Melatonin promotes skeletal muscle mitochondria-associated endoplasmic reticulum membrane contacts and restores organellar membrane integrity and phospholipid composition in Zücker diabetic fatty rats of both sexes

Abstract

Background

Mitochondria-associated endoplasmic reticulum (ER) membranes (MAMs) are essential for cellular homeostasis and play crucial roles in lipid metabolism and calcium signaling. In skeletal muscle (SKM), MAMs contribute to metabolic flexibility, and their disruption is implicated in metabolic disorders such as diabesity (obesity and its associated type 2 diabetes). Melatonin, known for its antioxidant, circadian rhythm and metabolic regulatory effects, has been shown to preserve ER and mitochondrial function. This study explored whether melatonin could restore organellar membrane integrity, phospholipid composition in SKM, and MAM reassembly under diabesity conditions.

Methods

Five-week-old male and female Zücker diabetic fatty (ZDF) and lean (ZL) rats were subdivided into untreated control group (C) and treated with oral melatonin group (M, 10 mg/kg/day in drinking water for 12 weeks). MAM integrity and ER‒mitochondria interactions in the vastus lateralis (VL) muscle were assessed via both in situ proximity ligation assay (PLA) and transmission electron microscopy. Mitochondrial membrane integrity was evaluated by measuring mitochondrial permeability transition pore (mPTP) opening and cytochrome C release. Mitochondrial and ER phospholipid composition was analyzed using ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC‒MS/MS).

Results

In this study, we observed for the first time that melatonin promoted MAM assembly and ER‒mitochondria contacts in the VL from both sex obese diabetic rats. It also improved mitochondrial membrane integrity and reduced cytosolic cytochrome C release. In mitochondria, melatonin restored the oxidized cardiolipin (CL_ox) to total cardiolipin (CL) ratio by decreasing CL_ox and increasing total CL and its species (CL (72:7), CL (76:12)). In the ER, melatonin restored the phosphatidylcholine (PC) to phosphatidylethanolamine (PE) ratio by reducing total PC, and specific PC species (PC (38:2) and PC (42:6)) while increasing total PE and specific PE species (PE (34:0), PE (36:2), PE (40:7), and PE (40:8)).

Conclusions

These findings highlight the role of melatonin as a therapeutic agent for preserving SKM mitochondrial membrane integrity and its potential effect on interorganellar communication via MAM assembly during diabesity. Furthermore, melatonin contributes to the maintenance of proper phospholipid composition, promoting an ideal environment for proper MAM assembly.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12964-026-02799-y.

Background

There is increasing evidence for a link between organellar dysfunction, interorganellar miscommunication, and the pathogenesis of metabolic disorders such as obesity and type 2 diabetes mellitus [1]. Mitochondria-associated endoplasmic reticulum membranes (MAMs) are critical sites of interaction between mitochondria and endoplasmic reticulum (ER), playing crucial roles in maintaining organellar communication and cellular homeostasis [2]. These specialized junctional membranes facilitate a range of essential cellular processes, such as maintenance of integrity, phospholipid composition and function of the organellar membranes, interorganellar phospholipid exchange, and cellular stress response [1].

In the context of diabesity (obesity and its related type 2 diabetes), skeletal muscle (SKM), a major key site for insulin action and energy expenditure, often exhibits defects in mitochondrial function and ER stress [3, 4]. This organellar dysfunction contributes to the progression of diabesity and is strongly associated with compromised organellar and MAM integrity, leading to altered cellular metabolism [5]. These disruptions underscore the critical importance of maintaining MAM function and membrane phospholipid composition for SKM metabolic health, providing a novel approach for the development of an effective treatment against diabesity.

MAMs are specialized interfaces that form tight connections between the mitochondria and the ER, leaving a gap of less than 30 nm between these organelles [6]. This proximity is maintained by specific membrane proteins that physically connect the two organelles at designated sites. The complex of key proteins that form MAMs involves the voltage-dependent anion channel 1 (VDAC1), which is located in the outer mitochondrial membrane, and the inositol 1,4,5-trisphosphate receptor 1 (IP3R1), which is located on the ER. Both proteins are linked via the molecular chaperone glucose-regulated protein 75 (GRP75), which facilitates the exchange of critical molecules such as phospholipids, allowing rapid communication between these organelles [6]. In addition, MAMs comprise a diverse array of proteins, underscoring their structural and functional complexity [7]. Notably, the mitochondrial inner and outer membranes also possess peculiar protein, fatty acid and phospholipid compositions [8]. Disturbances in these components can compromise membrane properties, integrity and fluidity; alter mitochondrial function and energy production; and negatively affect cell viability [9, 10]. Cardiolipin (CL) is a phospholipid located mainly in the inner mitochondrial membrane and is crucial for ATP production, membrane integrity maintenance, cytochrome c (Cyt C) binding, and control of mitochondria-induced apoptosis [11]. Under pathophysiological conditions involving increased oxidative stress, such as in diabesity [12], CL is oxidized (CL_ox), leading to a conformational change and translocation into the outer mitochondrial membrane [13]. Then, Cyt C loses its interaction with CL, decreasing mitochondrial membrane integrity and allowing Cyt C release into the cytosol, which activates a cascade of apoptotic signals and cell death [11]. In addition, CL_ox induces the opening of the mitochondrial permeability transition pore (mPTP), enabling mitochondrial calcium release [11].

In the ER, similarly to mitochondria, changes in membrane phospholipid composition can impact the protein folding, function, and activity of ER channels, further contributing to metabolic dysfunction [14]. One of the most studied channels is the Sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA), which internalizes calcium into the ER [15]. SERCA is an ER transmembrane pump with a key function in thermogenesis and SKM metabolic rate regulation [16], whose efficiency is altered depending on the phospholipid methylation rate and composition [17]. Phosphatidylethanolamine (PE) and phosphatidylcholine (PC) are the predominant phospholipids found in mammalian membranes, including the SKM ER [18]. Unbalanced synthesis of muscle PC, PE, and fatty acids in obesity alters the PE/PC ratio, increases muscle insulin resistance, and favors the progression of type 2 diabetes [18]. In addition, reduced PE levels inhibit SERCA activity, disrupting cellular homeostasis and muscle function [17]. Consequently, this synergistic effect of MAM disruption, alterations in the ER PC/PE ratio, and compromised mitochondrial membrane integrity (Cyt C release into the cytosol, together with mitochondrial accumulation of CL_ox) could have important consequences for the progression of diabesity.

Both mitochondrial and ER membrane integrity are essential for MAM architecture maintenance and ER–mitochondria contact assembly, and their membrane stability and fluidity are largely determined by their phospholipid composition [9, 10]. Alterations in membrane phospholipids can modify membrane curvature, fluidity, and stability, thereby influencing organelle apposition and the formation of functional ER–mitochondria contact sites [19]. In this context, obesity-associated lipid remodeling may contribute to MAM disruption not only through changes in tethering proteins but also via compromised membrane integrity. Increased susceptibility to mPTP opening reflects intrinsic mitochondrial membrane vulnerability [20], which may limit the ability of mitochondria to establish or maintain stable contacts with the ER, being an indirect indicator of mitochondrial membrane disruption.

Melatonin is a hormone that is produced mainly during the night in the pineal gland, but also it is generated in various other tissues and can be found in a range of dietary sources derived from animals, fungi, and plants, including spices [21]. While melatonin is widely recognized for its role in managing circadian rhythms, it is also associated with pharmacological effects because of its strong antioxidant properties, anti-inflammatory effects, and role as a metabolic regulator [22–24]. Our previous research demonstrated that melatonin restores organellar homeostasis by alleviating ER stress in the SKM of Zücker Diabetic Fatty (ZDF) rats [25], partially by increasing SERCA activity and expression [26], improving mitochondrial function and oxidative capacity, and restoring muscle mass and function [27, 28]. Melatonin was also found to improve, in male rodents, the plasma lipid profile [29] and the mitochondrial functionality of other insulin-sensitive tissues, including the liver [30] and heart, by inhibiting cardiolipin peroxidation and preventing mPTP opening and Cyt C release into the cytosol [11], revealing its role in regulating fatty acid and lipid synthesis and mitochondrial membrane integrity. Moreover, biophysical studies using phospholipid model systems have shown that melatonin, as an amphiphilic molecule, interacts with lipids in cellular membranes, increasing membrane fluidity and indicating a direct bilayer-modulating effect and a mitochondrial protective effect of melatonin, stabilizing organelle membranes [31].

Experimental evidence suggests that the protective effects of melatonin on organellar function are mediated by the regulation of MAM assembly, which contributes to improved function of both the liver [32] and heart in vitro [33]. However, the role of melatonin in regulating MAM assembly within SKM in obese-diabetic rats has not yet been investigated. Therefore, the present study aims to provide a descriptive characterization of obesity-associated alterations and melatonin effects on MAM assembly and number of contacts, membrane lipid composition, and mitochondrial membrane stability in the vastus lateralis (VL) muscle of both male and female obese-diabetic ZDF rats, without addressing the causality between these parameters. This animal model mirrors key features of human diabesity, including progressive insulin resistance, glucose intolerance, hyperinsulinemia, dyslipidemia, and elevated blood pressure [34].

Methods

Reagents

All reagents and chemicals used in this study were of analytical grade and the highest available purity. 2-Propanol, triethylamine, acetonitrile, and glacial acetic acid were purchased from Scharlau (Madrid, Spain). Standard (18:2)₄ cardiolipin, (18:1)₂ (∆9-Cis) phosphatidylethanolamine, and (18:1)₂ (∆9-Cis) phosphatidylcholine were purchased from Avanti Polar Lipids (Alabama, United States). Melatonin and other chemicals were obtained from Sigma‒Aldrich (Madrid, Spain).

Animals and experimental protocol

This research was conducted after obtaining ethical approval, following the protocols established by the University of Granada (Granada, Spain) and in line with the animal care and protection guidelines set forth by the European Union. The ethical approval code assigned to this research is 23/06/2021/096-CEEA.

Five-week-old ZDF rats (fa/fa) and lean littermates (ZL, fa/-) from both sexes were paired and housed in transparent plastic cages within a climate-controlled environment, maintained at 30‒40% relative humidity and a thermoneutral temperature range of 28–30 °C, with a 12-hour light/dark cycle (lights on at 7 a.m.) and ad libitum access to food and tap water. The animals were acclimatized to room conditions for one week. ZL and ZDF rats of both sexes were then randomly subdivided into four groups: untreated control rats (female C-ZL and C-ZDF, and male C-ZL and C-ZDF; n = 8) and chronically treated for 12 weeks with 10 mg/kg/day oral melatonin (0.066% v/v ethanol) in the drinking water (female M-ZL and M-ZDF, and male M-ZL and M-ZDF; n = 8).

Melatonin dose and treatment protocol were selected on the basis of previous dose‒response and time‒course studies conducted by our group [26, 29, 35], which revealed dose‒dependent mitochondrial and metabolic benefits and indicated that sustained, long-term treatment is necessary to fully elicit ER- and mitochondria-associated adaptations. Based on these findings, melatonin was administered chronically to ensure the stabilization of mitochondrial function and the subsequent structural remodeling of ER–mitochondria contact sites.

To ensure that both sexes of ZDF rats develop early and progressively parallel obesity, dyslipidemia, hyperglycemia, hyperinsulinemia, hypertension, and endothelial dysfunction, the animals were fed the manufacturer’s detailed diet. Type 2 diabetes manifests in these rats at approximately 9‒10 weeks of age, followed by insulinopenic diabetes at approximately 20 weeks of age, mimicking the clinical characteristics of human diabesity. Therefore, ZDF rats constitute a robust rodent model for metabolic syndrome research [36]. Melatonin solutions were freshly prepared three times a week, with dosages adjusted according to body weight and water intake throughout the study. The water bottles were wrapped in aluminum foil to prevent light exposure. At the end of the study, the animals were anesthetized with sodium thiobarbital (thiopental) and then euthanized.

Transmission electron microscopy (TEM)

The tissue samples were preserved in 2% paraformaldehyde (PFA) and 0.5% glutaraldehyde in 0.05 M cacodylate buffer (pH 7.4) at 4 °C for 4‒18 h. Following fixation, the tissues were treated with reduced osmium tetroxide followed by dehydration in ethyl alcohol. Spurr’s resin (Electron Microscopy Sciences) was prepared, mixed 1:1 with propylene oxide, and incubated overnight on a rotator. A 3:1 mixture of Spurr’s resin: propylene oxide was made, and the samples were infiltrated overnight. The final embedding was completed in pure Spurr’s resin (hard mixture) for 12–18 h, and the mixture was polymerized at 60 °C. The embedded samples were thin-sectioned (60–80 nm thickness) for electron microscopy using a diamond knife (Diatome) on a Reichert Jung E and analyzed using a TALOS F200C G2 transmission electron microscope from the Center for Scientific Instrumentation (CIC) at the University of Granada at an accelerating voltage of 80 kV. The ER‒mitochondria contacts were quantified using Fiji Software (v2.9.0, ImageJ, National Institutes of Health, Stapleton, NY, USA) by the proportion of the mitochondrial surface closely associated with the ER and standardized to the mitochondrial perimeter, and the number of MAM contacts per mitochondria. We also calculated the different types of MAMs according to the width of the contacts (0‒10, 10‒20, and 20–30 nm) [37]. For each animal (n = 5 per group), 8‒10 muscle fibers were randomly selected, and 10‒20 electron micrographs were acquired per fiber at ×10,000‒×25,000 magnification. This provided an average of 150‒200 mitochondria analyzed per rat.

In situ proximity ligation assay (PLA)

MAM contacts were quantified by measuring the proximity between VDAC and IP3R using in situ PLA. When these two proteins are in close proximity (< 30 nm), forming a complex, a fluorescent red dot appears in the histological sections, indicating a MAM contact between the mitochondria and the ER. VL muscles were fixed in 4% PFA overnight after tissue sample collection. The samples were subsequently dehydrated in an increasing series of ethanol from 70 to 100%, cleared with xylene, and embedded in paraffin. Four-micron-thick sections were cut using a LEICA RM2125 RTS Microtome, deparaffinized, and rehydrated. For heat-induced antigen retrieval, tissue sections were boiled in Citrate Buffer (10 mM Trisodium Citrate and 0.05% Tween-20, pH 6.0) at 100 °C for 20 min. When the slides were cooled, the samples were placed in H₂O for 5 min at room temperature. Then, the sections were blocked (Duolink^® Blocking Solution, cat# DUO82007, Sigma‒Aldrich, France) for 30 min at 37 °C and incubated overnight at 4 °C in a humidity chamber with a binary primary antibody mixture of anti-IP3R (rabbit, 1/10000, cat# ab5804, Abcam, Cambridge, UK) and VDAC1 (mouse, 1/20000, cat# ab14734, Abcam, Cambridge, UK). After one wash with Tris-Buffered Saline (TBS, 150 mM NaCl, 50 mM Tris-HCl) with 0.05% Tween-20 (TBS-T), the slides were incubated for 1 h at 37 °C with the complementary secondary antibody (Rabbit PLUS or Mouse MINUS). Proximity ligations were performed according to the manufacturer’s protocols (Duolink© cat# DUO92101, Sigma‒Aldrich, France). Preparations were mounted with Ibidi Mounting Medium with DAPI (cat# 50011, Ibidi, Germany), and fluorescence was analyzed with an Axiovert200M Zeiss inverted fluorescence microscope using the AxioVision program (z-stack imaging, microscopy magnification ×40). Fluorescent dot signals were quantified using Fiji Software (v2.9.0, ImageJ, National Institutes of Health, Stapleton, NY, USA). The negative controls included sections incubated with one primary antibody only. The muscle fiber surface was obtained from the intrinsic autofluorescence of the tissue in the green channel (488 nm excitation). Autofluorescence provided sufficient contrast to delineate individual fiber borders. Results were expressed as the number of dots per 100 µm² fiber surface. A minimum of 20 images were taken per sample of each rat, at least five rats from each experimental group were analyzed, and three independent series were performed to ensure reliability.

Isolation of the cytosolic, mitochondria-rich, and ER-rich fractions

For VL cytosolic, mitochondria-rich and ER-rich fraction isolation, 300 mg wet tissue samples were homogenized with a Teflon pestle (4 cycles of 6 s each) in 1 mL isolation medium (IM; 10 mM Tris, 250 mM sucrose, 0.5 mM Na₂EDTA, 1 g/L fatty acid-free bovine serum albumin (BSA), 1% protease inhibitor cocktail, and 1% phosphatase inhibitor cocktail, pH 7.4). VL fractions were isolated from the tissue homogenate by sequential centrifugation as previously described [26], following specific protocols for preserving membrane integrity [38]. After fraction isolation, the samples were aliquoted and stored at -80 °C. A protein concentration assay was performed using the Bradford method with BSA as the standard.

Measurement of calcium retention capacity

The calcium retention capacity, an indirect indicator of mitochondrial membrane stability, was assessed by measuring the sensitivity of mitochondrial permeability transition pore (mPTP) opening. Mitochondrial membrane integrity is required for correct MAM assembly. Therefore, mitochondria isolated from VL samples were used to specifically evaluate intrinsic mitochondrial membrane stability and calcium retention capacity under controlled conditions. Isolated mitochondria were resuspended in respiration buffer (250 mM sucrose, 10 mM HEPES, 5 mM KH₂PO₄, 10 mM KCl, 1 mM MgCl₂, 0.1 mM EGTA and 0.1% fatty acid-free BSA, pH 7.2) at a final concentration of 0.5 mg/mL. Then, 10 µL of respiratory substrate was added to maintain mitochondrial membrane potential to a final concentration of 5 mM glutamate and 2.5 mM malate. Finally, 0.25 µM of the fluorochrome BAPTA-Oregon Green 488 (cat# O6807, Thermo Fisher Scientific, Spain) was added to a final volume of 100 µL as previously described [39]. The mixture was incubated for 5 min at 37 °C under gentle agitation, and the baseline fluorescence (F₀) was initially measured without the addition of calcium. Subsequently, 5 pulses of 0.50 µM CaCl₂ were added, which produced an increase in fluorescence. Calcium was retained and accumulated in the mitochondria in the presence of ATP, decreasing the fluorescent signal. Mitochondrial calcium overload causes opening of the mPTP, leading to the release of large amounts of calcium, which can be quantified. Calcium release was quantified using the Infinite F200 fluorometer (TECAN, Swiss). In parallel, 1 µM cyclosporine A (CsA) (cat# 30024, Sigma‒Aldrich, Spain), a standard mPTP inhibitor, was added to each sample control. mPTP opening was quantified from signals over time and expressed as the area under the curve (AUC), after subtracting the baseline fluorescence (ΔF/F₀), to represent the total calcium released from the mitochondria from four replicates for each sample.

Protein expression analysis by Western blot

For Western blot analysis, 30 µg of cytosolic and mitochondria-rich fraction proteins underwent sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (SDS-PAGE). Then, the gels were transferred into a nitrocellulose membrane (Bio-Rad Trans-Blot SD, Bio-Rad Laboratories, Hercules, CA, USA) and blocked for 1.5 h at room temperature using Blocking Buffer (Phosphate Buffer Saline (PBS) containing 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 1.8 mM KH₂PO₄, supplemented with 0.1% Tween-20 (PBS-T) and 5% free-fatty milk). Overnight incubation in a cold chamber was carried out with anti-β-actin (cat# sc-81178; Santa Cruz Biotechnology, USA), as a loading control, and anti-Cytochrome C (cat# sc-13156, Santa Cruz Biotechnology, USA) primary antibodies, both generated in mice; and primary antibodies raised in rabbit against OPA1 (cat#SAB-5700860; Sigma‒Aldrich), DRP1 (cat#SAB-5700783; Sigma‒Aldrich), MFN2 (cat#SAB210801; Sigma‒Aldrich), and FIS1 (cat#HPA017430; Sigma‒Aldrich). All primary antibodies were used at dilutions of 1:1000 in the PBS-T with 10% Blocking Buffer. Following overnight incubation, the membranes were washed three times for 15 min in PBS-T and subsequently, incubated at room temperature for 1.5 h with an anti-mouse (cat# MBS674947, Donkey Anti-Mouse IgG, MyBiosource, USA) or anti-rabbit (cat# A16035, Donkey anti-Rabbit IgG, Thermo Fisher Scientific, Spain) horseradish peroxidase (HRP)-conjugated secondary antibody at a dilution of 1:2000 in PBS-T with 10% Blocking Buffer. Then, the membranes underwent three additional 15-minute washes in PBS-T, and specific proteins were detected using the Clarity Western ECL commercial kit (Bio-Rad, Spain) following the manufacturer’s protocol. Signal intensities were obtained by the Image Station 4000MM Pro Molecular Imaging system (Kodak, USA) and analyzed using Image J 1.33 software (National Institutes of Health, Bethesda, MD, USA).

Lipid extraction and characterization by UHPLC-MS/MS

Lipids were extracted from mitochondria-rich and ER-rich fractions using the Folch method with little variation [40]. Fractions were obtained from 500 mg of dissected tissue and cut into pieces of less than 1 mm³. The mitochondrial and ER fractions were dissolved in 5 mL of chloroform: methanol (2:1). The samples were subsequently vortexed for 15–20 min, and the resulting homogenate was centrifuged at 500 × g for 5 min. The liquid phase was recovered, and a 0.2 volume of 0.9% NaCl solution was added. After vortexing, the samples were centrifuged at 500 × g for 5 min, and the lower chloroform phase containing lipids was evaporated under a nitrogen stream.

UHPLC-MS/MS was performed using an ACQUITY UPLC I CLASS liquid chromatograph (Waters Corporation, United States) with a ZORBAX Eclipse XDB-C18, 5 μm, 400 bar, 4.6 × 150 mm analytical column (Agilent Technologies, United States). The column heater operated at 35 °C, and eluents A (180 mL acetonitrile, 20 mL water, 1 mL triethylamine, and 1 mL glacial acetic acid) and B (180 mL 2-propanol, 20 mL water, 1 mL trimethylamine, and 1 mL glacial acetic acid) were used. The gradient was formulated following the Minkler & Hoppel protocol, consisting of 50% B for 5 min, 50% B to 80% B over 10 min, 80% B to 100% B over 15 min, and holding 100% B for 10 min [41]. The flow rate was 400 µL/min. The standards and samples were reconstituted at 200 and 100 µL respectively, 50/50 eluent A/eluent B, and injected (5 µL) into a UPLC ACQUITY I CLASS (Waters Corporation, United States). The mass spectrometer used was the XEVO TQ-XS (Waters Corporation, United States) triple quadrupole with a spectra range of 1300‒1600 and 600‒900 Da for the mitochondrial and ER fractions, respectively. Molecular species of cardiolipin, phosphatidylethanolamine, and phosphatidylcholine were identified according to their m/z ratios. Data acquisition was conducted on a MassLynx SCN 943 data system version 4.2 for Windows (Waters, Spain).

Statistical analysis

All the experimental procedures were conducted at least in triplicate, with each experiment repeated a minimum of two times for each rat to increase the reliability of the results. The data obtained were expressed as the means ± standard deviations (SD). Comparative analyses among the different experimental groups were performed to evaluate the effects of obesity, melatonin treatment, and sex, as well as their potential interactions, via three-way analysis of variance (ANOVA), followed by Tukey’s post hoc test, after confirming assumptions of normality and homogeneity and including obesity status, treatment, and sex as independent factors. Differences between groups were considered statistically significant when the P-value was less than 0.05. Statistical analyses were conducted using SPSS version 22 for Windows (SPSS, Michigan, IL, USA).

Results

Melatonin enhances MAM assembly and ER‒mitochondria interactions in VL

MAM assembly was evaluated by in situ PLA targeting the VDAC1‒IP3R complex. As illustrated in Fig. 1a, immunofluorescence-based PLA analysis showed that in the VL from both sex obese-diabetic rats, MAM interactions were highly decreased compared to their respective lean control rats, also shown by the marked reduction in the number of fluorescent dots/100 µm² seen in Fig. 1b (female C-ZDF, 0.06 ± 0.03 vs. C-ZL, 0.25 ± 0.07; and male C-ZDF, 0.07 ± 0.03 vs. C-ZL, 0.27 ± 0.06; P < 0.01). Melatonin completely restored MAM integrity in both sex ZDF rats, reaching similar ER‒mitochondria contact levels to healthy ZL rats (female M-ZDF, 0.33 ± 0.08 and male M-ZDF, 0.29 ± 0.10; P < 0.01; Fig. 1b). No effect of melatonin on MAMs was observed in VL of male and female lean control rats (Fig. 1b).

Fig. 1.

Melatonin effects on VL MAM assembly in female and male obese-diabetic rats. a Representative images of MAM interactions measured by in situ proximity ligation assay (PLA). The red dots represent MAM assembly (< 30 nm) by ER‒mitochondria contacts (VDAC1‒IP3R complex). b MAM quantitative analysis of VDAC1‒IP3R complex dots per 100 µm² of skeletal muscle fibers. Scale bar = 50 μm, magnification ×40. Results are expressed as the means ± SD of 5 animals/group (n = 25‒30 images per sample). Dots represents the average value obtained from all images analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (**P < 0.01 for melatonin vs. control rats; ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats)

Moreover, ER‒mitochondria distances and MAM type according to the proximity of MAM contacts (0‒10, 10‒20, and 20–30 nm) were also evaluated by TEM, as shown in Fig. 2a and electronic Supplementary Material (ESM) Fig. 1. In obese-diabetic rats from both sexes, ER‒mitochondria minimum and mean distance in nm (minimum distance: female C-ZDF, 7.30 ± 2.55 and male C-ZDF, 5.88 ± 1.66; mean distance: female C-ZDF, 20.89 ± 3.71 and male C-ZDF, 16.40 ± 2.47) were increased compared to their respective lean control rats (minimum distance: female C-ZL, 2.07 ± 0.98 and male C-ZL, 2.21 ± 1.17; mean distance: female C-ZL, 11.47 ± 2.18 and male C-ZL, 10.37 ± 1.17; P < 0.01; Fig. 2b, c, respectively). Chronic melatonin treatment enriched ER‒mitochondria contacts during diabesity, as shown by the reduction of ER‒mitochondria minimum distance in the VL from both sex ZDF rats (female M-ZDF, 3.31 ± 1.10 and male M-ZDF, 2.59 ± 1.03; P < 0.01; Fig. 2b). Moreover, as shown in Fig. 2c, the VL ER‒mitochondria mean distance was also diminished after melatonin treatment in both sex ZDF (female M-ZDF, 9.89 ± 1.51 and male M-ZDF, 8.70 ± 1.06; P < 0.01) and ZL rats (female M-ZL, 6.90 ± 0.85 and male M-ZL, 7.87 ± 0.82; P < 0.01 and P < 0.05, respectively).

Fig. 2.

Melatonin effects on VL MAM contact restoration in female and male obese-diabetic rats. a Representative images of ER‒mitochondria interactions measured by transmission electron microscopy (TEM). Structural analysis by TEM of (b) the ER‒mitochondria minimum and (c) mean distance; the % MAM/mitochondria perimeter in the (d) total interface (0–30 nm), e 0–10 nm range, and (f) 10–20 nm range; and the MAM counts per mitochondria in the (g) 0–10 nm range and (h) 10–20 nm range. Scale bar = 500 nm. i Structural analysis by TEM of VL mitochondrial surface. Arrows indicate the mitochondrial matrix electron density. Results are expressed as the means ± SD at the total interface or according to the indicated ranges (n = 150‒200 mitochondria analyzed in n = 5 rats/group). Dots represents the average value obtained from all mitochondria analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (*P < 0.05 and **P < 0.01 for melatonin vs. control rats; ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats)

As illustrated in Fig. 2d, the mean proportion of VL mitochondrial membranes in contact with ER (% MAM/mitochondria perimeter) in total interface (0–30 nm) was lower in both sex obese-diabetic rats compared to their respective lean controls (female C-ZDF, 9.53 ± 2.79 vs. C-ZL, 22.76 ± 4.61; and male C-ZDF, 10.67 ± 2.81 vs. C-ZL, 43.18 ± 5.29; P < 0.01), as very close (0–10 nm range MAM) and close (10–20 nm range MAM) ER‒mitochondria contacts were also found to be decreased in the VL from both sex ZDF rats (% 0–10 nm MAM/mitochondria perimeter: female C-ZDF, 4.32 ± 1.09 and male C-ZDF, 3.80 ± 2.54; % 10–20 nm MAM/mitochondria perimeter: female C-ZDF, 2.35 ± 1.11 and male C-ZDF, 2.53 ± 1.21) compared to control ZL rats (% 0–10 nm MAM/mitochondria perimeter: female C-ZL, 8.83 ± 2.69 and male C-ZL, 26.45 ± 6.01; % 10–20 nm MAM/mitochondria perimeter: female C-ZL, 6.24 ± 1.73 and male C-ZL, 8.93 ± 2.58; P < 0.01; Fig. 2e, f, respectively). No differences in the proportion of VL mitochondrial membrane in contact with ER at 20–30 nm range were found among any of the groups (ESM Fig. 2a). Melatonin completely restored the proportion of VL mitochondrial membranes in contact with ER in total interface in both sex ZDF rats, increasing the 0–30 nm % MAM/mitochondria perimeter and reaching levels comparable to those of the ZL control rats (female M-ZDF, 24.41 ± 4.99 and male M-ZDF, 46.59 ± 5.66; P < 0.01; Fig. 2d). Furthermore, as shown in Fig. 2d, melatonin treatment also increased the VL 0–30 nm % MAM/mitochondria perimeter in female ZL rats (M-ZL, 37.93 ± 6.57; P < 0.05). The 0–10 nm % MAM per mitochondria perimeter was also augmented after melatonin treatment in the VL from both sex ZDF (female M-ZDF, 13.58 ± 5.18 and male M-ZDF, 29.66 ± 8.16; P < 0.05 and P < 0.01, respectively; Fig. 2e) and female ZL rats (M-ZL, 29.52 ± 5.90; P < 0.01; Fig. 2e). In ZDF rats from both sexes, the proportion of VL mitochondrial membranes in contact with ER in 10–20 nm range was also increased after melatonin treatment as shown in Fig. 2f, achieving values similar to the % MAM/mitochondria perimeter at same range found in lean control rats (female M-ZDF, 6.97 ± 2.17 and male M-ZDF, 8.41 ± 2.18; P < 0.01).

In addition, a structural analysis of MAM contact sites per mitochondria was performed by TEM. When considering the total interface distance (0–30 nm), no significant differences were observed in MAM counts per mitochondria within the VL across any experimental group (ESM Fig. 2b). However, at shorter distances, specifically the very close (0–10 nm) and close (10–20 nm) interfaces, MAM counts per mitochondria were significantly reduced in VL muscle from both female and male ZDF rats. In the 0–10 nm range, ZDF females and males exhibited counts of 1.06 ± 0.94 and 1.79 ± 0.21, respectively, compared to their ZL controls (female C-ZL, 3.52 ± 0.48 and male C-ZL, 4.22 ± 0.78; P < 0.01; Fig. 2g). Similarly, in the 10–20 nm range, ZDF females and males showed counts of 2.38 ± 0.63 and 2.63 ± 1.19, respectively, compared to their lean control rats (female C-ZL, 4.91 ± 1.11 and male C-ZL, 6.05 ± 0.95; P < 0.01; Fig. 2h). Notably, melatonin treatment restored MAM counts per mitochondria in ZDF rats of both sexes, increasing contacts in the 0–10 nm range (female M-ZDF, 3.34 ± 0.66 and male M-ZDF, 4.94 ± 1.06; P < 0.01; Fig. 2g) and in the 10–20 nm range (female M-ZDF, 5.46 ± 1.29 and male M-ZDF, 6.83 ± 1.17; P < 0.01; Fig. 2h). In the 20–30 nm range, no differences were observed between ZDF and ZL rats of either sex, neither was any effect of melatonin detected (ESM Fig. 2c).

Collectively, these intensive structural results point a reduction of MAMs in VL of male and female obese diabetic rats, which is restored by chronic melatonin treatment.

Melatonin restores the integrity and cardiolipin amount of the VL mitochondrial membrane

Loss of mitochondrial membrane integrity was evaluated by morphological alterations of mitochondrial membranes, alterations in Ca²⁺ retention capacity and mPTP activity (opening), and the Cyt C release from the mitochondrial membrane into the cytosol. Also, we explored whether melatonin treatment enhances sensitivity to, or works synergistically with, CsA, the standard mPTP inhibitor.

TEM images showed that obese-diabetic rats of both sexes had disrupted VL mitochondrial membrane integrity, as evidenced by the loss of mitochondrial matrix electron density (Fig. 2a). It was also observed in the VL of both sex ZDF rats that mitochondria were larger than in their respective lean control rats as observed by the increase in mitochondrial surface in µm² (female C-ZDF, 0.52 ± 0.15 vs. C-ZL, 0.13 ± 0.06; and male C-ZDF, 0.64 ± 0.12 vs. C-ZL, 0.14 ± 0.09; P < 0.01; Fig. 2i). Furthermore, both sex ZDF rats presented increased mitochondrial fusion and reduced fission as evidenced by the FIS1, DRP1, OPA1 and MFN2 protein expression patterns (ESM Fig. 3). In both sex ZDF rats, a higher accumulation of glycogen granules was found (ESM Fig. 1) and the VL mitochondria were more elongated and elliptical, whereas in ZL rats the VL mitochondria were rounded and circular as shown in Fig. 2a. After melatonin treatment, the VL mitochondria of both sex obese-diabetic rats practically recovered typical matrix density (Fig. 2a) and mitochondrial fission/fusion balance (ESM Fig. 3). Melatonin also decreased the percentage area of glycogen granules in muscle fibers (ESM Fig. 1c) and VL mitochondrial circular morphology and mean size in both male and female ZDF (female M-ZDF, 0.09 ± 0.08; and male M-ZDF, 0.05 ± 0.03; P < 0.01; Fig. 2i) and female ZL rats (M-ZL, 0.02 ± 0.02; P < 0.05; Fig. 2i).

As shown in Fig. 3a, the AUC of mitochondrial PTP opening curves (ESM Fig. 4) in VL mitochondria from both sex obese-diabetic rats was significantly higher compared to their respective lean control rats (female C-ZDF, 1219.8 ± 52.5 vs. C-ZL, 629.8 ± 53.2; and male C-ZDF, 1147.9 ± 37.9 vs. C-ZL, 732.0 ± 8.5; P < 0.01), indicating a lower Ca²⁺ retention capacity due to an earlier mPTP opening. Oral melatonin treatment causes a notable inhibition of Ca²⁺-induced opening of mPTP in isolated mitochondria from the VL muscle in both sex ZDF rats (female M-ZDF, 508.4 ± 82.3; and male M-ZDF, 291.8 ± 33.2; P < 0.01) and also in lean female (M-ZL, 405.5 ± 48.1; P < 0.05) and male rats (M-ZL, 221.8 ± 28.3; P < 0.01). As expected, CsA inhibited the mPTP opening in both sex and phenotype rats by 49 ± 4%, 69 ± 2%, 79 ± 4%, and 83 ± 1% in VL mitochondria from female ZL, female ZDF, male ZL, and male ZDF, respectively, when compared to their respective controls (P < 0.01). CsA also induced an even greater inhibition of the mPTP opening in isolated VL mitochondria from both sex and phenotype rats previously treated with oral melatonin (female M-ZL + CsA, by 64 ± 6%; and M-ZDF + CsA, by 78 ± 3%; and male M-ZL + CsA, by 83 ± 2%; and M-ZDF + CsA, by 88 ± 1%) compared to untreated-control (P < 0.01) and melatonin-treated rats (P < 0.05; Fig. 3a).

Fig. 3.

Melatonin effects on VL mitochondrial membrane integrity in female and male obese-diabetic rats. a Calcium-induced mPTP opening measurement. The effect of cyclosporine A (CsA) was included as a classical inhibitor of mPTP opening. b Densitometric quantification of the amount of mitochondrial Cytochrome C (Cyt C) (preserved form). c Densitometric quantification of the amount of cytosolic Cyt C (released form). d Ratio of Cytosolic Cyt C/Mitochondrial Cyt C. e Representative blots of Cyt C protein expression in the mitochondrial and cytosolic fractions. Results are expressed as the means ± SD of 8 animals/group in triplicate. Dots represents the average value obtained from the triplicates analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (*P < 0.05 and **P < 0.01 for melatonin vs. control rats; †P < 0.05 and ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats)

Moreover, Cyt C amount in VL mitochondrial fractions from both sex obese-diabetic rats (female, 0.17 ± 0.03; and male, 0.25 ± 0.05) was decreased compared to their respective lean control female (0.68 ± 0.19; P < 0.01) and male littermates (0.44 ± 0.05; P < 0.05; Fig. 3b). Melatonin increased mitochondrial Cyt C levels in obese-diabetic female (0.37 ± 0.02; P < 0.05) and male rats (0.75 ± 0.13; P < 0.01; Fig. 3b). Regarding VL cytosolic Cyt C, illustrated in Fig. 3c, both sex ZDF rats presented higher levels than control ZL rats (female C-ZDF, 1.17 ± 0.24 vs. C-ZL, 0.60 ± 0.06; and male C-ZDF, 1.55 ± 0.18 vs. C-ZL, 0.81 ± 0.05; P < 0.01). Consequently, the ratio of cytosolic/mitochondrial Cyt C was also increased in both sex obese-diabetic rats (female, 5.26 ± 0.82; and male, 5.97 ± 0.41) compared to lean control ones (female, 0.99 ± 0.13; and male, 1.50 ± 0.15; P < 0.01; Fig. 3d). Melatonin completely restored cytosolic Cyt C amount in both sex ZDF rats (female, 0.33 ± 0.18; and male, 0.76 ± 0.07; P < 0.01; Fig. 3c), also recovering the cytosolic/mitochondrial Cyt C ratio (female, 0.87 ± 0.17; and male, 1.01 ± 0.12; P < 0.01; Fig. 3d). Furthermore, melatonin treatment inhibited Cyt C release into the cytosol in lean female (0.34 ± 0.04; P < 0.05) and male rats (0.29 ± 0.11; P < 0.01), decreasing the cytosolic/mitochondrial Cyt C ratio (female M-ZL, 0.43 ± 0.13; and male M-ZL, 0.83 ± 0.11; P < 0.05). Melatonin effects on VL mitochondrial and cytosolic Cyt C amount can also be observed in blots from Fig. 3e.

CL content and its peroxidation rate are highly related to mitochondrial membrane integrity and function. A representative ion chromatogram of separated CL molecular species from VL mitochondria-rich fractions is illustrated in ESM Fig. 5a, in which a clear peak is observed and corresponds to the retention time of the CL standard used.

As shown in Fig. 4a, VL mitochondria from both sex obese-diabetic rats presented a higher ratio CL_ox/CL compared to lean counterparts (female C-ZDF, 0.116 ± 0.007 vs. C-ZL, 0.063 ± 0.002; and male C-ZDF, 0.124 ± 0.006 vs. C-ZL, 0.064 ± 0.003; P < 0.01) due to an increase in CL_ox amount (female C-ZDF, 13.21 ± 0.41; and male C-ZDF, 14.11 ± 0.36; Fig. 4b) and a decrease in total CL levels (female C-ZDF, 114.06 ± 0.81; and male C-ZDF, 113.71 ± 2.22; Fig. 4c) compared to their respective lean controls (female CL_ox, 8.57 ± 0.40, and CL 136.14 ± 3.12; and male CL_ox, 8.64 ± 0.16, and CL 136.03 ± 3.79; P < 0.01). After melatonin treatment, the ratio CL_ox/CL in VL mitochondria from both sex ZDF rats was completely restored (female M-ZDF, 0.066 ± 0.002; and male M-ZDF, 0.071 ± 0.002; P < 0.01), also being lowered in both sex ZL rats (female M-ZL, 0.051 ± 0.003; and male M-ZL, 0.050 ± 0.004; P < 0.05) as observed in Fig. 4a. Melatonin also reduced mitochondrial CL_ox amount in the VL from both sex obese-diabetic (female, 8.38 ± 0.83; and male, 8.98 ± 0.86; P < 0.01) and lean animals (female, 6.82 ± 0.33; and male, 7.37 ± 0.24; P < 0.05; Fig. 4b). Furthermore, as shown in Fig. 4c, melatonin increased total CL levels in VL mitochondria from both sex obese-diabetic rats (female, 126.88 ± 2.69; and male, 126.73 ± 2.72; P < 0.01) almost reaching CL levels of lean littermates.

Fig. 4.

Melatonin effects on VL phosphatidylcholine (PC) and phosphatidylethanolamine (PE) amounts in both sex obese-diabetic rats. a PC/PE ratio. b Total PC molecular species amount. c Total PE molecular species amount. Results are expressed as the means ± SD of 8 animals/group in triplicate. Dots represents the average value obtained from the triplicates analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (*P < 0.05 and **P < 0.01 for melatonin vs. control rats; ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats)

Regarding CL molecular species, only three CL species (CL (70:7), CL (72:7), and CL (76:12)) were significantly reduced in VL mitochondria from both sex obese-diabetic ZDF rats (CL (70:7) female, 1.85 ± 0.28; and male, 1.47 ± 0.16; CL (72:7) female, 10.50 ± 1.98; and male, 9.64 ± 0.24; CL (76:12) female, 0.84 ± 0.06; and male, 0.95 ± 0.06) compared to their respective lean control animals (CL (70:7) female, 6.48 ± 0.45; and male, 5.39 ± 0.86; CL (72:7) female, 23.65 ± 0.49; and male, 25.23 ± 2.48; CL (76:12) female, 3.73 ± 0.73; and male, 4.13 ± 0.13; P < 0.01; Fig. 4d, e,f, respectively). However, melatonin-treated obese-diabetic rats from both sexes presented higher levels of CL (72:7) in the VL mitochondrial membranes, as shown in Fig. 4e (female, 18.39 ± 1.08; and male, 15.06 ± 1.39; P < 0.05). Furthermore, melatonin also increased CL (76:12) levels in VL mitochondria from both sex ZDF rats (female, 2.76 ± 0.20; and male, 2.63 ± 0.40; P < 0.01; Fig. 4f). These melatonin effects on CL molecular species characterization in VL mitochondria from both sex lean and obese-diabetic rats can be also observed in ESM Tables 1 and in the analysis of the mass spectra derived from the ion chromatogram at the same retention time as CL standard used (ESM Fig. 6).

In summary, chronic melatonin treatment restores the integrity and cardiolipin amount of the VL mitochondrial membrane of both sex obese diabetic ZDF rats.

Melatonin increases the PC/PE ratio of the VL endoplasmic reticulum membrane phospholipids

Phospholipid content in ER membranes, in particular PC and PE, and their proportions are key for maintaining ER function. As shown by the representative ion chromatogram of separated phospholipids from VL ER-rich fractions illustrated in ESM Fig. 5b, all phospholipids were separated in less than 30 min. Peaks between approximately 18 and 25 min represent molecular species of PE and PC, corresponding to the retention time of the respective PE and PC standards employed.

As illustrated in Fig. 5a, the measured PC/PE ratio of VL ER-rich fractions was higher in both sex obese-diabetic ZDF rats (female, 5.21 ± 0.15; and male, 5.30 ± 0.15) than in their lean counterparts (female, 2.85 ± 0.20; and male, 2.96 ± 0.17; P < 0.01) due to an increase in total PC amount (female C-ZDF, 4650.67 ± 82.52 vs. C-ZL, 3330.70 ± 205.63; and male C-ZDF, 4808.54 ± 72.95 vs. C-ZL, 3507.87 ± 166.50; P < 0.01; Fig. 5b). Furthermore, total PE levels were found to be lowered in VL ER-rich fractions from both sex obese-diabetic animals (female, 892.54 ± 18.55; and male, 907.67 ± 23.59) compared to control lean ones (female, 1167.70 ± 36.87; and male, 1183.62 ± 84.41; P < 0.01; Fig. 5c). Interestingly, melatonin diminished the ratio PC/PE of the ER membranes in VL from both sex ZDF rats (female, 3.47 ± 0.17; and male, 3.48 ± 0.11; P < 0.01; Fig. 5a), decreasing total PC (female M-ZDF, 4248.02 ± 148.10; and male M-ZDF, 4377.53 ± 135.63; P < 0.05; Fig. 5b) and completely reversing total PE levels in obese-diabetic rats (female, 1224.61 ± 57.75; and male, 1258.98 ± 83.12; P < 0.01; Fig. 5c).

Fig. 5.

Melatonin effects on VL cardiolipin (CL) levels in female and male obese-diabetic rats. a Relation between the amount of oxidized CL (CL_ox) and the total amount of CL expressed as CL_ox/CL ratio. b Total CL_ox molecular species amount. c Total CL molecular species amount. d Relative CL (70:7) amount. e Relative CL (72:7) amount. f Relative CL (76:12) amount. Results are expressed as the means ± SD of 8 animals/group in triplicate. Dots represents the average value obtained from the triplicates analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (*P < 0.05 and **P < 0.01 for melatonin vs. control rats; ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats)

Concerning PE and PC molecular species characterization, PE (34:0), PE (36:2), PE (40:7), and PE (40:8) levels were significantly reduced in VL ER-rich fractions from both sex obese-diabetic rats (PE (34:0) female, 19.78 ± 5.07; and male, 27.93 ± 7.01; PE (36:2) female, 136.19 ± 8.15; and male, 147.07 ± 15.07; PE (40:7) female, 150.39 ± 9.02; and male, 157.44 ± 13.38; PE (40:8) female, 15.04 ± 0.51; and male, 23.05 ± 2.76) compared to their respective lean control animals (PE (34:0) female, 64.62 ± 4.27; and male, 64.16 ± 4.69; PE (36:2) female, 253.40 ± 18.11; and male, 244.34 ± 23.89; PE (40:7) female, 267.45 ± 16.36; and male, 269.93 ± 16.90; PE (40:8) female, 37.47 ± 0.86; and male, 41.90 ± 4.02; P < 0.01; Fig. 6a, b,d, e, respectively), while PE (36:4), PC (34:1), PC (38:2), and PC (42:6) levels were higher in both sex ZDF rats than in ZL ones (PE (36:4): female C-ZDF, 239.91 ± 26.05 vs. C-ZL, 101.58 ± 5.38; and male C-ZDF, 265.42 ± 21.19 vs. C-ZL, 124.32 ± 5.64; PC (34:1): female C-ZDF, 312.40 ± 2.91 vs. C-ZL, 146.34 ± 3.25; and male C-ZDF, 313.83 ± 4.67 vs. C-ZL, 143.55 ± 28.84; PC (38:2): female C-ZDF, 247.28 ± 10.54 vs. C-ZL, 120.54 ± 29.37; and male C-ZDF, 330.63 ± 14.55 vs. C-ZL, 134.92 ± 20.11; PC (42:6): female C-ZDF, 107.43 ± 6.31 vs. C-ZL, 35.84 ± 7.54; and male C-ZDF, 117.19 ± 4.15 vs. C-ZL, 44.59 ± 2.23; P < 0.01; Fig. 6c, f,g, h, respectively). Moreover, male obese-diabetic ZDF rats presented higher levels of PE (40:8) and PC (38:2) than female ones (P < 0.05; Fig. 6e, g, respectively). Melatonin treatment in both sex obese-diabetic animals restored the levels of most of the characterized phospholipids in the VL ER membrane, increasing PE (34:0) (female, 82.18 ± 10.30; and male, 77.91 ± 7.85; P < 0.01; Fig. 6a), PE (36:2) (female, 258.93 ± 29.84; and male, 240.46 ± 7.37; P < 0.01; Fig. 6b), PE (40:7) (female, 277.97 ± 9.66; and male, 316.47 ± 20.37; P < 0.01; Fig. 6d), and PE (40:8) (female, 38.54 ± 0.93; and male, 40.05 ± 2.62; P < 0.01; Fig. 6e).

Fig. 6.

Melatonin effects on VL phosphatidylethanolamine (PE) and phosphatidylcholine (PC) species in both sex obese-diabetic rats. a Relative PE (34:0) amount. b Relative PE (36:2) amount. c Relative PE (36:4) amount. d Relative PE (40:7) amount. e Relative PE (40:8) amount. f Relative PC (34:1) amount. g Relative PC (38:2) amount. h Relative PC (42:6) amount. Results are expressed as the means ± SD of 8 animals/group in triplicate. Dots represents the average value obtained from the triplicates analyzed per animal. Three-way ANOVA followed by Tukey’s post-test was performed to compare differences between groups (**P < 0.01 for melatonin vs. control rats; ††P < 0.01 for Zücker Diabetic Fatty (ZDF) vs. Zücker Lean (ZL) rats; ‡P < 0.05 for female C-ZDF vs. male C-ZDF)

Furthermore, melatonin significantly reduced the levels of PC (38:2) and PC (42:6) in the ER-rich fractions of the VL muscle in both male and female obese diabetic ZDF rats (PC (38:2): female M-ZDF, 156.42 ± 19.46 and male M-ZDF, 140.12 ± 15.36; PC (42:6): female M-ZDF, 47.04 ± 8.02 and male M-ZDF, 38.21 ± 11.48; P < 0.01; Fig. 6g, h), restoring phospholipid levels comparable to those of lean controls. Detailed characterization of melatonin’s effects on PE and PC molecular species in VL ER membranes across sexes and phenotypes is provided in ESM Table 1 and in the mass spectra derived from ion chromatograms at retention times corresponding to standard phospholipids (ESM Fig. 7).

Collectively, these results highlight the capacity of chronic melatonin treatment to restore the phospholipid content of ER membranes from the VL of both sex obese and diabetic ZDF rats.

Discussion

Melatonin is a mitochondrion-targeted antioxidant known for its protective effects on mitochondrial function [42]. In addition to its mitochondrial benefits, melatonin also enhances ER function by reducing ER stress and inhibiting apoptosis, thereby promoting cell survival [43–45]. However, the precise mechanisms underlying the protective effects of melatonin at the ER level have not yet been fully elucidated.

We previously demonstrated that chronic melatonin treatment confers organelle-specific protection in the VL muscle of obese-diabetic ZDF rats of both sexes, improving mitochondrial function [28] and alleviating ER stress [25]. In the present study, as expected, we further confirmed that melatonin restores MAM assembly disrupted by diabesity, increasing the number of ER‒mitochondria contact sites in the VL muscle of male and female ZDF rats. More specifically, here we demonstrated that chronic melatonin administration increases both the number of MAM counts per mitochondria and the proportion of the outer mitochondrial membrane in close contact with the ER (within < 20 nm), although the functionality of these promoted ER–mitochondria connections has not been explored. In addition, melatonin promotes ER‒mitochondria rapprochement by reducing both the minimum and mean distances between both organelles, especially very close contacts (0–10 nm). These findings may suggest a heightened regulatory role of melatonin in facilitating phospholipid exchange between these organelles [46], although this role has not been directly measured here, and we cannot distinguish between altered phospholipid biosynthesis or membrane trafficking, which is a limitation of the current study. A recent human study investigating the expression and genetic polymorphisms of canonical phospholipid biosynthesis genes (PCYT1A, PCYT2, PTDSS2, and PEMT) revealed modest correlations between expression levels of some gene and obesity but, overall, did not provide strong evidence that alterations in de novo phospholipid synthesis are determinants of metabolic phenotypes in skeletal muscle. Only PEMT showed notable associations with adiposity traits and regulatory polymorphisms [47], suggesting that the regulation of specific enzymes in phospholipid exchange or remodeling pathways may be more relevant than global changes in phospholipid biosynthesis. However, further mechanistic experiments based on isotope tracers and key phospholipid biosynthesis enzyme activities are needed to corroborate this hypothesis and better understand how melatonin counteracts the effects of diabesity on membrane phospholipid composition. Similar modulatory effects of melatonin on MAM formation have been reported in models of ischemic heart injury [33] and mechanical trauma-induced hepatic damage [32], where melatonin restored normal ER‒mitochondria interactions. Under these pathological conditions, MAM contacts are abnormally increased, and melatonin decreases them to physiological levels. Taken together, our findings and those from our previous studies suggest that melatonin, by enhancing ER‒mitochondria interactions, may exert a MAM modulatory effect, fine-tuning organelle contacts to respond to cellular demands and maintain organellar homeostasis and function, although further studies focusing on MAM function will help decipher the effects of melatonin on ER–mitochondria interorganellar communication.

The role of MAMs in obesity-induced insulin resistance, the first step in type 2 diabetes, remains a topic of ongoing debate. Conflicting evidence exists regarding MAM alterations in two primary insulin-sensitive tissues: the SKM and, notably, the liver under conditions of obesity and/or diabetes [48–52]. These discrepancies highlight the dynamic nature of MAMs, whose assembly may be upregulated during the early stages of diabesity as an adaptive response. In this context, increased MAM formation may enhance mitochondrial function and/or facilitate processes such as autophagy and damaged cell apoptosis, both of which occur through increased calcium transfer into the mitochondria [49, 50]. Conversely, in more advanced stages of disease, MAM integrity may be compromised, leading to reduced ER‒mitochondria contacts and contributing to cellular dysfunction [48, 51, 52]. In the livers of mice fed high-fat and high-sucrose diets, alterations in MAM contacts within the 0–10 nm range were observed during the early weeks, whereas changes in distances greater than 30 nm emerged at later stages of diabesity [52]. Furthermore, in the gastrocnemius muscle of these mice, ER‒mitochondria interactions were reduced as soon as one week of high-fat and high-sucrose diet feeding [48]. Therefore, disrupted MAMs could also be observed in both SKM and liver in the early stages of the disease. Consistent with these findings, the present study revealed a reduction in the number of MAM contacts and an increase in the minimum and mean ER‒mitochondria distances in the VL muscle of 18-week-old obese-diabetic rats of both sexes. Here, we also showed that, under obese-diabetic conditions, enlarged mitochondrial sizes were found, leading to a lower proportion of the mitochondrial membrane in close proximity (within the < 20 nm range) to the ER. These findings together suggest a mid-to-advanced pathological stage, aligning with our previously published observations in the same rat strain, where by 12 weeks of age, key hallmarks of progressive lipotoxicity and glucotoxicity in diabesity, such as impaired glucose homeostasis, increased insulin resistance [53], dyslipidemia [29] and exacerbated low-grade inflammation [54], were evident.

Although similar alterations in MAMs have already been described by other groups in diabetic mouse models and in muscle biopsies from patients, notably, this study is the first to investigate the pharmacological modulation of MAM assembly by melatonin in both in the SKM of both sex obese and diabetic rodents. Our current data revealed that structural and functional impairments in ER–mitochondria interactions were consistently observed in both sexes. Importantly, melatonin treatment restored MAM integrity in male and female ZDF rats in a comparable way. These findings show that male and female ZDF rats exhibit similar degrees of diabesity development, including VL MAM architecture disruption, as expected by maintaining the animals under a sex-specific diet according to the manufacturer’s recommendation [55], and a sex-independent regulatory effect of melatonin on MAM dynamics, supporting the physiological relevance and translational potential of this intervention across sexes. Considering the increasing emphasis on sex as a biological variable in biomedical research, the present data highlight the robustness and generalizability of melatonin’s actions on SKM MAM remodeling.

Furthermore, in the VL of obese-diabetic ZDF rats, we previously reported an increase in mitochondrial fusion [27] associated with decreased mitochondrial functionality [28]. The findings of the present study also support this evidence of increased mitochondrial fusion under diabesity condition in the VL muscle for both sexes through descriptive mitochondrial structural analysis and fusion and fission marker measurements. In the mitochondrial structural analysis, although the magnification used does not allow a detailed morphometric analysis of cristae architecture, which represents a limitation of the present study, qualitative alterations in mitochondrial ultrastructure were observed in obese and diabetic animals. Specifically, mitochondria exhibit a swollen, enlarged and elongated appearance, increased mitochondrial clumping, and a reduction in matrix electron density commonly associated with the loss of cristae and internal membrane organization, supporting our previous studies in the same rat strain showing that VL mitochondria from obese-diabetic rats present a loss of function and oxidative capacity, leading to bioenergetic failure [28]. Therefore, the aberrant morphology of VL mitochondria found in obese animals could also be explained in part by the abovementioned mitochondrial dysfunction, in addition to the imbalance in mitochondrial dynamics. These structural abnormalities are also consistent with additional data from the present study showing compromised mitochondrial membrane integrity, as evidenced by elevated mPTP opening and increased Cyt C release into the cytosol in the VL of both male and female ZDF rats. Moreover, here, we report an increase in the CLox/CL ratio in the VL of the obese-diabetic rats of both sexes caused by elevated CL peroxidation and a reduction in total CL levels. Given that CL is a key phospholipid of the inner mitochondrial membrane essential for cristae formation, Cyt C stabilization, and proper respiratory complex function, the observed decrease in CL and increase in CL_ox levels could explain the membrane integrity and cristae loss observed in TEM images, mitochondrial swelling, and functional decline [13]. Similarly, when CL is peroxidized, it translocates to the outer mitochondrial membrane, where it promotes Cyt C release, mPTP opening, and the activation of apoptotic pathways. In addition, previous results from our group in the same rat strain also demonstrated an increase in the oxidative stress status [27] and in the activation of ER stress-induced apoptotic pathways [25] in the VL muscle of obese-diabetic ZDF rats. Furthermore, an increase in systemic oxidative and proinflammatory stress in ZDF rats and enhanced susceptibility to oxidative stress induced by nonselective hydroxyl radicals were found, increasing malondialdehyde blood levels and, therefore, lipid peroxidation [54]. All these data, together with previously published and current data, suggest that increased muscle oxidative stress in the dysfunctional mitochondria of ZDF rats increases nonselective lipid peroxidation, enhances membrane phospholipid peroxidation, including that of the CL, and promotes the activation of not only ER stress-mediated apoptosis but also mitochondrial stress-mediated apoptosis through mPTP opening and Cyt C release. However, in the present study, the peroxidation of other types of phospholipids was not evaluated in depth, and only oxidized phospholipid species with a relative abundance greater than or close to 1 could be identified, which is a limitation of the present study. Therefore, more studies are needed that focus on phospholipid peroxidation and apoptotic pathway activation to demonstrate the role of obesity in the lipid peroxidation of membrane phospholipids and to elucidate the critical role of the ER and mitochondria in muscle apoptosis.

These CL alterations are well-documented in SKM under obesity-related conditions [56, 57]. Recent findings indicate that CL synthesis in SKM is modulated by CLOCK genes [58], highlighting the potential regulatory role of circadian modulators such as melatonin in CL homeostasis. These findings, along with the present data and those of previous studies from our group and others [11, 43, 59], support the protective role of melatonin in enhancing CL synthesis and preventing lipid peroxidation, mPTP opening, and Cyt C release through its antioxidant and mitochondrial protective properties and not through direct effects on MAM regulation, thereby helping to restore mitochondrial function compromised by diabesity. In terms of specific CL species, melatonin was found to increase the levels of CL (72:7) and CL (76:12), effectively restoring the basal levels that were reduced in the VL of both male and female ZDF rats. Similar effects have been observed in rodent models of diet-induced obesity, where physical exercise increased these same CL species in the soleus muscle [60], further reinforcing the link between CL remodeling and metabolic interventions. Moreover, in the serum of patients with obesity and type 2 diabetes, a reduction in these CL species was also observed [61], suggesting that CL species profiles could be potential biomarkers of mitochondrial damage. Therefore, our present lipidomic analysis provides insights into how obesity and melatonin modulate membrane composition, with potential implications for MAM integrity. CL is essential for mitochondrial membrane curvature, integrity and structural stability [62], and its specific species were found to be broadly affected by diabesity without a clear pattern according to chain length or degree of unsaturation, suggesting a generalized disruption of mitochondrial membrane architecture and consequently contributing to an unfavorable environment for proper ER–mitochondria contact assembly. Notably, CL(72:8), the most abundant CL species and probably corresponding to tetralinoleoyl cardiolipin, presented the highest degree of peroxidation, which might be partly explained by its greater relative abundance in muscle mitochondrial membranes than other CL species, promoting its susceptibility to peroxidation by increasing the probability of interaction with reactive oxygen species. Furthermore, these findings are consistent with previous reports linking long-chain polyunsaturated fatty acids, specifically in CL species [63], with greater susceptibility to oxidative damage [64]. Melatonin partially restored the levels of key CL species, potentially contributing to the improved membrane stability and curvature necessary for MAM formation.

In addition, we previously demonstrated that melatonin increased mitochondrial fission in the VL of ZDF rats, thereby improving their function [27, 28]. Similarly, we also report here the role of melatonin in restoring a circular mitochondrial phenotype, decreasing its size and clumping, and improving matrix electron density, suggesting that melatonin promotes healthy cristae morphology and supporting our published data on melatonin’s effects on mitochondrial dynamics and function. However, additional structural analyses are needed to fully understand the effects of melatonin on mitochondrial size and cristae organization. A reduction in glycogen granule content was also observed in the current study by TEM analysis after melatonin treatment in the VL from both sex ZDF rats, being consistent with our previous results in the same rats in which melatonin promoted a switch from white glycolytic fibers to red oxidative ones [28]. Taken together, these results suggest that melatonin, by increasing oxidative metabolism, decreases the demand for glycogen and therefore its accumulation in the fibers of obese-diabetic rats, which, having a glycolytic phenotype, present large accumulations of glycogen granules. These metabolic changes in muscle phenotype are typically associated with a reorganization of mitochondrial positioning [65], which may also explain the differences in mitochondrial size and surface area observed in the present study. Further studies comparing longitudinal and transverse muscle sections are needed to decipher the effects of diabesity and melatonin on mitochondrial morphology and the reorganization of their positioning in muscle fibers in parallel with mitochondrial dynamics by obtaining a 3D virtual map of muscle fibers.

Recent studies have linked mitochondrial dysfunction in various diseases to impaired ER‒mitochondria communication, MAM disruption, compromised phospholipid synthesis and transfer between the two organelles, reduced membrane integrity and organelle function, and diminished cellular protection against lipotoxicity [66, 67]. Consequently, altered levels of PC and PE, the main phospholipids of the ER, in SKM are associated with impaired mitochondrial function, increased insulin resistance, and disrupted glucose homeostasis, contributing to glucotoxicity [68]; together, these alterations contribute to the progression of diabesity. Moreover, an increased PC/PE ratio in the ER has been observed in rodent models with impaired SKM SERCA activity [68, 69], which is associated with an altered PEMT-mediated conversion of PE to PC [47, 70, 71] and a decreased metabolic rate and increased obesity [17], as SERCA function is well known to play a key role in metabolism, thermogenesis and obesity prevention [72]. Consistent with these findings, our present results demonstrate that the PC/PE ratio in the VL from both sex ZDF rats is increased, driven by elevated PC and reduced PE levels in the ER. These findings support the hypothesis that MAM disruption is closely associated with altered membrane phospholipid composition. Furthermore, our previously published data revealed reduced SERCA expression and activity in the VL of the same rat strain, reinforcing the link between impaired phospholipid imbalance and mitochondrial dysfunction [26] and suggesting that the phospholipid composition in the ER modulates SERCA activity and the metabolic rate. In addition, a previous study from our group indicated that melatonin exerts direct and receptor-independent effects on cellular and organellar membranes, mainly through its antioxidant properties and physicochemical interaction with lipid bilayers, preserving organellar membrane fluidity and integrity, preventing lipid peroxidation and stabilizing phospholipid packing in hepatic ischemia–reperfusion injury [73]. Therefore, we suggest that melatonin, through its well-established antioxidant and mitochondrial protective effects, stabilizes organelle membranes by improving their integrity and composition and restores PC and PE levels and an optimal lipid environment, thereby promoting an improvement in SERCA function. Further mechanistic studies are still needed to decipher the complex correlation between phospholipid composition, membrane integrity, MAM assembly, and improved SERCA function.

For the specific species of PC and PE, ZDF rats presented higher levels of PC (34:1), PC (38:2) and PC (42:6) and lower levels of PE (34:0), PE (36:2), PE (40:7) and PE (40:8). PE promotes the negative membrane curvature critical for ER–mitochondria contacts [68, 74], and several specific PE species are dysregulated in the VL from both sex obese and diabetic ZDF rats. Reductions were observed across short-, medium-, and long-chain polyunsaturated species, whereas PE (36:4) levels were increased, highlighting that obesity affects multiple PE species without a simple trend in chain length or saturation. This perturbation of PE species likely compromises the local ER curvature and flexibility required for efficient ER–mitochondria contact formation, which may also lead to MAM architecture disruption. Increased levels of PC in membranes are related to the presence of a greater proportion of rigid bilayer-stabilizing domains [68, 74]. Specific PC species generally increased across short and long species, particularly the unsaturated forms. This elevation could reduce membrane flexibility and impair ER–mitochondria coupling by opposing the negative curvature promoted by PE. Similarly, an increased PC/PE ratio in human SKM is associated with metabolic syndrome and obesity, accompanied by a decrease in PE species containing highly polyunsaturated fatty acids such as PE (36:4) and PE (40:7) [75, 76]. Importantly, the obesity-induced increase in the PC/PE ratio in the ER of the VL from both sex ZDF rats was normalized following melatonin treatment, which restored specific PE and PC species to levels comparable to those in healthy lean control rats, notably by decreasing the abundance of long and polyunsaturated PC species, supporting a rebalancing of the ER membrane geometry that promotes an optimal environment conducive to proper ER–mitochondria proximity and MAM assembly and stability. Taken together, these results indicate that obesity induces broad and nonselective remodeling of mitochondrial and ER membrane lipids, destabilizing both the curvature and fluidity required for MAM assembly. However, melatonin restores phospholipid disturbances, thereby potentially enhancing MAM integrity. ER phospholipid balance restoration after melatonin treatment may also explain our previous findings showing that melatonin enhances SERCA expression and activity in the VL of the same rat strain, thereby contributing to the mitigation of obesity-related metabolic dysfunction [26]. Consistent findings have been reported in humans undergoing exercise-based anti-obesity and/or anti-diabetes interventions, where a reduction in the PC/PE ratio was observed, either through an increase in PE levels [77] or a decrease in PC levels [78]. The close interplay between mitochondrial morphology, ER‒mitochondria contact sites, and phospholipid homeostasis supports the hypothesis that melatonin’s protective effects on mitochondria in the VL of obese-diabetic rats contribute to the stabilization of organellar membranes, MAM assembly and the normalization of MAM-mediated phospholipid transport between both organelles. This likely results in an improvement in cellular homeostasis and muscle function and the mitigation of diabesity progression (as muscle glucose uptake is essential for blood glucose control), although these systemic effects were not measured in the present study.

A recent study revealed that melatonin may act by regulating the direct binding of MAM to IP3R, stimulating multiple cellular functions in the liver, such as the regulation of mitochondrial dynamics, autophagy, calcium flux, apoptosis, and lipid accumulation [79].

However, on the basis of the integrated analysis of our present results, we suggest that the mechanism by which melatonin modulates MAM assembly is not through pharmacological action on MAM but rather through an improvement in mitochondrial membrane stability. Melatonin is a well-established antioxidant and mitochondria-protective molecule, and its ability to reduce oxidative stress may contribute to the maintenance of mitochondrial membrane integrity and resistance to permeability transition. By stabilizing mitochondrial membranes, melatonin generates a structural environment that may favor the formation and maintenance of ER–mitochondria contacts and the restoration of membrane phospholipid composition.

Although our present and previous findings demonstrate that melatonin promotes both structural and functional recovery of mitochondria and, by extension, structural MAM integrity in the VL of male and female ZDF rats, especially in the 0–10 nm range which is mostly associated with phospholipid exchange-related MAM function, direct measurements of other key MAM structural proteins (such as GRP75) and MAM function, mechanistic experiments and quantitative assessments of ER-to-mitochondria calcium transport were not conducted, which is a limitation of the present study. Therefore, further studies are needed to establish a causal relationship between melatonin and MAM reassembly, as well as to elucidate the mechanisms involved in its role in improving interorganellar communication, MAM functionality, insulin sensitivity, muscle glucose uptake, lipid synthesis, transport and accumulation, apoptosis, and muscle contractile function, thereby clarifying the therapeutic potential of melatonin in the treatment of diabesity.

Moreover, as also reported in previous studies from our group in the same rat strain and treatment protocol, no detectable behavioral or locomotor alterations were found in treated animals [25]. Despite these favorable safety and behavioral observations, several limitations related to the experimental model should be considered when interpreting the translational relevance of the present findings. First, ZDF rats constitute a well-established and highly reproducible model of genetic diabesity that does not fully represent the multifactorial etiology and heterogeneity of human obesity. In addition, skeletal muscle is also a heterogeneous and flexible tissue with different fiber-type compositions, which may influence MAM architecture. Therefore, alterations identified in the VL may not apply uniformly to all human skeletal muscles or fiber populations. Importantly, the dose used in this study should be interpreted within a pharmacological rather than a physiological framework, as the observed effects are unlikely to be achieved at endogenous systemic melatonin concentrations. Although human equivalent dose estimations suggest that relatively high oral doses (15–150 mg/day) are required to elicit comparable effects in obese patients, melatonin displays a remarkably favorable safety profile in both preclinical and clinical settings [80]. Nevertheless, given the reported variability in melatonin’s circadian and metabolic effects depending on dose and timing, further well-controlled randomized clinical trials and chronopharmacological studies, as well as in vitro human studies, will be essential to define the optimal therapeutic window, safety, and clinical applicability of high-dose melatonin treatment and to validate the relevance of melatonin-mediated MAM regulation in humans.

Conclusion

In conclusion, this study demonstrated that melatonin restores MAM assembly in the VL of both male and female obese-diabetic ZDF rats, reversing key structural and functional deficits associated with diabesity. Chronic melatonin treatment fully reestablished ER‒mitochondria contacts and improved the mitochondrial ultrastructure and circularity, thereby normalizing the membrane phospholipid composition. Given the growing evidence that MAM regulation is essential for maintaining metabolic homeostasis, further studies directly assessing ER–mitochondria calcium transfer and causality using genetic and/or pharmacological MAM inhibitors and/or linkers will be crucial to further define the functional impact of MAM alterations observed in obesity and to determine whether melatonin could represent a potential therapeutic agent to preserve SKM organellar membrane integrity and interorganellar communication via MAM stabilization. This may be critical for sustaining muscle phospholipid synthesis and transport, and for mitigating metabolic dysfunctions driven by diabesity. However, deeper studies are needed to fully understand the mechanisms underlying melatonin effects on phospholipid synthesis, transport and peroxidation, and the effects of melatonin not only on MAM architecture assembly but also on MAM functionality, apoptosis, phospholipid synthesis and trafficking, mitochondrial dynamics and calcium homeostasis.

Moreover, further studies in human muscle models, along with mechanistic investigations, are necessary to elucidate the signaling pathways through which melatonin modulates MAM architecture and function. Collectively, these findings support the clinical potential of melatonin or melatonergic compounds in counteracting diabesity-induced SKM dysfunction, improving insulin sensitivity, and limiting disease progression.

Abbreviations

CL

Cardiolipin

CL_ox

Oxidized cardiolipin

CsA

Cyclosporine A

Cyt C

Cytochrome C

ER

Endoplasmic reticulum

GRP75

Glucose-regulated protein 75

IP3R

Inositol 1,4,5-trisphosphate receptor

MAM

Mitochondria-associated membranes

mPTP

Mitochondrial permeability transition pore

PC

Phosphatidylcholine

PE

Phosphatidylethanolamine

PLA

Proximity ligation assay

SERCA

Sarco/endoplasmic reticulum Ca2+-ATPase

SKM

Skeletal muscle

VDAC

Voltage-dependent anion channel

VL

Vastus lateralis

ZDF

Zücker diabetic fatty

ZL

Zücker lean

Authors’ contributions

Conceptualization and design, A.A.; Formal analysis, A.A.; Funding acquisition, A.A.; Methodology, D.S. and M.V; Data acquisition and analysis, D.S. and M.V.; Data interpretation, D.S., J.R., M.V. and A.A.; Project administration, A.A.; Software, D.S. and J.R.; Supervision, A.A. and J.R.; Writing-original draft, D.S.; Writing – review & editing, A.A. and J.R.

Funding

This work was supported by grant PID2024-157763OB-I00, funded by MICIU/AEI /10.13039/501100011033 andby the ERDF, EU.

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information. The data generated are available from the corresponding author upon request.

Declarations

Ethics approval and consent to participate

The study was performed according to the guidelines of the Declaration of Helsinki and approved by the Ethical Committee of the University of Granada (Granada, Spain) according to the European Union guidelines. The protocol reference number is 23/06/2021/096-CEEA.

Consent for publication

All the authors have read and agreed to the published version of the manuscript. AA is the guarantor of the present study and is responsible for the integrity of the work as a whole.

Competing interests

The authors declare no competing interests.