Gelsolin dysfunction causes photoreceptor loss in induced pluripotent cell and animal retinitis pigmentosa models

Mutations in the Retinitis Pigmentosa GTPase Regulator (RPGR) cause X-linked RP (XLRP), an untreatable, inherited retinal dystrophy that leads to premature blindness. RPGR localises to the photoreceptor connecting cilium where its function remains unknown. Here we show, using murine and human induced pluripotent stem cell models, that RPGR interacts with and activates the actin-severing protein gelsolin, and that gelsolin regulates actin disassembly in the connecting cilium, thus facilitating rhodopsin transport to photoreceptor outer segments. Disease-causing RPGR mutations perturb this RPGR-gelsolin interaction, compromising gelsolin activation. Both RPGR and Gelsolin knockout mice show abnormalities of actin polymerisation and mislocalisation of rhodopsin in photoreceptors. These findings reveal a clinically-significant role for RPGR in the activation of gelsolin, without which abnormalities in actin polymerisation in the photoreceptor connecting cilia cause rhodopsin mislocalisation and eventual retinal degeneration in XLRP.

2, In figure 3g, increase in phallodin signaling in photoreceptors is not phenocopied in RPGR mutant mice. This may be caused by the difference in the developing stages. hiPS-derived photoreceptor cultured for 100 days is not matured one, that is still embryonic stage. So, authors should examined palloidin staining in RPGR mutant mice at more early stage.
3, It is not clear how quantify the intensity of phalloidin immunostaining. Fluorescent intensity was carefully compared with any internal control signals.
4, Authors show that phosphorylated coffin proteins are increased in RPGR/XLRP retinal tissues by western blot. But, it is not clear where are phospho-cofilins localized in developing photoreceptors, and whether they are indeed increased in RPGR/XLRP photoreceptors ? To clarify these, they should also show the results of phopho-cofilin immunostaining in hiPSC-derived retinal tissues (control and patient derived), gelsolin mutant mice and rpgr mutant mice, respectively. 5, In figure 6d, authors say that over expression of active gelsolin could rescue RPGR-loss phenotype in ciliogenesis. Although this result is directly support their claim that gelsolin is functionally related with RPGR, they only show few images from each conditions. To eliminate the possibility of artifacts, quantification analysis of ciliagenesis is also required.

Reviewer #2 (Remarks to the Author)
My review is focused on the use of phospho-arrays. The data shown in Fig. 4 appear to support the conclusion that cortacin is hyperphosphorylated in the absence of RPGR. The subsequent work on gelsolin being a partner of RPGR and excessive actin polymerization being responsible for retinal degeneration is an interesting hypothesis that will stimulate discussion in the field.
Reviewer #3 (Remarks to the Author) In the manuscript entitled "Gelsolin dysregulation causes photoreceptor loss in induced pluripotent stem cell and animal models of retinitis pigmentosa" Megaw et al. propose that impaired F-actin disassembly in photoreceptors carrying mutations in the ciliopathy gene RPGR could be the basis for rhodopsin mislocalization and subsequent photoreceptor cell death. The authors showed that photoreceptor (PR) differentiated from iPSCs from patients carrying a similar RPGR mutation as well as PRs from a RPGR-KO mouse model exhibit increased actin polymerization. At the molecular level, they identified an increase in phosphorylation of cofilin in RPGR mutant which is known to bind to and promote actin disassembly in its unphosphorylated state. Since phosphorylation of cofilin is known to be regulated, among other proteins, by actin-severing protein gelsolin, the authors further examined the photoreceptor morphology in a gelsolin knockout mouse and showed mislocalization of rhodopsin, loss of outer nuclear nuclear layer (ONL) thickness, and disruption of outer segments (OS) similar to that observed in the RPGR-KO mouse. Co-immunoprecipitation studies revealed that gelsolin interacts with WT but not mutant RPGR. Finally, the authors claim restoration of cilia formation in an RPGR-depleted RPE cell line by expression of a constitutively active form of gelsolin. The authors conclude that under WT conditions RPGR activates gelsolin in photoreceptors and thus, is involved in regulation of actin dynamics. This is the first paper describing the interaction of RPGR with gelsolin and proposes a novel possible pathogenic mechanism linking RPGR mutations with rhodopsin mislocalization. While it does not explain the molecular events that subsequently trigger PR cell death, it adds to our understanding of changes in protein trafficking in the connecting cilium of mutant RPGR/XLRP rods and may also contribute towards better explaining other retinal ciliopathies.
The following comments/suggestions are proposed that would significantly strengthen the value of this paper.
Major comments: 1. Statistical tests used have neither been described in the Material & Methods section nor in the figure legends (Fig 2, Fig3, Fig 5). What statistical parameter of central tendency (mean, median) was used? What do the error bars show ( {plus minus}SD, {plus minus}2SD, {plus minus}SEM, {plus minus}2SEM)?
2. The manuscript falls short of identifying the binding site of Gelsolin on RPGR. Is there anything that can be deduced from the Gelsolin co-IP result shown in Figure 6? How do you reconcile the absence of both the constitutive RPGR1-19 and RPGR1-ORF15 bands from the patient derived RPGR mutant cultures? Well-conceived experiments (e.g. co IP studies or yeast two hybrid assays) utilizing RPGR variants that have deleted regions should be included to narrow down the region of interaction between these 2 proteins. Such data may inform as to whether the proposed pathogenic mechanism (and potential gelsolin augmentation therapeutic strategy) would only apply to the subset of RPGR mutations that alter this binding region.
3. The authors have shown some preliminary evidence supporting the claim that overexpression of gelsolin can correct the absence of ciliogenesis (Fig 6d). However there are some limitations to the current results: a) The conclusion is purely based on qualitative results showing some α-tubulin labeled structures. The authors should consider providing quantitative data AND electron microscopy imaging to unequivocally state that a cilium (with its microtubule doublets organization) is being formed. b) The cell used in this study is not a photoreceptor but an RPE cell line. As the authors have access to the RPGR KO mouse why not consider demonstrating whether gelsolin gene augmentation corrects RHO mislocalization and the RPGR-induced cell death process? 4. A well-described feature of RPGR disease is (as reported by the authors) rhodopsin mislocalization, however mislocalization of cone opsins has also been found in mice, dogs, and humans carrying RPGR mutations. Because of the importance of cone-mediated visual impairment in this condition, and current efforts at therapeutically addressing the defect also in this PR cell population, it seems as an oversight to not have investigated in the gelsolin KO mouse whether cone opsin mislocalization is also seen. Regardless of the results, this would be very informative, and these findings should be discussed.
Minor comments: a) The title of the paper does not truly capture the conclusions that can be drawn from the results of this paper. While the studies show an interaction between RPGR and gelsolin, there is no experiment demonstrating a "dysregulation" of gelsolin or even establishing gelsolin dysregulation as a direct "cause" for PR cell loss. It is recommended that the title be changed to better reflect the conclusions that can be drawn based on the results presented in the paper.
b) The order of the panels from f) For figures 3b, 3c and 3h, the asterisks indicate significant change when compared to which value? Please us brackets to indicate this as used for figures 2b, 5b and 5c. g) In the RPGR-KO mouse photoreceptors, there is an increase in the length of actin fibers but a decrease in the phalloidin staining intensity (Fig.3g). What is a possible explanation for this?
h) The result title "RPGR activates the actin-severing protein Gelsolin" is again misleading. The results only show decreased binding of gelsolin to F-actin in RPGR/XLRP; it does not establish RPGR mutation by itself to be a cause of this decrease and certainly does not support the direct involvement of RPGR in activating gelsolin. i) In Figure 4a, there is a slight discrepancy between the fold-change value for ser3 phosphorylation on cofilin stated by the authors in the text (3.16 fold change; RPGR activates the actin-severing protein Gelsolin, line 4) and that indicated by the color key in the figure. Based on the figure 4a, there appears to be higher than 4-fold change in the cofilin phosphorylation. k) It would be informative to calculate the change in actin fiber length in the photoreceptors of the gelsolin-KO mice (Fig. 5), similar to the measurements made in the RPGR-KO photoreceptors (Fig.3g). l) In the co-IP immunoblots using gelsolin IP (Fig.6b), the authors should include IB for gelsolin in order to show whether gelsolin was immunoprecipitated at equivalent amounts in the control and patient cell samples. They should also indicate the MW of gelsolin in the legend of Fig 6. m) While the gelsolin misregulation could possibly explain the cofilin hyper-phosphorylation, a number of other cytoskeletal-modifying proteins are also hyper-phosphorylated in the RPGR/XLRP iPSC-derived photoreceptors. The authors should experimentally prove that it is indeed a disruption of gelsolin-cofilin interaction that is leading to hyper-phosphorylation of cofilin. Is cofilin hyper-phosphorylated in the gelsolin-KO photoreceptors or in gelsolin depleted RPE as well? And does expression of activated gelsolin in RPE alter cofilin phosphorylation in these cells?n) A final diagram, either in the main paper or in the supplementary material, with a model explaining the possible role of RPGR-gelsolin binding and downstream molecular interactions will be a useful addition to clearly explain the conclusions.

Dear Referees
Many thanks for your comments and feedback on our article. You have suggested further experiments and revisions to enhance the body of work. Here we describe these experiments and revisions in turn, beginning with Reviewer #1.

Reviewer #1:
(#1) There is only one example of organoids displaying characteristic phenotype of photoreceptor. How reproducible is this phenotype? The authors should show a range of images displaying the variation in defect of photoreceptors. As well as that, authors show only high magnification of view of retinal tissue. Lower magnification images are also required to clearly show that actin polymerization was increased in the hiPSC-derived differentiating photoreceptors.

RESPONSE #1
We have included, as a supplementary figure (Supplementary Figure 3), further lower magnification images of control and patient iPSC-derived photoreceptor cultures which demonstrate the increased phalloidin staining in recoverin-positive cells in RPGR-mutant lines.

RESPONSE #2
We believe that the increase in phalloidin staining seen in human iPSC-derived photoreceptors was in fact phenocopied in our Rpgr-mutant mice, with increased actin in the connecting cilium (Fig 3g). As the reviewer suggests, we have now examined the Rpgrmutant mice in development (P2 and P10; new Supplementary Figure 4). We found that no difference in the photoreceptor actin cytoskeleton was seen in the P2 or P10 Rpgr KO mouse retina when compared to control. Thus, Rpgr's role in actin regulation does not appear during these developmental stages in vivo -indeed we suggest that it does not appear until after eye opening at P14. We propose that the accelerated actin phenotype seen in hiPSC-derived photoreceptor cultures reflects the stress the organoids are under in the artificial in vitro environment, just as iPS cells from individuals with late onset neurodegenerative diseases such as MND reveal phenotypes much earlier than in vivo.

RESPONSE #3
Examining the same regions of interest on our images as those used for the quantification of phalloidin immunostaining, and using the same parameters, we compared the intensity of nuclei staining (Hoechst 5ug/ml) as an internal control. We found no difference in the intensity of Hoechst staining in our RPGR-mutant iPSC derived photoreceptors compared to control iPSC-derived photoreceptors (data not shown). Given this internal control, we conclude that the difference seen in phalloidin staining between RPGR-mutant and control iPSC-derived photoreceptor cultures is significant.

RESPONSE #4
This is an excellent suggestion, but unfortunately not one we could address as no reliable commercial antibody exists for immunostaining of phospho-cofilin. We have however looked at the overall levels of cofilin expression in gelsolin mutant mice and Rpgr mutant mice. There is strong staining in the photoreceptor layer of the wild type retina. There does not appear to be any difference in protein localization between wild type and mutant retinas / cultures (Supplementary Figure 5), although, the degenerate nature of the retina means the outer segments are not as well defined.
(#5) 5, In figure 6d, authors say that over expression of active gelsolin could rescue RPGRloss phenotype in ciliogenesis. Although this result is directly support their claim that gelsolin is functionally related with RPGR, they only show few images from each conditions. To eliminate the possibility of artifacts, quantification analysis of ciliagenesis is also required.

RESPONSE #5
We have added to figure 6 the graphs (6e) containing the quantification of rescue by constitutively active gelsolin of the ciliogenesis defect seen when RPGR is knocked down in the RPE cell line. The means, SEM and p values are now documented in the legend of figure 6.

Reviewer #2:
(#1) My review is focused on the use of phospho-arrays. The data shown in Fig. 4 appear to support the conclusion that cortacin is hyperphosphorylated in the absence of RPGR. The subsequent work on gelsolin being a partner of RPGR and excessive actin polymerization being responsible for retinal degeneration is an interesting hypothesis that will stimulate discussion in the field.

RESPONSE #1
We were pleased that Reviewer #2 felt the article to be interesting and note that he/she had not suggested we perform any further experiments. figure legends (Fig 2, Fig3, Fig 5).

What statistical parameter of central tendency (mean, median) was used? What do the error bars show ( {plus minus}SD, {plus minus}2SD, {plus minus}SEM, {plus minus}2SEM)? RESPONSE #1
We have amended all our figure legends to include the statistical information required (Mean +/-SEM) (#2) The manuscript falls short of identifying the binding site of Gelsolin on RPGR. Is there anything that can be deduced from the Gelsolin co-IP result shown in Figure 6? How do you reconcile the absence of both the constitutive RPGR1-19 and RPGR1-ORF15 bands from the patient derived RPGR mutant cultures? Well-conceived experiments (e.g. co IP studies or yeast two hybrid assays) utilizing RPGR variants that have deleted regions should be included to narrow down the region of interaction between these 2 proteins. Such data may inform as to whether the proposed pathogenic mechanism (and potential gelsolin augmentation therapeutic strategy) would only apply to the subset of RPGR mutations that alter this binding region.

RESPONSE #2
We agree that defining the binding site for gelsolin on RPGR is important, and like the reviewer we were also struck by the observation that both the constitutive RPGR1-19 and RPGR1-ORF15 forms of RPGR are missing from the gelsolin co-IP in the mutant cells. To address these issues we collaborated with Hemant Khanna at the University of Massachusetts. Using his established RPGR Co-IP profile, we found an RPGR-Gelsolin interaction in the constitutive splice variant RPGR1-19 but not in RPGR1-ORF15 (Supplementary Figure 6). This suggests that the Gelsolin interacting domain of RPGR is within the basic domain encoded by the C terminal region of RPGR1-19. The lack of binding between RPGR1-ORF15 and Gelsolin is at first sight confusing, but based on preliminary data from the Khanna lab that the two forms of RPGR form a complex we hypothesise in the discussion that this complex is required for gelsolin binding and that the XLRP mutation that leads to a disordered ORF15 prevents the formation of this complex and so prevents gelsolin binding and activation.

(#3)
The authors have shown some preliminary evidence supporting the claim that overexpression of gelsolin can correct the absence of ciliogenesis (Fig 6d). However there are some limitations to the current results: a) The conclusion is purely based on qualitative results showing some -tubulin labeled structures. The authors should consider providing quantitative data AND electron microscopy imaging to unequivocally state that a cilium (with its microtubule doublets organization) is being formed. b) The cell used in this study is not a photoreceptor but an RPE cell line. As the authors have access to the RPGR KO mouse why not consider demonstrating whether gelsolin gene augmentation corrects RHO mislocalization and the RPGR-induced cell death process? RESPONSE #3 Please see our response to Reviewer #1's comment #5. We have added to figure 6 the graphs (6e) containing the quantification of rescue by constitutively active gelsolin of the ciliogenesis defect seen when RPGR is knocked down in the RPE cell line. The means, SEM and p values are documented in the legend of figure 6. Reviewer #3 has also asked for confirmation of cilia formation by electron microscopy. Given that hTERT-RPE are ciliated cells, we could certainly use EM to produce a picture of a cilium in our cultures. We are not sure however what this would add. EM is not a quantitative technique and so we instead have used a standard cilia marker and 3 dimensional imaging (in the z plane) to quantify cilia. Finally, Reviewer #3 has asked for demonstration of rescue of the photoreceptor degeneration seen in the Rpgr KO mouse by using gene augmentation. We agree that set of translational experiments are the next logical step for the project. They will not only confirm our proposed mechanism but also offer a possible novel therapeutic for RPGR/XLRP and have amended our discussion to this end. However, we see these experiments as a very substantial undertaking and out with the scope of this paper.
(#4) A well-described feature of RPGR disease is (as reported by the authors) rhodopsin mislocalization, however mislocalization of cone opsins has also been found in mice, dogs, and humans carrying RPGR mutations. Because of the importance of cone-mediated visual impairment in this condition, and current efforts at therapeutically addressing the defect also in this PR cell population, it seems as an oversight to not have investigated in the gelsolin KO mouse whether cone opsin mislocalization is also seen. Regardless of the results, this would be very informative, and these findings should be discussed.

RESPONSE #4
Thank you for suggesting this extension to our study. We have carried out the suggested experiment and report that there is no cone opsin mislocalisation in the Gelsolin KO mouse (this data is not included in the manuscript but provided as a figure for the reviewer). This interesting results illustrates key differences between rod and cone biology in the regulation of opsin localization.

Minor Comment a
The title of the paper does not truly capture the conclusions that can be drawn from the results of this paper. While the studies show an interaction between RPGR and gelsolin, there is no experiment demonstrating a "dysregulation" of gelsolin or even establishing gelsolin dysregulation as a direct "cause" for PR cell loss. It is recommended that the title be changed to better reflect the conclusions that can be drawn based on the results presented in the paper.

RESPONSE a
The reduced RPGR-gelsolin interaction in RPGR-mutant photoreceptor cultures, together with western blot demonstration of reduced Gsn activation, evidences the dysregulation of gelsolin. The retinal degeneration seen in the Gsn ko mouse evidences gelsolin loss as a direct cause for photoreceptor cell loss. We presume that the reviewer is highlighting the difference between a failure of gelsolin activation and the complete absence of gelsolin as reason for caution in assuming causality. However, given that active gelsolin is required for F-actin cleavage, we argue that the two can be equated. In light of the reviewer's comments, however, we have tried to be more precise in a new title "Loss of gelsolin function as a cause of photoreceptor loss in induced pluripotent cell-and animal models of retinitis pigmentosa"