The scaffolding protein flot2 promotes cytoneme-based transport of wnt3 in gastric cancer

Article activity feed

1. 

   Version published to 10.7554/elife.77376 on eLife

   Aug 30, 2022
2. 

   eLife

   Jul 12, 2022

   Author Response

   Reviewer #1 (Public Review):

   1. The quality of many data and some experimental should be improved. Specifically, most experiments used the overexpression approach. Genetic approaches would need to be employed, particularly in embryos.

   We have improved the experimental data as suggested. Specifically, we have added more overview pictures and have selected more representative images. In addition, we have added a new set of experiments using the paracrine colony formation assays to support the requirement for Wnt3 cytonemes in proliferation and growth as well as Flot2 function in Fig.2 and 4.

   As suggested, we have added various loss-of-function approaches. For example, we show the effect of a dominant-negative Flot2 construct and a siRNA-mediated knockdown of Flot2. We further added an F0 Crispant approach in the zebrafish …

   Author Response

   Reviewer #1 (Public Review):

   1. The quality of many data and some experimental should be improved. Specifically, most experiments used the overexpression approach. Genetic approaches would need to be employed, particularly in embryos.

   We have improved the experimental data as suggested. Specifically, we have added more overview pictures and have selected more representative images. In addition, we have added a new set of experiments using the paracrine colony formation assays to support the requirement for Wnt3 cytonemes in proliferation and growth as well as Flot2 function in Fig.2 and 4.

   As suggested, we have added various loss-of-function approaches. For example, we show the effect of a dominant-negative Flot2 construct and a siRNA-mediated knockdown of Flot2. We further added an F0 Crispant approach in the zebrafish embryo explained in the following: The Flotillin genes have undergone teleost-specific genome duplication in the zebrafish. Therefore, there are four Flotillin genes present in zebrafish: namely Flot1a,1b and 2a,2b (von Phillipsborn et al., 2005). The Flotillins prominently expressed during zebrafish blastula, and gastrula stages are Flot1a and Flot2b. Therefore, we have designed the appropriate gRNAs to knock out Flot1b and Flot2a. In detail, we used CRISPR-mediated knock-out of Flot1b and 2a by microinjections of a combination of 3 different gRNAs per gene and Cas9 to block Flot1b/2a function. We have used this approach regularly to generate F0 Crispants for individual genes or in combination (Winter et al., 2022). Our experiments showed a significant reduction of Wnt8a cytonemes in zebrafish gastrulation after KO of Flot1b/2a. We have included this new data set in Fig. 6.

   1. The dominant-negative Flot2 is the key tool utilized in the paper, but it is unclear whether this construct has been characterized in the system used and how it affects endogenous protein function. Has its impact on the endogenous Flot2 been examined?

   The construct has been characterised by Neumann-Giesen et al., 2004. We have further characterised this construct and added a confocal analysis (Supp. Fig. 1E) showing the effect of DN-Flot2-GFP expression on endogenous Flot2 localisation, assessed by IF. When comparing to surrounding untransfected cells, we show that the Flot2 mutant construct causes mislocalisation and accumulation of WT Flot2 (and loss of punctate staining).

   1. Similarly, the effectiveness and specificity of siRNA for example, the expression level of Flot2 would need to be assessed in all experiments.

   Western Blot showed successful knockdown of Flot2 expression in Supp. Fig. 2D. Furthermore, we have added an image (Supp. Fig. 2E) of Lrp6 and Flot2 antibody staining after Flot2 siRNA treatment. Minimal Flot2 staining further indicates successful and specific knockdown of Flot2.

   1. Furthermore, it is unclear whether the tagged constructs (e.g., Flot2-GFP, Wnt8a-mcherry) have been characterized and whether the tags affect the protein function.

   All fluorescently tagged proteins used have been characterised in previous papers, and any constructs cloned by us have been assessed by their ability to induce reporter activation. See cited sources in the methods section.

   1. Most images show one single cell. Could more cells be presented? The nature of the images should be disclosed. For example, are those confocal images (single plane or Z-stack)?

   Based on this comment (and comments from the other reviewers), we have used more zoomed-out images to show more than one cell, specifically in Fig. 1, Fig. 3 and Suppl. Figures. All images shown are Z-stack images unless stated, and a statement has been added to materials and methods to reflect this.

   1. The P values for which group are not clear in many panels. It is not clear which groups were analyzed. For example, Fig. 4C, D, H and many other panels.

   We have added p values for all conditions in all graphs. Usually, the experimental values are compared to the control in that graph unless indicated otherwise with bars.

   1. Statistical analyses are lacking for some panels. For example, χ2 test is needed for many panels, including Fig. 3E, and many others.

   As the reviewer suggested, we added a Pearson’s χ2 test for Fig. 3E and the new Fig. 6K.

   1. Figure 1: 1) AGS is supposed to compare with control (HFE-145). These data are missing in the chart. The cell number in AGS is significantly higher than that in other cells (25 vs, 7 and 8, line 666), which can compromise the statistical analysis.

   We have added the quantification of HFE filopodia length has been added to Fig. 1B.

   1. 1. Qualification data are needed to support the statement in Line 71-72.

   We show quantification of the effects of the expression of IRSp534K on cell number and proliferation (BrdU assay) are shown in Fig. 4. The appropriate reference has been added to this in the text.

   1. 1. Fig1D: Wnt3-positive filopodia in AGS is double compared to that in HFE-145, which is not consistent with the image shown in Fig1C. 4) Fig1H-I: The red channel is overexposed. The authors should explain why a-myox and a-evi signals were detected outside the cell (or just the background)? The more appealing evidence should be the co-localization of Myox or Evi with wnt3a on the filopodia.

   We are grateful for this comment. To address these points, we selected images more representative of the quantitative data in Fig. 1C, D. We further improved the staining of Myosin-X and Evi to reduce the background. Zoomed panels are overexposed to highlight the localisation on cytonemes. Furthermore, co-staining of Evi and Wnt3, showing co-localisation on cytonemes, has also been added (Fig. 1J)

   1. 1. Figure 2 nicely showed the impact of paracrine Wnt signaling induced by producing cells. However, there are many issues with this experiment. 1) The reporter plasmids are transiently transfected, which inevitably leads to the expression at different expression levels . How could the authors compare the expression levels as a readout in different conditions if this is the case? Better and reliable methods should use stable cell lines. Thus, the authors should make a stable 7xTCF-NLS-mCherry stable line or co-transfect the cell with GFP to show the relative transfection level. This concern also applied to other figures using 7xTCF-NLS-mCherry reporter assay.

   The HFE-145 (TCF-mCherry) cells were transfected in separate dishes but were pooled together and counted before re-plating for co-cultivation. Therefore, the transfection efficiency should be consistent. For the AGS cells, we tried generating a stable cell line, but this was unsuccessful. Nevertheless, transfection efficiencies were consistently above 70%. To strengthen our line of arguments, we have added a new set of data to monitor the consequences of Wnt3 dissemination by cytonemes in a paracrine colony-forming assay. These data support our conclusion that Wnt3 is transported on cytonemes to neighbouring cells to regulate the survival and proliferation of gastric cancer stem cells.

   1. Thus, the mCherry positive cells in Fig2B, D and F cannot present all receiving cells, as the transfection rate should not be 100%. Also, did all experiments start with a similar cell number ? Thus, the Chart in D is not accurate, and the reason that assesses the number of receiving cells is not clear. It is not clear what "per image" means in D? Is the number correct (1 cell vs 1.5 cells) in D?

   The reviewer is correct that the mCherry-positive cells cannot represent 100% of cells due to transfection efficiency. But efficiency was consistently high (>70%), and all quantifications were shown in relative values. Furthermore, all experiments started with the same number of cells (counted for transfection and again for plating during co-cultivation). Therefore, we believe the comparison of the quantifications is valid.

   1. Additionally, is it possible to image Wnt3 is being transported to the receiving cells ?

   We have added an antibody staining against Wnt3 of STF-mCh-expressing HFE-145 cells after co-cultivation with AGS cells to show strong Wnt3 staining in the recipient cells, indicating transport (Supp. Fig. 1C).

   1. Fig2 C: Western blot could be added to show the mCherry expression level in each group.

   All experiments started with the same number of cells (counted for transfection and again for plating during co-cultivation). Furthermore, we believe measuring the fluorescence in the nuclei in the receiving cells gives an accurate measure of the reporter activity.

   1. It is better to include the red channel only in E. It is difficult to see the red signal in the current images.

   The BrdU stains are brightfield images and do not use any fluorescence. Therefore, there are no channels to split.

   1. How was the qualification conducted? Could the whole population be analyzed more quantitatively?

   Quantification was conducted by counting BrdU+ cells as a percentage of total cells, counterstained with haematoxylin, as outlined in the materials and methods section. We quantified 10 separate locations with roughly 50 cells per repeat. Based on these measurements, a representation of the population was displayed.

   1. Figure 4: 1) The critical data should be that the formation of wnt3a cytoneme (length, number) is impaired in Flot2-deficient cells, which are missing in the figure and the manuscript

   We agree that the impact of Flot2 siRNA on Wnt3 should be assessed; however, because Flot2 KD reduces filopodia number and length, it would not be reliable or accurate to compare the percentage of Wnt3-positive cytonemes to controls, as this would naturally record a reduction regardless of direct effects on Wnt3. Therefore, we performed antibody staining against Wnt3 after Flot2 KD (and in DN-Flot2-expressing cells) to assess its localisation (Supp. Fig. 2). However, quantifying these observations is difficult due to the abovementioned points.

   1. 1. A-D: The expression of Flot2 should be presented in separated images. The membrane localization is not clear. Fig4D shows flot2 occasionally localized with Wnt3. Time-lapse experiments will provide additional evidence of the constant localization of Flot and Wnt3

   For clarification, we have added time-lapse images of Flot2-GFP/Wnt3-mCh expressing cells (Supp. Fig. 2G), where Flot2 and Wnt3 can be seen travelling together intracellularly.

   1. E: This panel has similar issues described in Figure 2. How was the transfection rate in E? Did all cells express Flot2 or dnFlot2? Their expression should be examined at the same time.

   As highlighted in the previous comment, whilst we cannot guarantee all cells are expressing constructs, transfection efficiency was consistently high, >70% throughout experiments. Additionally, imaging of BrdU+ cells requires a brightfield camera that can capture colour. Microscopes which can capture fluorescent images only show brightfield in black and white, whereby BrdU staining cannot be distinguished from counterstains. Therefore, showing fluorescence of cells in these BrdU stains is not possible, but cells were checked for transfection efficiency before the staining process.

   1. 1. Figure 5 is one of the key figures. However, the quality of the images is not high enough to support the conclusion. A-D: The membrane co-localization is not convincing. Better images with a membrane marker are needed. Also, it is better to present images in separate channels. The red color A should be magenta

   We are sorry that the resolution of the images seemed to be reduced in the version uploaded. We have therefore improved the resolution of the images and added an analysis of AGS cells expressing Ror2-BFP, Flot2-GFP and membrane-mCherry, which shows the membrane localisation of Flot2 and Ror2 (Supp. Fig. 3a). As suggested by the reviewer, Fig. 5A has also been changed to magenta.

   1. Did the dominated-negative Flot2 affect the expression of endogenous Flot2? Similarly, the expression of endogenous Flot2 in siRNA expressing cells should be shown

   Yes, the expression of endogenous Flot2 is reduced in mislocalised as described in previous comments. In addition, we provide further evidence showing the consequences of Wnt3 dissemination by Flot2-dependent cytonemes in the paracrine colony-forming assay. These data support our conclusion that Flot2 is required for Wnt3 transported on cytonemes to neighbouring cells to regulate the survival and proliferation of gastric cancer stem cells.

   1. Instead of showing the image of single-cell, additional experiments, for example, the western blot should provide additional evidence to show Ror2 expression on the membrane is lost

   We respectfully disagree with the reviewer. Our data suggest that Ror2 expression is not reduced upon reduction of Flot2 function. Instead, we find that Ror2 localisation is altered. Therefore, a Western Blot analysis would not be able to show a shift in the localisation of the protein.

   1. 5E. The current images are too small to appreciate the co-localization. Similarly, separated channels should be presented

   As suggested, the Golgi/Ror2-mCh images have been increased to allow appreciation of the co-localisation. In addition, images for Rab5, Rab7 and LAMP1 have been moved to the Supp. Fig. 3, with separate channels shown (including for Golgi).

   1. How many experiments have been conducted? It seems that the cell number is not high

   As advised by the reviewer, we repeated this experiment to increase the n number for each group to 10.

   1. It is necessary to describe how E-Co-efficient (PCC) was determined in more detail

   A description of how the PCC was measured has been added to the material and methods section.

   1. F: The label for the X-axis is missing

   We have added the label for the X-axis.

   1. G: The nuclei and cell boundaries are not clear ; the markers for these should be included to give confidence where and how the quantification was conducted

   We have marked the nuclei as well as the adjacent cytoplasm with asterisks - to show the localisation in which the fluorescence of KTR-mCherry was measured.

   1. Similarly, the expression of Flot2 should be examined in these experiments as it is likely not all cells express those constructs at similar levels

   Flot2-GFP (and other constructs) expression was analysed, and only expressing cells were selected for quantifying the JNK reporter. We decided against showing these channels in the images as it made the visualisation of the JNK reporter difficult.

   1. Additional experiments, for example, Western bolt to show pJNK levels, are necessary to support the conclusion further

   The KTR-mCherry reporter is a valid reporter to quantify JNK activity. We are convinced that this reporter is a better tool for measuring pJNK levels. The reporter allows visualizing JNK activity in specific cells within minutes. This advantage has been demonstrated in numerous publications, including in our own (Brunt et al., 2021). Therefore, we believe the KTR-mCherry reporter is a fast and reliable tool to measure JNK activity in individual cells.

   1. . H: The P values for which group are not clear.

   All p values above the bars are compared to the untransfected control unless indicated otherwise with bars. p values for all groups have been added

   1. I-J: The mem-mCherry shows the protrusions but not the cytoneme because these did not show wnt3 labeling.

   Cytonemes are defined as cellular protrusions transporting signalling components (Kornberg et a., 2014; Zhang and Scholpp, 2019). Therefore, cytonemes, as well as by the presence of Wnt3, can be defined by the presence of other Wnt signalling components, such as Ror2 (Mattes et al., 2018). Therefore, we term the protrusions in this figure cytonemes due to the presence of Ror2.

   1. Figure 6: The experimental designs are problematic. 1) Is Flot2 expressed in zebrafish embryos at the stage analyzed? The results in panels A-B using the overexpression approach do not reflect the endogenous expression of Flot2

   A previous paper (Philipsborn et al., 2005) looking at flotillin expression in zebrafish embryos shows that Flotillins are expressed early in development (observed as early as 1 hpf), which continues throughout gastrulation. We have added the following reference to the text to highlight this: "we addressed Flot2 function during zebrafish development, where Flotillins are highly expressed in early developmental stages and can be visualised on the tips of cellular protrusions (von Philipsborn et al., 2005)".

   1. Overexpression of Flot2-GFP could cause unintentional consequences.

   We agree with the reviewer that Flot2 OE could have consequences. To test for these consequences, the microinjection of membrane-mCherry was used as a control to confirm that the observed phenotypes are specific to Flot2 function rather than a side effect on injection. Furthermore, we have added a new data set in which we generated F0 Crispants for Flot1b/2a. We could show that these cells display shorter and fewer Wnt8a cytonemes.

   1. Also, where were those cells that were imaged ?

   We have added a schematic drawing to indicate where the cells were imaged within the embryo (Fig. 6a).

   1. Could the authors show more cells? Images of separated channels should be shown. The cell in B seems to be round. Was the cell at the mitosis stage?

   We show more images from cells with the typical morphology of a zebrafish epiblast cell.

   1. The authors injected various DNAs to show the consequence of the expression. This method is very unreliable, as injection of DNA likely leads to mosaic expression of the proteins at different expression levels thus, the expression levels are very hard to be controlled. Has the expression of various constructs been compared in different conditions? RNA injection experiments are recommended, as these usually lead to uniform and reliable protein expression

   We have used mRNA expression as well as DNA expression in our previous works. We found that DNA expression has several advantages for analysing small cell clones. After DNA injection, only a few cells express the construct at sufficient levels allowing good imaging. Therefore, we used a Flot2-GFP pCS2+ construct to generate a sparse, mosaic expression. In these experiments, we find a reliable, stable expression of Flot2-GFP in a subset of cells, which is very important for imaging.

   1. 1. Did overexpression of Flot2 or Wnt8 cause severe developmental defects? Were those embryos healthy? Could the authors show live images of group embryos? The authors need to explain the "0" values in some columns (+wnt8a, flot2/wnt8s) in G. Did these results indicate those embryos did not express pax6a at all?

   Indeed, overexpression of Flot2 and Wnt8a led to an arrest in development in development with obvious malformation of the body axis and frequent oedema formation. Therefore, we focussed our analysis on gastrulation and stopped at 24hpf to visualize a potential effect on AP patterning. "0" values represent embryos where the forebrain is entirely missing (Pax6a expression could only be seen in the hindbrain). This has previously been observed (Brunt et al., 2022). We have added the following sentence to clarify this: " In some cases, the FB was completely lacking (no Pax6a expression), which was recorded as "0"."

   Reviewer #2 (Public Review):

   1. The case for Flot2 being a modulator of Wnt cytonemes is made, but its characterization as a specific regulator of Wnt cytonemes is over-stated given the data provided. To make the case for Wnt specificity, the authors need to show that Flot2 modulation does not impact signaling filopodia housing other signaling molecules.

   We agree with this reviewer and have changed the term "regulator" to "modulator". Furthermore, we have added a short paragraph to the discussion highlighting that Flot2 can also promote Hh cytonemes, and thus Flot2 could modulate cytonemes in general. We further changed the text accordingly to describe "specificity" as having a specific role related to morphogen receptors.

   1. For some experiments, cell viability appears to be a complicating factor. The cells in which IRSp53 function is targeted look very unhealthy , so it is not clear reliable results can be obtained using the experimental parameters described.

   We agree with the reviewer. Therefore, we have revisited this data set. We find that IRSp534K-GFP expressing cells were healthy in general. Consequently, we have replaced the previous image with a cell with a morphology seen in most cells (Suppl. Fig. 1).

   Reviewer #3 (Public Review):

   1. Whether Flot2 manipulation specifically affects Wnts on cytonemes, or it could have a more general effect should also be considered

   We agree with this reviewer and have addressed this point in our response to reviewer 2. For example, we have changed the term "regulator" to "modulator" in the manuscript. Furthermore, we have added a short paragraph to the discussion highlighting that Flot2 can also promote Hh cytonemes, and thus Flot2 could modulate cytonemes in general. We further changed the text to describe "specificity" as having a specific role related to morphogen receptors.

   1. Statistical analysis should be done more consistently. It is either missing for some samples or the comparison between samples is not given

   p values have been added to all groups and some additional comparisons have been added

   1. The part of the manuscript related to the analysis of Ror2 and Flot2 in cytoneme formation and PCP pathway could be better connected with the central theme of the work/title, which is mainly on canonical signalling by Wnt3. Perhaps, directly analyzing the effect of Ror2 manipulation on Wnt3 levels on the cytoneme could be useful.

   Our experimental data suggest that the Ror2 function is required for Wnt3 cytoneme formation. We further show that Flot2 is needed to transport Ror2 to the plasma membrane to promote cytoneme formation. Therefore, we believe that we have addressed the sequence of events in this manuscript by showing the requirement for Flot2 in cytoneme formation and providing a possible molecular mechanism.

   Read the original source
3. 

   eLife

   Jul 12, 2022

   Evaluation Summary:

   The manuscript by Routledge et al. describes the role of Reggie-1/Flottilin2 in the formation of filopodia-like membrane protrusions called cytonemes and which were shown to be conserved between gastric cancer cells and Zebrafish. Authors demonstrate that Flot2 is present on the cytoneme along with Wnt3 in gastric cancer and with Wnt8a in Zebrafish. Furthermore, Flot2 is also present with Ror2 on the cytoneme and together they are believed to modulate cytoneme formation. This study extended the previous studies and provides new details about regulatory events controlling a cell biological process that will be of interest to those in the Wnt and cytoneme fields.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested …

   Evaluation Summary:

   The manuscript by Routledge et al. describes the role of Reggie-1/Flottilin2 in the formation of filopodia-like membrane protrusions called cytonemes and which were shown to be conserved between gastric cancer cells and Zebrafish. Authors demonstrate that Flot2 is present on the cytoneme along with Wnt3 in gastric cancer and with Wnt8a in Zebrafish. Furthermore, Flot2 is also present with Ror2 on the cytoneme and together they are believed to modulate cytoneme formation. This study extended the previous studies and provides new details about regulatory events controlling a cell biological process that will be of interest to those in the Wnt and cytoneme fields.

   (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. The reviewers remained anonymous to the authors.)

   Read the original source
4. 

   eLife

   Jul 12, 2022

   Reviewer #1 (Public Review):

   Studies published by the authors and others have shown that many cells extend prolonged filipodia to transport Wnts to exert their functions in a distance. In this current study, the authors identified Flot2, a scaffolding protein, which interacts with Ror2 to regulate the formation of signaling filipodia to transport Wnt3 in GC cells and Wnt8a in zebrafish. However, the role of Flot2 in promoting cytoneme formation to spread morphogens has been reported in other animal models and cancer cells; and the data that support the role of Flot2 in Ror2 are not convincing. Additionally, the paper employed similar published approaches. Thus, the significance and novelty of this work are not very high.

   The quality of many data and some experimental should be improved. Specifically, most experiments used the …

   Reviewer #1 (Public Review):

   Studies published by the authors and others have shown that many cells extend prolonged filipodia to transport Wnts to exert their functions in a distance. In this current study, the authors identified Flot2, a scaffolding protein, which interacts with Ror2 to regulate the formation of signaling filipodia to transport Wnt3 in GC cells and Wnt8a in zebrafish. However, the role of Flot2 in promoting cytoneme formation to spread morphogens has been reported in other animal models and cancer cells; and the data that support the role of Flot2 in Ror2 are not convincing. Additionally, the paper employed similar published approaches. Thus, the significance and novelty of this work are not very high.

   The quality of many data and some experimental should be improved. Specifically, most experiments used the overexpression approach. Genetic approaches would need to be employed, particularly in embryos. The dominant-negative Flot2 is the key tool utilized in the paper, but it is unclear whether this construct has been characterized in the system used and how it affects endogenous protein function. Has its impact on the endogenous Flot2 been examined? Similarly, the effectiveness and specificity of siRNA for example, the expression level of Flot2 would need to be assessed in all experiments. Furthermore, it is unclear whether the tagged constructs (eg, Flot2-GFP, Wnt8a-mcherry) have been characterized and whether the tags affect the protein function.

   Other general points:
   1. Most images show one single cell. Could more cells be presented? The nature of the images should be disclosed. For example, are those confocal images (single plane or Z-stack)?
   2. The P values for which group are not clear in many panels. It is not clear which groups were analyzed. For example, Fig. 4C, D, H and many other panels.
   3. Statistical analyses are lacking for some panels. For example, χ2 test is needed for many panels, including Fig. 3E, and many others.

   Comments on figures:

   1. Figure 1: 1) AGS is supposed to compare with control (HFE-145). These data are missing in the chart. The cell number in AGS is significantly higher than that in other cells (25 vs, 7 and 8, line 666), which can compromise the statistical analysis. 2) Qualification data are needed to support the statement in Line 71-72. 3) Fig1D: Wnt3-positive filopodia in AGS is double compared to that in HFE-145, which is not consistent with the image shown in Fig1C. 4) Fig1H-I: The red channel is overexposed. The authors should explain why a-myox and a-evi signals were detected outside the cell (or just the background)? The more appealing evidence should be the co-localization of Myox or Evi with wnt3a on the filopodia.

   2. Figure 2 nicely showed the impact of paracrine Wnt signaling induced by producing cells. However, there are many issues with this experiment. 1) The reporter plasmids are transiently transfected, which inevitably leads to the expression at different expression levels. How could the authors compare the expression levels as a readout in different conditions if this is the case? Better and reliable methods should use stable cell lines. Thus, the authors should make a stable 7xTCF-NLS-mCherry stable line or co-transfect the cell with GFP to show the relative transfection level. This concern also applied to other figures using 7xTCF-NLS-mCherry reporter assay. 2) Thus, the mCherry positive cells in Fig2B, D and F cannot present all receiving cells, as the transfection rate should not be 100%. Also, did all experiments start with a similar cell number? Thus, the Chart in D is not accurate, and the reason that assesses the number of receiving cells is not clear. It is not clear what "per image" means in D? Is the number correct (1 cell vs 1.5 cells) in D? Additionally, is it possible to image Wnt3 is being transported to the receiving cells? 3) Fig2 C: Western blot could be added to show the mCherry expression level in each group. 4) It is better to include the red channel only in E. It is difficult to see the red signal in the current images. 5) F: How was the qualification conducted? Could the whole population be analyzed more quantitatively?

   3. Figure 4: 1) The critical data should be that the formation of wnt3a cytoneme (length, number) is impaired in Flot2-deficient cells, which are missing in the figure and the manuscript. 2) A-D: The expression of Flat2 should be presented in separated images. The membrane localization is not clear. Fig4D shows flot2 occasionally localized with Wnt3. Time-lapse experiments will provide additional evidence of the constant localization of Flot and Wnt3. 3) E: This panel has similar issues described in Figure 2. How was the transfection rate in E? Did all cells express Flot2 or dnFlot2? Their expression should be examined at the same time.

   4. Figure 5 is one of the key figures. However, the quality of the images is not high enough to support the conclusion. A-D: The membrane co-localization is not convincing. Better images with a membrane marker are needed. Also, it is better to present images in separate channels. The red color A should be magenta. Did the dominated-negative Flot2 affect the expression of endogenous Flot2? Similarly, the expression of endogenous Flot2 in siRNA expressing cells should be shown. D: Instead of showing the image of single-cell, additional experiments, for example, the western blot should provide additional evidence to show Ror2 expression on the membrane is lost. E: High magnification images should be presented to show the localization. The current images are too small to appreciate the co-localization. Similarly, separated channels should be presented. How many experiments have been conducted? It seems that the cell number is not high. For siRNA experiments, was the expression of Flot2 validated? It is necessary to describe how E-Co-efficient (PCC) was determined in more detail. F: The label for the X-axis is missing. G: The nuclei and cell boundaries are not clear; the markers for these should be included to give confidence where and how the quantification was conducted. Similarly, the expression of Flot2 should be examined in these experiments as it is likely not all cells express those constructs at similar levels. Additional experiments, for example, Western bolt to show pJNK levels, are necessary to support the conclusion further. H: The P values for which group are not clear. I-J: The mem-mCherry shows the protrusions but not the cytoneme because these did not show wnt3 labeling.

   5. Figure 6: The experimental designs are problematic. 1) Is Flot2 expressed in zebrafish embryos at the stage analyzed? The results in panels A-B using the overexpression approach do not reflect the endogenous expression of Flot2. Overexpression of Flot2-GFP could cause unintentional consequences. Also, where were those cells that were imaged? Could the authors show more cells? Images of separated channels should be shown. The cell in B seems to be round. Was the cell at the mitosis stage? 2) B: The authors nicely showed that Wnt8a-mCherry is clustered on Flot2-GFP-expressing filipodia. Because of the nature of overexpression of Wnt8a-mCherry, it is possible that Wnt8a-mCherry and Flot2-GFP were expressed in the same spots. Could the authors perform time-lapse experiments to show Wnt8a-mCherry is being delivered to the neighboring cells by Flot2-GFP-expressing filipodia? 3) The authors injected various DNAs to show the consequence of the expression. This method is very unreliable, as injection of DNA likely leads to mosaic expression of the proteins at different expression levels thus, the expression levels are very hard to be controlled. Has the expression of various constructs been compared in different conditions? RNA injection experiments are recommended, as these usually lead to uniform and reliable protein expression. 4) Did overexpression of Flot2 or Wnt8 cause severe developmental defects? Were those embryos healthy? Could the authors show live images of group embryos? The authors need to explain the "0" values in some columns (+wnt8a, flot2/wnt8s) in G. Did these results indicate those embryos did not express pax6a at all?

   Read the original source
5. 

   eLife

   Jul 12, 2022

   Reviewer #2 (Public Review):

   The manuscript presents a study focused on the Flotillin-2 (Flot2) membrane-associated scaffolding protein and its effects on Wnt delivery through signaling filopodia called cytonemes. Strengths of the manuscript include the importance of the problem being investigated, the combination of in vitro cell biological analyses with in vivo Wnt phenotypic analyses, and the identification of Flot2 as a conserved regulatory partner contributing to Wnt cytoneme activity. Mechanistic studies executed in control and gastric cancer cell lines suggest Flot2, which is overexpressed in some gastric cancers, may promote cancer cell cytoneme signaling. In vivo experiments in which Wnt phenotypes are evident in zebrafish with altered Flot2 activity are convincing and support a conserved role for Flot2 in Wnt signaling activity.

   Reviewer #2 (Public Review):

   The manuscript presents a study focused on the Flotillin-2 (Flot2) membrane-associated scaffolding protein and its effects on Wnt delivery through signaling filopodia called cytonemes. Strengths of the manuscript include the importance of the problem being investigated, the combination of in vitro cell biological analyses with in vivo Wnt phenotypic analyses, and the identification of Flot2 as a conserved regulatory partner contributing to Wnt cytoneme activity. Mechanistic studies executed in control and gastric cancer cell lines suggest Flot2, which is overexpressed in some gastric cancers, may promote cancer cell cytoneme signaling. In vivo experiments in which Wnt phenotypes are evident in zebrafish with altered Flot2 activity are convincing and support a conserved role for Flot2 in Wnt signaling activity.

   Key conclusions are generally supported by the data provided but many individual results are overinterpreted or overstated, making the proposed mechanistic model premature. This should be corrected by the inclusion of additional controls and experimental parameters. More information is needed to clarify how results were quantified throughout the study because oftentimes, results read as though all filopodial extensions, not exclusively cytonemes, were surveyed. Some of the cellular extensions visible in cell images provided look like retraction fibers. If they are included in quantification, results may be skewed. If all extensions shown are being counted as cytonemes, supporting evidence that they are behaving as signaling filopodia needs to be provided.

   The case for Flot2 being a modulator of Wnt cytonemes is made, but its characterization as a specific regulator of Wnt cytonemes is over-stated given the data provided. To make the case for Wnt specificity, the authors need to show that Flot2 modulation does not impact signaling filopodia housing other signaling molecules.

   For some experiments, cell viability appears to be a complicating factor. The cells in which IRSp53 function is targeted look very unhealthy, so it is not clear reliable results can be obtained using the experimental parameters described.

   Read the original source
6. 

   eLife

   Jul 12, 2022

   Reviewer #3 (Public Review):

   Strength/novelty: The manuscript is overall well written and the authors have convincingly supported their main conclusion that Wnt3, which was previously shown to be upregulated in several GCs, is present on cytonemes produced by these cells. While Wnt proteins are known to utilise diverse modes and extracellular carriers for their secretion, this work indicates that cytonemes might be the preferential mode for the dispersion of Wnt3 by GC cells. This observation has important implications on our understanding of Wnt3-mediated activation of canonical signaling. Furthermore, they show that Flot2 levels are also higher in the GC, which further enhances Wnt3 levels on the cytonemes and consequently proliferation of the GC cells.

   Scope for improvement: 1) Whether Flot2 manipulation specifically affects Wnts on …

   Reviewer #3 (Public Review):

   Strength/novelty: The manuscript is overall well written and the authors have convincingly supported their main conclusion that Wnt3, which was previously shown to be upregulated in several GCs, is present on cytonemes produced by these cells. While Wnt proteins are known to utilise diverse modes and extracellular carriers for their secretion, this work indicates that cytonemes might be the preferential mode for the dispersion of Wnt3 by GC cells. This observation has important implications on our understanding of Wnt3-mediated activation of canonical signaling. Furthermore, they show that Flot2 levels are also higher in the GC, which further enhances Wnt3 levels on the cytonemes and consequently proliferation of the GC cells.

   Scope for improvement: 1) Whether Flot2 manipulation specifically affects Wnts on cytonemes, or it could have a more general effect should also be considered. 2) Statistical analysis should be done more consistently. It is either missing for some samples or the comparison between samples is not given. 3) The part of the manuscript related to the analysis of Ror2 and Flot2 in cytoneme formation and PCP pathway could be better connected with the main theme of the work/title, which is mainly on canonical signaling by Wnt3. Perhaps, directly analyzing the effect of Ror2 manipulation on Wnt3 levels on the cytoneme could be useful.

   Read the original source
7. 

   Version published to 10.1101/2022.01.07.475396v1 on bioRxiv

   Jan 8, 2022