A protective role for Drosophila Filamin in nephrocytes via Yorkie mediated hypertrophy

Nephrocytes lacking key components of the filtration barrier respond by increasing Drosophila Filamin expression, which accumulates at the cell cortex to elicit a protective role.

The manuscript by Koehler and Denholm uses Drosophila nephrocytes to test the hypothesis that the Integrin-actin crosslinking protein Cheerio/Filamin serves a protective role during injury. Using genetic variants affecting Cheerio's ability to interact with Integrin, the authors find that activated Cheerio leads to cellular hypertrophy, which they go on to suggest is mediated by TOR and Hippo signaling. In the context of nephrocyte injury (involving disruption of slit diaphragms and filtration function), Cheerio can partially rescue nephrocyte filtration and a related delay in pupation. They also show some ability of human FilaminB to improve certain aspects of the nephrocyte injury model. The experiments are well-conducted, including quantification of multiple phenotypes across multiple genotypes. The study presents several novel findings that will be of interest to the field, and is therefore appropriate for publication in Life Science Alliance, however some of the conclusions seem overstated based on the current data. As discussed below, either additional experimental support should be provided, or the writing should be modified to reflect that some of the data are simply consistent with the author's interpretation/model, but are not themselves conclusive evidence.

Major comments
Cheerio localization: In Figure 1 and associated text, the authors conclude that Cheerio partially colocalizes with the slit diaphragm protein Duf, which they suggest may indicate Cheerio is in complex with slit diaphragm-associated proteins. However, the images provided ( Figure  1A,B) are very low magnification, which makes determining its precise localization difficult. Can the localization of Cheerio be determined in high-resolution images similar to those in panel E? This would be particularly useful in examining the localization of the wildtype, inactive, and active forms of the protein. Similar high magnification images would be helpful for Human FilB, which they suggest also localizes to slit diaphragms (Supplemental Figure 4A). If these types of images cannot be provided, it seems necessary to significantly tone down the conclusion that these proteins are located at slit diaphragms. The methods section (Table 2) suggests the three Cheerio MSR variants from the Huelsmann 2016 paper are UAS, however the description in that paper indicates they are knock ins at the endogenous locus (attB-mediated replacement of the 3' end). If this is correct, it will change the wording in places. For example, the text indicates these constructs are being driven by Sns-Gal4, but instead they would be expressed from the endogenous cheerio promoter. It may also be more accurate to indicate that expression of the Cheerio-MSR Active form of the protein causes hypertrophy, as opposed to "elevated" levels of Cheerio-MSR Active protein.
There are notable differences in the localization patterns of Cheerio:GFP and the anti-Cheerio staining, particularly in panel A. Is there an explanation for this? They appear more similar (cortical) in panel B. I am also confused by the wildtype Cheerio:GFP localization patterns in panel B and C. I believe the Cheerio:GFP in B is the GFP trap line mentioned in the methods section (this was not clear in the figure legend). Is there an explanation for why the wildtype Cheerio:GFP in panel B appears cortical while the Cheerio:GFP with wildtype MSR (essentially the wildtype protein I believe) is cytoplasmic?
Cheerio exerts a protective effect in damaged nephrocytes: Figure 2. Panel D suggests that Cheerio-Active can rescue FITC:Albumin uptake caused by knockdown of slit diaphragm proteins. Knockdown of these proteins disrupts the pattern of slit diaphragms on the nephrocyte surface, and presumably reduces uptake of fluorescent tracers due to subsequent loss of labyrinthine channel area, where the majority of endocytic uptake of these tracers occurs. However, the current data ( Figure 2B) suggest that slit diaphragms are not restored by Cheerio-Active. Is Cheerio-Active directly affecting rates of endocytosis through some other mechanism? It would be helpful if the authors could provide more discussion of possible mechanisms underlying these findings. In their previous study (Koehler et al., JASN 2020), Cheerio knockdown had no apparent effect on otherwise normal nephrocytes. The model proposed here suggests that Cheerio has a protective effect in injured podocytes/nephrocytes. Thus, if Cheerio protects nephrocytes in response to injury, one would predict that knockdown of Cheerio in a damaged nephrocyte should significantly dampen that protective response, leading to stronger defects, be it cell morphology, size, or function. If the authors agree with this rationale, can they perform a double knockdown of Cheerio and Sns or Duf to assess whether this is correct?
TOR and Yki are downstream of Cheerio: Page 21, line 16. The subsection title of "TOR and Yorkie signaling induce nephrocyte hypertrophy and appear to be downstream targets of Cheerio" seems an overly strong conclusion because, as the authors note in the manuscript, the expression of TOR DN or HIPPO alone decrease cell size, thus the apparent rescue of Cher Active hypertrophy may simply be the averaged effects of two independent processes acting on cell size. It may support their conclusion if they could show increased activity for these pathways in Cher Active cells. There are available genetic reporters for Hippo activity (Parker and Struhl, PLOS Biology 2015). TOR activity has been assessed in flies using various readouts, such as phospho-4E-BP antibody staining (Teleman et al., Developmental Cell. 2005). Similarly, based on the current data, the proposed relationship between Cheerio and TOR/Yki in Figure 6 should be made less conclusive; perhaps by making the arrows dashed or adding question marks. The figure legend text should also reflect that these relationships are somewhat tentative.
Minor comments Figure 1 Is titled: "Elevated Cheerio-MSR levels result in nephrocyte hypertrophy", however, because the hypertrophy only results from the active form, perhaps more specific language could be used, such as "Elevated levels of Cheerio-MSR-Active result in nephrocyte hypertrophy". Figure 2 The order of genotypes in the images in panel B is different from the x-axis of graphs in panels C/D. Similarly, the order of genotypes within panels C/D is different from the x-axis in F. Since these include all of the same genotypes, it would be helpful to the reader if the order of genotypes could be kept consistent. The order of genotypes in these graphs is also different from that in Supplemental Figure 2. This makes it confusing when trying to compare the effects of active vs inactive Cheerio in the different genotypes. It is not possible to distinguish the genotypes by symbols in Figure2 panel D. Can the symbols and lines also be color coded or the symbols enlarged or otherwise made more obvious? The figure legend for panels E,F state that Cheerio Active rescues delayed pupation of Duf or Sns RNAi, but the data only indicate a delayed pupation for Duf (and associated rescue). Thus it seems reference to a rescue of Sns RNAi should be omitted here. Figure 3 In Panel B, it appears that expressing hFilB in Duf kd cells is somehow restoring Duf expression. Is this correct? If so, can the authors provide any explanation? In Panel D, there is a "****" but no line indicating which genotypes are statistically different. The figure legend suggests none of them are different, so the "****" probably just needs to be removed. Figure 4 Can the data in panels F and G be merged together with D and E, respectively, to make comparisons between all genotypes more straightforward?
Page 19, line 8 indicates that the Cheerio constructs do not possess the ACB, but as discussed above, I do not believe this is accurate. The mutant Cheerio knock-ins described in the Huelsmann 2016 paper suggest the MSR variants should be part of the full-length protein (long isoform), which should therefore contain the ACB. I believe this also affects the interpretation of the data as presented on page 29, line 10, where the inference is made that hFilB rescues cells size due to possessing the ACB, thus the ACB is important for cell size, however if it is correct that the Cheerio variants are indeed full-length proteins possessing the ACB, this conclusion does not seem supported.
Yki is the common abbreviation for Yorkie, not "Yrk".
Page 24, the text does not make it clear that the phenotypes result from homozygosity for the Cheerio MSR Active or Inactive variants, not wildtype Cheerio. Table 2 lists UAS-hFilB constructs delta ACB and delta MSR, but I do not believe there are any data or discussion of these lines in the manuscript. Probably remove these from table, along with associated primer information.
Reviewer #2 (Comments to the Authors (Required)): The paper describes the effects of modulating the fly ortholog of human filamins. This is done in fly nephrocytes, which are used to study aspects of human kidney podocytes and proximal tubule cells. The paper focuses on nephrocytes as a model of podocytes but uses endocytic function as a marker, a model of proximal tubule cell function. Nephrocyte are both a model of both human podocyte and proximal tubules cell and this should really be mentioned. Also, the localisation of filamin B in human proximal tubules cells might suggest the drosophila model is better suited to establishing filamin's role in tubules cell endocytosis -given that its at the cell surface. Its not so simple to say that Drosophila filamin (Cher) models only filamin biology in human podocytes. Apologies in advance if I've misunderstood or missed something I should have seen or considered.
1. The paper deals with Filamin's 'mechano-protective role' in nephrocytes however there needs to be experimental data presented in the paper to support this mechanotransduction role. At present there is no experimental data relating mechantransduction, the mechano-protective role is inferred from other studies and not demonstrated directly. 2. Neither the introduction nor the discussion mentions a considerably important paper about YAP/TAZ singling in mammalian podocytes and Drosophila nephrocytes by Hurcombe et al Nature Communications 2019. That paper's findings need to be mentioned and the current data examined within the context of Hurcombe's work. 3. Nephrocyte numbers should be stated for all experimental genotypes. If modulations are protective then nephrocyte numbers should be similar -this needs to be shown. 4. Nephrocyte area need to be stated in square microns, not as relative values. 5. Figure 1 and other images. The images of nephrocytes are very small, most of the image is extraneous area and the images do little to support the quantified data. From the images, the nephrocytes do not appear to be larger in the Cher active experimental group. 6. There is a large discrepancy between the albumin binding as quantified in figure 3 compared to albumin binding for similar genotypes in Figure 2 (70-80% reduced in figure 3 vs 30% reduced in Figure 2 for both the sns KD and duf KD flies). 7. Figure 3B the images should be presented in the same order as the data in the graphs (Fig 3C and D). 8. Figure 3D the asterisks needs an accompanying line. 9. Figure 3. Cells really need to be counterstained with some general cell marker. They cannot be seen in the Duf KD and other images. 10. Figure 3's figure legend: this is a little confusing me: the title states 'Human Filamin B rescues nephrocyte size and function', then for 3B it states 'Expression of human Filamin B wildtype (human Fil B WT) in Duf and Sns depleted nephrocytes did not restore morphology', then 3C states, 'Nephrocyte size was restored by expressing human Filamin B wildtype, then 3D states 'FITC Albumin uptake is not rescued by expression of the human Filamin B wildtype. If I've understood this, it looks like Human Filamin B rescues the AgNO3 toxicity independently of rescuing albumin binding / endocytosis; which is odd. Yet function is stated as being restored -which is only partly the case. 11. Supp Fig 4 -colocalization data is required to support the claim that Human filamin B localises to the slit diaphragm. Currently it shows the HA tag and Pyd are at the cell surface but that's far from showing colocalization at the slit diaphragm. If the same logic was applied more generally, all cell surface proteins could be claimed to localise to the slit diaphragm, and that is not the case. 12. Figure 5. The control appears to comprise a line with two chromosome 3 balancers (MKRS/Tm6B). That's a lot of mutations that cannot be regarded as representative of wild type and therefore another control really needs to be used. 13. Discussion: Filamin B is expressed at the apical surface of human proximal tubule cells -cells modelled in the nephrocyte via the endocytosis assays. Proximal tubule cells are never mentioned in the paper and that needs addressed, the nephrocyte is not just a model of podocytes. It is unclear who these findings from the fly model relate to podocytes alone. 14. There is persistent reference to 'cell injury' yet there is no insult to the cells that constitutes an injury as such; injury is due to gene over-expression that modulate components of the slit diaphragm. An injury model would align more with transient chemical or hydrostatic pressure provocations. Perhaps avoid using the term injury. 15. Mechanical force is mentioned in the discussion yet there is no methodological approach that addressed forces being exerted on nephrocytes nor the effect this had on nephrocyte function or phenotype. Without such a provocation, it is hard to state that the modifications to nephrocytes were linked to mechanotransduction. 16. Figures 4 and 5 have no images of nephrocytes to support the quantified data. Can these be presented?

Major comments
Cheerio localization: In Figure 1 and associated text, the authors conclude that Cheerio partially colocalizes with the slit diaphragm protein Duf, which they suggest may indicate Cheerio is in complex with slit diaphragm-associated proteins. However, the images provided ( Figure 1A,B) are very low magnification, which makes determining its precise localization difficult. Can the localization of Cheerio be determined in highresolution images similar to those in panel E? This would be particularly useful in examining the localization of the wildtype, inactive, and active forms of the protein. Similar high magnification images would be helpful for Human FilB, which they suggest also localizes to slit diaphragms (Supplemental Figure 4A). If these types of images cannot be provided, it seems necessary to significantly tone down the conclusion that these proteins are located at slit diaphragms.
We added higher resolution microscopy images for the endogenously tagged Cheerio, the different Cheerio variants and the human Filamin B expressing nephrocytes.
High resolution imaging using confocal microscopy combined with Airyscan revealed co-localisation of Duf with endogenously tagged Cheerio (Cher:GFP) in garland cells isolated from 3 rd instar larvae. The newly generated images are now included in Figure   1B of the manuscript.
Our data reveals different localisation patterns for the three variants. While Wildtype and inactive Cheerio can be found largely in the cytoplasm, active Cheerio showed an 1st Authors' Response to Reviewers May 25, 2022 alternating pattern with Duf confirming our hypothesis that Cheerio in particular the active version localizes to the slit diaphragm complex. The data is integrated in Supp.
We also performed high resolution imaging with Airyscan and confocal microscopy for nephrocytes expressing the human Filamin B wildtype construct. The human Filamin B has a C-terminal HA tag, which was used for co-localisation studies with Duf. We did not observe a co-localisation of human Filamin B with Duf, which is in contrast to Cheerio (see above). However, the HA pattern resembles the nephrocyte diaphragm pattern we observe for Duf and Pyd, hence human Filamin B does localize to the nephrocyte diaphragm. Whether there is a direct interaction with another nephrocyte diaphragm protein remains unknown. We have amended the manuscript to reflect these new data.
Supp. Figure 4: Human Filamin B localizes to the nephrocyte diaphragm and results in only a mild nephrocyte phenotype. D High resolution imaging revealed human Filamin B wildtype localizes close to the nephrocyte diapraghm, which is visualized by anti-Duf staining (cyan). human Filamin B wildtype has a C-terminal HA tag. HA: magenta Scale bar = 5 µm.
The methods section (Table 2) suggests the three Cheerio MSR variants from the Huelsmann 2016 paper are UAS, however the description in that paper indicates they are knock ins at the endogenous locus (attB-mediated replacement of the 3' end). If this is correct, it will change the wording in places. For example, the text indicates these constructs are being driven by Sns-Gal4, but instead they would be expressed from the endogenous cheerio promoter. It may also be more accurate to indicate that expression of the Cheerio-MSR Active form of the protein causes hypertrophy, as opposed to "elevated" levels of Cheerio-MSR Active protein.
We do apologize for the confusion. The fly strains used in this manuscript are indeed UAS-driven and were kindly provided by Jari Ylänne. They are not mentioned in the Huelsmann et al, 2016 manuscript, but the mutations described in this manuscript are the same. We changed this in the manuscript.
There are notable differences in the localization patterns of Cheerio:GFP and the anti-Cheerio staining, particularly in panel A. Is there an explanation for this? They appear more similar (cortical) in panel B.
We agree with the reviewer, that there are notable differences in localisation of Cheerio, in particular in embryonic nephrocytes. A few explanations could be an unspecific binding of the Cheerio antibody, which causes an additional staining pattern, when compared to Cheerio:GFP. We used an anti-GFP antibody to enhance the Cheerio:GFP signal and the antibody efficiency of the anti-Cheerio and the anti-GFP antibody might be different. The localisation pattern is much more similar in larval garland cells, which are used for all experiments in the manuscript. We apologize for the confusion and changed the figure legend accordingly. One potential explanation for the different localisation pattern might be the overexpression of Cheerio in panel C. We used UAS-driven Cheerio variants, which resulted in higher expression levels, than the endogenous Cheerio. For the GFP-trap line we expect endogenous levels. Hence, the increased protein levels seem to result in an additional cytoplasmic localisation. Moreover, the UAS-driven Cheerio WT is lacking the actinbinding domain, which might also partially impact on the localisation of Cheerio. This potential impact seems to be of minor importance, as the active variant, which is also lacking the actin-binding domain appears to be cortical. We added these observations into the results section.
Cheerio exerts a protective effect in damaged nephrocytes: Figure 2. Panel D suggests that Cheerio-Active can rescue FITC:Albumin uptake caused by knockdown of slit diaphragm proteins. Knockdown of these proteins disrupts the pattern of slit diaphragms on the nephrocyte surface, and presumably reduces uptake of fluorescent tracers due to subsequent loss of labyrinthine channel area, where the majority of endocytic uptake of these tracers occurs. However, the current data ( Figure 2B) suggest that slit diaphragms are not restored by Cheerio-Active. Is Cheerio-Active directly affecting rates of endocytosis through some other mechanism? It would be helpful if the authors could provide more discussion of possible mechanisms underlying these findings.
To address, whether Cheerio, in particular the active variant, influences endocytosis, thereby causing a functional rescue of the Duf and Sns knockdown phenotype, we performed additional experiments and literature research.
We agree, that the FITC-Albumin uptake assay cannot be used to distinguish between filtration at the nephrocyte diaphragm and uptake/endocytosis in the lacunae. Hence, we do not know the exact mechanism, why and how the rescue fly strains (Cheerio Active and Duf/Sns RNAi) perform better than single knockdown flies. We tried to address this question in more detail. Our data describing the effects of overexpression of the Cheerio variants ( Figure 1G) does not show an increase of FITC-Albumin uptake in either of the genotypes, suggesting no direct effect on endocytosis.
Moreover, changes of the slit diaphragm and the lacunae structure might be very subtle and might require higher resolution microscopy such as electron microscopy. Cheerio exhibits an actin-binding domain and is an actin cross-linker. Therefore, the overexpression of active Cheerio might stabilize the foot processes and lacunae and result in a better result in the FITC-Albumin uptake assay.
To investigate the link between endocytosis and Cheerio we also performed Rab7 Supp. Figure 3: FITC-Albumin uptake assays and Rab7 immunofluorescence staining of Cheerio rescue strains. C Overexpression of either inactive or active Cheerio in nephrocytes did not cause any changes in Rab7 mean intensity. Duf: cyan; Rab7: magenta; GFP: green. Scale bar = 25 µm. D Immunofluorescence stainings in control and Cheerio kd nephrocytes did not reveal changes in Rab7 mean intensity. HRP: green; Rab7: magenta. Scale bar = 25 µm.
We also performed Rab7 immunofluorescence stainings in nephrocytes overexpressing either inactive or active Cheerio. Neither of the variants caused a significant increase of Rab7-positive vesicles. This data is now included in Supp. Cheerio. Hence, we concluded that our FITC-Albumin results primarily reflect filtration function rather than endocytosis.
In their previous study (Koehler et al., JASN 2020) We agree with this rational and thank the reviewer for this comment. We  RNAi results in an additional and significant size decrease. One-way ANOVA plus Tukey's multiple comparisons test: ***: p < 0.001 ****: p < 0.0001. C Filtration function was assessed by FITC-Albumin assays and reveals an increased FITC-Albumin intensity in Cheerio depleted nephrocytes (Cher kd), while loss of Duf and Sns results in a significant decrease of the FITC-Albumin intensity. The combination of either Cheerio and Duf RNAi or Cheerio and Sns RNAi causes a decreased FITC-Albumin intensity when compared to control cells, but also a significantly higher intensity when compared to the Duf and Sns single knockdown cells.  (Teleman et al., Developmental Cell. 2005). Similarly, based on the current data, the proposed relationship between Cheerio and TOR/Yki in Figure 6 should be made less conclusive; perhaps by making the arrows dashed or adding question marks. The figure legend text should also reflect that these relationships are somewhat tentative.
We agree that it is of greatest interest to investigate whether expression of active We hypothesised that if Yorkie and/or TOR are in the same pathway as Cheerio no additional size increase should be observed in nephrocytes. Indeed, co-expression of active Cheerio and overactive Yorkie did not cause an additional size increase ( Figure   4J). However, expressing active Cheerio and overactive TOR resulted in a significant size increase when compared to active Cheerio expressing cells ( Figure 4J). Taken together, our data therefore shows, that TOR does not seems to be activated by Cheerio and does not seem to be a downstream target. Yorkie translocates into the nucleus upon active Cheerio expression and seems to be a downstream target of Cheerio. We added the new data to the manuscript and also changed the graphical summary in Figure 6 accordingly. We added the new data and conclusion in the manuscript. These experiments were very informative and have helped us hone in on the Cheerio pathway; we thank the reviewer for these suggestions.

Figure 4: TOR and Hippo signalling mediate the hypertrophy phenotype in nephrocytes. H
Immunofluorescence stainings using an antibody against phosphorylated 4E-BP1 did not reveal any significant differences in mean intensity in nephrocytes expressing active or inactive Cheerio in comparison to control cells. I Immunofluorescence staining using an antibody against Yorkie revealed a significantly higher mean nuclear intensity in nephrocytes expressing active Cheerio when compared to control cells. One-way ANOVA plus Tukey's multiple comparisons test: ****: p < 0.0001. J Nephrocyte size was significantly increased when combining active Cheerio with over-active TOR, while the combination of active Cheerio with over-active Yorkie did not cause a significant size increase. Oneway ANOVA plus Tukey's multiple comparisons test: **: p < 0.01.

Figure 6: The mechano-protective role of Cheerio in Drosophila nephrocytes.
Cheerio is active in diseased nephrocytes and accumulates at the cell periphery. This activation and the increased protein levels result in a mechano-protective effect via hypertrophic growth, mediated via TOR and Hippo signalling. However, Cheerio levels need to be tightly controlled, as excessive increase of protein levels results in a shift from the protective to a pathogenic effect, including a morphological and functional phenotype.

Figure 1 Is titled: "Elevated Cheerio-MSR levels result in nephrocyte hypertrophy", however, because the hypertrophy only results from the active form, perhaps more specific language could be used, such as "Elevated levels of Cheerio-MSR-Active result in nephrocyte hypertrophy".
We changed the title of Figure 1 accordingly.

Figure 2 The order of genotypes in the images in panel B is different from the x-axis of graphs in panels C/D. Similarly, the order of genotypes within panels C/D is different from the x-axis in F. Since these include all of the same genotypes, it would be helpful to the reader if the order of genotypes could be kept consistent. The order of genotypes in these graphs is also different from that in Supplemental Figure 2. This makes it confusing when trying to compare the effects of active vs inactive Cheerio in the different genotypes. It is not possible to distinguish the genotypes by symbols in Figure2 panel D. Can the symbols and lines also be color coded or the symbols enlarged or otherwise made more obvious? The figure legend for panels E,F state that Cheerio Active rescues delayed pupation of Duf or Sns RNAi, but the data only indicate a delayed pupation for Duf (and associated rescue). Thus it seems reference to a rescue of Sns RNAi should be omitted here.
We agree and changed the order of genotypes and symbols as proposed.
We also changed the figure legend for panels E and F, as there is no significant rescue of the Sns phenotype.

Figure 3 In Panel B, it appears that expressing hFilB in Duf kd cells is somehow restoring Duf expression. Is this correct? If so, can the authors provide any explanation?
In the images shown it might look like there is more Duf in the FilB rescue nephrocytes.
However, we think this is mainly due to the strong background signal we also see in In Panel D, there is a "****" but no line indicating which genotypes are statistically different. The figure legend suggests none of them are different, so the "****" probably just needs to be removed.

Figure 4 Can the data in panels F and G be merged together with D and E,
respectively, to make comparisons between all genotypes more straightforward?
We could merge the files, but decided to not do this for the following reasons.
Experimental setups are different. Due to the GFP-tag of the Cheerio variants, the rescue with either TOR-DN or HIPPO were preformed using the FITC-Albumin in combination with the anti-Albumin antibody, while TOR-DN and HIPPO alone were done with the normal FITC-Albumin assay protocol. This resulted in additional incubation times and procedures for the rescue cells; hence the absolute values cannot be compared directly, but they could be compared after normalisation to the respective control. We merged the normalized data-sets and show them below. Performing statistics here is critical, as these are independent experiments with different protocols as mentioned above. If the reviewer and the editors think, merged data-sets would be better, we can of course provide them for the normalized data and change the figure accordingly. Figure: Active Cheerio mediated hypertrophy is reversed by blocking TOR and activating HIPPO signalling. A Cher Active causes a significant size increase when compared to inactive Cheerio, which can be reversed by blocking TOR signalling (TOR DN). However, blocking TOR alone also results in a significant size decrease. One-way ANOVA plus Tukey's multiple comparisons test: ***: p < 0.001, ****: p < 0.0001. B Activating Hippo resulted in a rescue of the hypertrophy phenotype observed in nephrocytes expressing active Cheerio. Activating Hippo alone also caused a significant size decrease. One-way ANOVA plus Tukey's multiple comparisons test: ****: p < 0.0001.

Page 19, line 8 indicates that the Cheerio constructs do not possess the ACB, but as discussed above, I do not believe this is accurate. The mutant Cheerio knock-ins described in the Huelsmann 2016 paper suggest the MSR variants should be part of the full-length protein (long isoform)
, which should therefore contain the ACB. I believe this also affects the interpretation of the data as presented on page 29, line 10, where the inference is made that hFilB rescues cells size due to possessing the ACB, thus the ACB is important for cell size, however if it is correct that the Cheerio variants are indeed full-length proteins possessing the ACB, this conclusion does not seem supported.
Again, we do apologize for make this mistake. The fly strains used are UAS-driven strains without the actin-binding domain. They are not described in the mentioned manuscript (Huelsmann et al., 2016). We changed the manuscript accordingly.
Yki is the common abbreviation for Yorkie, not "Yrk".
We changed that accordingly.

Page 24, the text does not make it clear that the phenotypes result from homozygosity for the Cheerio MSR Active or Inactive variants, not wildtype Cheerio.
We rephrased the text to make our approach clearer and to emphasize that we only generated homozygous flies for the active and inactive variant. The wildtype construct was not assessed. We are sorry for this mistake and deleted the information.

The paper describes the effects of modulating the fly ortholog of human filamins. This is done in fly nephrocytes, which are used to study aspects of human kidney podocytes and proximal tubule cells. The paper focuses on nephrocytes as a model of podocytes but uses endocytic function as a marker, a model of proximal tubule cell function. Nephrocyte are both a model of both human podocyte and proximal tubules cell and this should really be mentioned. Also, the localisation of filamin B in human proximal tubules cells might suggest the drosophila model is better suited to establishing filamin's role in tubules cell endocytosis -given that its at the cell surface. Its not so simple to say that Drosophila filamin (Cher) models only filamin biology in human podocytes. Apologies in advance if I've misunderstood or missed something I should have seen or considered.
We agree, that there are evidences for similarities between nephrocytes and proximal tubular cells. We included these findings and our thoughts in the discussion.

The paper deals with Filamin's 'mechano-protective role' in nephrocytes however there needs to be experimental data presented in the paper to support this mechanotransduction role. At present there is no experimental data relating mechantransduction, the mechano-protective role is inferred from other studies and not demonstrated directly.
We agree with the reviewer and performed additional experiments. We induced mechanical stress via hypo-osmotic stress and could show that nephrocytes respond with swelling and a significant size increase in control cells (Figure A below).
We hypothesized to see morphological changes as well, which might be rescued in nephrocytes expressing active Cheerio. However, we did not observe changes in morphology upon hypo-osmotic stress in control cells ( Figure B above), which is most likely due to the short time frame of the experimental setup; 5mins stress and immediate fixation afterwards. Also, changes might be very subtle and cannot be visualized by normal confocal microscopy as done here.
However, we decided to keep the term 'mechano-protective' as the Cheerio variants we used in our experiments express the mechanosensor domain and lack the actinbinding domain. Moreover, the active mechanosensor variant resulted in the best rescue and the putative protective hypertrophy phenotype. Based on our findings we argue that these protective effects are elicited by the activation of the mechanosensor region. Hence, we kept the term mechano-protective.

Neither the introduction nor the discussion mentions a considerably important paper about YAP/TAZ singling in mammalian podocytes and Drosophila nephrocytes by Hurcombe et al Nature Communications 2019. That paper's findings need to be mentioned and the current data examined within the context of Hurcombe's work.
We agree that this is a very interesting paper and examined our data in the context of  In their manuscript, Hurcombe et al. provide data showing that HIPPO signalling is responsible for the pathology observed upon loss of shaggy.
In detail, proteomic analysis revealed an upregulation of Ajuba and YAP/TAZ targets after depleting GSK3ß. Ajuba blocks phosphorylation of YAP/TAZ, which causes translocation of the two proteins into the nucleus, activation of downstream signalling and switching HIPPO off. Verteporfin was used to prevent translocation of YAP/TAZ into the nucleus in shaggy depleted and LiCL treated nephrocytes and had a beneficial effect on the cells. Taken together, the phenotype observed upon loss of shaggy is YAP/TAZ mediated, as switching HIPPO on is beneficial in this context.
Our data show that hyperactive Yorkie (YAP), which switches HIPPO off, results in a hypertrophy phenotype (Figure 4B), while blocking Yorkie by expressing HIPPO caused a nephrocyte size decrease ( Figure 4G). Moreover, the activation of HIPPO reversed the increased cell size in nephrocytes expressing active Cheerio ( Figure 4D).
In addition, we provide data showing that Yorkie translocates into the nucleus upon activation of Cheerio, suggesting Yorkie is a downstream target of Cheerio.
Taken together, both studies show that HIPPO signalling plays an important role in nephrocyte biology and is downstream of GSK3ß/shaggy and Cheerio. We summarized these findings and conclusion in the discussion.

Nephrocyte numbers should be stated for all experimental genotypes. If modulations are protective then nephrocyte numbers should be similar -this needs to be shown.
For our experiments and analysis, we did a cell-by-cell analysis. This means, that we measured nephrocyte size and fluorescent intensity for every single cell in this image.
The individual values are then used to calculate the mean for each fly, which means that one data point is representing one fly. We added this in the material and methods section. Based on our analysis approach, we think numbers do not need to be equal to be able to observe protective effects as we measure and quantify single cells and compare mean values between flies and genotypes.
However, we assessed the mean number of nephrocytes per fly after expressing the different Cheerio variants and did not observe any significant differences in nephrocyte number (Figure A below). We also assessed whether loss of Duf and Sns resulted in less nephrocytes and whether this can be rescued by expression of active Cheerio ( Figure B below). We did not observe any significant changes in any of the genotypes assessed. We apologize if we misunderstood the comment from the reviewer.

Nephrocyte area need to be stated in square microns, not as relative values.
We added all graphs showing raw data (microns) in Supp. Figure 7. We decided to keep the normalized relative values for easier comparison. However, we agree that it is of interest to show the absolute values. As experimental setups were different due to the GFP-tag of the Cheerio variants, it is difficult to compare raw data throughout.
In experiments which included the Cheerio variants, we performed the FITC-Albumin uptake assay and subsequent anti-Albumin immunofluorescence. This caused additional incubation times and solutions and has an impact on the FITC-Albumin signal.
Supp. Figure 7: Absolute values of nephrocyte size. A-N For two variables we used Student's t-test and considered a p-value < 0.05 as significant. For more than two variables we used One-way ANOVA plus Tukey's multiple comparisons test and also considered a p-value < 0.05 as significant.
5. Figure 1 and other images. The images of nephrocytes are very small, most of the image is extraneous area and the images do little to support the quantified data.
From the images, the nephrocytes do not appear to be larger in the Cher active experimental group.
We agree, that in some of the pictures we have a lot of extraneous area, but we decided to keep the magnification equal for all genotypes for comparison. Often one of the images cannot be cropped. We showed a different representative image for Cher Active. We also added magnifications or high-resolution imaging where it is possible to support our quantification data.
6. There is a large discrepancy between the albumin binding as quantified in figure 3 compared to albumin binding for similar genotypes in Figure 2 (70-80% reduced in figure 3 vs 30% reduced in Figure 2 for both the sns KD and duf KD flies).
This discrepancy is a result of different experimental setups. The Cheerio variants are fused to a GFP tag, hence the FITC-Albumin cannot be quantified after incubation, as the GFP signal is overlapping with the FITC signal. Therefore, we performed Albumin immunofluorescence stainings (with anti-Albumin) after incubation with FITC-Albumin.
The anti-Albumin antibody was then labelled with a Cy3 secondary antibody to enable quantification of Albumin within the cell. The Filamin constructs are fused to a HA-tag, therefore FITC-Albumin assays could be performed without additional antibody-based staining. The genotypes (Duf kd and Sns kd and respective controls) were done in the same way (Figure 2 plus additional antibody-based staining, Figure 3 without additional staining) to enable comparisons with the respective rescues. For normal FITC-Albumin uptake assays isolated nephrocytes were incubated in FITC-Albumin for 1 min, followed by a 1 min washing step and subsequent fixation for 20 mins and mounting.
To perform antibody based stainings, nephrocytes were also Incubated with FITC-Albumin for 1 min, followed by a washing step for 1 min and fixed for 20 mins with formaldehyde. Instead of mounting cells were then incubated for 1 h with methanol, followed by 3 washing steps for 10 mins and incubation with the albumin antibody overnight. The next day, cells were again washed for three times, incubated with the Cy3 secondary antibody for 45 mins, and washed again before they were mounted.
Due to this long and intense staining protocol, the Albumin signal will be lower, if compared to cells immediately fixed after incubation with FITC-Albumin. Moreover, due to weak antibody binding of the Albumin antibody, signal intensity was much lower. We think that these different protocols are causative for the discrepancy observed. As all genotypes, which we compare, were done with the same protocol, this should not be a problem.
7. Figure 3B the images should be presented in the same order as the data in the graphs (Fig 3C and D).
We changed that accordingly.
8. Figure 3D the asterisks needs an accompanying line.
We deleted the asterisks, as there was no difference between the rescue and single knockdown cells. The significance described the effect resulting from comparing the control (h Fil B WT) with the rescues and single knockdowns. For simplification, we removed the asterisks.
9. Figure 3. Cells really need to be counterstained with some general cell marker. They cannot be seen in the Duf KD and other images.
To make the cells more visible we added circles to mark the outlines of the cells. We changed the figure title and legend to make our observations and the resulting conclusion clearer. Fig 4 - 12. Figure 5. The control appears to comprise a line with two chromosome 3 balancers (MKRS/Tm6B). That's a lot of mutations that cannot be regarded as representative of wild type and therefore another control really needs to be used.

Supp
We agree, that the ideal control is a homozygous Sns-Gal4 line without the Dicer and without any balancers. However, we failed to generate this fly strain.
We compared the absolute values for the nephrocyte size of w;sns-Gal4/+;UAS-dicer2/+ and w;sns-Gal4; MKRS/TM6B (control) and did not see a significant difference. We added this genotype to the graph in Figure 5A and 5D. We also did not observe a difference in size when we compared the homozygous Cher Inactive and homozygous Cher Active flies, supporting or finding that the hypertrophy phenotype is lost with increasing Cher Active expression. In addition, the presence of the balancers also did not impact on morphology as shown in Supp Figure 5A.
Also, considering that the flies without the balancers would likely be healthier, the effect observed in the FITC-Albumin assay would be even bigger, which means homozygous Cheerio flies would present with an even worse phenotype. As a second functional read-out we performed the AgNO3 toxin assay. Here we compared the homozygous flies to their respective heterozygous controls. We did the same for the homozygous human Filamin B WT flies and experiments and found no significant difference in cell size when comparing w;sns-Gal4/+;UAS-dicer2/+ and w;sns-Gal4; MKRS/TM6B (control). The above-mentioned argument regarding the FITC-Albumin assay also applies for the Filamin B expressing flies. In addition, to further confirm our findings we performed experiments at 28°C to increase Cheerio levels. Here, we used w;sns-Gal4/+;UAS-dicer2/+ flies as control, which is the correct one. We observed a severe morphological phenotype in flies expressing the different Cheerio variants (Supp. Figure 5B).
Therefore, we decided to keep the data in the manuscript.
13. Discussion: Filamin B is expressed at the apical surface of human proximal tubule cells -cells modelled in the nephrocyte via the endocytosis assays. Proximal tubule cells are never mentioned in the paper and that needs addressed, the nephrocyte is not just a model of podocytes. It is unclear who these findings from the fly model relate to podocytes alone.
We added a paragraph in the discussion addressing this point.
14. There is persistent reference to 'cell injury' yet there is no insult to the cells that constitutes an injury as such; injury is due to gene over-expression that modulate components of the slit diaphragm. An injury model would align more with transient chemical or hydrostatic pressure provocations. Perhaps avoid using the term injury.
We agree that the term injury might be misleading and therefore changed it to 'diseased nephrocytes' throughout the manuscript. All changes are marked in red.
15. Mechanical force is mentioned in the discussion yet there is no methodological approach that addressed forces being exerted on nephrocytes nor the effect this had on nephrocyte function or phenotype. Without such a provocation, it is hard to state that the modifications to nephrocytes were linked to mechanotransduction.
As mentioned above, we performed additional experiments inducing hypo-osmotic stress and could show, that nephrocytes respond to these changes in their physical environment by swelling, but our approach did not reveal morphological changes.
However, as outlined above, we used fly strains expressing the mechanosensor region domain of Cheerio, hence we kept the term mechanotransduction.
16. Figures 4 and 5 have no images of nephrocytes to support the quantified data. Can these be presented?
We added images for Figure 4 as Supplementary Figure 5. Images to support the quantification of the data shown in Figure 5 are now shown in Supplementary Figure   6. Thank you for submitting your revised manuscript entitled "Drosophila Filamin exhibits a mechano-protective role in nephrocytes via Yorkie mediated hypertrophy" to Life Science Alliance. The manuscript has been seen by the original reviewers whose comments are appended below. While the reviewers continue to be overall positive about the work in terms of its suitability for Life Science Alliance, some important issues remain. Please address the final Reviewer 1 minor points and the points raised by Reviewer 2 regarding the change in the language (and in the title as well) throughout the manuscript to better reflect the conclusions.
Our general policy is that papers are considered through only one revision cycle; however, given that the suggested changes are relatively minor, we are open to one additional short round of revision. Please note that I will expect to make a final decision without additional reviewer input upon resubmission.
Please submit the final revision within one month, along with a letter that includes a point by point response to the remaining reviewer comments.
To upload the revised version of your manuscript, please log in to your account: https://lsa.msubmit.net/cgi-bin/main.plex You will be guided to complete the submission of your revised manuscript and to fill in all necessary information.
Please pay attention to adhere to our editorial requirements for revisions: A. THESE ITEMS ARE REQUIRED FOR REVISIONS --A letter addressing the reviewers' comments point by point.
--An editable version of the final text (.DOC or .DOCX) is needed for copyediting (no PDFs).
--High-resolution figure, supplementary figure and video files uploaded as individual files: See our detailed guidelines for preparing your production-ready images, https://www.life-science-alliance.org/authors --Summary blurb (enter in submission system): A short text summarizing in a single sentence the study (max. 200 characters including spaces). This text is used in conjunction with the titles of papers, hence should be informative and complementary to the title and running title. It should describe the context and significance of the findings for a general readership; it should be written in the present tense and refer to the work in the third person. Author names should not be mentioned.

B. MANUSCRIPT ORGANIZATION AND FORMATTING:
Full guidelines are available on our Instructions for Authors page, https://www.life-science-alliance.org/authors Thank you for this interesting contribution to Life Science Alliance.