Write a comment

Review coordinated via ASAPbio’s crowd preprint review

This review reflects comments and contributions by Chen Yang, Andreia Faria-Pereira, Madhulika Rai, Joachim Goedhart, Anna Oliveras, and Luciana Gallo. Review synthesized by Joseph Biggane.

This study investigated hepatic lipid accumulation in mice fed a normal diet, fasting mice, and in mice fed a western diet. The authors reported that dietary conditions led to extensive lipid droplet remodeling and changes in mitochondrial topology. The authors reported corresponding changes in PLIN5 expression, a mediator of lipid drop-mitochondrial interactions, using a novel single-cell tissue imaging approach together with spatial proteomics. Further investigation revealed that overexpression of PLIN5 variants, including PLIN5 S1551 and PLIN5 S155E, impacted lipid drop-mitochondria interactions. Specifically, the authors reported extensive organelle interactions in mice overexpressing PLIN5 S155A fed a control diet and reduced steatosis and improved redox state in western diet-fed mice overexpressing PLIN5 S155E.  The findings underscore the importance of organelle interactions in hepatocyte lipid metabolism, with emphasis on the potential implications for understanding metabolic dysfunction-associated steatotic liver disease (MASLD).

Positive aspects of this study:

• This study combined several model systems to produce a very thorough study of the research question.

• The breadth of techniques utilized is impressive, and the authors reported a novel single-cell tissue imaging technique.

• This study presents several novel conclusions about the cellular mechanisms underlying hepatic lipid accumulation, which may lead to novel therapeutic strategies.

Crowd Review Comments:

• Regarding the Abstract:

  • The authors may consider less emphasis on naming specific mutations in the abstract and focus on explaining the biology for greater clarity.

• Regarding the Introduction:

  • It would be helpful to add a scheme or complement the existing scheme in Fig. S1A elucidating the localization of pericentral hepatocytes (close to CV) and periportal hepatocytes (close to PV).

• Regarding the Results:

  • In subsection ‘Mitochondria and LD topologies mirror the spatial zonation of hepatic lipid utilization and synthesis’:

    • Following the sentence “The average LD area gradually increased from bins 3-9 before slightly decreasing in pericentral regions, where LD numbers were the highest (bins 10-12; Fig. 1G).”, a summary would be appreciated indicating that LD increase in size but not in number through the PP-PC axis, and that hepatocytes directly close to PC the LDs are much more abundant but smaller.

• Following the sentence “Similarly, LD features were also zonated along the PP-PC axis in both 2D and 3D datasets (Fig. S1E)”, in 3D analysis, the initial cells (1 and 2) present similar LD volumes and counts than latter cells (11, 12) (Fig. S1E), which is not reflected in 2D data (Fig. 1 G). The authors should include and discuss this discrepancy in the text.

• In subsection ‘Increased lipid flux rewires organelle circuits in the zonated liver’:

  • In reference to “...recapitulates spatial signatures associated with the functional dichotomy…”, the authors should consider elaborating more on how the spatial signatures correlate with function, or provide further evidence, to support this claim.

• In reference to “...scPhenomics workflow analysis (Fig. 2A-C)...”, there was confusion about what the new workflow analysis includes. It seems to be referring to Figures 2 and 3 altogether as an entire analysis scheme. It would be useful to show a step-by-step general diagram in a supplementary figure to clarify the analysis being proposed. This will help others to reproduce it.

• In reference to the sentence “Mitochondria in periportal (PP) and mid-lobular regions appeared longer and thicker (Fig. 2B, insets)”, it would be useful to mention if there is mitochondrial fusion. The images seem to show mitochondrial fusion.

• In reference to the sentence ”Notably, long and thick mitochondria frequently wrapped around LDs, forming extensive contact sites between the two organelles (Fig. 2B, insets).”, are contact sites between LDs and mitochondria only evaluated by fluorescence signal overlap? There was skepticism about this approach and it was suggested that the authors could consider using specific fluorescent probes, such as split construct systems.

• In reference to the sentence “The WD-fed mice had more spherical mitochondria with reduced LD interactions (Fig. 2C).”, in this Figure, no quantification is presented to sustain this claim. In Figure S3, the authors provide a quantification of the colocalizing mito-LD pixels, providing evidence for this sentence. Therefore, the authors should refer to Fig. S3 here.

• In reference to the sentence “Fasted WD-fed mice showed notable changes in mitochondrial morphology and a significant increase in mitochondrial interactions with LDs compared to unfasted WD-fed mice.”, it is not clear from the images presented that mitochondrial morphology is changed between both Western diet (WD) conditions (not fasted vs fasted). In both conditions, mitochondria are round and dispersed throughout the hepatocytes. The authors should clarify this statement and provide mitochondria circularity quantification.

  • Another comment suggested that it is not clear from the images that there is a change in mitochondrial morphology. Additionally, with previous data, it seems like the mitochondrial-LD interaction sites have longer mitochondria whereas that is not seen here. Can this even be called interaction in Fig S3?

• In reference to the sentence “Notably, LD size increased, with larger droplets observed in fasted WD-fed mice, suggesting that mitochondria-LD contact sites may facilitate LD expansion (Fig. S3).”, LD expansion could be an independent metabolic phenotype. There is not enough evidence to suggest that mitochondria-LD contact sites facilitate LD expansion. The authors should provide more clarity on this statement.

• In subsection ‘Comparative proteomics of sorted hepatocytes from fed and fasted mice’:

  • In reference to the sentence “However, unlike cultured cells, mitochondrial respiration and ATP production under these conditions was significantly reduced (Fig S4D-F).”, it would help to provide a short introduction that they are not only using proteomics but also live bioenergetic measurements.

• In reference to the sentence “Notably, PLIN5, a LD- associated protein that recruits mitochondria to form mitochondria-LD contact sites, was significantly upregulated in the fasted liver (Fig. 4E).” " it would support their hypothesis to show that PLIN5 is indeed located in the areas of LD-mito overlap by IHC.

• In subsection ‘Mitochondria-LD contact sites promote LD expansion in fed mice’:

  • In reference to the sentence “Overexpression of the truncated PLIN5 CΔ (1-424) variant, which lacks the mitochondria binding domain, did not affect mitochondrial morphology but led to the formation of small LDs throughout the lobule, likely due to a dominant negative effect on lipolysis 19. Lastly, when comparing the ability of PLIN5 variants to promote mitochondria-LD interactions, PLIN5 S155A demonstrated the most pronounced effect across the entire lobule (Fig 5E)”, the PLIN5 overexpression was performed in a WT  background where PLIN5 continues to be expressed. The authors propose that the phenotype of CΔ (1-424) PLIN5 variant could result from a dominant negative effect. However, couldn't this also be the underlying cause for increased LD counts and density under PLIN5-S155A variant?

    • A subsequent comment suggested that the size of LDs in all conditions except S155A looks similar. Although, quantification supports their claim, better representative images could be helpful here.

• In reference to the sentence “Overexpression of WT PLIN5 in WD-fed mice noticeably affected mitochondria length, similar to that observed in the CNTR diet-fed mice. However, under these conditions, mitochondria-LD contact sites appeared more frequent than in CNTR diet-fed mice…”, similar quantification to Fig. 5C, D and E should be presented for WD-fed mice.

• In reference to the sentence “Interestingly, WD-fed mice overexpressing WT PLIN5 had reduced glutathione levels comparable to those observed in mice on the CNTR diet (Fig 6D).” " it would be helpful if the authors could add a sentence explaining the rationale for assessing glutathione and lipid peroxidation levels.

• In reference to the sentence “Interestingly, WD-fed mice overexpressing WT PLIN5 had reduced glutathione levels comparable to those observed in mice on the CNTR diet (Fig 6D).” " the authors should revise this sentence, as glutathione levels in WD-fed mice overexpressing WT PLIN5 are not different from CNTR-fed mice. Glutathione levels are indeed shown to be reduced in WD-fed "PLIN5 null".

• In response to the sentence “The overexpression of PLIN5 variants had no effect on body weight, serum FA levels, and cholesterol (Fig S7B-D).” " it would be helpful to explain how glutathione levels and lipid peroxidation are measures of redox state at the end of this section.

• Regarding figures:

  • Figure 1E:

    • In this figure, it is not clear what M1, M2, and M3 refer to. Perhaps, the figure is referencing individual mice. Nevertheless, this could use clarification.

• Additionally, as the authors are comparing more than two groups in these graphs, the authors should revise the statistical test applied here (t-test).

• Figure 1F

  • The authors should consider reformatting this, and similar, figures, where the y-axis is altered to start at 0. For example, the mitochondria circularity graphs range from 0 to 1, and only part of the scale is represented (e.g. Fig. 5D).

• Figure 3

  • Pearson’s correlation coefficients are difficult to interpret. They could be shown in a supplementary manner. To simplify, the authors might consider the use of the color index for this figure.

• Regarding the Discussion:

  • The following article was mentioned with the comment below: https://www.nature.com/articles/s41598-025-87871-2

    • This recently published article overlaps quite a lot with the work in this preprint.  

• In response to the sentence “Our findings highlight the utility of scPhenomics in identifying spatially relevant phenotypes and its potential to uncover novel adaptive mechanisms driving liver disease when integrated with other single-cell methods”, the scPhenomics is a novel approach for sc phenotypic profiling that promises to be a great contribution to the field and future diagnostic tools, but the authors should consider further documentation and a concise explanation of the pipeline.  The authors might also consider publishing their code in open repositories. 

• In response to the sentence “Examining mitochondria-LD interactions in the zonated liver may also play a role in hepatocellular carcinoma”, it is difficult to understand what the authors mean by this sentence.

• Regarding the Materials and Methods:

  • Subsection ‘Confocal microscopy and scPhenomics’:

    • The authors may consider detailing the training of the Cellpose model or sharing the Cellpose model altogether.

Conflicts of interest:

• None declared

Competing interests

The authors declare that they have no competing interests.