Intravascular Hepatocyte Lipid Droplets Are Commonly Present in Hepatic Sinusoids, Portal Venules, and Central Veins in MASH

In metabolic dysfunction-associated steatotic liver disease with severe steatosis, hepatocytes are characterized by very large lipid droplets (LDs), which expand the cell and take up a large proportion of the cytoplasm. When steatotic hepatocytes undergo necrosis, the cell membrane ruptures, releasing LDs into the extracellular space. Kupffer cells (resident liver macrophages) surround and phagocytose these extracellular LDs, forming crown-like structures.¹ However, in cases of profound hepatocyte steatosis and necrosis, the rate of LD release may overwhelm the capacity of Kupffer cells to process them by phagocytosis. We aimed to determine whether such extracellular LDs can leak into adjacent blood vessels including sinusoids, portal venules, and central venules.

Clinical data, histological slides, and liver tissue from 6 patients with histologically confirmed metabolic dysfunction-associated steatohepatitis (MASH) were derived from a biorepository and prospective cohort study at Veterans Affairs Puget Sound Health Care System (VAPSHCS).1, 2, 3 The study was approved by the institutional review board at VAPSHCS.

Liver tissue was harvested from 13 male C57Bl6/J mice fed a high-fat, high-cholesterol diet for 6 months, after which they developed fibrosing steatohepatitis, as we demonstrated in multiple prior studies.^4,5 Experimental procedures were approved by the Institutional Animal Care and Use Committee of VAPSHCS.

Traditional formalin-fixed, paraffin-embedded specimens do not allow direct visualization of LDs because lipids are lost during the dehydration and clearing steps. What is usually indirectly observed is the “ghost” of the LD counterstained by the surrounding cytoplasm. However, if LDs are released after cell death into the extracellular space and leak into adjacent portal venules or sinusoids, they would be invisible by regular staining techniques that do not fix and stain lipids.

We used a method of both fixing and staining for the lipids with osmium tetroxide to allow visualization of LDs together with appropriate counter stains, as follows. Samples of human or mouse liver tissue were placed in Trump's fixative (4% formaldehyde and 1% glutaraldehyde in a phosphate buffer) for 24 hours, and then small pieces (∼2 mm) were submerged for 1 hour in 1% osmium tetroxide for lipid staining and fixation. Osmium tetroxide binds at the carbon-carbon double bonds of unsaturated fatty acids in triglycerides and cholesterol esters. After washing, samples were embedded in methacrylate plastic resin, which enables very thin sectioning. Blocks were sectioned at 1 μm thickness, counterstained with methylene blue or Masson’s trichrome, and viewed with a light microscope. Thinner sections (<100 nm) were prepared for transmission electron microscopy. We also sectioned frozen liver tissue embedded in optimal cutting temperature and stained sections with osmium tetroxide.

We identified LDs within portal venules in 4 out of the 6 human liver MASH specimens stained with osmium. Figure 1a–d shows many, large LDs in the middle of blood vessels, clearly identified as intravascular by the presence of adjacent or adherent red blood cells (RBCs), with diameters as large as 10–30 μm (ie, more than an order of magnitude larger than the biggest chylomicrons)—see also Figure A1a for more examples from different patients. Multiple large LDs (identified by yellow stars in panel a) appear to be in the process of being released from hepatocytes into an adjacent sinusoid. Figure 1e shows clusters of small LDs along the walls of sinusoids (also identified by RBCs), which was also a commonly observed pattern (Figure A1b). Human specimens fixed in osmium and counterstained with Masson’s trichrome revealed perisinusoidal (“chicken-wire”) fibrosis typical of MASH, with intravascular LDs as well as LDs that were within hepatocytes directly adjacent and projecting into sinusoids and portal venules (Figure 1f and Figure A2). Transmission electron microscopy confirmed the presence of LDs within sinusoids, surrounded by adherent RBCs (Figure A3a). Staining for perilipin 2, the major LD protein in hepatocyte,⁶ also confirmed that these were intravascular LDs (Figure A3b).

Figure 1.

Human liver sections of four patients with MASH fixed and/or stained with osmium tetroxide (black) and counterstained with eosin (a–e) of Masson trichrome (f) demonstrating evidence of black, osmium-stained LDs in portal venules and sinusoids. Footnote: Panels (a–d) show black, osmium-stained intravascular LDs (red arrows) in the middle of portal venules and sinusoids, identified as intravascular by the presence of adjacent or adherent pink, eosin-stained RBCs (orange stars). The sections show innumerable, identical-looking black LDs within hepatocytes. Multiple large lipid droplets (identified by yellow stars in panel a) appear to be in the process of being released from hepatocytes into an adjacent sinusoid. Panel (e) shows clusters of small LDs (blue arrows) along the walls of sinusoids, also identified by adjacent RBCs (orange stars); also hepatocyte LDs that project into and distort adjacent sinusoids (white arrows). Panel (f) shows perisinusoidal fibrosis with Masson’s trichrome (blue) typical of MASH and hepatocyte LDs that project into and distort adjacent sinusoids (white arrows).

In mice fed a high-fat, high-cholesterol diet, we identified LDs within portal veins and sinusoids in liver sections from 11 out of 13 the liver sections that had identifiable portal veins. Figure 2 shows many LDs in the middle of a portal vein (identified by adjacent hepatic artery and bile duct (Figure 2a and d), with large numbers of adherent RBCs (Figure 2c and e)—see additional examples in Figure A4a. Figure 2f shows an intravascular LD (green arrow) that appears to be embolizing and obstructing a sinusoid. Mice fed a high-fat diet (without any additional cholesterol) developed severe steatosis without steatohepatitis and also had intravascular LDs within portal veins and sinusoids, but not mice on a normal diet (Figure A4b). Multiple LDs were also visualized in a central vein (Figure A4b).

Figure 2.

Mouse liver sections from mice fed an HFHC diet, which develop fibrosing steatohepatitis, fixed, and/or stained with osmium tetroxide demonstrating evidence of brown-black, osmium-stained intravascular LDs in portal venules and sinusoids. Footnote: Panels (a–d) show liver tissue fixed in osmium tetroxide, sectioned, and counterstained with methylene blue. A large portal vein (PV) is demonstrated with a nearby hepatic artery (HA) and bile duct (BD). Multiple, brown osmium-stained LDs (red arrows) are seen within the portal vein, with adjacent or adherent RBCs (orange stars)—as well as innumerable, identical-looking LDs within hepatocytes. Panels (e and f) show frozen liver sections stained with osmium tetroxide, again demonstrating intravascular brown-black, osmium-stained LDs (red arrows) with adherent RBCs (orange stars). Panel (f) shows an LD (green arrow) that appears to be embolizing and obstructing a sinusoid. HFHC, high-fat, high-cholesterol.

In conclusion, our findings suggest that release of intact hepatocyte LDs into adjacent sinusoids, portal venules, and central veins (ie, intravascular LDs) after hepatocyte necrosis is common in human and experimental MASH. Intravascular LDs were noted to embolize small sinusoids or to cluster along the walls of sinusoids and were frequently surrounded by adherent RBCs. This extends prior reports of large LDs being released into the space of Disse and subsequently into the sinusoidal lumen.⁷

The presence of intravascular LD within sinusoids, portal venules, and central veins may have several pathophysiological consequences. At the local hepatic level, sinusoidal obstruction may occur when lipid emboli impede blood flow through the sinusoids, potentially leading to increasing sinusoidal pressure (including early portal hypertension, an increasingly recognized contributor to fibrosis development in MASH), localized hypoxia, and further hepatocellular injury.⁸ Furthermore, an inflammatory response may occur due to the exposure of lipid material to the sinusoidal environment, contributing to subsequent stellate cell activation, fibrosis, and progression to fibrosing MASH.⁶

The presence of LDs within central veins means that LDs have access to the systemic circulation and could potentially exert systemic effects. These LDs may represent a previously unrecognized pathway of hepatic lipid export following hepatic necrosis, distinct from classical lipoproteins. Once in the circulation, these LDs could interact with circulating immune cells, serve as substrates for lipoprotein remodeling, or deliver oxidized lipids and cholesterol esters that promote systemic inflammation, insulin resistance, or atherogenesis.