Exploring multiorgan mitochondrial dysfunction in the switch toward progressive MASLD in AMLN mice

Summary

Hepatic mitochondrial maladaptation features the transition from metabolic dysfunction-associated steatotic liver disease (MASLD) to Steatohepatitis (MASH) up to fibrosis/cirrhosis. However, it is still unexplored whether mitochondrial alterations also affect adipose tissue, muscle and heart during disease progression.

C57Bl/6 mice were fed an AMLN diet to recapitulate the human MASLD spectrum. In the liver, TEM depicted a progressive morphologic dysfunction of mitochondria, which appeared swollen in MASH, with disorganized cristae/matrix loss in MASH-fibrosis. The mitophagy pathway was reduced in MASH-fibrosis, thus explaining the accumulation of damaged mitochondria, whereas mitochondrial complexes activities alongside OXPHOS protein levels and ATP production were dampened across the disease in liver, adipose, muscle, and cardiac tissues. Finally, the release of cell-free circulating mitochondrial DNA into the bloodstream reflected tissue mitochondrial impairment.

In sum, we demonstrated that alterations in mitochondrial morphology, life cycle, and activity feature all disease stages in the liver but also in other tissues engaged in MASLD evolution.

Graphical abstract

Highlights

• •Hepatic mitochondrial dysfunction features MASLD onset and progression
• •The impairment of mitochondria in adipose and muscle tissues is unexplored
• •Mitochondrial morphology, mitophagy, and ATP production were affected in all tissues
• •Mitochondrial failure in MASLD is a systemic issue involving multiple tissues

Biochemistry; Molecular biology; Systems biology

Introduction

Metabolic dysfunction-associated steatotic liver disease (MASLD) is one of the most alarming global health challenges, whereby affecting around 25% of the adult population and 10% of the pediatric one.^1,2 This condition entails a plethora of clinical phenotypes, starting from simple and reversible steatosis to its progressive form, steatohepatitis (MASH). The latter is portrayed by serious histo-pathological aberrances, among which are advanced triglyceride overload, inflammation, spotty necrosis, and hepatocellular degeneration, which could escalate into fibrosis, cirrhosis, and hepatocellular carcinoma (HCC), in predisposed individuals.^3,4 Compelling evidence points toward inadequate mitochondrial adaptation as a central player in the transition from steatosis to MASH and so forth to HCC.^5,6 Increased mitochondrial mass and biogenesis are described in the livers of patients with MASLD, and even more in those with MASH.⁷ This notion supports the theory that mitochondria exhibit an elevated degree of adaptability, modulating their quantity and functional activity in the presence of excessive metabolic substrates. Indeed, in the early stages of MASLD, mitochondrial activity increases in response to hepatic insulin resistance (IR) and free fatty acid (FFA) accumulation, in an attempt to compensate for the heightened energy demand. Conversely, mice and humans with MASH displayed megamitochondria with crystalline inclusions, respiratory chain dysfunction, lower maximal respiration, mitochondrial uncoupling, and ultimately declining mitochondrial efficiency.⁸ The sustained mitochondrial morphological/functional detriments result in the overproduction of harmful by-products, as reactive oxygen species (ROS), strictly associated with the lipid peroxidation of organelle membranes, mitochondrial DNA (mtDNA) damage, endoplasmic reticulum (ER) stress, cell death, and tissue inflammation, all of which may contribute to MASLD evolution.⁹

Specifically, the loss of mitochondrial flexibility leads to the unbalanced mitobiogenesis, causing the assembly of misshapen mitochondria characterized by low β-oxidation, OXPHOS capacity, and ATP production.¹⁰ Furthermore, the inhibition of mitophagy exacerbates the accumulation of exhausted mitochondria and favors the release of circulating cell free mitochondrial (mt-) DNA (ccf-mtDNA) into the bloodstream. The latter acts as damage-associated molecular patterns (DAMPs) signals, thus amplifying the hepatic injuries.^5,11,12

Although the primary site of MASLD is the liver, multiple tissues, in particular fat and muscle mass, have now emerged as central players in its onset and worsening, by participating in insulin signaling sensing, and releasing pro-inflammatory mediators, hormones, and so on, to the point that current trends support the need to face MASLD by a multi-systemic point of view.^13,14

Despite the extensive efforts employed in this research context, the precise event cascade that may precipitate simple steatosis to steatohepatitis is not completely clarified. Thus, a deep understanding of the landscape of the organelle abnormalities that occur after fat deposition into the hepatocytes remains an uncharted territory. In particular, impairments in mitochondrial function, activity, and turnover may represent the missing link between steatosis onset and hepatocyte degeneration. Similarly, we can hypothesize that mitochondrial alterations could also affect the other tissues involved, thus further complicating the pathological framework of MASLD. We postulated that mitochondrial dysfunction in different sites of the body may possibly contribute to the intrinsic systemic nature of MASLD, further boosting the disease progression and worsening its course with different comorbidities.

Therefore, we aimed to investigate the role of mitochondrial dysfunction in the switch from simple steatosis to MASH and fibrosis by exploiting a murine model of MASLD induced by high-fat, high-fructose, high-cholesterol diet, also referred to as Amylin liver NASH (AMLN) diet, to resemble human MASLD. Our purpose was to define whether mitochondrial maladaptation tracks the disease severity, firstly in the liver, and secondarily also in the periphery, thus mirroring the hepatic damage.

Here, we corroborated the impairment of mitochondrial activity in the liver, and we extended these observations beyond to other tissues. Indeed, we outlined that mitochondrial maladaptation is not restricted to the liver, but it is also encountered in adipose, skeletal muscle, and cardiac tissues, further boosting the progression of liver disease.

Results

AMLN diet recapitulates the histo-pathological hallmarks of human MASLD

It has been previously demonstrated that the long-term administration of AMLN diet in male mice mimics the main features of human MASLD, encompassing simple steatosis until MASH and fibrosis.¹⁵ In this regard, we decided to feed C57Bl6 male mice standard (SD) or AMLN diets for 14, 22 and 28 weeks to resemble the different stages of human MASLD, hereafter renamed as steatosis, MASH and MASH-fibrosis groups, respectively.

According to the time of AMLN diet exposure, we highlighted a progressive increase in body and liver weight from the steatosis group to the MASH-fibrosis one (Table S1). In keeping with this result, the intrahepatic triglyceride content was strongly enhanced upon AMLN administration (∗∗∗p-adj<0.001 for steatosis, MASH, MASH-fibrosis group vs. each control group; Figure 1A). Intrahepatic and serum cholesterol levels increased only in mice belonging to MASH and MASH-fibrosis groups, compared to their SD counterparts (∗∗∗∗p-adj<0.0001; Figures 1B and S1A), although hepatic Apolipoprotein B (ApoB) synthesis tended to be higher already in the steatosis group (Figure S1B). Accordingly, serum ApoB and VLDL/LDL were enhanced after AMLN feeding (∗p-adj<0.05 for steatosis, MASH, and MASH-fibrosis group vs. each control group; Figures S1C and S1F). Conversely, HDL cholesterol was strongly dampened in all disease conditions (∗p-adj<0.05; Figures S1D and S1E), thus featuring the human pro-atherogenic dyslipidemia that accompanied MASLD. In addition, we measured the concentrations of hepatic and serum oxidized LDL (ox-LDLs), which reflect the membrane lipid damage, especially for poly-unsaturated phospholipid, and we found that they were enhanced in MASH and MASH-fibrosis. In the latter condition, hepatic ox-LDL levels were even higher compared to serum ones (Figures S1G and S1H).

Figure 1.

AMLN diet resembles the histo-pathological features of human MASLD

(A) The intrahepatic triglycerides amount was measured by exploiting the Triglycerides Colorimetric Assay Kit in homogenate livers (n = 10/group).

(B) Total cholesterol was measured in homogenate liver samples by using the Cholesterol Colorimetric Assay Kit (n = 10/group).

(C) LDs accumulation was assessed by Hematoxylin and Eosin (H&E) staining of hepatic specimens (Original magnification 100×, scale bar 100 μm).

(D) Hepatic protein levels of pAKT (ser473), AKT and InsR were assessed by Western blot and normalized to the Vinculin as housekeeping gene.

(E) Insulin was measured in the sera of mice belonging to SD and AMLN groups (n = 10/group) by using the Mouse INSULIN ELISA Kit.

(F) Lipolysis was assessed by exploiting the Lipolysis Colorimetric Assay Kit in homogenized liver samples.

(G) Red Sirius positive areas (%) were quantified in hepatic tissues (n = 10/group) embedded in paraffin within formalin fixation by ImageJ in 10 random non-overlapping micrographs per condition by calculating the percentage of pixels above the threshold value with respect to total pixels per area (Original magnification 100×, scale bar 100 μm). The black arrow indicates the pericellular chicken wire fibrosis.

(H) Hepatic hydroxyproline production was determined by the spectrophotometric method of Bergman and Loxley. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in duplicate (Western Blot analysis, lipolysis, and hydroxyproline production).

Data are expressed as data points and SD (adjusted (adj) ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction).

At the histological level, the steatosis group was characterized by a mixed micro/macro pattern of lipid droplets (LDs), mainly gathered in the centrilobular zone 3, whereas they spread toward the entire lobule during MASH (Figure 1C). Alongside severe steatosis, MASH was featured by the presence of severe immune infiltration, testified by enormous inflammatory foci, close to parenchymal sites of spotty necrosis and hypertrophic hepatocytes. An even extreme subversion of hepatic tissue has been encountered in the MASH-fibrosis group, in which we discriminated mainly widespread giant LDs (macro pattern), accompanied by severe fibrosis (Figure 1C). In keeping with these findings, circulating pro-inflammatory interleukin (IL)-6, 1β, and TNFα were noticeably raised in the MASH-fibrosis group compared to the other conditions (∗∗∗∗p-adj<0.0001; Figures S1I–S1K).

Moreover, InsR and pAKT/AKT protein levels lowered over MASLD progression (∗p-adj<0.05 for steatosis, MASH, and MASH-fibrosis group vs. each control group; Figures 1D, S2A, and S2B), and were paralleled by higher serum insulin and HOMA-IR in the MASH-fibrosis group (∗∗∗∗p-adj<0.0001; Figure 1E; Table S1). As a consequence of reduced insulin sensitivity, serum ketone bodies were augmented in AMLN mice with the sharper increase in MASH-fibrosis (Table S1). Although we did not perform oGTT or an ITT to test glucose and insulin tolerance, all these findings indicated the presence of severe hepatic IR mainly in the MASH-fibrosis group. In line with the enhanced fat content and IR, we also found an enhanced lipolysis, mostly in the early stages of MASLD, probably due to the attempt of handling excessive fat (∗∗∗p-adj<0.0001 for steatosis, MASH and MASH-fibrosis group vs. control group; Figure 1F). According to the disease progression across AMLN diet exposure, we observed an intensified percentage of Red Sirius positive areas, thus clearly highlighting the presence of perisinusoidal/pericellular chicken wire fibrosis, higher hydroxyproline levels, and positivity for alpha-SMA in MASH-fibrosis (∗∗∗∗p-adj<0.0001; Figures 1G, 1H, and S1L).

Finally, we confirmed the presence of fat accumulation at Oil Red O (ORO) staining in primary hepatocytes isolated from mice fed an AMLN diet (n = 2 mice/group) (∗p-adj<0.05 for steatosis group vs. control group; ∗∗p-adj<0.01 for MASH, MASH-Fibrosis group vs. control group; Figure S3A).

The progressive loss of hepatic mitochondrial dynamics parallels MASLD severity

To further dissect the physiology of hepatocytes, we next sought to investigate intracellular organelles’ morphology by using TEM. Intriguingly, TEM analysis confirmed the progressive enlargement of LDs, with the micro ones that convey in the largest during the course of the disease (Figure 2A). These alterations were associated with a severe disruption in mitochondrial structures. Indeed, we noticed that mitochondria appeared normo-shaped with rare misshapen or elongated ones in steatosis, while they were completely swollen in MASH. Furthermore, in the MASH-fibrosis group, we discriminated aberrant organelles with disorganized cristae and matrix loss, and this mitochondrial derangement was paralleled by a progressive fragmentation of ER and apoptotic nuclei (Figure 2B).

Figure 2.

The loss of hepatic mitochondrial dynamics as signature of progressive MASLD

(A and B) The intrahepatic lipid droplets (LDs) accumulation and aberrant mitochondrial architecture was assessed by transmission electron microscopy (TEM, magnification 3000× and 7000× where the scale bar indicates 3.33 μm or 1.43 μm, respectively).

(C) Hepatic protein levels of PGC1α, DRP1, Mfn1, Mfn2, Opa1, Parkin, PINK1, BNIP3-L, BNIP3, and LC3-I and LC3-II were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(D) Hepatic protein levels of CIV-ATP5A, CIII-UqRC2, CIV-MTCOI, CII-SDHB (OXPHOS complexes), citrate synthase, and COXIV were assessed by Western blot and normalized to the Vinculin as a housekeeping gene. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in duplicate (for Western Blot analysis).

(E–H) The hepatic citrate synthase, complex I, complex III, and ATP synthase activities were measured by colorimetric assays in isolated mitochondria. Data were normalized on μg of mitochondrial proteins.

(I) ATP production was measured by colorimetric assays in homogenate liver samples (n = 10/group). Data were normalized based on the amount of proteins obtained from the homogenized tissues. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in quadruplicate (for all mitochondrial enzymatic activities and ATP production). At least 4 independent experiments were conducted. Data are expressed as data points and SD (adjusted (adj) ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction.

For this reason, we decided to explore whether the structural defects were mirrored by functional anomalies in these organelles. We demonstrated that while Drp1 (fission) protein levels were increased in mice treated with AMLN diet, Pgc1a (one of the key regulators of mitobiogenesis), Mfn1/2, Opa1 (fusion) and Parkin, Pink1, Bnip3/3L, Lc3 I-II (mitophagy) were reduced, thus suggesting an unbalanced mitobiogenesis, linked to a suppression of mitophagy in particular in MASH-fibrosis (∗p-adj<0.05 for steatosis, MASH and MASH-Fibrosis group vs. each control group; Figures 2C and S2C–S2L). This data overall implies the accumulation of failed mitochondria, which was further confirmed by the suppression of OXPHOS sub-unities, citrate synthase and cytochrome c oxidase (Complex IV) protein levels (∗p-adj<0.05 for steatosis, MASH and MASH-Fibrosis group vs. each control group; Figures 2D and S2M–S2S).

Secondly, we observed that these changes were reflected by the loss of functional assets after AMLN diet, thereby suppressing the citrate synthase, complex I, III, V (ATP synthase) enzymatic activities (∗p-adj<0.05; Figures 2E–2H) and ATP production, especially in the presence of MASH and MASH-fibrosis (∗∗∗p-adj<0.001; Figure 2I). Overall, these data strongly interconnected the defects in mitochondrial morphology with the imbalance of mitobiogenesis and an impaired function as features of progressive MASLD.

The failed hepatic mitochondrial respiration associates with enhanced oxidative stress

Next, we investigated whether the unbalance of mitobiogenesis observed above was paralleled by the lack of mitochondrial function. Coherently, Seahorse assay showed a reduction of hepatic oxygen consumption rate (OCR) going toward MASH-fibrosis (∗∗∗∗p-adj<0.0001 for MASH and MASH-fibrosis group vs. control group; Figure 3A), supporting the linkage between an impaired hepatic bioenergetic profile and advanced disease. To further validate this finding, the mitochondrial respiration was measured even in primary hepatocytes isolated from mice fed the AMLN diet, reaching the lowest levels in MASH-fibrosis (∗∗p-adj<0.01; Figure S3B). In line with what observed in hepatic tissue (Figure 3A), mitochondrial OCR was reduced also in blood of MASH-fibrosis group, representative of mitochondrial respirometry of leukocytes and platelets (∗∗∗p-adj<0.001; Figure 3B). This data suggests an alteration of mitochondrial function even in peripheral cells during MASLD progression. In furtherance, the ccf-mtDNA (COXIII and ND1 fragments), DAMPs signals hallmarks of mitochondrial derangement, were broader released in the sera of MASH-fibrosis mice (∗p-adj<0.05; Figure 3C), thereby boosting MASLD worsening. Consistently, we observed a remarkable elevation of hepatic oxidative stress, evidenced by enhanced ROS/RNS, H2O2 levels, lipid peroxidation (measured by aldehyde derivatives, MDA) and apurinic/apyrimidinic (AP) sites - the main ROS-induced DNA damage (∗p-adj<0.05; Figures S4A–S4D). In addition, we observed high LDH enzymatic activity as well as serum lactate secretion in AMLN mice compared to their littermates fed SD whereas hepatic lactate was reduced. The latter result may be ascribable to the rapid conversion of lactate into pyruvate or to an enhanced export (∗∗∗ p-adj<0.001 for steatosis, MASH and MASH-fibrosis groups vs. control group; Figure S4E). Overall, these data strongly support the pivotal role of hepatic mitochondrial failure encompassing morphology, dynamics and activity as driven of MASLD evolution. Interestingly, as in the liver, these aberrancies seem to affect even other cell types, as peripheral cells thus supporting that mitochondrial derangements in MASLD may be extended to extra-hepatic tissues and opening spyholes for MASLD monitoring.

Figure 3.

The mitochondrial bioenergetic profile was altered toward progressive MASLD

(A and B) 1 to 3 mg of frozen livers and 50 μL of blood samples (n = 10/group) were manually homogenized in 250 μL of MAS buffer (70 mM sucrose, 220 mM mannitol, 5 mM KH2PO4, 5 mM MgCl2, 1 mM EGTA, 2 mM HEPES) and 80 μg of homogenates were loaded in Seahorse XF24 plate. Data show the oxygen consumption rate (OCR) during the time course (OCR pmol O2/min ± SD, kinetic graph) and the total area under the curve (AUC) of OCR. Each dot represents the mean value ± SD. All reactions were performed in quadruplicate. At least 2 independent experiments were conducted.

(C) The release of cell-free mitochondrial DNA fragments (ccf-mtDNA), encompassing COXIII and ND1 ones, was quantified through quantitative real-time PCR and normalized on the standard curve obtained from serial dilutions of a sample pool at a known concentration. An equal amount of sera samples (n = 10/group) were pooled prior to the DNA extraction, and gene expression analyses were performed in quadruplicate. ccf-mtDNA was measured as copies/μL∗10ˆ4. At least 4 independent experiments were conducted. Data are expressed as data points and SD (adjusted (adj) ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction).

Mitochondrial dysfunction affects adipose tissue during MASLD

We then decided to assess whether the mitochondrial alterations observed in the liver could be generalized to other tissues involved in MASLD pathophysiology starting from the adipose tissue. We detected a progressive reduction of pAKT/AKT and InsR protein levels during disease course (∗∗∗∗p-adj<0.0001; Figures 4A, S5A, and S5B). As we expected, we found an escalation in lipolysis activation, according to the time of AMLN exposure, thus culminating in the MASH-fibrosis group (∗∗∗∗p-adj<0.0001; Figure 4B). Intriguingly, in these mice we found positive Red Sirius areas (Figure 4C) and a peak in hydroxyproline content (∗∗∗p-adj<0.001 for MASH-fibrosis group vs. control group; Figure 4D), possibly suggesting a stiffening of adipose tissue, which is a phenomenon related to adipocyte loss of function and linked to systemic IR.¹⁶

Figure 4.

Mitochondrial dysfunction injures the adipose tissue during MASLD progression

(A) Protein levels of pAKT (Ser473), AKT and InsR were assessed by Western blot and normalized to the Vinculin housekeeping gene in homogenated adipose tissues (n = 10/group).

(B) Lipolysis was assessed by exploiting the Lipolysis Colorimetric Assay Kit in homogenated adipose samples (n = 10/group).

(C) Red Sirius positive areas were assessed in frozen adipose tissues embedded in OCT (Original magnification 200×, scale bar 100 μm).

(D) Adipose hydroxyproline production was determined by the spectrophotometric method of Bergman and Loxley.

(E) Adipose protein levels of Pgc1a, Mfn1, Drp1, Parkin, Pink1, BNIP3-L, BNIP3, and LC3-I and LC3-II were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(F) Adipose protein levels of CIV-ATP5A, CIII-UqRC2, CIV-MTCOI, CII-SDHB (OXPHOS complexes) were assessed by Western blot and normalized to the Vinculin as a housekeeping gene. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in duplicate (Western Blot, lipolysis, and hydroxyproline measure).

(G) 1 to 3 mg of frozen adipose samples were homogenized and plated in the Seahorse XF24 plate. Data show the oxygen consumption rate (OCR) during the time course (OCR pmol O2/min ± SD, kinetic graph) and the total area under the curve (AUC) of OCR. Each dot represents the mean value ± SD. All reactions were performed in quadruplicate. At least 2 (Seahorse) or 4 (for all the other analyses) independent experiments were conducted. Data are expressed as data points and SD (adjusted (adj) ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction).

Differently from the liver, here we observed increased levels of Mfn1 (fusion) protein compared to their SD counterparts, while Pgc1a (mitobiogenesis) and Drp1 (fission) remained unchanged during the MASLD course. Although we observed this opposite trend, the mitophagy pathway (Parkin, Pink1, Bnip3/3L, and LC3I-II) hugely declined in MASH-fibrosis mice, thereby suggesting an unbalanced mitobiogenesis related to the failure of discarding damaged mitochondria even in this tissue (Figures 4E and S5C–S5J). In line with this result, mitochondrial OXPHOS complexes V (ATP5A) and III (UQRC2) were almost abolished in adipose tissue from the MASH-fibrosis group (Figures 4F and S5K–S5O). Accordingly, citrate synthase activity and total ATP production were lower in all diet-induced MASLD groups and more so in those affected by MASH-fibrosis, in which the total ATP concentration was quite abrogated (∗∗∗∗p-adj<0.0001; Figures S6A and S6B). Consistently, Seahorse assay revealed low oxygen consumption rate in MASH and MASH-fibrosis mice, thus emphasizing that mitochondrial activity fails during MASLD progression even in adipose tissue (∗∗p-adj<0.01 for MASH and MASH-fibrosis group vs. control group; Figure 4G). Finally, these events primed oxidative stress with ROS/RNS and MDA overproduction (∗p-adj<0.05; Figures S6C and S6D) as observed in hepatic tissues isolated from mice with severe MASLD, thus pointing out the involvement of mitochondrial derangement in promoting advanced damage also in adipose tissue.

AMLN fed mice resemble histological MASLD phenotype in muscle tissues

During the onset of MASLD, also skeletal and cardiac muscle mass may play a pivotal role, thus contributing to the systemic impact of the disease.¹⁷ Since both these tissues are strongly enriched in mitochondria, we hypothesized that mitochondrial structure/function impairments could also affect them during the worsening of MASLD.

In both muscle tissues, we highlighted that the triglyceride content was early enhanced upon AMLN exposure, already accumulating in the steatosis group and remaining higher in the others compared to their littermates fed SD (∗∗∗∗p-adj<0.0001 for steatosis, MASH and MASH-fibrosis group vs. each control group; Figures 5A and 6A). In keeping with the enhanced triglyceride content, at histological level and at ORO staining (Figures 5B, 6B, and S7A), we found fat infiltration alongside the skeletal muscle fibers and an increase in pericardial fat mass (12 mg in steatosis, 15 mg in MASH, and 19 mg in MASH-fibrosis vs. almost absent in control groups) in AMLN-fed mice. Notably, fat overload in muscle mass participates in the disruption of insulin signaling activation, exacerbating peripheral IR and promoting hepatic steatosis.^18,19 In our model, we observed lower protein levels of pAKT/AKT as well as InsR in skeletal tissue of AMLN-fed mice compared to their SD controls (∗p-adj<0.05; Figures 5C, S7B, and S7C). On the contrary, in cardiac muscle, we observed a modest effect of the diet on insulin signaling, probably due to the partial preservation of cardiac function (Figures 6C, S8A, and S8B). However, tissue lipolysis was enhanced in both type of muscles during MASLD, mainly in the MASH-fibrosis group compared to SD (∗p-adj<0.05; Figures 5D and 6D), thus exacerbating the release of fatty acids from both tissues. Triglyceride build-up can lead to reduced muscle mass, atrophy, and predisposes over the long-term to fibrosis.²⁰ Indeed, Red Sirius positive areas (Figures 5E and 6E) along with the hydroxyproline concentration were enhanced in MASH and in MASH-fibrosis in both skeletal and cardiac specimens (∗p-adj<0.05; Figures 5F and 6F), thus confirming the presence of fibrosis also in these tissues.

Figure 5.

AMLN diet exposure outlines histologic MASLD progression and mitochondrial dysfunction in skeletal muscle

(A) The triglyceride amount was measured by exploiting the Triglycerides Colorimetric Assay in homogenate muscle tissues (n = 10/group).

(B) LDs accumulation was assessed by Hematoxylin and Eosin (H&E) staining of muscle specimens (Original magnification 100×, scale bar 100 μm).

(C) Skeletal muscle protein levels of pAKT (Ser473), AKT, and InsR were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(D) Lipolysis was assessed by exploiting the Lipolysis Colorimetric Assay Kit in homogenate muscle samples (n = 10/group).

(E) Red Sirius positive areas were assessed in muscle tissues embedded in paraffin within formalin fixation (Original magnification 100×, scale bar 100 μm).

(F) Muscle hydroxyproline production was determined by the spectrophotometric method of Bergman and Loxley.

(G) Aberrant mitochondrial architecture was assessed by transmission electron microscopy (TEM, magnification 5000× where the scale bar indicates 2.0 μm and the inset was 7000×). We observed an enrichment in small and healthy mitochondria with intact double mitochondrial membranes that are localized adjacent to the z-line in all conditions except in the MASH-fibrosis. In the latter, we found an increased number of failed and enlarged mitochondria with the loss of internal matrix.

(H) Skeletal protein levels of Pgc1a, Mfn1, Drp1, Parkin, Pink1, Bnip3, Bnip3L, and LC3 were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(I) Skeletal protein levels of CIV-ATP5A, CIII-UqRC2, CIV-MTCOI, CII-SDHB (OXPHOS complexes) were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(J) Skeletal citrate synthase activity was measured by a colorimetric assay in homogenized muscle tissues.

(K) Skeletal muscle ATP production was measured by a colorimetric assay in homogenized muscle tissues. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in duplicate (for triglycerides quantification, Western Blot analysis, lipolysis, and hydroxyproline production) quadruplicate (for citrate synthase activity and ATP production).

(L) 1 to 3 mg of frozen skeletal muscles were homogenized and plated in the Seahorse XF24 plate. Data show the oxygen consumption rate (OCR) during the time course (OCR pmol O2/min ± SD, kinetic graph) and the total area under the curve (AUC) of OCR. Each dot represents the mean value ± SD. All reactions were performed in quadruplicate. At least 2 (Seahorse) or 4 (for all the other analyses) independent experiments were conducted. Data are expressed as data points and SD (adjusted (adj) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction).

Figure 6.

Mitochondrial dynamics are impaired in cardiac tissue during MASLD progression

(A) The triglyceride amount was measured by exploiting the Triglycerides Colorimetric Assay in homogenate cardiac muscles (n = 10/group).

(B) LDs accumulation was assessed by Hematoxylin and Eosin (H&E) staining of heart specimens (Original magnification 100×, scale bar 100 μm).

(C) Cardiac muscle protein levels of pAKT (ser473), AKT, and InsR were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(D) Lipolysis was assessed by exploiting the Lipolysis Colorimetric Assay Kit in homogenate cardiac muscle samples (n = 10/group).

(E) Red Sirius positive areas were assessed in cardiac muscle tissues embedded in paraffin within formalin fixation (Original magnification 100×, scale bar 100 μm).

(F) Heart hydroxyproline production was determined by the spectrophotometric method of Bergman and Loxley.

(G) Cardiac muscle protein levels of Pgc1a, Mfn1, Drp1, Parkin, Pink1, Bnip3, Bnip3L, and LC3 were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(H) Cardiac muscle protein levels of CIV-ATp5A, CIII-UqRC2, CIV-MTCOI, and CII-SDHB (OXPHOS complexes) were assessed by Western blot and normalized to the Vinculin as a housekeeping gene.

(I) Cardiac muscle citrate synthase activity was measured by colorimetric assays in homogenates.

(J) Cardiac muscle ATP production was measured by colorimetric assays in homogenate heart tissues. An equal amount of tissue homogenates (n = 10/group) was pooled prior to the analyses, and all reactions were performed in duplicate (for triglycerides quantification, Western Blot analysis, lipolysis, and hydroxyproline production) quadruplicate (for citrate synthase activity and ATP production).

(K) 1 to 3 mg of frozen cardiac muscle samples were homogenized and plated in the Seahorse XF24 plate. Data show the oxygen consumption rate (OCR) during the time course (OCR pmol O2/min ± SD, kinetic graph) and the total area under the curve (AUC) of OCR. Each dot represents the mean value ± SD. All reactions were performed in quadruplicate. At least 2 (Seahorse) or 4 (for all the other analyses) independent experiments were conducted. Data are expressed as data points and SD (adjusted (adj) ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 one-way ANOVA, Tukey’s correction).

Muscle mitochondrial flexibility and function are lost during MASLD progression

The impairment of mitochondrial function in muscle and cardiac tissues upon MASLD progression has been scantily investigated. In skeletal muscle, we found an increased amount of deranged interfibrillar mitochondria in the MASH-fibrosis group at TEM analysis (Figure 5G).

Moreover, in both tissues, we observed a mitochondrial life cycle imbalance, favoring the process of fusion at the expense of mitophagy, testified by Mfn1 protein levels elevation and by Pgc1a, Drp1, Pink, Parkin, Bnip3/3L, LC3 softening mainly in MASH-fibrosis (∗p-adj<0.05; Figures 5H, 6G, S7D–S7K, and S8C–S8J). In the latter group, OXPHOS complexes (∗∗∗∗p-adj<0.0001 for MASH-fibrosis group vs. control group; Figures 5I, 6H, S7L–S7P, and S8K–S8N), along with citrate synthase and ATP production, were strongly reduced compared to mice fed SD for 28 weeks (∗p-adj<0.05; Figures 5J, 5K, 6I, and 6J). Notably, we revealed a progressive decrease in cardiac ATP production even in the SD groups, probably due to mice aging, which is per se related to mitochondrial deficits (Figure 6J). Accordingly, Seahorse assay highlighted the main drop of oxygen consumption rate in mice fed with AMLN diet for 28 weeks compared to their SD control, thus supporting the involvement of failed muscle bioenergetic profile over disease progression (∗∗p-adj<0.01 for MASH-Fibrosis group vs. control group for skeletal muscle, Figure 5L; ∗∗∗∗p-adj<0.0001 for MASH-fibrosis group vs. control group for cardiac muscle; Figure 6K). Finally, organelle dysfunction was accompanied by ROS/RNS and MDA overproduction during fibrotic MASLD in both tissues, but even more in the cardiac one (∗∗∗∗p-adj<0.0001 for MASH-fibrosis group vs. control group; Figures S9A–S9D), thus sustaining the increased cardiovascular risk which frequently occurs in patients with fibrosing MASLD.²¹

Discussion

Although MASLD is emerging as one of the most critical health issues of this century, no resolutive therapeutic approaches are still available. Currently, only Resmetirom, a liver-targeted thyroid hormone receptor-β agonist, has been recently approved for the treatment of MASH-fibrosis in clinical trials albeit the effects are modest and follow-up studies are needed to draw definite conclusions.²²

Our study is aimed to explore the hepatic mitochondrial defects that occur during MASLD progression and to investigate whether they may be extended to adipose tissue, skeletal, and cardiac muscles by exploiting a mouse model fed AMLN diet, which resembles the human disease spectrum. Here, we demonstrated a significant decline of mitochondrial physiology, in terms of structure and activity, across disease severity in all the assessed tissues, with the greatest effect in the liver. Indeed, during MASLD progression and more so in the latest stages, we noticed an exacerbated build-up of misshapen/swollen mitochondria, an impaired mitochondrial respiration, a higher ccf-mtDNA release into the bloodstream and escalated oxidative stress, overall supporting that mitochondrial failure is the key driver in the transition from simple steatosis to MASH and fibrosis. Moreover, since mitochondrial respiration was highly reduced also in blood cells (leukocytes and platelets) in MASH-fibrosis, these data pointed out the clinical utility to assess the circulating bioenergetic profile to trace a signature of disease severity.²³

In the liver, we reported that the spreading and the enlargement of LDs occurred during MASH and in MASH-fibrosis was furtherly accompanied by a serious disruption of mitochondrial structure and function, including organelle swelling, internal cristae disruption with matrix loss. All these structural aberrancies were causative of impaired ATP production due to the derangement in respiratory chain complexes, enhanced ROS production, and membrane lipid peroxidation, which in all emphasize mitochondrial dysfunction. This data is in line with recent studies from our group in which we described a premature block of mitophagy and a mitochondrial loss of function across the disease course in patients with MASLD.²³

Notwithstanding, oxidized lipids (MDA and ox-LDLs) may be critical markers of severe MASLD.²⁴ Indeed, these compounds may reflect oxidative damage to membrane lipids, and the generation of the latter is closely entangled with the oxidation of polyunsaturated fatty acids in LDL particles. Moreover, ox-LDLs correlate with oxidative stress and inflammation, which are common during MASLD progression. As a consequence, it has been reported that increased concentrations of hepatic ox-LDLs prime hepatic stellate cells (HSCs) activation, by exerting immunomodulatory properties. Here, we reported that hepatic and circulating lipids were strongly enhanced in all disease conditions, albeit did not further increase during MASH-fibrosis development, maybe due to the severe liver failure. Concerning ox-LDLs, we demonstrated that during fibrosis, they were mainly located in the liver compared to the blood compartment. These observations were in line with the recent lipidomic study, which revealed a peculiar lipidic signature in patients with fibrosis.²⁴

Then, we decided to expand these observations to other tissues involved in MASLD. In adipose tissue, we identified an increased lipolysis, suggestive of IR related to a significant reduction in mitochondrial functionality. Interestingly, these events were also matched with the scarring of the adipose tissue, which is a sign of systemic IR and adipocytes degeneration.¹⁶ Indeed, it has been reported that in adipose tissue, dysfunctional mitochondria are linked to the presence of IR, leading to an impairment in fatty acid oxidation and to lipid accumulation and inflammatory cell infiltration.²⁵ This occurrence coupled with pro-inflammatory mediators and adipokine secretion may aggravate systemic inflammation and fibrosis. Even more, the pro-inflammatory milieu amplifies IR and inhibits the OXPHOS machinery.²⁶ As a consequence, adipose tissue fibrogenesis has been associated with clinically significant hepatic fibrosis, suggesting a bidirectional relationship between liver and adipose tissue in MASLD.^27,28,29 Interestingly, a 12-month follow-up study with matched liver and adipose tissue biopsies showed that bariatric surgery enables not only the reduction of adipose mass but also ameliorates hepatic OXPHOS and mitochondrial biogenesis.³⁰

In addition, in line with the reduction in the expression of mitochondrial complexes observed in our model, we previously reported a hepatic/adipose tissue comparative transcriptome analysis highlighting the oxidative phosphorylation as the only significantly downregulated pathway both in liver and adipose tissue during MASLD.³¹ Alongside, we outlined also a common deregulation of lysosomal-autophagic pathway in hepatic and adipose tissue isolated from patients affected by MASH-fibrosis, thereby fostering defective autophagic clearance and multivesicular bodies buildup.³² This data may explain the accumulation of misshapen mitochondria, due to the block of mitophagy, which is a cargo-specific form of autophagy to eliminate damaged mitochondria.³³ Our observations sustained the importance of adipose tissue in MASLD initiation and agreed with the recent study which demonstrated the presence of blunted mitochondrial respiration, associated with IR and inflammation, in visceral adipose tissue from obese patients who developed MASLD.³⁴

Next, we demonstrated that skeletal and cardiac muscles also exhibited prominent mitochondrial defects, entailing detrimental mitochondrial dynamics and dampened OXPHOS capacity, thus fostering triglyceride storage also in non-classical tissues. To sustain physical activity and promote glucose utilization, skeletal muscle is strongly enriched in mitochondria, and during MASLD, a partial substitution of muscle mass with fat has been reported, paralleled by atrophy, weakness, and fatigue. All-in-all, these alterations are accompanied by mitochondrial exhaustion, which induces low energy expenditure, IR, myocellular apoptosis and an adverse secretory pattern of myokines.³⁵ Indeed, an association between severe IR, muscular accumulation of triglycerides and diminished oxidative enzymes has been reported in obese and type 2 diabetic subjects.³⁶ Accordingly, the activity of the electron transport chain was found to be reduced by ∼40% in skeletal muscle from patients with type 2 diabetes.³⁷ Therefore, mitochondrial dysfunction may represent a shared pathogenic mechanism by MASLD and sarcopenia and in particular, in patients with cirrhosis the quick shift to fat and protein catabolism as energy source, leads to impaired bioenergetics, rapid muscle breakdown and frailty.³⁵ On the contrary, aerobic exercise is correlated with improved skeletal muscle oxidative capacity, mitobiogenesis, synthesis of mitochondrial proteins and bioenergetic metabolism³⁸ along with it relies on hepatic mitophagy.³⁹

Likewise, mitochondria are essential for the proper cardiac functions. Accordingly, mitochondrial network and turnover are hampered in a mouse model of MASLD with IR, thus contributing to hypertrophy and cardiac dysfunction. Therefore, failures in mitochondrial physiology may be responsible for cardiac defects, including myocardial injuries and cardiomyopathy, often associated with MASLD.

Intriguingly, we highlighted that ATP production from mitochondrial respiration is hampered by aging and mainly by MASH-fibrosis in our experimental model. Notwithstanding, different from the progressive demise of mitochondria identified in the liver, we demonstrated that the blunted OXPHOS capacity paired with low ATP generation is noteworthy of the latter stages of MASLD in all the peripheral tissues analyzed.

Our findings exceeded the “liver-centric” view of MASLD, supporting the importance of a comprehensive assessment of mitochondria maladaptation across multiple tissues, in particular in those patients affected by advanced disease. While previous studies were predominantly focused on hepatic organelle abnormalities, here we demonstrated that these defects are extended to peripheral tissues. The multi-tissue mitochondrial detriments may trigger the progression of hepatic disorders and their related comorbidities by altering lipid and glucose utilization and by releasing proinflammatory and pro-fibrotic mediators, including ccf-mtDNA that acts as mito-DAMPs, thus creating an unfavorable systemic milieu. Indeed, our result could suggest that the mitochondrial failure is present in the liver at the earlier stages of MASLD. The latter is featured by a severe inflammatory response whose main promoter and mediator is oxidative stress, which in turn leads to the release of reactive species. The loss of hepatic energetic homeostasis along with mitochondrial dysfunction impact on metabolic pathways (i.e., glucose signaling) thus triggering adipose and muscle tissues detriment. These ones, initially compensate for the bioenergetic demand, whereas during MASH-fibrosis the mitochondrial adaptability is failed.

Hence, we corroborate the idea to the pinpoint mitophagy pathway, which is crucial for injured mitochondria disposal, as a druggable target to test in clinical studies. In this regards, mitophagy stimulation through the induction of PINK1/Park2 signaling has been recently pointed out to relieve MASLD in in vitro models,⁴⁰ whereas liver specific Parkin knock-down hastens steatosis onset, inflammation, and fibrosis.⁴¹

In addition, we also demonstrated that an impaired OXPHOS machinery and a defective mitophagy may be generalized to other mouse models of MASLD, further reinforcing the key role of systemic mitochondrial dysfunction in the disease evolution (Figure S10).

To sum up, mitochondrial failure in MASLD could be considered as a systemic issue affecting multiple tissues, favoring the onset of metabolic comorbidities and driving the evolution of liver disease over time. Overall, the study of mitochondria across different tissues offers the opportunity to discover diagnostic tools and helpful opportunities to manage MASLD through a holistic approach.

Limitations of the study

The main limitation of this study is that we did not perform oral glucose and insulin tolerance tests to assess IR, although we have formally proven the presence of reduced insulin signaling by measuring insulin, HOMA-IR, and serum ketone bodies. In addition, further investigations aimed at restoring mitochondrial homeostasis are needed to establish whether targeting organelle health could mitigate advanced MASLD and its complications.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Paola Dongiovanni (paola.dongiovanni@policlinico.mi.it).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •All data reported in this article will be shared by the lead contact upon request.
• •No unpublished custom code, software, or algorithm was used in this study.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We would like to thank Dr. Stefano Gatti and Dr. Maurizio Moggio for supporting us in data generation and interpretation and Taconic Biosciences for the Academic Grant Award to Paola Dongiovanni.

This study was supported by Italian Ministry of Health (Ricerca Corrente 2025 - Fondazione IRCCS Cà Granda Ospedale Maggiore Policlinico), by Italian Ministry of Health (Ricerca Finalizzata Ministero della Salute GR-2019-12370172; RF-2021-12374481; PNRR-MCNT2-2023-12378295; 5x1000 2020 RC5100020B; Bando Ricerca Corrente and Piano Nazionale Complementare Ecosistema Innovativo della Salute - Hub Life Science-Diagnostica Avanzata (HLS-DA)’ - PNC-E3-2022-23683266 – “INNOVA.” The Department of Pathophysiology and Transplantation, University of Milan, is funded by the Italian Ministry of Education and Research (MUR): Dipartimenti di Eccellenza Program 2023 to 2027).

Author contributions

The authors’ responsibilities were as follows:

MM and EP, study design, data analysis and interpretation, article drafting; ML data analysis and interpretation; MB and DD, data generation; MR and LN, acquisition of TEM images; SC, data interpretation; SG, data generation and article revision; PD, study design, funding acquisition, article revision, and has primary responsibility for final content. All authors read and approved the final article.

Declaration of interests

The authors declare that they have no conflict of interest.

Declaration of generative AI and AI-assisted technologies in the writing process

This study did not exploit generative AI and AI-Assisted technologies.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit polyclonal Anti-PGC1 alpha antibody, WB	Abcam	Cat#AB191838; RRID: AB_2721267
Rabbit monoclonal Anti-DRP1 antibody, WB	Abcam	Cat#AB184247; RRID: AB_2895215
Mouse monoclonal Anti-Mitofusin 1 antibody, WB	Abcam	Cat#AB126575; RRID: AB_11141234
Mouse monoclonal Anti-Mitofusin 2 antibody, WB	Abcam	Cat#AB56889; RRID: AB_2142629
Rabbit monoclonal Anti-OPA1 antibody, WB	Abcam	Cat#AB157457; RRID: AB_2864313
Mouse monoclonal Parkin antibody, WB	Cell Signaling	Cat#4211; RRID: AB_2159920
Rabbit polyclonal Pink1 antibody, WB	Novus	Cat#BC100-494; RRID: AB_10127658
Rabbit monoclonal Bnip3L antibody, WB	Cell Signaling	Cat#12396; RRID: AB_2688036
Rabbit polyclonal BNIP3 antibody, WB	Cell Signaling	Cat#3769; RRID: AB_2259284
Rabbit polyclonal Lc3 antibody, WB	Cell Signaling	Cat#2775; RRID: AB_915950
Rabbit polyclonal Anti-Citrate synthase antibody, WB	Abcam	Cat#AB96600; RRID: AB_10678258
Rabbit polyclonal Phospho-Akt antibody (S473), WB	Cell Signaling	Cat#9271S; RRID: AB_329825
Rabbit monoclonal AKT antibody (pan, CD7E7), WB	Cell Signaling	Cat#4691; RRID: AB_915783
Mouse monoclonal InsR antibody, WB	Santa Cruz	Cat#sc-57342; RRID: AB_784102
Rabbit monoclonal Anti-Vinculin antibody [EPR8185]	Abcam	Cat#AB129002; RRID: AB_11144129
Mouse monoclonal Total OXPHOS Rodent WB Antibody Cocktail	Abcam	Cat#AB110413; RRID: AB_2629281
Rabbit polyclonal Anti-alpha smooth muscle Actin antibody	Abcam	Cat#AB5694; RRID: AB_2223021
Rat monoclonal Brilliant Violet 605 anti-mouse CD45 Antibody	BioLegend	Cat#157217; RRID: AB_3097472
Chemicals, peptides, and recombinant proteins
Dulbecco’s modified Eagle’s medium (DMEM)	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#41965039
Fetal Bovine Serum (FBS)	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#A5670701
Phosphate-Buffered saline (PBS)	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#70011044
L-Glutamine	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#A2916801
Penicillin/Streptomycin	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#15070063
Trypsin/EDTA	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#25200056
Hank’s balanced salt solution (HBSS)	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#12350039
Insulin-Transferin-Selenium (ITS-G100X)	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#41400045
Dexamethasone, 10 mM in DMSO	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#A13449
Oil Red O	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#189400250
Hematoxylin	BioOptica (Milano, Italy)	Cat#05-06012
RIPA Lysis and Extraction Buffer	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#89901
Protease Inhibitor Cocktail Tablets	Roche	Cat#04693116001
Phosphatase Inhibitor Cocktail Tablets	Roche	Cat#04906845001
Sucrose	Sigma Aldrich	Cat# S0389
KH2PO4	Sigma Aldrich	Cat# RDD037
MgCl2	Sigma Aldrich	Cat#8147330100
Hepes	Sigma Aldrich	Cat# PHR1428
EGTA	Sigma Aldrich	Cat#324626
BSA free fatty acids	Sigma Aldrich	Cat#126575
Cytochrome C	Sigma Aldrich	Cat# C2506
NADH	Sigma Aldrich	Cat# N8129
Antimycin A	Sigma Aldrich	Cat# A8674
TMPD	Sigma Aldrich	Cat#8790
Ascorbic acid	Sigma Aldrich	Cat# PHR1008
Sodium azide	Sigma Aldrich	Cat# S8032
Mannitol	Sigma Aldrich	Cat# PHR1007
Critical commercial assays
Mitochondria Isolation Kit for Cultured Cells	Abcam	Cat#AB110170
Citrate Synthase Colorimetric Assay Kit	Abcam	Cat#AB239712
Mitochondrial Complex I Activity Colorimetric Assay Kit	Abcam	Cat#AB287847
Mitochondrial Complex III Colorimetric Activity Assay Kit	Abcam	Cat#AB287844
ATP synthase Enzyme Activity Microplate Colorimetric Assay Kit	Abcam	Cat#AB109714
ATP Assay Kit (Colorimetric/Fluorometric)	Abcam	Cat#AB83355
DCF ROS/RNS Assay Kit (biofluids, culture supernatant, cell lysates) (Fluorometric)	Abcam	Cat#AB238535
Hydrogen Peroxide Assay Kit (Colorimetric/Fluorometric)	Abcam	Cat#AB102500
Lipid Peroxidation (MDA) Assay Kit (Colorimetric/Fluorometric)	Abcam	Cat#AB118970
DNA Damage Assay Kit (AP sites, Colorimetric)	Abcam	Cat#AB211154
Lactate Dehydrogenase (LDH) Assay Kit (Fluorometric)	Abcam	Cat#AB197000
Lactate Assay Kit (Colorimetric)	Sigma Aldrich	Cat#MAK064
Lipid peroxidation (MDA) Assay Kit (Colorimetric)	Sigma Aldrich	Cat#AB118970
QIAamp DNA Micro Kit	QIAGEN	Cat#56304
Human APOB / Apolipoprotein B-100 ELISA Kit	Sigma Aldrich	Cat#RAB0610
HDL and LDL/VLDL Quantification Kit	Sigma Aldrich	Cat#MAK045
Oxidized LDL Assay Kit	Abcam	Cat#AB242302
Ketone body Assay Kit	Abcam	Cat#AB272541
OptiPrep™ Density Gradient Medium	Sigma Aldrich	D1556
Collagenase type IV from Clostridium Histolyticum	Sigma Aldrich	Cat#C5138
Clarity Western ECL Substrate	Bio-Rad	Cat#1705061
Hydroxyproline Assay Kit (Colorimetric)	Abcam	Cat#AB222941
Triglyceride Assay Kit - Quantification	Abcam	Cat#AB65336
Total Cholesterol & Cholesteryl Ester Colorimetric Assay Kit	Abcam	Cat#AB282928
DNA Damage Assay Kit (AP sites, Colorimetric)	Abcam	Cat#AB211154
Mouse TNF-alpha DuoSet ELISA	R&D System	Cat#DY410
Mouse IL-6 DuoSet ELISA	R&D System	Cat#DY406
Mouse IL-1 beta/IL-1F2 DuoSet ELISA	R&D System	Cat#DY401
Lipolysis Colorimetric Assay Kit	Sigma Aldrich	Cat#MAK211
Mouse Insulin ELISA Kit	Life Technologies-ThermoFisher Scientific (Waltham, United States)	Cat#EMINS
Experimental models: Organisms/strains
C57BL/6NTac	Taconic Biosciences	MASH-B6-CONTROL-M
C57BL/6NTac	Taconic Biosciences	MASH-B6-M
Software and algorithms
ImageJ analysis software	version 1.52 a	https://imagej.nih.gov/ij/
JMP (SAS, Cary, NC)	16.1 Pro	https://www.jmp.com/it/software
Graphpad Prism (San Diego, CA)	version 10.0	https://www.graphpad.com/

Experimental model and study participant details

Animals and diet information

C57Bl/6 male mice (Taconic Biosciences, NY, USA) were housed at constant room temperature (23°C) under 12-hour light/dark cycles with ad libitum access to water in compliance with the Principles of Laboratory Animal Care (NIH publication 86-23). C57BL/6N mice were fed an AMLN diet (40 kcal% fat, 20 kcal% fructose and 2% cholesterol) or standard diet (SD) for 14-22-28 weeks, starting from 6 weeks of age (n=10 mice/group). Food intake and body weight were recorded weekly. Before sacrifice, mice were fasted for 16 hours, and the interventions were done during the light cycle. Blood, liver, epididymal adipose tissue (EAT), heart and skeletal muscle from hind-limbs (gastrocnemius) samples were collected at the time of sacrifice. The experimental protocol was approved by Fondazione IRCCS Cà Granda, Ospedale Maggiore Policlinico Milano, and Italian Ministry of Health Review Boards (protocol 10/2017-UT).

Method details

Samples collection and biochemical evaluation

Blood samples were collected by tail vein puncture. Hepatic triglycerides and total cholesterol were measured by colorimetric assays (Abcam), according to the manufacturer’s instructions. At sacrifice, liver samples, epididymal adipose tissue (EAT), heart and skeletal muscle from hind-limbs (gastrocnemius) were rapidly removed, harvested and snapped frozen in liquid nitrogen. Samples of liver, heart and skeletal muscle were fixed in formalin and embedded in paraffin, for histological evaluation. 30 mg of tissue were homogenized in 0.5% Nonidet P-40 (NP-40) lysis buffer and triglyceride and cholesterol content were quantified by exploiting the Triglycerides Colorimetric Assay Kit and Cholesterol Colorimetric Assay Kit, respectively. To measure triglyceride concentration, lysates were incubated with lipase for 20 minutes and then with the reaction mix for 30 minutes. The triglycerides were converted to free fatty acids and glycerol. Glycerol is then oxidized to generate a product which reacts with a probe to generate color (spectrophotometry at λ= 570 nm). Data were interpolated to the standard curve and concentration of triglycerides in nmol/μl was calculated as: (B/V) ∗ D. B= amount of triglycerides in the sample well calculated from the standard curve in nmol. V= sample volume added in the sample well. D=dilution factor. Results were normalized on the amount of protein obtained from the homogenized tissues. Data were expressed as nmol/mg. To assess cholesterol amount, lysates were incubated with the total cholesterol reaction mix with a specific probe for 60 minutes at 37 C generating color (570 nm). Data were interpolated to the standard curve and concentration of cholesterol (μg/l) in the samples was calculated as follow: (A/V) ∗ D. A=amount of cholesterol determined from Standard Curve. V= sample volume added in the sample well. D=dilution factor. Results were normalized on the amount of protein obtained from the homogenized tissues. Data were expressed as μg/mg. Lipolysis was measured by using Lipolysis Colorimetric Assay Kit (Sigma Aldrich, MAK064, St. Louis, MO) which contains a synthetic catecholamine (isoproterenol) that activates β-adrenergic receptors. This activates adenylate cyclase, converting ATP to cAMP. cAMP then activates the hydrolysis of triglycerides by hormone-sensitive lipase. Lipolysis is determined by measuring a colorimetric product with absorbance at 570 nm (A570) proportional to the amount of glycerol present. Data were interpolated to the standard curve and concentration of glycerol (nmol/μl) in the samples was calculated as follows: S_a/S_v. S_a=amount of glycerol determined from Standard Curve. S_v= sample volume added in the sample well. Results were normalized on the amount of protein obtained from the homogenized tissues. Data were expressed as nmol/μg.

30 mg of hepatic tissue were homogenized in 0.5% Nonidet P-40 (NP-40) lysis buffer and 5 μl of sera samples were diluted in the Assay Diluent to measure ApoB with APOB / Apolipoprotein B-100 ELISA Kit (Sigma Aldrich; RAB0610) through enzyme linked immunosorbent assay (ELISA), which detects the hepatic ApoB-100, according to the manufacturer’s instructions. The lower limit of ApoB detection was 2.5 ng/mL. Data were interpolated to the standard curve and concentrations of ApoB in ng/mL were calculated.

In another experimental setting, 5 μl of sera samples were diluted in the Assay Buffer to quantify HDL and LDL fractions by exploiting the colorimetric HDL and LDL/VLDL quantification Kit (Sigma Aldrich; MAK045). HDL and LDL/VLDL were separated by using precipitation procedure and were measured at 570 nm). Data were interpolated to the standard curve and total serum cholesterol, HDL and LDL/VLDL separated fractions have been calculated as follows: (S_a/S_V) ∗D. S_a=amount of cholesterol in the sample well calculated from the standard curve in μg. S_V= sample volume added in the sample well in μl. D=dilution factor. Data were expressed as μg/μl.

Then, other 5 μl of sera samples were diluted in the Assay Buffer to quantify Oxidized LDL by using Oxidized LDL Assay Kit (Abcam; ab242302) through ELISA, which detects oxidized phospholipids associated with LDL in plasma, serum or other biological fluid samples. The lower limit of oxidized phospholipids detection was 15 ng/mL. Data were interpolated to the standard curve and concentration of oxidized LDL in ng/mL was calculated as follows: B/V∗D. B=amount of LDL-ox in the sample well calculated from the standard curve in μl/ml. V= sample volume added in the sample well in μl. D=dilution factor.

Histology

Hepatic, muscle and cardiac tissue samples were fixed in 10% PBS buffered formalin. All tissues were embedded in paraffin within 24 hours of formalin fixation. Tissue sections were stained by Hematoxylin and Eosin (H&E) to assess liver pathology, whereas liver fibrosis was revealed by Red Sirius staining, and it has been staged according to Kleiner et al. (2). For adipose tissues, we employed frozen samples embedded in OCT to stain collagen fibers. The deposition of the latter was quantified by ImageJ analysis software (https://imagej.nih.gov/ij/) in 10 random micrographs per each sample (magnification 100x) by calculating the Red Sirius positive area, as percentage of pixels above the threshold value with respect to the total pixels per area.

Isolation of murine primary hepatocytes

In n=2 mice/group for each time point primary mouse hepatocytes were isolated by a multi-step ethylene glycol tetra-acetic acid (EGTA)/collagenase perfusion technique, through microcannulation of portal vein to ensure efficient perfusion of tissues. Cells were centrifuged at 50g x 3 minutes for three times to separate hepatocytes (on bottom) and non-parenchymal cells (NPCs, on top). Next, cell suspensions were FACS sorted by using FACSanto to remove dead cells and cellular debris, exploiting 7-Aminoactinomycin D (7-AAD) staining. CD45+positive cells were removed by hepatocyte fractions [50, 51].

Hepatocytes were cultured on plastic in DMEM medium supplemented with 10% FBS, 2 mM L-glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin: 100 nM insulin and 100 nM dexamethasone. Isolated cells’ vitality was evaluated using flow cytometry and was around 99.5%. Isolated hepatocytes were fixed in 4% paraformaldehyde, and then, Oil Red O (ORO) was exploited to visualize lipid accumulation, and ORO positive areas were quantified by ImageJ software (version 1.52 a) in 10 random micrographs (magnification 400x).

Transmission electron microscopy (TEM)

From each mouse, hepatic and skeletal muscle biopsies were isolated and fixed in 2.5% glutaraldehyde in cacodylate buffer pH 7.4, overnight. Afterwards, they were postfixed in 2% osmium tetroxide (OsO4) for 1 hour. Finally, liver specimens were dehydrated with increasing ethanol series, embedded in an Epon resin and polymerized in an oven at 62°C for 48 hours. Ultrathin (70–90 nm) sections were collected on nickel grids, stained with Uranyl Acetate Replacement (UAR), Lead Citrate and observed with a Hitachi HT7800.⁴²

Hydroxyproline content evaluation

Tissue samples (between 30 and 50 mg) were homogenized in water and hydrolyzed in 6 M HCl at 120°C overnight. After centrifugation at 10’000 g for 3 minutes, 30 μl of samples were moved to a 96-well plate and oxidized with a solution containing 6% Chloramine T, for 5 minutes at room temperature. Thereafter, a solution with 4-dimethylamino-benzaldehyde (DMAB) and perchloric acid/isopropanol was added. The final mixture was incubated at 60°C for 90 min and then the absorbance was determined at 560 nm, by the spectrophotometric method of Bergman and Loxley (Bergman and Loxley). Standard solutions containing 0 (blank), 0.2, 0.4, 0.6, 0.8, 1.0 μg/well of 4-hydroxy-L-proline were treated likewise. The standard curve was linear in this range (r = 0.98). Results were normalized on the amount of protein obtained from the homogenized tissues. The value of the liver hydroxyproline level was expressed as μg/mg.

Serum insulin detection

5 μl of sera samples were diluted in the Assay Diluent and 400 μIU/mL of standard solution was prepared to produce a dilution series and proceed with Mouse INSULIN ELISA Kit (Invitrogen, ThermoFisher-Scientific, USA). Both samples and standard curve were added to Mouse Insulin Antibody Coated wells and incubated for 2.5 hours at room temperature with gentle shaking. Then, wash 4 times with 1X Wash Buffer and incubate for 1 hour at room temperature with prepared biotin conjugate. Add 100 μL of prepared Streptavidin-HRP solution and incubate 45 minutes at room temperature. Finally, add 100 μL of TMB Substrate, incubate for 30 minutes at room temperature in the dark, add 50 μL of Stop Solution and read the absorbance at 450 nm. The value of insulin in the sera was expressed as μlU/ml.

Cytokine secretion

5 μl of sera samples were diluted in Assay Diluent to quantify cytokine secretion through ELISA assays to detect TNF-α (DY410), IL6 (DY406), and IL1β (DY401) (R&D Systems, Minneapolis, USA). The lower limits of detection were 15.6, 3.9, and 9.4 pg/mL, respectively. Data were interpolated to the standard curves and concentration of TNF-α, IL6, and IL1β in pg/mL were calculated.

Mitochondria isolation and mitochondrial enzyme activity

Mitochondria were isolated from tissue lysates (30 mg) by exploiting a Mitochondria Isolation Kit (ab110170; Abcam, Cambridge, UK). The isolated mitochondria were suspended in the RIPA buffer (25 mM Tris-HCl pH 7.6, 150 mM NaCl, 1%NP-40, 1% sodium deoxycholate, 0.1% SDS). Mitochondrial protein content was estimated by BCA protein assay kit from ThermoFisher-Scientific and 10 μg of mitochondrial protein were seeded in a 96-well plate in quadruplicate for the measurement of enzymatic activities (Citrate synthase, complex I, III and ATP synthase). In detail, isolated mitochondria were used to evaluate changes in maximal citrate synthase activity using a colorimetric assay (ab239712, Abcam, Cambridge, UK). Citrate synthase is the initial enzyme of the TCA cycle that catalyzed the reaction to form citric acid from acetyl CoA and oxaloacetic acid. This reaction results in the formation of CoA with a thiol group that reacts with the DTNB, generating a yellow product (λ=412nm). The citrate synthase activity was calculated by interpolating data on the standard curve, and results were normalized on sample volume (μL). Data were expressed as nmol/min/μl. Similarly, mitochondrial complex I and complex III activities have been evaluated by colorimetric assays (at 600 nm and 550 nm respectively), according to manufacturer’s protocols (ab287847 and ab287844, Abcam). The absorbance was measured in kinetic mode at an interval of 1 min. The specificity of enzymatic activities has been tested by using Rotenone and Antimycin A as specific inhibitors of the mitochondrial complex I and III, respectively. Data were interpolated to the standard curves and then net complexes activities were measured as the difference between the activity assessed in the reaction without the inhibitor and that in the reaction with the inhibitor. Results were normalized on the amount of mitochondrial proteins (μg). The activities were expressed as nmol/min/μg.

ATP synthase activity

The oligomycin-sensitive ATPase activity was evaluated by using an assay coupled with pyruvate kinase which converts the ADP to ATP and produces pyruvate from phosphoenolpyruvate on isolated mitochondria (ab109714, Abcam). The absorbance was measured at 340 nm in kinetic mode at an interval of 1 min. The specificity of enzymatic activity has been tested by using Oligomycin as a specific inhibitor of the mitochondrial complex V (ATP synthase). Data were interpolated to the standard curve and then net complex V activity was measured as the difference between activity measured in the reaction without the inhibitor and that in the reaction with the inhibitor. Results were normalized on the amount of mitochondrial protein (μg). Data were expressed as OD/min/μg.

Total ATP levels were measured in tissue lysates by using the colorimetric ATP assay kit (ab83355, Abcam). 30 mg of tissue were homogenized in 0.5% Nonidet P-40 (NP-40) lysis buffer, deproteinized, and neutralized with perchloric acid (PCA) and KOH, following the manufacturer’s instructions. ATP concentration was determined by phosphorylating glycerol, resulting in a colored product (570 nm) proportional to the amount of ATP present. The latter was calculated by interpolating data on the standard curve and the concentration of ATP (nmol/μL) was calculated as: (B/V∗D) ∗ DDF where B = amount of ATP in the sample well calculated from standard curve (nmol or mM), V = sample volume added in the sample well, D = sample dilution factor and DDF = deproteinization dilution factor. Results were normalized on the amount of protein (μg). Data were expressed as nmol/μg.

OXPHOS complexes were tested in homogenized tissues by using total OXPHOS Rodent Western Blot Antibody Cocktail (ab110413, Abcam) and western blot analysis.

Seahorse assay

By Seahorse XF Cell Mito Stress Test (Agilent, Santa Clara, CA, USA) on live isolated hepatocytes, the oxygen consumption rate (OCR), a measurement of mitochondrial respiration and the extracellular acidification rate (ECAR) were determined in presence of specific mitochondrial activators and inhibitors. ATP synthase blocker (oligomycin) was used to determine proton leak. Mitochondrial uncoupler carbonyl cyanide 4-trifluoromethoxy-phenylhydrazone (FCCP) was used to measure maximum respiratory function (maximal OCR). Reserve capacity was calculated as a maximal OCR minus the basal respiration. The inhibitor of complex I (rotenone) and a blocker of complex III (antimycin A), were injected to completely abolish the mitochondrial respiration and to confirm that any changes in respiration is mediated by mitochondria. Data was analyzed by the Seahorse Wave Desktop Software (version 2.4).

Seahorse assay on frozen tissues

Frozen tissues (1 to 3 mg) were manually homogenized in 250 μl MAS buffer (70 mM sucrose, 220 mM mannitol, 5 mM KH2PO4, 5 mM MgCl2, 1 mM EGTA, 2 mM HEPES) and then the homogenates were centrifuged 5 min at 1000 × g, 4°C. The resulting supernatants were collected and stored at −80°C. Next, 80 μg of tissue homogenates were loaded in a Seahorse XF24 plate with other 20 μl MAS buffer. The plate was centrifuged 5 min at 2000 × g, 4°C without brake, using a plate carrier. 350 μl MAS buffer containing cytochrome c (10 μg/ml, final concentration) were added to allow complete membrane permeabilization to the substrates. The cartridge has been prepared by injecting 70 μl of different drugs as follow: port A: NADH (1 mM) (Complex I substrate) for complex I assessment; port B: Antimycin A (4 μM) (Complex III inhibitor); port C: TMPD (Electron donor to cytochrome c/Complex IV) + ascorbic acid (Maintains TMPD in the reduced state) (0.5 mM TMPD + 1 mM ascorbic acid); port D: Azide (Complex IV inhibitor) (50 mM). All compound injections were diluted in MAS buffer. The measure times for the Seahorse run protocol was customized as follow: cycle 1: wait 2 min, mix 2 min, wait 2 min; measurement period 1: mix 50 sec, measure 4 min (x 4 cycles); STATE 2: injection 1 (port A), measure after injection: mix 50 sec, measure 4 min (x 4 cycles); STATE 3: injection 2 (port B), measure after injection: mix 50 sec, measure 4 min (x 4 cycles); STATE 4: injection 3 (port C), measure after injection: mix 50 sec, measure 4 min (x 4 cycles); STATE 5: injection 4 (port D), measure after injection: mix 50 sec, measure 4 min (x 4 cycles) (Kinetic representation). The total area under the curve (AUC) and standard deviation (SD) have been calculated.

Oxidative stress

30 mg of tissue were homogenized in 0.5% Nonidet P-40 (NP-40) and oxidative stress was evaluated by assessing reactive oxygen species (ROS)/reactive nitrogen species (RNS) and Hydrogen Peroxide concentrations in tissue lysates, following the manufacturer’s instructions (ab238535 and ab102500, respectively, Abcam). Lipid peroxidation was quantified by the measure of reactive aldehydic derivatives (i.e., Malondialdehyde, MDA (ab118970, Abcam). Results were normalized on the amount of protein obtained from the homogenized tissues. Data were expressed as follows: μM/μg (ROS/RNS, H₂O₂) and nmol/mg (MDA). The apurinic/apyrimidinic (AP) sites were assessed at 450 nm in 100 μg/mL of purified genomic DNA extracted from liver samples by using the DNA damage – AP sites – Assay Kit (ab211154, Abcam, Cambridge, UK). Conversely, mitochondrial DNA damage induced by ROS was revealed by measuring the amount of ccf-mtDNA in serum samples.

Evaluation of lactate dehydrogenase activity and lactate production

30 mg of tissue were homogenized in 0.5% Nonidet P-40 (NP-40) and lactate Dehydrogenase (LDH) activity was measured by LDH Assay Kit (ab197000, Abcam) which exploits the conversion of lactate into pyruvate and NADH. 50 μL of hepatic homogenized tissues were plated in quadruplicate in 96 well and then, the reaction mix containing the LDH assay buffer, the PicoProbe and LDH substrate mix was added to measure the fluorescence reaction at Ex/Em 535/587 nm for 10-30 minutes at 37°C. LDH activity was normalized on the amount of protein obtained from the homogenized tissues. Data were expressed as mU/mg. The amount of hepatic lactate production and the lactate release were measured in total tissue lysates or in serum samples by employing colorimetric (570 nm) Lactate assay kit (MAK064, Sigma Aldrich). Results were drawn by the standard curve and lactate concentrations were calculated according to the protocol and expressed as nmol/μL.

Measurement of cell free circulating mitochondrial DNA (ccf-mtDNA)

Ccf-mtDNA was extracted from 200 μl of diluted serum samples through QIAmp DNA Mini Kit (Manchester, UK) as previously described.⁵ The Protease solution and AL Buffer were added to sera and incubated 56°C for 10 minutes to disrupt protein-DNA interactions. The DNA fragments up to 50 kb were trapped in columns providing silica membrane followed by two rinsing steps, while contaminants were removed in the flowthrough. Subsequently, DNA was eluted in water and ccf-mtDNA concentration was measured by Nanodrop 1000 microvolume 42 spectrophotometer (ThermoFisher-Scientific, USA). 20 ng of ccf-mtDNA and PowerUp SYBR Green Master Mix were exploited for the quantitative real time PCR (RT-qPCR) assay with QuantStudio 3 Real-time PCR system. Specifically, the thermocycling setup was set as follows: 95°C for 10 minutes, 40 cycle: 95°C for 10 minutes, 60°C for 1 minute, 95°C for 15 second, 60°C for 1 minute and 95°C for 15 second. All reactions were delivered in quadruplicate. The following primers were designed for amplifying the mitochondrially-encoded cytochrome C oxidase III (ccf-COXIII) and the Mitochondrially Encoded NADH:Ubiquinone Oxidoreductase Core Subunit 1 (ccf-ND1) fragments, coding subunits of the mt-respiratory chain: forward 5’-TGACCCACCAATCACATGC -3’ and reverse 5’- ATCACATGGCTAGGCCGGAG -3’ (ccf-COXIII); forward 5’- CCCTAAAACCCGCCACATCT -3’ and reverse 5’- GAGCGATGGTGAGAGCTAAGGT -3’ (ccf-ND1).

Ccf-COXIII and ccf-ND1 serum concentrations were obtained by interpolating data to a DNA (mice) standard curve with the following points: 163.000 picograms (pg), 58.000 pg, 29.999 pg, 14.500 pg and 7.250 pg. The amount of ccf-mtDNA encompassing the sum of COXIII and ND1 quantities, was divided with the size of PCR fragments (bp; 102 bp COXIII, 109 bp ND1) and the molar mass per base pair (g mol ⁻ 1).⁴³ The product was finally multiplied with Avogadro’s constant to obtain the quantity (picogram) of DNA fragments expressed as logarithmic. Data were expressed as copies/μL∗10ˆ4.

Western Blot analysis

Total protein lysates were extracted from 50 mg of tissues, using a RIPA buffer containing protease and phosphatase inhibitors (Roche). Equal amounts of proteins (50 μg) were separated by SDS-PAGE, transferred electrophoretically to the nitrocellulose membrane (BioRad, Hercules, CA) and incubated with specific antibodies. An equal amount of samples (n=10/group) were pooled prior to electrophoretic separation, and all reactions were performed in duplicate and ran together in the same gel. At least four independent experiments were conducted. Protein levels were quantified by ImageJ software (version 1.52 a) and normalized on vinculin (Rabbit polyclonal Anti-Vinculin 1:5000, ab130007 Abcam), as a housekeeping gene. Antibodies and concentration used are listed in Table S2.

To optimize protein extraction from adipose tissue, we applied a Removal of Excess Lipids (RELi) method to reduce lipid contamination.

Quantification and statistical analysis

Statistical analysis of animal experiments

Data are represented as mean ± standard deviation (SD). Statistical analyses were performed using JMP 16.1 Pro (SAS, Cary, NC) and Graphpad Prism (version 10.0, San Diego, CA) by using one or two-way analysis of variance (ANOVA) or chi-square test, where appropriate. p-values were corrected for multiplicity by Tukey’s honestly significant difference (HSD) multi-comparison post hoc test, and adjusted p-values < 0.05 were considered statistically significant.

Published: August 27, 2025