Galectin-9 interacts with Vamp-3 to regulate cytokine secretion in dendritic cells

  1. Rui Santalla Méndez
  2. Andrea Rodgers Furones
  3. René Classens
  4. Manon Haverdil
  5. Marta Canela Capdevila
  6. Anne van Duffelen
  7. Cornelia G. Spruijt
  8. Michiel Vermeulen
  9. Martin ter Beest
  10. Annemiek B. van Spriel
  11. Laia Querol Cano

Reviewed by

Read the full article

Listed in

Log in to save this article

Abstract

Intracellular vesicle transport is essential for cellular homeostasis and is partially mediated by SNARE proteins. Endosomal trafficking to the plasma membrane ensures cytokine secretion in dendritic cells (DCs) and the initiation of immune responses. Despite its critical importance, the specific molecular agents that regulate DC cytokine secretion are poorly characterised. Galectin-9, a ß-galactoside-binding protein, has emerged as a novel cellular modulator although its exact intracellular roles in regulating (immune) cell homeostasis and vesicle transport are virtually unknown. We investigated galectin-9 function in primary human DCs and report that galectin-9 is essential for intracellular cytokine trafficking to the cell surface. Galectin-9-depleted DCs accumulate cytokine-containing vesicles in the Golgi complex that eventually undergo lysosomal degradation. We observed galectin-9 to molecularly interact with Vamp-3 using immunoprecipitation-mass-spectrometry and identified galectin-9 was required for rerouting Vamp-3-containing endosomes upon DC activation as the underlying mechanism. Overall, this study identifies galectin-9 as a necessary mechanistic component for intracellular trafficking. This may impact our general understanding of vesicle transport and shed new light into the multiple roles galectins play in governing cell function.

Article activity feed

  4. Review Commons

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

    Learn more at Review Commons

    Reply to the reviewers

    Manuscript number: RC-2023-01899

    Corresponding author(s): Laia Querol Cano

    1. General Statements [optional]

    We thank the reviewers for their positive evaluation and constructive feedback and appreciate the reviewers’ assessment that our findings are highly interesting and novel. We have already addressed some of their questions (see section 3 below and highlighted changes in the manuscript text) and initiated experiments to address the reviewers’ suggestions. We anticipate these experiments can be completed within three months.

    2. Description of the planned revisions

    Reviewer #1

    Summary This paper is focused on the role of galectin-9 in dendritic cells using monocyte-derived DCs. To study the functional characteristics of galectin-9, they depleted galectin-9 with the use of gal9 siRNA. They found galectin-9 is required for TNF-alpha, IL-6, IL-10 and IL-12 secretion. Galectin-9 was found to be involved in TNF trafficking and interacted with Vamp-3 to regulate the release of TNF. They concluded that galectin-9 controls cytokine intracellular trafficking and secretion through functional interaction with the SNARE protein Vamp-3 in DCs through an endosomal compartment.

    The findings seem significant in understanding the role of galectin-9 in dendritic cells which has not been previously explored, and they also expand our understanding of the roles of galectins and their potential function in intracellular cytokine trafficking. However, there are some concerns about the findings and conclusions of the study.

    Major comments

    Supplementary Figure 2A. No obvious decrease in TNF, IL-6, or IL-12 was observed in gal9 siRNA-treated cells that were stimulated with LPS and zymosan, and results in this panel are inconsistent with those of Suppl. Fig. 2B. An explanation for this discrepancy should be provided.

    *We thank the reviewer for pointing this out. The discrepancy can be explained by the different time points used in Supplementary Fig. 2A versus Supplementary Figure 2B and main Figure 1. Supplementary Figure 2A depicts the optimisation carried out to screen the best stimuli for each of the cytokines analysed after 36h, which is considerably longer than the timepoints used throughout the manuscript (4-16 h). To make this consistent, we will repeat these experiments using 16 h stimulation. *

    The data is sufficient to support the notion that galectin-9 is required for cytokine secretion, but intracellular staining of an additional cytokine (IL-6, IL-10, and IL-12) would be a good addition with controls included (non-transfected cells vs WT vs galectin-9-depleted).

    This is a good suggestion and we performed intracellular stainings to detect IL-6, IL-10 and IL-12 on WT and galectin-9 depleted moDCs. Whereas intracellular IL-6 or IL-12 could not be detected with any of the commercial antibodies (IL-6 clone MQ2-13A5, Biolegend #501103; IL-12 clone 20C2, BD Biosciences #557020), we set-up intracelullar IL-10 stainings in primary DCs. IL-10 intracellular accumulation was observed in gal-9 KD DCs compared to WT cells upon LPS treatment although to a lesser extent than TNFα (see figure 1 for reviewer). We are now repeating this assay on multiple donors and this data will be incorporated to the revised manuscript.

    Figure 1 for reviewer. Intracellular flow cytometry showing IL-10 levels in NT siRNA (black) and gal9 siRNA (light grey) moDCs treated with LPS for 6 h. Isotype control is depicted with unfilled dashed line. Numbers represent geometric mean intensity.

    Addition of an immunofluorescent experiment using Vamp-3 and TNF co-localization in gal9 siRNA-treated cells would strengthen observations regarding galectin-9 association with Vamp-3 in immunoprecipitation.

    • We agree and will address this by:*
    • Immunofluorescence studies with moDCs (three independent donors with galectin-9 depletion (and Non Targeting siRNA counterparts) stimulated with LPS for 2, 4 and 6 h. Staining of Vamp-3, TNF-α, galectin-9 and DAPI is already established as have used these antibodies throughout the manuscript and thus do not anticipate any issues when performing these assays.
    • *Co-localisation of galectin-9 and Vamp-3 will be determined by quantifying both the Manders’ and Pearson’s correlation coefficients (See Major point #4 from reviewer #1 in section 3). *

    Figure 7A. Vamp-3 does not appear to redistribute towards the cell membrane following LPS stimulation in this figure. Either a different set of images needs to be selected or the text needs to be revised.

    *To address this issue, we will include an enlarged zoomed in image of a representative cell in the revised version of the manuscript. Furthermore, we will include a Golgi marker (GM130) in our staining panel to quantify Vamp-3- Golgi co-localisation in WT and Galectin-9 depleted moDCs treated with LPS. *

    All findings in this study regarding galectin-9 immunoreactivity are dependent on a single goat anti-galectin-9 antibody (AF2045). Findings would be strengthened by the use of a second galectin-9-specific antibody in at least one additional experiment (either immunofluorescence or immunoprecipitation).

    *We have purchased another anti galectin-9 antibody (clone 9M1-3, BioLegend #348902) that will be used for immunofluorescence experiments to confirm our findings. *

    Minor comments

    Figure 2D needs at least one more experiment before results can be depicted. An n=2 is not sufficient to merit publication.

    • We agree and will conduct the same experiment as described in Figure 2D with one additional donor to obtain n=3. Data from panel 2E will be also updated to incorporate the new data set.*

    Supplementary Figure 2 has additional small square symbols in panels A and B that should be removed.

    • We apologise for this. Supplementary Figure 2A will be re-made with new donors for the revised version of the manuscript without the square symbols. Supplementary Figure 2B has been remade to include four donors for each time point and stimulation and the square symbols have been removed.*

    Supplementary Figure 2 legend. This legend has repetitive text regarding representative data from one donor. How many donors were tested for these experiments?

    • Please see minor comment #5 above. Only a representative donor was included for panel A but experiments are being conducted to replace this figure with a more complete one including data from at least three independent experiments. Panel B has been remade and now includes data from four independent donors.*

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

    This paper shows that knock down of galectin-9 suppresses secretion of certain cytokines in activated dendritic cells. Then they show that this correlates with failure of the peripheral localization of the SNARE- protein VAMP3 and suggest that this is due to a direct interaction with galectin-9.

    Most of the data consist of comparing wt and galectin-9 KD cells regarding secretion and trafficking of selected cytokines (with focus on TNFa), and of cellular localization of the vSNARE VAMP3. The cells are mainly LPS activated human dendritic cells derived from differentiation on blood monocytes from donors, but differentiated THP-1 cells are also used. These data show convincingly that cytokine secretion is inhibited in galectin-3 KD cells, and for TNFa studied on detail, this is due to failure of post-Golgi trafficking of the, shown in different ways, including a RUSH assay. The trafficking of the vSNARE VAMP3 to the periphery of the cells is also inhibited by absence of galecytin-9, leading to its retention in the Golgi nearer the nucleus. Thus this couples in a surprising the unsolved question of how SNAREs themselves are trafficked to their correct destinations, to the function of a cytosolic galectin.

    The weak part of the paper is the molecular interaction between galectin-9 and VAMP3. This is based on co-immunoprecipitations followed by proteomic characterization and Western blots as summarized in Fig. 6. These data show that galectin-9 and VAMP3 occur in the same precipitated complex, but not that they interact directly. Many other proteins are also in these complexes, and the Western blot data are not very strong. Thus additional experiments would be needed to claim the direct interaction as depicted in Fig.8 for example using purified recombinant proteins, or sharpened focus using mutants of the two interactors.

    • We thank the reviewer for these positive comments and agree that protein co-immunoprecipitation does not warrant direct interaction. To discern whether galectin-9 and Vamp-3 directly interact or are part of a bigger protein complex, we will use purified recombinant proteins in GST-pull down assays. Briefly, we will generate and express Vamp-3-GST constructs that will be incubated with recombinant Galectin-9 protein, which has been performed in a similar manner in Miller et al., Cell. 2011. Protein complexes will be resolved and analysed by SDS-Page. GST-only beads will be used as negative control and a known SNARE complex (Syntaxin 4 together with Snap23) will be used as positive control in these experiments. We have experience in producing recombinant proteins in HEK293 cells And GST-pull down experiments within our department. *

    3. Description of the revisions that have already been incorporated in the transferred manuscript

    Reviewer #1:

    Major Comments

    1. Page 4, Results. The assertion of 90% downregulation of galectin-9 protein is not substantiated by the data shown in Supplementary Figure 1 (it is not indicated which data, the flow or Western blot, provides the source for this statement). It must be assumed it is from Western blots, as the data from flow cytometry does not show a reduction of 90%. This assertion would be strengthened by density analysis of Western blots shown as a graph beside the blots in Supplementary Figure 1C. *We thank the reviewer for addressing this point. This refers to the Western blot data. We have now quantified four independent donors and have added this graph as a new panel in Supplementary Figure 1 (new panel D). *

    Mander's correlation coefficient is typically not advised for use in co-localization of immunofluorescence since it has been found to downplay associations in low intensity fluorescent staining and favors high intensity and co-occurrence. It also ignores blank pixels. What is your reasoning for using Mander's instead of Pearson's correlation coefficient in your study?

    • Mander’s correlation coefficient describes co-occurrence and is used to determine the fraction of the protein of interest that co-localises with another protein (Arruda et al., 2014. Nature Medicine; Horner et al., 2011. PNAS). Our question was to quantify to what extend do cytokines (using TNFα as proof of principle) localise within a particular organelle and therefore we believe using Mander’s correlation coefficient is appropriate. The reviewer is right in saying that Mander’s calculations ignore blank pixels although Pearson coefficients are also highly dependent on pixels coming from unlabelled regions, which can be an issue when quantifying the presence of cytokines in vesicles and requires the use of a threshold to discern background from relevant signal. We have calculated the Pearson’s correlation coefficient for all the relevant figure panels, which shows similar co-localisation as the Mander’s quantification. This data can be found in new Supplementary Figure 5.*

    Page 8, Discussion. It would be interesting to suggest a mechanism that explains galectin-9-mediated depletion of Vamp-3 protein levels (which is suggestive of transcriptional repression), particularly since experiments suggest that gal9 siRNA treatment did not affect transcription of cytokines.

    We have modified the discussion in page 10 to further speculate on the mechanisms by which galectin-9 may reduce Vamp-3 protein levels. Based on our RNAseq data of WT and galectin-9 depleted moDCs that show no differences in Vamp-3 gene expression, we believe galectin-9 is important for Vamp-3 stabilisation rather than participating in Vamp-3 transcriptional regulation.

    Page 8, Discussion. The statement that the findings from this study is in line with the report that extracellular recombinant galectin-9 enhances IL-6 and IL-8 secretion in mast cells and IL-12 secretion in moDCs is somewhat confusing. Does extracellular galectin-9 cross the cell membrane into the cytoplasm? What is the evidence that it is capable of acting intracellularly when exogenously applied?

    • We regret that this explanation did not come across clearly in the manuscript. We agree with the reviewer that there is no evidence that exogenously added galectins can re-enter the cytoplasm and be functionally active. To avoid any confusion, we have rephrased the text to “in addition, other studies report…”. *

    Minor Comments

    1. Figure 2A needs to have better colors chosen to indicate gal9 KD_LPS and gal9 KD_LPS/BFA+Mon samples, as they show similar colors. *We agree and have modified the figure accordingly. *

    Figure 6D. Vamp-3 immunoreactivity shown in the blot does not appear to diminish on gal9 siRNA treatment as suggested in Fig. 6E. A better representative blot should be shown.

    *For clarification purposes, we have added the quantification of the Vamp-3 representative blot to Figure 6D. To emphasise the differences in Vamp-3 levels upon galectin-9 depletion we have also added the quantification of Snap23 to Figure 6E. *

    Supplementary Figure 1 A and B need a legend to show NT and gal9 siRNA-treated samples.

    • We have now added this legend to Supplementary Figure 1A and B.*

    Supplementary Figure 4, is it possible to have a merged image without galectin 9 (nuclei + TNF) for better clarity of TNF localization relative to nuclei?

    We addressed this point and replaced the previous merged image with one only containing the fluorescent signals corresponding to nuclei (DAPI) and TNF*α. *

    Please explain more explicitly why TNF+ cell % increased in gal9 siRNA-treated cells while secretion trended downward in these cells. Presumably TNF is retained in the GM130+ Golgi apparatus following knockdown of galectin-9 but this is not clearly explained in the text on page 6.

    • We have rephrased the text on page 6 to address the reviewer’s concern. The text has been modified to “Overall, these results demonstrate TNFα is retained in the GM130 positive Golgi complex following galectin-9 depletion thus establishing galectin-9 as essential for cytokine trafficking to the plasma membrane via the endosomal machinery”.*

    Lack of literature/rationale support for the use of CD80, CD86, CD83, HLA-DR as markers for plasma membrane protein trafficking being unaffected by galectin 9 depletion. These require further support to explain their use as good markers for general cellular trafficking.

    Dendritic cells are well-known to upregulate CD80, CD86, CD83 and HLA-DR membrane expression upon maturation (Immunobiology of Dendritic Cells., Banchereau et al., 2000. Annual Review of Immunology; Reis e Sousa., 2006. Nat Rev Immunology*). In resting dendritic cells, these proteins are stored in the endosomal compartment but traffic to the membrane upon activation **and are therefore well-established markers for intracellular transport and membrane re-organisation upon dendritic cell activation (Klein et al., 2005. International Immunology; Baravalle et al., 2011. Journal of Immunology). *We have added a sentence to the results section (page 4) to explain the rationale for choosing these markers.

    Figure 8. Please show cortical actin cytoskeleton in this figure and correct spelling for vesicle in left panel.

    • We have added the actin cytoskeleton to the figure and corrected the spelling mistake. *

    Reviewer #2:

    A brief summary/discussion of the time aspect would also be of interest as the experimental setups are quite complex. The KD after siRNA obviously takes some hours, but after that stimulation with LPS appears to take many more hours to affect VAMP3 distribution in wt cells (Fig. 7). What is it that takes so much time? The time from ER - plasma membrane is more like 30 minutes for a constitutively secreted protein. The RUSH experiment (Fig. 5) also show a relatively fast passage of TNFa out of ER-Golgi in wt cells.

    • We thank the reviewer for raising this point. It takes 36-48 h for the gal9 siRNA to be effective after which galectin-9 levels stay depleted for up to 72 h. We chose 6 h to analyse Vamp-3 redistribution based on our ELISA experiments (Supplementary Figure 2B) that show cytokine secretion peaks 16 h after LPS stimulation. We agree with the reviewer that cytokine gene transcription, translation and protein trafficking also occurs earlier but endogenous intracellular cytokine levels are not high enough to detect them using confocal microscopy. Similarly, the RUSH experiment (Figure 5) was done using an over-expression system in which much higher levels of TNFα are being produced. *

    • To further clarify this we have included a schematic depicting the experimental setup and times (see figure 2 for reviewer). *

    Figure 2 for reviewer. Experimental setup. Monocyte-derived dendritic cells (moDC) are obtained from blood samples and differentiated for 6 days to generate immature dendritic cells (DCs). At day 3 of the differentiation moDCs are transfected with either non-targeting (NT) or gal9 siRNA to deplete galectin-9 protein levels and obtain wild type or galectin-9 knockdown (gal-9 KD) DCs. Six days after isolation, cells were treated with LPS for 6 h (to allow sufficient endogenous cytokine to accumulate intracellularly) prior to being fixed and immunofluorescent experiments (IF) conducted.

    4. Description of analyses that authors prefer not to carry out

    *This section is not applicable. *

  5. Review Commons

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

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    This paper shows that knock down of galectin-9 suppresses secretion of certain cytokines in activated dendritic cells. Then they show that this correlates with failure of the peripheral localization of the SNARE- protein VAMP3 and suggest that this is due to a direct interaction with galectin-9.

    Most of the data consist of comparing wt and galectin-9 KD cells regarding secretion and trafficking of selected cytokines (with focus on TNFa), and of cellular localization of the vSNARE VAMP3. The cells are mainly LPS activated human dendritic cells derived from differentiation on blood monocytes from donors, but differentiated THP-1 cells are also used. These data show convincingly that cytokine secretion is inhibited in galectin-3 KD cells, and for TNFa studied on detail, this is due to failure of post-Golgi trafficking of the, shown in different ways, including a RUSH assay. The trafficking of the vSNARE VAMP3 to the periphery of the cells is also inhibited by absence of galecytin-9, leading to its retention in the Golgi nearer the nucleus. Thus this couples in a surprising the unsolved question of how SNAREs themselves are trafficked to their correct destinations, to the function of a cytosolic galectin.

    The weak part of the paper is the molecular interaction between galectin-9 and VAMP3. This is based on co-immunoprecipitations followed by proteomic characterization and Western blots as summarized in Fig. 6. These data show that galectin-9 and VAMP3 occur in the same precipitated complex, but not that they interact directly. Many other proteins are also in these complexes, and the Western blot data are not very strong. Thus additional experiments would be needed to claim the direct interaction as depicted in Fig.8 for example using purified recombinant proteins, or sharpened focus using mutants of the two interactors.

    A brief summary/discussion of the time aspect would also be of interest as the experimental setups are quite complex. The KD after siRNA obviously takes some hours, but after that stimulation with LPS appears to take many more hours to affect VAMP3 distribution in wt cells (Fig. 7). What is it that takes so much time? The time from ER - plasma membrane is more like 30 minutes for a constitutively secreted protein. The RUSH experiment (Fig. 5) also show a relatively fast passage of TNFa out of ER-Golgi in wt cells.

    Referees cross-commenting

    I agree with all the comments of REviewer 1. Especially major comment 1 and 2 about the confusion of the figures. Major comment 8 is also highlöy relevant, as I also raiused but phrased ion a different way. There is so far no evidence that externally added galectin can reenter the cytosolic compoartment; it is only taken up inside vesicles,

    Significance

    This is highly interesting and novel. Very few papers have coupled SNARE function to galectins before, with key exception from group of Deretic et al regarding secretory autophagy suggesting association of galectin-3 to TRIM16 (not a SNARE, but can be SNARE associated).

    Many follow up questions come up which could have been addressed in this paper or in a future paper. For example, could the Galectin-9 KD phenotype be rescued by added galectin-9 from the outside, as seen for many cases with other galectins? Were any other SNAREs affected?

  6. Review Commons

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

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    Summary

    This paper is focused on the role of galectin-9 in dendritic cells using monocyte-derived DCs . To study the functional characteristics of galectin-9, they depleted galectin-9 with the use of gal9 siRNA. They found galectin-9 is required for TNF-alpha, IL-6, IL-10 and IL-12 secretion. Galectin-9 was found to be involved in TNF trafficking and interacted with Vamp-3 to regulate the release of TNF. They concluded that galectin-9 controls cytokine intracellular trafficking and secretion through functional interaction with the SNARE protein Vamp-3 in DCs through an endosomal compartment.

    The findings seem significant in understanding the role of galectin-9 in dendritic cells which has not been previously explored, and they also expand our understanding of the roles of galectins and their potential function in intracellular cytokine trafficking. However, there are some concerns about the findings and conclusions of the study.

    Major Comments

    1. Page 4, Results. The assertion of 90% downregulation of galectin-9 protein is not substantiated by the data shown in Supplementary Figure 1 (it is not indicated which data, the flow or Western blot, provides the source for this statement). It must be assumed it is from Western blots, as the data from flow cytometry does not show a reduction of 90%. This assertion would be strengthened by density analysis of Western blots shown as a graph beside the blots in Supplementary Figure 1C.
    2. Supplementary Figure 2A. No obvious decrease in TNF, IL-6, or IL-12 was observed in gal9 siRNA-treated cells that were stimulated with LPS and zymosan, and results in this panel are inconsistent with those of Suppl. Fig. 2B. An explanation for this discrepancy should be provided.
    3. The data is sufficient to support the notion that galectin-9 is required for cytokine secretion, but intracellular staining of an additional cytokine (IL-6, IL-10, and IL-12) would be a good addition with controls included (non-transfected cells vs WT vs galectin-9-depleted)
    4. Mander's correlation coefficient is typically not advised for use in co-localization of immunofluorescence since it has been found to downplay associations in low intensity fluorescent staining and favors high intensity and co-occurrence. It also ignores blank pixels. What is your reasoning for using Mander's instead of Pearson's correlation coefficient in your study?
    5. Addition of an immunofluorescent experiment using Vamp-3 and TNF co-localization in gal9 siRNA-treated cells would strengthen observations regarding galectin-9 association with Vamp-3 in immunoprecipitation.
    6. Figure 7A. Vamp-3 does not appear to redistribute towards the cell membrane following LPS stimulation in this figure. Either a different set of images needs to be selected or the text needs to be revised.
    7. Page 8, Discussion. It would be interesting to suggest a mechanism that explains galectin-9-mediated depletion of Vamp-3 protein levels (which is suggestive of transcriptional repression), particularly since experiments suggest that gal9 siRNA treatment did not affect transcription of cytokines.
    8. Page 8, Discussion. The statement that the findings from this study is in line with the report that extracellular recombinant galectin-9 enhances IL-6 and IL-8 secretion in mast cells and IL-12 secretion in moDCs is somewhat confusing. Does extracellular galectin-9 cross the cell membrane into the cytoplasm? What is the evidence that it is capable of acting intracellularly when exogenously applied?
    9. All findings in this study regarding galectin-9 immunoreactivity are dependent on a single goat anti-galectin-9 antibody (AF2045). Findings would be strengthened by the use of a second galectin-9-specific antibody in at least one additional experiment (either immunofluorescence or immunoprecipitation).

    Minor Comments

    1. Figure 2A needs to have better colors chosen to indicate gal9 KD_LPS and gal9 KD_LPS/BFA+Mon samples, as they show similar colors.
    2. Figure 2D needs at least one more experiment before results can be depicted. An n=2 is not sufficient to merit publication.
    3. Figure 6D. Vamp-3 immunoreactivity shown in the blot does not appear to diminish on gal9 siRNA treatment as suggested in Fig. 6E. A better representative blot should be shown.
    4. Supplementary Figure 1 A and B need a legend to show NT and gal9 siRNA-treated samples.
    5. Supplementary Figure 2 has additional small square symbols in panels A and B that should be removed.
    6. Supplementary Figure 2 legend. This legend has repetitive text regarding representative data from one donor. How many donors were tested for these experiments?
    7. Supplementary Figure 4, is it possible to have a merged image without galectin 9 (nuclei + TNF) for better clarity of TNF localization relative to nuclei?
    8. Please explain more explicitly why TNF+ cell % increased in gal9 siRNA-treated cells while secretion trended downward in these cells. Presumably TNF is retained in the GM130+ Golgi apparatus following knockdown of galectin-9 but this is not clearly explained in the text on page 6.
    9. Lack of literature/rationale support for the use of CD80, CD86, CD83, HLA-DR as markers for plasma membrane protein trafficking being unaffected by galectin 9 depletion. These require further support to explain their use as good markers for general cellular trafficking.
    10. Figure 8. Please show cortical actin cytoskeleton in this figure and correct spelling for vesicle in left panel.

    Referees cross-commenting

    I agree with the comments of Reviewer #2 and especially that galectin-9 and VAMP-3 co-IP does not necessarily indicate that they are bound together. I also concur that a galectin-9 rescue experiment would be valuable.

    Significance

    General Assessment.

    Strengths of study: Novel findings showing a role for galectin-9 in regulating cytokine trafficking and release from dendritic cells. This is a new observation which has not be reported before for galectin-9.

    Weaknesses: Some results need verification and additional experiments are required to confirm the findings.

    Advance: There is no existing knowledge for a role for galectin-9 in cytokine trafficking and this study fills this gap in existing published knowledge. The kind of advance that the study makes is fundamental and conceptual.

  7. Version published to 10.1101/2022.05.13.491792v2 on bioRxiv
  8. Version published to 10.1101/2022.05.13.491792v1 on bioRxiv