Whole-body regeneration and polarity repatterning in polyps of Xenia umbellata (Octocorallia) Authors: Elinor Nadir, Vyctoria Marillac, Tamar Lotan, Yehuda Benayahu Published: 2026 ABSTRACT Cnidarians are renowned for their regenerative abilities, up to whole-body regeneration (WBR), which supports recovery from disturbances and contributes to their ecological success. However, most mechanistic insight into morphogenesis and polarity in WBR still derives from cnidarian solitary polyp models. As such, WBR and polarity repatterning in polyps of colonial cnidarians, especially octocorals, is largely understudied. The pulsating zooxanthellate octocoral Xenia umbellata is increasingly recognized for its ecological resilience and invasive potential, with regenerative abilities that may contribute to its spreading success. Here, we experimentally evaluated regeneration from miniature fragments of both tentacles and polyp bodies and assessed polarity repatterning following double amputation of polyps, removing both oral and basal structures. We show that X. umbellata exhibits WBR, regenerating complete polyps from fragments as small as ∼0.1 mm2. Regeneration from both polyp parts proceeded through a conserved sequence of stages, including tissue ball formation, ultimately producing fully functional polyps. Early regeneration dynamics differed between fragment types: tentacle-derived fragments required structural reorganization, including the initial elimination of differentiated tentacle structural features and de novo mesentery formation, whereas polyp body fragments retained their internal architecture and regenerated more rapidly. Fragment size had only a limited effect on regeneration success and rate of both polyp parts. All successful regeneration led to pulsating polyps with oral disc diameters 5–10% the size of original polyps. In contrast to solitary cnidarians, over 60% of double-amputated X. umbellata polyps regenerated two oral discs. However, non-synchronized timing of oral disc formation and biased placement of attachment discs revealed patterns consistent with partial retention of original positional information despite loss of axial landmarks. Together, these findings suggest that regeneration in X. umbellata combines internal polarity cues with additional colony-level patterning processes absent from solitary models, and different from currently known colonial cnidarians. The findings expand the comparative framework of cnidarian regeneration and further establish X. umbellata as a powerful model for studying WBR, polarity, and coloniality in cnidarians. Introduction Regeneration ability is widespread across animal phyla, but to very different degrees, ranging from local tissue regeneration to regrowing missing body parts (Poss & Tanaka, 2024). Most extreme among regeneration abilities is whole-body regeneration (WBR), in which animals can regrow from fragments of their original bodies (Bely & Nyberg, 2010; Bideau et al., 2021). This ability is particularly common in in colonial organisms, such as urochordates, bryozoans, and cnidarians, and is intimately tied to their asexual reproductive abilities, which are necessary for colony growth (Hiebert, Simpson & Tiozzo, 2021). Cnidarians—hydroids, corals, sea anemones, and jellyfish—are particularly adept at regeneration, which provides key advantages for survival of mechanical disturbances, predation, or environmental stressors common in their natural habitats, as well as their colonization capabilities (Elchaninov, Sukhikh & Fatkhudinov, 2021; Bideau et al., 2021; Hiebert, Simpson & Tiozzo, 2021). In recent years, regeneration ability was linked to invasive success in fast-growing Tubastrea corals (Luz, Miller & Kitahara, 2021), while phase shifts in benthic coral reef communities favor stress-tolerant, fast-growing, weedy species, which often possess extreme forms of asexual reproduction and regeneration (Darling et al., 2012; Kubicek et al., 2019; Creed et al., 2021). As anthropogenic stressors such as climate change push further phase shifts in coral reefs, octocorals are emerging as increasingly dominant taxa (Reverter et al., 2022). In particular, the Indo-Pacific soft coral family Xeniidae exhibit key traits that allow them to thrive under increasingly unstable oceanic conditions, such as resilience to elevated temperature, eutrophication, and acidification (Gabay et al., 2014; Vollstedt et al., 2020; Mezger et al., 2022). High fecundity and prolific asexual reproduction also contribute to their ability to rapidly colonize substrates (Benayahu & Loya, 1985; Ben-David-Zaslow et al., 1999), including degraded reefs (Stemmer et al., 2013; Bruckner & Dempsey, 2015). These traits have enabled xeniids to recently become aggressive invasive species throughout the Caribbean and South America (Schizas et al., 2026; Mantelatto et al., 2018; Ruiz-Allais, Benayahu & Lasso-Alcalá, 2021; Bolick, Lee & Bolick, 2023; Toledo-Rodriguez et al., 2025). Additionally, prolific regeneration has been reported in these soft corals (Hu et al., 2020; Nadir, Lotan & Benayahu, 2023), though it is still understudied. As fragmentation and clonal propagation have already been shown to enhance spread of invasive xeniids (Ruiz-Allais, Benayahu & Lasso-Alcalá, 2021), regeneration has therefore emerged as a key trait that is hypothesized to heighten invasion severity as well as lead to expansion in their natural habitats after disturbance (Toledo-Rodriguez et al., 2025). Our previous study (Nadir, Lotan & Benayahu, 2023) showed that polyps of Xenia umbellata can regenerate from detached tentacles, an extreme form of WBR previously reported in very few cnidarians, none of them octocorals (Lesh-Laurie, Hujer & Suchy, 1991; Pearse, 2002; Toh & Ng, 2016; Kaposi, Courtney & Seymour, 2021). Yet, the resulting rapid development of colonies contained multiple polyps, suggesting that the limits of their regenerative potential had not been fully examined. In scleractinians, WBR has been reported from miniature fragments of several species (e.g., Fungia granulosa, Oculina patagonica, and Dipsastraea favus) as small as 0.3–0.5 mm in diameter (Vizel et al., 2011), but this ability has never been tested in octocorals. As such, regeneration warrants closer investigation in the latter group both within the framework of xeniids as successful colonizers and invaders of the reef as well as within their application as model cnidarians (Nadir, Lotan & Benayahu, 2023). Positional memory, or maintenance of polarity, is a key feature of regeneration across all animal phyla (Bideau et al., 2021; Poss & Tanaka, 2024). Coral morphogenesis, and thus positional information, is complicated by the fundamental unit of selection being the colony (Hiebert, Simpson & Tiozzo, 2021), with colonies possessing complex growth forms that vary due to intrinsic as well as environmental factors (Re et al., 2024). In our previous study, amputation experiments in settled individual polyps excised from X. umbellata colonies showed that positional information was relevant for regeneration. When most of the original polyp was maintained, or a clearly defined region remained, the missing part was regenerated to form a complete polyp that could bud daughter polyps to form a small colony (Nadir, Lotan & Benayahu, 2023; Nadir et al., 2025). Thus, the excised individual polyps could reform a basal attachment disc, amputated oral discs could be regenerated by the attached polyp body, and oral discs cut from the polyp could themselves reform the attachment disc and settle. Experiments on Hydra vulgaris and Nematostella vectensis showed that when both oral and aboral ends are removed, polyps reform them in their original locations, indicating that positional information is maintained (Bode, 2012; Röttinger, 2021). In contrast to these solitary cnidarians, X. umbellata is colonial, which may dictate regeneration progression towards coloniality. Experiments in the distantly related colonial hydrozoans show a variety of polarity effects, with some, such as Podocoryne carnia, retaining polarity and regenerating both distal (oral disc) and proximal (colony tissue or stolon) parts after polyp amputations, and some, such as Hydractinia echinata, expressing differing, colony-specific regeneration patterns where original polarity is destroyed (Müller, Plickert & Berking, 1986; Mayorova et al., 2015; Bradshaw, Thompson & Frank, 2015). As such, further research on polarity patterning and positional information in X. umbellata, as a novel model for regeneration in colonial corals, is expected to provide insights into regenerative patterning in cnidarians. Here, we investigate WBR and polarity repatterning in X. umbellata polyps. First, we hypothesized that WBR can occur from multiple polyp tissues, specifically the tentacles and the polyp body, but that regeneration progression would differ depending on fragment origin (tentacle vs. polyp body) and initial size. Second, we hypothesized that positional information would be retained following complete removal of both oral and basal structures, such that double-amputated polyps would exhibit biased morphogenesis and polarity during regeneration. By testing these hypotheses, our study extends the comparative framework of cnidarian regeneration to octocorals and highlights how modular body organization shapes regenerative outcomes. These findings also have implications for understanding coral resilience and invasive potential, and provide a foundation for future studies of regeneration and polarity in colonial anthozoans. Materials & Methods Rearing of X. umbellata and regeneration experiments Colonies of X. umbellata were maintained in a flow-through water table at the Mevo’ot Yam Aquaculture Center, Israel (32.404°N, 34.868°E) at 23–25 °C, pH 8.1, and salinity 38 ppt. Individual polyps were prepared in 12-well plates, brought to the lab at Tel Aviv University, and maintained in MRC Bod Model 150 incubators with continuous air circulation at 25 °C, in a 12L:12D light regime with light level of 50–57 µmol m−2 s−1 as described previously (Nadir, Lotan & Benayahu, 2023). Settled individual polyps were prepared from colonies, brought to the lab, and acclimated for 3–4 days as described previously (Nadir, Lotan & Benayahu, 2023). Specifically, each polyp was haphazardly selected from a different colony to maximize variability. These polyps were anesthetized using 0.2 M linalool (Goel et al., 2019) (Sigma-Aldrich L2602). Initially, a small group of intact, healthy polyps (n = 4) were preserved after anesthesia in 4% glutaraldehyde and used for histology as controls (see below). To determine whether regeneration was possible from miniature fragments of the polyp, two regeneration experiments were conducted from: (1) tentacle fragments and (2) polyp body fragments. Once fully immobilized, tentacles and polyp bodies were excised using Storz E3360 scissors. For the tentacle experiment, one tentacle from each polyp was selected, for a total of 40 tentacles, then tentacles were each transversely sectioned under an Olympus SZX16 dissecting microscope with the scissors into five fragments. Polyp bodies were cut by first amputating polyps ∼3–5 mm above their attachment discs, then removing the oral disc similarly 3–5 mm below the connection to the oral disc. A total of 20 polyp bodies were selected and cut in this way. Polyp bodies were each transversely sectioned into four to six cross sections about 150–200 micrometers long, then each further halved, yielding eight to 12 fragments per polyp. This method of fragmentation allowed cutting the polyp body into small pieces with minimal loss of mesenterial tissue and body fluids. Fragments for each experiment were pooled together during the fragmentation process until ∼200 fragments had been generated, then fragments were mixed and randomly placed individually in tissue culture plates. Each experiment was monitored for 30 days (tentacle experiment, March–April 2022, n = 122, polyp body experiment, n = 120, April–May 2022). Fragments were maintained in 24-well tissue culture plates filled with 0.45 µm filtered seawater (FSW). Plates were monitored and imaged three times weekly, with 50% FSW replaced after each observation. Following each handling, plates were cyclically rotated within the incubator to minimize positional environmental effects. The initial fragment projection area (mm2) at commencement of experiments (day 0) was measured for both tentacle and polyp body fragments using ImageJ software (Abràmoff, Magalhães & Ram, 2004), and was defined as the area contained within the fragment when viewed from above. Thereafter, regeneration was determined three times per week using an Olympus SZX16 dissecting microscope equipped with an Olympus DP74 camera until the end of the experiments (day 30). In addition, to characterize the morphological features of different developmental stages, 4–5 representative samples for each regeneration stage from both experiments were examined by scanning electron microscopy (SEM: see below). To investigate the histological features of the regenerated fragments removed both from the polyp bodies and the tentacles additional experiments were repeated as detailed above (July–October 2025, n = 216 for tentacle fragments and n = 180 for polyp body fragments). These were compared to samples of intact control animals. Polyp body double amputation To determine effects of removing both the attachment disc and the oral disc, settled individual polyps were carefully detached from the bottom of the tissue wells, anesthetized and cut on both ends so that the remaining polyp bodies (1–2 cm in length after cut) were maintained without the oral disc or the attachment disc. Double-amputated polyp bodies (n = 36) were then placed into 12-well tissue culture plates with fresh 0.45 µm FSW and similarly maintained in incubators (see above). As attempts to mark the polyp ends with dye were unsuccessful, original polarity could not be accurately determined. Regeneration was assessed three times per week using an Olympus SZX16 scope as stated above and monitored for 29 days, after which the experiment was ended. Photographs and video recordings of pulsation development were taken throughout the experiments to document morphogenetic processes and regeneration over time. At the end of the experiment, all surviving regenerates were preserved in 4% paraformaldehyde (PFA, Bar-Naor 15710) and maintained at 4 °C until processed for dissection. Animals were then sliced in half longitudinally using Storz E3360 scissors to display their gastrovascular cavity and imaged with the scope. Histology Samples were fixed in 4% glutaraldehyde (Sigma, 354400) in filtered sea water, and kept at 4 °C until processed for histology. Following rinsing in distilled water, samples were decalcified for 15 min to dissolve the sclerites using a 1:1 mixture of 50% formic acid in distilled water and 20% trisodium citrate. Samples were then washed 3 times in distilled water and kept in 70% EtOH. Paraffin-embedded samples were sectioned into 6 µm-thick cross-sections and stained with hematoxylin and eosin and observed under a Nikon Eclipse 80i Microscope and filmed with attached Olympus DP74 camera. Scanning electron microscopy (SEM) SEM samples were fixed in 4% glutaraldehyde in FSW. Samples were kept until processed at 4 °C, then washed three times in PBS (Biological industries, 02-023-1A), dehydrated through a graded series of ethanol concentrations up to 100%, and critically point-dried with liquid CO2. Finally, the samples were gold-coated and examined under a JCM 6000 NeoScope Benchtop scanning electron microscope (JEOL Tokyo, Japan). Statistics Time to regenerate an early polyp from fragments of both polyp bodies and tentacles was analyzed using linear regression. Data normality for differences in initial projection area between fragments that successfully regenerated and those that failed to regenerate was assessed using the Shapiro–Wilk test. As data did not meet assumptions of normality, it was analyzed using the non-parametric Mann–Whitney U test. A significance threshold of p < 0.05 was applied. Fragments that did not regenerate (i.e., remained in the “cut” stage or died) or were sampled prior to reaching the early polyp stage were excluded from these analyses. Animal welfare and ethics X. umbellata colonies used in the current study were obtained from an educational coral farming facility. This species is not under CITES and, therefore, does not require ethics approval. Results Tentacle and polyp body fragments formed tissue balls before initiating regeneration into polyps Fragments from both the tentacles and the polyp body (n = 122 and 120, respectively) exhibited similar morphological stages of WBR into new polyps (Figs. 1A and 1B). The results enabled assignment of the key regeneration stages of the animals occurring during progression of the experiments (Table S1): (1) cut—day 0, no regeneration was initiated, (2) initial regeneration—showing wound closure and initial rounding of the fragment outline, (3) tissue ball—rounded, smooth structure, (4) elongated structure, (5) early polyp—appearance of mouth and tentacle buds, and (6) late polyp—elongated pulsating tentacles (Figs. 1A–1B). Development of the tentacle pulsation of the regenerated animals was analyzed following Nadir et al. (2025), and showcased the same pulsation regeneration stages as described for oral disc regeneration, namely: (a) initial pulsation—slight motion of the tentacles, (b) asynchronous pulsation and (c) synchronous pulsation similar to an original polyp (ranging from 0.4–0.6 Hz, Video S1) all of which were pooled as the late polyp stage for the purposes of this study. These regenerated polyps had oral discs 0.5–1 mm in diameter, corresponding to about 5–10% of the original polyp diameter. Notably, on day 30 of both experiments, over 75% of surviving fragments in both experiments had developed into late polyps (Figs. 1C and 1D) (n = 37 for the tentacle fragments, n = 35 for polyp body fragments). Early regeneration was apparent 1–2 days following amputation and at this stage sloughed off tissue pieces, symbiotic algae, and sclerites were observed around most fragments, indicating tissue disintegration or death (Fig. S1). Tentacle and polyp body fragments exhibited morphological differences in appearance during the first two stages, with tentacle fragments still showing their original pinnules (Figs. 1A and 1B), but both appeared externally identical from the tissue ball stage onwards. Whole body regeneration from tentacle and polyp body fragments in Xenia umbellata. (A and B) SEM of regeneration of tentacle fragments (A) and polyp body fragments (B). Cut, early regeneration, tissue ball, elongated structure, early polyp and late polyp stages shown. Fragments lose differentiated structures (e.g., pinnules) then form into tissue balls and attach to the surface, reforming the polyp oral disc at the opposite end of attachment. Scale bar 0.5 mm. Arrows show pinnules in original fragment and regenerated polyps. (C and D) Percentage of different regeneration stages over experiment time (days) in regenerates originating from tentacle fragments (C) and polyp body fragments (D) throughout 30 days of regeneration. (E and F) Scatter box plots for initial size of fragments, comparing initial projection areas for fragments that began regeneration but died before forming polyps, and those that managed to regenerate into polyps for the tentacle (E) and polyp body (F) experiments. Asterisks represent significant difference in initial size between polyp body fragments that developed into polyps (Mann–Whitney-U test, p < 0.01). Although regeneration progressed similarly in both experiments, the timing differed. By day 4, most tentacle fragments were still in early regeneration (45%), while only 18% reached tissue balls; in contrast, 63% of polyp body fragments reached the tissue ball stage, with just 3% remaining in the early regeneration stage (Figs. 1C and 1D). Corresponding differences in regeneration timing persisted throughout the experiment time, reflected for instance in a continuously higher proportion of early polyps to late polyps in the tentacle experiment (Figs. 1C and 1D). Notably, the overall results were similar: the earliest polyps appeared by day 6 of both experiments, and after 2 weeks around half of polyp fragments had already transformed into polyps in both experiments. Mortality was almost similar for both experiments, with a total of 55 (45%) and 51 (43%) fragments disintegrating (dying) throughout the tentacle and polyp body experiments, respectively. The majority died in initial stages of regeneration before day 10 (Table 1). Those that persisted were mostly ‘arrested’ in early stages, such as tissue balls, before finally disintegrating at the end of the experiment (day 30), though a minority of fragments died after advancing to a polyp stage (Table 1). 10.7717/peerj.21503/table-1Table 1Fragment regeneration stage of Xenia umbellata at death.Mortality shown by tissue origin (tentacle or polyp body), and presenting number and proportion of fragment mortality per regeneration stage out of total mortality in that experiment. Total mortality shown out of total fragments per experiment. Stage of death Tentacle Polyp body Cut26/55 (47%)7/51 (14%)Early regeneration2/55 (4%)0/51 (0%)Tissue ball16/55 (29%)31/51 (60%)Elongated structure5/55 (9%)4/51 (8%)Early polyp6/55 (11%)8/51 (16%)Late polyp0/55 (0%)1/51 (2%)Total mortality55/122 (45%)51/120 (43%) Fragment regeneration stage of Xenia umbellata at death. Mortality shown by tissue origin (tentacle or polyp body), and presenting number and proportion of fragment mortality per regeneration stage out of total mortality in that experiment. Total mortality shown out of total fragments per experiment. Initial fragment size has minimal effect on regeneration To determine effects of initial fragment size on successful regeneration, we tested whether initial projection area (mm2) affected ability to develop into polyps. Fragments that developed into early polyps (“successful”) and those that failed, but managed to begin regeneration, and remained in the intermediate, tissue ball, or smooth tissue stages (“unsuccessful”) were compared (Figs. 1E and 1F). Initial fragment projection areas varied between 0.06 and 0.93 mm2. No difference was found in the tentacle experiment between successful and unsuccessful fragments (Mann–Whitney U = 480, P = 0.7999). However, successful polyp-body fragments were significantly larger than unsuccessful ones (0.33 mm2 vs. 0.26 mm2; Mann–Whitney U = 1597, P = 0.0061). To determine whether initial size is related to the rate of regeneration for fragments that form polyps, linear regression was employed to compare initial size to time to form a polyp (i.e., the first day an early polyp was recorded for that fragment). While a slight inverse relationship was found between time to form a polyp and initial projection area, R2 was low for both experiments (tentacle linear regression, R2 = 0.16, F (1,47) = 8.79, p = 0.005, polyp body linear regression, R2 = 0.09, F (1,63) = 6.22, p = 0.015) indicating that initial fragment size had a minor but not negligible effect on regeneration time, supporting our hypothesis that size influences regeneration progression, albeit weakly. Early regeneration differs between tentacle and polyp body fragments owing to distinct initial structural remodeling requirements Further analysis of early regeneration stages of tentacle and polyp body fragments revealed that early day 2 tissue balls were spherical and unattached, whereas from day 4, tissue balls and all regenerating fragments, all regenerating fragments for both experiments were attached to the well surface. The histological examination showed that all fragments maintained the characteristic tissue and cellular components of the original polyps throughout regeneration (Fig. 2). Cell layers, namely distinguishable epidermis and gastrodermis, with the mesoglea in between, were maintained, with sclerites in the mesoglea, symbiotic algae (zooxanthellae) in the gastrodermis, and gland cells in the epidermis (Fig. 2). Tentacle fragments (Fig. 2A) lost differentiated tentacle structures, with pinnules still visible but reduced in size by day 1 (Fig. 2B), and day 2 tissue balls forming a hollow, spherical shape devoid of pinnules (Fig. 2C). As tentacle fragments initially lacked mesenteries (Fig. 2A), early regeneration and early tissue ball stages were also mesentery-free, but by day 4 attached tentacle tissue balls had reformed mesenteries (Fig. 2D). Mesenteries were then maintained in the elongated structure phase, which was hollow and cylindrical (Fig. 2E), while a distinct pharynx was visible in early polyps (Fig. 2F). In the polyp body fragments (Fig. 2G), regeneration appeared to progress more rapidly, with samples from day 1 already forming a rounded shape (Fig. 2H) and turning spherical by the following day (Fig. 2I). Mesenteries which were already present in the original polyp body fragment (Fig. 2G) were maintained throughout regeneration. Day 4 tissue balls showed structural changes both externally and internally, as fragments lost their rounded shape and mesenteries were less pronounced (Fig. 2J). As regeneration progressed, the elongated structure phase appeared hollowed out, though mesenteries were still present (Fig. 2K). Early polyps exhibited distinct pharynx and tentacle buds (Fig. 2L). Morphogenesis of both tentacle and polyp body fragments appeared identical from the elongated structure stage onwards (day 6) (Figs. 2E–2F, 2K–2L). Distinct early remodeling between tentacle and polyp body fragments confirms that regeneration progression is affected by fragment origin. Histological sections of early regeneration stages of tentacle (A–F) and polyp body (G–L) fragments of Xenia umbellata. (A) Closeup of tentacle from intact control animal. (B) Early regeneration stage of tentacle. (C) Early, unattached tissue ball of tentacle. (D) Advanced, attached tissue ball of tentacle. (E) Elongated structure of tentacle. (F) Early polyp from tentacle fragment. (G) closeup of polyp body cross section from intact control animal. (H) Early regeneration from polyp body. (I) early tissue ball from polyp body. (J) Advanced, attached tissue ball from polyp body. (K) Elongated structure from polyp body. (L) Early polyp from polyp body fragment. ep, epidermis; gc, gastrovascular cavity; gd, gastrodermis; gl, gland cells; m, mesoglea; ms, mesentery; p, pinnule; ph, pharynx; sc, sclerites (space left by decalcified sclerites); zx, zooxanthellae (photosymbionts). Scale bars 100 μm. Scale bar at B applies to (B–F) and (H–L). Polyp polarity is partially maintained by double amputation Polyp bodies cut on both ends (Fig. S2) regenerated oral discs on one (25%) or both ends (64%) (Fig. 3, Table 2), which followed the same regeneration stages as previously reported oral disc regeneration (Nadir, Lotan & Benayahu, 2023). Namely, by day 2 polyp bodies had healed the cuts from amputation and formed mouths (Figs. 3A–3C). By day 5 tentacle buds had formed (Figs. 3D–3F), and by day 7 tentacles were clearly visible (Figs. 3G–3I). By day 9 most tentacles had developed 6–9 pairs of pinnules along each tentacle (Figs. 3J–3L) and by day 12, most were pulsating regularly as in the intact polyps, indicating that regeneration of the oral disc was functionally complete (0.5 Hz, Figs. 3M–3O, Video S2). The process was followed for an additional 17 days, during which many regenerates produced polyp buds, thus confirming reformation of a colonial structure. Regeneration process of double amputated polyps of Xenia umbellata. Rows indicate days (D) post amputation, columns indicate regeneration outcomes. Regenerates develop asymmetrically, and can develop into an attached, single oral disc polyp (A, D, G, J, M, P), or a double-oral disc polyp that is either attached (B, E, H, K, N, Q) or unattached (C, F, I, L, O, R). P1, Q1, and R1 show corresponding internal structure in longitudinally halved regenerates of P, Q, and R. P1 shows separate mesenterial filaments between the regenerated polyp and its daughter bud, which do not intersect and are perpendicular to each other. Inset shows closeup of attachment disc and daughter bud. Q1 and R1 show a continuous shared polyp body between the double-polyp, with shared mesenteries. Scale bar one mm. m developing mouths in early regeneration of double-amputated polyps. Arrowheads point to attachment discs. Asterisks denote the smaller oral disc in double-oral disc regeneration. Arrows point to mesenterial filaments. 10.7717/peerj.21503/table-2Table 2Proportion of outcomes for double-amputated polyp bodies of Xenia umbellata.Proportion refers to described groups on the left. Groups referred to include polyps that developed one or two oral discs, or died, two-oral disc polyps that attached or did not attach to the well surface, and percentage of attached polyps that developed their attachment discs asymmetrically, closer to the smaller oral disc. Total refers to total double-amputated polyp bodies in experiment, n = 36. Outcome/group Proportion One oral disc/total9/36 (25%)Two oral disc/total23/36 (64%)Attached/two oral disc15/23 (65%)Unattached/two oral disc8/23 (35%)Asymmetrical attachment/attached two oral disc13/15 (87%)Dead/total4/36 (11%) Proportion of outcomes for double-amputated polyp bodies of Xenia umbellata. Proportion refers to described groups on the left. Groups referred to include polyps that developed one or two oral discs, or died, two-oral disc polyps that attached or did not attach to the well surface, and percentage of attached polyps that developed their attachment discs asymmetrically, closer to the smaller oral disc. Total refers to total double-amputated polyp bodies in experiment, n = 36. All polyps with two oral discs initially began regenerating their two sides asymmetrically (Figs. 3E–3F), with one side developing an oral disc at a delay (e.g., Fig. 3K), indicating that positional information from the original polarity was partially retained. As regeneration progressed, most oral discs eventually ‘caught up’ to each other and appeared symmetrical (e.g., Fig. 3R). Single-oral disc polyps were attached to the well surface by a median of day 12 (Fig. 3M). Most two-oral disc polyps began developing attachment discs along the bottom of the polyp body structure (Fig. 3K) by a median of day 9 (65%), while the rest remained unattached (Figs. 3L, 3O and 3R). Attachment discs became fully attached by a median of day 16, though for some attachment started as early as day 12 (Fig. 3N). Attachment disc development was biased; appearing closer to the smaller oral disc in 87% of attached two oral disc polyps (Figs. 3K, 3N and 3O, Table 2), while in the rest it appeared in the middle of the polyp body. Preserved samples of final regenerates were cut in half longitudinally to reveal differences in internal structure between the three main experimental outcomes, namely one-oral disc polyps that underwent budding (Fig. 3P), and two-oral disc polyps with (Figs. 3Q) and without (Fig. 3R) an attachment disc. Notably, while polyp buds connected to the original polyp through the attachment disc (Fig. 3P1), polyp bodies that developed two oral discs maintained connected mesenteries between the polyps, reflecting their initial origin as a single polyp body, irrespective of the presence of an attachment disc (Fig. 3Q1 and Fig. 3R1). In contrast, daughter bud mesenteries were perpendicular to the main mesenterial axis (Fig. 3P1, inset). Overall, asymmetrical regeneration of oral discs and biased development of attachment discs are patterns consistent with partial retention of positional information, supporting our hypothesis that positional information would be retained, though validating these findings by tracking original polyp polarity requires further studies. Discussion This study demonstrates WBR in polyps of X. umbellata, exhibiting complete regeneration from miniature fragments with projection areas as small as 0.1 mm2 derived from both tentacles and polyp bodies. Although we hypothesized that regeneration would depend on both fragment origin and size, our results only partially supported this prediction. Initial size significantly affected regeneration, but the effect was relatively weak. Notably, size was approximated by projection area, a two-dimensional measure that does not capture other factors such as volume or internal structure. Thus, alternative measures of size may reveal stronger effects. In addition, fragmentation-related factors, such as loss of internal fluids or wound perimeter (Henry & Hart, 2005) may also shape the trajectory of regeneration. In contrast, fragment origin had a clear effect on regeneration progression, strongly supporting our initial hypothesis. Early regeneration dynamics differed between fragment types: polyp body fragments regenerated more rapidly and largely retained internal organization, whereas tentacle-derived fragments underwent extensive structural remodeling, including pinnule resorption, formation of a gastrovascular cavity, and de novo mesentery development (Figs. 1 and 2). Despite these early differences, regeneration trajectories converged after the tissue ball stage, resulting in externally and internally similar, functional polyps, a fraction of the original polyp size. This convergence indicates a shared downstream regeneration program once missing structures are re-established. Comparable patterns of divergent early regeneration followed by convergence have been reported in WBR in some marine invertebrates. In adult Cassiopea xamachana, fragments lacking oral arms regenerated more slowly at early stages but showed no differences at later stages once oral structures had reformed (Ostendarp et al., 2022). As in our study, C. xamachana also showed limited effects of initial size on regeneration success. In sponges, fragments originating from different body regions initially follow distinct morphogenetic pathways but reorganize into a common functional body plan once the aquiferous system is reformed (Ereskovsky et al., 2021). In X. umbellata, internal structures such as mesenteries, which provide mechanical support in intact polyps (Fabricius & Alderslade, 2001), may facilitate progression through early regenerative stages, consistent with findings in N. vectensis, where mesenteries are required for WBR (Amiel et al., 2019). Tissue ball formation suggests conserved regeneration mechanisms across cnidarians The formation of tissue balls during regeneration in anthozoans has previously been reported in scleractinian corals, where it was initially described as a response to extreme stress when conventional tissue regeneration is not possible (Kramarsky-Winter & Loya, 1996). In the species F. granulosa, O. patagonica, and D. favus, excised tissue slivers contract into spherical tissue balls, and subsequently settle and regenerate new polyps (Vizel et al., 2011). Similar processes have been documented in the invasive scleractinians T. coccinea and T. tagusensis, where excised tissue fragments form spherical aggregates on top of skeletal tissue, after which these spheres form mouths and bud new polyps over weeks to months (Luz, Miller & Kitahara, 2021). Intriguingly, some scleractinian corals, such as Pocillopora damicornis and Goniopora lobata, have been reported to form tissue balls from tentacle or colony fragments, though these have not been reported to regenerate (Feuillassier et al., 2014; Lu et al., 2017). The tissue balls observed during WBR in X. umbellata appear, at least from our morphological results, strikingly like those described in scleractinians, both externally and internally. In both systems, tissue balls maintain organized cell layers and retain symbiotic zooxanthellae within the gastrodermis. One notable difference, however, is that scleractinian tissue balls initially lack a calcium carbonate calyx, whereas X. umbellata tissue balls retain their skeletal sclerites throughout regeneration (Fig. 2), reflecting fundamental differences in skeletal architecture between the two groups (Fabricius & Alderslade, 2001). As in scleractinians, tissue ball attachment to the substrate represented a critical transition during regeneration in X. umbellata, coinciding with substantial internal reorganization (e.g., Fig. 2D). Despite these similarities, regeneration in X. umbellata proceeded considerably faster, with most surviving fragments forming functional polyps within one month, whereas mouth formation in scleractinian tissue balls reached only ∼45% after nine weeks even under optimal conditions (Vizel et al., 2011). Furthermore, in both X. umbellata (Fig. S1) and the studied scleractinians, early tissue sloughing and reorganization into rounded aggregates preceded attachment, internal restructuring, and polyp reformation, suggesting that partial tissue loss and large-scale tissue rearrangement are integral components of successful regeneration in both of these anthozoan groups (Kramarsky-Winter & Loya, 1996; Luz, Miller & Kitahara, 2021). Differences in regeneration rate and skeletal context between X. umbellata and scleractinians can be hypothesized, therefore, to reflect lineage-specific constraints rather than fundamentally distinct mechanisms. However, tissue ball formation is not exclusive to colonial anthozoans, and has been studied extensively in hydrozoans as well as other basal metazoans. In both colonial and solitary hydrozoans, rings or cross-sections of tissue from the polyp body column have been reported to become spherical before regenerating into polyps or stolons (Müller, Plickert & Berking, 1986; Ferenc et al., 2021). Indeed, WBR in some sponges also exhibits tissue fragments folding into tissue spheres (Ereskovsky et al., 2021), and placozoan regeneration observations show an intermediate spherical phase (Schwartz, 1984; Osigus et al., 2022). This specific form of regeneration could be dependent on small starting size or a specific shape of the tissue fragment; for instance, in X. umbellata, entire tentacles will regenerate colonies but will not go through a tissue ball stage (Nadir, Lotan & Benayahu, 2023). Similarly, Hydra regeneration from body fragments depends on size and shape of original fragments, with some fragments turning into spheroids (Bode & Bode, 1984), and sponge tissue ball formation depends on the fraction of the original sponge in the regenerating fragment (Ereskovsky et al., 2021). Together, such results suggest the tantalizing hypothesis that regeneration from miniature tissue fragments via a tissue ball intermediate represents a conserved morphogenetic strategy among cnidarians, and perhaps even other basal metazoans. Tissue ball formation may simply be a factor of mechanical, or combined mechano-chemical constraints on small tissue fragments, as has been shown in Hydra (Wang et al., 2019; Ferenc et al., 2021; Perros et al., 2024). Notably, while some bilaterians are highly regenerative, this specific mechanism of WBR may not be maintained in these groups. For instance, planarians don’t form tissue balls even during regeneration from equivalently tiny fragments as in the current study (Montgomery & Coward, 1974; Reddien & Sánchez Alvarado, 2004; Ivankovic et al., 2019). While many colonial urochordates are capable of WBR from small fragments of vascular tissue, and spherical structures may be transiently formed, regeneration occurs due to circulating totipotent blood cells, and frequently results in multiple transient buds which are reabsorbed, reflecting a potentially very different regeneration strategy from cnidarians (Rinkevich, Shlemberg & Fishelson, 1995; Rinkevich & Rinkevich, 2025). Future comparative studies of the tissue balls across different taxonomic groups may reveal intriguing similarities or differences. Partial retention of positional information shapes polarity during regeneration in Xenia umbellata Positional information plays a central role in polarity patterning during regeneration across animal phyla and is often retained during WBR in highly regenerative species, including oligochaetes, planarians, and cnidarians such as H. vulgaris and N. vectensis (Bode, 2012; Reddien, 2018; Kostyuchenko & Kozin, 2021; Boilly, Hondermarck & Boilly-Marer, 2022; Mazloumi Gavgani et al., 2025). Although our study did not track original polarity, our results show patterns consistent with partial retention of positional information following double amputation in X. umbellata, with influences on polarity patterning during regeneration (Fig. 3). Several observations provide support to this conclusion. First, polyps that regenerated a single oral disc likely retained their original axial polarity, similar to double-amputated N. vectensis polyps, which regenerate both ends according to pre-existing positional information (Reitzel et al., 2007). Second, in polyps that regenerated two oral discs, formation of the second disc was consistently delayed, resulting in transient oral disc asymmetry during early regeneration. This suggests that residual positional cues may initially inhibit or bias secondary oral disc formation. Third, attachment discs most frequently formed closer to the smaller regenerated polyp (Table 2), further indicating that positional information from the original polyp continued to influence colony-level polarity and morphogenesis. Although attachment to the substrate is a vital component of xeniid biology, attachment disc formation occurred after oral disc regeneration in double-amputated polyps (Fig. 3), suggesting a certain hierarchy in regenerative priorities. Early oral disc and tentacle formation may be advantageous for harvesting photosynthates and thus supplying energy for regeneration, whereas attachment likely reflects a subsequent thigmotropic response to contact with the substrate, consistent with previous observations in both single polyps and colonies of X. umbellata (Kim, Wild & Tilstra, 2022; Nadir, Lotan & Benayahu, 2023). Notably, two-oral-disc polyps remained internally continuous above the attachment disc and displayed shared gastrovascular cavities and mesenteries (Fig. 3), indicating that while an early colony-like structure emerged, the shared origin of the two oral discs was retained. In summary, the current findings support our hypothesis that positional information is maintained following double amputation in polyps of X. umbellata, but also reveal a level of complexity not predicted from studies in some solitary cnidarians (Bode, 2012; Van der Burg et al., 2020; Röttinger, 2021). Rather, studies on colonial hydrozoans may provide a more relevant framework for understanding organizing principles underlying colonial regeneration in X. umbellata. Hydrozoan colonies exhibit diverse regenerative patterns. For example, in P. carnea, polyps can regenerate the stolon (the colony tissue connecting polyps) and the oral disc (Sanders, Travert & Cartwright, 2020) whereas in H. echinata, polyps excised from the colony transform into stolons (Bradshaw, Thompson & Frank, 2015). Remarkably similar results to this study were reported following double amputation in H. echinata where double-oral disc polyps initially formed, followed by the later development of stolon-like tissue between them, which eventually attached to the substrate, though in this case no asymmetry was reported between the developing oral discs (Müller, Plickert & Berking, 1986). These observations suggest that colony-derived signaling, or the loss of such signaling may partly explain our results, though H. echinata is distantly related to X. umbellata and possesses a markedly different colonial organization. Further work will be required to identify the mechanisms underlying the partial retention of positional information in X. umbellata. Positional cues may be encoded through molecular gradients, as proposed for axial patterning in other regenerating systems (Simsek & Özbudak, 2022), and maintained by organizer-like signaling centers analogous to the head organizer described in H. vulgaris (Bode, 2012). Notably, colonial hydrozoans have been shown to possess organizers both in the ‘head’ region as well as in shoots of growing stolons, with specific axial patterning genes expressed in colony tissues (Mayorova et al., 2015; Sanders, Travert & Cartwright, 2020). Alternatively, loss of the oral disc may trigger a default oral regeneration program activated by wounding in the absence of inhibitory cues, as reported in both N. vectensis and H. vulgaris (Cazet, Cho & Juliano, 2021; Mazloumi Gavgani et al., 2025). These proposed mechanisms fit our observations, but remain to be verified by molecular testing. The nervous system may represent an additional candidate mechanism underlying the observed regeneration. In X. umbellata, neural organization includes a specialized oral pacemaker structure (Nadir et al., 2025), and neural signaling contributes to positional information during regeneration across multiple animal phyla (Sinigaglia & Averof, 2019). Supporting this, recent work in N. vectensis showed that neuronal ablation induces double-headed regeneration following basal amputation, indicating that neurons normally suppress ectopic head formation (Mazloumi Gavgani et al., 2025). However, this role is not conserved across cnidarians, as nerve-free Hydra regenerate normally (Marcum & Campbell, 1978), suggesting that neural contributions to polarity are lineage-specific and context-dependent. As the current study did not describe neural morphology during regeneration, its relevance in this species remains unresolved. In colonial organisms, additional layers of organization complicate polarity regulation compared to unitary organisms. In hydrozoan and anthozoan colonies, colony form and polarity is affected by morphogenetic factors, but also ecological and physiological factors (Cartwright, Travert & Sanders, 2021; Re et al., 2024). Several studies have reported evidence for apical dominance in highly integrated coral colonies—a form of polarity control previously thought to be largely restricted to plants, including seagrasses, involving inhibition of growth of lateral branches by the apical branch (Sánchez & Lasker, 2004; Re & Duarte, 2025). Despite variability in early regeneration outcomes, double-amputated X. umbellata polyps ultimately converged on a similar colony-like morphology and produced daughter polyp-buds by the end of our experiment. Together, these findings suggest that polarity patterning in X. umbellata cannot be fully explained by frameworks derived from solitary cnidarians, including other anthozoans, or by known mechanisms in more distantly related colonial cnidarians. Elucidating how molecular, neural, and colony-level processes interact during regeneration in this species will therefore provide valuable insight for comparative studies of regeneration, polarity, and coloniality among cnidarians and beyond. Conclusions The octocoral X. umbellata exhibits impressive WBR, regenerating complete, functional polyps from miniature fragments derived from both tentacles and polyp bodies. The development of tissue balls during WBR revealed superficial similarities to other cnidarian tissue balls that should be further tested for potential shared mechanisms of WBR. We observed patterns consistent with partial retention of positional information following double amputation, as over 60% of fragments regenerated two oral discs, often exhibiting asymmetric development and biased attachment-disc positioning despite the complete loss of axial landmarks. These patterns suggest that polarity repatterning in X. umbellata integrates residual positional cues with colony-level patterning processes. Together, these findings highlight fundamental differences between regeneration in this species and previously studied solitary and colonial cnidarians, thus establishing X. umbellata as an inspiring comparative model for studying WBR, polarity, and colonial organization. Future research should focus on deciphering the molecular and cellular basis of the regenerative processes described here, supported by comparative analyses with established cnidarian models. Additionally, the remarkable regenerative plasticity observed in X. umbellata may contribute to its ecological success. Based on our study, it is suggested that high regenerative capacity and flexible polarity repatterning may enhance the previously studied octocoral resilience to increasingly frequent environmental disturbances, facilitating X. umbellata’s spread and invasive potential in non-indigenous biogeographic regions, though this remains to be determined in naturalistic settings. Supplemental Information Sloughed off dead tissue near or surrounding newly formed tissue balls from (A) tentacle fragments and (B) polyp body fragments Arrowheads point at sloughed off tissue next to or under tissue balls, in which brown zooxanthellae can be clearly seen. Scale bar 0.5 mm. Tissue balls showcase range of sizes of tissue balls formed. Both tissue balls shown later regenerated into polyps. Original polyp before and after amputation for double-amputation experiment (A) Attached polyp before amputation, amputation planes marked with dashed lines. (B) Polyp body after double amputation. Scale bar two mm. Pulsating miniature polyps from WBR Final miniature polyps from WBR from tentacle (left) and polyp body (right) experiments, corresponding to Figs. 1A and 1B. Pulsation in double amputated polyps with two oral discs Top—double oral disc polyp with attachment disc, corresponding to Fig. 3Q. Bottom—double oral disc polyp without attachment disc, corresponding to 3R. Asterisks mark the smaller polyp. Raw data Diagnostic descriptions of regeneration stages for WBR Whole body regeneration stages for regeneration from tentacle or polyp body fragments, as shown in Figs. 1A and 1B, with corresponding detailed diagnostic descriptions of how those stages were defined.