Planarian stem cells sense the identity of missing tissues to launch targeted regeneration

  1. Tisha E. Bohr
  2. Divya A. Shiroor
  3. Carolyn E. Adler

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

In order to regenerate tissues successfully, stem cells must first detect injuries and then produce missing cell types through largely unknown mechanisms. Planarian flatworms have an extensive stem cell population responsible for regenerating any organ after amputation. Here, we compare stem cell responses to different injuries by amputation of a single organ, the pharynx, or removal of tissues from other organs by decapitation. We find that planarian stem cells adopt distinct behaviors depending on what tissue is missing: loss of non-pharyngeal tissues increases numbers of non-pharyngeal progenitors, while removal of the pharynx specifically triggers proliferation and expansion of pharynx progenitors. By pharmacologically inhibiting either proliferation or activation of the MAP kinase ERK, we identify a narrow window of time during which proliferation, followed by ERK signaling, produces pharynx progenitors necessary for regeneration. Further, unlike pharynx regeneration, eye regeneration does not depend on proliferation or ERK activation. These results indicate that stem cells tailor their proliferation and expansion to match the regenerative needs of the animal.

Article activity feed

  1. eLife

    ###Reviewer #3:

    In this manuscript, Bohr et al examine how the pluripotent stem cell system of planarians responds to organ-specific damage. If and how the differentiation of specific cell types is dynamically regulated is a conceptually fascinating problem in planarians and in general stem cell research. The authors address this problem by comparing the stem cell response between a single-organ amputation (the pharynx) versus broad tissue loss (decapitation). Their findings indicate that only the removal of the pharynx triggers the increased differentiation of pharyngeal cell types, while the loss of non-pharyngeal tissues upregulates the differentiation of progenitors of multiple organs, but not the pharynx. Further, the authors implicate temporally restricted ERK signaling as a regulatory component in the differentiation of pharyngeal cell types. These observations are also important because they contrast with the previously proposed "target blind" model (LoCascio et al., 2017) that posits the differentiation of different cell types at constant relative proportions, with the rate of stem cell divisions as global production rate regulator. In contrast, the observations by Bohr et al provide further evidence for more flexibility and specificity within the planarian stem cell system ("target consciousness") in the sense of lineage-specific adjustments in the production rates of specific cell types.

    That said, the manuscript generally suffers from an overly narrow focus. Important questions remain regarding the specificity of the stem cell response to pharynx amputation and multiple experiments lack important controls (see below). Moreover, the authors have overlooked that a "target conscious" progenitor response has already been demonstrated by the selective proliferation of protonephridial markers expressing neoblasts in response to protonephridial damage by RNAi (Vu et al., 2015).

    Major points:

    1. Specificity of the stem cell response:

    The central premise of the paper is the selective amplification of pharyngeal progenitors in response to pharynx amputation. The authors conclude this based on i) an increase in the absolute number of foxA+/piwi-1+ cells in a specific area, while ii) de-capitation has no effect on the absolute number of foxA+/piwi-1+ cells in the same area. This approach is an insufficient demonstration of specificity, as the known phenomenon of wound-induced stem cell activation might also change the absolute number of specific neoblast subclasses and might do so in an injury-dependent manner.

    To account for this important caveat, the authors need to i) quantify RELATIVE proportions of foxA+/piwi-1+ cells out of total piwi cells (or of total H3P+ cells) and ii) they need to include other organ progenitors in the initial analysis. The latter is also critical because the pharynx is a complex organ comprising descendants of multiple lineages (e.g., muscle, neurons, epidermis) and it is not clear whether the foxA+/piwi-1+ cells indeed serve as a singular origin of all constituting lineages (as assumed by the authors), or if they only provide a subset of pharyngeal cells with a rate-limiting role in pharynx assembly (e.g., pharyngeal muscle). In the face of such uncertainty, iii) the quantification of new cell incorporation into the pharynx versus other tissues via BrdU labeling would be necessary to address this caveat and to provide a global perspective on the specificity of the response independent of incompletely characterized marker genes.

    In addition, the following experimental design problems need to be addressed or better documented, including:

    • The authors provide insufficient methodological detail on progenitor quantifications, even though the entire manuscript rests on this assay. What are their criteria for scoring a piwi-1+ cell as double-positive for the often weak and noisy lineage labels? If done "by eye", was double-blind scoring used? Were all cells in a given Z-stack counted or only specific planes? If the latter, by which criteria were image planes selected for quantification? How were the specific specimens out of an experimental cohort selected for imaging/quantification? Though not necessarily a unique shortcoming of this particular study, these points simply need to be adequately addressed in order to rigorously support quantitative differences between experimental conditions (e.g., specificity).

    • The authors appear not to distinguish at all between technical replicates (e.g., multiple specimens within an experimental cohort) versus biological replicates (independent experimental cohorts). This is significant, because i), the use of the standard error of the mean (SEM) that the authors employ throughout is not really an appropriate measure for a single biological replicate with 3 animals - the standard deviation (SD) would seem a more appropriate measure in this context (SD). ii), the number of worms quantified for each experiment is generally low (n=3 animals in Fig. 1, 3, 5, 6; n=5 animals in the rest of the figures) given the observed variability in the data (e.g., ~25-30 foxA+/piwi-1+ cells 3d after pharynx amputation in Fig. 1E versus 50 cells in figure 1H). Similarly, for kinetic experiments as in Fig. 1H or 2C, it is simply crucial to ensure that the error bars include the variation in response dynamics between multiple replicates due to the drift in the baseline fraction of H3P+ cells or varying staining efficiencies (e.g., different batches of animals on different days), rather than the technical variation in a single experimental cohort only. Please address these concerns by adding more specimens and a thorough description of the experimental design.

    1. Timeline of pharynx regeneration: The pharynx regeneration timeline and associated events that the authors present are insufficiently supported by experimental data. The conclusion in line 215 "that proliferation in a critical window of 1 to 2 days after pharynx amputation produces a population of progenitors that are likely essential for pharynx regeneration" rests i) on the diagnosis of a "proliferative peak of FoxA+ stem cells that occurs after pharynx amputation (Figure 2C)" (line 202). However, rather than a "proliferative peak", Fig. 2C shows a broad "proliferative plateau" of FoxA+ stem cells between 6h and 3 days after amputation. Similarly, the foxA+/H3P+ quantification after pharynx amputation in Fig. 4G also displays a lack of a peak of foxA+/H3P+ from 1d to 2d after amputation. ii), the associated nocodazole experiments suffer from the fact that the authors did not quantify the impact of the drug on the abundance of foxA+/piwi-1+ cells during treatment intervals from 0-1d and 2-3d after pharynx amputation. Therefore, the authors cannot rule out that nocodazole treatment might have similar effects on the abundance of foxA+/piwi-1+ cells throughout the 1-3 d post-amputation time interval, with the more severe organ-level phenotype of the 1-2d treatment window being caused by some other effect of the drug (e.g., on the differentiation of another rate-limiting cell type for pharynx regeneration or, conceivably, inhibition of priming neuronal activity ). Similar concerns apply regarding the statement in line 242, "... a window 1-2 days after amputation in which activation of ERK signaling is important for pharynx regeneration." Here, i) the quantification of the end-stage phenotype of drug treatment during the 1-2d time interval (regain of feeding ability) is missing. ii) Similarly, the examination of the consequences of PD treatment on foxA+ expression in piwi-1+ cells in panel 4D-H employs drug soaking for 3 days, yet the corresponding end-stage phenotype of 3-day drug treatment is not shown. iii) the implications of ERK in pharynx regeneration are tentative. Even though the PD compound is initially correctly introduced as "MEK inhibitor", the authors subsequently switch to the factually wrong "ERK inhibitor" designation (e.g., line 358). Further, additional experimental evidence for the assumed Erk inhibition as the cause of the observed phenotypes would be desirable to rigorously support the conclusion.

    These caveats need to be addressed if a cell biological timeline is to remain part of this manuscript.

    1. Integration with the existing literature:

    The authors need to better integrate their findings with the literature. First, they need to cite the findings of Vu et al, which explicitly demonstrated a specific increase in the fractional abundance of piwi-1+/protonephridial marker+ cells in response to RNAi-mediated damage to protonephridia (Vu et al., 2015). As such, this study already demonstrates the main point of Bohr et al., namely that the planarian stem cell system is capable of "target conscious" progenitor provision. At the very least, the authors should credit these results as additional evidence for their model. A further finding that they should discuss is the demonstration by LoCascio et al (LoCascio et al., 2017) that flank region cut-outs cause a significant increase in pharynx cell incorporation over baseline, despite the absence of injury to the pharynx. How do the authors reconcile the discrepancy between these data and their own? In general, the discussion would benefit greatly from a more explicit comparison between the "target blind" model versus their data, as well as a broader perspective on the regulation of stem cell homeostasis.

  2. eLife

    ###Reviewer #2:

    Bohr, Shiroor, and Adler investigate how stem cells respond to the loss of specific tissues in planarians. The planarian stem cell population (neoblasts) are distributed throughout the planarian body and include pluripotent stem cells and a wide range of lineage-committed progenitor cells. How this heterogenous pool of cells behave post-injury or amputation is incompletely understood. The discovery of markers to label stem cell progeny have opened the door to investigate how stem cells respond to tissue loss. However, the anatomy of planarians makes it difficult to surgically remove or damage specific organs. The PI of this study developed an assay to remove the pharynx by "chemical amputation" to study the mechanisms underlying regeneration of this organ without drastically perturbing or injuring other tissues. Using this approach, this paper investigates how a well-defined population of FoxA+ progenitors respond to pharynx removal at early time-points during the regeneration. Their data suggest that stem cells are able to detect loss of the pharynx and respond by generating significantly more cells fated to become pharynx, whereas amputation of non-pharyngeal tissues does not have an obvious effect on pharynx progenitor specification dynamics. In addition, using pharmacological treatments, the authors show that cell proliferation and ERK signaling are required for the expansion of pharynx progenitors and cell differentiation. In contrast, other cell types in the planarian eye do not appear to require proliferation or ERK signaling, suggesting that stem cell responses "target blind" as suggested in a previous study, but are rather tuned to specific missing tissues.

    This work has the potential to make a significant contribution to the field by advancing our understanding of how the planarian heterogenous stem cell population responds to the loss of a specific organ. However, the report is preliminary as presented. It appears that the authors performed many experiments a single time. In addition, description of the methods is insufficient. The authors need to chiefly demonstrate the reproducibility of the data and robustness of the observations.

    Concerns:

    1. The authors need to replicate experiments to increase the sample size for most experiments.

    2. Details for imaging and quantification should be explicitly stated in the methods, and the reported cell count numbers should be normalized as appropriate for each set of experiments.

    3. Although the authors mention "lineage-tracing" experiments (see minor comments), they do not perform DNA analog pulse-chase experiments to analyze a temporal progression and spatial localization of stem cells to FoxA+ progenitors after pharynx removal. The authors rely on PH3-staining in conjunction with FoxA, and supplementary experiments using the pluripotent stem cell marker tgs-1 (which was only examined at 1 dpa). Could the authors clarify what they think FoxA+ stem cells represent? Are these self-renewing pluripotent stem cells or lineage-committed progenitors? Can the authors get some insight by scanning their images of PH3+ cells expressing FoxA visibly undergoing metaphase? Are daughter cells uniformly FoxA+ as reflected in the model? At least in one of the cells shown in the nocodazole treated controls it appears that both daughter cells express FoxA (Figure 3). I suggest showing some higher magnification images to support the interpretations/conclusions. Others have posited (e.g., Rink, Chapter 2 of Planarian Regeneration: Methods and Protocols), that not every dividing cell may be a long-term self-renewing stem cell and whether transient amplifying cells exist or contribute to regeneration in planarians is unknown. Adler and Sánchez Alvarado (2015) discuss the role of transient states and how the transcriptional profiles change in response to regeneration. It wasn't clear to me how the authors think about these cells based on the limited number of experiments and analysis, and there are a few places where the terminology is inconsistent, especially in reference to proliferating ovo+ progenitors (P. 14). The authors need to be clear, and it might be helpful to illustrate their model in one of the early figures or to include it in the final model, which omits tgs-1 due to the limited number of experiments performed with this marker gene (Figure 7).

    4. The pharynx is complex and there is no data to assess what the contribution of other progenitor populations might be. I don't think or think it is unlikely that FoxA+ progenitors are solely responsible for reconstructing the pharynx. The authors should examine how other progenitor populations behave during the process of pharynx regeneration by extending the timeline of progenitor cell analysis. This would reveal if there is fluctuation in progenitor dynamics as animals regenerate the pharynx or re-scale proportions after pharynx regeneration. For example, can the authors test if they are able to detect a contribution of neural progenitors to regeneration of the pharyngeal nervous system? And if so, when during the regeneration process does it take place in the context of their study?

  3. eLife

    ###Reviewer #1:

    This manuscript explores how planarian stem cells respond to the loss of a specific organ: the pharynx. The previously proposed "target-blind" model of planarian regeneration (LoCascio et al. 2017) posited that stem cells do not respond directly to missing tissues, but rather replace missing cell types based on their normal rates of homeostatic turnover. In contrast, the authors of this manuscript suggest that planarian stem cells can sense and respond to the loss of specific missing tissues, using the pharynx as a case study. The authors conclude that planarians may use more than one mode of regeneration, depending upon the target being regenerated (eye vs. pharynx).

    The question explored in this paper is of fundamental importance, and providing an alternative model by which planarian stem cells regenerate missing tissues should be of interest to a broad readership. Unfortunately, in its current form, the manuscript presents enticing preliminary findings, rather than robust experimental observations. Because the authors are attempting to refute a previously published model, it is critical that the data are clear and convincing. If not, these findings could be summarily dismissed without appropriate debate. If the authors can demonstrate that their results are robust across larger samples sizes and experimental replication, and address the major issues listed below, this manuscript would represent a significant contribution to our understanding of planarian regeneration.

    Major Issues:

    1. Throughout the manuscript, experiments were either not repeated, or the number of biological replicates was not reported. In most cases, it appears that experiments were done only once (with the exception of the drug treatments). Numbers of biological replicates and sample sizes should be explicitly stated and the data from different replicates reported for Figs. 1D-G, 2B-D, 3C, 3E-F, 4B, 4D-H, 5C-D, and 6A-E.

    2. The authors do not sufficiently describe their methods for imaging and quantifying cells (Figs. 1E, 1G, 2C-D, 3F, 4E-H, 5D, 6B). The size of the area covered to collect these data is unclear. High-magnification images are shown: are these the areas that were imaged? If so, their results could be biased by choosing small regions of interest. Ideally, the authors should quantify more than one region per animal. Also, they do not describe the depth of the z-stacks collected or how these stacks were normalized/standardized across conditions. All their conclusions hinge on the quantification of progenitor populations in response to different amputation paradigms or chemical treatments, so the standards for imaging and quantification must be clearly reported.

    3. Inappropriate statistical tests were used throughout. The use of multiple t-tests amplifies the chance of a Type I error and is especially problematic when up to 7 comparisons were made! The authors should use one-way ANOVA with multiple comparison corrections for all experiments with more than two groups.

    4. Figs. 1D-E show that upon pharynx amputation but not head amputation, FoxA+ piwi+ pharynx progenitors increase. These data suffer from the quantification issues highlighted above: how the data were quantified is not sufficiently described, only 3 data points were taken (one per animal), the experiment appears to have been performed only one time, and the wrong statistical test was used. Rather than reporting the number of FoxA+ piwi-1+ cells counted, the authors should quantify the total number of doubly positive cells as a percentage of piwi-1+ cells, as was previously published (Adler et al. 2014). The authors also fail to specify whether the change observed between "3 dpa phx" and "3 dpa head" is significant, which is a material point.

    5. Fig. 2D also suffers from the inadequate quantification practices described above. Ideally, FoxA+ cells should be quantified as a percentage of the H3P+ cells observed.

    6. The authors use "stem cell", "progenitor", "stem cell progenitor", and "progenitor stem cells" in a mixed and confusing way throughout the paper. For example, in lines 174-175 the authors state that "proliferation of FoxA+ stem cells precedes the increase in pharynx progenitors." This refers to FoxA+ H3P+ cells vs. FoxA+ piwi-1+ cells, but the only difference is that the former are stem cells in the act of mitosis. Is a distinction being made? Elsewhere in the paper, FoxA+ piwi+ cells are referred to as stem cells. The terminology used needs more clarity and consistency.

    Along these lines, what is a FoxA+ PSC (Line 168)? Are the authors suggesting that FoxA is a pluripotency marker? Or are the authors saying that tgs-1 has a more expansive expression pattern and is co-expressed with progenitor markers? If double FISH was performed with the progenitor markers reported (ovo, myoD, gata-4/5/6, six-1/2, and pax6), would they all overlap with tgs-1? These experiments need to be performed to make any claims about FoxA expression in the context of pluripotency.

    1. In Fig. 3C, how does pharynx regeneration occur over such a long period of time after nocodazole treatment in the 1-2 day window? Does target-blind regeneration occur once this window is missed? The authors should repeat analysis of FoxA+ progenitors at later time points in this condition, and/or show rates of BrdU incorporation into the pharynx with and without nocodazole treatment in this window.

    2. The use of the inhibitor PD in Fig. 4 is problematic. No data are shown to verify that phosphorylation of ERK was inhibited in these experiments. A citation of previous use is not sufficient. The effect of PD on WT uncut animals with regards to FoxA+ cells is not shown and is a necessary control. To address questions of drug specificity, the authors should corroborate their findings with a second inhibitor of ERK signaling like U0126, which has already been shown to work in planarians (Owlarn et al. 2017).

    3. The conclusion that ERK signaling functions in regulating differentiation but not proliferation is premature (line 249-264). Figs. 4E-H should be quantified as percentages of piwi-1+ and H3P+ cells, especially since there is decreased proliferation overall with PD treatment.

    4. The use of head fragments to compare eye regeneration vs. pharynx regeneration is inappropriate. Previous studies have already shown that the absence of eyes is not required to induce ovo+ progenitor amplification (LoCascio et al. 2017). Thus, this is not a surprising result (line 343-345) and the authors are mis-citing previous observations (in the earlier work, head fragments were never described). The region where ovo+ cells was quantified in Fig. 6A is not justified or explained. The yellow box is placed in a medial region where ovo+ cells do not normally reside. The authors should image within the laterally positioned ovo+ streams that have been previously described (Lapan & Reddien 2012; LoCascio et al. 2017).

    5. Rather than using head fragments, the authors should repeat the flank resection experiments shown previously (LoCascio et al. 2017). This previous study showed that increased BrdU incorporation into the pharynx occurred following flank resection even though the pharynx was present. That result may have been 1) an artifact of increased BrdU staining due to stimulation of proliferation upon injury, 2) caused by unintentional damage to cells associated with the pharynx, or 3) a response to the loss of FoxA+ progenitor populations that surround the pharynx rather than the loss of the differentiated organ. The authors have the opportunity to revisit this published observation by quantifying the FoxA+ progenitor response during flank resection +/- pharynx. Without these data, this story is incomplete and therefore the conclusion of a targeted regeneration response is not yet convincing.

    6. The negative results that proliferation and ERK signaling are dispensable for eye regeneration in Fig. 6 are weak and unconvincing. The regenerated eyes appear smaller; this should be quantified (number of PRNs per eye). If the small pharynges that form in Figs. 3D and 4C are considered a deleterious phenotype, why is the same standard not applied to the eye? Also, existing eye progenitors could have been sufficient for eye regeneration in these drug treatments. Furthermore, eyes did not regenerate after nocodazole treatment 50% of the time. Is it not more likely that the observations reported are dosage and timing artifacts? How has proliferation been affected? These observations do not live up to the claims made.

    7. The authors claim that ovo+ cells are not proliferative (H3P+) even in cases where there is eye progenitor amplification (head amputation), but the data are not shown (line 321). They should be. Indeed, previous publications have never shown that ovo+ cells proliferate. This might mean that there are proliferating eye progenitors that precede expression of ovo. The authors should discuss this alternative.

    8. The authors claim that eye regeneration does not require proliferation or ERK signaling but pharynx regeneration does. This conclusion hinges on the gross observation that eyes can regenerate in the presence of nocodazole and PD (see point 12 above). These data are coarse and the interpretations are unconvincing. Instead, their model can be directly tested using BrdU pulse-chase experiments. According to the authors' model, one would predict that following pharynx amputation, the rate of incorporation of BrdU+ cells into the regenerating pharynx should be higher than in uninjured controls. Conversely, the rate of BrdU incorporation into the regenerating eye should remain unchanged between injured (i.e. eye-resected) and control animals (LoCascio et al. 2017). Once the authors have established the prediction above, they have the opportunity to show the effects of nocodazole, PD, and U0126 on BrdU incorporation in the regenerating eye vs pharynx following eye resection and pharynx amputation in the same animal. This way, the authors can directly test the requirement for proliferation and/or ERK signaling in both tissues.

  4. eLife

    ##Preprint Review

    This preprint was reviewed using eLife’s Preprint Review service, which provides public peer reviews of manuscripts posted on bioRxiv for the benefit of the authors, readers, potential readers, and others interested in our assessment of the work. This review applies only to version 1 of the manuscript.

    ###Summary

    All three reviewers recognized the potential significance of this work, but shared concerns about sample sizes, lack of biological replicates, and insufficient details about cell quantification.

  5. Version published to 10.1101/2020.05.05.077875 on bioRxiv