Light microscopic and transmission microscopic findings in MASLD model mice

Liver tissue from CDAA-diet mice bred for 8 weeks showed hepatocellular ballooning, liver fibrosis, and inflammatory cell infiltration along with fatty deposits and were classified as Type 4 according to the Matteoni classification (Fig. 1a–c). In addition, this mouse model was considered to meet the diagnostic criteria for MASLD because it had fatty liver and insulin resistance, although it did not show obesity [14]. Figure 1d shows mitochondrial bilayer and cristae surrounded by rER and glycogen granules. TEM revealed circularly expanded and deformed mitochondria at 4 weeks after feeding CDAA diet (Fig. 1e). Some internal cristae were shortly truncated and formed autophagosomes. After another 8 weeks, an increase in the number of degenerated mitochondria, irregular expansion of cristae, and irregular bilayer membranes were observed (Fig. 1g). Autophagosomes were formed to envelop these structures. rER was vacuolated and exhibited an abnormal morphology (Fig. 1f). Glycogen granules in the cytosol decreased as the disease progressed. In some cases, the membrane structure was disrupted. The surrounding glycogen granules were reduced, and rER was open.

Fig. 1

Microscopic images of liver tissue from MASLD model mice. Asterisk mitochondria; arrow, rER endoplasmic reticulum. a, d Normal-diet mice (8 weeks), b, e CDAA-diet mice (4 weeks), c, f, g CDAA-diet mice (8 weeks)

Liver tissue from HF-diet and Ath+HF-diet mice also showed hepatocellular ballooning, liver fibrosis, and inflammatory cell infiltration, along with fatty deposits, and were classified as Type 4 according to the Matteoni classification (Supplementary Fig. 1a, c). In addition, this mouse model was considered to meet the diagnostic criteria for MASLD because it had fatty liver and insulin resistance [16, 17]. LM showed that ballooning, inflammation, and fatty liver were more severe in Ath+HF-diet mice than in HF-diet mice. TEM showed that Ath + HF-diet mice exhibited severe abnormalities in mitochondria, glycogen granules, and rER, similar to those observed in CDAA-diet mice at 8 weeks (Supplementary Fig. 1d). In contrast, HF-diet mice showed only mild mitochondrial abnormalities, and their glycogen granules and rER were almost normal like those in normal diet-fed mice and CDAA-diet mice at 4 weeks (Supplementary Fig. 1b).

Patient profiles

The patient backgrounds and hematological findings are shown in Table 1. Based on the diagnostic criteria for MASLD, 40 patients were classified into two groups: 11 in the non-MASLD group and 29 in the MASLD group. Of the non-MASLD patients, six were males and five were females. The median age of the patients was 71 years. The mean BMI was 23.5 kg/m2; TG was 106 mg/dL; HDL-C was 58 mg/dL; fasting blood sugar (FBS) was 115 mg/dL; alanine aminotransferase (ALT) was 36 U/L; and gamma glutamyl transpeptidase (γ-GTP) was 44 IU/L. Only one patient had CK18 > 260 U/L, which is the cut-off for diagnosis of steatohepatitis, and no patient showed liver steatosis in imaging studies. Of the patients with MASLD, 16 were males and 13 were females. The median age of the patients was 60 years. The mean BMI was 27.2 kg/m2; TG was 160 mg/dL; HDL-C was 45 mg/dL; FBS was 139 mg/dL; ALT was 56 U/L; and γ-GTP was 78 IU/L. Five patients had CK18 > 260 U/L and all patients showed liver steatosis in imaging studies.

Table 1 Patient characteristics and hematological findingsLight microscopic findings in human samples

Table 2 shows the results of LM. The diagnoses in the non-MASLD group were as follows: normal in four cases, HBV (Hepatitis B virus)-related liver disease in three cases, HCV (Hepatitis C virus)-related liver disease in three cases, and other liver disease in one case. In this group, the Matteoni classification and all NAS components were normal. In contrast, all cases in the MASLD group exhibited findings consistent with SLD. In the Matteoni classification; five were diagnosed with Type 1, two with Type 2, one with Type 3, and 21 with Type 4. The breakdown of the NAS for steatosis, inflammation, and ballooning are listed in Table 2. Nine patients were classified as 0–2, 15 as 3–4, and five as ≤ 5.

Table 2 Light microscopic findingsTEM findings of mitochondria

Mitochondria are intracellular organelles with a size of approximately 2 μm and exhibit a characteristic double membrane and gourd-shaped morphology. The classification of mitochondria was based on preserved morphology, characteristic double membrane structure, and distinct internal cristae. The cells with spherical morphology, ruptured internal cristae, autophagosome formation, or decreased number of mitochondria were classified as abnormal and weighted in this order (Fig. 2a–c).

Fig. 2

Classification of TEM findings by intracellular organelles. Asterisk mitochondria; inside dotted line glycogen granules; arrow rER. As shown below, 0 was defined as normal, 1 as mildly abnormal, and 2 as highly abnormal. TEM findings of liver tissue are summarized in Table 2. a–c TEM of mitochondria. a Clear membrane/normal cristae = score 0, b Round shape/ruptured cristae = score 1, c Autophagosome formation/decrease in number = score 2, d–f TEM of glycogen granules. d Rich = score 0, e Decrease in number = score 1, f Complete loss = score 2, g–i TEM of rER. g Normal = score 0, h Expansion/degeneration = score 1, i Dissolution/spherical shape (loss) = score 2

TEM findings of glycogen granules

Glycogen granules that filled the cytosol were considered normal, while those with decreased numbers and replaced by cavities were considered abnormal (Fig. 2d–f).

TEM findings of rER

rER with a parallel-layered structure was considered normal, while that with an open lumen and wavy deformations was considered abnormal. The extremely small or laminar structures that changed to spherical shape were considered as abnormalities (Fig. 2g–i).

TEM findings in human samples

Mitochondria were abnormal in 27 of 40 cases, with a score of 1 in 15 cases and 2 in 12 cases. The mitochondrial abnormalities were found in 4 of 11 non-MASLD individuals, with a score of 1, and most patients with MASLD (23 of 29) had abnormal mitochondria; 12 of the 23 patients had severe abnormalities with a score of 2. Glycogen granules were abnormal in 22 of 40 samples, and 7 of 22 samples had severe abnormalities with a score of 2, with all cases being MASLD. The non-MASLD patients did not have abnormal glycogen granules. Furthermore, 22 of 29 patients with MASLD had abnormal glycogen granules. Of the 40 patients, 25 had abnormal rER, with 11 having a score of 1, and 14 with severe abnormalities had a score of 2. Abnormalities of rER were identified in 4 of 11 non-MASLD patients, and 21 of 25 patients with MASLD exhibited abnormal rER morphology. The breakdown of these scores is shown in Table 3. Example of TEM findings in human samples is shown (Supplementary Fig. 2). In liver tissue from non-MASLD patients, there are numerous mitochondria with intact membrane structures, and visible cristae. Glycogen granules are abundant in the cytoplasm and the rough endoplasmic reticulum (rER) is not expanded and remains parallel. In liver tissue from MASLD patients, as seen under low magnification, the number of mitochondria is decreased, and under high magnification, the membrane structure is disrupted, and the internal cristae are no longer visible. Glycogen granules are reduced and the rER surrounding the mitochondria is expanded and degenerated. In liver tissue from MASLD patients, fat droplets are observed under low magnification.

Table 3 Electron microscopic findingsComparison of TEM findings

TEM findings of liver tissues between the non-MASLD and MASLD groups were compared using the Mann–Whitney U test. Significant differences in mitochondria (p = 0.003), and glycogen granules (p < 0.0001) were observed between the two groups (Fig. 3a, b). However, no significant difference was observed in rER between the two groups (Fig. 3c). Furthermore, the MASLD group was divided into two subgroups, non-metabolic dysfunction-associated steatohepatitis (non-MASH) (n = 6) and metabolic dysfunction-associated steatohepatitis (MASH) (n = 23) groups based on the presence or absence of hepatitis findings, and an analysis was conducted. The non-MASH group corresponded to simple steatosis without inflammation or fibrosis. Comparison between the non-MASLD and non-MASH groups showed no significant differences in the scores of TEM findings (Fig. 3d–f). In contrast, comparison between the non-MASLD and MASH groups showed significant differences in the scores for mitochondria (p = 0.0024), glycogen granules (p < 0.0001), and rER (p = 0.0322). Additionally, comparison between the non-MASH and MASH groups showed significant differences in the scores for glycogen granules (p = 0.0409) and rER (p = 0.0291); however, no significant difference was observed in the scores for mitochondria.

Fig. 3

Comparison of TEM findings. a–c Comparison of TEM findings between the non-MASLD and MASLD groups. d–f Comparison of TEM findings between among the non-MASLD, non-metabolic dysfunction-associated steatohepatitis (non-MASH), and metabolic dysfunction-associated steatohepatitis (MASH) groups. The MASLD group was divided into two subgroups, non-MASH (n = 6) and MASH (n = 23) groups based on the presence or absence of hepatitis findings. The non-MASH group corresponded to simple steatosis without inflammation or fibrosis

Analysis of the association between TEM and LM findings

Next, we compared the TEM findings of human liver tissue with each item of the NAS (steatosis, inflammation, and ballooning) and the total NAS and used Spearman’s correlation coefficient to determine the correlation between these indices (Fig. 4). The mitochondrial abnormalities were strongly correlated with steatosis (r = 0.585, p = 0.0000727), inflammation (r = 0.603, p = 0.0000376), and ballooning (r = 0.693, p = 0.000000719621) (Fig. 4a–c). Abnormalities in glycogen granules were strongly correlated with steatosis (r = 0.656, p = 0.00000428), inflammation (r = 0.793, p = 0.00000000109), and ballooning (r = 0.723, p = 0.000000141) (Fig. 4d–f). Abnormalities in rER correlated with steatosis (r = 0.401, p = 0.0103), inflammation (r = 0.438, p = 0.00472), and ballooning (r = 0.565, p = 0.000146). The total NAS was highly correlated with abnormalities in mitochondria (r = 0.711, p = 0.00000027), glycogen granules (r = 0.793, p = 0.00000000106), and rER (r = 0.507, p = 0.00085) (Fig. 4j–l).

Fig. 4

Correlation between NAS and transmission electron microscopic findings. The correlation of each index was examined using Spearman’s correlation coefficient. The correspondence among TEM findings, NAS items, and the number of patients is presented in Tables 2, 3. The number in each square indicates the number of patients who presented with the relevant findings. The intensity of the red color in each box increases with an increase in the number of individuals with each score