Genotype-environment interaction drives the onset of riboflavin-responsive multiple acyl-CoA dehydrogenase deficiency in carriers of single heterozygous ETFDH variants Authors: Jing Ma, Huiqiu Zhang, Jiaqi Jiao, Menghan Su, Junsen Zhao, Dan Liu, Qi Wen, Jianli Wang, Juan Wang, Weisong Duan, Xueli Chang, Junhong Guo, Wei Zhang Published: 2026 ABSTRACT AbstractBackgroundRiboflavin-responsive multiple acyl-CoA dehydrogenase deficiency (RR-MADD) is an autosomal recessive disorder of fatty acid oxidation predominantly caused by variants in the ETFDH gene. However, approximately 10% of patients carry a single heterozygous variant. We hypothesize that ETFDH haploinsufficiency may contribute to the development of RR-MADD, especially under certain environmental stressors.MethodsSkin fibroblasts derived from one RR-MADD patient carrying a heterozygous variant in ETFDH and his asymptomatic father carrying the same variant were cultured. Under varying concentrations of riboflavin, ETFDH gene expression profiles, intracellular free fatty acid (FFA) levels, mitochondrial function, cellular viability, and accumulation of reactive oxygen species (ROS) and lipid droplets were assessed and compared between the two cell lines. These phenotypic comparisons were subsequently extended to HEK293 cell models engineered to carry either heterozygous or homozygous c.917G > A variants in ETFDH. Etfdh knock-in mice carrying c.250G > A (p.A84T) variants were used to evaluate the contribution of single heterozygous variants to disease susceptibility at the organismal level.ResultsUnder identical culture conditions, fibroblasts from the RR-MADD patient and his father exhibited comparable levels of ETFDH gene expression and FFA, along with similar mitochondrial function, cellular viability, ROS production, and accumulation of lipid droplets. This finding suggests that environmental factors rather than genotype cause phenotypic differences between the two individuals. Moreover, exposure to severe riboflavin deficiency induced a similar RR-MADD phenotype in HEK293 cells regardless of whether they carried heterozygous or homozygous variants. Mice harboring the heterozygous Etfdh c.250G > A variant exhibited a classic RR-MADD phenotype and pathological changes following combined dietary interventions consisting of riboflavin deficiency and high-fat diet.ConclusionsThis study demonstrates that single heterozygous ETFDH variants confer increased susceptibility to RR-MADD in the presence of specific environmental stressors, offering new genetic insights into the inheritance pattern of the disease.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12964-026-02972-3.Keywords: Multiple acyl-CoA dehydrogenase deficiency, Single heterozygous variant, ETFDH, Environmental stressors, Inheritance pattern Introduction Multiple acyl-CoA dehydrogenase deficiency (MADD), also known as glutaric aciduria type II, is an autosomal recessive disorder of fatty acid oxidation [1, 2]. MADD is clinically heterogeneous and can be classified into neonatal-onset forms with or without congenital anomalies and late-onset form [3, 4]. Late-onset MADD, the most common lipid storage myopathy, is characterized by muscle weakness, exercise intolerance, hepatomegaly, episodes of vomiting, hypoglycemia, and metabolic acidosis [3]. A considerable number of late-onset MADD patients respond to riboflavin therapy, and therefore this clinical phenotype is defined as riboflavin-responsive MADD (RR-MADD) [1, 5]. RR-MADD is predominantly caused by defects in electron transfer flavoprotein dehydrogenase (ETFDH), which is encoded by the ETFDH gene and transfers electrons from acyl-CoA dehydrogenases to the respiratory chain [1, 3, 6, 7]. The mutant ETFDH protein disrupts fatty acid β-oxidation and compromises mitochondrial function, resulting in intracellular lipid accumulation [1]. Causative genes for RR-MADD include, albeit in a minority of cases, electron transfer flavoprotein (ETF) genes (ETFA and ETFB), riboflavin transporter genes (SLC52A1, SLC52A2, SLC52A3, and SLC25A32), flavin adenine dinucleotide (FAD) synthase gene (FLAD1), and coenzyme A synthase gene (COASY) [5, 7, 8]. RR-MADD is inherited in an autosomal recessive manner, and most patients carry biallelic loss-of-function variants in ETFDH [3, 6, 9–11]. However, in about 10% of patients, only a single heterozygous ETFDH variant is identified [12]. Zhang et al. described a father and his son (Patients 13 and 14) suffering from RR-MADD, who carried the same heterozygous c.1285 + 1G > A variant in ETFDH [11]. We hypothesize that single heterozygous variants in ETFDH might be associated with RR-MADD. Nevertheless, first-degree relatives carrying the same single heterozygous variant in ETFDH as the probands are typically unaffected, suggesting that additional modifying factors are necessary for disease onset. Environmental factors, particularly riboflavin deficiency, play a significant role in the pathogenesis of RR-MADD [3, 9, 10, 13–15]. As an active form of riboflavin, FAD promotes the folding and conformational stabilization of ETFDH [16]. Thus, riboflavin deficiency reduces the ETFDH protein levels in fibroblasts from both RR-MADD patients and healthy individuals [13]. A high-fat diet represents another key environmental factor for RR-MADD. Etfdh knock-in (KI) mice carrying homozygous c.250G > A (p.A84T) variants required a combined riboflavin-deficient and high-fat diet to manifest the RR-MADD phenotype [13]. Family members carrying the same heterozygous variant in ETFDH may be exposed to a range of environmental factors varying in type and intensity, and thus phenotypic heterogeneity may arise from differential exposure to environmental factors. This study aims to investigate whether single heterozygous ETFDH variants confer increased susceptibility to RR-MADD under specific environmental stressors, such as riboflavin deficiency and high-fat intake, utilizing muscle tissues and fibroblasts derived from patients, HEK293 cell lines, and Etfdh(h)A84T knock-in model mice. Materials and methods Study participants From our cohort of RR-MADD patients [15], three unrelated patients (P1, P2, and P3) were recruited for the present study. The diagnosis of RR-MADD was based on clinical symptoms, biochemical abnormalities, muscle pathological findings, and response to riboflavin treatment. Whole exome sequencing (WES) revealed that the three patients carried neither biallelic variants in ETFDH nor variants in other genes associated with MADD, including ETFA, ETFB, SLC52A1, SLC52A2, SLC52A3, SLC25A32, FLAD1, and COASY. WES detected the heterozygous ETFDH c.1205C > T variant in P1 and the heterozygous ETFDH c.917G > A variant in P3, whereas no candidate MADD-associated variants were detected in P2. This study has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments and approved by the institutional ethics committee of the First Hospital of Shanxi Medical University (K073). Written informed consent for participation was obtained from all participants. Comprehensive genetic analyses Long-read sequencing (LRS) by Oxford Nanopore Technology Genomic DNA was extracted from peripheral blood. Whole-genome LRS was performed to detect potential structural variants (SVs) and intronic variations using a PromethION sequencer (Oxford Nanopore Technologies, UK). The LRS achieved an average sequencing read depth of 28.59 ×, with 99.91% coverage of the target regions. The DNA quantity and quality were assessed using a Qubit 3.0 Fluorometer (Invitrogen, USA) and electrophoresis, respectively. Library preparation was carried out using a 1D Genomic DNA ligation kit (SQK-LSK110) according to the manufacturer’s protocol. For each patient, one PRO-002 (R9.4.1) flow cell was used. PromethION data base-calling was performed using guppy v.3.3.0, and only pass reads (Qscore ≥ 7) were used for subsequent analysis. Reads were aligned to the human reference genome (GRCh37/hg19) using minimap2. SVs and short tandem repeats (STRs) were detected using Sniffles and RepeatHMM, respectively. The single-nucleotide polymorphisms (SNPs) and small insertions/deletions (indels) were analyzed using the Genome Analysis ToolKit (GATK) HaplotypeCaller. Variants with low quality (depth < 10 or genotype quality < 20) were filtered out. Variants with minor allele frequency (MAF) > 0.5% in the Genome Aggregation Database (gnomAD) were also excluded. The numbering of variants in ETFDH was based on the reference sequence NM_004453. SpliceAI software was used to predict splicing outcomes (https://spliceailookup.broadinstitute.org/). The candidate causal variants discovered by LRS were then confirmed by Sanger sequencing. Sequencing of cDNA and TA cloning cDNA sequencing was performed to detect potential abnormal splicing in ETFDH. The PAXgene Blood RNA System (QIAGEN, Germany) was used to isolate full-length mRNA from white blood cells. Subsequently, cDNA was generated using the HiScript III 1 st Strand cDNA Synthesis Kit (R312-01, Vazyme, China) according to the manufacturer’s protocol. cDNA was amplified with specific primers (Supplementary Table 1), and PCR products were purified by 2% TAE agarose gel electrophoresis and sequenced with an automated sequencer (Applied Biosystems 3730xl DNA Analyzer, Thermo Fisher Scientific, USA). The targeted DNA products containing abnormal splicing were then inserted into the pEASY®-T1 Cloning Vector (TransGen Biotech, China) according to the manufacturer’s protocol. After ligation, the products (5 μL) were transformed into E. coli Trans5α competent cells (50 μL, Tsingke Biotechnology, China) and incubated for 30 min on ice, 1 min at 42 °C, then 2 min on ice. Luria-Bertani (LB) broth was added to each tube (500 μL) and the cells were cultured at 37 °C, 150 rpm for 60 min. Then, 100 μL of the transformed cells were spread evenly on LB-agar plates containing ampicillin and cultured at 37 °C for 12 h. Cells transformed with vectors containing recombinant DNA produced white colonies, which were selected for further culturing at 37 °C, 250 rpm for 6 h. The plasmid DNA was then extracted and Sanger sequencing was performed to detect variants in introns corresponding to abnormal splicing. Validation of the identified variants was performed in family members. Those who carried the same single heterozygous variant in the ETFDH gene as the proband, but without any related clinical manifestations upon clinical examination were defined as carriers. Assessment of muscle pathology in patients with single heterozygous variants in ETFDH Muscle biopsy was performed in all patients before riboflavin treatment. For histopathological analysis, serial frozen sections (12 μm) were stained with hematoxylin and eosin (HE) and Oil Red O (ORO), and then observed under a microscope (BX53, Olympus, Japan). Cultivation of human primary skin fibroblasts and construction of HEK293 cell lines with c.917G > A variants in ETFDH Cultivation and identification of human primary skin fibroblasts Human primary skin fibroblasts were obtained from P2, his asymptomatic father carrying the same variant in ETFDH (C2), and a healthy control (HC). C2 was clinically evaluated by an experienced neurologist and had no symptoms or signs of MADD. His serum riboflavin level was within the normal range (2.45 ng/mL, reference range 1–19 ng/mL). Based on these findings, he was classified as an asymptomatic carrier. Skin tissue samples were dissected into 0.5 mm sized pieces and placed into a 24-well tissue culture plate containing Dulbecco’s modified Eagle medium (DMEM) (11995500, Thermo Fisher Scientific, USA) supplemented with 20% fetal bovine serum (FBS) (FSP500, ExCell Bio, China), 1% antibiotic–antimycotic solution (15240112, Thermo Fisher Scientific, USA), and 1% GlutaMAX (35050061, Thermo Fisher Scientific, USA). The cultures were maintained for 2 weeks at 37 °C in a 5% CO2 incubator. The primary culture at 70% confluency was digested with 0.25% trypsin (C25200072, Thermo Fisher Scientific, USA) and passaged for the following experiments. All the experiments were carried out using fibroblasts within 15 passages. For identification of fibroblasts, the cells were fixed with 4% paraformaldehyde (PFA) for 10 min at room temperature (RT). After being blocked with blocking buffer (PBS containing 0.1% Triton X-100 and 10% FBS) for 1 h at RT, the fibroblasts were then incubated with anti-vimentin antibody (1:50, sc-32322, Santa Cruz, USA) at 4 °C overnight. The fibroblasts were then incubated with Alexa Fluor 488-conjugated anti-mouse IgG secondary antibody (1:5000, AB_2534069, Thermo Fisher Scientific, USA) for 1 h at RT. The nuclei were stained with DAPI (AR1176, BOSTER, China) for 1 min at RT in the dark and the stained fibroblasts were observed under a fluorescence microscope (BX53, Olympus, Japan). Construction of HEK293 cell lines carrying c.917G > A variants in ETFDH HEK293 cell lines are frequently employed in investigating the pathogenesis of MADD [16–18]. P3 carried a heterozygous variant c.917G > A in ETFDH. Hence, HEK293 cell lines carrying heterozygous or homozygous c.917G > A variants were generated by GENECARER Bio (China) using a prime editing system. Prime editors contain a nCas9 (H840A) tethered to an engineered reverse transcriptase programmed by prime editing gRNAs (pegRNAs). The pegRNA was designed to mediate site-specific nicking by nCas9, and then served as a template for reverse transcriptase to install desired variants [19]. The pegRNA was cloned into the puromycin-resistant backbone plasmid using restriction enzyme digestion (BsmBI, EG22507V; XhoI, EG15583S; both from Yugong Biotech, China) and ligation (T4 DNA Ligase, 2011 A, Takara, Japan). The resulting construct was named pCDNA3.1-U6-ETFDH-twinpegRNA-SFFV-COGFP-puro (GC011101308, GENECARER Bio, China) and verified by sequencing. HEK293 cells were seeded in 24-well plates until 80% confluency, then a mixture of plasmid DNA (0.8 µg), Lipofectamine 3000 Transfection Reagent (1 µL) (L3000015, Thermo Fisher Scientific, USA), and 100 µL of serum-free DMEM was incubated at RT for 20 min and applied to the cells. After 48 h of transfection, the medium was replaced with culture medium containing 2 μg/mL puromycin to select for puromycin-resistant colonies. Then the monoclonal cell lines were isolated and propagated. Genotype identification of HEK293 cells was performed to select clones carrying heterozygous or homozygous c.917G > A variants in ETFDH, and the primers are shown in Supplementary Table 2. All the experiments were carried out using HEK293 cells within 25 passages. Experimental interventions The cells were cultured until 50% confluency, and then the medium was replaced with fresh riboflavin-depleted DMEM (DZPYG0321, BOSTER, China) containing 10% Knockout™ serum supplement (10828028, Thermo Fisher Scientific, USA), with riboflavin (102529139, Sigma, USA) added to a final concentration of 530 nmol/L (fully supplemented medium), 50 nmol/L (supplemented medium) or 12.6 nmol/L (severely deficient medium) [13, 16]. The concentration of 530 nmol/L is the standard riboflavin concentration in normal DMEM, which is sufficient to support normal cell growth and serves as a control condition. The concentrations of 50 nmol/L and 12.6 nmol/L were selected based on previously published studies [13], approximating the serum riboflavin concentrations observed in healthy controls and RR-MADD patients before receiving riboflavin treatment, respectively [14]. Then the cells were cultured for an additional 6 days for the following experiments. To further elucidate the pathogenic mechanism of single heterozygous ETFDH variations, the rescue experiments involving ETFDH gene overexpression or riboflavin repletion were performed on HEK293 cells carrying heterozygous ETFDH variants. The ETFDH overexpression plasmid was designed and synthesized by Public Protein/Plasmid Library (China), using pcDNA3.1-T2A-EGFP as the vector. After being cultured in severely riboflavin-deficient medium for 3 days, HEK293 cells were either transfected with the ETFDH overexpression plasmid using jetPRIME transfection reagent (101000046, jetPRIME, France) or treated with riboflavin at a final concentration of 530 nmol/L. Cell samples were collected 72 h after treatment for subsequent experiments. Etfdh(h)A84T KI mice and dietary interventions Mice harboring homozygous Etfdh c.250G > A (p.A84T) variants were kindly gifted by Professor Yan Chuanzhu (Department of Neurology, Qilu Hospital, Shandong University, China), and detailed information about construction of the KI mouse model was provided in a previous publication [13]. The homozygous mice were crossed with wild-type (WT) C57BL/6 mice, and the resulting heterozygotes were intercrossed to generate WT (WT group), heterozygous (Het group), and homozygous (Hom group) offspring for experiments. The primers for mouse genotype identification are shown in Supplementary Table 2. All procedures were approved and strictly carried out under the guidelines of the Shanxi Provincial Animal Ethics Association. Starting at 16 weeks of age, the mice were assigned to the control diet or vitamin B2-deficient (B2D) diet for 12 weeks. Beginning at 24 weeks of age, a subset of mice in the B2D group received a 4-week oral gavage of high-fat (HF) emulsion at a dose of 10 mL/kg once daily, forming the B2D + HF diet group. The remaining mice fed with the B2D diet and control diet received daily gavage of an equal volume of physiological saline for 4 weeks, forming the B2D diet group and Control diet group, respectively. Ultimately, with dietary interventions ending at 28 weeks of age, nine groups (n = 8 mice per group) were formed, encompassing three genotypes under three different dietary conditions (Supplementary Fig. 2a). The special pellet feed and HF emulsion were provided by Huafukang Biological Technologies (Beijing, China) and Xiao Shu You Tai Biotechnology (Beijing, China), respectively. The ingredients are shown in Supplementary Tables 3–5. Comprehensive analyses of functional and phenotypic alterations in human muscle tissues, cells, and mice Reverse transcription and real-time quantitative polymerase chain reaction (RT-qPCR) for ETFDH expression in human muscle tissues and fibroblasts Total RNA was extracted from human muscle tissues using TRIzol reagent (15596018CN, Thermo Fisher Scientific, USA) and from fibroblasts with a FastPure Cell/Tissue Total RNA Isolation Kit (RC112-01, Vazyme, China). Reverse transcription was performed using HiScript III-RT SuperMix Kit (R323, Vazyme, China), and RT-qPCR was performed using ChamQ Universal SYBR qPCR Master Mix (Q711, Vazyme, China) in a LightCycler 480 II real-time PCR system (Roche, Switzerland). The threshold cycle (Ct) values for ACTB and ETFDH were measured in human muscle tissues and fibroblasts. The RT-qPCR primers are listed in Supplementary Table 2. Expression of ETFDH protein in muscle tissues and cells Muscle tissues from a healthy control, patients carrying single heterozygous variants (n = 3) and those with biallelic variants in ETFDH (n = 4) (detailed clinical and genetic characteristics are shown in Supplementary Table 6), fibroblasts, HEK293 cells, and mouse muscle tissues after dietary interventions were lysed with sodium dodecyl sulfate (SDS) lysis buffer (P0013G, Beyotime, China). Protein concentrations were quantified with a BCA assay kit (AR1189, BOSTER, China) and the absorbance at 562 nm was measured using a multifunctional microplate reader (BioTek Synergy H1, USA). Protein samples (23 μg of protein for human muscle tissues, 10 μg of protein for cells, and 8 μg of protein for mouse muscle tissues) were separated in 10% SDS-polyacrylamide gels (PG112, Yamei, China) and then transferred onto polyvinylidene difluoride (PVDF) membrane (IPVH00010, Millipore, USA). After being blocked, the PVDF membrane was incubated with anti-ETFDH antibodies (1:1000, ab126576, Abcam, UK; 1:3000, 11109-1-AP, Proteintech, China) and anti-GAPDH antibody (1:8000, A19056, ABclonal, China), followed by peroxidase-conjugated goat anti-mouse or goat anti-rabbit IgG (1:10,000, BA1050/BA1054, BOSTER, China). Protein bands were visualized with enhanced chemiluminescence (ECL) detection reagents (AR1171, BOSTER, China), and images were quantified using ImageJ (version 1.52, Bethesda, MD, USA). Relative expression levels were calculated as ratios normalized to GAPDH. Intracellular free fatty acid (FFA) assay After cells were cultured for 6 days with different concentrations of riboflavin, the medium was removed. Intracellular FFA levels were determined using an Amplex Red Free Fatty Acid Assay Kit (S0215M, Beyotime, China). Fluorescence intensity in the supernatant was detected with a multifunctional microplate reader, and FFA levels were quantified accordingly. Results were presented as μM per 1 × 105 cells. Assessment of mitochondrial oxygen consumption rate (OCR) Mitochondrial oxygen consumption rate of cells was measured using a XF Cell Mito Stress Test Kit (103015-100, Seahorse Bioscience, USA) on a Seahorse XFe24 extracellular flux analyzer (Seahorse Bioscience, USA). The optimal seeding density was determined to be 3 × 104 cells/well for fibroblasts and 5 × 104 cells/well for HEK293 cells. The optimal concentrations of the injection compounds for both fibroblasts and HEK293 cells were determined to be 1.5 μM oligomycin, 1.0 μM carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP), and 0.5 μM rotenone & antimycin A. After being cultured for 6 days with different concentrations of riboflavin, cells were digested, seeded in XFe24 cell culture plates coated with poly‐D‐lysine (PDL) (ST508, Beyotime, China) in 100 μL pre-warmed assay medium (XF base RPMI medium supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose), and then incubated at 37 °C in a CO2-free incubator for 30 min to facilitate cell adhesion. After the cells had adhered, 400 μL of assay medium was added and the cells were incubated in a CO₂‑free incubator at 37 °C for an additional 15 min. Prewarmed oligomycin, FCCP, and rotenone & antimycin A were loaded into the injector ports A, B, and C of the sensor cartridge, respectively. The cartridge was calibrated by the XFe24 analyzer and the assay was carried out using the Mito Stress Test protocol. Each condition involved three cycles of mixing (180 s), waiting (120 s), and measuring (180 s). This series was repeated following each injection. The basal OCR and maximal OCR were calculated from the bioenergetics profile. After detection, cell numbers were determined using Hoechst 33342 staining (C1022, Beyotime, China) and a high-content microscopy imaging system (ImageXpress Micro Confocal, Molecular Devices, USA). The OCR values were normalized accordingly. Adenosine triphosphate (ATP) assay in cell and muscle tissues ATP levels of fibroblasts, HEK293 cells, and mouse muscle tissues were determined using an ATP Assay Kit (S0027, Beyotime, China). Cells (5 × 105 cells for each group) were lysed and centrifuged to collect the supernatant for ATP measurement. For mouse muscle tissues, 100 μL of lysis buffer was added per 20 mg of tissue. A portion of the supernatant was boiled for 2 min to ensure complete release of ATP, while the remaining supernatant was reserved for subsequent protein quantification. The supernatant was detected by enhanced chemiluminescence using a multifunctional microplate reader, and total ATP levels were calculated accordingly. ATP levels were presented as μM per 5 × 105 cells or μmol per mg protein. Cellular viability assay Cellular viability was assessed with a CCK-8 kit (K1018, APExBIO, USA). Briefly, after being cultured for 6 days under different concentrations of riboflavin, cells were seeded at a density of 5 × 104 cells in 100 μL medium per well in a 96-well plate. Then, 10 μL of CCK-8 reagent was added to the samples, followed by incubation at 37 °C for 2 h. Absorbance at 450 nm was measured by a multifunctional microplate reader. Intracellular reactive oxygen species (ROS) measurement The detection of intracellular ROS levels was performed using the ROS-sensitive probe H2DCFDA (HY-D0940, MCE, USA). Cells were incubated with 10 µM H2DCFDA staining solution in serum-free DMEM in the dark at 37 °C for 30 min, then harvested with 0.25% trypsin and resuspended in fresh riboflavin-depleted DMEM. Lastly, a flow cytometer (BD FACSCanto™ II, BD Biosciences, USA) was used to estimate the cellular fluorescence intensity at an excitation wavelength of 488 nm and an emission wavelength of 530 nm. FlowJo software (version 10.8, FlowJo, Ashland, OR, USA) was used to analyze the data. Cellular ORO and BODIPY staining Intracellular lipid droplets were determined by ORO and BODIPY staining. After fixation with 4% PFA at RT for 10 min, the cells were stained with the ORO working solution at RT for 20 min (C0158S, Beyotime, China), and then washed with the washing solution for 30 s, followed by counterstaining with hematoxylin staining solution (C0107, Beyotime, China) for 1 min. For BODIPY staining, the cells were fixed as described above, and then stained with BODIPY™ 493/503 (D3922, Invitrogen, USA) in the dark at 37 °C for 30 min, followed by counterstaining with DAPI (AR1176, BOSTER, China) for 1 min. The stained cells were observed under a fluorescence microscope. Assessment of mouse body weight and motor function The body weight of mice was measured weekly starting from 8 weeks of age until the end of dietary interventions. Exercise endurance and muscle strength were assessed using the treadmill test, rotarod test, and grip strength test before (at 16 weeks of age) and after (at 28 weeks of age) dietary interventions. The treadmill was started at 4 m/min, with the speed increased incrementally by 2 m/min every 2 min until the mice reached exhaustion. A consistent stimulation current of 0.2 mA was applied to motivate the mice to run. Exhaustion was defined as remaining on the electric stimulus platform for 10 s without attempting to reengage the treadmill belt as reported previously [20]. The total running distance, running time, and maximal running speed were recorded. Following three days of brief training, the rotarod was initiated at 4 rpm and gradually accelerated to 40 rpm. The latency to fall from the rod was recorded up to a maximum of 300 s. Each mouse was tested five times with at least 5-min breaks between trials [21]. Depressive-like and anxiety-related behaviors were assessed by the tail suspension test and open field test, respectively. Assessment of serum biochemical changes in mice At the end of dietary interventions, blood samples were collected via eyeball extirpation after mice were euthanized by isoflurane inhalation. The serum FFA and acylcarnitine levels were measured by ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) (Metabo‐Profile Biotechnology, China). The serum creatine kinase (CK) enzyme activity was measured using an Amplex Red Creatine Kinase Assay Kit (S0287S, Beyotime, China) and quantified with a multifunctional microplate reader. Histopathology of mouse skeletal muscle After blood collection, gastrocnemius and quadriceps femoris muscles were harvested for further analyses. The skeletal muscles were frozen in isopentane, which was precooled in liquid nitrogen. The gastrocnemius muscles of all groups were sectioned at 10 μm thickness, and then stained with HE, ORO, and succinate dehydrogenase (SDH). Myofiber cross-sectional area (CSA) was measured using ImageJ, and the coefficient of variation (CV) of CSA was calculated as follows: CV = standard deviation of CSA/mean of CSA. Metabolic characterization of mouse skeletal muscle FAD levels in quadriceps femoris were measured by UHPLC-MS/MS (Metabo-Profile Biotechnology, China), and triglyceride (TG) levels in quadriceps femoris were measured using a tissue/cell triglyceride kit (E1013, Pulilai, China) to assess the impact of the dietary interventions on skeletal muscle. TG levels were presented as μmol per mg protein. Statistical analysis Continuous data were presented as mean values ± standard error of the mean (SEM) and categorical variables were presented as percentages. Data analysis and graphing were performed using SPSS (version 24.0, IBM Corp., USA) and GraphPad Prism (version 9.0, GraphPad Software, USA). One-way analysis of variance (ANOVA) with Tukey post hoc test was performed for multiple group comparisons involving one independent variable, and two-way ANOVA with Tukey post hoc test for those involving two or three independent variables. A P value < 0.05 was considered statistically significant. * indicates P < 0.05, ** indicates P < 0.01, *** indicates P < 0.001, and ns indicates no significant difference. For cell-based experiments, the number of independent experiments is provided in the figure legends. Experiments involving animal tissues were performed with at least three independent experimental replicates per assay condition, and the number of mice per group is indicated in the figure legends. Results Clinical manifestations All three patients included in this study presented clinical symptoms consistent with RR-MADD, including muscle weakness, exercise intolerance, and myalgia. Muscle biopsy revealed pathological findings consistent with lipid storage myopathy (Fig. 1a–1f). Detailed clinical characteristics and biochemical profiles of these cases have been previously reported in our publication (No.7, No.12, and No.13) [15]. None of the family members of the three patients were affected by RR-MADD. Hematoxylin and eosin (HE) (a, c, e) and Oil Red O (ORO) (b, d, f) staining of human muscle tissues revealed pathological findings consistent with lipid storage myopathy in P1 (a, b), P2 (c, d), and P3 (e, f). Comparison of ETFDH protein (g) and ETFDH mRNA (h) in muscle tissues among a healthy control, patients carrying single heterozygous variants, and those with biallelic variants. The error bars represent standard error of the mean. * indicates P < 0.05, ** indicates P < 0.01, *** indicates P < 0.001, and ns indicates no significant difference (n = 3) Molecular diagnosis LRS by Oxford Nanopore Technologies revealed abnormal splicing in P2 LRS was performed on the three patients. An average of 100.3 Gb WGS ONT long reads per sample with an average N50 read length of 36.10 kb were generated. After mapping all cleaned reads to the human reference genome GRCh37/hg19, we obtained an average depth of 28.59 ×. In P1 and P3, no additional variants were identified in the candidate genes other than the heterozygous c.1205C > T variant and c.917G > A variant in ETFDH, respectively, as revealed by WES (Fig. 2a, 2b). Following the criteria proposed by the American College of Medical Genetics and Genomics (ACMG) [22], both variants were classified as likely pathogenic: c.1205C > T (p.T402I) in P1 (PM2_Supporting + PM3_Strong + PP3_Moderate + PP4) and c.917G > A (p.G306E) in P3 (PM2_Supporting + PM3_Strong + PP3_Strong + PP4). Whole-exome sequencing (WES) revealed the heterozygous c.1205C > T variant in ETFDH in P1 (a) and c.917G > A variant in P3 (b). Long-read sequencing (LRS) identified a heterozygous variant (c.488-205A > T) in intron 4 of ETFDH in P2 (c, d). cDNA sequence analysis identified the abnormal splicing in P2 between exon 3 and exon 7 of ETFDH (e). TA cloning and sequencing confirmed that the c.488-205A > T variant in P2 resulted in a 133 bp heterozygous insertion between exon 4 and exon 5 (f). Pedigree diagrams show the segregation of these variants among family members (g). The proband is indicated by an arrow. The squares represent males and the circles represent females. Asymptomatic carriers are marked by half-filled symbols, while unaffected and affected individuals are marked by unfilled and filled symbols, respectively In P2, WES did not identify any candidate variants. However, LRS detected a heterozygous variant in intron 4 of ETFDH (chr4:159,606,048, c.488-205A > T) (Fig. 2c, 2 d), which was absent in gnomAD. SpliceAI analysis of this variant predicted a high probability of affecting splicing with a ΔScore of 0.86. Sequencing of cDNA and TA cloning confirmed the variant in P2 Accordingly, no abnormal splicing was detected in ETFDH cDNA from P1 and P3. In P2, cDNA sequence analysis identified abnormal splicing between exon 3 and exon 7 of ETFDH (Fig. 2e). TA cloning sequencing confirmed that the variant c.488-205A > T at chr4:159606048 in ETFDH resulted in a 133 bp heterozygous insertion between exon 4 and exon 5, which originated from intron 4 (chr4:159605914_159606046) (Fig. 2f). The insertion further resulted in the occurrence of a premature termination codon. Based on the ACMG criteria, c.488-205A > T was classified as likely pathogenic (PS3 + PM2_Supporting + PP3 + PP4). As no SVs and additional pathogenic variants were identified in the three patients, these findings indicated that each individual carried only a single heterozygous variant in ETFDH. The segregation of these variants among family members is illustrated in Fig. 2g. RR-MADD patients with single heterozygous ETFDH variants exhibited higher ETFDH protein expression in muscle tissues than those with biallelic variants The majority of RR-MADD patients carry missense variants, while a small proportion have null variants [12], both of which lead to reduced ETFDH protein levels [23–25]. Consequently, individuals carrying a single heterozygous variant would be expected to have higher residual levels of ETFDH protein than those carrying biallelic variants. Our analysis of muscle tissues confirmed that patients with single heterozygous variants in ETFDH demonstrated higher ETFDH protein levels than those with biallelic variants (P < 0.01), although both groups showed lower levels compared with the healthy control (P < 0.001) (Fig. 1g). Moreover, ETFDH mRNA levels in muscle specimens showed no significant differences among the three groups (P > 0.05) (Fig. 1h). The above results suggest that the patients carrying a single heterozygous ETFDH variant are unlikely to harbor an additional, undetected variant. Fibroblasts derived from P2 and C2 maintained comparable phenotypes under both supplementary and deficient riboflavin conditions The morphology and vimentin expression in fibroblasts are shown in Supplementary Fig. 1a. Comprehensive genetic sequencing identified a single heterozygous variant c.488-205A > T in P2. However, his father (C2), who carried the same variant, was not affected by RR-MADD. This phenotypic discrepancy may be attributed to differences in the severity of riboflavin deficiency or the presence of an additional, undetected ETFDH variant in P2. Therefore, to assess the contributory role of single heterozygous variants in RR-MADD susceptibility, we examined whether fibroblasts derived from P2 and C2 exhibited similar phenotypes when cultured at matching riboflavin concentrations. ETFDH mRNA expression levels were comparable between fibroblasts derived from P2 and C2, and both were approximately half the level observed in fibroblasts from HC (Fig. 3a), which indicated that the c.488-205A > T variant induced nonsense-mediated mRNA decay. Moreover, regardless of the riboflavin concentration, no significant differences were observed between the two cell lines across multiple key metrics, including ETFDH protein abundance (Fig. 3b), FFA content (Fig. 3c), mitochondrial function (basal OCR, maximal OCR, and ATP production) (Fig. 3d–3f), cellular viability (Fig. 3g), ROS accumulation (Fig. 3h), and lipid droplet deposition (Fig. 4). The results imply that P2 does not harbor a second pathogenic variant and single heterozygous variants in ETFDH may predispose individuals to RR-MADD. Comparison of ETFDH mRNA (a), ETFDH protein (b), free fatty acid (FFA) content (c), basal and maximal oxygen consumption rate (OCR) (e), adenosine triphosphate (ATP) production (f), cellular viability (g), and ROS accumulation (h) in fibroblasts from P2, his asymptomatic father (C2), and a healthy control (HC) under different riboflavin conditions. A representative graph displays the OCR profile of fibroblasts following sequential addition of 1.5 μM oligomycin, 1.0 μM carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP), and 0.5 μM rotenone & antimycin A (d) (n = 5) Lipid droplet deposition in fibroblasts visualized by ORO (a) and BODIPY (c) staining. Comparison of the percentage of ORO-positive cells (b) and BODIPY fluorescence intensity (d) in fibroblasts under different riboflavin conditions (n = 5) Fibroblasts derived from P2 and C2 developed the phenotype of RR-MADD under riboflavin-deficient conditions Fibroblasts from P2, C2, and HC displayed no significant differences in ETFDH protein expression (Fig. 3b), FFA content (Fig. 3c), mitochondrial function (Fig. 3d–3f), cellular viability (Fig. 3g), ROS accumulation (Fig. 3h) or lipid droplet deposition (Fig. 4) under riboflavin-sufficient conditions. However, when cultured in riboflavin-deficient medium, fibroblasts derived from P2 and C2 exhibited the phenotypic features of RR-MADD compared with fibroblasts from HC—including reduced ETFDH protein expression, elevated FFA levels, mitochondrial dysfunction, decreased cellular viability, increased ROS production, and increased lipid droplet deposition. The data indicate that under riboflavin-deficient conditions, single heterozygous variants in ETFDH might be related to RR-MADD. HEK293 cell lines with heterozygous variants in ETFDH exhibited a similar phenotype of RR-MADD caused by homozygous variants under riboflavin-deficient conditions After identification of the ETFDH c.917G > A variants in HEK293 cells (Supplementary Fig. 1b), parallel analyses in HEK293 cell lines with heterozygous and homozygous ETFDH variants confirmed the findings in fibroblasts. With decreasing concentrations of riboflavin in the culture medium, HEK293 cells in both Het and Hom groups progressively exhibited RR-MADD-like phenotypic features. These features included reduced ETFDH protein expression (Fig. 5a), elevated FFA levels (Fig. 5b), mitochondrial dysfunction (Fig. 5c–5e), decreased cellular viability (Fig. 5f), increased ROS production (Fig. 5g), and increased lipid droplet deposition (Fig. 6), all of which differed significantly from those in the wild-type group. Under conditions of severe riboflavin deficiency, the changes observed in HEK293 cells with heterozygous and homozygous variants were comparable, demonstrating that a heterozygous variant in ETFDH confers increased susceptibility to RR-MADD in the context of riboflavin deficiency. Comparison of ETFDH protein (a), FFA content (b), basal and maximal OCR (d), ATP production (e), cellular viability (f), and ROS accumulation (g) in HEK293 cells under different riboflavin conditions. A representative graph displays the OCR profile of HEK293 cells following sequential addition of 1.5 μM oligomycin, 1.0 μM FCCP, and 0.5 μM rotenone & antimycin A (c) (n = 5) Lipid droplet deposition in HEK293 cells visualized by ORO (a) and BODIPY (c) staining. Comparison of the percentage of ORO-positive cells (b) and BODIPY fluorescence intensity (d) in HEK293 cells under different riboflavin conditions (n = 5) Both overexpression of the ETFDH gene and riboflavin repletion significantly increased ETFDH protein levels (P < 0.01), alleviated intracellular FFA accumulation (P < 0.01), increased ATP content (P < 0.001), improved cellular viability (P < 0.01), and decreased lipid droplet accumulation (P < 0.001) in heterozygous HEK293 cells following 3 days of intervention under severely riboflavin-deficient conditions (Fig. 7). These findings indicate that single heterozygous ETFDH variants confer increased susceptibility to RR-MADD through ETFDH haploinsufficiency under riboflavin-deficient conditions, rather than nonspecific effects of riboflavin deprivation. Comparison of ETFDH mRNA (a), ETFDH protein (b), FFA content (c), ATP production (d), cellular viability (e), the percentage of ORO-positive cells (g), and BODIPY fluorescence intensity (i) in HEK293 cells carrying the heterozygous ETFDH variant following ETFDH overexpression and riboflavin repletion. Lipid droplet deposition in HEK293 cells visualized by ORO (f) and BODIPY (h) staining (n = 3) Mice carrying either heterozygous or homozygous Etfdh variants exhibited RR-MADD-like characteristics after B2D + HF dietary interventions After the 12-week dietary interventions, two-way ANOVA revealed significant main effects of genotype (P < 0.001) and dietary interventions (P < 0.001) on the body weight of mice, with no significant interaction effect between the two factors (P = 0.210). Both B2D and B2D + HF diets induced significant weight loss across genotypes (P < 0.05), and body weight was significantly lower in the Het and Hom groups than in the WT group (P < 0.05) (Fig. 8a, 8b, and Supplementary Fig. 2b, 2c). As assessed by the treadmill and rotarod tests, the Het and Hom mice in the B2D + HF diet group exhibited reduced exercise endurance compared with WT mice, characterized by shorter running distance (P < 0.05) (Fig. 8c) and decreased latency to fall (P < 0.05) (Fig. 8d). Further comparisons regarding exercise endurance, muscle strength, and motion-related behaviors are detailed in Supplementary Fig. 2d–2i. Mice carrying either heterozygous or homozygous Etfdh variants exhibited characteristics of riboflavin-responsive multiple acyl-CoA dehydrogenase deficiency after B2D + HF dietary interventions. Representative photographs of mice in different groups after dietary interventions (a). Comparison of body weight changes (b) (n = 7 mice per group), total running distance in the treadmill test (c) (n = 6 mice per group), and latency to fall in the rotarod test (d) (n = 6 mice per group) across different groups before and after dietary interventions The serum FFA and acylcarnitine pattern analyses revealed impaired fatty acid β-oxidation in the Het and Hom mice following B2D + HF dietary interventions, with elevated amounts of medium- and long-chain FFAs (C6, C8, C12, C13, C16, and C20) and acylcarnitines (C6, C10, C14, C16, and C18) when compared with the WT group (P < 0.05) (Fig. 9a, 9b). Consistently, serum CK enzyme activity was significantly higher in the Het and Hom groups than that in the WT group under B2D + HF dietary conditions (P < 0.05), indicating muscle damage (Fig. 9c). Comparison of serum FFA (a), acylcarnitine (b), creatine kinase (CK) enzyme activity (c), and the coefficient of variation (CV) of myofiber cross-sectional area (CSA) (e) in mice across different groups following dietary interventions. Representative images of skeletal muscle pathology assessed by HE, ORO, and succinate dehydrogenase (SDH) staining (d) (n = 3 mice per group) Histological analyses revealed that skeletal muscle morphology remained normal across all genotypes following Control and B2D dietary interventions. However, in response to B2D + HF interventions, the Het and Hom groups exhibited marked intramyofibrillar lipid droplet deposition, alongside increased SDH enzyme activity (Fig. 9d) and an increased CV of myofiber CSA (Fig. 9e) compared with the WT group (P < 0.05). Taken together, mice harboring heterozygous and homozygous Etfdh variants develop key clinical characteristics of RR-MADD patients following B2D + HF dietary interventions, with most phenotypic parameters being comparable between the two groups. Molecular and metabolic alterations in skeletal muscle of mice with Etfdh variants after dietary interventions Comparison of FAD content in skeletal muscle between Het mice in the Control diet and B2D diet groups showed a trend toward lower levels in the B2D diet group, although the difference did not reach statistical significance (P = 0.081) (Fig. 10a). The B2D + HF dietary interventions resulted in significant accumulation of TG in skeletal muscle across all genotypes (P < 0.05), with markedly higher levels of TG in the Het and Hom groups than those in the WT group (P < 0.001) (Fig. 10b). Comparison of flavin adenine dinucleotide (FAD) (a), triglyceride (TG) (b), ETFDH protein (c), and ATP production (d) in skeletal muscle from mice across different groups following dietary interventions (n = 3 mice per group) Following B2D + HF dietary interventions, ETFDH protein expression levels (Fig. 10c) and ATP production (Fig. 10d) were significantly reduced in the Het and Hom groups relative to the WT group (P < 0.01), consistent with the findings observed in fibroblasts and HEK293 cells. Discussion Approximately 10% of patients with RR-MADD carry single heterozygous variants in ETFDH, which are often identified through WES, and some researchers have suspected that these patients may harbor another variant, which is missed by available sequencing technologies [3, 6, 9, 10]. The influence of environmental factors further complicates the clinical picture of RR-MADD [3, 9, 10, 13]. Therefore, in this study, we investigated the response to different riboflavin concentrations in the culture medium (as an environmental factor) using fibroblasts derived from P2 carrying a single heterozygous ETFDH variant and his asymptomatic father C2 carrying the same variant, along with HEK293 cell lines carrying heterozygous or homozygous ETFDH variants. Additionally, we explored the phenotypic changes in mice carrying Etfdh variants under dietary interventions. The results showed that fibroblasts from P2 and C2 exhibited similar phenotypes across different riboflavin concentrations, ruling out the possibility of an additional variant in P2. Under riboflavin-dicient conditions, both P2 and C2 fibroblasts demonstrated an RR-MADD-like phenotype. Similarly, riboflavin deficiency induced a comparable RR-MADD phenotype in HEK293 cells with heterozygous and biallelic ETFDH variants. Consistent with our in vitro findings, mice carrying either heterozygous or homozygous Etfdh variants developed an RR-MADD-like phenotype and pathology under the combined riboflavin-deficient and high-fat diet interventions, whereas wild-type mice remained unaffected. These results provide compelling evidence that a single heterozygous ETFDH variant predisposes carriers to RR-MADD in conjunction with certain environmental factors. Environmental factors significantly contribute to the pathogenesis of RR-MADD [3, 9, 10, 13]. Serum riboflavin deficiency is present in almost all patients and intermittent administration of riboflavin is sufficient to maintain an asymptomatic state [14, 15]. Riboflavin is taken up and transported by cells via specific transporters to target tissues, subsequently converted to flavin mononucleotide (FMN) and FAD through riboflavin kinase and FAD synthase in cells [26]. In RR-MADD patients, decreased muscular FAD levels have been documented [15]. The present study demonstrated that riboflavin deficiency directly led to reduced skeletal muscle FAD concentrations, which was consistent with a previous study [13]. At the cellular level, under conditions of severe riboflavin deficiency, reduced ETFDH protein expression has been observed in cultured fibroblasts, particularly in cells carrying ETFDH variants [13]. Other environmental factors that increase the mitochondrial fatty acid β-oxidation flux, including prolonged exercise, cold exposure, pregnancy, and the use of weight-loss drugs, also influence the onset and severity of the disease [27–33]. As previously demonstrated, the homozygous Etfdh knock-in mouse model successfully recapitulated the RR-MADD phenotype only under combined interventions consisting of riboflavin deficiency and high-fat diet [13]. Collectively, these observations underscore the important role of environmental factors in this disease. It is therefore not surprising that a single heterozygous variant in ETFDH, under defined environmental stressors, could lower the threshold for developing RR-MADD. Heterogeneity in environmental exposures could explain why the family members of these three patients, who carry the same variant, remain unaffected by RR-MADD. In our previous study, among six RR-MADD patients, P2 (identified as P12 in the original cohort) demonstrated the lowest muscle FAD level, accounting for only 56.3% of the average value observed in the other five patients and merely 33.8% of the average level measured in six healthy controls [15]. Consistent with the normal phenotype, his father (C2) exhibited a normal serum riboflavin level. The other two patients, P1 and P3, exhibited higher serum triglyceride levels compared with patients carrying biallelic ETFDH variants, suggesting another environmental factor that warrants close attention [15]. Apart from environmental factors, the impact of variant type on the enzyme activity of ETFDH protein varies significantly. Null variants in ETFDH often lead to lower residual ETFDH enzyme activity, whereas missense variants cause either severe misfolding of the protein or milder folding defects [16, 34]. Heterozygous null variants may thus confer greater risk for RR-MADD compared with missense variants. Our previous meta-analysis revealed that patients with single heterozygous variants in ETFDH exhibited a higher proportion of null variants than those with biallelic variants (34%, 95% CI [23%–48%] vs 21%, 95% CI [16%–27%]) [12]. This study has several limitations. First, fibroblasts were obtained from a single patient and his father, which limited the assessment of inter-individual variability. Studies involving a broader range of patient-derived cells carrying different ETFDH variants are needed to validate these findings. Second, while LRS significantly enhances the detection of deep intronic and regulatory variants, interpreting the pathogenicity of non-coding variants remains challenging. Furthermore, variants located in regions with extreme GC-content or complex repetitive structures may still be difficult to resolve reliably depending on the sequencing platform. Third, the phenotypic presentation in our patients may be influenced by modifier genes, compensatory pathways, or other unexamined environmental factors, which could explain the clinical heterogeneity observed despite identical ETFDH genotypes. This represents an important area for future investigation. In summary, this study identified three RR-MADD patients carrying single heterozygous variants in ETFDH, as revealed by WES, LRS, and cDNA sequencing, offering new insights into the genetic diagnosis of RR-MADD. Subsequent cell-based experiments and in vivo analyses confirmed that single heterozygous ETFDH variants confer increased susceptibility to disease manifestation under defined environmental stressors, suggesting an alternative inheritance pattern of RR-MADD. Supplementary Information Supplementary Material 1: Figure 1. Immunofluorescence assay confirms the morphology and expression of vimentin in fibroblasts (a). Sanger sequencing confirms the homozygous or heterozygous ETFDH c.917G>A variants in HEK293 cells (b) (n = 3). Supplementary Material 2: Figure 2. Dietary intervention protocol for experimental mice (a). The line chart shows changes in body weight during dietary intervention of male mice (b) (n = 4 mice per group) and female mice (c) (n = 3 mice per group). Comparison of motor function as assessed by total running time (d) and maximal running speed (e) in the treadmill test, and grip strength (f) across different groups before and after dietary interventions. Comparison of emotional behavioral indicators, including immobility time in the tail suspension test (g), and time in central zone (h) and zone transition number (i) in the open field test across different groups before and after dietary interventions (n = 6 mice per group). Supplementary Material 3. Supplementary Material 4. Acknowledgements We thank all participants for their interest and participation in the study. Authors’ contributions ZW and GJH designed the study; MJ wrote the original manuscript and performed the statistical analysis; MJ and ZHQ completed the cell culture and experiments; MJ, ZHQ, and JJQ completed the animal experiments; SMH and ZJS helped with cell culture; WQ and WJL helped analyze the data; LD and CXL collected the clinical information; WJ and DWS performed data curation; ZW, MJ, and JJQ reviewed and edited the manuscript. Funding This research was funded by the National Natural Science Foundation of China (grant number 82471432). Data availability The data supporting the results and conclusions are available in the paper and the Supplementary Materials, or from the corresponding author upon reasonable request. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Jing Ma, Huiqiu Zhang and Jiaqi Jiao are contributed equally to this work. Contributor Information Junhong Guo, Email: neuroguo@163.com. Wei Zhang, Email: zhangvey@126.com. References Associated Data Supplementary Materials Supplementary Material 1: Figure 1. Immunofluorescence assay confirms the morphology and expression of vimentin in fibroblasts (a). Sanger sequencing confirms the homozygous or heterozygous ETFDH c.917G>A variants in HEK293 cells (b) (n = 3). Supplementary Material 2: Figure 2. Dietary intervention protocol for experimental mice (a). The line chart shows changes in body weight during dietary intervention of male mice (b) (n = 4 mice per group) and female mice (c) (n = 3 mice per group). Comparison of motor function as assessed by total running time (d) and maximal running speed (e) in the treadmill test, and grip strength (f) across different groups before and after dietary interventions. Comparison of emotional behavioral indicators, including immobility time in the tail suspension test (g), and time in central zone (h) and zone transition number (i) in the open field test across different groups before and after dietary interventions (n = 6 mice per group). Supplementary Material 3. Supplementary Material 4. Data Availability Statement The data supporting the results and conclusions are available in the paper and the Supplementary Materials, or from the corresponding author upon reasonable request.