Distinct mechanisms underlie oral vs aboral regeneration in the cnidarian Hydractinia echinata

1) The phenotypes shown by the authors are quite convincing but the results shown in Figure 2 should be supported by quantitative data and statistical analysis regarding the different morphological and cellular phenotypes and kinetics of regeneration are missing. Similarly results shown in Figure 3 could be more precise, as deduced from the combination of EDU pulse chase labeling to PH3 detection that could provide a temporal analysis of the contribution of cycling cells in blastema formation. In fact the pattern of cycling cells on detached but intact polyps appears variable, in some cases forming a clear and unique band relatively low along the body column with sharp boundaries (as in Figure 3A), while in other cases, it appears rather spread all along the body column of the polyps as shown in Figure 3–figure supplement 1 (panels B and C). As this band pattern is quite important in the blastema formation mechanism proposed by the authors, this variability should be quantified and discussed by the authors. Chiefly, we would like to see a more rigorous quantification of the cell cycle parameters measured during regeneration.

Regeneration experiments as described in Figure 2 have been carried out in our lab hundreds of times over the past years. These experiments have established that over 90% of healthy Hydractinia polyps regenerate their heads following decapitation within two-three days, but regeneration times in unhealthy or malnourished animals may last substantially longer. By contrast, very young polyps (i.e. shortly post metamorphosis) may regenerate an amputated head even faster, perhaps because at this stage they contain many more i-cells than fully grown ones. We have added some clarification statements to the text to address this point (in the first Results section).

As suggested, we performed anti-pH3 analysis of EdU pulse chased cells during head regeneration. Our results, presented in a new supplemental figure to Figure 5 (Figure 5–figure supplement 1), show that the blastema includes double positive cells, i.e. cells that were in S-phase in the band before decapitation, and in mitosis in the head blastema 24 and 48 hours later. These cells had migrated from the i-cell band to the blastema within this time frame.

The distribution of cycling cells in Hydractinia polyps largely depends on developmental stage and state of health. In young, post metamorphosed polyps one can find cycling cells throughout the body column and tentacles although the majority is still found in an aboral band. This is also true for newly growing, clonal polyps in a sexually mature colony. However, fully-grown polyps generally display a clear proliferation ‘band’ at the lower polyp body column, co-localizing with cells expressing stem cells associated genes. The width of the ‘band’ is variable, but cycling cells are only rarely observed in the head area or tentacles. As requested, we have now quantified it in an additional experiment and found that over 90% of all randomly selected polyps had displayed this pattern of proliferating cells. This is now also mentioned in the text.

2) It seems important to the arguments made by the authors to be able to monitor the interstitial to epithelial transition in the blastema. We would like to suggest that the authors monitor the appearance of epithelial markers in cells that are still GFP positive at the protein level but no longer at the transcript level. That way the authors could show that i-cells from the blastema differentiate into epithelial cells when the Piwi promoter becomes inactive. To get this result, the GFP protein should persist at least one or two days after transcripts disappear, that way the Piwi:GFP(+) i-cells would lose their interstitial GFP signature at the time they gain the epithelial one.

Evidence for interstitial to epithelial transition in Hydractinia has been provided in at least two publications in the past (Müller et al. 2004, Dev Biol 275, 2015-224; Künzel et al. 2010, Dev Biol 348, 120-129). Therefore, demonstrating that Piwi1+ i-cells can give rise to epithelial cells would not constitute a major scientific novelty in our opinion. In the regenerative context, however, we agree that we cannot exclude that epithelial cell proliferation and/or transdifferentiation had also contributed to the new head. Unfortunately, in the absence of a verified, exclusive marker for Hydractinia epithelial cells we are, at present, unable to perform the suggested experiments. Furthermore, we feel that the question of possible contribution of differentiated epithelial cells to head regeneration is not central to the point we are making in our paper. We have therefore toned-down our statement on the role of epithelial cells in head regeneration in the manuscript and hope to address this question in a future study.

3) Some specificity controls for the RNAi experiments need to be presented. Since RNAi of the stem cell factors all have the same phenotype, it is suggested to knockdown a key cell cycle regulator and show that this affects EdU. It is also suggested to knockdown a gene that is not expressed in the i-cells and show that it has a different phenotype.

To strengthen our claim of RNAi specificity we used two new approaches in addition to the non-coding dsRNA (control RNAi) that has already been presented as control at first submission.

First, we performed qPCR on Piwi1 RNAi animals to confirm a specific reduction of Piwi1 transcript level (normalized to 18S). Second, as suggested, we knocked down two additional genes we expected to be important for cell cycle progression: histone H2A and histone H4. Indeed, treatment with H2A or H4 dsRNA almost completely abolished EdU incorporation comparing to non-coding dsRNA we used as control. These new results are now presented in a new supplemental figure to Figure 4 (Figure 4–figure supplement 1).

Knocking down of any confirmed non-i-cell gene during regeneration might very well result in compromised regeneration. In particular, Ncol1 (an exclusive nematoblast marker) knockdown did affect regeneration as we have shown and this would probably also be true for neural genes because Hydra head regeneration requires de novo neurogenesis (Miljkovic-Licina et al. 2007, Development 134, 1191-1201). Hence, it would be difficult to a priori select a gene that does not have a role in head regeneration since this complex process probably involves many genes.

We would also like to add that in the past we have used RNAi to knockdown various Hydractinia genes in various contexts and monitored distinct phenotypes for each knocked down gene (e.g. Duffy et al. 2010, Development 137, 3057-3066; Duffy et al. 2011, Dev Biol 362, 271-281; Millane et al. 2011, Development 138, 2419-2439).

We therefore suggest that the new experiments described above, together with our published work, provide strong evidence for the specificity of RNAi in Hydractinia.