4D Single-Cell Spatial Transcriptomics Reveals Dynamic Morphogenetic Gradients and Regenerative Domains in Planarians

Article activity feed

1. 

   GigaScience

   Jul 10, 2026

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is …

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is crucial for polarity remodeling and blastema formation. This study provides a comprehensive spatial molecular and cellular map of regenerative processes, highlighting injury-induced spatial domains and key regulatory factors in planarian regeneration. We also provide an interactive web portal, offering a valuable resource for exploring and analyzing regeneration mechanisms in a spatiotemporal context.

   This work has been peer reviewed in GigaScience (see https://doi.org/10.1093/gigascience/giag064), which carries out single-anonymized peer review. These reviews are published under a CC-BY 4.0 license and were as follows:

   Reviewer 3：

   In the manuscript entitled "4D Single-Cell Spatial Transcriptomics Reveals Dynamic Morphogenetic Gradients and Regenerative Domains in Planarians", Han, Chen, Li et al. use high-resolution Stereo-seq on regenerating planarians to reconstruct a 4D spatiotemporal transcriptomic map of planarian regeneration. In their analysis, they recognize most of the cell types identified in planarians and are able to recover the gene expression dynamics during regeneration of the body axes at the cellular, tissue, and organismal scales. One of the main findings is the identification of injury-induced spatial domains, specifically the Anterior Regenerative Zone (ARZ). Interestingly, the ARZ is enriched in positional control gene expression in several cell types, not only in muscle. The authors identify Mediator 8 as a gene expressed in the ARZ and required for proper blastema formation. The study also provides an interactive web portal with the corresponding data. The analysis of the regeneration process in planarians using Stereo-seq provides a new and very useful strategy to understand the dynamics of gene expression integrated with cell and tissue types. Through this strategy, the authors corroborate the expression patterns of several genes already described as essential for planarian regeneration, and they identify new blastema regions comprising different expression patterns, both anterior and posterior. Among them, the study focuses on the ARZ and performs trajectory analysis, which provides a very informative view of the cellular movements and changes that occur at early stages of regeneration. The finding that Med8 is required to initiate regeneration in both wounds validates the utility of the strategy followed. The publication of an open and interactive web portal with the dataset will be a useful tool for the planarian community and for research on regenerative processes in general. However, in its present form, the study presents several weaknesses and issues that should be addressed before publication.

   Main concerns:

   The 3 main concerns are 1) the presentation of the strategy of each analysis performed is not detailed and not clear enough; 2) essential data for the present manuscript is supposedly found it Han et al. submitted, when it should be available in the present manuscript; and 3) the conclusions from the functional analysis of med8 are not accurate.

   Regarding concern 1:

   • The authors explain that they analyzed 16 animals (2 per time point) processed into 10 μm thick sections, producing a total of 353 sections. However, they do not specify whether the same number of sections were analyzed per animal, nor do they indicate the total size (or thickness) and cell number of the animals analyzed. In this regard, it may be that the number of sections analyzed per animal is shown in Figure S2A (this is not clear from the figure legend or the text). If so, why is there variation in the number of sections per animal? Is it due to differences in animal size? If so, how were these differences addressed in order to integrate the data from different animals?
   • In Figure 1F, the different parts of the blastema are divided according to the pigmented/unpigmented area. What were the criteria used to divide each blastema into three parts (proximal, middle, and distal)? Gene expression? Length of the region? This should be clarified.
   • A schematic overview of each strategy used to obtain the results would help the reader understand the procedures.

   Regarding concern 2:

   • Throughout the study, the authors refer to Han et al., submitted for the custom framework used to create the atlas. Since that study is not currently available, this important information is missing from the current manuscript.
   • Similarly, from the section "Spatiotemporal dynamics of positional gradients during whole-body regeneration" onward, the study relies on the so-called SBGs (spatially biased genes), also described in Han et al., submitted, which is not published. This is very important data that, if not published elsewhere, must be included in the current manuscript. A description of how the SBGs were obtained, together with a list and explanation, is required. Otherwise, essential information about the SBGs—on which the subsequent analyses depend—is entirely missing.
   • According to the authors, PCGs are a subset of SBGs. What is the essential difference between PCGs and SBGs?

   Regarding concern 3:

   • Functional analysis of med8. The authors refer to a role of this gene in polarity remodeling. First, this is an unclear concept, because polarity establishment and polarity remodeling are two different processes during regeneration. Second, the RNAi results presented do not support a role for med8 in polarity. The phenotype suggests that med8 is required for cell-fate specification during early stages of regeneration, since neoblasts increase but differentiated cells decrease in general. However, the results do not support a role for med8 in pole formation. The authors report a decrease in sfrp-1, but only at later stages, and no data are shown for notum. The conclusion that med8 regulates blastema growth and positional information is therefore not accurate and does not align with the results presented.

   Additional main concerns:

   • A spatiotemporal transcriptomic atlas of planarian regeneration was already published in Cui et al. 2023. Although the authors cite it in the introduction, they should also compare their results with those published previously. Do they observe similar cell types and gene expression dynamics at the time points analyzed? This comparison should appear in both the Results and Discussion sections.
   • The authors state that the different clusters of SBGs fluctuate during regeneration and then stabilize. First, which genes belong to each cluster? This information should be included. Second, according to Figure 2A, some clusters fluctuate (e.g., A1 and A2) but others do not (e.g., A5 or M9). A more accurate interpretation is required. Furthermore, the starting point of the analysis is t0, when head and tail are missing, and the endpoint is 14 days of regeneration. How can one assess whether gene expression stabilizes if the starting and ending states are completely different? Importantly, in Figure 2A, the cluster labels in the image do not correspond to the bars, and there are 15 bars but 16 clusters. The figure should be corrected. The colormap labels are also missing.
   • The authors find that predictions from the Gierer-Meinhardt model are consistent with the expression of some genes (ARNT, Ndk, EGR1…). First, a description and reference for these genes are required. Second, what about other genes involved in Wnt signaling and AP patterning, previously proposed to follow this model in Werner et al. 2015? Third, how can transcription factors such as Hox4b follow the model if they are not secreted?
   • In general, the manuscript does not refer to specific PCGs known to be critical for regeneration and patterning, such as notum and wnt1. Is this because they could not be identified in the dataset? If so, is it not possible to perform FISH and integrate these results with the Stereo-seq data?

   Additional issues to be addressed:

   • Lines 161-163: Why do the authors conclude that the increase in dorsal epidermal progenitors, neural progenitors, and pharyngeal lineages indicates an active wound response or the generation of new tissue? This statement is not clear.
   • The 4D atlas annotates 36 refined cell types, which is a noteworthy result when compared with published scRNA-seq databases. The authors should compare their results with scRNA-seq in terms of the number and identity of cell types.
   • In Figure 1F, the blastema is divided according to the pigmented/unpigmented region. However, in Figure 1G the dividing lines do not follow this curvature. Should they not follow the pigment curvature as well? Additionally, why do the blastemas in Figure 1G show such pronounced lateral deviation? Is this because the incisions were not perpendicular to the AP axis? If so, with this variability, it is difficult to understand how the datasets could be integrated. A detailed explanation is required.
   • Lines 202-203: "To test this, we analyzed the spatiotemporal patterns of SBGs by mapping representative samples from each time point onto clusters along the A/P axis using logistic regression (Fig. 2A)." What is meant by "representative samples"? The procedure should be specified more clearly.
   • Data S4 is not mentioned in the text.
   • smed03831, caveolin3, and smed01640 are the genes enriched in the ARZ domain. However, the authors perform functional analysis only with med8. Is there any specific reason for this?

   Read the original source
2. 

   GigaScience

   Jul 10, 2026

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is …

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is crucial for polarity remodeling and blastema formation. This study provides a comprehensive spatial molecular and cellular map of regenerative processes, highlighting injury-induced spatial domains and key regulatory factors in planarian regeneration. We also provide an interactive web portal, offering a valuable resource for exploring and analyzing regeneration mechanisms in a spatiotemporal context.

   This work has been peer reviewed in GigaScience (see https://doi.org/10.1093/gigascience/giag064), which carries out single-anonymized peer review. These reviews are published under a CC-BY 4.0 license and were as follows:

   Reviewer 2：

   In the manuscript '4D single-cell spatial transcriptomics reveals dynamic morphogenetic gradients and regenerative domains in planarians,' Han and colleagues generate a truly stunning spatial transcriptomics dataset of planarian regeneration from the species Schmidtea mediterranea. The authors' dataset includes whole 3D reconstructions of two regenerating planarian fragments at 8 different timepoints during regeneration, a fantastic accomplishment and resource of broad interest to the regenerative biology community. The authors analysis of the dataset includes characterization of spatially biased genes (SBGs) and exploration of an anterior regenerative zone (ARZ) and the role of the gene med8 in its' regulation. While the authors' dataset is remarkable and their analysis of spatially biased genes and med8 function is interesting, I'm not yet convinced that their conclusions are fully tested by the included experiments. In addition, I think that the authors have not included sufficient quality control metrics for their spatial dataset, which makes determining the limitations or caveats of their analysis and conclusions more difficult. However, my concerns could be addressed by additional analysis and minor experiments, or by softening the conclusions of the authors to include alternative models. I've detailed the areas of analysis/discussion that I believe require improvement below:

   Major Criticisms:

   1. Stereo-seq resolution and capture efficiency: The authors assert that their spatial approach is high enough resolution to resolve cell types and they claim to have characterized 36 cell types in their abstract. However, the 'cell type' in their dataset that they choose to focus on - Clu.31 - has gene markers expressed in three different cell types that have been shown to be distinct in the literature and prior planarian atlases. The authors should analyze gene expression signatures of other stereo-seq 'cell types' to determine if they also show mixed expression signatures. In addition, I am curious if stereo-seq is more likely to capture highly expressed genes (like those expressed in parenchymal cell types) than more lowly expressed genes (like the transcription factors expressed in stem cells). If it exists, this bias could influence annotation of cell types in highly heterogeneous regions of the worm like the parenchyma or parapharyngeal region. Finally, there is very little QC data in the supplementary materials (Size/volume of segmented cells, UMIs and features per cell, variability in features/UMIs per section, per replicate, and per cell type, etc.) I think this analysis would be highly valuable for the reader to interpret the data and the 36 identified 'cell types'.

   2. Dynamics of spatially biased genes: The authors analysis on the dynamics of spatially biased genes (SBGs) is very interesting, but the 'oscillations' the authors referred to were not clear to me in the data across all or even most of the pattern clusters in Figure 2A. In general, it seemed more like the pattern cluster was 'noisy' or more broad before stabilizing to its final location. In addition, the PCA analysis in Figure 2B seems to show that Intact and 14dpa transcriptomics is very similar, but 0h, 12h, and 36h timepoints are very distinct from 3, 5, 7, and 10 day fragments. This would suggest that early wound response gene expression is highly distinct (even opposing) the gene expression programs active during late in regeneration. More exploration of this idea, as well as clarified language on exactly what the author means by 'oscillations' and which gene groups follow this pattern would greatly improve this section and better support the author's conclusions.

   3. The Cellular/Functional identity of Clu.31: The authors state throughout the manuscript that Clu.31 (the ARZ) is an injury-induced anterior state enriched for SBGs and regulating polarity establishment. However, it is also possible that this spatial state represents the anterior peripheral nervous system (numerous sensory neurons and surface epithelial cells that help sense mechanical and chemical cues). SBGs could be enriched because this combination of cell types is only present in the anterior of the animal. Indeed, the authors show that the ARZ is localized to the anterior in intact animals in the absence of an injury (Figure 3) and enriched genes (S4Aii) strongly indicate that Clu.31 contains gabrg+ mechanosensory neurons. If Clu.31 is regenerating nervous system, this would also explain its ventral bias and expression of tgs-1 and other nb2 genes, since nb2 neoblasts have been suggested to be both an amputation responsive neoblast subset (Zeng et al. Cell) and a neural progenitor state (Raz et al. Cell Stem Cell). Clarifying how the composition of the tri-lineage region changes during regeneration may help distinguish if Clu.31 is truly an injury induced region vs. the regenerating sensory nervous system. For example, it is known that agat-1+ cells transcriptionally responsive and enriched at the wound site a 2-4 days post amputation, but less so at later timepoints (Benham-Pyle et al Nature Cell Biology, Kent et al. Developmental Biology). This shift in composition should be observable in Clu.31 since it contains agat+ epidermal cells. Such a shift in composition or the identification of a regeneration-specific marker expressed in Clu.31 would add support to the author's conclusions. Regardless of the outcome of these experiments/analyses, the discussion and interpretation of the data could be modified to address the hypothesis that Clu.31 represents the cellular neighborhood created when the peripheral nervous system intercalates with the anterior DV boundary epithelium and body wall muscle, which needs to be regenerated in amputated worms. As is, the comparison to the apical epithelial cap considered in the discussion (Line 438) may be pre-mature.

   4. Med8 function: Med8 produces a clear phenotype in the authors' experiments, and their data indicates that it is required for ARZ formation. However, I am not sure that the authors data supports the claim that Med8 is directly regulating blastema and PCG expression, as opposed to regeneration of the nervous system (which is highly interconnected with formation of the anterior pole and the size of the anterior blastema) and stem cell function more broadly. The fact that Med8 RNAi also leads to head degeneration in intact worms (Figure S6F) strongly suggests a more fundamental defect in neural differentiation or stem cell function. The strongest evidence presented by the authors supporting a broader function in polarity establishment is the disruption of posterior Wnt expression, (Figure 5F and G), but these in situs are single representative images with no quantitation and could also be explained by a stem cell defect. Additional data could be provided (e.g. visualization of wound-induced gene expression, quantitation of anterior or posterior stem cell numbers and proliferation rates at 2dpa) to support regulation of PCGs or blastema formation. The authors could also leverage their single cell sequencing to determine if Med8 RNAi impacts neural progenitor abundance more than other progenitor cell types. Together, these experiments would determine if Med8 is important for amputation-induced blastema formation and polarity re-establishment vs. stem cell function and neural differentiation more broadly.

   Minor Criticism/Feedback:

   1. In Figure 1I, the authors show DEGs enriched in each cluster/region. In the blastema regions, I was surprised by the number of DEGs for each time point. It appears that there are ~10K upregulated and 10K downregulated DEGs by the later time points, which suggests that 2/3 of the transcriptome is differentially expressed… The authors should clarify in the text or methods what cutoff they used for the DEGs and how significant the DEGs are in this figure.
   2. For readability, I really think that all figures should be on a white background.
   3. How do gene expression profiles from the stereo-seq compare to bulk rnaseq at similar timepoints?
   4. It is very interesting that there are some cell types that appear to contract and then expand during regeneration (Cluster 0, 23) or that aggregate/become more targeted during regeneration (pharynx pouch, cluster 29). Molecular differences between early and late cells within these cell types would be particularly interesting for understanding different phases of regeneration, but this may be beyond the scope of the current study.
   5. The authors frequently reference Han et al. submitted, but this manuscript would need to be pre-printed or published in order for this work to reference it.
   6. The Y axis of Figure 2E should be labeled

   Read the original source
3. 

   GigaScience

   Jul 10, 2026

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is …

   AbstractRegeneration relies on precise spatiotemporal gene expression and cellular responses to establish tissue identity and body patterning. Using high-resolution Stereo-seq (715 nm) on 353 sections from 16 whole animals at 8 regeneration timepoints, we constructed a 4D spatiotemporal transcriptomic map of planarian regeneration. Our analysis captured 36 refined cell types from 3,508,004 segmented cells, enabling genome-wide transcriptional imputation of gene expression dynamics across body axes at cellular, tissue, and organismal scales. We identified dynamic positional gradients and distinct spatially distributed cell types during regeneration, including an injury-induced Anterior Regenerative Zone (ARZ). The ARZ exhibited enriched positional signals in epidermal, muscle, and neural cells and was regulated by Mediator 8, which is crucial for polarity remodeling and blastema formation. This study provides a comprehensive spatial molecular and cellular map of regenerative processes, highlighting injury-induced spatial domains and key regulatory factors in planarian regeneration. We also provide an interactive web portal, offering a valuable resource for exploring and analyzing regeneration mechanisms in a spatiotemporal context.

   This work has been peer reviewed in GigaScience (see https://doi.org/10.1093/gigascience/giag064), which carries out single-anonymized peer review. These reviews are published under a CC-BY 4.0 license and were as follows:

   Reviewer 1：

   The authors employs high-resolution Stereo-seq technology combined with multi-timepoint spatial transcriptomic data to construct a 4D spatiotemporal transcriptomic map of planarian regeneration. This work significantly advances the understanding of spatial gene expression dynamics during planarian regeneration, overcoming the limitations of traditional two-dimensional and planar spatial transcriptomics. Furthermore, the authors identify a novel injury-indced Anterior Regenerative Zone (ARZ) and, through functional validation of Mediator 8 (Med8), deepen insights into the mechanisms underlying polarity remodeling in planarians. The study also provides an interactive online database, enriching spatial molecular and cellular data resources for the regenerative biology. The work is notably innovative, and the authors present convincing evidence supporting their conclusions. The manuscript overall is written well and data is presented clearly. The discussion and conclusions has done well to highlight the potential problems in this study. I have a few points that should be addressed before publishingI have a few points that should be addressed before publishing.

   1.In lines 233-236, it is reported that the positional control gene (PCG) like ndk restores its spatial expression pattern as early as 12 hpa, whereas its expression level only significantly increases at 36 hpa. Given this pronounced temporal discordance between early recovery of spatial patterning and the later peak in mRNA levels, the authors should analyze and discuss possible molecular mechanisms that could account for this discrepancy, and consider the biological implications of this phenomenon for understanding how spatial information and gene-expression regulation are coordinated during regeneration.

   2.In lines 356-360, Med8 knockdown markedly reduces the ARZ cell lineages and the expression of anterior-posterior polarity markers (e.g., sfrp-1, wnt1, wnt11-1), producing a clear effect on regeneration polarity formation.However no gross disruption of the whole-body AP axis was observed. Please further analyze and discuss the possible regulatory scope and mechanisms of Med8. Specifically, do other redundant pathways or compensatory mechanisms exist in planarians that maintain global positional information despite loss of Med8? What is the hierarchical and cell-type specificity of Med8's role in polarity regulation? Transfer Authorization Response

   Read the original source
4. 

   Version published to 10.64898/2026.02.18.706529 on bioRxiv

   Feb 19, 2026
5. Version published to 10.1093/gigascience/giag064

   Jan 1, 2026