henhaobio
Sign in

Nature MetabolismCell BiologyPeer-reviewed journal article

Spatially resolved rewiring of mitochondria-lipid droplet interactions in hepatic lipid homeostasis

Sun Woo Sophie KangiD, Lauryn A. Brown, Colin B. Miller, Katherine M BarrowsiD, Jihye L. GolinoiD, Hanyang LiuiD, Constance M. CultraroiD, Daniel FelicianoiD, Mercedes B. Cornelius-Muwanuzi, Kirsten RemmertiD, Jonathan Matthew HernandeziD, Andy D. TraniD, Michael J. KruhlakiD, Alexei V. LobanoviD, Maggie C. CamiD, Natalie Porat‐ShliomiD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

Hepatic lipid accumulation, or Metabolic Dysfunction-Associated Steatotic Liver Disease (MASLD), is a significant risk factor for liver cancer. Despite the rising incidence of MASLD, the underlying mechanisms of steatosis and lipotoxicity remain poorly understood. Interestingly, lipid accumulation also occurs during fasting, driven by the mobilization of adipose tissue-derived fatty acids into the liver. However, how hepatocytes adapt to increased lipid flux during nutrient deprivation and what occurs differently in MASLD is not known. To investigate the differences in lipid handling in response to nutrient deficiency and excess, we developed a novel single-cell tissue imaging (scPhenomics) technique coupled with spatial proteomics. Our investigation revealed extensive remodeling of lipid droplet (LD) and mitochondrial topology in response to dietary conditions. Notably, fasted mice exhibited extensive mitochondria-LD interactions, which were rarely observed in Western Diet (WD)-fed mice. Spatial proteomics showed an increase in PLIN5 expression, a known mediator of LD-mitochondria interaction, in response to fasting. To examine the functional role of mitochondria-LD interaction on lipid handling, we overexpressed PLIN5 variants. We found that the phosphorylation state of PLIN5 impacts its capacity to form mitochondria-LD contact sites. PLIN5 S155A promoted extensive organelle interactions, triglyceride (TG) synthesis, and LD expansion in mice fed a control diet. Conversely, PLIN5 S155E expressing cells had fewer LDs and contact sites and contained less TG. Wild-type (WT) PLIN5 overexpression in WD-fed mice reduced steatosis and improved redox state despite continued WD consumption. These findings highlight the importance of organelle interactions in lipid metabolism, revealing a critical mechanism by which hepatocytes maintain homeostasis during metabolic stress. Our study underscores the potential utility of targeting mitochondria-LD interactions for therapeutic intervention.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (9 Dec 2025) · CC0 1.0

Imaging of mitochondria and lipid droplets (LDs) across mouse liver lobules shows fasting, not Western diet, expands mitochondria-LD contacts; PLIN5 phosphomutants tune contacts and lipid storage.

How hepatocytes handle lipid influx during fasting versus nutrient excess is unclear, partly because organelle remodeling is hard to study in intact tissue. This work profiles mitochondria and LD features per hepatocyte along the periportal-pericentral axis and tests PLIN5 as a mediator of their contacts.

Key results

  1. ShownMitochondria and LD features vary systematically from portal to central vein; hepatocyte-averaged features fall into three spatial groups in PCA, consistent with zones of oxidation versus synthesis and storage.

    18 PP-PC axes from three mice, 12 bins, mean axis length 250 microns with 12-16 cells per axis; PCA clusters R1-R3, R4-R10 and R11-R12; organelle features agreed between 2D and Imaris 3D z-stacks. · Mitochondria and LD topologies mirror the…

  2. ShownFasting greatly increased mitochondria-LD overlap across the lobule, whereas Western diet (WD) gave only a modest rise; overnight fasting of WD-fed mice increased contacts and enlarged LDs.

    Overlap increased in fasted mice in all bins except R11 and R12; WD overlap similar to CNTR pericentral values; fasted WD vs unfasted WD colocalization higher (n = 5 independent experiments). · Increased lipid flux rewires organelle circuits…; Figure legends

  3. SuggestedProteomics of sorted periportal and pericentral hepatocytes suggested zonation was largely preserved during fasting; PLIN5 was upregulated, and ketogenesis rather than ATP production appeared favored.

    4,999 proteins; 1,865 (37%) fed and 1,734 (35%) fasted zonated; PLIN5 doubled by western blot (n=5); beta-hydroxybutyrate rose more than 15-fold; respiration and ATP reduced. · Comparative proteomics of sorted hepatocytes…; Figure legends

  4. SuggestedIn control-diet mice, PLIN5 S155A promoted elongated mitochondria tightly associated with large LDs, while S155E gave fewer, smaller LDs and round mitochondria; the truncated variant produced small LDs.

    AAV8 overexpression; scPhenomics from 12 PP-PC axes from two mice; S155A had the strongest mitochondria-LD overlap across the lobule; S155A raised liver TG in CNTR mice. · Mitochondria-LD contact sites promote LD…; Overexpression of WT PLIN5 in the liver of…; Figure legends

  5. ShownIn WD-fed mice, WT PLIN5 overexpression restored glutathione to control-diet levels and lowered lipid peroxidation, while S155E blunted the WD-induced rise in glucose.

    Liver GSH and MDA, glucose and TG by two-way ANOVA, n = 5 independent experiments; WD TG did not rise beyond a ceiling across variants; body weight, serum FA and cholesterol unaffected. · Overexpression of WT PLIN5 in the liver of…; Figure legends

Methods, data and reproducibility

Cellpose-based segmentation of hepatocytes, mitochondria and LDs in confocal lobule images, binned along the portal-central axis (scPhenomics), combined with TMT proteomics of FACS-sorted periportal and pericentral hepatocytes and AAV8 PLIN5 variant overexpression.

  • Confocal liver sections from mtDendra2 mice (control, fasted, WD) analyzed by scPhenomics · mouse (mtDendra2 excised, male) · n = 18 PP-PC axes from three mice
  • TMT-based proteomics of FACS-sorted periportal and pericentral hepatocytes from fed and fasted mice · mouse · n = four ad-lib-fed and overnight-fasted mice; 4,999 proteins
  • PLIN5 variant overexpression (Null, WT, S155A, S155E, CΔ (1-424)) in control and WD-fed mice, imaging · mouse · n = 12 PP-PC axes from two mice (control diet)
  • Biochemical readouts (TG, GSH, MDA, glucose, serum lipids) after PLIN5 overexpression · mouse (C57BL/6J) · n = n = 5 independent experiments
  • Seahorse respiration assays of sorted periportal and pericentral hepatocytes · mouse · n = four independent experiments
Methods & tools
Cellpose (custom-trained models), scikit-image regionprops, Python 3.10, Imaris 9.9.0, Leica SP8 confocal, FACSAria Fusion, TMTpro 16plex, Orbitrap Eclipse, Proteome Discoverer 2.4 (SEQUEST, Percolator), Limma R package v3.40.6, Seahorse XFe96, AAV8 vector
Reagents & environment
Python v. 3.10; Imaris 9.9.0; Proteome Discoverer 2.4; Limma v3.40.6; GraphPad Prism 9.
Reproducibility gaps
No code repository, trained Cellpose models, or proteomics accession is given in the text; proteomics results are cited only as a supplementary table.
Limitations (5)
  • The direction of fatty acid transfer between mitochondria and LDs was not measured in vivo because of the assay's difficulty, so the storage versus oxidation role of the contacts is inferred. · stated by the authors
  • Mitochondria-LD contacts were not increased in WD-fed mice, which the authors attribute likely to the shorter feeding duration. · stated by the authors
  • 2D sections altered cell size measurements relative to 3D: periportal hepatocytes were smaller in 3D but not in 2D. · stated by the authors
  • PLIN5 variant imaging rests on 12 axes from two mice with only male mice used; axes appear to be the unit, so inter-animal variability is poorly captured. · noted in this summary
  • Redox benefit of WT PLIN5 under WD is not directly tied to contact sites, since no contact-defective control is shown in WD and ER-mediated effects are not excluded. · noted in this summary

Read this if you study hepatic lipid metabolism, liver zonation, or organelle contact sites and want a spatially resolved imaging-plus-proteomics view of PLIN5 function.

Citations
29
FWCI
47.31
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISMiner '22Zhang '26Ross '25Cuñarro '26Garcia-Maci '19Tie '25
75% semantic matchCell Biology
G. E. Miner, C. M. So, W. Edwards+4 more·bioRxiv·Feb 3, 2022