Growth/differentiation factor 15 controls number of ependymal and neural stem cells in the ventricular-subventricular zone

  1. Katja Baur
  2. Carmen Carrillo-García
  3. Şeydanur Şan
  4. Manja von Hahn
  5. Jens Strelau
  6. Gabriele Hölzl-Wenig
  7. Claudia Mandl
  8. Francesca Ciccolini

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Late in neural development, the expression of growth/differentiation factor (GDF) 15 increases in the germinal epithelium of the murine ganglionic eminence (GE), especially in progenitors with characteristics of neural stem cells (NSCs). However, the function of GDF15 in this region is still unknown. We here show that apical progenitors in the E18 GE also express the GDF15 receptor and that ablation of GDF15 promotes proliferation and cell cycle progression of apically and subapically dividing progenitors. A similar phenotype was also observed in the adult ventricular subventricular zone (V-SVZ). At both ages, increased proliferation leads to the transient generation of more neuronal progenitors, which is compensated by cell death, and to a permanent increase in the number of ependymal cells and apical NSCs. We also found that GDF15 receptor-expressing cells display immunoreactivity for the epidermal growth factor receptor (EGFR), which is also involved in progenitor proliferation, and that manipulation of GDF15 affects the expression of EGFR in mutant progenitors. Moreover, our data indicate that EGFR signalling in WT and mutant progenitors relies on distinct transduction modes. However, only exposure to exogenous GDF15, but not to EGF, normalized proliferation and the number of apical progenitors, indicating that alteration in EGFR signalling is not the main mechanism by which GDF15 affects proliferation in the embryonic GE.

Taken together, GDF15 directly regulates proliferation of apical progenitors in the developing GE, thereby affecting the number of total ependymal cells and NSCs in this region.

Article activity feed

  1. Version published to 10.1101/2022.12.02.518869v3 on bioRxiv
  5. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    *Summary: This paper illustrates the role of GDF15 in ganglionic eminence featuring the influence on neurogenesis and progenitor proliferation. Using the conventional knockout GDF-15 mouse lines, the authors provide a series of counting data indicating that GDF15 controls neurogenesis in the late embryonic or adult neural stem cells in the ventral forebrain. *

    We would like to thank the reviewer to reading our manuscript and providing comments. As the reviewer highlights, we have used a conventional GDF15 knock-in knock-out model since the growth factor is expressed not only at tissue levels but also at a systemic level. Therefore, targeted ablation of GDF15 would be complicated and the results difficult to interpret. Moreover, GDF15 and its receptor GFRAL in normal conditions are expressed at very low levels in adult mice and mutant mice do not display any obvious developmental phenotype. However, GDF15 is characteristically expressed during development and our previous data have shown that its expression is particularly increased at later developmental ages and particularly in neural precursors, which are both the focus of our analysis. Although these previous analyses have long highlighted that GDF15 is particularly expressed in the V-SVZ and in the choroid plexus, its physiological role in this system remains to the best of our knowledge unknown. It is important to investigate this issue because, as we mention below, GDF15 expression is increased, including in NSCs, upon brain injury and aging. As the reviewer rightly mentions, using this conventional approach and straightforward quantitative analyses, instruments available and normally used for the investigation of biological phenomena, we have discovered that GDF15 directly affects the number of ependymal cells and neural stem cells thereby providing a first function for the expression of the growth factor in this region.

    Major comments:

    The entire story in this manuscript seems similar to the previous findings where the authors demonstrated the role of GDF15 in the hippocampal neural stem cells in relation to EGF and CXCR4.

    As the reviewer rightly mentions we have in a previous paper investigated the effect of GDF15 on the stem cells of the hippocampus. However, we would like respectfully to disagree with the reviewer that our current manuscript describes similar findings. Whereas we have previously shown that the effect of GDF15 in hippocampal stem cells and neurogenesis is a transient inhibition of proliferation due to reduced EGFR expression, we here found that absence of GDF15 leads instead to increased proliferation and a permanent increase of stem cell number, besides of ependymal cells. As the reviewer rightly mentions, on the basis of our previous observations, also in this study we have analyzed EGFR expression and signalling. However, although our data clearly show that EGFR expression and more importantly signalling are altered also in this niche, we could show that the effect of GDF15 is more complex than altering EGFR signalling. Since we show for the first time in this study, that besides GDF15, neural progenitors also express GFRAL, our data point at a selective effect of GDF15 in the development of neural stem cell in the GE and in the deriving adult niche, which include change in EGFR signalling.

    *The data and the conclusion presented here sound reasonable to me. The current manuscript, however, gave me the impression that the story is less impactful and rather descriptive. To improve the quality of the current draft, the authors may wish to clearly highlight the novelty of the findings, not simply apply the previous strategy to the other anatomical brain regions. Alternatively, the draft could emphasize the similarity of utilising the same biological strategies to control the number of adult stem cells in the distinct stem cell niche. *

    We would like to thank again the reviewer for the positive consideration of our quantitative analyses, although we cannot agree with the referee’s conclusion about the impact of our current study. In particular, we would like to focus here not only on the specific findings of the two studies that, as we mentioned above, reach different conclusions, but also on the general meaning of our new study in the context of understanding the role of GDF15. Besides during development and aging, GDF15 expression is promptly upregulated in several pathologies including, cancer, injury and neurodegeneration. Indeed, a growing body of evidence has highlighted the possible role of GDF15 as stress hormone and mitokine, which could be part of a conserved mechanism of the body to signal and respond to stress. Within this context, it would be very important to understand the effect of GDF15 on stem cells, as it may prompt not only understanding of the physiological role of GDF15 but also of the mechanism underlying the response and contribution of stem cell to stress and injury. Supporting this view, it was recently shown that GDF15 is upregulated in quiescent neural stem cells following brain injury (Llorens-Bobadilla et al., 2015). However, the main problem in investigating the physiological role of GDF15 is that the expression of its recently discovered receptor is very limited in the brain. Therefore, our data here are important not only because of the novel effect that they describe for GDF15 in NSC development, but also because they show that GDF15 directly affects stem cell behavior. We have now followed the suggestion of the reviewer and re-edited several parts of the manuscript including editorial changes to highlight the relevance of our findings and the figures to better illustrate them. In particular, we have added also additional analyses to support our conclusions. These include data illustrated in Fig.1B, C; Fig. 4B, G and Fig. 5G.

    *I also have two concerns which may help to improve the current draft. Firstly, I would suggest considering the data presentation as many of the counting data are not accompanied by representative images or a detailed description of the methods, which could impede the credibility of the data. For instance, it is not clear to me how the authors judged the apical vs subapical progenitor counting as neither the pictures nor the methods clearly specified how these are distinguished. Same to the Figures 3 and 8. *

    We thank the reviewer for these suggestions, which we have now implemented in our revised manuscript. These changes include addition of representative images to Figs. 1, 2, 3, 4, 5, 6, as well as an illustration of how apical vs subapical cells were counted in Fig. 2A, an illustration showing how ependymal and single-ciliated cells were determined in Fig. 8A, and more detailed descriptions in the materials and methods section.

    *I would also suggest the authors may wish to carefully review the text as many abbreviations are not properly stated (for example, what are P+ progenitors?), or not properly explained why the particular gene expression is analysed (EGF, Sox2, CXCR4). This is also applied to the anatomical jargon (like apical, subapical etc), which can be specified in the figure by introducing the cartoons or point-on images. *

    We thank the reviewer for pointing these shortcomings out. We have now done a careful proofreading of the text and implemented changes in the relative figures.

    The changes include: Explanation of Prominin-1-expressing (P+) progenitors (p. 10, line 35) and apical and subapical progenitors (p.3, line 11-23), as well as more detailed reasoning for the analysis of gene expression for EGFR, Sox2 and CXCR4 (p. 9, line 5-7; p.12, line 44 and following)

    *Last not least, I should point out that the author uses less commonly used terminology. To my knowledge, the SVZ progenitors in the GE are now called basal progenitors (Bandler et al 2017 for example) and the word intermediate progenitors is used for the Tbr2+ IP cells in the developing cortex. MASH-1 is an old gene name as it is revised as Ascl1 (please refer to any recent papers and web databases such as Mouse Genome Informatics, and NCBI, for instance). The use of prevailed wording will help the readers to understand the presented story. *

    We have now revised the terminology according to the reviewer’s suggestion.

    Minor comments:

    Abstract Line 12-15 is confusing: What does "genotype" means?

    We used the term to refer to the genetic differences between WT and GDF15 mutant mice with respect to GDF15 expression. We apologize for the lack of clarity. We have now modified the paragraph in an effort to improve its clarity.

    Introduction *Despite the focus of this paper being on the proliferation in GE, the introduction mixed up the references describing the dorsal telencephalon. It's better to cite the ventral GE as some progenitor behaviours are different from the ones in the dorsal. *Maybe it's better to dedicate more to describing the lineage trajectory in the ventral GE and the molecular players (such as EGF), which makes it harder to understand the rationales of the several experiments.

    We would like to thank the reviewer for making this helpful comment. In the introduction we have tried to make two points: firstly, to clarify how the different progenitor types in the VZ can be distinguished based on the localization of their site of mitosis and secondly the importance of studying GDF15 in the context of NSCs in the subependymal zone of the lateral ventricle. For the first point we have several studies referring to dividing dynamics of radial glia in the developing cortex. This reflect the fact that many papers have studied both apical and subapical radial glia within the context of the developing cortex, unlike subapical progenitors, which were first discovered in the developing ganglionic eminence. A similar problem applies to the analysis of EGFR expression in the context of VZ progenitors, although we agree with the reviewer that it should introduced for a better understanding of our analyses. Therefore, we have now introduced several changes in our introduction to eliminate the shortcomings and to offset the imbalance in terms of citations.

    Results

    Fig1: V/SVZ -> VZ? I think V means ventricles while VZ is for the ventricular zone. Single-channel images should be presented to demonstrate the positive or negative cells for each antigen. Only a subset of progenitors in the adult SVZ is GDF15 positive although this is not described in the text.

    We have now replaced V/SVZ with V-SVZ, meaning ventricular-subventricular zone, throughout the manuscript and added single channel images to all figures where it is relevant.

    *Why the GDP15 staining was performed only in the adult sections, but not in E18 while the GFRAL is shown in both stages? The text claims "GDF15 is particularly expressed in the germinal region of the GE" but I did not find the data shown in this draft. *

    The fact that GDF15 is expressed in the choroid plexus and in the subependymal region of the lateral ventricle was first observed in the neonatal rat brain and prompted us to investigate the hypothesis that GDF15 may affect NSCs. Moreover, in our previous manuscript we have confirmed that GDF15 is expressed in neural progenitors of the embryonic murine GE (Carrillo-Garcia et al., 2014). In this new manuscript, we have complemented these data by adding the missing information concerning the expression of the protein in the adult V-SVZ. Notably, we also investigate for the first time the expression of the receptor in this area. This is key issue in the field, since the expression of GFRAL has been reported only in few regions of the brain, which is in apparent contrast with the growing list of effects in which GDF15 has been involved. For completeness of information and to further strengthen our conclusions we have now added new set of images in figure 1B, C showing co-expression of GFRAL and EGFR.

    *Line 24-25: I did not understand this statement. *

    The sentence refers to the results published in our previous paper, as mentioned in our reply above, which illustrate expression of GDF15 in the GE at different ages of development and in the adult V-SVZ. In an effort to improve its clarity, we have now modified the sentence into: “Consistent with these observations, we have previously reported that in the GE, *Gdf15 *transcripts increase at late developmental age and remain high in the adult V-SVZ.”

    Fig. 2 Line 33: what are apical P+ progenitors?

    We apologize for this shortcoming. P+ is the abbreviation for Prominin-1 immunopositive progenitors. This information has been now added to the text.

    *Fig. 2A: The total analysed cells are not described in M&M. *

    We have now added this information in the relevant section of the manuscript (p. 7, lines 36-43).

    *Fig 2 C and D. While the counting of apical or subapical progenitors has been done respectively, the representative images of which regions are judged as apical or subapical are not shown. This comment also applies to Line 41: I did not get the logic of how this analysis will be able to distinguish apical or subapical cell division. *

    Mitotic apical and subapical progenitors have been detected on the basis of the position of their nuclei. Namely, mitosis was considered apical if the nucleus of the dividing cell was within two nuclei distance (~ 10 µm) of the apical surface, and considered subapical if the nuclei of the dividing cells was at a greater distance from the apical surface. Besides adding this information to the manuscript, see “Image analysis” in the “Materials and Methods” section, we have now illustrated our approach in the new Fig. 2B.

    *Fig. 2 E and F: I am not sure why the proliferation was assessed in vitro whole mount cultures. IP injection in vivo animals would be more convincing. *

    We have used the same whole mount preparation to determine changes in proliferation upon acute fixation of the tissue. We have then determined the effect of growth factors and pharmacological modulators in whole-mount explants preparation as this would allow us to test their effect in standardized conditions. For the sake of consistency, we have then used the same whole mount explant setting to investigate proliferation by means of IdU incorporation. We selected this mean of analysis because changes in proliferation were already detected upon tissue fixation, and direct exposure of the tissue to the pharmacological modulators allowed us to investigate the direct effect of the drugs on proliferation behavior. Using this setting, we have obtained data that are compatible and consistent with our analysis on acutely fixed preparations. We agree with the reviewer that these experiments could be also repeated by injecting the IdU in vivo, however this would be against the current animal 3R guidelines that prompt to minimize the use of animal in vivo experiments and only when they cannot be replaced by alternative approaches in vitro or ex vivo.

    *Fig. 3 I am not sure the mitotic spindle orientation analysis is very informative to stand out as one independent figure. In some contexts (Noctor et al 2008), it does not correlate to the asymmetric or symmetric division modes. *

    We would like to like to respectfully disagree on this issue. The reason why we think this data set is important is twofold. Firstly, previous papers pointing at changes in the number of NSCs in the GE, have established that this was caused by a change in the spindle orientation leading to the generation of extra SNP (Falk et al., 2017), indicating a role for the orientation of the mitotic spindle in this context. Since we observed that GDF15 promotes not only progenitor proliferation but also apical divisions, it is important to show that this effect does not reflect a change in spindle orientation. Secondly, these data set highlights an age-dependent effect on the orientation of the mitotic spindle that is fully consistent with previously published data supporting the solidity of our findings. However, since we did not see any significant differences between the WT and Gdf15-/- animals, we have decided to move this data to a supplementary figure (new supplementary figure S3).

    *Fig. 4 I am not convinced by this data since how the fluorescent intensity is measured is not described. If the internal controls to adjust the staining variation among samples are not used, the data is not convincing to me. The representative pictures are not convincing either to claim the substantial differences. Perhaps immunoblotting is better to be employed to quantify the protein expression difference. *

    We have now added additional pictures with higher magnification to show the difference in EGFR intensity, including a calibration bar (Fig. 4). Quantitative analysis showed a trend decrease at E18 which is strongly significant in the adult V-SVZ. We now also show analysis of phEGFR and modified extensively the relative result section (see also our reply to the comment on *Evidence, reproducibility and clarity *of reviewer 2). Furthermore, we have added the following paragraph to the methods section:

    “For fluorescence intensity measurements of EGFR, slices stained at the same time with the same antibody solutions, and imaged on the same day with constant confocal microscope settings (laser intensity, gain, pixel dwell time), were measured using Fiji/ImageJ. Raw immunofluorescence intensity was normalized by subtracting background fluorescence levels, i.e. fluorescence in cells considered negative for EGFR. To rule out any unspecific secondary antibody binding, fluorescence was compared to slices incubated with secondary, but not primary antibodies (2nd only control); no difference was found between 2nd only control and cells considered negative in EGFR-labelled samples, or between 2nd only controls of different genotypes.”

    *Fig. 5 This is very busy figure composed of mainly counting graphs of different experiments. I think at least it is better to separate the data in vivo or culture. *

    We have now rearranged the figure according to the reviewer’s suggestion. We have moved, also according to Reviewer 2’s suggestions, some less relevant data to supplementary figure S4 (that is, previous panels G-J) and added confocal images to illustrate the results of previous panels C-E. We hope that this improved the focus and clarity of this figure.

    *Fig. 6. Pictures of Day 2 and Day 7 should be presented to highlight the difference between them. *

    We have now added pictures showing the cell culture at DIV2 and DIV7, as well as the different treatments, in Fig. 6A.

    *Fig. 7 Mash-1 should be rephrased as Ascl1. *

    We have now changed the name of the gene throughout the manuscript.

    Fig. 8 A: I am not sure why these pictures are B&W even though the two antigens are stained. The main text needs more description since no explanation of FOP, b-catenin etc. The picture of GFD15 KO looks having massive numbers of FOP+ cells, which is not correlated to the counting, I guess?

    We apologize for the lack of clarity. We have now added additional panels (Fig. 8A) to illustrate the rationale behind the analysis and to demonstrate how ependymal and single-ciliated cells were counted. We have added the following sections to the manuscript text:

    Materials and methods:

    “Both the β-catenin and fibroblast growth factor receptor 1 oncogene partner (FOP) primary antibodies are mouse monoclonal antibodies of the same immunoglobulin class. Therefore, for this double immunostaining both antigens were revealed using the same secondary antibody and each was distinguished based on the localization and morphology of the labelling, which is lining the cell boundaries or at the basal body of the cilia for β-catenin and FOP, respectively.”

    Results:

    “We here used β-catenin to label cell-cell contacts, thereby visualizing cell boundaries, and fibroblast growth factor receptor 1 oncogene partner (FOP), a centrosomal protein, to visualize the basal body of the cilia. As both β-catenin and FOP-antibodies where derived from the same host species, the antigens were labelled in a single fluorescent channel and differentiated based on label localisation and intensity (Fig. 8A). Cells with a single centrosome or centrosome pair (one to two FOP+ dots) were counted as single-ciliated (SC), whereas cells with more than two centrosomes, i.e. multiciliated cells, were counted as ependymal (Epen; Fig. 8A).”

    We have also added the following paragraph to the figure legend:

    “(A) Schematic showing counting of ependymal (Epen) and single-ciliated (SC) cells using FOP and β-catenin as markers. (a’) Closeup of WT image in (B), showing β-catenin, indicating cell-cell-contacts, and FOP, indicating ciliary basal bodies/centrosomes, in a single channel. Scale bar = 10 µm. (a’’) β-catenin and FOP labels are distinguished by location, morphology and label intensity, with FOP being single dots that are more intense than β-catenin and located within the cell boundaries. (a’’’) Cells containing one or two centrosomes were considered SC cells (red), while cells with more than two centrosomes were considered multiciliated and therefore Epen (blue).”

    We would also like to point out to the reviewer that since FOP was used as a label for the ciliary base, the “massive numbers of FOP+ cells” (i.e., multiciliated cells) were indeed quantified in Fig. 8B (now 8C) as ependymal cells (Epen).

    ** Referees cross-commenting**

    I agree with Reviewer #2's comment that despite the amount of data presented, they are not presented in a coherent manner. I would suggest revising carefully before submitting to any journals. As detailed above the manuscript has been revised to improve clarity and coherence according the reviewer’s suggestions.

    Reviewer #1 (Significance (Required)):

    *The presented finding of the role of GDF15 in the ventral progenitors are evident and a new finding has not been reported. Since the same effects and signalling pathways involved in adult hippocampus neurogenesis are previously published by the same authors, the impact of the current manuscript is limited. I think heightening the role of GDF15 in the biological context of ventral progenitors, or alternatively, making a comparison to the previous finding would greatly improve the quality of the draft. In my opinion, this work would be appealing to the community of neural stem cells but maybe not to the broad audience. My expertise is neurodevelopmental biology focusing on neuronal lineages and neurogenesis. *

    We have already clarified that the effect that we report here of GDF15 on NSCs is not only novel, but is also very different from what we have previously observed in the hippocampus (see also our reply above to the comment on evidence, reproducibility and clarity of reviewer 1). Although many environmental signals and growth factors have been implicated in the regulation of NSC proliferation and self-renewal, GDF15 is, to our knowledge, one of the few factors directly regulating the number of apical NSCs. Following the suggestion of the reviewer, we have now revised our manuscript in order to highlight the difference between the two studies. Besides being important within the field of NSCs, we believe that our data are also important for understanding the physiological role of GDF15, whose expression is increased during development and in the response to a growing list of stressors. For such an understanding, it is essential to identify target cell populations which can directly respond to the growth factor. Our finding that the GDF15 receptor is expressed in NSCs provide first evidence that GDF15 can directly modulate stem cell development, providing a first function for the increase in its expression. Moreover, our observation that GFRAL continues to be expressed in adult NSCs opens up to the possibility that the increase in the expression of the growth factor in the stress response is to recruit/modulate stem cell behavior. Consistent with this scenario, it was recently observed that brain injury promoted and increase of GDF15 expression in NSCs (Llorens-Bobadilla et al., 2015).

    • Reviewer #2 (Evidence, reproducibility and clarity (Required)): *

    Here the authors explore the role of GDF15 during development of the adult neural stem cell niche at the lateral wall of the lateral ventricle using GDF15 knock-out mice. They find increased progenitor proliferation at neonatal stages and at 8weeks, compensated by neuronal death. Further they report that EGFR+ cells are arranged differently in the GDF15 mutants (in clusters rather than columns) with also lower levels of EGFR. This is surprising to me, as the authors observe an increase in proliferation. They then report that addition of EGF leads to an increase in prominin+ progenitors in the GDF15KO, but not the WT, but there is lower levels of EGFR in the KOs. They then block CXCR4, which is allegedly required for GDF15 to modulate EGFR expression, and they find that this blocking reduces proliferation mostly in WT cells. As can be seen from this summary, to me, the model of how GDF15 loss is supposed to increase proliferation is not clear. Even less so, when in adult SVZ, EGFR+ progenitors were increased, while EGFR was reduced at postnatal stages. Beyond this, the authors show convincingly that ependymal cells are increased at adult stages, while I see no data supporting their claim of NSCs to be increased (at least not reaching levels of significance).*

    • Taken together, this manuscript contains a lot of data, but to me no coherent picture emerges. If the picture is that the higher proliferation rate of apical progenitors at E18 generates more ependymal cells, then this should be shown (by including analysis e.g. at P5 when ependymal cells emerge). How GDF15 would affect proliferation in general is also not clear to me - maybe an unbiased analysis by RNA-seq could help to separate the main effects from diving into known candidates that seem not to explain the main aspects.*

    The reviewer mentions multiple aspects of our study that we would like to clarify. Therefore, we apologize for our lengthy reply.

    Firstly, the apparent contradiction between the increase in progenitor proliferation despite the concomitant decrease in EGFR levels. It is long known activation of EGFR regulates multiple aspects of progenitor behavior including proliferation, migration and differentiation. It is also known that responsiveness of NSCs to EGF increases during development, a process that is paralleled by an increase in expression of EGFR expression increase with developmental age, peaking around the perinatal age. However, since NSCs start to slow their cell cycle and enter quiescence exactly at the same time in which the increase in responsiveness to EGF and in EGFR expression in NSCs during this same period, it is clear that there is not a linear correlation between NSC proliferation and EGFR expression. A possible explanation for this apparently counterintuitive observation is that the increase in EGFR expression is paralleled by an increase in the expression of Lrig1, a developmental negative regulator of EGFR (Jeong et al., 2020). Moreover, in our study we report that lack of GDF15 leads to a decrease in the expression of EGFR protein and not in the levels of EGFR mRNA. In light of the well-known feedback mechanism by which in the presence of high EGFR activation the receptor transits to late endosome for degradation, a decrease in the protein levels could actually represent a higher level of activation EGFR signalling in the mutant progenitors than in the wild type counterpart. This is consistent with the characteristics of punctuate EGFR immunostaining we see in the mutant tissue and our analysis of EGFR activation. Our data show that despite difference in activation kinetics, wild type and mutant progenitors similarly respond to exogenous stimulation with EGF. Moreover, there is no difference between the two genotypes in the expression of EGFR in mitotic cells, and blockade of EGFR more dramatically affects the proliferation of mutant than WT progenitors. Finally, exposure to exogenous EGF promotes the proliferation of WT but not mutant progenitors. Taken together, these observations suggest that endogenous activation of EGFR driving proliferation is higher in mutant than WT progenitors. Consistent with this hypothesis, our new data illustrated in Fig. 5A-G of the revised manuscript, show that EGFR is similarly phosphorylated in mutant and in WT progenitors and that levels of Phospho-EGFR are observed in regions with low levels of EGFR expression, especially in the postnatal V-SVZ.

    With respect to the effect of CXCR4, we have investigated its effect with respect to the ability of GDF15 to promote EGFR expression at the cell surface and secondly with respect to affect proliferation in vivo and in vitro. Both experiments reveal a permissive role of the receptor, whose activity is necessary for GDF15 to promote EGFR expression at the cell surface and for cell to undergo proliferation. While these observations confirm in part our previously published data, the molecular mechanisms underlying these effects remain unclear. However, since AMD on its own does not affect EGFR expression in either WT or mutant progenitors, the two effects are not related. Despite the absence of a clear mechanism by which CXCR4 affects proliferation, our data indicate that the permissive effect of CXCR4 is more important for the proliferation of TAPs rather than for NSCs. Therefore, the different effect of CXCR4 inhibition between WT and mutant progenitors likely reflect the fact that the latter are enriched in NSCs.

    Finally, the evidence that NSCs are significantly increased in the mutant V-SVZ is reported in Fig. 8. In this figure, it is clearly reported that compared to the WT counterpart at early postnatal stages, only the multiciliated ependymal cells are significantly increased in the mutant niche, whereas uniciliated progenitors display only a trend increase (Panel C). However, the total number of apical cells is increased in the mutant V-SVZ, indicating that the number of both cell types are likely increased. Consistent with this hypothesis, in panel G of the same figure we show that in the adult V-SVZ, when the ependymal cells are fully differentiated, also apical GFAP+ NSCs are significantly increased. Notably, in this figure we show that GFAP+ NSCs also display a primary cilium, an elongated morphology and lack multiple cilia, and therefore are not atypical ependymal cells. Finally, in supplementary table S6, we show no difference in terms of % of clone forming cells between dissociated cell preparations of the WT and mutant V-SVZ. These observations and our finding of increased Ki67 apical expression in the adult V-SVZ, illustrated in supplementary figure S2B, clearly show that apical NSCs are increased.

    We have now introduced multiple modification in text and figures to clarify the mechanisms underlying the effect of GDF15 ablation on EGFR expression and activation and the differential effect of CXCR4 on WT and mutant progenitors. The new data set illustrating phosphoEGFR are illustrated in figure 4B, G. We have also modified figure 8 in an effort to illustrate more clearly the effect of lack of GDF15 on NSC number.

    *Major comments: *

    *1) Inconsistencies start already in Figure 1: The authors show expression of the receptor at neonatal stages (much higher) and adult stages, but GDF15 is shown only in adult stages and the citations of their previous work suggests that indeed it may not be present at this early stage in the GE VZ (p.8, line 10). If it is, please show. If it isn't, could it be that it is in the CSF and signals only to apical cells? *

    As the reviewer rightly mentions, GDF15 is present in the CSF and signals mainly to apical cells, as it is known to be secreted by the choroid plexus (Böttner et al., 1999; Schober et al., 2001). However, we would like to respectfully disagree with the reviewer. In our previous work, we clearly showed that GDF15 expression increases at E16 and is highest at E18 in the GE, which is the reason we specifically chose this timepoint for analysis (compare Carrillo-Garcia et al. (2014), Fig. 1A). In this manuscript we have also shown that EGFR expressing progenitors in the GE express GDF15. As the expression of GDF15 at embryonic and neonatal ages has already been investigated by us and others for two decades (see also Schober et al. (2001)), we refrained from showing expression of GDF15 at these ages again. However, we have now modified the relevant result section to clearly highlight the existence of this previous findings.

    *2) An overview of the KO phenotype by lower power pictures would be helpful. For example an overview over the GE and PH3 immunostaining WT and KO at comparable section levels. *

    Our analysis is based on whole-mount preparation of the whole GE. To standardize our analysis the same number of pictures were taken at similar locations to obtain a quantification representative of the apical surface of the whole GE. Therefore, the areas of interest were not selected on the basis of the number of mitotic cells and the differences observed do not reflect a positional effect. Lower power pictures illustrating the whole GE, are unlikely to be helpful, because they would not show the nuclear immunostaining. However, we have now modified the relevant Material and Methods section as follows to describe the standardization of our quantitative analysis: “Whole mounts were imaged using a Leica TCS SP8 confocal microscope with a 40x or 63x oil immersion objective and LASX software (Leica). For the quantification, an average of three different regions of interest were chosen at fixed rostral, dorsal and ventral position of the GE or V-SVZ and averaged for the collection of a single data set.”

    1. Figure 2B- where is the apical surface, where are we in the GE? Where was quantification done? *

    We have now added images detailing the localization of apical and subapical cells in new Fig. 2A, as well as further clarifications of the imaging and quantification in the materials and methods section.

    1. Clarify the part with EGF signaling and/or take a more comprehensive view by a proteomic or transcriptomic approach, as EGFR and CXCR4 which were already investigated previously, may not explain the phenotype. *

    We agree with the reviewer that our data should prompt a more comprehensive approach. However, this is surely a work worthy of a separate manuscript, since we agree with the reviewer that changes in EGFR and CXCR4 do not fully explain the effect of GDF15 on proliferation. We have now clarified our conclusions about EGF signalling, modifying the relevant part in the result section. We have also modified the abstract as follows:

    “From a mechanistic point of view, we show that active EGFR is essential to maintain proliferation in the developing GE and that GDF15 affects EGFR trafficking and signal transduction. Consistent with a direct involvement of GDF15, exposure of the GE to the growth factor normalized proliferation and EGFR expression and it decreased the number of apical progenitors. A similar decrease in the number of apical progenitors was also observed upon exposure to exogenous EGF. However, this effect was not associated with reduced proliferation, illustrating the complexity of the effect of GDF15.”

    *5) Do the authors actually think that the effects on EGFR are in the cells expressing the GDF15 receptor? Then please show co-localization. *

    As both EGFR and GFRAL are widely expressed in the embryonic GE (see Figs 1B and 4B), making overlap inevitable, we did previously not assume the need to show co-localization. We have now added images showing co-localization of EGFR and GFRAL in E18 and adult brain sections in Fig. 1B.

    *6) Figure 5D shows virtually no apical mitosis in WT, but indeed there are apical mitosis in WT E18 GE as one can also see in panel 5A. *

    We apologize for the confusion. In the manuscript, we use Ki67 and analysis of nuclear morphology to determine the number of cells undergoing cell division, i.e. in meta-, ana- or telophase and immunostaining with antibody with phH3+, which stains additionally cells also at late G2 and early mitotic stages. Consistent with this, the number of mitotic cells scored with Ki67 and quantified in Fig. 5D is smaller than the number of phH3+ cells that is illustrated in Fig. 5A. Throughout the manuscript, cells labeled by phH3 immunoreactivity are named “phH3+ cells”, as quantified in Fig. 5B, whereas with “dividing cells” we refer to cells with Ki67 labeling that show nuclear morphology of meta-, ana- or telophase. We have, also according to the suggestions of reviewer 1, added images of the whole mounts analyzed for Fig. 5D, as well as the following text in the materials and methods section: “Cells were considered dividing if the nuclei were labelled with Ki67 and the nuclear morphology showed signs of division, i.e. meta-, ana- or telophase, in DAPI and Ki67-channels. For the sake of clarity, “dividing cells” only refers to this way of detection, while cells positive for phH3 are termed “phH3+ cells”, as phH3 also labels cells in interphase and prophase, as well as late G2 phase.”

    *7) For the effect on ependymal cell generation it could be good to include an intermediate age, such as P5-7, when ependymal cells differentiate, staining e.g. for Lynkeas or Mcidas, known fate determinants regulating ependymal cell differentiation at that time. *

    Most of our research was performed in either E18 or adult animals, where ependymal cells are either not yet present or already fully differentiated. Since ependymal cell differentiation starts at birth, we used P2 animals to look at ependymal cell differentiation. As shown in Fig. 8B, C this age is appropriate to study early ependymal differentiation, as a lot of multiciliated ependymal cells are already present at this age, and the difference between WT and Gdf15-/- animals is clearly visible and significant. While another age or additional markers might be interesting, we argue that it would not add to the conclusion or significance of this paper, as we can see this phenotype already at age P2 and it can still be detected it in adult animals.

    *Minor comments: *

    *-) p. 8, subapical progenitors are mentioned in line 42 without explaining how they are defined. *

    We have now added more detailed definitions of apical and subapical progenitors to the introduction.

    *-) p.8, line 44: the word increased in mentioned 2x *

    We have removed the additional word.

    *-) In the description of Figure 8 C and D seem to have been mixed up. *

    We have changed the description of Fig. 8.

    ** Referees cross-commenting***

    *I also fully agree with the point that this manuscript is very difficult to read. I think that anyhow the results have to be reorganized to focus on the most important data, so rewriting will have to be done for clarity either way. *

    We apologize for the lack of clarity we have now extensively modified and re-edited the manuscript in an effort to improve its clarity.

    Reviewer #2 (Significance (Required)): *

    Exploring signmalling factors important for the stem cell niche is important, and the GDF15 indeed seems to have an effect there. The problem is, that much has been done with this factor already, but of course a mechanistic understanding of whats going on is important and could be the strength of this manuscript. However, it is really not clear, which mechanisms causes what. What is clear, is that the increased proliferation of neuronal progenitors is counterbalanced by death. Its also clear that ependymal cells are increased, which is an interesting effect. But how and why is not clear and may be the best to focus in this paper. *

    As the reviewer mentions, several publications focus on GDF15. However, there is only one publication investigating the effect of GDF15 on neural stem cells and this focuses on the hippocampus. Therefore, we would like to respectively disagree with the conclusion of the reviewer “that much has been done with this factor”. Moreover, a serious problem with previous studies investigating GDF15 is the fact that its receptor is scarcely expressed and therefore it is not clear if these studies investigate direct or indirect effects of the growth factor. Since we here for the first time show that neural stem cells in the GE and V-SVZ express GDF15-receptor GFRAL, our study for the first time show a direct involvement of GDF15 on proliferation, number of ependymal cells and, as detailed in our reply above, apical NSCs. This knowledge is not only relevant to the field of normal and cancer stem cells, but also within the context of the role of GDF15 as mitokine and as stress hormone (see also our reply to major comments 2 of reviewer 1). Therefore, although we agree with the reviewer that the molecular mechanisms underlying the effect of GDG15 need further investigation, our data are novel and of relevance to the general scientific community.

    References:

    Böttner, M., Suter-Crazzolara, C., Schober, A., Unsicker, K., 1999. Expression of a novel member of the TGF-beta superfamily, growth/differentiation factor-15/macrophage-inhibiting cytokine-1 (GDF-15/MIC-1) in adult rat tissues. Cell Tissue Res 297, 103-110.

    Carrillo-Garcia, C., Prochnow, S., Simeonova, I.K., Strelau, J., Hölzl-Wenig, G., Mandl, C., Unsicker, K., von Bohlen Und Halbach, O., Ciccolini, F., 2014. Growth/differentiation factor 15 promotes EGFR signalling, and regulates proliferation and migration in the hippocampus of neonatal and young adult mice. Development 141, 773-783.

    Falk, S., Bugeon, S., Ninkovic, J., Pilz, G.A., Postiglione, M.P., Cremer, H., Knoblich, J.A., Gotz, M., 2017. Time-Specific Effects of Spindle Positioning on Embryonic Progenitor Pool Composition and Adult Neural Stem Cell Seeding. Neuron 93, 777-791 e773.

    Jeong, D., Lozano Casasbuenas, D., Gengatharan, A., Edwards, K., Saghatelyan, A., Kaplan, D.R., Miller, F.D., Yuzwa, S.A., 2020. LRIG1-Mediated Inhibition of EGF Receptor Signaling Regulates Neural Precursor Cell Proliferation in the Neocortex. Cell Rep 33, 108257.

    Llorens-Bobadilla, E., Zhao, S., Baser, A., Saiz-Castro, G., Zwadlo, K., Martin-Villalba, A., 2015. Single-Cell Transcriptomics Reveals a Population of Dormant Neural Stem Cells that Become Activated upon Brain Injury. Cell Stem Cell 17, 329-340.

    Schober, A., Böttner, M., Strelau, J., Kinscherf, R., Bonaterra, G.A., Barth, M., Schilling, L., Fairlie, W.D., Breit, S.N., Unsicker, K., 2001. Expression of growth differentiation factor-15/ macrophage inhibitory cytokine-1 (GDF-15/MIC-1) in the perinatal, adult, and injured rat brain. J Comp Neurol 439, 32-45.

  6. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Here the authors explore the role of GDF15 during development of the adult neural stem cell niche at the lateral wall of the lateral ventricle using GDF15 knock-out mice. They find increased progenitor proliferation at neonatal stages and at 8weeks, compensated by neuronal death. Further they report that EGFR+ cells are arranged differently in the GDF15 mutants (in clusters rather than columns) with also lower levels of EGFR. This is surprising to me, as the authors observe an increase in proliferation. They then report that addition of EGF leads to an increase in prominin+ progenitors in the GDF15KO, but not the WT, but there is lower levels of EGFR in the KOs. They then block CXCR4, which is allegedly required for GDF15 to modulate EGFR expression, and they find that this blocking reduces proliferation mostly in WT cells. As can be seen from this summary, to me, the model of how GDF15 loss is supposed to increase proliferation is not clear. Even less so, when in adult SVZ, EGFR+ progenitors were increased, while EGFR was reduced at postnatal stages. Beyond this, the authors show convincingly that ependymal cells are increased at adult stages, while I see no data supporting their claim of NSCs to be increased (at least not reaching levels of significance). Taken together, this manuscript contains a lot of data, but to me no coherent picture emerges. If the picture is that the higher proliferation rate of apical progenitors at E18 generates more ependymal cells, then this should be shown (by including analysis e.g. at P5 when ependymal cells emerge). How GDF15 would affect proliferation in general is also not clear to me - maybe an unbiased analysis by RNA-seq could help to separate the main effects from diving into known candidates that seem not to explain the main aspects.

    Major comments:

    1. Inconsistencies start already in Figure 1: The authors show expression of the receptor at neonatal stages (much higher) and adult stages, but GDF15 is shown only in adult stages and the citations of their previous work suggests that indeed it may not be present at this early stage in the GE VZ (p.8, line 10). If it is, please show. If it isn't, could it be that it is in the CSF and signals only to apical cells?
    2. An overview of the KO phenotype by lower power pictures would be helpful. For example an overview over the GE and PH3 immunostaining WT and KO at comparable section levels.
    3. Figure 2B- where is the apical surface, where are we in the GE? Where was quantification done?
    4. Clarify the part with EGF signaling and/or take a more comprehensive view by a proteomic or transcriptomic approach, as EGFR and CXCR4 which were already investigated previously, may not explain the phenotype.
    5. Do the authors actually think that the effects on EGFR are in the cells expressing the GDF15 receptor? Then please show co-localization.
    6. Figure 5D shows virtually no apical mitosis in WT, but indeed there are apical mitosis in WT E18 GE as one can also see in panel 5A.
    7. For the effect on ependymal cell generation it could be good to include an intermediate age, such as P5-7, when ependymal cells differentiate, staining e.g. for Lynkeas or Mcidas, known fate determinants regulating ependymal cell differentiation at that time.

    Minor comments:

    • p. 8, subapical progenitors are mentioned in line 42 without explaining how they are defined.
    • p.8, line 44: the word increased in mentioned 2x
    • In the description of Figure 8 C and D seem to have been mixed up.

    ** Referees cross-commenting**

    I also fully agree with the point that this manuscript is very difficult to read. I think that anyhow the results have to be reorganized to focus on the most important data, so rewriting will have to be done for clarity either way.

    Significance

    Exploring signmalling factors important for the stem cell niche is important, and the GDF15 indeed seems to have an effect there. The problem is, that much has been done with this factor already, but of course a mechanistic understanding of whats going on is important and could be the strength of this manuscript. However, it is really not clear, which mechanisms causes what. What is clear, is that the increased proliferation of neuronal progenitors is counterbalanced by death. Its also clear that ependymal cells are increased, which is an interesting effect. But how and why is not clear and may be the best to focus in this paper.

  7. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary:

    This paper illustrates the role of GDF15 in ganglionic eminence featuring the influence on neurogenesis and progenitor proliferation. Using the conventional knockout GDF-15 mouse lines, the authors provide a series of counting data indicating that GDF15 controls neurogenesis in the late embryonic or adult neural stem cells in the ventral forebrain.

    Major comments:

    The entire story in this manuscript seems similar to the previous findings where the authors demonstrated the role of GDF15 in the hippocampal neural stem cells in relation to EGF and CXCR4.

    The data and the conclusion presented here sound reasonable to me. The current manuscript, however, gave me the impression that the story is less impactful and rather descriptive. To improve the quality of the current draft, the authors may wish to clearly highlight the novelty of the findings, not simply apply the previous strategy to the other anatomical brain regions. Alternatively, the draft could emphasize the similarity of utilising the same biological strategies to control the number of adult stem cells in the distinct stem cell niche.

    I also have two concerns which may help to improve the current draft.

    Firstly, I would suggest considering the data presentation as many of the counting data are not accompanied by representative images or a detailed description of the methods, which could impede the credibility of the data. For instance, it is not clear to me how the authors judged the apical vs subapical progenitor counting as neither the pictures nor the methods clearly specified how these are distinguished. Same to the Figures 3 and 8.

    I would also suggest the authors may wish to carefully review the text as many abbreviations are not properly stated (for example, what are P+ progenitors?), or not properly explained why the particular gene expression is analysed (EGF, Sox2, CXCR4). This is also applied to the anatomical jargon (like apical, subapical etc), which can be specified in the figure by introducing the cartoons or point-on images.

    Last not least, I should point out that the author uses less commonly used terminology. To my knowledge, the SVZ progenitors in the GE are now called basal progenitors (Bandler et al 2017 for example) and the word intermediate progenitors is used for the Tbr2+ IP cells in the developing cortex. MASH-1 is an old gene name as it is revised as Ascl1 (please refer to any recent papers and web databases such as Mouse Genome Informatics, and NCBI, for instance). The use of prevailed wording will help the readers to understand the presented story.

    Minor comments:

    Abstract Line 12-15 is confusing: What does "genotype" means?

    Introduction Despite the focus of this paper being on the proliferation in GE, the introduction mixed up the references describing the dorsal telencephalon. It's better to cite the ventral GE as some progenitor behaviours are different from the ones in the dorsal.

    Maybe it's better to dedicate more to describing the lineage trajectory in the ventral GE and the molecular players (such as EGF), which makes it harder to understand the rationales of the several experiments.

    Results

    Fig1: V/SVZ -> VZ? I think V means ventricles while VZ is for the ventricular zone

    Single-channel images should be presented to demonstrate the positive or negative cells for each antigen. Only a subset of progenitors in the adult SVZ is GDF15 positive although this is not described in the text. Why the GDP15 staining was performed only in the adult sections, but not in E18 while the GFRAL is shown in both stages? The text claims "GDF15 is particularly expressed in the germinal region of the GE" but I did not find the data shown in this draft.

    Line 24-25: I did not understand this statement.

    Fig. 2 Line 33: what are apical P+ progenitors ?

    Fig. 2A: The total analysed cells are not described in M&M.

    Fig 2 C and D. While the counting of apical or subapical progenitors has been done respectively, the representative images of which regions are judged as apical or subapical are not shown. This comment also applies to Line 41: I did not get the logic of how this analysis will be able to distinguish apical or subapical cell division.

    Fig. 2 E and F: I am not sure why the proliferation was assessed in vitro whole mount cultures. IP injection in vivo animals would be more convincing.

    Fig. 3 I am not sure the mitotic spindle orientation analysis is very informative to stand out as one independent figure. In some contexts (Noctor et al 2008), it does not correlate to the asymmetric or symmetric division modes.

    Fig. 4 I am not convinced by this data since how the fluorescent intensity is measured is not described. If the internal controls to adjust the staining variation among samples are not used, the data is not convincing to me. The representative pictures are not convincing either to claim the substantial differences. Perhaps immunoblotting is better to be employed to quantify the protein expression difference.

    Fig. 5 This is very busy figure composed of mainly counting graphs of different experiments. I think at least it is better to separate the data in vivo or culture.

    Fig. 6. Pictures of Day 2 and Day 7 should be presented to highlight the difference between them.

    Fig. 7 Mash-1 should be rephrased as Ascl1.

    Fig. 8 A: I am not sure why these pictures are B&W even though the two antigens are stained. The main text needs more description since no explanation of FOP, b-catenin etc. The picture of GFD15 KO looks having massive numbers of FOP+ cells, which is not correlated to the counting, I guess? written

    ** Referees cross-commenting**

    I agree with Reviewer #2's comment that despite the amount of data presented, they are not presented in a coherent manner. I would suggest revising carefully before submitting to any journals.

    Significance

    The presented finding of the role of GDF15 in the ventral progenitors are evident and a new finding has not been reported. Since the same effects and signalling pathways involved in adult hippocampus neurogenesis are previously published by the same authors, the impact of the current manuscript is limited. I think heightening the role of GDF15 in the biological context of ventral progenitors, or alternatively, making a comparison to the previous finding would greatly improve the quality of the draft. In my opinion, this work would be appealing to the community of neural stem cells but maybe not to the broad audience. My expertise is neurodevelopmental biology focusing on neuronal lineages and neurogenesis.

  8. Version published to 10.1101/2022.12.02.518869v2 on bioRxiv
  9. Version published to 10.1101/2022.12.02.518869v1 on bioRxiv