The patient was 56-year-old male. His developmental milestones were normal. The limb strength was normal before he was 51-year-old, when he had progressive muscle weakness on lower extremities. His upper extremities were involved one month before first consultation. When consulted, he complained with difficulty in standing up from squatting position and lifting arms over his shoulder, as well as tiredness after walking for 100 meters and muscle tenderness. There was a past history of cigarettes smoking. There was no family history of muscle disease and the pedigree was shown in Fig. 1A. Neurologic examination revealed that only motor systerm involvement (the medical research council scale): Deltoids 3/5, biceps 4/5(left) and 4 + /5(right), triceps 5-/5(left) and 5/5(right), hip flexion 4/5, others were normal. No muscle atrophy. He had continuously elevated creatine kinase(CK) fluctuating between 383 to 928 U/L (38–174 U/L). Needle EMG showed only myopathic pattern on deltoids and quadriceps; Nerve conductions were normal. The ultrasound of heart, liver, gallbladder, pancreas, spleen and kidney as well as electrocardiogram was normal.

The clinical features of this patient. A The pedigree of this patient; B Compound heterozygous variants in FLAD1; C The Bam picture of sequence analysis showed the variant bases T and C were not on the same allele. D T2WI-FSE showed muscle atrophy and fatty degeneration of the posterior thigh group, and the fat signal was suppressed on T2WI-FSE-FS. FSE: Fast spin echo; FS Fat suppression

The next generation sequencing using blood sample was carried out for gene analysis, and it revealed compound heterozygous variants in FLAD1 (c.1588 C > T p.Arg530Cys reported previously [17]; c.1589 G > C p.Arg530Pro; Fig. 1B). Although DNA sample from the patient’s parents who passed away at an advanced age was not available, sequence analysis on Bam picture proved that variants c.1588 C > T and c.1589 G > C were from two isolated homologous chromosomes (Fig. 1C). The allele frequency for those two variants was 0.00002121 and 0 in gnomAD database, separately. The bioinformatic tools were conducted to predict the pathogenicity. Pathogenic prediction score by REVEL was 0.628 and 0.759 ( > 0.644 is deleterious); by SIFT was 0.000 and 0.003 (Both are damaging); by PROVEAN was −6.07 and −5.57 (Both are deleterious); by PolyPhen-2 was 0.986 and 0.988 (Both are probably damaging). The variant p.Arg530Cys was registered as pathogenic in ClinVAR database. According to the ACMG guidelines, the two variants are considered as pathogenic (PM2_supporting + PM3_very strong + PS3_supporting) and likely pathogenic (PM2_supporting + PM3 + PP3 + PM5), separately. Western blot against FADS (FLAD1 gene) in muscle sample before riboflavin treatment was significantly reduced in this patient (Supporting information 1).

His muscle strength was normal after 2 months treatment with riboflavin 80 mg/day. Then he was placed on a reduction dose of riboflavin 30 mg/day in the following 7 years. During follow-up, his CK level, needle EMG on deltoids, the ultrasound examination of heart and electrocardiogram were normal. The riboflavin level in serum was 10.8 ng/ml (Reference range: 2.3–14.6 ng/ml) after treatment. Acylcarnitine profile from dried blood spots revealed moderate elevations of C4-, C5-, C6-, C8-, C10-, C10:1-, C12-, C14:1-, C14:2- and C18:2-acylcarnitines after riboflavin regimen for 5 years, meanwhile his urine organic acid analysis showed mild elevation of ethylmalonic acid. Muscle MRI showed muscle atrophy and fatty degeneration on the posterior thigh group on T2WI-FSE (Fast spin echo) after 5 years of riboflavin treatment, moreover, the fat infiltration signal was suppressed on T2WI-FSE-FS (Fat suppression) (Fig. 1D).

Light-microscopic assessment of sections showed a marked variation in fiber size due to a large number of small atrophic fibers. Fibers with centrally placed nuclei accouted for about 20%. A few highly atrophic fibers formed nuclear clumps (Black arrow in Fig. 2A). There were a lot of fibers with small round vacuoles (Stars in Fig. 2A) which were proved as type I fiber in ATPase staining (Fig. 2G, H). There were a few necrotic and regenerating fibers and fibers with subsarcolemmally red staining mimicking ragged red fibers (atypical RRFs, Stars in Fig. 2B), as well as numerous COX-negative fibers(Stars in Fig. 2C). The ORO stain showed that lipid droplets were prominently increased predominantly in type I fibers (Fig. 2D). Diffuse SDH enzyme deficiency was noticed compared to normal control(Fig. 2E, F, separately). There was no group atrophy and the atrophic fibers are mainly type I (Fig. 2G, H). The immunohistochemistry studies did not show any abnormality.

Pathological features of this patient. A H&E staining showed marked variation in fiber size and fibers with centrally placed nuclei. A few nuclear clumps (Black arrow), necrotic and regenerating fibers and a lot of fibers with small round vacules (Stars) were documented; B MGT staining showed several atypical ragged red fibers(Stars); C COX-negative fibers(Stars); D ORO staining showed numerous fibers with lipid droplets; E SDH enzyme deficiency in patient; F SDH enzyme activity in normal control; G, H ATPase 4.3 and 10.4 proved the atrophic fibers are mainly type I

To elucidate the underlying mechanisms of RRMADD, the whole transcriptome sequencing together with integrated bioinformatics analysis were conducted in this patient with FLAD1-RRMADD, three patients with ETFDH-RRMADD and three normal controls. Venn diagram was used to display expressed gene between three groups shown in Fig. 3A. Summary of DEGs was shown in Fig. 3B. The results showed a total of 1610 DEGs between FLAD1-RRMADD and control group, among these 709 up-regulated and 901 down-regulated DEGs were identified, involving numerous signaling pathways (Supporting information 2-1); for that between ETFDH-RRMADD and control group, a total of 1143 DEGs, 623 up-regulated and 520 down-regulated DEGs were identified (Supporting information 2-2); for FLAD1-RRMADD and ETFDH-RRMADD group, a total of 1019 DEGs, 354 up-regulated and 665 down-regulated DEGs were identified (Supporting information 2-3).

Differential gene expression on FLAD1- RRMADD, ETFDH - RRMADD and normal controls. A Venn diagram was used to display expressed gene between samples; B Numbers of differential genes between groups. X axis represented the sample. Y axis represented the DEGs. C Volcanic maps showed that GDF15 gene was significant up-regulation in FLAD1-RRMADD vs control group. D Volcanic maps showed that GDF15 gene was significant up-regulation in ETFDH-RRMADD vs control group. E GDF15 level in serum of FLAD1-RRMADD after riboflavin treatment (N = 1) and ETFDH-RRMADD (before and after riboflavin treatment, N = 5), and normal controls (N = 10). *p ≤ 0.05, **p < 0.01, ***p < 0.001. F Pathway classifications of DEGs. X axis represents number of DEGs. Y axis represents functional classification of GO classification. The main differences for the co-expressed genes of FLAD1- RRMADD vs control and ETFDH- RRMADD vs control were fatty acid metabolism process. G Pathway classifications of DEGs. X axis represents number of DEGs. Y axis represents functional classification of KEGG pathway classification. The main differences for the co-expressed genes of FLAD1- RRMADD vs control and ETFDH- RRMADD vs control were biosynthesis of amino acids and fatty acid metabolism

Volcanic maps showed that the expression of GDF15 mRNA was overwhelmingly significantly up-regulated in both FLAD1-RRMADD vs control and ETFDH-RRMADD vs control group (Fig. 3C, D). We next detected the GDF15 protein level in serum in FLAD1-RRMADD after riboflavin treatment (N = 1) and ETFDH-RRMADD (before and after riboflavin treatment, N = 5). The GDF15 level in serum was an obvious increase in ETFDH-RRMADD group before and after riboflavin treatment comparing to normal controls (N = 10, p < 0.001, Fig. 3E). For FLAD1-RRMADD, the GDF15 level in the only patient after riboflavin treatment was significantly higher than the average of 10 normal controls (862.75 vs 342.68). The protein expression of GDF15 in muscle samples before riboflavin treatment was proved significantly increased in both FLAD1-RRMADD and two ETFDH-RRMADD patients (Supporting information 1).

Volcanic maps also showed that the expression of TRIB3 and KLHDC7B mRNA was significantly up-regulated in both FLAD1-RRMADD vs control and ETFDH-RRMADD vs control group (Fig. 3C, D). Nevertheless, the protein study for those two genes failed to support the results that were found in mRNA study.

GO classification for the co-expressed genes of FLAD1-RRMADD vs control and ETFDH-RRMADD vs control showed that the main differences were fatty acid metabolism process (Fig. 3F). KEGG pathway classification analysis for the co-expressed genes of FLAD1-RRMADD vs control and ETFDH-RRMADD vs control showed that the main differences were biosynthesis of amino acids and fatty acid metabolism (Fig. 3G).

We reviewed literature and listed a summary of 18 cases with MADD/GAII caused by FLAD1 variants including the present case (Supporting information 3), in which 10 patients were proved with full or partial riboflavin responsiveness (Full riboflavin responsiveness: After riboflavin treatment, the symptoms including muscle weakness disappear and the patient returns to a normal life. Partial riboflavin responsiveness: After riboflavin treatment, the symptoms are partially improved, the patient’ condition is stable but not normal) [9, 17, 19,20,21,22,23]. The comparison between 10 FLAD1-RRMADD (9 cases from literature and the present case) and 106 ETFDH-RRMADD (in our neuromuscular center) was shown in Table 1.

The onset age of FLAD1-RRMADD ranges from birth to 51 years, in which 6 out of 10 cases presented in infancy, in contrast, cases with ETFDH-RRMADD tend to be late-onset. On muscle pathology, 100% of FLAD1-RRMADD or ETFDH-RRMADD were proved with lipid storage myopathy. The atypical RRFs were more frequent in ETFDH-RRMADD, while the fibers with faint COX staining were more common in FLAD1-RRMADD (p < 0.001, respectively). On muscle weakness symptom, accompanying symptoms such as vomiting, electromyography, serum CK level, and other muscle pathology features such as lipid storage or decreased SDH activity, it was similar between FLAD1-RRMADD and ETFDH-RRMADD.