Boosting the peripheral immune response in the skeletal muscles improved motor function in ALS transgenic mice.

Monocyte chemoattractant protein-1 (MCP1) is one of the most powerful pro-inflammatory chemokines. However, its signalling is pivotal in driving injured axon and muscle regeneration.

progression and overall survival [24,25]. We found that, despite the same extent of MN loss during the disease progression [24], the fast-progressing mice (129SvSOD1 G93A ) showed earlier muscle denervation and higher axonal dysregulation that correlated with a poor in ammatory response and reduced macrophages in ltration in the periphery [26,27] compared with the slow-progressing ALS mice (C57SOD1 G93A ). Further analyses showed that fast-progressing ALS mice failed to activate Monocyte Chemoattractant Protein 1 (MCP1) in MN perikarya and peripheral axons compared to C57SOD1 G93A mice [26,28]. This evidence suggests that MCP1 signalling and immune cell recruitment might be pivotal in delaying muscular denervation and triggering regeneration in the PNS [29,30], thus regulating the speed of the disease progression of the two ALS models.
MCP1 levels are increased in serum and cerebrospinal uid of sporadic and familial ALS patients [39]. Besides, studies in mSOD1 mice have shown that MCP1 is signi cantly upregulated in the spinal cord and peripheral nerves at the early disease stage, suggesting a pathogenic role of this chemokine [26,40].
While the effective role of immune cell recruitment in the PNS is still controversial [15,16,54,55], no experimental inference is available on the immune response in skeletal muscles during ALS course.
In the present study, we investigated the therapeutic e cacy of a scAAV9 vector engineered with the Mcp1 gene injected in the skeletal muscles of fast and slow progressing SOD1 G93A mice. We found that an early boosting of immune response in the peripheral compartment is crucial in countering the denervation atrophy and slowing down the ALS progression in slow-but not fast-progressing mSOD1 mice. Moreover, our data described a pleiotropic role of MCP1 in the CNS as a protective factor, able to modulate the neuroin ammation, possibly reducing the MN loss. This evidence is instrumental in comprehending the contribution of the immune response in ALS, shedding light on its worth in governing the speed of the disease progression.
Adult (6 or 8 weeks-old) SOD1 G93A mice underwent a single bilateral intramuscular injection of 2,18x1010 vg/μL of the scAAV9 vector. The scAAV9 opportunely diluted in sterile PBS was injected in both hindlimb (Tibialis Anterior, TA; Gastrocnemius Caput Medialis, GCM; Gluteus Maximus, GM) and forelimb (Triceps Brachii, TB) muscles following the protocol previously described by Gruntman et al. [56]. Brie y, mice were anaesthetised with iso urane inhalation, fur shaved to visualise the target muscles, and a 30-gauge needle was inserted in the muscle centre to inject the scAAV9 (10μL/muscle). Mice were divided into the treated (scAAV9_MCP1) and control (scAAV9_(empty)) groups through a block-randomisation in which the blocks are de ned on the body weight, sex and sibling separation.

Behavioural analysis
Starting from 8 weeks of age, motor onset and disease progression were monitored bi-weekly in C57SOD1 G93A and 129SvSOD1 G93A scAAV9-treated mice by a blind operator recording the body weight and the motor performance at the paw grip endurance (PaGE) test. In the PaGE test, mice are placed on a horizontal grid at 30 cm height, and the tail is gently pulled until they grasp the grid with their fore and hind paws. The grid is gently turned upside down, and the latency time of the mouse to fall on the table is recorded for a maximum of 90s. Each mouse is given up to three attempts, and the most prolonged latency is recorded. The onset of muscle strength de cit is considered when the mice showed the rst signs of impairment in the PaGE test. In C57SOD1 G93A mice, the latency was evaluated as previously described [57]; conversely, in 129SvSOD1 G93A animals, the performance obtained in the grip strength test was assessed through a score, calculated as indicated by Lauranzano et al. [58].

Immunohistochemical analysis
Mice were anaesthetised with a mix of ketamine (1.75 mg/Kg) and medetomidine (1 mg/Kg) and transcardially perfused with 50 ml of 0.1M PBS pH 7.4. Following the blood removal, skeletal muscles (TA, GCM, GM and TB) and nerves were dissected and immediately frozen in cooled isopentane or npentane, respectively. At the same time, the vertebral column was post-xed overnight in a solution of 4% paraformaldehyde in 0.1M PBS. The following day the vertebrae were removed, and the spinal cord was transferred to 30% sucrose solution with 0.1% sodium azide in 0.1 M PBS at 4°C for cryoprotection before mounting in optimal cutting temperature compound (Tissue-Tek, Sakura).
Fluorescence-labelled spinal cord sections were analysed under a sequential scanning mode to avoid bleed-through effects with an IX81 microscope equipped with a confocal scan unit FV500 with three laser lines: Ar-Kr (488 nm), He-Ne red (646 nm), and He-Ne green (532 nm) (Olympus, Tokyo, Japan) and a UV diode using a 10x objective. For lumbar motor neurons count (one every ten sections), a total of 12 serial ChAT-stained sections were analysed. The neuron areas were analysed with Fiji software (Image J, U. S. NIH, Bethesda, Maryland, USA). As previously indicated [59], only neuronal somas with an area ≥ 400μm 2 were considered for quantitative analysis of MN numbers. Fluorescence-labelled sections images (3/5 per animal) of the Tibialis Anterior and Triceps Brachii muscle were analysed with an Olympus virtual slide system VS110 (Olympus, Center Valley, USA) and acquired at 20x magni cation. A systematic random sampling procedure was applied as previously described [57,60]. Brie y, a grid of equivalent sampling elds was outlined on the muscle slice pro le. To ensure that every part of the slice had an equal chance of being sampled, a bidimensional stereological sampling procedure was applied analysing equivalent elds placed at a xed distance from each other on the tissue slice, using the "grid" function in Fiji (Image J, U. S. NIH, Bethesda, Maryland, USA). The same approach was used to evaluate the neutrophil elastase staining by calculating the percentage of covered area (Area fraction %) per eld for each section in the analysis with Fiji software.

Morphometric analysis of muscles
Tibialis Anterior and Triceps Brachii muscles were dissected out and snap-frozen in liquid nitrogen. For the muscle bres composition (SDH staining), 10 μm-thickness serial coronal cryosections from the midbelly region of the TA muscle were air-dried and then incubated at 37°C for 30′ in phosphate buffer (0.2 M, pH 7.6) containing 13.5 mg/mL Na-succinate (Sigma-Aldrich) and 0.5 mg/mL of nitro blue tetrazolium (Sigma-Aldrich, 0.29 mg/mL of buffer solution). After staining, sections were xed with 4% paraformaldehyde, dehydrated in 15% alcohol for 5′ and nally mounted with DPX compound (Sigma-Aldrich).
Images were acquired with an Olympus virtual slide system VS110 (Olympus, Center Valley, USA) at 20x magni cation and analysed through Fiji (Image J, U.S. NIH, Bethesda, Maryland, USA) on 3/5 serial sections per animal. For the SDH staining, a systematic random sampling procedure was applied as described above. For the muscle bre cross-sectional area and centralised nuclei, the entire TA or TB muscles section was analysed with the "MuscleJ" plug-in of Fiji software as previously described [61].

Western blot
Mice were anaesthetised with a mix of ketamine (1.75 mg/Kg) and medetomidine (1 mg/Kg) and transcardially perfused with 50ml of 0.1M PBS pH7.4. Following blood removal, skeletal muscles were dissected out and immediately frozen in cooled isopentane. The spinal cord was uxed from the vertebral column employing sterile physiological solution (0.9% NaCl) and dissected in the three main segments (i.e. cervical, thoracic and lumbar). Spinal cord segments and nerves were immediately frozen on dry ice. Protein lysates were obtained by homogenisation of mice skeletal muscles, sciatic nerves and spinal cords in lysis buffer as previously described [57]. Brie y, tissues were powdered in liquid nitrogen then homogenised by sonication in ice-cold homogenisation buffer (Tris HCl pH 8 50 mM, NaCl 150 mM, EGTA pH 8.5 mM, MgCl2 1.5 mM, Triton x-100 1%, anhydrous glycerol 10%, phosphatases and proteases inhibitor cocktail Roche), centrifuged at 13000 rpm for 15 min at 4°C and the supernatants were collected and stored at -80°C.

Real-Time PCR
Tissues (spinal cords, sciatic nerves and muscles) were freshly collected and immediately frozen on dry ice after mouse perfusion with 0.1 M PBS. The total RNA from tissues was extracted using the Trizol method (Invitrogen) and puri ed with PureLink RNA columns (Thermo Fisher) following the manufacturer's instructions. RNA samples were treated with DNase I, and reverse transcription was done with a High Capacity cDNA Reverse Transcription Kit (Thermo Fisher). For Real-time PCR, we used the Taq Man Gene expression assay (Applied Biosystems) following the manufacturer's instructions on cDNA specimens in triplicate, using SensiFAST Probe Hi-ROX Kit (BioLine) and 1x mix containing the speci c probes (Thermo Fisher). The following probes (Thermo Fisher) were used for the real-time PCR assay: MCP1 (Mcp1, Mm00441242_m1); cholinergic receptor nicotinic gamma subunit (Chrng, Mm00437419_m1); CD8 alpha receptor (CD8a; Mm01182107_g1); CD4 alpha receptor (CD4a; Mm00442754_m1); Forkhead box P3 (Foxp3; Mm00475162_m1); Insulin-like Growth Factor 1 (Igf1; Mm00439560_m1); Tumour Necrosis Factor-alpha (Tnfα, Mm00443258_m1); Macrosialin (CD68; Mm03047343_m1); Interleukin 4 (Il4, Mm00445259_m1); Interleukin 1 beta (Il1β, Mm00434228_m1). Relative quanti cation was calculated from the ratio between the cycle number (Ct) at which the signal crossed a threshold set within the logarithmic phase of the given gene and that of the reference β-actin gene (Mm02619580_g1). Mean values of the triplicate results for each animal were used as individual data for the Livak relative gene expression analysis (2 -ΔΔCt ).

Statistical analysis
All the statistical analyses were performed using Prism 9 for Windows (GraphPad Software Inc.). Values are reported as mean ± SEM. For each analysis, the dependant and group variable are respectively named on the y-and x-axis of the graph.
The sample size for behavioural analysis was de ned according to the "Guidelines for preclinical animal research in ALS/MND: A consensus meeting" [64]. Parameters (body weight and PaGe test) used to evaluate disease progression in SOD1 G93A mice were analysed by repeated-measures ANOVA followed by Sidak's post-analysis. Symptoms onset was analysed by Log-rank Mantel-Cox test, and Kaplan-Meier plots were generated.
Mean values ± standard deviation were used for statistical analysis by Student's t-test for two groups or by One-way ANOVA followed by Fisher's multiple comparison test for more than two groups. In the case of two independent variables, two-way ANOVA followed by Fisher's LSD multiple comparison test was performed. D'Agostino & Pearson omnibus normality test and relative QQ plots were used to assess the assumption of normality. In the event of populations with unequal variance, the Brown-Forsythe ANOVA test followed by the unpaired t-test with Welch's correction was applied.
For all analyses, a p-value < 0.05 was considered statistically signi cant. The asterisk * indicates the comparison with the non-transgenic littermates, while the dot ° indicates the comparison between scAAV9_MCP1-and scAAV9_(empty)-treated mice. Further details, including p-values and number of samples, are documented in the Results, Figures, and relative captions.

Results
MCP1 is more expressed in the central and peripheral nervous system of C57SOD1 G93A than 129SvSOD1 G93A mice We previously found that MCP1 was signi cantly upregulated by MNs and peripheral axons of slowprogressing than fast-progressing SOD1 G93A mice at the disease onset [26,28]. The immunohistochemical analysis supported the higher activation of MCP1 in the spinal cord of C57SOD1 G93A compared with 129SvSOD1 G93A model (Fig. 1A, B), con rming the chemokine expression by MNs (Fig. 1C Fig. 1G-I) in the CNS of mSOD1 mice [40]. Notably, a temporal change in the chemokine expression pattern was found in C57SOD1 G93A as the disease progresses, characterised by a signi cant MCP1 expression by microglia and MNs at the pre-symptomatic and onset disease stage, respectively, followed by a widespread activation at the advanced phases (Fig. 1E). Specularly, immunohistochemical analysis revealed a progressive increase of MCP1 expression in the sciatic nerve of C57SOD1 G93A mice as the disease progresses ( To hit the neuromuscular system of ALS mice, we selected the self-complementary adeno-associated virus serotype 9 (scAAV9) in light of its ability to target the spinal cord hijacking the axonal transport machinery and travel along the nerve following intramuscular (i.m.) injection [65]. A single bilateral i.m. To analyse the effect of the MCP1 boosting on the disease progression, the scAAV9 vector properly engineered with the murine sequence of Mcp1 (scAAV9_MCP1) was injected in the hindlimb (Gastrocnemius Caput Medialis, GCM; Tibialis Anterior, TA; Gluteus Maximus, GM) and forelimb (Triceps Brachii, TB) skeletal muscles of pre-symptomatic (8 weeks-old) C57SOD1 G93A mice ( Fig. 2A). Given the early muscle de cit in mSOD1 mice [66,67], eight mice per group were sacri ced ~two weeks before the motor onset (14 weeks), whilst ten mice were monitored until the symptomatic disease stage (20 weeks). A scAAV9(empty) vector was used as control ( Fig. 2A).
During the study, no difference in the body weight was observed between the two experimental groups, excluding any major side effect upon the induction of a pro-in ammatory factor in ALS mice (Fig. 2B). Notably, in the scAAV9_MCP1 treated mice the impairment of muscle strength was delayed and progressed slowly up to 20  The MCP1 boosting in the hindlimb skeletal muscles of C57SOD1 G93A mice delayed the denervation atrophy and prompted muscle re-innervation The impairment of skeletal muscles is an early event in the ALS pathogenic cascade [68][69][70], pivotal in determining the motor ability of mSOD1 mice. To dissect the effect of the MCP1 boosting on muscular degeneration, we rst investigated the TA muscle given the advanced susceptibility of the lower motor units in mSOD1 mice [71] and the high composition in fast-fatigable bres, which are early affected by the disease [66,72]. In keeping with the ameliorated clinical phenotype, at 14 weeks, the TA muscle of scAAV9_MCP1-treated mice was less jeopardised (Fig. 3A). Our analysis recorded a reduction of 38.8±2.6% (mean±SEM) of the muscle mass in the scAAV9(empty)-treated mice compared with nontransgenic (Ntg) littermates, which decreased at 25.9±2.3% (mean±SEM) upon scAAV9_MCP1 injection.
Moreover, a signi cant downregulation of the foetal gamma-subunit of the acetylcholine receptor (AChRγ) was recorded upon MCP1 induction, indicating considerable preservation of muscle innervation compared with the scAAV9(empty)-treated mice [73] (Fig.3B). Accordingly, the histological analysis revealed a reduced percentage of denervated neuro-muscular junctions (NMJs) in the hind paw muscle of scAAV9_MCP1-compared with the scAAV9(empty)-treated mice ( Supplementary Fig. 3A, C). Intriguingly, the histological examination uncovered a higher proportion of motor axons expressing the Growth Associated Protein 43 (GAP43) upon MCP1 boosting, indicating the reinstatement of the nerve sprouting in the fast-twitch muscle of SOD1 G93A mice [74,75] (Supplementary Fig. 3A, D). A higher rate of GAP43 + regenerating motor axons and innervated NMJs was still detectable in the scAAV9_MCP1-treated group at the symptomatic disease stage ( Supplementary Fig. 3B-D), albeit without impinge on the AChRγ transcription level nor TA muscle atrophy compared with the control group (Fig. 3A, B). Altogether these data align with previous evidence suggesting maladaptive axonal sprouting in mSOD1 mice as the disease progresses [76,77].
The early MCP1 boosting favoured the establishment of an anti-in ammatory milieu in the skeletal muscles of C57SOD1 G93A mice Multiple evidence correlates the protective effect of the MCP1-mediated in ammation to its chemoattractant activity towards immune cells, particularly macrophages [29,49,78], which are pivotal at sustaining skeletal muscle healing upon damage [79]. Accordingly, we next analysed the chemokine levels within the TA muscle of the scAAV9_MCP1-treated and control group at the pre-symptomatic and symptomatic disease stages. Mcp1 transcript resulted signi cantly upregulated in the TA muscle of C57SOD1 G93A mice compared with the NTg littermates at 20 but not 14 weeks of age and dramatically increased by scAAV9_MCP1 injection at both time points (Fig. 3C). Notably, the extent of chemokine induction recorded in scAAV9_MCP1-treated mice at 14 and 20 weeks of age was similar, con rming the ability of the scAAV9 at inducing the chemokine several weeks after the single i.m. injection (FC vs Ntg 14wks: 548.8±81.7; 20wks: 542.1±82.5. mean±SEM). Accordingly, we found a signi cant increase in the recruitment of phagocytic CD68 + macrophages within the TA muscle of scAAV9_MCP1-treated mice compared with the control groups. This effect was signi cant at 14 but not 20 weeks of age, when macrophages massively in ltrate the skeletal muscle of mSOD1 mice [80] (Fig. 3D-F).
Therefore, we investigated the in ammatory ngerprint acquired by the MCP1-recruited immune cells in the skeletal muscle of 14 weeks-old C57SOD1 G93A mice. The histological analysis revealed that the percentage of the M1 iNOS + myeloid cells in ltrated in the TA muscle signi cantly dropped upon MCP1 boosting, whereas a remarkably increase of the M2 CD206 + counterpart was recorded compared with the scAAV9(empty)-treated mice (Fig. 4 A-D). The examination of the muscular in ammatory milieu in the scAAV9_MCP1 treated mice revealed a signi cant downregulation of the Insulin-like Growth Factor 1 (Igf1) compared with the control group (Fig. 4E), a cytokine released by M1-macrophages exerting an autocrine function pivotal to trigger the M2-gene programme [81]. In keeping with this, in the TA muscle of scAAV9_MCP1-treated mice, we found a higher increase of Sirtuin 1 (Sirt1) deacetylase protein level (Fig.  4H), whose overexpression in skeletal muscle is associated with macrophage polarisation shift towards the anti-in ammatory phenotype [82,83]. This, together with the signi cant Tumour necrosis factor-α (Tnfα) downregulation ( CCR2 is also expressed by activated T lymphocytes [84], instrumental in the regenerative mechanisms of skeletal muscles [52]. The analysis of CD4 and CD8a transcripts did not reveal any difference between the two groups of C57SOD1 G93A mice ( Supplementary Fig. 4A, B), whilst signi cant upregulation of FoxP3 transcript was recorded upon MCP1 induction, suggesting an increased in ltration of T regulatory lymphocytes (T regs) compared with the scAAV9(empty) group ( Supplementary Fig. 4C). As indicated by the ne kinetic governing the immune response within the injured muscle [53,85,86], T regs are the last immune cells in ltrating into the injured tissue pivotal at sustaining the pro-healing programme [87,88].
Altogether, this evidence suggests that the MCP1 induction has anticipated the physiological immune response within the hind paw muscles of C57SOD1 G93A mice. In keeping with this, a reduced release of the elastase enzyme was recorded in the TA muscle of scAAV9_MCP1-treated mice compared with the control group ( Supplementary Fig. 4D, E), indicating that neutrophils, which are the rst immune cells recruited by the chemotactic gradient established within the damaged muscle [85,89,90], have already given way to leucocytes.
Despite the signi cant Mcp1 upregulation (Fig. 3C), no difference in macrophages (Fig. 3D, F) and T cells ( Supplementary Fig. 4A-C) recruitment was found within the TA muscle of scAAV9_MCP1-and scAAV9(empty)-treated mice at 20 weeks of age. Besides, no variation was registered in the in ammatory response between the two experimental groups other than a signi cant reduction in Igf1 levels ( Supplementary Fig. 4F) and the Cytochrome b-245 heavy chain (gp91 PHOX ) expression in the scAAV9_MCP1-treated group ( Supplementary Fig. 4G, H).
The early MCP1-mediate boosting of the immune response triggered myogenic progenitor cell differentiation in the hindlimb skeletal muscle of C57 SOD1 G93A mice The ability of recruited immune cells at governing the myogenic programme upon damage is strictly dependent on the acquired in ammatory ngerprint. While the M1-macrophages promote activation and proliferation of the myogenic progenitors, their switch towards the M2 phenotype is fundamental to sustain the nal commitment of satellite cells (SCs) towards myogenesis [85,[91][92][93][94].
We showed that scAAV9_MCP1 injection anticipated the immune response in the TA muscle of C57SOD1 G93A mice, favouring the establishment of an anti-in ammatory pro-regenerative milieu. Therefore, we next assessed the impact of MCP1 boosting on the expression of two critical myogenic factors in the TA muscle of 14 weeks-old C57SOD1 G93A mice: Paired Box 7 (Pax7), the hallmark of SCs stemness [95], and Myogenin (MyoG), a marker of early commitment and differentiation [96]. MyoG, but not Pax7, resulted signi cantly upregulated in the TA muscle of scAAV9_MCP1-treated mice compared with the control group ( Fig. 5A-C). Besides, our analysis revealed that the expression of the Myoblast Determination protein 1 (MyoD), a transcription factor critical at de ning the activated SCs fate [97,98], was signi cantly increased upon MCP1 boosting (Fig. 5A, D). Suitably, the histological examination showed a reduction in the percentage of quiescent (Pax7 + /MyoD -) and a signi cant increase of differentiating (Pax7 -/MyoD + ) SCs in the TA muscle of scAAV9_MCP1-treated mice compared with controls ( Supplementary Fig. 5A, D). According to the increased myogenic activity, a higher percentage of centralised myonuclei was recorded in the hind paw muscle of C57SOD1 G93A mice upon MCP1 boosting (% Vs Ntg: 191.0±54.7, Empty; 380.7±44.2, MCP1. mean±SEM), indicating an intense regenerative process compared with the scAAV9(empty) group [99] (Supplementary Fig. 5B, E).
Given that the shift from large fast-twitch to small slow-twitch bres is a common hallmark of ALS [100,101], we next estimated the TA oxidative muscle bre composition by the succinic dehydrogenase (SDH) histochemical assay. As expected, SOD1 G93A mice exhibited a higher percentage of slow-twitch oxidative bres (87.3±1.4%, mean±SEM) compared with Ntg littermates (52.9±0.1%, mean±SEM), which signi cantly decreased upon MCP1 boosting (74.7±1.9%, mean±SEM) ( Supplementary Fig. 5C, F). In keeping with this, the TA muscle bres mean cross-sectional area (CSA) was higher in the scAAV9_MCP1treated mice compared with the control group (Fig. 5E, F). Besides, the histological analysis revealed a signi cant reduction in the percentage of the small (<205μm 2 ) and compensatory preservation of medium (1000-2000μm 2 ) bres in the TA muscle of C57SOD1 G93A mice upon MCP1 boosting (Fig. 5E, G).
Altogether, this evidence suggests a possible correlation between the M2 polarisation of the muscular in ammatory milieu and the enhanced myogenic activity in the TA muscle of the scAAV9_MCP1-treated mice. This effect progressively weakened during the disease progression as, at 20 weeks, we did not record any differences between the two groups of SOD1 G93A mice in the myogenic programme, despite the signi cant Mcp1 upregulation in the skeletal muscles (data not shown).
The MCP1-mediated boosting preserved motor axon from demyelination in the sciatic nerves of C57SOD1 G93A mice We previously reported an association between the activation of the MCP1-mediated pathway in the PNS and a slower disease progression of SOD1 G93A mice [26]. Therefore, we analysed the effect of MCP1 induction and the eventual immune cell recruitment within the sciatic nerve of C57SOD1 G93A mice. Unlike skeletal muscle, the treatment modestly increased Mcp1 levels in the sciatic nerves at 14 weeks (Fig. 6A). However, this did not further enhance the macrophages nor cytotoxic T cells recruitment, as demonstrated by the unchanged levels of CD68 and CD8a transcripts than the scAAV9(empty)-treated group (Fig. 6B, C). In keeping with this, no difference in the Tnfα transcription was recorded in the sciatic nerve of SOD1 G93A mice compared with the NTg littermates (Fig. 6D).
At the symptomatic disease stage, the gene expression analysis showed a signi cant and similar upregulation of Mcp1 and CD68 transcripts in the PNS of both groups of SOD1 G93A mice compared with NTg littermates (Fig. 6A, B). Intriguingly, our analysis revealed a signi cant decrease in the CD68 + macrophages recruitment (Fig. 6B) and a reduction trend in the CD8 + lymphocytes in ltration (Fig. 6C) along motor axons of C57SOD1 G93A mice upon MCP1 boosting. The diminished leucocytes recall signi cantly abated the PNS in ammation in scAAV9_MCP1-treated mice, as demonstrated by the Tnfα downregulation compared with the control group (Fig. 6D). This effect translated into an increased expression of the p75-neurotrophin receptor (p75 NTR ) within the sciatic nerves of scAAV9_MCP1-treated mice compared with controls ( Figure 6E, F), suggesting increased motor axon regeneration and remyelination upon MCP1 boosting [17,102]. Suitably, whilst the heavy neuro lament (NF200) and myelin basic protein (MBP) levels were signi cantly downregulated in the PNS of symptomatic SOD1 G93A mice, their expression resulted unchanged in the scAAV9_MCP1-treated mice compared with the NTg littermates (Fig. 6E, G, H). Altogether, these observations suggest the almost complete maintenance of the axonal structure and myelin ensheathment of the sciatic nerve of symptomatic C57SOD1 G93A mice upon MCP1 boosting.
The MCP1 induction within spinal motor neurons of C57SOD1 G93A mice is protective by decreasing neuroin ammation We showed that the scAAV9_GFP spreads retrogradely from the injected muscles alongside the motor unit of mSOD1 mice, nally transducing MN soma ( Supplementary Fig. 2). Therefore, we analysed the effect of chemokine induction on the neurodegenerative signature of ALS.
The gene expression analysis con rmed a signi cant Mcp1 upregulation in the lumbar spinal cord of scAAV9_MCP1-treated mice compared with the control group at 14 weeks (Fig. 7D). Conversely, at 20 weeks, no difference was recorded between the two groups of SOD1 G93A mice (Fig. 7D), suggesting that, at the full-blown stage, the massive chemokine expression by microglia ( [103,104], Fig. 1E, Supplementary Fig. 1C, F) (Fig. 7A-C). At 14 weeks, the anti-in ammatory markers Interleukin 4 (Il4) (Fig. 7E) and Arg1 (Fig. 7G, H) were signi cantly upregulated in the CNS of scAAV9_MCP1-treated mice; whereas, at the symptomatic disease stage, the treatment resulted in a signi cant downregulation of the pro-in ammatory factors Interleukin 1β (Il1β) (Fig. 7F) and gp91 PHOX (Fig. 7G, I) compared with the scAAV9(empty) group. These modi cations did not alter the glia activation state, as demonstrated by the unchanged expression levels of the Ionised calcium-binding adapter molecule 1 (Iba1) and glial brillary acidic protein (GFAP) between the two groups of C57SOD1 G93A mice ( Supplementary Fig. 6A, B). These data suggest that the MCP1 boosting in spinal MNs might have extended the so-called "stable phase" of the disease in SOD1 G93A mice [105,106], preserving the glia towards an anti-in ammatory phenotype followed later by the inhibition of pro-in ammatory environment, which re ected in MN preservation.
The MCP1 boosting in forelimb skeletal muscles of C57 SOD1 G93A mice delayed the denervation atrophy through immune-related myogenesis The mSOD1 mice rst develop hindlimb tremors, then progressive hindlimb weakness with rapidly deteriorating gait, eventually culminating in the paralysis of one or both hindlimbs [71,[107][108][109]. Forelimbs function remains comparatively spared throughout the disease progression, indicating a distinct susceptibility of the upper motor unit in mSOD1 mice [57,71,110]. This evidence highlighted the importance of the forelimbs contribution in the disease progression of ALS mice, particularly at the advanced disease stage [57].
Likewise the hindlimbs, alterations in the forepaws could be detectable before evident motor impairment [67]. Accordingly, the muscle weight measurement showed that at 14 weeks the TB muscle of SOD1 G93A mice has already lost the 19.3±2.6% (mean±SEM) of its mass compared with the NTg littermates, which increased to 45.8±4.6% (mean±SEM) at 20 weeks. Notably, the MCP1 boosting signi cantly preserved the forepaw muscle of C57SOD1 G93A mice from the atrophic phenomenon reducing the muscle mass loss to 2.5±1.9% and 33.3±3.2%, respectively (mean±SEM) (Fig. 8A). Suitably, starting from the 14 weeks, our analysis showed a signi cant AChRγ upregulation in the TB muscle of SOD1 G93A mice, an effect magni ed at 20 weeks corroborating the early and progressive NMJ alteration prior to the appearance of any sign of motor impairment. Notably, the MCP1 boosting remarkably prevented the NMJ denervation as demonstrated by the signi cant AChRγ downregulation compared with the scAAV9(empty)-treated mice at both time points (Fig. 8B).
As for the TA muscle, a single scAAV9_MCP1 i.m. injection resulted in a long-lasting Mcp1 upregulation compared with the control groups and with the same extent at both the considered time points (Fig. 8C). Suitably, macrophages recruitment dramatically increased in the TB muscle of scAAV9_MCP1-treated mice compared with the control groups at 14 and 20 weeks (Fig. 8D-F). Besides, MCP1 boosting fostered the in ltration of cytotoxic CD8 + T cells, but not CD4 + T lymphocytes and FoxP3 + Tregs, only at 14 weeks, suggesting an early in ammatory response within the forepaw muscle. Indeed, in the scAAV9(empty)treated mice, a slightly heightened of T cells and Tregs in ltration was recorded only at the symptomatic disease stage (Fig. 8G-I).
This evidence was corroborated by the Tnfα upregulation recorded in the TB muscle of 14 weeks-old scAAV9_MCP1-treated mice compared with controls (Fig. 9A), and by the unchanged expression of Arg1 between the two groups of SOD1 G93A mice (Fig. 9C, D), suggesting a massive in ltration of M1-polarised leucocytes six weeks after the scAAV9_MCP1 injection. Notably, our analysis revealed a signi cant increase of Igf1 transcript in the TB muscle of scAAV9_MCP1-treated mice (Fig. 9B), indicating the ongoing switching of in ltrated M1 cells towards the M2 pro-healing phenotype [81]. Accordingly, at 20 weeks, the Tnfα downregulation (Fig. 9A), the heightened Arg1 expression and the decreased Igf1 transcription compared with controls ( Fig. 9B-D) suggested the establishment of an anti-in ammatory muscular milieu twelve weeks after the scAAV9_MCP1 injection.
The histological analysis of transverse TB muscle sections showed that the MCP1-mediated immune cells in ltration did not signi cantly modify the quiescent status of the SCs at the pre-symptomatic disease stage (Supplementary Fig. 7A, C). Accordingly, no signi cant difference in the percentage of centralised myonuclei was recorded between the two groups of SOD1 G93A mice at 14 weeks, albeit an increasing trend was noticeable upon MCP1 boosting ( Supplementary Fig. 7B, D). Conversely, at 20 weeks, the switch of the recruited leucocytes towards the M2 pro-healing phenotype promoted the TB regeneration in scAAV9_MCP1-treated mice, as demonstrated by the increased percentage of The data collected demonstrated the bene cial action of MCP1 within the motor unit of C57SOD1 G93A mice. Therefore, we assessed whether the chemokine induction in 129SvSOD1 G93A mice, which show a faint activation of the MCP1 axis, was able to ameliorate the disease progression.
Eight-week-old 129SvSOD1 G93A mice (No.12 per group) were i.m. injected with the scAAV9_MCP1 and monitored until the clear symptomatic disease stage (17 weeks). The behavioural analysis showed that neither the bodyweight nor the motor onset was modi ed by the chemokine boosting (Empty, 14.1±0.4 weeks; MCP1, 13.4±0.4 weeks. mean±SEM) (Fig. 10 A, C). Nevertheless, a worsening of the grip strength impairment was recorded in the scAAV9_MCP1-treated mice at the advanced disease stages (Fig. 10B). However, the histological examination did not reveal any difference between the two groups of 129SvSOD1 G93A mice in the extent of NMJ denervation and TA muscle atrophy (Fig. 10D-F).
Compared to C57SOD1 G93A mice, at the symptomatic disease stage, fast-progressing mSOD1 mice strongly upregulated the chemokine within the TA muscle compared to respective Ntg littermates (FC = 65.1±12.4, 129SvSOD1 G93A mice; FC = 3±0.2, C57SOD1 G93A mice. mean±SEM), which dramatically increased upon scAAV9_MCP1 injection (FC = 752.1±208.4. mean±SEM) (Fig. 10G). Interestingly, the analysis of CD68 + cells density and CD8a and CD4 transcripts demonstrated massive recruitment of macrophages and T lymphocytes in the hind paw muscle of scAAV9_MCP1-but not scAAV9(empty)treated group compared with the NTg littermates (Fig. 10H-K), suggesting a drastic alteration of the muscular in ammatory response in 129SvSOD1 G93A mice upon chemokine boosting. Notwithstanding the signi cant Foxp3 upregulation (Fig. 10L), an intense in ammation characterised the TA muscle of scAAV9_MCP1-treated mice, arguably due to the impaired immunomodulatory capability of Tregs of mSOD1 mice at the advanced disease stage [106]. Accordingly, the pro-in ammatory markers Tnfα and gp91 PHOX were signi cantly upregulated in scAAV9_MCP1-treated mice ( Supplementary Fig. 8A, C, D).
Conversely, the expression level of the anti-in ammatory factor Arg1 was unchanged between the two groups of 129SvSOD1 G93A mice, whereas the Igf1 was downregulated, indicating that the in ltrated M1macrophages ( Supplementary Fig. 8A, B) were not apt to switch towards the M2 pro-healing phenotype [81] (Supplementary Fig. 8E). Besides, the MCP1 boosting was unable to upregulate further the Pax7 and MyoG expression in 129SvSOD1 G93A mice ( Supplementary Fig. 8A, F, G), suggesting a worthless myogenic response by mSOD1 mice at the advanced disease stage.
Since in C57SOD1 G93A scAAV9_MCP1-treated mice the chemokine-mediated protective role arose early in the disease course, we assessed the effect of MCP1 boosting in fast-progressing mSOD1 mice ~2 weeks before the motor symptoms appearance. Six-week-old 129SvSOD1 G93A mice (5 per group) were i.m.
injected with the scAAV9 vectors, and the analysis of TA muscle was performed at 12 weeks of age.
Intriguingly, although Mcp1 resulted markedly upregulated ( Supplementary Fig. 9C), 129SvSOD1 G93A mice appeared insensitive to the chemokine boosting as demonstrated by the unchanged CD68 + macrophages recruitment compared with the scAAV9(empty) group ( Supplementary Fig. 9A, B). Indeed, no difference in the extent of the TA muscle atrophy was recorded between the two groups of 129SvSOD1 G93A mice at 12 weeks ( Supplementary Fig. 9D).
Altogether, the data collected indicate a laggard activation of the muscular immune response by fastprogressing SOD1 G93A mice, culminating in an exacerbated in ammation upon MCP1 boosting that might be responsible for the worsened clinical phenotype at the advanced disease stage.

Discussion
In this study, we examined the involvement of the MCP1-mediated axis in governing the speed of ALS progression in two SOD1 G93A models characterised by remarkable differences in the disease progression rate.
Our observations revealed that, albeit the scAAV9_MCP1 i.m. injection boosted the chemokine to the same extent along the neuromuscular system of the two ALS models, the treatment led to an opposite effect on the clinical phenotype of C57 compared with 129Sv mSOD1 mice. Slow-progressing C57SOD1 G93A mice responded positively to MCP1 boosting, anticipating the recruitment and phenotypic switch of leucocytes within the peripheral compartment. This sustained the activation of the myogenic programme and nerve regeneration, nally slackening off the motor symptoms. Conversely, fast-progressing 129SvSOD1 G93A mice exhibited an adverse response to the treatment, exacerbating the toxic in ammatory response in the periphery, resulting in worsened motor ability late in the disease. Intriguingly, our data showed a novel immune-unrelated role for MCP1 in promoting motor axon regeneration and modulating neuroin ammation in the nervous system of mSOD1 mice, with the overall effect of slackening MN degeneration.
We recently reported a different activation of MCP1 within MN soma and peripheral compartment of fastversus slow-progressing SOD1 G93A models [26,28]. Our studies revealed that fast-progressing mSOD1 mice exhibited earlier muscle denervation and motor axon deterioration correlated with lower immune cells in ltration in the peripheral compartment than slow-progressing mSOD1 mice [26,27]. We speculated that this defective immune response underpinned the greater peripheral degeneration and more rapid disease course of 129SvSOD1 G93A mice. This evidence put the MCP1-mediated immune cell recruitment forward as a discriminating factor of the different speed in the disease progression of the two mSOD1 models.
MCP1 is a chemokine with a renowned pro-in ammatory capability [111]. In the neurological context, the increased expression of MCP1 is usually associated with neurodegenerative/neuroin ammatory diseases [112][113][114], including ALS [115][116][117]. Accordingly, in the spinal cord of SOD1 G93A mice, we recorded a gradual increase of MCP1 levels as the disease progresses, characterised by a strong expression by microglia at the advanced disease stage.
Besides its classic toxic in ammatory activity, several evidence indicated a pivotal role of the MCP1mediated axis at orchestrating nerve [118][119][120][121] and muscle [29,49,79,122] regeneration. In keeping with this, we recorded a gradual increase of chemokine expression along motor axons and Schwann cells as the disease progresses, suggesting the protective role of MCP1 in the PNS of mSOD1 mice.
Immune cells in ltration has been reported within nerves and skeletal muscles in ALS [16,54,55,123,124], albeit its contribution to the disease progression is still elusive. Here we assessed the in uence of peripheral immune response in fast-and slow-progressing SOD1 G93A mice through the i.m. injection of scAAV9_MCP1, which neatly boosted the chemokine along the motor unit of both ALS models.
The data herein collected highlighted a delayed activation of the immune response in the muscular compartment of mSOD1 mice. Indeed, at 14 weeks (i.e. ~2 weeks before the overt muscle strength impairment), in concomitance with a pronounced TA muscle denervation atrophy, C57SOD1 G93A mice have just launched the in ammatory response activating resident macrophages and recruiting neutrophils, which are the rst immune cells entering within the damage site to amplify the in ammation and promote the recruitment of haematogenous leucocytes [89,90]. Conversely, the early MCP1 boosting anticipated the in ammatory response within the muscular compartment of slow-progressing mSOD1 mice. Indeed, a "second wave" of immune cell in ltration, in which neutrophils gave way to macrophages and T lymphocytes [53,85,86], characterised the TA muscle of C57SOD1 G93A mice six weeks after the scAAV9_MCP1 injection. Intriguingly, our analysis revealed that the immunosuppressive capability of MCP1-recruited Tregs [125] dampened the in ammation within the damaged tissue, sustaining the switch of M1 phagocytic macrophages towards the M2 pro-regenerative phenotype [88,126,127]. In keeping with the Tnfα and Igf1 downregulation and increased Sirt1 expression [81,82], a higher percentage of CD206 + M2-macrophages was recorded in the TA muscle of scAAV9_MCP1-treated mice compared with the scAAV9(empty) group, which got stuck on the " rst wave" of the immune response [53,85,86].
The anticipated peripheral immune response of scAAV9_MCP1-treated C57SOD1 G93A mice re ected into increased myogenic activity by virtue of the lack of the inhibitory action of neutrophils [128] and inductive action of the Tregs and M2 macrophages on myogenic progenitor cells [85,87,91,129]. This translated into TA muscle regeneration and preservation from denervation atrophy and metabolic dysregulation compared with the control group.
The same extent of TA muscle atrophy registered in scAAV9_MCP1-and scAAV9(empty)-treated mice at 20 weeks indicated gradual exhaustion of the elicited muscle pro-healing immune response in the hindlimbs of C57SOD1 G93A mice as the disease progresses. However, the preservation of the forepaw muscles, which are belatedly affected in the mSOD1 model [71,109,110], might be partially responsible for the ameliorated motor performance of slow-progressing mSOD1 scAAV9_MCP1-treated mice at the advanced disease stage [57]. The early MCP1 boosting within TB muscle of C57SOD1 G93A mice forced and sustained the pro-in ammatory response at 14 weeks, which was decisive at countenancing its transition towards the anti-in ammatory and pro-regenerative state [50,130], eventually preventing the forepaw muscle denervation atrophy at the symptomatic disease stage.
The evidence herein collected highlighted the pivotal role of the peripheral immune response in triggering skeletal muscle regeneration and its temporal activation as a limiting factor in achieving a signi cant effect to slacken off the disease progression in mSOD1 mice.
The delayed activation of the peripheral immune response was exacerbated in fast-progressing SOD1 G93A mice, whose genetic background is associated with a poor ability at recruiting immune cells during phlogosis [131,132]. Accordingly, although the chemokine resulted massively upregulated six weeks after the scAAV9_MCP1 injection, 129SvSOD1 G93A mice were unable to promptly and properly react to the chemotactic gradient established within the TA muscle fostering the haematogenous macrophage recruitment. The de cient activation of the immune response early in the disease led to its mismanagement at the advanced stage. Indeed, pursuant to the dramatic macrophages and T lymphocytes recruitment, a massive and persistent in ammation characterised the TA muscle of 129SvSOD1 G93A scAAV9_MCP1-treated mice at 17 weeks, nally hampering skeletal muscle regeneration and function [53,130]. This result mirrored the recent ndings by Rizzo et al. [133], who demonstrated that, in splenectomised mdx mice, the delayed macrophage in ltration impaired their shift towards the M2 pro-healing ngerprint, eventually hindering muscle bres regeneration. Therefore, we can surmise that the tardive and maladaptive activation of the peripheral immune response, even upon MCP1 boosting, might be the chief culprit of faster disease progression of 129SvSOD1 G93A mice.
The data collected in the hind paw muscle of mSOD1 mice suggested that preserving the muscular compartment since the early disease stage might have slackened off ALS dying-back degeneration of the motor system [11]. Nevertheless, we demonstrated that the retrograde overexpression of MCP1 within the sciatic nerves of mSOD1 mice directly affected the stability and regeneration of peripheral motor axons.
Indeed, the chemokine upregulation within the sciatic nerves of C57SOD1 G93A scAAV9_MCP1-treated mice resulted in increased sprouting of GAP43 + motor axon terminal branches, accounting for the reduced denervation atrophy of hindlimb muscles across the disease progression.
The MCP1 dispensable capability to promote axonal outgrowth was previously described in aSMN1expressing NSC34 cells cultures [44] and DRG explants obtained from MCP1-treated [120,134] or genetically depleted mice [119]. According to this information, the chemokine plays an immune-unrelated role in amplifying and maintaining the regenerative capacity of peripheral axons, promoting the expression of the regeneration-associated genes (e.g. GAP43), which is concurrent to the MCP1-mediated neuron-macrophage interaction [45,119,134].
Based on our evidence, the MCP1 pro-regenerative effect in the PNS of mSOD1 mice became clear at the symptomatic disease stage when the chemokine overexpression preserved the motor axon cytoarchitecture and myelination, decreased toxic in ammation and sustained the collateral innervation of hindlimb skeletal muscles.
Albeit the MCP1 pleiotropic mechanism is far from being elucidated, recent studies reported the direct in uence of the chemokine at modulating the neuroin ammation by governing the recruited myeloid cells ngerprint. For instance, several observations obtained in animal models of spinal cord injury demonstrated that MCP1 released by neurons attracts and activates macrophages through CCR2 to drive them toward the M2 pro-healing phenotype. In turn, MCP1-activated macrophages establish a permissive environment [119,134], eventually preventing neurodegeneration [45]. Here we showed that a similar mechanism could be prompted within the spinal cord of SOD1 G93A mice where MCP1 resulted signi cantly upregulated by MNs at the disease onset. Albeit it is now clear that haematogenous monocytes cannot penetrate the CNS of mSOD1 mice [16,135,136], we can suppose that the scAAV9speci c induction of MCP1 within MN perikaryon might have modulated the activation state of the CNSresident myeloid cells (i.e. microglia). Suitably, our data demonstrated that the scAAV9_MCP1 injection in pre-symptomatic mSOD1 mice extended the so-called "stable phase" of the disease [105,106], maintaining the M2 polarisation of the neuroin ammatory milieu and reducing MN loss along the disease course.

Conclusions
For the rst time in the ALS context, we demonstrated the pivotal role of the immune response in promoting and governing skeletal muscle regeneration and thus the speed of the disease progression. Our observations suggest that, although potentially protective, the immune response is delayed in ALS mice and, hence, ineffective at sustaining a substantial recovery of the peripheral compartment. Notably, the dichotomic effect recorded in the two SOD1 G93A strains following MCP1 boosting pointed out the nature and temporal activation of the immune response as discriminating factors to foster skeletal muscle regeneration, slacken the dying-back degeneration and slow down ALS course. This also emphasizes the different immune response due to genetic background as a key determinant of the variability in the disease progression as reported in ALS patients carrying the same SOD1 mutation [3,4]. Altogether, these observations nominate the muscular compartment as a primary target for developing effective therapeutic interventions in ALS capable of interfering with the speed of the symptoms progression and the dying-back degeneration tangibly. Besides, the comprehension of the mechanisms underlying the protective role ful l from MCP1 in the motor unit of mSOD1 mice might provide innovative evidence regarding the contribution of the immune response in ALS.
Despite in vitro and in vivo models of the disease have generated different potential pharmacological targets, ALS still lacks an adequate therapy able to delay or even halt its development [137,138]. We think that this is mainly due to the poor knowledge of the temporal and spatial mechanisms by which the immune response governs the pattern of the disease [139,140]. Our ndings provide a possible explanation for the failure of unspeci c immunomodulatory treatments [23,141] and suggest new potential strategies to prevent ALS progression.
Altogether, the evidence herein provided demonstrated the crucial role of the so-far-underestimated peripheral compartment in ALS pathoprogression straightforwardly. Although the latest clinical studies reported a defective monocyte/macrophages in ltration at the site of nerve degeneration [142] and a direct correlation between the PNS in ammation and longer disease duration [21], no observations are still available on the rst body compartment affected by ALS: the skeletal muscle. Therefore, in virtue of the easy accessibility of bioptic samples, the characterisation of the immune muscle ngerprint to assess a potential correlation with biomolecular pathways underlying atrophy and myogenesis might produce a combination of muscle-derived, immune-related molecular signatures that will be useful as a clinical adjunct in the prognostic evaluation of ALS patients.

Availability of data and materials
The authors con rm that the data, materials, and software information supporting the ndings of this study are available within the article and its supplementary materials.         The scAAV9_MCP1 injection reduces the forelimb muscles degeneration promoting leucocytes recruitment in C57SOD1G93A mice. A Muscle wasting was calculated by measuring the TB muscle weight of scAAV9_MCP1-and scAAV9(empty)-treated mice compared to relative Ntg littermates at each time points analysed. The percentage of muscle atrophy was calculated relative to Ntg mice. Data are reported as mean±SEM. 14 weeks: n=8 per group; 20 weeks: n=10 per group. B, C The real-time PCR The MCP1-mediated immune response triggers the myogenic programme in the forelimb muscles of

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.