Lysosome-Dependent Sphingolipid Regulation as a potential therapeutic Target for Cohen Syndrome
This article has been Reviewed by the following groups
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
- Evaluated articles (Review Commons)
Abstract
Cohen Syndrome (CS) is a rare autosomal recessive disorder caused by biallelic mutations in the VPS13B gene, affecting approximately 50,000 individuals worldwide. Clinical features include postnatal microcephaly, developmental delay, intellectual disability, neutropenia, and retinal dystrophy. VPS13B belongs to the bridge-like lipid transfer protein (BLTP) family, which also includes VPS13A, VPS13C, and VPS13D in mammals. Although its precise function remains unclear, VPS13B localizes to the Golgi complex, and its loss leads to Golgi fragmentation, a consistent cellular phenotype observed in VPS13B-deficient models. We used the rescue of this cell-autonomous phenotype as the basis for a microscopy-based high-throughput screening assay, through which we identified several small molecules capable of restoring Golgi morphology. Most of these compounds shared a common mechanism of action, relying on lipid accumulation in acidic organelles due to their cationic amphiphilic properties (CADs). Lipidomic profiling revealed a reduction in C18-N-acyl sphingolipids as a characteristic feature of VPS13B knockout (KO) cells, a defect that was reversed by the majority of the identified compounds. To evaluate the physiological relevance of these findings, we tested two compounds, azelastine and raloxifene, in cortical organoids (COs) derived from VPS13B KO human pluripotent stem cells. These organoids exhibited smaller size and reduced neurite outgrowth, reminiscent of the secondary microcephaly observed in CS patients. Treatment with either compound significantly recovered the neurite outgrowth phenotype, reinforcing physiological relevance of the compound effect. Taken together, our findings highlight a potential effect of the CAD on lysosome-dependent sphingolipid regulation, allowing the recovery of Golgi integrity and partial rescue in the cortical organoid CS model. Although additional studies are required to delineate the exact molecular targets, this work uncovers a potential mechanism that could be leveraged for the treatment of CS.
Article activity feed
-
Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We thank the Reviewers for their positive assessment of the quality and significance of our work, as well as for their insightful comments, which have helped us to further improve the manuscript. We have addressed the majority of the comments in the revised version and, for those that require additional time, we outline below a detailed plan of the experiments we intend to perform.
We agree with Reviewer #2 that a more detailed mechanistic understanding of the drug effects would further strengthen the study, and we are grateful to both reviewers for the constructive experimental suggestions provided to address this point. In particular, we are highly …
Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.
Learn more at Review Commons
Reply to the reviewers
We thank the Reviewers for their positive assessment of the quality and significance of our work, as well as for their insightful comments, which have helped us to further improve the manuscript. We have addressed the majority of the comments in the revised version and, for those that require additional time, we outline below a detailed plan of the experiments we intend to perform.
We agree with Reviewer #2 that a more detailed mechanistic understanding of the drug effects would further strengthen the study, and we are grateful to both reviewers for the constructive experimental suggestions provided to address this point. In particular, we are highly motivated to better define the causal role of C18 sphingolipid alterations in mediating the effects of the drugs, as suggested by Reviewer #2, as well as to investigate the involvement of the retromer complex in the lysosome-to-Golgi connection, as suggested by Reviewer #1.
Below, we provide a point-by-point description of the revisions already incorporated into the manuscript, along with the planned experiments that will address the remaining comments
REVIEWER #1:
VPS13B is a bridge-like lipid transfer protein, the loss or mutation of which is associated with Cohen syndrome (CS) involving Golgi fragmentation. In this study, the authors performed image-based chemical screens to identify compounds capable of rescuing the Golgi morphology in VPS13B-KO HeLa cells. They identified 50 compounds, the majority of which are lysosomotropic compounds or cationic amphiphilic drugs (CADs). Treatment of cells with several of these compounds causes lysosomal lipid storage, as assessed by BMP/LBPA staining, filipin staining, or LipidTOX staining. Interestingly, most LipidTOX puncta colocalized with transferrin receptor-positive compartments but not lysosomes. Similar to lysosomotropic compounds, knocking down NPC1 or SMPD1, mimicking lysosomal storage disease, also substantially rescued Golgi morphology. The authors show that VPS13B-KO cells have reduced C18 sphingolipids, which is reversed by treatment with CADs. Finally, the authors show that two CADs partially rescue neurite outgrowth in neuronal cultures. However, these drugs do not rescue the size of VPS13B KO organoids.
Overall, this is an impressive study identifying CADs as potential therapeutics for CS and suggesting sphingolipid upregulation as a general strategy for CS treatment. The morphological and lipidomics analyses unravel important molecular basis of CS pathology. This study will be of high interest to the field of lipid biology and organelle homeostasis. I have a few comments to help improve the quality of this study.
- The reverse of lipid changes in VPS13B-KO cells by CADs is intriguing. Are CAD-mediated benefits such as Golgi morphology recovery permanent or only transient within 24 hours of treatment? How do the CADs affect the Golgi morphology in WT HeLa cells?
RESPONSE:
We thank the reviewer for this insightful question Indeed, the effects of CADs on Golgi organization are most evident in VPS13B KO cells, where the Golgi apparatus is severely fragmented and becomes more compact upon drug treatment, whereas the effect is much less apparent in wild-type cells. Nevertheless, a careful quantitative analysis of the images (now presented in the new Fig. S7) demonstrates that the impact of these compounds on Golgi morphology is not restricted to KO cells but is likely more general, supporting a link between lysosomal storage and Golgi organization. Although this observation indicates an indirect effect (consistent with the proposed mechanism of action), rather than a direct correction of VPS13B loss, it does not compromise in our opinion their potential beneficial effect for KO cells as shown also from the results obtained in organoid-derived neurons.
Under continuous treatment, azelastine keeps the Golgi in a compact state for 72 hours without any noticeable deleterious effect on the cells (see new Fig. S10) Raloxifene, on the contrary proved to be toxic over the same time period. We believe this difference reflects the mechanism of action of CADs, which progressively accumulate within acidic organelles and may eventually reach a toxic threshold upon prolonged exposure. For this reason, lower drug concentrations administered over longer treatment periods may represent a viable alternative strategy. In this regard, we also refer the reviewer to our response to the comment on brain organoids below.
- Is it surprising that Azelastine-induced lipid storage in transferrin receptor compartments (early and recycling endosomes)? I suggest more controls to examine LipidTOX overlap with Golgi markers or other late endosome/lysosome markers such as LBPA and CD63.
RESPONSE:
We agree with the reviewer that this observation is somewhat unexpected. However, we would like to clarify that we do not intend to suggest that lipid storage occurs primarily in early or recycling endosomes, which would indeed contradict a substantial body of existing evidence. Rather, our data indicate that this particular dye (LipidTOX) labels recycling endosomes, at least in HeLa cells. This finding is consistent with the widely accepted view that lysosomal lipid storage exerts broader effects on intracellular trafficking, not limited to late endosomes/lysosomes. We corrected the text in order to clarify this concept.
LipidTOX was specifically developed to detect drug-induced phospholipidosis, and based on our data, it appears suitable for this purpose. To our knowledge, there is no published information detailing its intracellular localization, which motivated us to perform these control experiments. Unfortunately, the proprietary formulation of this product does not allow informed speculations to explain the observed localization or whether this could refer to the intact molecule or to a catabolite.
As suggested by the reviewer, we plan to perform co-staining with additional markers to further clarify this this point.
- Does the LipidTOX/TFRC overlap suggest potential roles of retrograde transport in supplying sphingolipids to the Golgi? The authors can quickly test if the knockdown of a retromer subunit (VPS35) blocks Azelastine-induced recovery of Golgi morphology.
RESPONSE:
We thank the reviewer for this insightful suggestion. Indeed, the retromer complex represents one of the best-characterized trafficking pathways from the endosomal system to the Golgi, and this relatively straightforward experiment could help to mechanistically clarify our observations. We plan to test whether VPS35 knockdown interferes with the effects of the drugs.
What is the rationale to use 500 nM to 1 uM azelastine and raloxifene for neuronal cultures and organoids? At such concentrations, no obvious changes in Golgi morphology or lipid storage were observed (Fig 4). Also, the lipidomics analysis was performed after 10 uM compound treatment. It might be worth trying dose-response experiments in organoid tests.
RESPONSE:
We thank the reviewer for this question. The rationale about this choice was indeed missing from our previous version of the manuscript. The reason of lowering the concentrations comes indeed from toxicity tests, preliminarily performed over long-term treatment of both WT and VPS13B KO organoids. This information has now been explicitly included in the Results section of the revised manuscript, and the broader implications are also discussed in the Discussion section.
MINOR COMMENTS:
It is important to know whether the authors used TGN or cis-Golgi markers for Golgi morphology analysis. Please label the two channels in Fig. 2C and throughout all figures. In many cases, it is not clear what is stained in the green channel to show the Golgi morphology. It was not even stated in the legend.
RESPONSE:
We now included the antibody staining in all figure legends where it was previously missing.
The authors stated that Recovery of Golgi morphology is dependent on lysosomal lipid storage. However, while the data show positive correlation between the two, no causal relationship is established by the data. It seems true that in all conditions (CADs or genetic knockdown) where lysosomal lipid storage was observed, the authors detect the Recovery of Golgi morphology. However, budesonide did not depend on lysosomal lipid storage to recover the Golgi morphology. Thus, the recovery of Golgi morphology is NOT dependent on lysosomal lipid storage, but inducing lysosomal lipid storage appears sufficient to recover Golgi morphology in VPS13B-KO HeLa cells.
RESPONSE:
We thank the reviewer for this comment and we agree that the previous title of the paragraph could have been misleading. This has been now changed in: “Lysosomal lipid storage mediates the recovery of Golgi morphology” which is probably less prone to ambiguous interpretations.
Obviously, in the previous version of the title we wanted to mean that Golgi recovery is dependent on lipid storage “in the context of CAD treatment” and not as a general statement.
With respect to the cause–effect relationship, we believe that the strongest evidence supporting this link is the observation that genetically induced lipid storage phenocopies the effects of drug treatment. We hope that this conclusion is now sufficiently clear from the revised text.
Each figure needs a title before the detailed legends for specific panels.
RESPONSE:
Titles have now been included to all figure legends.
Fig 8. Y axis labeling is missing.
RESPONSE:
Axes labels have now been included
Does U18666A rescues Golgi morphology in VPS13B-KO cells?
RESPONSE:
We thank the reviewer for this comment. U18666A indeed also corrects Golgi morphology. The result is now included in the new figure S5.
Please do not repeat the result section in discussion. Focus on the most important points.
RESPONSE:
We thank the reviewer for this comment. We shortened the descriptive part of the discussion trying as much as possible to avoid repetitions with the result session and keeping only the more essential information for the flow of the discussion.
Reviewer #1 (Significance (Required)):
This is an impressive study that identifies Cationic Amphiphilic Drugs (CADs) as potential therapeutics for Cohen syndrome (CS) and suggests sphingolipid upregulation as a general strategy for diseases driven by VPS13B loss-of-function. The unbiased approaches, notably the chemical screen and lipidomics, provide novel mechanistic insights into the underlying pathology of CS. This study will be of high interest to researchers in the fields of lipid biology and organelle homeostasis. It will also be highly valuable for clinical pediatricians managing CS patients.
REVIEWER #2:
This manuscript describes a compound screening aimed at identifying molecules that can restore Golgi organization in VPS13B knockout (KO) cells. The authors identify several compounds, most of which are lysosomotropic, and analyze their effects on Golgi morphology and lipid composition using multiple approaches. They report that VPS13B KO cells exhibit a reduction in C18-N-acyl sphingolipids, which can be restored by several of the identified compounds. Furthermore, two of these compounds, azelastine and raloxifene, promote neurite outgrowth in VPS13B KO cortical organoids. These findings are interesting and could potentially contribute to a better understanding of the pathophysiology of Cohen syndrome and the development of therapeutic strategies. However, despite the large number of analyses presented, the study remains largely descriptive, and there is no coherent mechanistic explanation for how these compounds restore Golgi structure in VPS13B KO cells. In addition to the reduction in C18-N-acyl sphingolipids, the KO cells display alterations in several other lipid species (LPC, LPE, PC40:1, PE42:1, TG, etc.), and treatment with the selected compounds induces further lipid accumulations, including cholesterol and BMP/LBPA. The relationship between these diverse lipid changes and the observed Golgi recovery lacks clarity and mechanistic consistency.
MAJOR COMMENTS:
The finding that compounds cannot prevent Golgi fragmentation caused by brefeldin A or nocodazole but can suppress statin-induced fragmentation is intriguing, but the underlying mechanism is not addressed. It is not evident whether this difference results from changes in membrane lipid composition or restoration of Rab/SNARE trafficking. The authors should examine Rab prenylation and SNARE localization by immunofluorescence or Western blotting to support their interpretation.
RESPONSE:
We thank the reviewer for this suggestion and agree that the ability of these compounds to counteract statin-induced Golgi fragmentation is indeed intriguing. The primary reason we did not further explore this aspect is that we evaluated the effects of statins not to be a central focus of the present study. Nevertheless, we fully agree that this observation represents a valuable opportunity to gain additional insight into the mechanism underlying drug-induced Golgi recovery.
To address this point, we plan to analyze Rab prenylation by Western blot and Rab localization by microscopy, focusing on a Golgi-associated Rab protein such as Rab6. In addition, we will employ downstream inhibitors of Rab prenylation, such as 3-PEHPC (an inhibitor of type II protein geranylgeranyltransferase (GGTase-II)), which should allow us to formally distinguish effects related to impaired Rab prenylation from those arising from inhibition of cholesterol biosynthesis.
Although restoration of C18 sphingolipids (SM 36:1, CER 36:1) is observed upon compound treatment, its causal role in Golgi recovery or neurite outgrowth is not established. The authors should test whether blocking the increase of C18 SM/CER prevents the rescue of Golgi or neuronal phenotypes.
RESPONSE:
We sincerely thank the reviewer for this comment. We agree that, based on the current data, a definitive cause–effect relationship between Golgi recovery and the increase in C18 sphingolipids cannot be firmly established, and we acknowledge that a deeper understanding of this issue will require further investigation. Furthermore, we believe that addressing this would not only provide a better mechanistic understanding of the biological processes behind the effect of the drugs but provide a potential avenue for therapeutic intervention. For these reasons, we are strongly motivated to pursue this aspect further.
With respect to the reviewer’s specific suggestion, we agree that preventing the increase in C18 sphingolipids would be an ideal experimental approach. However, the limited understanding of the regulatory mechanisms controlling C18 sphingolipid homeostasis currently precludes a fully informed strategy. In principle, if the observed increase were due to enhanced synthesis, one could envisage blocking it by silencing ceramide synthases with C18 selectivity, such as CERS1. The experiment shown in Fig. 7E (azelastine treatment in the presence of sphingolipid synthesis inhibitors) was designed with this rationale in mind. However, these results suggest that azelastine-induced C18 sphingolipid accumulation is unlikely to result from increased synthesis, and is instead more consistent with reduced degradation, in line with the proposed mechanism of action of CADs.
Based on these considerations, we propose to invert the experimental approach and test whether cellular re-complementation with C18 sphingolipids is sufficient to recapitulate the drug-induced Golgi recovery. We are aware of the technical challenges associated with the targeted delivery of exogenously supplied lipids, particularly given the likelihood that effective rescue would require lipid access to the Golgi apparatus. Based on current knowledge, we anticipate that externally supplied lipids would primarily traffic either to the ER via non-vesicular routes or to endosomes/lysosomes through endocytic uptake. From both locations they could eventually reach to some extent the Golgi. The route from endosomes to Golgi in particular as been intensively studied in the past with the use of fluorescent sphingolipid analogs1,2 and may well work also with native lipids.
Since we are not able to predict in advance which lipid species would be more effective or the optimal delivery strategy, we plan to test re-complementation using C18 sphingomyelin and some of its potential precursors, including C18 ceramide as well as using alternative delivery strategies such as incorporation in liposomes of different formulations and delivery at the plasma membrane with bovine serum albumin or cyclodextrins as carriers.
- Puri et al., (2001). J Cell Biol.154:535-47 (doi: 10.1083/jcb.200102084)
- Koivusalo et al.,(2007). Mol Biol Cell. 18:5113-23 (doi: 10.1091/mbc.e07-04-0330)
In Figure 7D, comparisons should include the LM and HM fractions isolated from WT cells.
RESPONSE:
Wild-type control were included in the figure as requested.
The subcellular fractionation experiment should be repeated using AZL and RAL, the compounds used in organoid experiments, rather than TFPZ, to assess whether similar results are obtained. The compounds used differ across experiments, making it difficult to draw consistent conclusions.
RESPONSE:
We thank the reviewer for this comment and apology for some inconsistencies in the selection of the compounds to highlight in the figures which are mostly remnants of the drug prioritization history over the progression of the project. We tried to make it more consistent in the current version.
In the new version of figure 7D, AZL is substituting TFPZ, while TFPZ data were moved to supplementary figure S19.
Golgi morphology in VPS13B KO cells is reported to recover in NPC1 KD and SMPD1 KD cells, but it is not shown whether SM 36:1, CER 36:1, or other lipid levels also increase or change in these conditions. If Golgi morphology recovery occurs via the same mechanism as with compound treatment, a similar lipid pattern should be observed.
RESPONSE:
We thank the reviewer for this question that allowed us to expand our study including new interesting findings. We agree that this is an important point to strengthen the link between CAD and genetic perturbation effects. Given the availability of several published lipidomic datasets modelling LDS in HeLa and in other cell lines, we decided to perform a re-analysis of those to specifically focus on C18 sphingolipids. We found a relative increase of 36:1 upon depletion of LSD genes in all analyzed datasets for NPC1 and SMPD1, but also for more than 15 other LSD genes including NPC2, recapitulating what we find with all the CAD molecules tested in our study. These changes, were not noticed or at least not discussed by most of the authors. This is not surprising since those studies are focused on different biological questions. We believe that these findings, besides reinforcing our hypothesis of a common mechanism between CAD and NPC1/SMPD1 KO, have of general interest for the regulation of C18 sphingolipids, which are among the relative few lipid species with a bona fide specific protein binding partner and proposed to play a crucial role in Golgi traffic.
MINOR POINTS:
The manuscript lacks sufficient information about the compound library used for screening (number and source of compounds, compound type).
RESPONSE:
We apologize if this information was not sufficiently visible in the original version of the manuscript. The data about source, catalog number, formulation and several additional identifiers is included in the File S1. This is now clearly indicated in the methods so that I can be more easily visible to the readers
Fig. 3A: a WT control image is required.
RESPONSE:
A WT control image is now included in the new version of Figure 3.
Fig. 4: include representative images at concentrations higher than 1.25 µM.
RESPONSE:
Representative images are now included for all concentrations higher than 1.25 µM, as requested.
Abbreviations such as BMP/LBPA should be defined when first mentioned.
RESPONSE:
The abbreviation of BMP/LBPA was already defined when first mentioned in the original version of the manuscript
The abbreviation for raloxifene is inconsistent (RLX vs RAL) and should be unified.
RESPONSE:
Raloxifene is now abbreviated as RLX all over the manuscript.
Fig. 5C: the meaning of the green and magenta bars is not explained.
RESPONSE:
Color code for figure 5C has been included.
The definitions and centrifugation parameters for light and heavy membrane fractions should be clearly stated in the Methods.
RESPONSE:
The centrifugation parameters were already defined in the original manuscript. It is not clear to us, which parameter the Referee is referring to. Below is the sentence in the methods section:
“Gradients were centrifuged at 165,000 g for 1.5 h at 4°C with a SW40Ti Swinging-Bucket rotor (Beckman-Coulter). The LM and HM fractions were collected at the 35%-HB and 35%-40.6% interfaces, respectively”
The concentration and incubation times for BFA and nocodazole should be included in the main text or figure legends.
RESPONSE:
Concentrations and incubation times of BFA and nocodazole were already present in the legend of figure 5.
Fig. 8C, D, G, H: y-axes lack labels and must be defined.
RESPONSE:
Axes labels have now been included
There are multiple typographical errors, including "VPS12" instead of "VPS13B", that should be corrected.
RESPONSE:
We corrected this specific mistake as well as others that we could identify after careful reading of the manuscript.
Reviewer #2 (Significance (Required)):
While the dataset is extensive and technically detailed, the manuscript lacks a clear mechanistic explanation connecting lipid changes to Golgi restoration. The choice and comparison of compounds are inconsistent across experiments, and the interpretation remains speculative. Substantial revision and additional experiments are required before the study can be considered for publication.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
This manuscript describes a compound screening aimed at identifying molecules that can restore Golgi organization in VPS13B knockout (KO) cells. The authors identify several compounds, most of which are lysosomotropic, and analyze their effects on Golgi morphology and lipid composition using multiple approaches. They report that VPS13B KO cells exhibit a reduction in C18-N-acyl sphingolipids, which can be restored by several of the identified compounds. Furthermore, two of these compounds, azelastine and raloxifene, promote neurite outgrowth in VPS13B KO cortical organoids. These findings are interesting and could potentially …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #2
Evidence, reproducibility and clarity
This manuscript describes a compound screening aimed at identifying molecules that can restore Golgi organization in VPS13B knockout (KO) cells. The authors identify several compounds, most of which are lysosomotropic, and analyze their effects on Golgi morphology and lipid composition using multiple approaches. They report that VPS13B KO cells exhibit a reduction in C18-N-acyl sphingolipids, which can be restored by several of the identified compounds. Furthermore, two of these compounds, azelastine and raloxifene, promote neurite outgrowth in VPS13B KO cortical organoids. These findings are interesting and could potentially contribute to a better understanding of the pathophysiology of Cohen syndrome and the development of therapeutic strategies. However, despite the large number of analyses presented, the study remains largely descriptive, and there is no coherent mechanistic explanation for how these compounds restore Golgi structure in VPS13B KO cells. In addition to the reduction in C18-N-acyl sphingolipids, the KO cells display alterations in several other lipid species (LPC, LPE, PC40:1, PE42:1, TG, etc.), and treatment with the selected compounds induces further lipid accumulations, including cholesterol and BMP/LBPA. The relationship between these diverse lipid changes and the observed Golgi recovery lacks clarity and mechanistic consistency.
Major comments
The finding that compounds cannot prevent Golgi fragmentation caused by brefeldin A or nocodazole but can suppress statin-induced fragmentation is intriguing, but the underlying mechanism is not addressed. It is not evident whether this difference results from changes in membrane lipid composition or restoration of Rab/SNARE trafficking. The authors should examine Rab prenylation and SNARE localization by immunofluorescence or Western blotting to support their interpretation.
Although restoration of C18 sphingolipids (SM 36:1, CER 36:1) is observed upon compound treatment, its causal role in Golgi recovery or neurite outgrowth is not established. The authors should test whether blocking the increase of C18 SM/CER prevents the rescue of Golgi or neuronal phenotypes.
In Figure 7D, comparisons should include the LM and HM fractions isolated from WT cells.
The subcellular fractionation experiment should be repeated using AZL and RAL, the compounds used in organoid experiments, rather than TFPZ, to assess whether similar results are obtained. The compounds used differ across experiments, making it difficult to draw consistent conclusions.
Golgi morphology in VPS13B KO cells is reported to recover in NPC1 KD and SMPD1 KD cells, but it is not shown whether SM 36:1, CER 36:1, or other lipid levels also increase or change in these conditions. If Golgi morphology recovery occurs via the same mechanism as with compound treatment, a similar lipid pattern should be observed.
Minor points
The manuscript lacks sufficient information about the compound library used for screening (number and source of compounds, compound type).
Fig. 3A: a WT control image is required. Fig. 4: include representative images at concentrations higher than 1.25 µM. Abbreviations such as BMP/LBPA should be defined when first mentioned. The abbreviation for raloxifene is inconsistent (RLX vs RAL) and should be unified. Fig. 5C: the meaning of the green and magenta bars is not explained. The definitions and centrifugation parameters for light and heavy membrane fractions should be clearly stated in the Methods. The concentration and incubation times for BFA and nocodazole should be included in the main text or figure legends. Fig. 8C, D, G, H: y-axes lack labels and must be defined. There are multiple typographical errors, including "VPS12" instead of "VPS13B", that should be corrected.
Significance
While the dataset is extensive and technically detailed, the manuscript lacks a clear mechanistic explanation connecting lipid changes to Golgi restoration. The choice and comparison of compounds are inconsistent across experiments, and the interpretation remains speculative. Substantial revision and additional experiments are required before the study can be considered for publication.
-
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
VPS13B is a bridge-like lipid transfer protein, the loss or mutation of which is associated with Cohen syndrome (CS) involving Golgi fragmentation. In this study, the authors performed image-based chemical screens to identify compounds capable of rescuing the Golgi morphology in VPS13B-KO HeLa cells. They identified 50 compounds, the majority of which are lysosomotropic compounds or cationic amphiphilic drugs (CADs). Treatment of cells with several of these compounds causes lysosomal lipid storage, as assessed by BMP/LBPA staining, filipin staining, or LipidTOX staining. Interestingly, most LipidTOX puncta colocalized with …
Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.
Learn more at Review Commons
Referee #1
Evidence, reproducibility and clarity
VPS13B is a bridge-like lipid transfer protein, the loss or mutation of which is associated with Cohen syndrome (CS) involving Golgi fragmentation. In this study, the authors performed image-based chemical screens to identify compounds capable of rescuing the Golgi morphology in VPS13B-KO HeLa cells. They identified 50 compounds, the majority of which are lysosomotropic compounds or cationic amphiphilic drugs (CADs). Treatment of cells with several of these compounds causes lysosomal lipid storage, as assessed by BMP/LBPA staining, filipin staining, or LipidTOX staining. Interestingly, most LipidTOX puncta colocalized with transferrin receptor-positive compartments but not lysosomes. Similar to lysosomotropic compounds, knocking down NPC1 or SMPD1, mimicking lysosomal storage disease, also substantially rescued Golgi morphology. The authors show that VPS13B-KO cells have reduced C18 sphingolipids, which is reversed by treatment with CADs. Finally, the authors show that two CADs partially rescue neurite outgrowth in neuronal cultures. However, these drugs do not rescue the size of VPS13B KO organoids.
Overall, this is an impressive study identifying CADs as potential therapeutics for CS and suggesting sphingolipid upregulation as a general strategy for CS treatment. The morphological and lipidomics analyses unravel important molecular basis of CS pathology. This study will be of high interest to the field of lipid biology and organelle homeostasis. I have a few comments to help improve the quality of this study.
- The reverse of lipid changes in VPS13B-KO cells by CADs is intriguing. Are CAD-mediated benefits such as Golgi morphology recovery permanent or only transient within 24 hours of treatment? How do the CADs affect the Golgi morphology in WT HeLa cells?
- Is it surprising that Azelastine-induced lipid storage in transferrin receptor compartments (early and recycling endosomes)? I suggest more controls to examine LipidTOX overlap with Golgi markers or other late endosome/lysosome markers such as LBPA and CD63.
- Does the LipidTOX/TFRC overlap suggest potential roles of retrograde transport in supplying sphingolipids to the Golgi? The authors can quickly test if the knockdown of a retromer subunit (VPS35) blocks Azelastine-induced recovery of Golgi morphology.
- What is the rationale to use 500 nM to 1 uM azelastine and raloxifene for neuronal cultures and organoids? At such concentrations, no obvious changes in Golgi morphology or lipid storage were observed (Fig 4). Also, the lipidomics analysis was performed after 10 uM compound treatment. It might be worth trying dose-response experiments in organoid tests.
Minor:
- It is important to know whether the authors used TGN or cis-Golgi markers for Golgi morphology analysis. Please label the two channels in Fig. 2C and throughout all figures. In many cases, it is not clear what is stained in the green channel to show the Golgi morphology. It was not even stated in the legend.
- The authors stated that Recovery of Golgi morphology is dependent on lysosomal lipid storage. However, while the data show positive correlation between the two, no causal relationship is established by the data. It seems true that in all conditions (CADs or genetic knockdown) where lysosomal lipid storage was observed, the authors detect the Recovery of Golgi morphology. However, budesonide did not depend on lysosomal lipid storage to recover the Golgi morphology. Thus, the recovery of Golgi morphology is NOT dependent on lysosomal lipid storage, but inducing lysosomal lipid storage appears sufficient to recover Golgi morphology in VPS13B-KO HeLa cells.
- Each figure needs a title before the detailed legends for specific panels.
- Fig 8. Y axis labeling is missing.
- Does U18666A rescues Golgi morphology in VPS13B-KO cells?
- Please do not repeat the result section in discussion. Focus on the most important points.
Significance
This is an impressive study that identifies Cationic Amphiphilic Drugs (CADs) as potential therapeutics for Cohen syndrome (CS) and suggests sphingolipid upregulation as a general strategy for diseases driven by VPS13B loss-of-function. The unbiased approaches, notably the chemical screen and lipidomics, provide novel mechanistic insights into the underlying pathology of CS. This study will be of high interest to researchers in the fields of lipid biology and organelle homeostasis. It will also be highly valuable for clinical pediatricians managing CS patients.
-
