Insulin resistance induced hepatic lipid deposition in db/db mice
To establish hepatic lipid deposition, we used the db/db mice model, which harbors a mutation in the diabetes (db) gene encoding the leptin receptor (ObR) and exhibits impaired leptin signaling closely related to liver lipid metabolism20. After eight weeks of ad libitum feeding with a standard diet, the db/db mice exhibited a significant increase in body weight (Fig. S1a) and blood glucose (Fig. S1b), alongside a marked decrease in insulin sensitivity (Fig. S1c, d). This indicates that the db/db mice have a serious disturbance in glucose and lipid metabolism, which leads to an abnormal increase in blood lipid levels. Studies have shown that after binding insulin in the liver, the insulin receptor autophosphorylates and goes on to phosphorylate several insulin receptor substrates, thereby activating classical IRS signaling pathways, including the phosphatidylinositol 3-kinase (PI3K)-phosphatidylinositol-dependent kinase (PDK)-protein kinase B (AKT) signaling pathway21. However, we found that protein levels of p-PI3K/PI3K and p-AKT/AKT were significantly lower in the livers of diabetic mice than in the livers of db/m mice, as detected by western blotting (Fig. 1a–h). This resistance leads to decreased glucose utilization, increased fatty acid uptake, and more severe hepatic lipid deposition.
Hepatic free fatty acids (FFA), TG, and liver indices of the db/db mice were significantly increased compared with the db/m group (Fig. 1k–m). The livers of db/m mice were bright red with sharp edges. By contrast, the livers of db/db mice had increased volume, a greasy texture, a lighter color, and blunted edges. Hematoxylin and eosin (H&E) staining showed that db/m mice had normally structured hepatic tissue with well-arranged cells, uniformed cytoplasm, and central nuclear position, whereas db/db mice had hepatic tissue with a disordered structure, with ballooned hepatocytes, numerous vacuoles, and extruded nuclei. Oil Red O staining revealed that the numerous big vacuoles in db/db mice hepatic tissue nearly filled up the entire cytoplasm (Fig. 1i, j) and were probably red-stained lipid droplets. Similarly, we found that in the db/db mice, the levels of serum cholesterol (CHOL), TG, high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and FFA were significantly increased (P < 0.05) compared to those in the db/m group (Table. 1). These findings further support the presence of increased lipid accumulation in the liver of the db/db mice compared to the wild-type group. In summary, our observations demonstrate that the liver of db/db mice exhibited significant insulin resistance, enhanced lipid deposition, and a substantial increase in the size and number of lipid droplets.
Table 1
Blood lipoprotein levels (mmol/L)
Group | db/m | db/db |
CHOL | 1.97 ± 0.45 | 3.93 ± 0.87** |
TG | 0.87 ± 0.28 | 1.74 ± 0.72* |
HDL-C | 1.10 ± 0.23 | 2.15 ± 0.43** |
LDL-C | 0.18 ± 0.06 | 0.43 ± 0.13** |
FFA | 1034 ± 480.6 | 1939 ± 363.9* |
Data are mean ± SEM.*P < 0.05, **P < 0.01 vs. db/m. CHOL: total cholesterol; TG: triglyceride; HDL-C: high-density lipoprotein cholesterol; LDL-C: low-density lipoprotein cholesterol; FFA: free fatty acid |
Increased LD-Mito contacts in isolated primary hepatocytes and fat layer in db/db mice liver
In previous studies, researchers used LipidTOX or BODIPY and Mito-tracker as fluorescent dyes for staining lipid droplets and mitochondria to demonstrate their interaction in primary adipose cells18. These interactions have been demonstrated in fat layers (FL) from rat livers17, but not yet in hepatocytes. We isolated primary hepatocytes by collagenase perfusion on livers22. and used BODIPY and Mito Tracker to stain lipid droplets and mitochondria and look for contact between them. In normal hepatocytes, the rod-shaped mitochondria were distributed uniformly throughout the cytoplasm, among which a few small lipid droplets were submerged, with no obvious contacts between the lipid droplets and mitochondria. By contrast, in db/db mice hepatocytes, some of the big lipid droplets compete with mitochondria for space within the cytoplasm, part of which was surrounded by rod-shaped mitochondria (Fig. 2a, b).
To characterize the features of LD-Mito contacts in livers in both db/m and db/db mice, hepatic tissues were milled and centrifuged at 900 xg to separate the lipid layer (FL)according to Talari’s work17. Detection of the lipid droplet internal marker protein Plin2 reveals more lipid droplets in the lipid layer isolated from the livers of db/db mice (Fig. 2c). The fluorescence in the lipid droplets and mitochondria of the lipid layer demonstrated that, in normal mice livers, the lipid droplets were of relatively uniform size, and many of them were attached to mitochondria. In db/db mice livers, lipid droplets were of varying sizes, with many giant lipid droplets among them. Furthermore, the number of LD-Mito contacts was significantly higher in the livers of db/db mice than in the livers of db/m mice (Fig. 2d, e). These phenomena were verified by TEM results that show that in normal mice livers, lipid droplets were scattered in the cytoplasm with few mitochondrial contacts, while in db/db mice livers, there were many large lipid droplets with a significant increase in mitochondrial contacts. TEM also revealed that the length and curvature of mitochondria bound to lipid droplets increased (Fig. 2f–i).
Considering that the elongated deformation of mitochondria is often accompanied by fusion and division, we tested the roles of key proteins Mfn2 and Drp1 in these processes. Mfn2, an integral component of the outer mitochondrial membrane, is instrumental in facilitating mitochondrial fusion23. Drp1 forms a ring-like structure when recruited to the outer mitochondrial membrane23, that constricts and eventually splits the mitochondria in the extracted fat layer and CM. The expression of Drp1 in CM of db/db mice decreased but the expression of Mfn2 in FL increased, indicating that mitochondrial fusion plays an important role in LD-Mito contact (Fig. 2j–l).
Morphological characteristics and status of isolated CM in normal and db/db mice livers
CM were extracted at a high centrifuge speed and stained with BODIPY and Mito-tracker. To our surprise, we found sporadic small lipid droplets in CM. We hypothesized that there may be lipid droplets in CM tightly bound to mitochondria that are so small that the fluorescence is easily quenched and difficult to observe (Fig. 3a, b). We therefore used Oil Red O and Janus Green B staining to reveal small LD-Mito contacts under bright field microscopy. As expected, a small amount of lipid droplets were drowned out in green mitochondrial signal in CM extracted from normal mice livers, while large amounts of green mitochondria were dotted with many small red lipid droplets in CM extracted from db/db mice livers (Fig. 3c, d). Under both conditions, no big lipid droplets were observed.
To further characterize the relationship between CM and lipid droplets, western blot was performed to quantify the lipid droplet marker protein Plin 2 in CM (Fig. 3g). These results confirmed that lipid droplets do exist in extracted CM and were richer in db/db mice liver. We then examined the differences in mitochondrial status and LD-Mito contact in the CM of both groups through Mitochondrial membrane potential assay kit with JC-1 and reactive oxygen species (ROS) tests (Fig. 3e, f, h, i). This revealed significant alterations in the hepatic lipid metabolism of db/db mice, characterized by an increase in the size and quantity of lipid droplets. Intriguingly, these lipid droplets were found in direct contact with mitochondria, with a tendency for larger droplets to form loose connections and smaller droplets to be tightly enveloped. Furthermore, the CM in lipid-accumulated livers demonstrated decreased activity and increased production of ROS compared to those in normal mice. These observations suggest a potential link between lipid accumulation, mitochondrial dysfunction, and oxidative stress, potentially contributing to the pathogenesis of conditions such as NAFLD. Further investigations are warranted to elucidate the underlying mechanisms and their implications for liver disease progression and treatment strategies.
Characteristics of two different LD-Mito contacts in isolated fat layers and CM in db/db mice livers
LD-Mito contacts exist in both fat layers and CM, which were dominated by large lipid droplets and very small droplets, respectively. Consequently, there may be two different models of LD-Mito contacts depending on the size of lipid droplets. Fat layers and CM from db/db mice livers were subjected to fluorescence staining to study the characteristics of these two possible types of LD-Mito contacts, as LD-Mito contacts were inadequate for study in normal mice livers. In the fat layer, the average area of lipid droplets in contact with mitochondria was 15 µm2 and the average diameter was 3.5 µm, with a large range in diameters (0.3–19.7 µm). In the CM, the average area of lipid droplets in contact with mitochondria was 0.3 µm2 and the average diameter was 0.14 µm. These small lipid droplets were uniform in size, with the largest no more than 2 µm in diameter (Fig. 4b-e). Altogether, more than 65% of the lipid droplets in the fat layer were larger than 2 µm in diameter, whereas the lipid droplets in the CM were smaller than 2 µm in diameter.
According to Talari’s work, the extracted fat layer was centrifuged at 10,400 xg to separate out the mitochondria, called PDM, and identify LD-Mito contacts17. We also extracted PDM and stained lipid droplets and mitochondria with BODIPY and Mito-tracker (Fig. 4a, b). To our surprise, we still found trace small lipid droplets in purified PDM. In contrast to the lipid droplets in CM that are surrounded by CM alone, the small lipid droplets in PDM stick together with mitochondria and form clumpy complexes resembling large lipid droplets surrounded by mitochondria at a density that floats in the fat layer under low centrifugation speeds (Fig. 4a, b). These small lipid droplets bind mitochondria so tightly in PDM that not even high-speed centrifugation can separate them. Interestingly, although these clumpy complexes filled with small lipid droplets and mitochondria were rare in PDM, TG analysis revealed that the total lipid droplet content was even higher in PDM than in CM (Fig. 4f). Therefore, we used TEM to observe the morphology of CM and PDM and found that the small lipid droplets were tightly wrapped by elongated mitochondria in CM, whereas the small lipid droplets in PDM were wrapped by multiple mitochondria surrounding by endoplasmic reticulum (ER) (Fig. 4i, j). We then used JC-1 and ROS tests to examine the oxidative status of these two types of mitochondria and found that membrane potential levels were significantly higher in PDM than in CM, though ROS levels were decreased in PDM (Fig. 4g, h).