populations, we used HCR-FISH to identify the cell types expressing the top gene in one of the clusters. We focused on the expression of FBCD1 , the gene that is the most represented by cells of C6. In situ hybridization identified a distinct cell type in both regenerating and non-regenerating tissues ( Fig 4D-E ). These cells showed round or oval morphologies with a central round nucleus. In some cases, short extensions could be observed. The cells were heterogeneously distributed in all tissues, including the nervous system and the body wall, and could be found associated with either epithelial tissues or with the extracellular matrix (ECM) in the normal intestine, the intestinal anlage and in the mesentery of normal or regenerating animals. The widespread distribution of this cell type hinted at a cell function consistent with patrolling the body to detect and respond to potential threats such as injury or bacterial invasion.

Mesenchymal cell populations

Mesenchymal cells of the intestinal anlage are yet to be well studied. They are known to be less proliferative than those in the overlying epithelium ( García-Arrarás et al., 2011 , 1998 ) and involved in ECM remodeling ( Quiñones et al., 2002 ). Some of the mesenchymal cells are thought to migrate from the connective tissue of the mesentery ( Cabrera-Serrano and García-Arrarás, 2004 ), while others have been shown to originate via EMT from the overlying epithelium ( García-Arrarás et al., 2011 ). Some of these cells will form a mesenchymal cellular layer associated with the luminal epithelial cells as the lumen forms ( García-Arrarás et al., 2011 ). As explained earlier, we have proposed that C2 and C7, form a separate supra-cluster, clearly identified by the increased expression of ERG with 5 of log 2 FC and a 90% representation ( Fig 2O ). These two clusters also share many marker genes, including multiple ECM genes ( Fig 5D ). For instance, highest expressed gene of C2 is TIMP3 (metalloproteinase inhibitor 3), followed by NPNT (nephronectin), which has been reported to be an integrin ligand during kidney development ( Sun et al., 2018 ). Additionally, this cluster has overexpression of ECM2 (extracellular matrix protein 2), KLKB1 (plasma kallikrein), MMP14 (matrix metalloproteinase 14), MMP24 (matrix metalloproteinase 24), HMCN1 (hemicentin-1), and ITA8 (integrin alpha-8), all of which are explicitly related to extracellular matrix component ( Bokel and Brown, 2002 ; Dolmatov and Nizhnichenko, 2023 ; Dong et al., 2006 ; Stamenkovic, 2003 ; Volkert et al., 2014 ). Many of these genes are also highly represented in C7. However, here we also find ITIH2 (inter-alpha-trypsin inhibitor heavy chain H2), DUOX1 (dual oxidase 1), SVEP1 (sushi, von Willebrand factor type A, EGF and pentraxin domain-containing protein 1), SEPP1 (selenoprotein P), and KLH20 (Kelch-like protein 20) suggesting that cells have gained some specialization and are more advanced in their differentiation when compared to those of C2. Along with this, when analyzing the gseGO, for C2 we obtain GO terms of numerous ECM processes, such as cell adhesion mediated by integrin , integrin-mediated signaling pathway , and regulation of cell-substrate junction assembly ( Fig 5F ). Similarly, C7 has an enrichment of regulation of cell-matrix adhesion , regulation of cell-substrate junction organization , and substrate adhesion-dependent cell spreading . Notably, these clusters also show enrichment of processes involved in cell proliferation, growth, and wound healing. To mention a few for C2 we find B cell differentiation , vascular endothelial growth factor receptor signaling , regulation of myeloid leukocyte differentiation, and negative regulation of myeloid cell differentiation , while C7 shows an enrichment of positive regulation of wound healing , regulation of neural precursor cell proliferation , positive regulation of response to wounding , humoral immune response , neuroblast proliferation, among others. This is not surprising considering that this tissue is undergoing active regeneration which requires growth and healing events to be active at distinct stages of the process.

Expression profile of mesenchymal cell types.

(A) Dot plot showing the top expressed genes of C2 and C7. Color intensity reflects expression levels in log2FC values. Dot size correspond to the percentage of representation of each gene in the respective cluster compared to all others. (B) Top GO enriched terms of biological processes of C2 and C7. Dot size corresponds to the enrichment score while the color reflects the adjusted p-value. (C) UMAP highlighting the clusters that correspond to the mesenchymal population (D-F) Cells expressing HMCNT1 mRNA in the connective tissue layer of (D) regenerating mesentery, (E) intestinal anlage and (F) normal intestine of H. glaberrima . Red = DAPI stain, green = HCR-FISH. Bar= 10μm

To confirm our prediction that cells in C2 and C7 were those present within the mesenchyme, we chose to localize the expression of HMCNT1 mRNA. The mRNA for this protein, known to code for an extracellular protein ( Dolmatov and Nizhnichenko, 2023 ; Lindsay-Mosher et al., 2020 ; Welcker et al., 2021 ), is present in both cell clusters. HCR-FISH showed that cells expressing the HMCNT1 mRNA were present in the connective tissue of the regenerating mesentery and certain regions of the connective tissue of the anlage. In the normal intestine, labeled cells were present in the connective tissue ( Fig 5A-C ). Cells in the mesentery and normal intestine connective tissue were somewhat distanced from each other and had an irregular morphology with intense labeling throughout the cytoplasm. A weaker labeling was observed in cells of the anlage, which were more densely packed and adjacent to the coelomic epithelium, suggesting that they correspond to cells undergoing EMT on their way to differentiate into ECM-producing mesenchymal cells. A very similar pattern of expression was observed with HCR-FISH for ERG (not shown). The ERG-expressing cells were also found in the connective tissue of normal intestine and in the cells of the anlage undergoing EMT.

Coelomic epithelia/mesothelial cell populations

The mesothelial layer of the intestine and mesentery is formed by cells in contact with the coelomic fluid (coelomic epithelium or peritoneocytes) together with myocytes and neurons. In the regenerating tissues, some of these cells dedifferentiate and form a coelomic epithelia that differs in morphology ( Mashanov et al., 2005 ) and gene expression ( Mashanov et al., 2017 , 2015 , 2012 , 2010 ) to the mesothelium that normally surrounds the organ. This dedifferentiated coelomic epithelium which is present mainly in the anlage and in areas of the adjacent mesentery, is also responsible for most of the cell division that takes place in the regenerating intestine ( García-Arrarás et al., 2011 ). Many of the changes observed in the mesothelium during regeneration take place in a gradient, beginning at the tip of the mesentery (where the anlage forms) and continuing in the adjacent regions of the mesentery mesothelium. Thus, the analysis of the scRNA-seq data in view of our knowledge of the ongoing events in the 9-dpe regenerating organ leads to the conclusion, as stated previously, that the 7 clusters within the major supra-cluster, represent the cells in the coelomic epithelium of the mesentery and the anlage. Here is our analysis:

C8 represents the proliferating cells

These cells mainly found within the anlage coelomic epithelia express proliferation markers such as PLK1 (serine/threonine-protein kinase PLK1), SMC2 (structural maintenance of chromosomes protein 2), PRI2 (PRIM2 - DNA primase large subunit), PCNA (proliferating cell nuclear antigen), and CDK1 (cyclin-dependent kinase 1) ( Locard-Paulet et al., 2022 ). In addition, this cluster has high expression of TOP2A (DNA topoisomerase 2-beta), a gene that has also been seen to be overexpressed in proliferating basal cells of the human gastrointestinal epithelia ( Busslinger et al., 2021 ). Other marker genes related to cell mitotic activity found here include CENPE (centromere-associated protein E), SMC4 (structural maintenance of chromosomes protein 4), K167 (proliferation marker protein Ki-67), and CCNB3 (G2/mitotic-specific cyclin-B3) ( Fig 6A ). Furthermore, this cluster expresses specific transcription factors such as E2F3 , MCM10 (protein MCM10 homolog), PAF15 (PCNA-associated factor) and BRCA1 (breast cancer type 1 susceptibility protein) that are also associated with control of cell division ( Humbert et al., 2000 ; Lõoke et al., 2017 ; Xie et al., 2014 ). The gseGO terms also confirm its proliferative identity with GO terms related to chromosome separation , condensation , mitotic cytokinesis , and regulation of cell cycle checkpoint ( Fig 6B ). Moreover, the dividing cell population is higher in the anlage samples (4%) than in the mesentery samples (2%), which is in accordance with what we have observed in regenerating animals, that while cell division does take place in the mesentery, more cells are proliferating in the epithelial layer of the intestinal anlage ( Bello et al., 2020 ; García-Arrarás et al., 2011 , 1998 ).

Expression profile of cell clusters undergoing differentiation or proliferation in the coelomic epithelium.

(A) Dot plot showing the expression profile of C5, C8, and C9. Color intensity reflects expression levels in log 2 FC values. Dot size represents percentage of representation of each gene. (B) Top GO enriched terms of biological processes of each cluster. (C) UMAP highlighting C5, C8, and C9, which along with clusters in color blue correspond to the coelomic epithelium. (D-F) Cells expressing MYH7 mRNA in the (D) regenerating mesentery, (E) basal area of the anlage coelomic epithelium and (F) muscle layer of the normal intestine. (G) Phalloidin labeled differentiating muscle cells in basal area of the coelomic epithelium. (H) Cell labeled with RN1 antibody (neuronal marker) in coelomic epithelia of the anlage. Red = DAPI stain, green D-F= HCR-FISH. G= Fluorescent Phalloidin, H= RN1 antibody. Bar-D-F= 25μm, G-H=10μm.

C5 represents muscle precursors

A second cell population known to originate from the coelomic epithelium and that can be identified in our data is that of muscle precursors. These cells are known to differentiate into enteric muscle during the second week of regeneration ( Murray and García-Arrarás, 2004 ). This population can be recognized in our data by the expression of muscle-specific markers present in C5, such as TITIN , MYL1 (myosin light chain 1/3), MYH7 (myosin 7), ACTG (actin, cytoplasmic 2), and CNN3 (calponin-3) ( Fig 6A ). Considering that this cell population is potentially a population of the coelomic epithelium actively undergoing differentiation toward muscle phenotype, they also share gene expression with other epithelial cell clusters (C0, C1, C3, C4), albeit at a lower fold-change. Similarly, this cluster shows a specific expression of transcription factors associated to muscle cells, namely FXL16 (F-box/LRR-repeat protein 16) and SCRT2 (transcriptional repressor scratch 2). Results of GO terms of this cluster also demonstrate enriched terms related to muscle tissue growth, such as muscle tissue morphogenesis , muscle development , myofibril assembly , and sarcomere organization ( Fig 6B ). The focus on development and morphogenesis is to be expected, considering that these are still undergoing differentiation toward a muscle phenotype. HCR-FISH of MYH7 corroborates the muscle precursor phenotype of the cells in this cluster ( Fig 6D-F ). Labeling is observed in the same regions as myoblasts or muscle cells were previously identified ( Murray and García-Arrarás, 2004 ), including differentiating muscle cells within or underlying the coelomic epithelia of the mesentery and the anlage ( Fig 6D & 6E ) and, as expected, the muscle layer in the normal intestine ( Fig 6F ). The same cell population found in the basal side of the coelomic epithelia of the regenerating intestine is labeled with the muscle marker, fluorescently-labeled phalloidin as shown in Fig 6G .

C9 represents the neuroepithelial cells

Cells in C9, represent another population of specialized cells that can be associated to cells previously described in the intestinal anlage. This small number (3%) of cells most likely corresponds to neuroepithelial cells that will eventually give rise to neurons. These cells are shown to be expressing neuroepithelial or neuronal genes such as neurotrypsin ( NETR ), potassium gated-voltage channels, PRD10 (PRDM10 - PR domain zinc finger protein 10 ) , ELAV2 and STA10 (STARD10 - START domain-containing protein 10). The latter is a protein that we have characterized as being expressed by enteric neurons and nerve bundles ( Rosado-Olivieri et al., 2017 ) ( Fig 6A ). The holothurian STA10 (STARD10) is recognized by our monoclonal antibody RN1. This antibody has been used to detect enteric neurons as they begin to differentiate in the coelomic epithelium during the second week of regeneration ( Tossas et al., 2014 ). Labeling of the 8-dpe regenerating mesentery shows one of these RN1 neuronal cells in the coelomic epithelium of the regenerating anlage ( Fig 6H ). Among its associated GO terms are positive regulation of ion transmembrane transporter activity , cyclic nucleotide metabolic process , regulation of muscle contraction, regulation of membrane potential and positive regulation of hormone secretion ( Fig 6B ). Moreover, there is also a great representation of processes involved in development, differentiation, and growth of nerve cells, all of which together would be expected of a neuroepithelial layer.

The characterization of these 3 clusters, that represent cells undergoing differentiation or proliferation, leaves a group of 4 clusters (C0, C1, C3, and C4) that show some overlap in expressed genes and at the same time share some gene expression with some of the previously described clusters. Nonetheless, as seen in Fig 3A &B, the cells in these four clusters still have high expression and representation of specific transcripts.

C4 represents the intestinal coelomic epithelial cells

The gene expression profile of cells in C4 sets them slightly apart from the other three clusters (C0, C1, C3). It identifies cells that are more advanced in their development toward a particular phenotype. C4 shows high expression of genes such KCNQ5 (potassium voltage-gated channel subfamily KQT member 5), SC6A9 (sodium and chloride dependent glycine transporter 1), EFNB2 (Ephrin-B2), and UNC5C (Netrin receptor UNC5C) ( Fig 7A ). Other than these, C4 also shows high expression of genes that are related to cell-cell interactions and ECM molecules such as LAMA2 (laminin subunit alpha-2) , MEGF6 (multiple epidermal growth factor-like domains protein 6) , FMN1 (formin-1) , NPHN (nephrin). This cluster, distinct from others, shows enrichment of GO terms related to more advanced stages of development, such as morphogenesis of epithelium , sensory perception , regulation of calcium ion transmembrane transport , among others ( Fig 7B ). Noteworthy, cells from this cluster correspond mostly to cell from the mesentery samples (67%) rather than from the anlage (37%). Additionally, from all the mesentery cells, about 11% are part of this cluster, while only 4% of the anlage cells are represented here. HCR-FISH of SC6A5 (sodium- and chloride-dependent glycine transporter 2), a gene differentially expressed in the cells of this cluster provided a surprising result. The expression of this mRNA was observed in a few cells in the coelomic epithelia of the regenerating mesentery and intestinal anlage ( Fig 7D-E ). However, intense labeling is observed in some cells of the normal intestine mesothelium. This pattern of labeling strongly suggests that the labeled cells of the regenerating intestine grouped in C4 (both mesenteryal and anlage) correspond to cells that are in a differentiation pathway to become part of the coelomic epithelia (possibly the peritoneocytes) in the regenerated organ. This conclusion is strengthened by the pseudotime analyses presented in the following section.

Expression profile of coelomic cell populations in the regenerating intestine anlage and mesentery.

(A) Dot plot showing the top expressed genes of C0, C1, C3 and C4. Color intensity reflects expression levels in log2FC values. Dot size correspond to the percentage of representation of each gene in the respective cluster compared to all others. (B) Top GO enriched terms of biological processes of each cluster. Dot size corresponds to the enrichment score while the color reflects the adjusted p-value. (C) UMAP highlighting the clusters of the coelomic epithelial cells of the regenerating intestine. (D-E) HCR-FISH for SC6A5 mRNA was performed as a marker for cells of C4. Labeling is observed in the (D) regenerating mesentery, (E) the coelomic epithelium of the anlage and (F) coelomic epithelium of the normal intestine. Red = DAPI stain, green = HCR-FISH. Bar= 25μm.

C0, C1, and C3 represent differentiation stages of coelomic epithelial cells

The three remaining clusters to be analyzed are C0, C1, and C3. These three clusters share many of their top representative genes, which are associated with developmental, regenerative, or oncogenic processes ( Bradford et al., 2009 ; Dunn et al., 2006 ; Lu et al., 2015 ; Nishimoto and Nishida, 2007 ; Oike et al., 2004 ; Zhao et al., 2008 ). These include, for cluster 0: TRFM (melanotransferrin), TIMP4 (metalloproteinase inhibitor 4), and DMBT1 , for cluster 1: FGF13 (fibroblast growth factor 13), TGFB3 (transforming growth factor beta 3), HS90A (heat shock protein HSP 90-alpha), and ANGL1 (angiopoietin-related protein 1) ( Fig 7A ), and for cluster 3: SEM5B (semaphoring-5B), LRIG3 (leucine-rich repeats and immunoglobulin-like domains protein 3), TUTLB (protein turtle homolog B), and NET1 (netrin-1). Moreover, C0 and C1 share an over-representation of ribosomal genes (not shown), and this is reflected in their GO analyses that highlights biological processes related to ribosomal activity such as cytoplasmic translation , and ribosome assembly , biogenesis , and assembly ( Fig 7B ). In addition to these, C3 also showed enrichment of processes involved in the development of tubular lumen-containing structures such as mesonephric and ureteric ducts , differentiation and regulation of cell growth , and negative regulation of axogenesis ( Fig 7B ).

C1 and C3 are also closely related in their localization, being overwhelmingly associated with the anlage. These two clusters are mostly composed of cells from the anlage tissue, where we expect to see the precursor cells that will give rise to specialized cells of the organ ( Fig 1D ). Precisely, C1 and C3 cells together correspond to 37% of all anlage cells, compared to 14.6% of all mesentery cells. In contrast, cells from C0 correspond to 27% and 18% of the mesentery and anlage cells, respectively.

To obtain insight into these cell clusters, we performed HCR-FISH for two different mRNAs; NET1 , a chemotropic protein highly represented in C1 and C3, and TRFM , the top represented gene in C0. Both in situ hybridization experiments labeled cells in the coelomic epithelium of the regenerating intestine, supporting our contention that the large supra-cluster represents the coelomic epithelium layer. However, their spatial pattern of expression was unpredictably different. While NET1 was highly expressed in most of the coelomic epithelial cells of the anlage, little expression was found in the regenerating mesentery or in the coelomic epithelium of the normal intestine ( Fig 8G - H ). TRFM , in contrast, was highly expressed in the coelomic epithelium of the normal intestine and poorly expressed in the intestinal anlage ( Fig 8A-E ). In the regenerating mesentery, a gradient in expression of the TRFM is observed, where high levels of expression were found in the coelomic epithelium close to the body wall and diminished as one approached the anlage. Thus, the HCR-FISH results show that C0, C1, and C3 correspond to cells of the coelomic epithelium, but strongly suggest that C0 differs from C1 and C3 both in their gene expression profile and in the localization where they are found, both in the regenerating and in the normal intestine.

HCR-FISH for TRFM and NET1 shows the differential gene expression of the anlage versus mesentery coelomic epithelia.

TRFM is mainly expressed in the mesentery coelomic epithelial cells that are distant from the regenerating anlage while NET1 is mainly expressed in the coelomic epithelial cells of the anlage TRFM mRNA expression (A-E) shows a gradient from (A) very strong expression in the coelomic epithelium at the attachment of the mesentery to the body wall, (B) high expression in the coelomic epithelia of the mesentery close to the body wall, (C) weaker expression in the mesentery close to the anlage to (D) weak or lack of expression in the coelomic epithelium of the anlage. (E) TRFM mRNA is expressed by the coelomic epithelium of the normal intestine. regenerating intestine. In contrast, few cells express (A) NET1 mRNA in the regenerating mesentery, while high expression is found in (B) the coelomic epithelium of the anlage (C) Little expression of NET1 is present in the coelomic epithelium of the normal intestine. Red = DAPI stain, green = HCR-FISH. Bar= 25μm.

Finally, it is essential to highlight the many signaling, or growth factors expressed by the coelomic epithelial clusters, in particular C0, C1 and C3 ( Fig 3 ). These include: Wnt, Hox, semaphorin, FGFs, TGF-beta, netrin, insulin-like growth factor (IGF), growth/differentiation factors (GDF), and angiopoietin related proteins (ANGL) (e.g. WNT9, SEM5B, FGF13 , NKx3 . 2 , TGFB3 , HOX9 , IGF1 , GDF8, ANGL1 , and FOXF1 ) ( Fig 3 ). This is important in view that the epithelium of the vertebrate blastema is characterized by its chemical modulation of the underlying mesenchyme, as will be discussed later. Likewise, other genes that serve as markers of specific cell types or cellular stages were also identified, including PIWL1 (piwi-like protein 1) in C1, YAP1 in C3 and C4, HES1 (transcription factor HES-1) in C1 and C3, and PRRX1 (paired mesoderm homeobox protein) in C0 and C4 ( Fig 3 and S5 Fig ).

In summary, of the 13 clusters identified, our data strongly suggest that four of them (C9, C10, C11, and C12) correspond to coelomocytes or immune cells, two of them (C2 and C7) correspond to cells with a mesenchymal phenotype and the remaining 7 to cells of the coelomic epithelia. Of these seven, C5 corresponds to differentiating muscle, C9 to differentiating neuroepithelium, C4 to differentiating coelomic epithelia and C8 to proliferating cells. C1 and C3 represent most of the cells found in the coelomic epithelium of the anlage. In contrast, C0 represents a coelomic epithelial phenotype more closely associated with the mesentery than the intestine.

Trajectory Analysis: What populations are driving cell specification?

Among the many mysteries of the holothurian intestinal regeneration process is the identification of the precursor cells. In simple terms, what are the cells from which all nascent cells derive from? We performed a trajectory analysis of the data to provide some insights into this issue. This type of analysis is usually performed with samples at different stages or time points. However, we considered it feasible to conduct this analysis because in the 9-dpe regenerating anlage/mesentery we find cells at various stages of differentiation. These cells could provide crucial information on how the cell populations are associated with each other. To address this, we initially employed RNA velocity analysis. This method describes the temporal dynamics of gene expression based on the relative abundances of spliced and unspliced mRNA across cell populations.

Our initial velocity analysis on all the clusters and samples ( Fig 9A-B ), provided three main results. First, the direction of arrows in our UMAP shows them flowing towards C5, C9, and C4. These arrows do not point towards any other cluster; thus, they are terminal arrows. Second, while arrows from C8 are not terminal, they are directed towards C1. This suggests that cells in C1 provide cells for the growth of the anlage via proliferation. Thus, these results support C5, C9 and C4 as terminal cell clusters which we have described as muscle, neuroepithelial, and the nascent coelomic epithelium cells, respectively. Third, velocity embedding shows shorter arrows that point from clusters 0 and 1 towards terminal populations previously described. Therefore, cells of C0 and C1 are not undergoing significant transcriptional changes. The RNA velocity results of the rest of the clusters are less interesting as they do not show directions towards any other clusters, mainly because of their individuality within the UMAP. However, it is interesting that arrows of the mesenchymal cell populations show distinct directions and lengths. Based on the results, it seems that portions of both mesenchymal clusters (C2 and C7) have gone or are undergoing more extensive differentiation changes.

Trajectory analysis of cell populations from the regenerating intestinal tissue.

(A) UMAP plot of all identified clusters. (B) RNA velocity embedded in UMAP of all main clusters. (C) UMAP of re-clustering of cells from C0, C1, C3, C4, C5, and C9. (D) RNA velocity analysis results from the subset from panel (C). (E) Jitter plot of Slingshot pseudotime of cells from C0, C1, C3, C4, C5 and C9. Pseudotime resulted in two lineages one containing C5 and the other C 9. (F) UMAP of re-clustered anlage cells corresponding to C0, C1, C3, and C4(G) RNA velocity results from the cluster results from panel F. (H) UMAP of panel F plot overlayed with pseudotime results of Slingshot. Color represent pseudotime values from 0 (blue) to 60 (yellow). (I) Jitter plot showing the Slingshot pseudotiume of cells from each cell cluster.

We then re-clustered C0, C1, and C3 cells, along with the differentiating cell populations (C4, C5, and C9) to better understand their relationship ( Fig 9C ). The velocity assessment of these newly clustered populations resulted in a similar pattern ( Fig 9D ). Here, we can see that the differentiating cells (C4, C5, and C9) have long arrows suggesting that these cells are going through an advanced stage of transcriptional change compared to others to some extent, C4. Moreover, it seems that C0 has a closer relationship to cells of C5 and C9 and that some cells of these clusters could potentially differentiate into cells of C4 (the coelomic epithelium cells). In this case, to complement and confirm our RNA velocity interpretation, we also performed a pseudotime analysis using Slingshot, which relies on the expression data of each cluster. This analysis showed C1 in an earlier pseudotime than C3, C4 and C0 in the resulting two lineages ( Fig 9D ). The resulting lineages differed by the terminal clusters, one containing C5 (muscle) and the other C9 (neuroepithelial). Thus, it supports what we have already visualized on the RNA velocity embeddings.

The results described so far show that C0, C1 and C4 are cells in distinct differentiation states, but we wanted to have a clearer view of the cell clusters that potentially have an essential role in the regeneration process. For this, we made another subset of cells that corresponded to C0, C1, C3 and C4, but uniquely from cells of the anlage ( Fig 9F ). The rationale was that cells from the mesentery are certainly at a different state from those of the anlage and thus could interfere with the pseudotime of cells from the anlage. The RNA velocity analysis using this subset strengthened our previous inferences. First, C1 seems to be the least dedifferentiated cell cluster, whose future state will be cells of C3 and part of the population of C4. Second, that C0 seems to be in a specialized state of differentiation that has a relationship to C4 ( Fig 9G ). This would explain the relationship of this cluster to that of the differentiating cells of C5 and C9. Interestingly, portions of C1, C0 and C4 appear to be in an advance process of differentiation based on their longer arrows compared to C3 and another portion of C4 close to C3 ( Fig 9G ). The Slingshot analysis of the anlage cells from C0, C1, C3 and C4 revealed that the pseudotime starts at a point of convergence that contains a portion of cells from C1 and C0. Yet, it further supports the cells from C1 as the least differentiated ( Fig 9H ). C1 is then followed by cells of C3, C4 and lastly C0, which for the most part seems to be at a more advanced differentiation state with a closer relationship to differentiating cells ( Fig 9I ).

In this study we employed scRNA seq and HCR-FISH techniques to examine the cellular phenotypes in regenerating intestinal tissues of the sea cucumber H. glaberrima . These techniques have been seldom used to characterize echinoderm cells, and the few studies available are mainly limited to embryonic stages of sea urchins ( Strongylocentrotus purpuratus, Lytechinus variegatus) and the sea stars ( Patiria miniata) ( Cocurullo et al., 2023 ; Foster et al., 2022 ; Meyer et al., 2023 ; Paganos et al., 2022a , 2022b , 2021 ; Tominaga et al., 2023 ). Nonetheless, these studies provide an excellent description of the cell population and dynamics arising from major germ lines during echinoderm development. We now apply the same techniques to explore cell phenotypes involved in intestinal regeneration in holothurians. This is, to our knowledge, the first-time scRNA-seq and HCR-FISH have been used in adult echinoderms to analyze the cellular and molecular basis of their amazing regenerative properties. This research integrates the extensive cellular and molecular information on intestinal regeneration in holothurians collected over the past two decades, offering a comprehensive view of the cellular phenotypes and molecular changes involved.

Cell types and differentiation stages in the 9-dpe regenerating intestine

Our study has focused on the description of cells present in the 9-dpe regenerating intestine of the sea cucumber, where cells are known to have gone through dedifferentiation and are, at this time-point, in the process of differentiating into specialized cells ( García-Arrarás et al., 2011 ). Based on our analysis we have identified 13 distinct populations that form part of the regenerating intestinal mesentery and anlage ( Fig 1B ). Among these clusters we have described the clusters corresponding to the coelomocyte, coelomic epithelium, and mesenchyme cell types ( Fig 3 ). Thus, we can now use the information that our laboratory and others have gathered to pinpoint the cellular mediators and their activity during regeneration. The coelomic epithelia, for example, is highly influential in the regeneration process as it is the site where major cellular events occur, particularly cell division, dedifferentiation, and differentiation ( García-Arrarás et al., 2011 ). Moreover, the mesenchyme ECM is known to undergo remodeling during the regeneration process, which is also critical for the proper growth of the new tissue ( Quiñones et al., 2002 ). However, while much of the events have been described using microscopic and histological tools, we need a more comprehensive understanding of the transcriptomic characteristics of specific cell populations involved in the regeneration process.

The sea cucumber contains various mesenchymal and coelomocyte populations

Amongst the phenotypes identified are clusters of mesenchymal (C2 and C7) and coelomocyte cell populations (C6, C10, C11, and C12). The mesenchymal populations demonstrated a unique expression of ERG and ETS-1 , markers of mesenchymal populations in other echinoderms. ETS-1 has long been associated with developmental processes of mesenchymal formation in sea urchin ( Koga et al., 2010 ; Rizzo et al., 2006 ). In other studies, ERG , a member of the ETS gene families, has been found to be necessary for controlling mesenchymal identity and differentiation ( Cox et al., 2014 ; Mochmann et al., 2013 ). Further analyses of their individual marker genes suggest that these cells might be involved in EMT. For instance, C2 expresses TIMP3 , a metalloproteinase inhibitor that aids in the extracellular matrix remodeling ( Dewing et al., 2020 ), and NPNT , a gene reported to be related to development and cancer processes ( Magnussen et al., 2021 ). These genes are also crucial for cells undergoing EMT as the cells need to detach from the other cells and the basal lamina that forms the epithelium. Comparatively, the expression profile of C7 with genes such as PA21B (phospholipase A2), TMPS9 (transmembrane protease serine 9), SEPP1 , DMBT1 , ITIH3 , and SVPE1 and its GO results suggest this mesenchymal population is undergoing different processes. Based on these contrasting expression profiles, we propose that C2 corresponds to cells that have recently undergone EMT from the coelomic epithelium and eventually differentiate into a more specialized phenotype (C7). Our pseudotime results further support this developmental transition as cells in C7 appears to be in a more advanced stage when compared to those of C2 ( S6 Fig ). Therefore, these two populations correspond to the first transcriptomic description of mesenchymal phenotypes reported in sea cucumber regenerating intestine.

Our dataset contains four distinct populations that we have characterized as coelomocytes. The coelomocyte populations reported in different holothuroid species ranged between 4 to 6 distinct types ( Hetzel, 1963 ; Ramírez-Gómez et al., 2010 ; Xing et al., 2008 ). Studies from our laboratory previously revealed four different coelomocyte populations in H. glaberrima, distinguished by their morphology, histochemistry, and phagocytic activity. These were lymphocytes, phagocytes, spherulocytes and a population named “giant cells” ( Ramírez-Gómez et al., 2010 ). The distinctive gene expression profile of each of the coelomocytes that we have identified can provide insights into the differences in their role as immune/circulating cells within the sea cucumber. For example, the C6 marker gene fibrinogen-like protein, which is part of a protein family known as FREP, makes this population of great interest.

Mainly because these molecules have been vastly studied across invertebrates, and multiple immune roles have been proposed, including phagocyte recognition and encapsulation ( Hanington and Zhang, 2011 ). A distinct example is that of C10, where the expression of HCK , FER and ITF8 markers suggest this might be a macrophage-like activity ( Chen et al., 2023 ; Dolgachev et al., 2018 ; Shuttleworth, 2018 ). Specifically, HCK , a member of Src family kinases (SFK), has been closely related to macrophage activation and polarization ( Bhattacharjee et al., 2011 ; Poh et al., 2015 ). Interestingly, recent studies in A. japonicus found that an Src homolog mediates the phagocytosis of Vibrio splendidus, which further supports C10 immune identity ( Wan et al., 2022 ). Thus, to our knowledge, this would be the first report of the expression profile of distinct coelomocyte populations in an adult echinoderm species, setting up the stage for integrating these populations with the previously described ones.

The coelomic epithelia of the intestinal anlage is composed of a heterogenous population of cells

Our research findings align with previous microscopic descriptions of cells in the normal and regenerating mesothelium and coelomic epithelia. In our data, we most easily identify the cell population that forms the cluster exhibiting a proliferative phenotype (C8). Proliferative cells in the regenerating intestine are primarily localized within the coelomic epithelium of the anlage ( García-Arrarás et al., 2011 , 1998 ; Reyes-Rivera et al., 2024 ), and the gene expression profile documented here unequivocally identifies these as proliferative.

A second population of cells that can be well correlated to previously described cells is that with a muscle cell phenotype (C5). The evidence suggests that this cell population represents those cells from the coelomic epithelium that are differentiating into myocytes. This evidence includes: (1) the top-expressed genes by the cells in this cluster are all muscle-associated genes; (2) the top-enriched terms are all related to muscle morphogenesis; (3) the cluster is mostly composed of cells that come from the anlage where muscle formation is known to be taking place at this stage ( Murray and García-Arrarás, 2004 ); (4) differentiated muscle cells were not dissociated by the enzymatic procedure strongly suggesting that the muscle cells that we have identified in our data are those that are in a differentiation process, rather than fully differentiated cells closer to the body wall; (5) the cells in this cluster were identified by HCR-FISH of MYH7. These cells are localized toward the basal region of the coelomic epithelium, the region where the differentiating myocytes are known to be present ( Murray and García-Arrarás, 2004 ). Moreover, among the top expressed genes of this cluster is Troponin I ( TNNI1 ), which has also been reported to be highly expressed in muscle precursor cells of the sea star embryo and immature cardiomyocytes of the chicken ( Mantri et al., 2021 ; Tominaga et al., 2023 ). The high differentiation activity of these cells is also supported by our pseudotime analysis, where distinct cells within the cluster are in individual differentiation states ( Fig 9D &E). Further analysis of these populations could allow us to understand the transcriptional changes these cells undergo to become fully specialized muscle cells.

An additional population in our dataset, is the neuroepithelial population (C9). This population has STARD10 , among its top expressed genes, known to be a phospholipid transfer protein present in a neural cell population localized in the coelomic epithelium of the intestine ( García-Arrarás et al., 2019 ; Rosado-Olivieri et al., 2017 ). Furthermore, cells of this cluster have high expression of other genes reported to be expressed by neuronal cells of regenerative and developmental tissues, such as beta-tubulin in developing human gut and sea urchin larva, and synapsin in planaria regenerating tissue ( Cocurullo et al., 2023 ; Elmentaite et al., 2020 ; King et al., 2024 ). Interestingly, an earlier study showing that beta-tubulin positive cells arise from dedifferentiated cells of the regenerating intestinal tissue of the sea cucumber ( García-Arrarás et al., 2019 ) also suggests that similar to the muscle population, these cells are neuroepithelial cells that are differentiating rather than fully specialized. Further support for this theory lies in the main contribution to C9 coming from anlage cells and their terminal differentiation state observed in the pseudotime analysis ( Fig 1D & 10D ).

The remaining four clusters of the coelomic epithelium supra-cluster are more challenging to characterize. Nonetheless we will explore some hypotheses regarding their cellular phenotypes. The identity of C4 was put forward based on its close transcriptional and pseudotime correlation with populations of the coelomic epithelia ( Figs 3 and 9 ). Additionally, in our in situ hybridization, its marker gene SC6A5 , is mainly expressed by the coelomic epithelial cells of the normal intestine, strongly suggesting that the cells expressing this gene in the anlage are those that will become the coelomic epithelial cells of the regenerated organ ( Fig 7 ).

C0 remains an intriguing cell population. On one hand it appears to be closely related, by its gene expression, to other coelomic epithelial populations ( Fig 3 ). Thus, it may represent an unknown cell population or cell stage. The in situ expression suggest that this cluster represents the cellular population of the mesentery or peritoneal coelomic epithelium. In this case we might be evidencing differences in the coelomic epithelium of the mesentery (exemplified by cluster 0) from those of the coelomic epithelium of the intestine (exemplified by cluster 4). Future experiments will be needed to address this controversy.

Lastly, C1 and C3 share many common genes, and represent distinct populations but are still closely associated, regardless of the clustering parameters or the statistical assessment performed. Although we considered combining them into a single cluster, we ultimately decided against it, as they likely represent different stages of cell development or plasticity in the regenerating intestine. One-to-one comparisons revealed that C1 expressed various ribosomal gene markers, while C3 expressed specific genes such as EPHA4 (ephrin type-A receptor 4) and LRIG3, indicating distinct transcriptomic states. Additionally, all trajectory analyses revealed that C1 cells appear to give rise to C3 cells.

Further characterization of the least differentiated cells, C1 and C3, within the anlage coelomic epithelium suggest that these cells probably serve as cellular precursors to differentiating cells. The evidence from pseudotime analyses and gene expression of C1 and C3 strongly support this conclusion. These cells appear to exhibit pluripotency with the potential to form muscle, neurons, coelomic epithelia and mesenchymal cells in the regenerating intestine. Their gene expression profiles include markers from gene families associated with embryonic development such as Hox, zinc fingers, basic helix-loop-helix and others, some of which are linked to stemness and pluripotency. For example, HSP90A , a molecular chaperone essential for stem cell pluripotency, and markers like HES1 and TGFb, essential for maintaining stem cell proliferation, are present in these cell populations ( Aztekin, 2021 ; Mishra et al., 2005 ). Moreover, the expression of PIWL1 (piwi-like protein 1) and YAP1 further support the classification of these clusters as precursor cells of the intestinal anlage. In hydra, piwi-like molecules are exclusively localized in stem/progenitor cells ( Juliano et al., 2014 ), and in mice, YAP1 is expressed only in multipotent cells during intestinal epithelium regeneration, vital for their emergence ( Ayyaz et al., 2019 ). Ultimately, trajectory analysis also supports C1 as a precursor cell population, as the dividing cells (C8) appear to give rise to C1, which then progresses toward C3. This analysis aligns with evidence of a proliferation center in the coelomic epithelia that provides precursor cells essential for the growth of the anlage.

C1 and C3 cells are probably dedifferentiated cells from the mesentery mesothelium, retaining markers common to all epithelial cells. Similar to what is known in other highly regenerative organisms, these dedifferentiated cells can proliferate and then re-differentiate into the cells of the new organ. Future experiments will determine whether they retain the memory of their previous phenotype (muscle or coelomic epithelium) or are completely pluripotent.

The coelomic epithelium of the intestinal anlage is pluripotent

Regardless of which cell population is responsible for giving rise to the cells of the regenerating intestine, our study reveals that the coelomic epithelium, as a tissue layer, is pluripotent. Thus, it is capable of giving rise to various cell types. Microscopy studies across different echinoderm species have consistently suggested that the coelomic epithelium can differentiate into various cell types. Among these are the formation of muscle cells and neurons in sea cucumbers (Yu Dolmatov et al., 1996), mesenchymal cells in brittle star ( Piovani et al., 2021 ), and even immune cells (coelomocytes) in sea star ( Byrne et al., 2020 ; Sharlaimova et al., 2021 ). In some species, the coelomic epithelium has even been proposed to transdifferentiate into intestinal luminal cells ( Mashanov et al., 2005 ).

The direct association of pluripotency with the coelomic epithelium is evident in both the homeostatic regenerative processes that maintain cellular numbers and in the regenerative responses to injury or autotomy. In some cases, pluripotency is directly associated with the coelomic epithelium of the regeneration anlage. This is exemplified in the brittle star Marthasterias glacialis, where it has been suggested that the coelomic epithelium is not only involved in arm regeneration but also serves as a source of immune cells ( Guatelli et al., 2022 ). This report indicated that “residential stem cells” in the regenerating arm of the brittle star originate from the coelomic epithelia ( Candia-Carnevali et al., 2009 ). Similarly, in the sea cucumber Holothuria forskali , the injured mesothelial layer has been identified as the source of undifferentiated cells that differentiate into the cells of the growing organ and into phagocytic cells, now recognized as coelomocytes ( Vandenspiegel et al., 2000 ).

Our scRNA-seq provides convincing evidence that various cell populations originate from the coelomic epithelia of the sea cucumber anlage, offering a more detailed view of the differentiation process. These cell populations, now classified by their gene expression data, correlate well with those previously characterized by microscopy descriptions. We have strong evidence for muscular, proliferating and neuroepithelial cells (C5, C8, and C9) arising from the coelomic epithelia of the anlage. We also show evidence of a differentiating cell population corresponding to the mature intestine coelomic epithelium (C4). Furthermore, our data suggest the presence of a mesenchymal population arising from the coelomic epithelium of the anlage through EMT ( García-Arrarás et al., 2011 ).

Of particular interest, our study reveals that the mesentery coelomic epithelia population (C0) exhibits localized expression of SAA1 , an immune response related gene. This finding, in line with previous reports of SAA1 localization in the coelomic epithelium of the regenerating intestinal tissue ( Santiago et al., 2000 ) ( S5 Fig ), raises the intriguing possibility that the coelomic epithelium of the sea cucumber regenerating intestine could also be the source of coelomocytes, as suggested by Guatelli (2022) ( Guatelli et al., 2022 ). This hypothesis is further supported by the observation that in the sea star Asteria rubens, coelomocytes arise from its coelomic epithelium (Vanden Bossche and Jangoux, 19976). However, our current results, while suggestive, do not yet provide conclusive evidence to support this in the sea cucumber, at least not in the coelomic epithelia of the regenerating intestine or the 9-dpe stage.

Currently, there is no definitive evidence identifying the precursor cells that give rise to these cell populations in echinoderms. However, given our gene expression data and the interactions observed among the cell populations, we postulate that cells from C1 stand as the precursor cell population from which the rest of the cells in the coelomic epithelium arise. Granted, much more experimental evidence will be needed to arrive at this conclusion.

Nonetheless, it provides a focus on the most likely, and probably most intriguing cell population. It also provides the path to explore multiple questions that arise from our results. Can C1 be subdivided into other cell populations? Are the cells in this cluster pluripotent? Do these cells originate from the same cells via dedifferentiation? These and many other questions will indeed be tested in future experiments.

Our data has allowed us to construct a model of the cell populations we identified in the 9-dpe intestinal anlage ( Fig 10 ). This model presents the coelomic epithelium of the anlage as a heterogeneous layer of cells comprising of six cell populations corresponding to distinct differentiation states. Among these are cell populations in the process of differentiation: muscle, neuroepithelium, and coelomic epithelium cells. We also propose the presence of undifferentiated and proliferating cell populations in the coelomic epithelia, which give rise to the cells in this layer. Underneath the coelomic epithelia are the mesenchymal cells, some of which may have originated from the coelomic epithelial layer via EMT. Not shown in our model is a population localized in the coelomic epithelium of the mesentery that we infer also plays an important role in the regeneration process of the intestine.

Model of cellular organization within the anlage of the regenerating intestine of the sea cucumber H. glaberrima at 9 dpe.

Schematic highlights the heterogeneity of the coelomic epithelia of the intestinal anlage and identified mesenchymal populations. The colors of cells and C# correlate with those used on the UMAP of Fig 1B .

The intestinal anlage as a regeneration blastema

The intestinal anlage is a particular regenerative structure that has long been described as a blastema-like structure due to its remarkable morphological resemblance to the “classical blastema” ( García-Arrarás et al., 1998 ). As stated previously a blastema is usually described as a transient structure composed of a mass of proliferating undifferentiated cells. This structure is found at the site of injury and will give rise to the regenerated organ ( Seifert and Muneoka, 2018 ). The blastema cells can originate from different lineages across species. For example, in amphibians, the blastemal cells originate from dedifferentiated muscle, cartilage, fibroblast and other tissues, while in the flatworm they originate from undifferentiated stem cells known as neoblasts ( Baguñà, 2012 ; Globus et al., 1980 ; Wagner et al., 2011 ). The blastema is overlayed by a wound epidermis that is formed by a re-epithelization process following injury, which develops into a specialized wound epidermis that in amphibians is known as the apical epithelial cap (AEC) ( Aztekin, 2021 ). Since its first description, more than 100 years ago, the blastema has been regarded as the best indicator of regeneration, and its presence is associated with tissues or organs that are highly regenerative. In fact, it has been proposed that the presence or absence of a blastema defines the regenerative success or failure of the regeneration process. However, as more regenerative species continue to be studied, the classical definition of a blastema has been reconsidered, moving towards its functional role rather than its histological or structural characteristics ( Seifert and Muneoka, 2018 ).

The blastema of salamanders and newts, have long served as models to describe a blastema cellular and molecular properties ( Globus et al., 1980 ; Scimone et al., 2022 ; Tajer et al., 2023 ). The sea cucumber intestinal anlage stands as a different structure in terms of the tissue compartmentalization of certain activities. Yet, when examined closely, much of the processes and signaling molecules described in the amphibian blastema also take place in the holothurian anlage, although their spatial occurrence might differ. For instance, like the blastema cells of the amphibians, the cells of the holothurian anlage coelomic epithelium are proliferative undifferentiated cells and originated via a dedifferentiation process. Moreover, some of the genes expressed by the amphibian blastema cells or the overlying AEC, such as Wnt, Tgf-beta, and Fgf, are also known to be expressed during sea cucumber regeneration ( Auger et al., 2023 ; Zeng et al., 2023 ). In amphibians, some of these factors are thought to be released by cells of the AEC, serving as a way of modulating the blastema cell activity. For example, FGF1 is expressed in the cells of the AEC while FGF receptors were localized to blastema cells( Zenjari et al., 1996 ). Moreover, both FGF and Wnt signaling have been shown to be vital for the formation of the blastema ( Makanae et al., 2014 ).

This expression of Fgf and Wnt and their possible functions correlates with what has been found in sea cucumbers. Not only are the same growth factors expressed in C1 and C3, but a recent study highlighted the role of FGF4 as a modulator of cell proliferation during the intestinal regeneration of the sea cucumber A. japonicus ( Zeng et al., 2023 ). This group demonstrated that inhibiting FGF4 and its receptor, FGFR2, negatively affected cell proliferation of the mesothelial layer during intestinal regeneration. Similar results have been shown for Wnt, primarily a study by our laboratory demonstrating that Wnt pathway inhibition, either by pharmacological drugs or by RNAi caused a reduction in cell proliferation ( Alicea-Delgado and García-Arrarás, 2021 ; Bello et al., 2020 ). Thus, in the sea cucumber coelomic epithelia of the anlage, cells are also involved in intercellular communications that modulate cellular dynamics through factors similar to those found in the AEC-blastema cell modulation.

The holothurian anlage and the amphibian blastema exhibit some differences. In the latter there is a clear separation of the AEC and the underlying blastema cells. In amphibians, the AEC has been shown to actively participate in the formation and maintenance of the blastemal cells, particularly in their proliferation and differentiation. In contrast, in holothurians, cells with both blastema and AEC characteristics can be found within the coelomic epithelium. In fact, both PRRX1 (paired mesoderm homeobox protein), a gene expressed by amphibian blastema cell precursors ( Gerber et al., 2018 ; Lin et al., 2021 ), and HES1 (transcription factor HES-1), a marker gene of the amphibian AEC, are expressed by the holothurian coelomic epithelium ( Aztekin, 2021 ) ( Fig 3 and S5 Fig ).

In conclusion, in all cases where it has been studied, the blastema appears as a dynamic structure where communication among its cells and the overlying epidermis or epithelium plays important roles in the regenerative process. In this context, the sea cucumber intestinal anlage could be considered a blastema since it uses similar mechanisms to fulfills the same role in the regeneration of the new organ. In this case, the cells of the coelomic epithelium would be those considered to be the blastemal cells.

Animals and Sample Collection

Two adult sea cucumbers were collected from the northeastern rocky region of Puerto Rico and were kept in aerated sea water for acclimatation prior to experiments. Evisceration of the sea cucumber was stimulated by intracoelomic injection of 0.35 M KCl as previously described ( Reyes-Rivera et al., 2024 ). They were maintained in sea water aquaria for 9 days to undergo regeneration until dissection. Before the dissection, animals were anesthetized by immersion in ice-cold water for 45 minutes. Dissection was done through an initial dorsal incision which allowed exposition of the internal organs, upon which the growing rudiment (or anlage) and mesentery were dissected and separated from each other. Each tissue was kept in CMFSS on ice and treated separately during the tissue dissociation process.

Tissue Dissociation

Mesentery and rudiment tissues were digested for 15 minutes, in rocking shaker at room temperature with 1mL of 0.05% Trypsin/0.02% EDTA solution prepared in CMFSS ( Bello et al., 2015 ). Digestions were quenched by adding 500µL of 0.2% BSA in CMFSS and then centrifuged for 2 minutes at 1500 rpm. The supernatants were discarded, and the pellets resuspended with 500µL of 0.04% BSA in CMFSS and the cells were gently separated by pipetting using glass pipets with fire blunted tips. The cell suspensions were filtered using a nylon cell strainer with 70mm mesh and aliquots were taken for cell counting. Cell viability was assessed through manual cell counting with hemocytometer to confirm that a viability higher than 90% was maintained. Sample suspensions were adjusted to 1000 cells/mL using CMFSS, as required for sequencing procedures.

Immuno- and cytochemistry

Fifty µL of dissociated cell sample were placed on polysine-treated slides and fixed with 50 µL of 4% paraformaldehyde and left to dry overnight. Slides were washed in PBS prior to use for immune and or cytochemistry. The methodology for the immunofluorescence techniques/preparation of slides was followed as published except for the time of PBS washes which were of 10 minutes instead of 15 minutes ( Diaz-Balzac et al., 2007 ; Reyes-Rivera et al., 2024 ). The primary and secondary antibodies used are in S7 Table . Samples to be probed with fluorescent phalloidin were treated directly with Phalloidin-TRITC (1:1000; Sigma P1951) for 1h as described previously ( Reyes-Rivera et al., 2024 ). DAPI was incorporated into the mounting medium as described previously ( Reyes-Rivera et al., 2024 ). Cells were observed in a Nikon DS-Qi2 fluorescent microscope.

Single Cell RNA Sequencing and Data Analysis

Single cell libraries were prepared using the 10X Genomics Chromium Next GEM Single Cell 3’ Kit v3.1 and the Chromium 10X instrument, following the protocol from manufacturer (CG000204). Sequencing was carried out with Illumina NextSeq 2000 at the Sequencing and Genotyping Facility of the University of Puerto Rico Molecular Science Building.

The raw sequencing reads from all four samples (two mesentery and two anlage) were processed individually using Cell Ranger (v7.1.0) ( Zheng et al., 2017 ) using the reference genome and gene models of H. glaberrima available at blastkit.hpcf.upr.edu/hglaberrima-v1. Gene models were annotated against human reference proteins from Uniprot and the entire Uniprot reference protein database. Gene IDs throughout the study correspond to the annotations with the human reference protein sequences from Unitprot.

Initial quality control was conducted in R (v4.3.2) using SoupX (v1.6.2) with default parameters to remove ambient RNA ( Young and Behjati, 2020 ). Subsequent data processing was carried out using Seurat (v4.3.0.1) ( Hao et al., 2024 , 2021 ; Satija et al., 2015 ; Stuart et al., 2019 ). After data normalization and identifying variable features (n=2000), 22,970 anchors were identified using 20 dimensions. The datasets were integrated using the canonical correlation analysis (CCA) approach of Seurat, followed by data scaling for principal component analysis (PCA) (npcs = 50) and Uniform Manifold Approximation and Projection (UMAP) dimensional reduction technique (dims = 1:20).

Neighbors were then identified with 20 dimensions, leading to cluster identification at a resolution of 0.5. Various resolution parameters (0.8, 1.2, and 1.4) and dimensions (up to 35 in increments of 5) were tested before the final resolution value. The statistical analysis tool scSHC was employed to assess the probability of each cluster being unique ( Grabski et al., 2023 ). Marker genes of each cluster or supra-clusters were identified using the FindMarkers() function in Seurat and mapped to the appropriate Uniprot IDs. Genes highlighted throughout the study where further validated using the NCBI non-redundant reference database and EchinoBase ( Telmer et al., 2024 ) via BLAST.

Differential expression data across clusters was used to perform gene set enrichment analysis of Gene Ontology (gseGO) biological processes terms using clusterProfiler (v.10.0) ( Yu et al., 2012 ) with a p-value cutoff of 0.05. For this analysis, BLASTp was used to map H. glaberrima gene models to human reference protein sequences from NCBI (GCF_000001405.40), facilitating the assignment of ENTREZ ID to the correct human GO terms in the AnnotationDbi (v1.64.1) human database (v2.1) ( Pagès et al., 2024 ).

RNA velocity loom files were generated with velocyto (v0.17), which relied on genome-masked regions obtained from RepeatModeler (v2.0.5) and RepeatMasker (v4.1.5) ( Smit et al., 2015 ; Smit and Hubley, 2015 ), along with the sea cucumber gene models and CellRanger dataset. These files were further analyzed with velocyto.R (v0.6) and SeuratWrappers (v0.2.0), using the ReadVelocity and RunVelocity functions. Pseudotime analysis for the distinct data subsets was conducted using Slingshot (v2.10.0) ( Street et al., 2018 ). The code of the data analysis has been made available at GitHub ( https://github.com/devneurolab/scRNAseq_Hglaberrima ).

Regenerating intestines from animals eviscerated 8-9 days previously and intestines from non-eviscerated (controls) animals were collected and fixed in 4% (v/v) paraformaldehyde with phosphate-buffered saline (0.01M PBS; 0.138M NaCl; 0.0027M KCl; pH7.4) overnight at 4 ℃. Tissues were then, washed three times by PBS, and treated overnight with 40% saccharose prior to cutting in the cryostat. Sections (20 mm) were prepared in a cryostat (Leica CM1850), as previously published for immunohistochemistry ( Reyes-Rivera et al., 2024 ). Gene spatial expression was determined by designing twelve (12) or twenty (24) set split probes for each gene marker. To decrease probe unspecific binding and increase its signal to background ratio, probes were designed as described by H. Choi and colleagues (Molecular Instruments) ( Choi et al., 2018 ). To ensure probe gene target specificity, all nucleotide regions selected for probe design underwent extensive validation using the newly developed H. glaberrima genome and transcriptome alignment tool ( https://blastkit.hpcf.upr.edu/hglaberrima-v1/ ) ( Medina-Feliciano et al., 2021 ).

Probes were used at a final concentration of approximately 20-25 nM. Immediately after slide preparation, HCR-FISH v3 was carried out using a modified version of Molecular Instrument’s fresh fixed frozen tissues protocol. Probe hybridization was conducted overnight at 37 ℃, while DNA hairpin amplification (B1-546nm or B2-546nm) was done at 25℃ overnight with 3pmol of h1 and h2, respectively. Our protocol modification involved replacing Ethanol with Methanol during sample permeabilization. Also, the use of proteinase K was omitted, and Tween 20 at 0.1% was added to all PBS wash buffers. Once slides were prepared, image acquisition was attained using a Nikon DS-Qi2 fluorescent microscope. Positive control probes (PolyA) were used as signal calibrators to define background from positive signal during image analysis.

Acknowledgements

We thank the funding support from the National Institute of Health (NIH) under the grant number 2R15GM124595. Also funding support was provided by the NIH Research Initiative for Scientific Enhancement (RISE) program to Y.M.N under grant number 5R25GM061151-22. We acknowledge the High-Performance Computing Facility of the University of Puerto Rico sponsored by the University of Puerto Rico and the Institutional Development Award (IDeA) INBRE grant P20 GM103475 from the National Institute for General Medical Sciences (NIGMS), a component of the NIH and the Bioinformatics Research Core of INBRE. We would also like to acknowledge our colleague Joseph F. Ryan for his constructive feedback and computational resources.

Data Availability

Data generated during this project has been made publicly available at Figshare ( Medina-Feliciano et al., 2024 ) including raw and processed sequencing data.

Labeling of dissociated cell phenotypes with cell markers.

Immunocytochemistry, fluorescently labeled phalloidin and Toluidene blue were used to identify various cell populations among the cells dissociated from the mesentery an anlage sample. These include A. a mesenchymal marker (KL4), B. a neuronal marker (RN1), C. a mesothelial marker (Meso1), D. a muscle marker (Phalloidin) and E & F two coelomocyte markers (The antibody SphAA12and toluidine blue). E. Immunocytochemistry using the SphAA12 antibody F. Overlay of toluidene blue labeling (see dark cell on lower left) identifies a different coelomocyte population.

Immunocytochemical labeling of dissociated cell phenotypes using three different antibodies against tubulin.

A. anti-acetylated tubulin labels around 7% of the cells. B. Anti-beta-tubulin labels around 70% of the cells. C. Anti-alpha tubulin labels around 80% of the cells.

UMAP of clusters after statistical assessment with scSHC.

Results reflect that each of the identified clusters are unique, except for C3 and C4, which it suggests they correspond to a single cluster. All clusters, except for C3/4 (black), are colored as in Fig 1B .

Expression of marker genes previously documented in the sea cucumber.

(A-C) UMAP highlighting the cells expressing (A) Wnt9, (B) SAA, and (C) Proteoglycan-4.

Expression of marker genes associated with cell types or state.

(A-D) Violin plot highlighting the level of expression of (A) PIWL1, (B) YAP1, (C) HES1, and (D) SAA1 across clusters.

Pseudotime analysis of all clusters within the 9dpe intestinal regeneration dataset.

Each color reflects the corresponding cluster form C0-C12. The x-axis represents slingshot pseudotime and y-axis the corresponding lineage.

Antibody/Markers used for immune- and cytochemical labeling of dissociated cell suspension.

• Alicea-Delgado M
• García-Arrarás JE
• Alvarado AS
• Medina-Feliciano JG
• Quispe-Parra DJ
• Colón-Marrero S
• Ortiz-Zuazaga H
• Sangiorgi B
• Gregorieff A
• Hudgell MA Barela
• Abreu-Irizarry RJ
• Torres-Gutiérrez V
• Rodríguez-Flores EJ
• Toledo-Román EJ
• Rodríguez N
• Díaz-Díaz LM
• Vázquez-Figueroa LD
• Grillo-Alvarado V
• Bhattacharjee A
• Cathcart MK
• Bradford DK
• Busslinger GA
• Weusten BLA
• Brosens LAA
• Cisternas P
• Davidson PL
• Cabrera-Serrano A
• Candelaria AG
• Candia-Carnevali MD
• Thorndyke MC
• Capdevila C
• Anandakrishnan N
• Azeloglu EU
• Schwarzkopf M
• Artavanis G
• Stegmaier J
• Cocurullo M
• Annunziata R
• Smith L Courtney
• Barela Hudgell MA
• Garcia-Arraras J
• Schillaci D
• Schrankel CS
• Appelboom BL
• Ratnayaka JA
• Desbiensb X
• Duterque-Coquillaud M
• Diaz-Balzac CA
• Santacana-Laffitte G
• San Miguel-Ruiz JE
• Valentin-Tirado G
• Rives-Sanchez M
• Garcia-Arraras JE.
• Diaz-Miranda L
• Garcia-Arraras JE
• Dolgachev V
• Balijepalli S
• McCandless LK
• Suresh M V.
• Raghavendran K
• Machado-Aranda D
• Dolmatov IY
• Nizhnichenko VA
• Hedgecock EM
• Baskakov I V.
• Bankfalvi A
• Suryo Rahmanto Y
• Richardson DR
• Elmentaite R
• Pritchard S
• Bayraktar OA
• Heuschkel R
• Teichmann SA
• Estrada-Rodgers L
• Díaz-Miranda L
• Torres-Avillán I
• Rodrígues-Ramírez R
• Valentín-Tirado G
• Rivera-Cruz A
• Gemberling M
• Masselink W
• Gac-Santel M
• Nowoshilow S
• Treutlein B
• Vethamany-Globus S
• Garrett-Sinha LA
• Irizarry RA
• Candia Carnevali MD
• Varela Coelho A
• Hanington PC
• Andersen-Nissen E
• Stoeckius M
• Srivastava A
• McElrath JM
• Kowalski MH
• Choudhary S
• Fernandez-Granda C
• Remoortere A
• Anderson LVB
• Trimarchi JM
• Dandapani S
• Enomoto-Iwamoto M
• Rosenbloom J
• Götzfried J
• Owusu-Boaitey KE
• Matsubara M
• Hochgerner H
• Lidschreiber K
• Kastriti ME
• Lönnerberg P
• van Bruggen D
• Sundström E
• Castelo-Branco G
• Linnarsson S
• Kharchenko P V.
• Taniguchi-Sugiura Y
• Lindsay-Mosher N
• Locard-Paulet M
• Magnussen SN
• Hadler-olsen E
• Steigedal TS
• Abedini-Nassab R
• De Vlaminck I.
• Mashanova D
• Mashanov VS
• Heinzeller T
• Rojas-Catagena C
• Luna-Martinez K
• Miranda-Negron Y
• Hatleberg WL
• Mochmann LH
• Von Der Heide EK
• Ortiz-Tanchez J
• Nieves-Ríos C
• Alvarez-Falcón S
• Rodriguez-Otero J
• Nishimoto S
• Ortiz-Pineda PA
• Ramírez-Gómez F
• Pérez-Ortiz J
• González-Díaz S
• Santiago-De Jesús F
• Hernández-Pasos J
• Del Valle-Avila C
• Rojas-Cartagena C
• Suárez-Castillo EC
• Méndez-Merced AT
• Roig-López JL
• García-Arrarás JE.
• Benvenuto G
• Ullrich-Lüter E
• Caccavale F
• Zakrzewski A
• Fournon-Berodia I
• Ramirez Flores RO
• Tripathi KP
• Monasterio G
• Saez-Rodriguez J
• Lundeberg J
• Villablanca EJ
• Czarkwiani A
• O’donoghue RJJ
• Akimenko MA
• Quiñones JL
• Quispe-Parra D
• Cruz-González S
• Aponte-Rivera F
• Méndez-Castaner L
• Reyes-Rivera J
• Soriano-López AE
• Fernandez-Serra M
• Squarzoni P
• Archimandritis A
• Ortíz-Pineda P
• Matos-Cruz V
• Rodríguez C
• Ortíz-Zuazaga H
• Rosado-Olivieri EA
• Ramos-Ortiz GA
• Díaz-Balzac CA
• Vázquez-Rosa E
• Biochem Physiol Biochem Mol Biol Author manuscript CB
• Roig-López L
• Santos-Ruiz L
• Santamaría JA
• Ruiz-Sánchez J
• Nishitsuji K
• Cloutier JK
• Sharlaimova N
• Shabelnikov S
• Martynova M
• Petukhova O
• Shuttleworth EE.
• Harrison FW
• Davidson EH
• Stamenkovic I
• Fletcher RB
• Hafemeister C
• Arshinoff BI
• Ettensohn CA
• Bossche JP Vanden
• Vandenspiegel D
• Flammang P.
• Vázquez-Vélez GE
• Rodríguez-Molina JF
• Quiñones-Frías MC
• Vlaeminck-Guillem V
• Duterque-Coquillaud M.
• Tsiairis CD
• Fahlbusch F
• Armistead J
• Kleinridders A
• Hammerschmidt M
• Dolmatov I Yu
• Eliseikina Alexander A
• Bulgakov MG
• Ginanova Nina E
• Lamash Vladimir P
• Korchagin TT
• Eliseikina M
• Ginanova Lamash TN
• Korchagin V
• Bulgakov A.
• Boilly-Marer Y
• Hondermarck H
• Boilly Benoni
• Tanegashima K
• Belgrader P
• McDermott GP
• Montesclaros L
• Underwood JG
• Masquelier DA
• Nishimura SY
• Schnall-Levin M
• Bharadwaj R
• McFarland C
• Mikkelsen TS

Article and author information

José e garcía-arrarás, for correspondence:, version history.

• Sent for peer review : July 1, 2024
• Preprint posted : July 5, 2024
• Reviewed Preprint version 1 : August 23, 2024

© 2024, Medina-Feliciano et al.

This article is distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use and redistribution provided that the original author and source are credited.

Views, downloads and citations are aggregated across all versions of this paper published by eLife.

Be the first to read new articles from eLife

"Experimental Studies of the Regeneration of Planaria Maculata", 1898

Scope and contents note.

This collection consists entirely of reprints of articles by biologist Thomas Hunt Morgan.

• Creation: 1898
• From the Collection: Morgan, Thomas Hunt, 1866-1945 (Person)

Conditions Governing Access

This collection is housed off-site and requires 48-hours' notice for retrieval. Please contact Special Collections for more information. Collection is open for use.

From the Collection: 2.19 Cubic Feet (1 record center carton, 2 legal size document boxes)

Language of Materials

From the Collection: English

Physical Storage Information

• Container: 1 (Mixed Materials)

Repository Details

Part of the Special Collections Repository

Collection organization

"Experimental Studies of the Regeneration of Planaria Maculata", 1898, Container: 1. Thomas Hunt Morgan reprints, MS-0483. Special Collections.

Cite Item Description

"Experimental Studies of the Regeneration of Planaria Maculata", 1898, Container: 1. Thomas Hunt Morgan reprints, MS-0483. Special Collections. https://aspace.library.jhu.edu/repositories/3/archival_objects/49228 Accessed August 24, 2024.

Reference this page: https://aspace.library.jhu.edu/repositories/3/archival_objects/49228

For guidance about how to search our finding aids: https://guides.library.jhu.edu/c.php?g=994933

Sheridan Libraries Policies: https://www.library.jhu.edu/policies/

We are committed to correcting and contextualizing these records as we identify them, and we invite you to contact us at [email protected] if you encounter harmful language in our finding aids or collections.

Special Collections [email protected] The Sheridan Libraries Special Collections 3400 North Charles Street Baltimore, Maryland 21218 USA

Peabody Archives [email protected] Peabody Institute Arthur Friedheim Library 1 East Mount Vernon Place Baltimore, Maryland 21202 USA

John Work Garrett Library (Evergreen Museum and Library) [email protected] 4545 North Charles Street Baltimore, Maryland 21218 USA

View Staff Interface | Powered by ArchivesSpace v3.4.1

• Question and Answer
• Open access
• Published: 08 November 2012

Q&A: What is regeneration, and why look to planarians for answers?

• Alejandro Sánchez Alvarado 1  

BMC Biology volume  10 , Article number:  88 ( 2012 ) Cite this article

36k Accesses

19 Citations

14 Altmetric

Metrics details

What is regeneration?

Historically, philosophers, naturalists and biologists alike have referred to the restoration of missing body parts after traumatic injury as regeneration. While still valid today, the concept of regeneration has expanded through the years to include a diverse set of phenomena. For instance, August Weisman considered physiological cell renewal to be regeneration and wrote so in a chapter dedicated to regeneration in his seminal 1893 book The Germ Plasm : 'the functions of certain organs depend on the fact that their parts continually undergo destruction, and are then correspondingly renewed. In this case it is the process of life itself, and not an external enemy, that destroys the life of a cell' [ 1 ]. Soon after, TH Morgan would also attempt to refine the precision of the concept of regeneration by coining terms that distinguish between regeneration requiring cell proliferation (epimorphosis) and regeneration effected by tissue remodeling (morphallaxis) [ 2 ]. Presently, regeneration is used to include multiple restorative processes manifested either as a result of physiological turnover (for example, the renewal of blood, skin and gut epithelial cells) or injury, and more recently has been used to define a branch of medical practice referred to as 'regenerative medicine'. Thus, rather than becoming more specific, the concept of regeneration has become much more general. This peculiarity can be attributed in great part to the fact that presently, and not unlike previous centuries, little unambiguous molecular, cellular, and evolutionary evidence exists to support a common or divergent mechanism controlling physiological and traumatic regeneration within and between species. That such diverse biological phenomena as adult neurogenesis and limb regeneration can be catalogued under the same umbrella is indicative of our limited mechanistic understanding of regenerative processes, and thus underscores how much more discovery research remains to be done.

Regenerative ability is broadly but unevenly distributed across species; why can't all animals replace tissues and organs after amputation?

A satisfactory explanation to this question is presently lacking. Many organisms known to regenerate body parts after injury have close relatives that have been subjected to similar if not identical selective pressures, and yet are incapable of regeneration [ 3 ]. Two possibilities are plausible: 1) the common ancestor to both species had regenerative capacities, but only one descendant retained such properties; and 2) the common ancestor had no such regenerative capacities and that speciation somehow resulted in the acquisition of regenerative properties in one, but not both descendants. Thus, to understand the seemingly random distribution of regenerative properties across animal species, it becomes essential to determine whether regeneration has evolved at a macroevolutionary (above and across species) or at a microevolutionary (within species) level. Brockes and colleagues [ 4 ] have recently proposed that limb regeneration in salamanders may have evolved locally in this organism, that is to say at a microevolutionary level. Their hypothesis is based on the observation that the three finger protein (TFP) family member Prod1, a key regulator of both patterning and growth in the regeneration of limbs, is likely unique to the salamanders. While tantalizing, this hypothesis needs to be tested in related, but phylogenetically more primitive, salamanders (Figure 1 ). If indeed Prod1 arose recently in Salamandridae (newts and salamanders) and Ambystomatidae (Axolotl) evolution (Figure 1 , in green) as the result of a local expansion of the TFP family, more basal salamanders such as Hynobiidae and Cryptobranchidae would be expected to lack Prod1 and thus the capacity for limb regeneration (Figure 1 , in red). However, there is some evidence that primitive salamanders (Cryptobranchidae) are nevertheless capable of regenerating appendages [ 5 , 6 ]. Still, whether evolution is ancestral or a species-dependent invention is a question that has yet to be conclusively resolved.

Phylogenetic relationships of the salamanders . In green are the species known to possess Prod1 and to display limb regeneration capacities. Red denotes the extant primitive groups of salamanders (see text for explanation). The tree is based on multiple sources [ 26 – 28 ].

Why is the evolutionary origin of regeneration an important issue?

The emerging field of regenerative medicine aims to identify strategies to repair tissues, organs, and human body parts that cannot be naturally replaced when damaged by either trauma or disease. Examples are spinal cord injury, loss of limbs and the loss of neurons to stroke and degenerative diseases like Parkinson's and Alzheimer's. Given that natural regeneration of such tissues occurs with frequency across vertebrate and invertebrate organisms alike, it stands to reason that if we can understand these processes, we should be able to extrapolate this knowledge to human health matters. If regeneration is evolutionarily ancestral and its mechanisms conserved across all animals with regenerative capacities, then it should be possible to coax mammalian tissues to launch a regenerative response by modulating pre-existing repair and regenerative mechanisms. On the other hand, if regeneration is an attribute invented independently multiple times in evolution, understanding which aspects of this process are unique, species-dependent inventions will also have an impact on how to apply the knowledge derived from animal regeneration studies to human health. For instance, understanding why a particular regenerative process takes place in a model system but not in human tissues may help identify new molecular pathways and cellular activities that could be extended to human cells and tissues to stimulate regeneration should endogenous mechanisms not be readily available. Either way, deciphering the modes and mechanisms driving regeneration in multiple model systems will not only help us resolve a long-standing question in biology and evolution, but also have clear ramifications for our understanding of wound healing and regeneration in humans.

Why are planarians a good model system to study regeneration?

There are many reasons why we chose planarians as a model system for the molecular and cellular dissection of animal regeneration [ 7 ]. Our decision was driven in great part by a need to bridge experimental gaps left exposed by traditional genetic model systems. The pronounced limitations of somatic tissue turnover and regenerative properties in standard invertebrate models such as Drosophila and nematodes, coupled with the difficulties of studying adult vertebrate somatic stem cells in vivo , were compelling reasons to examine and test the suitability of planarians, free-living members of the phylum Platyhelminthes, to inform both regeneration and stem cell biology. Planarians, which are non-parasitic flatworms, display remarkable regenerative capacities for all of their tissues, irrespective of whether these were derived from endoderm, mesoderm or ectoderm. Because of their evolutionary position, these bilaterally symmetric, triploblastic organisms were expected to share with vertebrates a large number of the molecular and cellular processes that make form and function possible in animals. We now know that this is indeed the case, as planarians share with vertebrates all of the major developmental signaling pathways responsible for the establishment of the bilateral body plan [ 8 , 9 ]. In addition to their remarkable powers of regeneration, and in contrast to vertebrate regeneration model systems, planarians are small (about the size of a toenail clipping), and rather easy and relatively inexpensive to rear in great numbers in the lab, allowing for genome-wide functional studies of regeneration. Planarians were also very attractive as a model system because an extensive body of literature spanning over two centuries exists, which describes in great detail the remarkable developmental plasticity of these animals [ 8 ]. This exquisite body of knowledge has, for the most part, just begun to be examined using the rigors and methods of modern molecular and cellular biology.

Why study one particular species - Schmidtea mediterranea ?

This particular species was selected because it met a number of criteria deemed necessary to perform molecular, cellular, and mechanistic studies successfully [ 7 ]. First, S. mediterranea is a stable diploid possessing four pairs of chromosomes (Figure 2 ). Second, it has a relatively small genome (approximately 800 Mb or the equivalent of the first four human chromosomes), making it relatively easy to sequence the genome [ 10 ]. Third, this species exists in two biotypes - one sexual, the other asexual - allowing for a comparison of both sexual and asexual reproduction and embryogenesis and regeneration. Fourth, because of its robust regenerative capacity, we were able to generate clonal lines that have limited polymorphisms in the population, thus facilitating gene isolation, and spatial and functional assays. Finally, the complex anatomy of planarians is well represented in S. mediterranea , allowing us to identify tissue-specific markers and thus define and visualize all organ systems (Figure 3 ).

The planarian Schmidtea mediterranea . Sexual (left) and asexual biotypes are shown with their corresponding diploid karyotypes. Modified from [ 17 , 22 ].

Sampling of the anatomical complexity displayed by the planarian S. mediterranea . Overlay of gut (blue, Smed-porcn-1 ), neurons (yellow, Smed-PC-2 ), axons, and pharynx (magenta, anti-α-tubulin antibody). Scale bar 200 μm. Modified from [ 8 ].

What triggers regeneration?

Across multiple species and phyla, the stimulus for regeneration is amputation. Planarians are no exception. Wounding and amputation in this organism leads to a coordinated cellular and molecular response that can be measured and is currently under intense investigation. We know, for example, that upon amputation, the body wall musculature undergoes depolarization, which in turn results in the contraction of the muscle fibers near the amputation plane, effectively reducing the surface area of the wound. This is followed by a loss of columnar morphology of the epidermal cells adjacent to the wound and their subsequent migration over the wound, until the exposed tissues are completely covered by a monolayer of these cells. This amputation-induced epithelial-mesenchymal interaction is likely involved in the signaling that triggers regeneration, as one of the earliest genes induced in response to wounding is part of the ancient, broadly conserved Wnt/β-catenin signaling pathway [ 11 – 13 ], which also plays a key role in wound healing [ 14 ].

Which types of tissue can regenerate?

All of them - that is why planarians are so attractive for the study of regenerative mechanisms. As such, it becomes possible to study how the differentiated derivatives of all embryonic germ layers (ectoderm, mesoderm and endoderm) can be restored in an adult context after they have been lost to amputation.

What is the smallest fragment of tissue capable of regenerating a complete worm?

This often-asked question was answered by TH Morgan in 1898 [ 15 ]. He reported that a fragment equivalent to 1/279th the size of the original animal was sufficient to produce a complete animal. He arrived at this number by first measuring the animal using eye-micrometers in his microscope, for which each division was 1/53 mm and 1/28 mm. After measuring the worm, Morgan would then draw, cut, and weigh a thin but larger cardboard scale replica of the intact animal. He would then cut the animal into the smallest possible pieces, measure each piece, and then cut an equivalent sized fragment form the cardboard replica. He followed the regeneration of the cut fragments, and then measured the weight of the cardboard pieces corresponding to the animal fragments that completed regeneration successfully. In other words, the cardboard replica was measured to weigh 279.5 mg, and the smallest planarian fragment that could regenerate corresponded to a cardboard piece weighing 1 mg, thus resulting in the 1/279th value for the smallest piece capable of regenerating a complete worm.

Is some sort of specialized stem cell required for regeneration?

Yes. Large numbers of small, undifferentiated cells populating the body plan of many flatworms were noticed towards the end of the 19th century [ 8 ]. These cells were also noted to be mitotically active, and their role in regeneration was confirmed by the pioneering work of Bardeen and Baetjer [ 16 ]. These investigators reported in 1904 that animals exposed to ionizing radiation lost their regenerative capacities. When the worms were inspected histologically, Bardeen and Baetjer reported a complete absence of both mitotic activity and undifferentiated cells. These specialized cells are referred to as neoblasts.

What are neoblasts?

Neoblasts are pluripotent, somatic stem cells that are broadly distributed across the planarian anatomy. In asexual animals they are the only cells capable of undergoing cell division and as such can be readily eliminated by gamma-irradiation to produce an animal that can survive for several weeks, but is incapable of mounting a regenerative response upon wounding. Neoblasts are small (approximately 5 µm in diameter) and by morphology alone correspond to approximately 25% of all cells in the organism. They share with other stem cells the characteristic of having a large nucleus containing highly decondensed chromatin and a scant, basophilic cytoplasm [ 17 ]. Molecular markers and genes affecting the function of neoblasts and their progeny have been identified [ 18 – 20 ], providing the field with novel molecular tools to characterize their biological functions in vivo .

Can a single neoblast generate a whole animal?

While the in vitro culture of neoblasts has yet to be established, single stem cell transplantation into adult planarians is possible, making the animal itself a tissue culture chamber in which to grow these cells. Recent experiments have unambiguously demonstrated that, with some frequency, single, transplanted neoblasts can restore viability and rescue many of the morphological defects of lethally irradiated adult animals [ 21 ]. Interestingly, under these conditions, the rescue of the irradiated animals occurs through a clonal expansion rather than migration of the injected cell, followed by expansion of the resulting colony of stem cells. These data would indicate that neoblasts are not migratory cells, a somewhat surprising result given how many niches (that is, the cellular microenvironment capable of supporting the maintenance of stem cells in plants and animals) were left vacant by the irradiation that would have been expected to promote stem cell mobilization.

Can neoblasts migrate?

Recently, we have shown that neoblasts can in fact migrate, but appear to do so only when a breach in structural integrity such as amputation is inflicted upon the animal. This stem cell behavior was discovered by selectively eliminating stem cells from only parts of the animal with gamma-irradiation. Essentially, the trunks of animals were protected from irradiation by a lead shield, while the head and tail were subjected to lethal doses of irradiation. When the animal is not amputated, the stem cells residing in the protected region do not mobilize to repopulate the irradiated tissues (Figure 4a ). However, if the partially irradiated animal is then decapitated, a marked mobilization of neoblasts towards the wound site becomes readily apparent (Figure 4b,c ) [ 22 ]. The fact that neoblasts do not appear to migrate in the absence of amputation, while at the same time continuing to effect tissue homeostasis [ 21 ], indicates that different mechanisms for restorative versus injury induced regeneration are likely to exist.

In vivo migration of stem cells in planarians . (a) Neoblasts labeled with the stem cell marker Smed-piwi-1 (purple) in a partially irradiated, unamputated (intact) animal. (b) Migrating neoblasts in a decapitated, partially irradiated animal. Arrow points to neoblasts at or near the site of amputation. (c) A decapitated, partially irradiated animal in which cells are visualized via fluorescent in situ hybridization. Neoblasts are in green ( Smed-piwi-1 ) and post-mitotic progeny in magenta ( Prog-1 ). (a,b) Arrowheads denote the boundary between irradiated (top) and unirradiated (bottom) tissues. Modified from [ 22 ].

Can the regenerative behavior of cells be traced to gene function in planarians?

Yes. RNA interference (RNAi) can be used to robustly abrogate specific gene function [ 16 ], which became possible in 1998, when we extended to planarians Dr Andy Fire (our then downstairs neighbor at the Carnegie Institution for Science in Baltimore, MD) and Dr Craig Mello's discovery that double-stranded RNA could silence gene expression in Caenorhabditis elegans . We demonstrated the efficiency and specificity of this method in planarians by targeting and measuring the protein products of the myosin and tubulin genes (Figure 5a ), which appeared in press a year later [ 23 ]. Presently, RNAi is the principal methodology being used by the planarian community to functionally interrogate genes and their functions in this organism. This method has allowed investigators to uncover remarkable phenotypes in RNAi-based screens [ 24 ] and signaling pathway perturbations (Figure 5b ) [ 13 , 25 ].

Functional perturbation of gene function by RNA-mediated genetic interference (RNAi) in planarians . (a) First RNAi effects reported in planarians show the specific loss of myosin (green) and tubulin (red) in regenerating tissues [ 23 ]. (b) Formation of multiple heads in an unamputated organism treated with β- catenin(RNAi) [ 25 ].

What lies ahead for planarians in particular and the field of regeneration in general?

Regeneration remains one of the last untamed frontiers of developmental biology. It is amongst the oldest biological problems known to humankind, dating back to antiquity in many cultures and, perplexingly, still awaiting a satisfactory mechanistic explanation. It is my firm belief that the time to plumb the molecular depths of regeneration is now. Tremendous strides have been made in the study of regeneration in Hydra, planarians, zebrafish, newts, and salamanders. Hence, a critical mass of knowledge is accruing that would permit a systematic interspecies comparison of regenerative capacities across very distant and diverse phyla. Equally important, a systematic and formal exploration of how the mechanisms of regeneration compare to embryogenesis can now begin in earnest. Such a comparison would help address the long-standing question of whether regeneration is simply a recapitulation of development or made possible by independent mechanistic innovations. In the case of planarians, are their embryonic stem cells functionally different from neoblasts? When during embryogenesis are neoblasts specified? To what extent are embryonic axes formation and organogenesis mechanisms similar or dissimilar between planarian embryogenesis and regeneration? Is the genetic toolkit required to organize body axes and facilitate organogenesis during embryogenesis the same as during regeneration? This is a particularly important question because many organisms such as the mouse, fruitfly, and frog can recover from ablation of numerous blastomeres or substantial injury to embryonic organs, yet display limited regenerative capacities as adults. Therefore, testing whether regulative development occurs in planarian embryos, for example, may help us identify key differences crucial to preserving regenerative abilities into adulthood. These and many more fascinating questions abound [ 8 ], so it is clear that when it comes to regeneration, we have but just begun to scratch the surface.

Weismann A: The Germ-Plasm. A Theory of Heredity. 1893, New York: Charles Scribner's Sons, First English

Google Scholar  

Morgan TH: Regeneration. 1901, New York: The Macmillan Company

Book   Google Scholar  

Brockes JP, Kumar A, Velloso CP: Regeneration as an evolutionary variable. J Anat. 2001, 199: 3-11. 10.1046/j.1469-7580.2001.19910003.x.

Article   PubMed   CAS   PubMed Central   Google Scholar  

Garza-Garcia AA, Driscoll PC, Brockes JP: Evidence for the local evolution of mechanisms underlying limb regeneration in salamanders. Integr Comp Biol. 2010, 50: 528-535. 10.1093/icb/icq022.

Article   PubMed   Google Scholar  

Griffin PC, Solkin VA: Ecology and Conservation of Onychodactylus fischeri (Caudata, Hynobiidae) in the Russian Far East. Asiatic Herpetol Res. 1995, 6: 53-61.

Developing a monitoring protocol for Siren and Amphiuma in the Southeastern United States. [ http://fl.biology.usgs.gov/posters/Herpetology/Sirens_and_Amphiuma/sirens_and_amphiuma.html ]

Sánchez Alvarado A: Planarian regeneration: its end is its beginning. Cell. 2006, 124: 241-245. 10.1016/j.cell.2006.01.012.

Elliott SA, Sánchez Alvarado A: The history and enduring contributions of planarians to the study of animal regeneration. WIREs Dev Biol. 2012, doi: 10.1002/wdev.1082

Sánchez Alvarado A: Regeneration and the need for simpler model organisms. Philos Trans R Soc Lond B Biol Sci. 2004, 359: 759-763. 10.1098/rstb.2004.1465.

Article   PubMed   PubMed Central   Google Scholar  

The Schmidtea mediterranea genome database (SmedGD). [ http://smedgd.stowers.org/ ]

Gurley KA, Elliott SA, Simakov O, Schmidt HA, Holstein TW, Sánchez Alvarado A: Expression of secreted Wnt pathway components reveals unexpected complexity of the planarian amputation response. Dev Biol. 2010, 347: 24-39. 10.1016/j.ydbio.2010.08.007.

Petersen CP, Reddien PW: A wound-induced Wnt expression program controls planarian regeneration polarity. Proc Natl Acad Sci USA. 2009, 106: 17061-17066. 10.1073/pnas.0906823106.

Rink JC, Gurley KA, Elliott SA, Sánchez Alvarado A: Planarian Hh signaling regulates regeneration polarity and links Hh pathway evolution to cilia. Science. 2009, 326: 1406-1410. 10.1126/science.1178712.

Whyte JL, Smith AA, Helms JA: Wnt signaling and injury repair. Cold Spring Harb Perspect Biol. 2012, 4: a008078-10.1101/cshperspect.a008078.

Morgan TH: Experimental studies of the regeneration of Planaria maculata . Arch Entw Mech Org. 1898, 7: 364-397.

Bardeen C, Baetjer F: The inhibitive action of the Roentgen rays on regeneration in planarians. J Exp Zoöl. 1904, 1: 191-195.

Article   Google Scholar  

Newmark PA, Sánchez Alvarado A: Not your father's planarian: a classic model enters the era of functional genomics. Nat Rev Genet. 2002, 3: 210-219.

Article   PubMed   CAS   Google Scholar  

Reddien PW, Oviedo NJ, Jennings JR, Jenkin JC, Sánchez Alvarado A: SMEDWI-2 is a PIWI-like protein that regulates planarian stem cells. Science. 2005, 310: 1327-1330. 10.1126/science.1116110.

Guo T, Peters AH, Newmark PA: A Bruno-like gene is required for stem cell maintenance in planarians. Dev Cell. 2006, 11: 159-169. 10.1016/j.devcel.2006.06.004.

Eisenhoffer GT, Kang H, Sánchez Alvarado A: Molecular analysis of stem cells and their descendants during cell turnover and regeneration in the planarian Schmidtea mediterranea. Cell Stem Cell. 2008, 3: 327-339. 10.1016/j.stem.2008.07.002.

Wagner DE, Wang IE, Reddien PW: Clonogenic neoblasts are pluripotent adult stem cells that underlie planarian regeneration. Science. 2011, 332: 811-816. 10.1126/science.1203983.

Guedelhoefer OC, Sánchez Alvarado A: Amputation induces stem cell mobilization to sites of injury during planarian regeneration. Development. 2012, 139: 3510-3520. 10.1242/dev.082099.

Sánchez Alvarado A, Newmark PA: Double-stranded RNA specifically disrupts gene expression during planarian regeneration. Proc Natl Acad Sci USA. 1999, 96: 5049-5054. 10.1073/pnas.96.9.5049.

Reddien PW, Bermange AL, Murfitt KJ, Jennings JR, Sánchez Alvarado A: Identification of genes needed for regeneration, stem cell function, and tissue homeostasis by systematic gene perturbation in planaria. Dev Cell. 2005, 8: 635-649. 10.1016/j.devcel.2005.02.014.

Gurley KA, Rink JC, Sánchez Alvarado A: Beta-catenin defines head versus tail identity during planarian regeneration and homeostasis. Science. 2008, 319: 323-327. 10.1126/science.1150029.

Weisrock DW, Harmon LJ, Larson A: Resolving deep phylogenetic relationships in salamanders: analyses of mitochondrial and nuclear genomic data. Syst Biol. 2005, 54: 758-777. 10.1080/10635150500234641.

Wiens JJ, BOnett RM, Chippindale PT: Ontogeny discombobulates phylogeny: paedomorphosis and higher-level salamander relationships. Syst Biol. 2005, 54: 91-110. 10.1080/10635150590906037.

Zhang P, Wake DB: Higher-level salamander relationships and divergence dates inferred from complete mitochondrial genomes. Mol Phylogenet Evol. 2009, 53: 492-508. 10.1016/j.ympev.2009.07.010.

Download references

Acknowledgements

ASA is an investigator of the Howard Hughes Medical Institute and the Stowers Institute for Medical Research and is a recipient of NIH NIGMS MERIT award R37GM057260.

Author information

Authors and affiliations.

Howard Hughes Medical Institute and Stowers Institute for Medical Research, 1000 E. 50th Street, Kansas City, MO, 64110, USA

Alejandro Sánchez Alvarado

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Alejandro Sánchez Alvarado .

Authors’ original submitted files for images

Below are the links to the authors’ original submitted files for images.

Authors’ original file for figure 1

Authors’ original file for figure 2, authors’ original file for figure 3, authors’ original file for figure 4, authors’ original file for figure 5, rights and permissions.

Open Access This article is published under license to BioMed Central Ltd. This is an Open Access article is distributed under the terms of the Creative Commons Attribution License ( https://creativecommons.org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article.

Sánchez Alvarado, A. Q&A: What is regeneration, and why look to planarians for answers?. BMC Biol 10 , 88 (2012). https://doi.org/10.1186/1741-7007-10-88

Download citation

Received : 12 October 2012

Accepted : 02 November 2012

Published : 08 November 2012

DOI : https://doi.org/10.1186/1741-7007-10-88

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Regenerative Capacity
• Limb Regeneration
• Somatic Stem Cell
• Irradiate Animal
• Limited Regenerative Capacity

BMC Biology

ISSN: 1741-7007

• General enquiries: [email protected]

The Online Books Page

Experimental studies of the regeneration of planaria maculata..

Help with reading books -- Report a bad link -- Suggest a new listing

Home -- Search -- New Listings -- Authors -- Titles -- Subjects -- Serials

Books -- News -- Features -- Archives -- The Inside Story

Edited by John Mark Ockerbloom (onlinebooks@pobox.upenn.edu) OBP copyrights and licenses .

• Skip to page content
• Skip to text-only view
• Skip to search in this text

Sorry, there was a problem.

Download the free Kindle app and start reading Kindle books instantly on your smartphone, tablet, or computer - no Kindle device required .

Read instantly on your browser with Kindle for Web.

Using your mobile phone camera - scan the code below and download the Kindle app.

Image Unavailable

• To view this video download Flash Player

Experimental Studies Of The Regeneration Of Planaria Maculata Hardcover – July 18, 2023

This classic study of planarian regeneration by the Nobel laureate Thomas Hunt Morgan is a must-read for students of biology and genetics. Morgan's experiments explored the fascinating ability of these flatworms to regenerate lost body parts, shedding light on the mechanisms of genetic inheritance and cellular differentiation. This book is a groundbreaking work in the field of genetics and a fascinating read for anyone curious about the mysteries of life.

This work has been selected by scholars as being culturally important, and is part of the knowledge base of civilization as we know it.

This work is in the "public domain in the United States of America, and possibly other nations. Within the United States, you may freely copy and distribute this work, as no entity (individual or corporate) has a copyright on the body of the work.

Scholars believe, and we concur, that this work is important enough to be preserved, reproduced, and made generally available to the public. We appreciate your support of the preservation process, and thank you for being an important part of keeping this knowledge alive and relevant.

• Print length 36 pages
• Language English
• Publisher Legare Street Press
• Publication date July 18, 2023
• Dimensions 6.14 x 0.25 x 9.21 inches
• ISBN-10 1019453273
• ISBN-13 978-1019453278
• See all details

Product details

• Publisher ‏ : ‎ Legare Street Press (July 18, 2023)
• Language ‏ : ‎ English
• Hardcover ‏ : ‎ 36 pages
• ISBN-10 ‏ : ‎ 1019453273
• ISBN-13 ‏ : ‎ 978-1019453278
• Item Weight ‏ : ‎ 8.1 ounces
• Dimensions ‏ : ‎ 6.14 x 0.25 x 9.21 inches

Customer reviews

• 5 star 4 star 3 star 2 star 1 star 5 star 0% 0% 0% 0% 0% 0%
• 5 star 4 star 3 star 2 star 1 star 4 star 0% 0% 0% 0% 0% 0%
• 5 star 4 star 3 star 2 star 1 star 3 star 0% 0% 0% 0% 0% 0%
• 5 star 4 star 3 star 2 star 1 star 2 star 0% 0% 0% 0% 0% 0%
• 5 star 4 star 3 star 2 star 1 star 1 star 0% 0% 0% 0% 0% 0%

Customer Reviews, including Product Star Ratings help customers to learn more about the product and decide whether it is the right product for them.

To calculate the overall star rating and percentage breakdown by star, we don’t use a simple average. Instead, our system considers things like how recent a review is and if the reviewer bought the item on Amazon. It also analyzed reviews to verify trustworthiness.

No customer reviews

• About Amazon
• Investor Relations
• Amazon Devices
• Amazon Science
• Sell products on Amazon
• Sell on Amazon Business
• Sell apps on Amazon
• Become an Affiliate
• Advertise Your Products
• Self-Publish with Us
• Host an Amazon Hub
• › See More Make Money with Us
• Amazon Business Card
• Shop with Points
• Reload Your Balance
• Amazon Currency Converter
• Amazon and COVID-19
• Your Account
• Your Orders
• Shipping Rates & Policies
• Returns & Replacements
• Manage Your Content and Devices

• Conditions of Use
• Privacy Notice
• Consumer Health Data Privacy Disclosure
• Your Ads Privacy Choices

A histological study of regeneration in planaria simplicissima, planaria maculata and planaria morgani

• Published: August 1907
• Volume 24 , pages 350–373, ( 1907 )

Cite this article

• N. M. Stevens 1  

141 Accesses

9 Citations

Explore all metrics

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save.

• Get 10 units per month
• Download Article/Chapter or eBook
• 1 Unit = 1 Article or 1 Chapter
• Cancel anytime

Price includes VAT (Russian Federation)

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

Bardeen , C. R., '01, On the Physiology of Planaria maculata, with Especial Reference to the Phenomena of Regeneration. Am. Journ. of Physiology. Vol. 5. 1901.

- '02, Embryonic and Regenerative Development in Planarians. Biol. Bulletin. Vol. 3. 1902.

Child , C. M., '06a, Contributions toward a Theory of Regulation. I. The Significance of the Different Methods of Regulation in Turbellaria. Archiv f. Entw.-Mech. Vol. 20. 1906.

- '06b, The Relation between Regulation and Fission in Planaria. Biol. Bulletin. Vol. 11. 1906.

- '06 c, The Relations between Functional Regulation and Form Regulation. Journ. of Exp. Zoöl. Vol. 3. 1906.

Curtis , N. C., '02, The Life History, the Normal Fission and the Reproductive organs of Planaria maculata. Proc. of Boston Soc. of Nat. Hist. Vol. 30. 1902.

Lillie , J. R., '01, Notes on Regeneration and Regulation in Planarians. II. Am. Journ. of Physiology. Vol. 6. 1901.

Morgan , T. H., '98, Experimental Studies of the Regeneration of Planaria maculata. Archiv f. Entw.-Mech. Vol. 7. 1898.

- '00, Regenerations in Planarians. Archiv f. Entw.-Mech. Vol. 10. 1900.

—, '01a, Regeneration. The Macmillan Co., New York. 1901.

Google Scholar  

- '01b. Growth and Regeneration in Planaria lugubris. Archiv f. Entw.-Mech. Vol. 13. 1901.

- '04a. An Attempt to Analyze the Phenomena of Polarity in Tubularia. Journ. of Exp. Zoöl. Vol. I. 1904.

- '04b. An Analysis of the Phenomena of Organic »Polarity«. Science. N. S. Vol. 20. 1904.

Morgan , T. H., '05, »Polarity« considered as a Phenomenon of Gradation of Materials. Journ. of Exp. Zoöl. Vol.2. 1905.

- '06, The Physiology of Regeneration. Journ. of Exp. Zoöl. Vol 3. 1906.

Randolph , H., '97, Observations and Experiments on Regeneration in Planarians. Archiv f. Entw.-Mech. Vol. 5. 1897.

Stevens , N. M., '01, Notes on Regeneration in Planaria lugubris. Archiv f. Entw.-Mech. Vol. 13. 1901.

- '04, On the Germ Cells and the Embryology of Planaria simplicissima. Proc. Acad. Nat. Sci. Philad. Feb. 1904.

Stevens , N. M., and Boring , A. M., '05, Regeneration in Polychoerus caudatus. Part I. Observations on Living Materials. By N. M. Stevens . Part II. Histology. By A. M. Boring . Journ. of Exp. Zoöl. Vol. 3. 1905.

- '06, Planaria morgani n. sp. Proc. of the Acad. Nat. Sci. Philad. Feb. 1906.

Thacher , H. F., '02, The Regeneration of the Pharynx in Planaria maculata. Am. Nat. Vol. 36. 1902.

Download references

Author information

Authors and affiliations.

Biological Laboratory, Bryn Mawr College, USA

N. M. Stevens

You can also search for this author in PubMed   Google Scholar

Rights and permissions

Reprints and permissions

About this article

Stevens, N.M. A histological study of regeneration in planaria simplicissima, planaria maculata and planaria morgani. Archiv für Entwicklungsmechanik der Organismen 24 , 350–373 (1907). https://doi.org/10.1007/BF02161840

Download citation

Received : 05 April 1907

Issue Date : August 1907

DOI : https://doi.org/10.1007/BF02161840

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Developmental Biology
• Histological Study
• Find a journal
• Publish with us
• Track your research