Autophagy is a self-degradative process responsible for the clearance of damaged or unnecessary cellular components. We have previously found that persistence of dysfunctional organelles due to autophagy failure is a key event in the pathogenesis of COL6/collagen VI-related myopathies, and have demonstrated that reactivation of a proper autophagic flux rescues the muscle defects of Col6a1-null (col6a1¡/¡) mice. Here we show that treatment with spermidine, a naturally occurring nontoxic autophagy inducer, is beneficial for col6a1¡/¡ mice. Systemic administration of spermidine in col6a1¡/¡ mice reactivated autophagy in a dose-dependent manner, leading to a concurrent amelioration of the histological and ultrastructural muscle defects. The beneficial effects of spermidine, together with its being easy to administer and the lack of overt side effects, open the field for the design of novel nutraceutical strategies for the treatment of muscle diseases characterized by autophagy impairment.

The key contribution of defective autophagy in the pathogenesis of muscle defects for these diseases is provided by studies in Col6a1-null (col6a1¡/¡) mice, in which abnormal signaling causes persistent AKT (v-akt murine thymoma viral oncogene homolog)-MTOR pathway activation and defective autophagosome formation in muscle fibers.8 Further studies in muscle biopsies of patients affected by Ullrich congenital muscular dystrophy (UCMD) and Bethem myopathy (BM), the 2 major human pathologies linked to mutations of COL6-encoding genes, confirm defective regulation of autophagy.8 Thereafter, defective autophagy induction has been reported in murine models for Duchenne muscular dystrophy and Emery-Dreyfuss muscular dystrophy, whereas increased autophagy flux contributes to the phenotype of dy3K/dy3K mice, the animal model of MDC1A congenital muscular dystrophy.5 Notably, the autophagic defects are reversible and rescue of a proper autophagic flux by either genetic, pharmacological or dietary approaches leads to a strong amelioration of the muscle defects, a finding which led to a pilot clinical trial aimed at restoring autophagy in UCMD and BM patients through a low-protein diet approach (ClinicalTrials.gov Identifier NCT01438788).

 Spermidine is a cationic polyamine naturally present in living cells at concentrations that decrease with aging. Spermidine supplementation to cell culture media or in vivo administration to entire animals prolongs life span in different model organisms, an effect which is linked to autophagy induction.9,10 Although the mechanisms by which spermidine acts are not yet fully understood, it has been shown that they involve a complex combination of acetylation and deacetylation events whose occurrence determines a shift in the acetylation pattern of the cell proteome without altering the total amount of acetylated proteins.9 Recently, the histone acetyltransferase EP300 has been described as the major target on which spermidine exerts an inhibitory activity, determining autophagy induction in vitro.11 Since in our previous studies we have found that reactivation of autophagy by either prolonged starvation or a low-protein diet is capable of rescuing the muscle pathology of col6a1¡/¡ mice,8 we evaluated in this myopathic model the efficacy of nutraceutical approaches able to forcibly induce autophagy. Here we show that spermidine administration by either addition to drinking water or intraperitoneal (i.p.) injection leads to a significant amelioration of the phenotype of col6a1¡/¡ mice. The ability to reactivate autophagy in COL6-deficient muscles, coupled with its low level of toxicity, points at spermidine as a nutraceutical with potential benefits for the treatment of these muscle diseases.

Spermidine reactivates autophagy in col6a1¡/¡ muscles To evaluate the therapeutic potential of spermidine in COL6- related muscle pathologies, we investigated wild-type and col6a1¡/¡ mice subjected to spermidine administration at different doses, durations and routes of administration. Spermidine was administered either by oral administration (per os), by addition to drinking water for 30 d, or i.p., by daily injection for 10 d. In both cases, 2 different doses were given (3 or 30 mM, per os; 5 or 50 mg/kg body weight, i.p.), in order to evaluate dosedependent effects (Table S1). Per os administration of spermidine at both 3 mM and 30 mM was effective in stimulating autophagy in skeletal muscles. Fluorescence microscope analysis of tibialis anterior showed several GFP-LC3 puncta in wild-type animals after 30 d spermidine treatment (Fig. 1A). Autophagy induction was further confirmed by immunoblotting analysis of cleavage and lipidation of MAP1LC3B/LC3B (microtubule-associated protein 1 light chain 3 b), quantified both as LC3B-II/-I and LC3B-II/ GAPDH ratios, which was increased at both low- and high-dose spermidine treatment (Fig. 1B and C; Fig. S1A and B). Of note, these treatments did not exert any overt change in the response to 24 h starvation in wild-type mice (Fig. 1A-C; Fig. S1A and B). In agreement with previous work,8 col6a1¡/¡ tibialis anterior displayed a markedly decreased autophagy in the 24-h fasting condition, when compared to the corresponding wild-type muscles.

 Interestingly, spermidine administration induced autophagy in col6a1¡/¡ muscles. Indeed, fluorescence microscopy showed the presence of GFP-LC3 puncta in tibialis anterior of spermidinetreated col6a1¡/¡ animals, which were clearly evident in mice treated with 30 mM spermidine and subjected to 24-h starvation (Fig. 1A). Immunoblotting analysis of tibialis anterior extracts confirmed a significant increase of LC3B-II/-I and LC3B-II/ GAPDH ratios in fasted col6a1¡/¡ mice treated with 30 mM spermidine, when compared to fasted col6a1¡/¡ mice without spermidine, whereas no significant change in LC3B-II/-I was detected in col6a1¡/¡ mice maintained under fed conditions or treated with the lower dose of spermidine (Fig. 1B and C; Fig. S2A and B). Interestingly, the LC3B-II/GAPDH ratio was instead significantly increased in fed col6a1¡/¡ mice following a low dose per os treatment (Fig. S1A). In order to further clarify the status of the autophagy machinery in the skeletal muscle of fed wild-type and col6a1¡/¡ mice subjected to per os spermidine treatment at different doses, we analyzed by immunoblotting the protein levels of RAB7, a small Ras-like GTPase with an essential role in lysosomal biogenesis and in the late phases of endosome and autophagosome fusion with lysosomes,12-13 and GABARAP, a yeast Atg8 ortholog playing a key role in the final stages of autophagosome biogenesis.14 Both the lipidated form of GABARAP (GABARAP-II) and RAB7 were increased in a dose-dependent fashion following spermidine treatment not only in wild-type but also in col6a1¡/¡ mice. These data indicate an increased rate of autophagosome and autolysosome formation, respectively, and are consistent with autophagy induction in fed animals of both genotypes (Fig. S2A-C).

 Altogether, these findings indicate that spermidine supplementation in drinking water for 30 d, promptly induces autophagy in wild-type muscles and also affects the autophagic process of col6a1¡/¡ muscles. In order to evaluate more precisely the effects of spermidine in inducing autophagy in skeletal muscles, we subjected wild-type and col6a1¡/¡ mice to daily i.p. injection with spermidine for 10 d (Table S1). Spermidine i.p. treatment at 5 mg/kg displayed no significant effects on fluorescent GFP-LC3 puncta and LC3B lipidation, both in wild-type and col6a1¡/¡ tibialis anterior (Fig. 1D and E; Fig. S1C). Slightly increased protein levels of GABARAP-II, but not RAB7, were detected in wild-type and col6a1¡/¡ tibialis anterior (Fig. S2D-F), further confirming that this dosage is unable to elicit a robust upregulation of autophagy. Conversely, spermidine i.p. treatment at 50 mg/kg triggered not only GFP-LC3 puncta formation and LC3B lipidation (Fig. 1D and F; Fig. S1D), but also a very strong increase in RAB7 and GABARAP-II protein levels (Fig. S2D-G and I) in muscles of wild-type mice maintained under fed conditions. Autophagy induction in this condition was further confirmed by the strong accumulation of the autophagosome markers LC3B-II and SQSTM1 (sequestosome 1) in tibialis anterior of wild-type mice treated simultaneously with spermidine i.p. 50 mg/kg and colchicine (Fig. S2G-K), an inhibitor of the late stages of autophagosome maturation and fusion with lysosome that allows investigating the autophagic flux in skeletal muscle.

15 Consistently with the role of RAB7 in the intracellular vesicle trafficking and in the fusion of autophagosomes with lysosomes, this protein displayed unaltered levels in spermidine plus colchicinetreated animals with respect to untreated control conditions (Fig. S2G-I). Unexpectedly, i.p. treatment with spermidine at 50 mg/kg combined with 24 h starvation led to decreased LC3B lipidation in wild-type muscles although autophagic vacuoles were clearly detectable (Fig. 1D and F; Fig. S1D). This latter finding may be explained by the intensive and prolonged autophagy induction in this condition, which would lead to the consumption of LC3B protein via autophagosome-lysosome fusion and degradation. Notably, treatment with spermidine at 50 mg/kg for 10 d led to an overt induction of autophagy in tibialis anterior of 24-hfasted col6a1¡/¡ mice, as indicated by the presence of a large number of GFP-LC3 puncta and by the conspicuous increase in LC3B lipidation (Fig. 1D and F; Fig. S1C). In fed conditions, autophagy appeared similarly induced in muscles from col6a1¡/¡ mice subjected to i.p. treatment with spermidine at 50 mg/kg, based on the increased presence of GFP-LC3 puncta (Fig. 1D) and on the higher levels of RAB7 and GABARAP-II (Fig. S2D to F). Moreover, combined treatment with colchicine and i.p. 50 mg/kg spermidine led to a significant accumulation of SQSTM1 and the lipidated form of LC3B in tibialis anterior of col6a1¡/¡ mice. Of note, differently from what we observed for wild-type mice, colchicine treatment alone did not determine an accumulation of any of these autophagic markers in muscles from col6a1¡/¡ mice (Fig. S2G-K).

 This evidence further confirms the autophagy impairment of COL6-deficient muscles, and represents an unambiguous proof of the concept that spermidine is capable of reactivating autophagy in muscles of col6a1¡/¡ mice. Altogether, these data indicate that spermidine is able to induce autophagy in skeletal muscle in a dose-dependent manner and is also effective in reactivating autophagy in col6a1¡/¡ muscles. Spermidine supplementation recovers the histological and ultrastructural defects of col6a1¡/¡ muscles To assess whether spermidine administration has any effect on the structural defects of COL6-deficient myopathic mice, we performed histological analysis of tibialis anterior muscles from wild-type and col6a1¡/¡ mice subjected to the different protocols of spermidine treatment. In wild-type mice, treatment with spermidine did not cause any sign of histological changes in muscle architecture at the tested doses and routes of administration (Fig. 2A and B). As expected,16 under untreated conditions col6a1¡/¡ tibialis anterior displayed an overt myopathic phenotype, with centrally nucleated myofibers, small atrophic fibers and giant hypertrophic ones. Notably, both i.p. and per os spermidine treatments led to a noticeable amelioration of the myopathic defects. The effects of spermidine seemed to correlate with the dose, as col6a1¡/¡ mice treated with the high doses displayed improved muscle histology (Fig. 2A and B). Indeed, 3 mM per os and 5 mg/kg i.p. spermidine treatments led to a partial rescue of histological alterations in col6a1¡/¡ mice and, albeit less extended, the myopathic defects were still present, with some atrophic and degenerating myofibers (Fig. 2A and B). 

However, col6a1¡/¡ animals displayed markedly improved muscle morphology both after 30 mM per os and 50 mg/kg i.p. spermidine treatment, with barely detectable signs of myofiber degeneration, more homogeneous myofiber size and rare atrophic fibers (Fig. 2A and B). Some centrally nucleated fibers were present in treated animals, consistent with myofiber regeneration (Fig. 2A and B). A well-known feature of col6a1¡/¡ muscles is the presence of altered organelles, in particular swollen mitochondria and dilated sarcoplasmic reticulum.8,17,18 Histological staining for SDH/succinate dehydrogenase, a mitochondrial marker, showed a more homogeneous labeling, pointing at an amelioration of the mitochondrial network in tibialis anterior of col6a1¡/¡ animals treated with higher spermidine doses (Fig. S3). Transmission electron microscopy of diaphragm ultrathin sections revealed a marked amelioration of sarcoplasmic reticulum and mitochondria ultrastructure in myofibers of spermidine-treated col6a1¡/¡ mice (Fig. 2C). Morphometric analysis confirmed a decreased percentage of aberrant mitochondria in muscles of spermidinetreated col6a1¡/¡ animals, which was significant after 30 mM per os treatment (P D 0.039) whereas it almost reached significance after 50 mg/kg intraperitoneal treatment (P D 0.077) (Fig. 2D). Furthermore, electron microscopy confirmed the presence of several autophagic vacuoles in spermidine-treated col6a1¡/¡ animals (Fig. 2C), consistently with the elimination of altered mitochondria through spermidine-induced autophagy. Altogether, these data indicate that spermidine-mediated reactivation of autophagy ameliorates the myopathic defects of col6a1¡/¡ animals, with a significant decrease of swollen mitochondria and a marked improvement of muscle structure.

Spermidine ameliorates the functional deficits of col6a1¡/¡ muscles Given the noticeable improvement of muscle structural defects, we next investigated whether the tested doses of spermidine were also capable of ameliorating the deterioration in other muscle features. A well-established feature of col6a1¡/¡ mice and of UCMD and BM patients is the increased incidence of apoptotic myofibers in skeletal muscles.8,17 TUNEL analysis for diaphragm revealed that spermidine administration elicited a strongly decreased incidence of apoptotic myofibers in col6a1¡/¡ mice. Indeed, both per os and i.p. treatments led to a significant decrease in the number of apoptotic myonuclei in col6a1¡/¡ diaphragms, and this effect was displayed at all the tested doses of spermidine (Fig. 2E). The decrease appeared stronger after i.p. treatment, where the effect positively correlated with the dose. However, per os treatment also displayed a clear effect, with about 35% decrease in the number of apoptotic myonuclei at both low and high doses (Fig. 2E). Conversely, wild-type animals displayed an increased number of TUNEL-positive myonuclei directly correlating with the dose both after i.p. and per os treatment (Fig. 2E), likely due to the excessive induction of autophagy elicited by spermidine in wild-type muscles. We have previously demonstrated that under physiological conditions col6a1¡/¡ mice develop significant less strength than wild-type mice.8 Therefore, we performed in vivo tetanic force measurements on gastrocnemius muscles of wild-type and col6a1¡/¡ mice under untreated conditions and following spermidine administration. Both per os and i.p. spermidine treatment led to a positive trend in normalized force in col6a1¡/¡ mice (Fig. S4),

 although the values obtained at a tetanic stimulation frequency of 100 Hz did not reach a statistically significant difference when compared to untreated animals (P D 0.138, per os treatment; P D 0.163, i.p. treatment). Of note, while the difference in the normalized force developed by wild-type and col6a1¡/¡ mice under standard conditions was highly significant, animals of the 2 genotypes displayed similar normalized force values after spermidine treatment (Fig. 2F; Fig. S4). As for apoptosis, high-dose i.p. spermidine administration caused a force decline in wild-type mice, likely due to the excessive autophagy induction elicited in these animals (Fig. 2F; Fig. S4B) as we previously observed in wild-type mice subjected to low-protein diet.8 Altogether, these data indicate that spermidine induction of autophagy recovers the apoptotic defects of col6a1¡/¡ animals, with also a positive trend in their muscle strength. Spermidine effects involve modulation of the AKT signaling pathway Spermidine induces autophagy through modifications at the level of the acetylproteome, thus modifying the expression levels of several genes.9,10 In previous studies aimed at mechanistic insight into the autophagic defects of COL6-related myopathies, we have demonstrated that col6a1¡/¡ muscles display a hyperactivation of the AKT-MTOR pathway and that this signaling defect plays a key role in the autophagy impairment.8 AKT kinase negatively regulates the activity of FOXO (forkhead box O) transcription factors by phosphorylation, and decreased levels of phosphorylated AKT represent a permissive condition for FOXO activation.1

9,20 To understand whether the autophagy-inducing effects of spermidine involve the AKT-FOXO axis, we investigated the activation state of AKT and the mRNA levels of several FOXO target genes in tibialis anterior of wild-type and col6a1¡/¡ mice subjected to i.p. treatment with spermidine at 50 mg/kg for 10 d. Chronic spermidine administration led to decreased AKT phosphorylation, an effect which was particularly strong in col6a1¡/¡ mice subjected to 24 h starvation (Fig. 3A). Real time PCR for Bnip3 (BCL2/adenovirus E1B 19kDa interacting protein 3), Ctsl (cathepsin L), and Map1lc3b, 3 autophagy-related genes regulated by the AKT-FOXO axis,19 showed significantly increased levels of their transcript in spermidine-treated, 24-h-fasted col6a1¡/¡ mice (Fig. 3B to D). These results indicate that spermidine elicits AKT dephosphorylation, which in turn triggers the translocation of FOXO transcription factors into the nucleus with the consequent increased transcription of FOXO target genes, including autophagy-related genes. Of note, high-dose spermidine i.p. treatment led to decreased AKT phosphorylation levels in tibialis anterior of wild-type mice even under fed conditions. Yet, AKT phosphorylation was restored when spermidine treatment was followed by 24-h starvation (Fig. 3A). This apparently paradoxical finding likely relies upon the stimulatory activity of several FOXO target genes, such as Sesn3 (sestrin 3) and Irs1 (insulin receptor substrate 1), on AKT. In turn, active AKT is known to upregulate protein synthesis and to inhibit FOXO activity, as a part of a feedback mechanism that maintains the balance between protein synthesis and protein degradation inside the cells.21

 In agreement with this concept, tibialis anterior of wild-type mice subjected to prolonged (48 h) fasting displayed significantly higher levels of phosphorylated AKT than those of 24-h-fasted wild-type animals (Fig. S5A). At the same time, expression of FOXO target genes was maintained at high levels in these muscles, as found for spermidine-treated and 24-h-fasted wild-type mice (Fig. 3B-D; Fig. S5B-D). Therefore, the excessive and prolonged stimulus for autophagy induction elicited in wildtype muscles by high-dose spermidine administration in combination with fasting induces a feedback mechanism that restores AKT phosphorylation. Altogether, these data indicate that spermidine affects AKT activation and enhances autophagy by promoting the transcription of FOXO target genes. Discussion Spermidine is a nontoxic autophagy-inducing natural compound, with an increasing interest for its applications as a nutraceutical tool for therapies. Spermidine counteracts aging and promotes longevity in different species by increasing the basal autophagy levels.9,10,22 Recent work suggests that this polyamine may be beneficial for counteracting several age-related pathologies, such as arterial aging and memory impairment.23-24 Moreover, spermidine shows a rapamycin-like effect in promoting the removal of brain protein aggregates and ameliorating the phenotype of a mouse model for TDP-43 proteinopathies,