Serotonin reuptake inhibitors improve muscle stem cell function and muscle regeneration in male mice

Serotonin reuptake inhibitor antidepressants such as fluoxetine are widely used to treat mood disorders. The mechanisms of action include an increase in extracellular level of serotonin, neurogenesis, and growth of vessels in the brain. We investigated whether fluoxetine could have broader peripheral regenerative properties. Following prolonged administration of fluoxetine in male mice, we showed that fluoxetine increases the number of muscle stem cells and muscle angiogenesis, associated with positive changes in skeletal muscle function. Fluoxetine also improved skeletal muscle regeneration after single and multiples injuries with an increased muscle stem cells pool and vessel density associated with reduced fibrotic lesions and inflammation. Mice devoid of peripheral serotonin treated with fluoxetine did not exhibit beneficial effects during muscle regeneration. Specifically, pharmacological, and genetic inactivation of the 5-HT1B subtype serotonin receptor also abolished the enhanced regenerative process induced by fluoxetine. We highlight here a regenerative property of serotonin on skeletal muscle.

Reviewer #1 (Remarks to the Author): In this manuscript the authors show that prolonged administration of the serotonin reuptake inhibitor, fluoxetine (FLX) in mice, increases the number of muscle stem cells and intramuscular angiogenesis, leading to improved regeneration as well as reduced fibrosis and inflammation.These effects were not observed in mice devoid of peripheral serotonin and upon pharmacological or genetic inactivation of the 5-HT1B subtype serotonin receptor While the primary observation that FLX promotes muscle progenitor cell expansion is of great interest in principle, the study is descriptive and mostly focused on cell counts, and mechanistic basis of this mitogenic effect remains unclear, thereby reducing somehow the interest of the finding.This reviewer recommends addressing the following points 1) Is this a cell autonomous effect or it is related to interactions with other FLX-responsive cell types -i.e.endothelial or vessel associated cells.This should be tested experimentally by alternative knocking down 5-HT1B specifically in MuSCs vs endothelial cells, using appropriate Crelox mice models.
2) Although the detailed investigation of the signal transduction and gene expression profile downstream to serotonin receptor might belong to an independent next investigation, it would be important to link the biological effect reported here to specific signaling cascade(s) and gene expression patterns in MuSCs and endothelial cells 3) Immunofluorescence analysis of myogenin/Pax7 in the same muscle stem cellswithin single fibers is recommended 4) The authors observed the same effect in both cultured C2C12 myoblasts and MuSCs, suggesting that the activation of serotonin receptor promotes an equivalent effect (i.e.mitosis) in committed myoblasts as well as muscle stem cells.Analysis of the physiological role of serotonin receptor in these two different cellular states is recommended.5) Is the mitogenic effect promoting asymetric cell division in MuSC? 6) Is there any physiological action of endogenous 5-HT1B?KO mouse model?Reviewer #2 (Remarks to the Author): This study reports that a chronic treatment with fluoxetine increases the proliferation of skeletal muscle progenitor cells and boosts angiogenesis in muscles.It then uses an injury model to evaluate whether this treatment can enhance muscle regeneration.Using cell culture assays and transgenic mice it also shows that this effect is dependent on serotonin signaling through the 5Ht1b receptor.
Given the broad, and ever increasing, use of selective serotonin reuptake inhibitors this study is both timely and potentially important.The results presented are intriguing and novel.The authors need, however, to address some important technical and conceptual points.
A major concern is that the study relies only on cellular data.Metabolic and exercise testing should be performed to determine whether the changes observed translate into functional changes.This is both important for the un-injured and regeneration studies.
Another issue with the regeneration experiments is that the mice pre-treated for 6 weeks with fluoxetine start with a higher number of progenitors and therefore are expected to be regenerating faster than untreated mice.It is unclear whether the regeneration itself is enhanced by the fluoxetine treatment or that the treated mice have such a head start that they appear to regenerate faster.Control groups of un-injured mice (treated and not) as well as mice for which the treatment was terminated at the time of injury should be analyzed.
The in vitro data with C2C12 cells are confusing.If the effect of fluoxetine is mediated by its binding to a serotonin receptor, the additive effect observed with fluoxetine and serotonin treatment is hard to explain since myoblasts do not synthesize serotonin; blocking SERT activity should not lead to increase extracellular serotonin levels as it happens in vivo.These experiments should be repeated with primary cell-sorted cells from the Tg:Pax7nGFP mice.

Additional points:
-The title should specify that the study was performed in mice.
-In Table 1, the SD for some important biomarkers such as MCP1, IL-6, TNFa, VEGF is very high.Given the low number of mice analyzed (n=4) this may have biased the statistics.Additional animals needs to be analyzed to validate these results.
-Likewise, some of the panels shown in several other Figures (2, supp.2, 3G, supp 4, sup 5) present data with large SD and low number of mice.Additional animals should be analyzed to strengthen the data.
-Serotonin is not a stable molecule.The statement that FBS comprises endogenous 5-HT should be sustained by experimental data.-As stated above, the effect shown with fluoxetine alone in sup. Figure 3A does not fit with a SERT/serotonin-mediated action of fluoxetine.These data should be experimentally strengthened, for example by using primary cells.
-In all experiments using GR127935 or other compounds, the data for these compounds alone should be shown.

Replies to the Reviewers' comments:
Please find the answers and corrections in blue below.
Reviewer #1 (Remarks to the Author): In this manuscript the authors show that prolonged administration of the serotonin reuptake inhibitor, fluoxetine (FLX) in mice, increases the number of muscle stem cells and intramuscular angiogenesis, leading to improved regeneration as well as reduced fibrosis and inflammation.These effects were not observed in mice devoid of peripheral serotonin and upon pharmacological or genetic inactivation of the 5-HT1B subtype serotonin receptor.
While the primary observation that FLX promotes muscle progenitor cell expansion is of great interest in principle, the study is descriptive and mostly focused on cell counts, and mechanistic basis of this mitogenic effect remains unclear, thereby reducing somehow the interest of the finding.This reviewer recommends addressing the following points.
We thank the reviewer for her/his positive comments on our work and constructive criticism of the manuscript.
1) Is this a cell autonomous effect or it is related to interactions with other FLX-responsive cell typesi.e.endothelial or vessel associated cells.This should be tested experimentally by alternative knocking down 5-HT1B specifically in MuSCs vs endothelial cells, using appropriate Cre-lox mice models.
Our current results strongly support a cell autonomous effect of fluoxetine (FLX) on satellite cells (SCs) via the action of 5-HT/5-HT1B receptor.
First, we showed that skeletal muscle and specifically primary SCs and immortalized myoblasts (C2C12) exhibited serotonergic identity through the expression of key players of the 5-HT system (Fig. 3a-b, Supplementary Fig. 3b,3j).
Furthermore, we previously demonstrated in vitro that FLX, in a 5-HT-dependent manner and upon 5-HT1B receptor activation, promoted C2C12 cell division (Supplementary Fig. 3c-f).
To unambiguously confirm the cell autonomous effects of FLX on SCs, we complemented our in vitro experimental approach by characterizing the effects of FLX on primary SCs sorted by FACS from Tg:Pax7nGFP mice.Because FBS serum is essential for culturing primary SCs, all these experiments included culture medium containing a concentration of endogenous 5-HT, which we measured by HPLC (Supplementary Fig. 3a).
Thus, we showed that FLX exposure promoted SCs proliferation with a significant increase in the number of SCs at 2-, 4-and 14-days post plating (Fig. 3C).Specifically, at 4 days post plating, FLXexposed cells expressed increased numbers of Pax7+/GFP+ cells, a marker of early stages of myogenesis, as well as increased numbers of myogenin+ cells, a marker or early stages of myogenic differentiation (Fig3d-e).Finally, at 14 days post plating, FLX-exposed cells showed a significant increase in the fusion index, a marker of the terminal stage of differentiation, associated with an increased number of single cells expressing Pax7+/GFP+, corresponding to the reserve pool of SCs (Fig. 3f-g).The overall cellular effects induced by FLX were counteracted in the presence of a 5-HT1B antagonist (Fig. 3c-g).
In total, FLX exerts autonomous effects by targeting primary SCs that express the 5-HT1B receptor and whose activation promotes the different stages of myogenesis with their proliferation, differentiation and self-renewal.
As suggested by the reviewer, we previously confirmed these results in vivo using an appropriate Cre-Lox model allowing specific deletion of 5-HT1B receptor expression in SCs (Supplementary Fig. 4i).Indeed, the effects of FLX targeting SCs were abolished in the basal state or post injury, after suppression of specific 5-HT1B receptor expression by SCs, suggesting a vessel-independent effect of FLX (Fig. 4g, Supplementary Fig. 4j).Similarly, the positive effects of FLX on post injury collagen deposition, one of the markers of effective wound healing, were also abolished in the absence of 5-HT1B receptor expression by SCs (Fig. 4h).
Associated with the specific effects of FLX on SCs, we showed that there were also other targets of this molecule such as vessels or immune system actors (Fig. 1d-g, Supplementary Fig. 1h-j,2d-f) that actively participate in the beneficial effects of FLX during muscle regeneration.These results are consistent with the literature since the effects of 5-HT/FLX on vascular tissue and the immune system have already been widely described in other organs (see References 9-10, 12-13, 56-63).More specifically, the positive effects of FLX during muscle regeneration were abolished when a 5-HT1B antagonist was delivered systemically to the mice, highlighting the crucial role of this receptor regardless of the actors or their interactions.
The question of the cellular interaction between closed partners of muscle regeneration, such as vessels, macrophages, and other immune cells, but also fibro-adipogenic progenitors, pericytes etc., is quite relevant and will be the subject of a new study.Indeed, the novelty of the present work is carried by the demonstration of an autonomous effect of FLX on SCs.
2) Although the detailed investigation of the signal transduction and gene expression profile downstream to serotonin receptor might belong to an independent next investigation, it would be important to link the biological effect reported here to specific signaling cascade(s) and gene expression patterns in MuSCs and endothelial cells As suggested by the reviewer, we performed different approaches on primary SC and C2C12 in order to identify signaling pathways underlying the biological effects of FLX/5-HT on SCs.
By RT-qPCR, we showed that FACS-sorted SCs from FLX-treated mice overexpressed genes related to myogenesis such as Pax7 and Myogenin, as well as those related to cell division such as Cyclin D1 (Supplementary Fig. 1g).
Then, we investigated the intracellular levels of cAMP, as a previously known second messenger of the 5-HT1 receptor family.After a short exposure to 5-HT, we showed a clear decrease in the intracellular level of cAMP in C2C12 cultured in a 5-HT-free medium and this effect was abolished with 5-HT1B antagonist (Supplementary Fig. 3g).Similarly, a short exposure to FLX in a regular medium resulted in a decrease in intracellular cAMP levels of C2C12 (Supplementary Fig. 3h).
Finally, using a screening on phosphorylation changes of major protein kinases, we showed that a short exposure to 5-HT positively or negatively modulated, in a 5-HT1B receptor-dependent manner, the activity of certain kinases (Supplementary Fig. 3i).
We propose as a mechanistic hypothesis underlying the cellular effects of FLX that it, by blocking the SERT transporter, increases the bioavailability of extracellular 5-HT allowing the reinforcement of the signal induced by the activation of the 5-HT1B receptor leading to the decrease of intracellular levels of cAMP, following the inhibition of adenylate cyclase, and to the positive or negative modulation of different key factors in the function of SCs such as ERK2 and Cyclin D1, which are well known to be involved in the proliferation process.
Furthermore, signaling pathways related to 5-HT1B receptor activation and involved in angiogenesis have already been described in the literature (References 58-59).Specifically, the intracellular mediators of these effects are the production of NO and the activation of the Src/PI3K/Akt/mTOR/ER pathway.
Further investigations would clarify all the transduction pathways underlying the effects of serotonin, whether dependent on the 5-HT1B receptor, on muscle stem cell function.
3) Immunofluorescence analysis of myogenin/Pax7 in the same muscle stem cellswithin single fibers is recommended In our initial study, we analyzed Pax7 and Myogenin immunostaining during muscle regeneration and more precisely 4 days post injury (Fig. 2b-d).At these early times of muscle regeneration, the muscle being too damaged, the isolation of single fibers is compromised.However, as suggested by the reviewer, using isolated fibers from uninjured muscles, we studied ex vivo the differentiation status of SCs exposed for 4 days to 5-HT as shown in Fig. 3i.We showed that 5-HT exposure promoted myogenic differentiation with an increase in the percentage of myogenin+ cells associated with a decrease in the percentage of Pax7+/GFP+ cells (Fig. 3h).
Furthermore, as presented above, we confirmed in vitro the effects of FLX on early stages of myogenesis with an increased level of Pax7+/GFP+ and Myogenin+ cells for SCs exposed to FLX (Fig. 3d-e).
4) The authors observed the same effect in both cultured C2C12 myoblasts and MuSCs, suggesting that the activation of serotonin receptor promotes an equivalent effect (i.e.mitosis) in committed myoblasts as well as muscle stem cells.Analysis of the physiological role of serotonin receptor in these two different cellular states is recommended.
We have previously shown in vitro that exposure to 5-HT and FLX, in a 5-HT-dependent manner, was associated with an increase in the rate of C2C12 cell division, suggesting a pro-proliferative effect of 5-HT/FLX (Supplementary Fig. 3c-d).These effects were abolished in the presence of 2 different 5-HT1B antagonists without any effect of 5-HT2A antagonist as a control, suggesting a crucial role of the 5-HT1B receptor in the proliferative effects of 5-HT/FLX (Supplementary Fig. 3e-f).In addition, 5-HT and 5-HT1B agonists also promoted terminal differentiation of myogenesis marked by an increase in the fusion index (former Figure S3D).
It should be first noted that the overall effects induced by FLX in a 5-HT/5-HT1B receptor-dependent manner, are not crucial for the physiological behavior of SCs as evidenced by the lack of abnormalities in muscle regeneration of basal SCs pool in TPH1 -/-or Pax7-Cre ER(T2) ::tetO1B mice compared to wildtype mice (Fig. 4a-b,g, Supplementary Fig. 4a,j).
Even if 5-HT/5-HT1B receptor stimulation is not mandatory for muscle, a pro-proliferative effect on SCs would be of major interest to improve muscle function in normal or pathological conditions.
As suggested by the reviewer and as mentioned previously, we deeply investigated the effects of FLX on FACS-sorted primary SCs cultures of 7 Tg:Pax7nGFP mice in medium with FBS containing endogenous 5-HT.
Thus, we showed that FLX exposure promoted SCs proliferation with a significant increase in the number of SCs at 2-, 4-and 14-days post plating (Fig. 3C).Specifically, at 4 days post plating, FLXexposed cells expressed increased numbers of Pax7+/GFP+ cells, a marker of early stages of myogenesis, as well as increased numbers of myogenin+ cells, a marker or early stages of myogenic differentiation (Fig3d-e).Finally, at 14 days post plating, FLX-exposed cells showed a significant increase in the fusion index, a marker of the terminal stage of differentiation, associated with an increased number of single cells expressing Pax7+/GFP+, corresponding to the reserve pool of SCs (Fig. 3f-g).The overall cellular effects induced by FLX were counteracted in the presence of a 5-HT1B antagonist (Fig. 3c-g).
In total, FLX exerts autonomous effects by targeting primary SCs that express the 5-HT1B receptor and whose activation promotes the different stages of myogenesis with their proliferation, differentiation and self-renewal.
As requested by the reviewer, we confirmed these results by another in vitro experimental approach with the culture of SCs in a medium comprising serum from FLX-treated and control mice instead of the usual FBS serum.In these sera, we previously characterized serotonin levels by HPLC that were increased in FLX-treated mice compared to controls (Figure 1, below).By Luminex, we also showed that FLX-treated mice had increased circulating levels of the growth factor FGFb compared to controls (Table 1).Although we did not have the opportunity to measure the plasma level of FLX in these sera, as this molecule has a long half-life ranging from 1 to 4 days 1 , we can assume that serum from FLXtreated mice also contained the molecule.Thus, we showed by videomicroscopy that SCs cultured in serum from FLX-treated mice initiated their entry into the first cell division more rapidly and had a significantly higher rate of cell division than SCs cultured in serum from control mice (Fig. 3j-k).Interestingly, the cell division promoting effects of FLX on SCs were abolished in the presence of a 5-HT1B antagonist (Fig. 3j-k).In addition, SCs exposed to serum from FLX-treated mice for 4 days expressed an increased number of Myogenin+ cells and this FLX-induced myogenic differentiation effect was abolished by a 5-HT1B antagonist (Fig. 3l).At 14 days post plating, exposure of SCs to serum from FLX-treated mice resulted in an increase in the number of Pax7+ cells, a marker of the reserve SC pool, and this effect was counteracted by a 5-HT1B antagonist (Fig. 3m).
In total, the cellular effects induced by exposure to serum from FLX-treated mice were thus similar to those induced by exposure to fluoxetine alone and promoted, in a stimulation-dependent manner of the 5-HT1B receptor, activation with cell cycle entry, proliferation, early differentiation, and, finally, selfrenewal of SCs.In vivo, we also showed that the beneficial effects of FLX during muscle regeneration included increased SCd proliferation and differentiation and were dependent on 5-HT and 5-HT1B receptor as demonstrated using pharmacological and genetic models (Fig. 4a-b,d-e,g).

5
) Is the mitogenic effect promoting asymetric cell division in MuSC?Indeed, our current results do not specify whether FLX through 5-HT has an exclusively proliferative effect targeting myoblasts by promoting symmetric cell division, which leads to an increase in cell density, notably in vitro, secondarily stimulating differentiation or whether these molecules show a dual effect stimulating both asymmetric and symmetric cell divisions promoting proliferation and differentiation, respectively.
Our ex vivo experimental approach analyzing the distribution of Pax7/GFP+ relative to Myogenin+ cells can be considered as an indirect estimation of the cell division state.Indeed, although the regulation of the cell division state remains complex and partially unknown, it has been described in SCs that the symmetric division state could result in identical daughter cells (both Pax7+ or Myogenin+, respectively), whereas the asymmetric division state could result in one Pax7+ and one Myogenin+ daughter cell 2 .
Thus, we previously found that exposure to 5-HT for 4 days resulted in an increase in the distribution of Myogenin+ cells associated with a decrease in Pax7+/GFP+ cells among SCs from single fibers, suggesting a 5-HT effect promoting asymmetric division (Fig. 3h-i).
This question would naturally deserve further investigation in a future study.Moreover, this issue has already been studied in other stem cell cascades.For example, it was shown that 5-HT increased the rate of symmetrical division of neural progenitors without effect on neural stem cells and increased the proliferation and differentiation of myeloid progenitors without effect on hematopoietic stem cells (References 5, 36).
6) Is there any physiological action of endogenous 5-HT1B?KO mouse model?
In our study, using the conditional 5-HT1B receptor deletion mouse model specifically within SCs, we showed no abnormality of muscle regeneration, including SC behavior or collagen deposition, in tamoxifen-treated Pax7-Cre ER(T2) ::tetO1B mice compared to wild-type mice (Fig. 4g-h, Supplementary Fig. 4j).
To our knowledge, no studies have characterized the physiological muscle phenotype or during muscle regeneration of constitutive knockout of 5-HT1B receptor animal models.Furthermore, in studies using 5-HT1B KO mice, no pathological muscle phenotype was described, whereas other abnormalities were reported such as weight gain, increased bone formation, increased aggressive behavior, increased locomotor activity, decreased anxiety-related behavior, early age-related motor decline etc [3][4][5][6][7][8] .
Reviewer #2 (Remarks to the Author): This study reports that a chronic treatment with fluoxetine increases the proliferation of skeletal muscle progenitor cells and boosts angiogenesis in muscles.It then uses an injury model to evaluate whether this treatment can enhance muscle regeneration.Using cell culture assays and transgenic mice it also shows that this effect is dependent on serotonin signaling through the 5Ht1b receptor.
Given the broad, and ever increasing, use of selective serotonin reuptake inhibitors this study is both timely and potentially important.The results presented are intriguing and novel.The authors need, however, to address some important technical and conceptual points.
We thank the reviewer for her/his positive comments on our work and for constructive comments in order to improve the manuscript.
A major concern is that the study relies only on cellular data.Metabolic and exercise testing should be performed to determine whether the changes observed translate into functional changes.This is both important for the un-injured and regeneration studies.
As suggested by the reviewer, we performed a functional study on muscle strength and physical performance parameters in FLX-treated mice in uninjured and injured muscle conditions.
Thus, we showed by an in vivo grip test that uninjured FLX-treated mice treated exhibited an increase in forelimb muscle strength (Fig. 1h).In a treadmill exercise test, uninjured mice treated with FLX showed an improvement in performance with an increase in maximal aerobic velocity, distance traveled and exercise endurance (Fig. 1i, Supplementary Fig. 1k-l).
To further investigate the changes in muscle physiology induced by FLX treatment, we assessed muscle fiber typing.Indeed, FLX-treated mice showed a change in fiber typing on uninjured TA sections with a significant decrease in the percentage of fiber type IIX associated with a significant increase in the percentage of fiber I, IIA, IIB (Fig. 1j, Supplementary Fig. 1m).
In order to investigate the effect of FLX on muscle physiology during regeneration, a functional study on the in situ contractile parameters of the injured TA muscle was performed, using nerve-muscle stimulation to induce twitch and tetanos mechanical responses.At 14 days post notexine, we showed that the TA muscles from FLX-treated mice showed 22% and 27% increased relative amplitudes of twitch and tetanos contractile responses, respectively (Fig. 2i-j).
Taken together, we have shown that FLX positively modifies muscle physiology with an increase in muscle strength and physical performance during exercise.These improved functional parameters are supported by changes in muscle metabolism with a change in the typing of muscle fibers, being enriched in fibers I, IIA, IIB corresponding to the fibers known as slow twitch with high oxidative activity, to the fibers known as fast twitch with high oxidative activity and to the fibers known as fast twitch with high glycolytic activity, respectively 9 .In addition, we showed that the in situ contractile parameters of injured muscle were improved after treatment with FLX.These results are consistent with a recent study conducted by Tutakhail's team which also demonstrated an improvement in muscle performance after FLX treatment in mice (reference 66).
Another issue with the regeneration experiments is that the mice pre-treated for 6 weeks with fluoxetine start with a higher number of progenitors and therefore are expected to be regenerating faster than untreated mice.It is unclear whether the regeneration itself is enhanced by the fluoxetine treatment or that the treated mice have such a head start that they appear to regenerate faster.Control groups of un-injured mice (treated and not) as well as mice for which the treatment was terminated at the time of injury should be analyzed.
As requested by the reviewer, we performed histological analysis at different time points of muscle regeneration in mice pre-treated with FLX at 18mg/kg/day for 6 weeks, mice treated with FLX only post notexine (NTX) injury and control mice (Figure 2a)." ()%$%'# "" ()%$%&# """()%$%%&# """"()%$%%%&$ At 4 days post injury, as expected, we observed that mice pre-treated with FLX showed an increase in SCs as well as Myogenin + differentiating cells (Figures 2b-c).Mice treated with FLX post injury showed an equal increase in the number of SCs but had a similar number of Myogenin+ cells compared to the control mice (Figures 2b-c).
At 14 days post injury, both FLX-pretreated and FLX-post-NTX-treated mice had increased numbers of muscle fibers and centro-nucleated fibers (Figures 2d-e), as well as decreased immune cell infiltration (Figures 2f-g).However, in contrast to the FLX-pretreated mice, the mice treated by FLX post NTX had similar calcium and collagen deposition as the control group (Figures 2h-j).
As previously shown, pre-treatment with FLX for 6 weeks prior to muscle injury supports a harmonious acceleration of muscle regeneration through stimulating muscle stem cells and myogenesis, modulating inflammation and resulting in efficient wound healing marked by reduced collagen deposition.
When FLX is delivered only post injury, it also exerts early positive effects targeting the pool of activated SCs and myoblasts.These results are consistent with the cell-autonomous effects of FLX demonstrated in vivo in the uninjured muscle (Supplementary Fig. 1e-f) and in vitro via faster entry into cell division, increased rate of cell division resulting in increased proliferation of SCs (Fig. 3c-d,j-k).Thus, these results suggest that FLX can rapidly induce a cell-autonomous effect stimulating the proliferation of already activated SCs post injury, unlike quiescent SCs in uninjured muscle, requiring prolonged exposure to fluoxetine (Supplementary Fig. 1e-f).
In contrast, a short post injury administration of FLX does not result in an effect targeting the early stages of myogenic differentiation, in contrast to our previous in vivo and in vitro results (Fig. 2d, Fig. 3e,l).Interestingly, despite the lack of effect promoting early myogenic differentiation, FLX delivered post injury also results in an increased number of fibers including regenerated ones, suggesting a similar acceleration of muscle regeneration as in case of 6 weeks FLX pretreatment.However, the muscle regeneration induced by FLX delivered post injury is less harmonious and efficient as shown by the similar or even increased number of collagen and calcium deposits compared to control mice, although the infiltration of immune cells is simultaneously decreased.
Taken together, these results suggest that fluoxetine requires prolonged pre-injury administration to any muscle injury, suggesting indispensable upstream tissue and cellular remodeling to support its full triple beneficial action on muscle regeneration.Interestingly, fluoxetine delivered post-injury may induce beneficial effects on muscle regeneration, specifically targeting SCs early.
The in vitro data with C2C12 cells are confusing.If the effect of fluoxetine is mediated by its binding to a serotonin receptor, the additive effect observed with fluoxetine and serotonin treatment is hard to explain since myoblasts do not synthesize serotonin; blocking SERT activity should not lead to increase extracellular serotonin levels as it happens in vivo.These experiments should be repeated with primary cell-sorted cells from the Tg:Pax7nGFP mice.
As suggested by the reviewer, we have clarified the mechanisms of action underlying the in vitro effects of FLX on SCs and C2C12.Indeed, our mechanistic hypothesis is that FLX, by blocking the SERT transporter known to be its main target of action, increases the bioavailability of extracellular 5-HT, leading to a prolonged stimulation of the 5-HT1B receptor, mediator of the cellular effects of FLX.
First, we showed that skeletal muscle and specifically primary SCs as well as immortalized myoblasts (C2C12) exhibited serotonergic identity through the expression of key players of the 5-HT system, including the SERT transporter, the main target of SSRIs (Fig. 3a-b, Supplementary Fig. 3b,3j).
Next, we previously showed in vitro that 5-HT exerted a proliferative effect on C2C12s marked by an increase in division rate in a dose-dependent manner (Supplementary Fig. 3d).Interestingly, when C2C12s were exposed to FLX in a medium devoid of endogenous 5-HT, no effect on the rate of cell division was observed, unlike C2C12s exposed to FLX + exogenous 5-HT, which resulted in an increase in the rate of cell division (Supplementary Fig. 3c).This proliferative effect was also observed when C2C12s were exposed to FLX in a culture medium comprising FBS with endogenous 5-HT (Supplementary Fig. 3a,f).These results thus highlight that 5-HT is crucial in mediating the cellular effects of FLX on C2C12s.They are further consistent with in vivo results showing abolished effects of FLX on SCs during muscle regeneration in TPH1 -/-mice lacking peripheral 5-HT (Fig. 4a-b).
As suggested by the reviewer and as mentioned previously, we deeply investigated the effects of FLX on FACS-sorted primary SCs cultures of 7 Tg:Pax7nGFP mice in medium with FBS containing endogenous 5-HT.
Thus, we showed that FLX exposure promoted SCs proliferation with a significant increase in the number of SCs at 2-, 4-and 14-days post plating (Fig. 3C).Specifically, at 4 days post plating, FLXexposed cells expressed increased numbers of Pax7+/GFP+ cells, a marker of early stages of myogenesis, as well as increased numbers of myogenin+ cells, a marker or early stages of myogenic differentiation (Fig3d-e).Finally, at 14 days post plating, FLX-exposed cells showed a significant increase in the fusion index, a marker of the terminal stage of differentiation, associated with an increased number of single cells expressing Pax7+/GFP+, corresponding to the reserve pool of SCs (Fig. 3f-g).The overall cellular effects induced by FLX were counteracted in the presence of a 5-HT1B antagonist (Fig. 3c-g).
As requested by the reviewer, we confirmed these results by another in vitro experimental approach with the culture of SCs in a medium comprising serum from FLX-treated and control mice instead of the usual FBS serum.Thus, we showed by videomicroscopy that SCs cultured in serum from FLX-treated mice initiated their entry into the first cell division more rapidly and had a significantly higher rate of cell division than SCs cultured in serum from control mice (Fig. 3j-k).Interestingly, the cell division promoting effects of FLX on SCs were abolished in the presence of a 5-HT1B antagonist (Fig. 3j-k).In addition, SCs exposed to serum from FLX-treated mice for 4 days expressed an increased number of Myogenin+ cells and this FLX-induced myogenic differentiation effect was abolished by a 5-HT1B antagonist (Fig. 3l).At 14 days post plating, exposure of SCs to serum from FLX-treated mice resulted in an increase in the number of Pax7+ cells, a marker of the reserve SC pool, and this effect was counteracted by a 5-HT1B antagonist (Fig. 3m).
Taken together, FLX exerts autonomous effects by targeting primary SCs that express the SERT transporter and the 5-HT1B receptor by promoting different stages of myogenesis, including their activation, proliferation, differentiation and self-renewal, in a 5-HT and 5-HT1B receptor dependent manner.

Additional points:
-The title should specify that the study was performed in mice.
As suggested by the reviewer, we have corrected the title to " Serotonin reuptake inhibitors improve muscle stem cell function and muscle regeneration in mice ".
-In Table 1, the SD for some important biomarkers such as MCP1, IL-6, TNFa, VEGF is very high.
Given the low number of mice analyzed (n=4) this may have biased the statistics.Additional animals needs to be analyzed to validate these results.
As requested by the reviewer, we wished to replicate the Luminex experiment using TA muscle homogenates under the different conditions presented.Unfortunately, due to the high sensitivity of the Luminex technique between the different batches of kits, some of our results were not consistent with the first series, despite our increase of the n value per condition.Thus, due to the lack of robustness of these results, we removed them from our study.
However, we demonstrated by Luminex on a sufficient n value (9 animals per condition) with low SD, the modifications induced by a prolonged treatment with FLX on the serum levels of cytokines, chemokines and growth factors.Thus, we showed that FLX-treated mice had increased circulating levels of the growth factor FGF compared to controls (Table 1).
-Likewise, some of the panels shown in several other Figures (2, supp.2, 3G, supp 4, sup 5) present data with large SD and low number of mice.Additional animals should be analyzed to strengthen the data.
As suggested by the reviewer, we increased the number of animals for the muscle regeneration experiments induced by intramuscular injection of notexine (Fig. 2b,d,g-h, Supplementary Fig. 2a,d-f).
Of note, because the results are from nonparametric statistical analysis, they are represented by median with interquartile range and not by mean with standard deviation.
For Figure 3G, we have strengthened and deepened our characterization of the effects of fluoxetine on MuSCs in vitro and replaced with the new Fig.3c-g, representing n=7 mice/ condition.
The experiments in Supplementary Fig. 4c-h and Supplementary Fig. 5a-g could be replicated and completed with the antagonist control condition alone.
-Serotonin is not a stable molecule.The statement that FBS comprises endogenous 5-HT should be sustained by experimental data.
As suggested by the reviewer, we measured by HPLC endogenous 5-HT concentration in FBS serum (Supplementary Fig. 3a).
-As stated above, the effect shown with fluoxetine alone in sup. Figure 3A does not fit with a SERT/serotonin-mediated action of fluoxetine.These data should be experimentally strengthened, for example by using primary cells.
As suggested by the reviewer and as mentioned previously, we have clarified the mechanisms of action underlying the in vitro effects of FLX and we deeply investigated the effects of FLX on primary SCs (Fig. 3c-g,j-m).
-In all experiments using GR127935 or other compounds, the data for these compounds alone should be shown.
Except for the in vitro experiments on C2C12s, all in vitro and in vivo experiments with 5-HT1B and 5-HT2A antagonists were completed with the antagonist alone condition (Fig. 3c-g, Supplementary Fig. 4c-h and Supplementary Fig. 5a-g).

REFERENCES:
In order to clarify our results and as suggested by the reviewer, we have corrected the Supplementary Figure 1h.Regarding the relative expression of the different markers in the SCs population at steadystate, the FACS-sorted cells used in RT-qPCR were isolated from Tg:Pax7nGFP mice and thus all expressed the transcription factor PAX7 (Supplementary Figure 1a).Although BrdU and Cyclin D1 are not the same markers, they are both still markers of cell division.Thus, we can estimate that when SCs expressed a higher level of the Cyclin D1 gene at 6 weeks of FLX treatment, 4% of these SCs are still dividing (Supplementary Figure 1g).Finally, we attempted to demonstrate myogenin immunostaining on muscle sections, but in the absence of muscle injury, we did not observe myogenin positive cells (data not shown).This result suggests that the effects of FLX on gene expression of myogenesis genes such as myogenin are not always consistent with their protein expression.Because the results are confusing, we corrected this figure (Supplementary Figure 1h).
Complementary to the cell-autonomous effects of FLX demonstrated in vivo, we showed in vitro that FLX exerted direct effects on SCs in a 5-HT/5-HT1B receptor axis-dependent manner by promoting: -Cell activation with a faster entry into the cell cycle, equivalent to a faster exit from quiescence (Figure 3j) -cell proliferation with an increased rate of cell division (Figures 3c-d,k).These results are consistent, for example, with the increased activity of the Akt and TOR signaling pathways by 5-HT demonstrated in C2C12s (Supplementary Figure 3i).
-an enhancement of early and terminal differentiation (Figures 3e-f,h-i,l).These results are also consistent with the increased activity of ERK2 and STAT5b signaling pathways by 5-HT demonstrated in C2C12s (Supplementary Figure 3i).
-A maintained self-renewal of SCs without depletion of the SCs pool (Figures 3g,m).
These results could be further consolidated, and as the reviewer rightly points out, important mechanistic questions remain.However, the precise mechanisms of action of serotonin on stem cells are complex, as evidenced by the vast literature on the effects of serotonin on adult hippocampal neurogenesis which have not hitherto been elucidated 1 .It has been demonstrated the expression of at least 10 5-HT receptors by neural stem cells and progenitors, notably cell type 1, 2a and 2b, and which can modulate multiple signaling pathways involved in this neurogenesis 1,2 .This complexity of the 5-HT system and its modes of action are also found in other organs, such as angiogenesis (reference 57).
Given the known complexity of the mechanisms of action of 5-HT, it seems relevant to continue our efforts to investigate the mechanistic and cellular effects of 5-HT/FLX on SCs but in a future study specifically targeting this issue.
Indeed, the present study has brought to light a major and unexpected discovery on the positive effects of FLX on muscle regeneration associated with a functional and metabolic improvement of striated muscle.It deserves to be shared with the broad readership of Nature Communication, encompassing scientists from different backgrounds, and it opens major perspectives of translational medicine with the investigation in clinical practice of the therapeutic potential of FLX in muscle diseases.
Reviewer #2 (Remarks to the Author): Unfortunately the authors have not satisfactorily addressed my concerns on the mechanism by which FLX promotes MuSC activity.They tried to provide a verbal explanation, but did not perform the requested analysis of Pax7, cyclinD1 and Myogenin expression in MuSCs within the 3 different stages -quiescence, proliferation and differentiation -in response to FLX.This analysis is absolutely feasible and can be done in vivo (on tissue sections) as well as ex vivo (by single fiber analysis)in wild type mice at different time points after injury (typically days 1, 3 and 5 postinjury)in untreated or FLX treated mice.Specific antibodies and experimental conditions are available from previous studies published by multiple investigators in the field.
Replies to the Reviewers' comments: Please find the answers and corrections in blue below.
Reviewer #1 (Remarks to the Author): Unfortunately the authors have not satisfactorily addressed my concerns on the mechanism by which FLX promotes MuSC activity.They tried to provide a verbal explanation, but did not perform the requested analysis of Pax7, cyclinD1 and Myogenin expression in MuSCs within the 3 different stages -quiescence, proliferation and differentiation -in response to FLX.This analysis is absolutely feasible and can be done in vivo (on tissue sections) as well as ex vivo (by single fiber analysis) in wild type mice at different time points after injury (typically days 1, 3 and 5 post-injury)in untreated or FLX treated mice.Specific antibodies and experimental conditions are available from previous studies published by multiple investigators in the field.
We thank the reviewer for her/his positive comments on our work and constructive criticism of the manuscript.
As suggested by the reviewer, we performed an immunofluorescence-based kinetic analysis over time of cell distribution in vivo of the expression of the markers Pax7, Ki67 (another well-known cell division marker) and Myogenin within the MuSC population on muscle tissue sections from FLX-treated and untreated C57Bl6 mice before injury and throughout muscle regeneration after injury.
In the absence of muscle injury, mice treated with FLX for 6 weeks showed an increase in the population of Pax7-positive/Ki67-positive MuSCs compared with control mice (Supplemental Figure 1h and Figure 1, below), while the percentage of Pax7-negative/Myogenin-positive MuSCs was low and similar in both groups (Supplemental Figure 1i and Figure 1, below).These results are consistent with our previous experiments showing that FLX led to an increase in MuSCs number by promoting cell division and upregulation of Cyclin D1 gene expression, allowing an almost doubling of the MuSCs pool in the absence of muscle injury (Figures 1b-c and Supplementary Figures 1b, c, f, g).Taken together, these results suggest that in the absence of muscle injury, FLX promotes the emergence from quiescence and the proliferation of MuSCs without affecting their differentiation capacity.
The kinetics over time of cell distribution of MuSCs' markers Pax7, Ki67 and Myogenin have previously been finely characterized during muscle regeneration 1,2 .In the early stages of muscle regeneration, it was shown that injury initially induced a drastic reduction in the MuSC pool and, two days after injury, most MuSCs remained in a quiescent state and a minority were activated without dividing.Five days after injury, MuSCs were dividing strongly, and the differentiation capacity of the myogenic cell population increased.The switch from quiescence to cell proliferation and differentiation of MuSCs occurred between the second day and the fifth post injury.Thus, we studied MuSC activity at four days post injury, enabling us to appreciate at one time point consistent with our previous experiments, the different stages of the MuSC cascade: emergence from quiescence, proliferation, and differentiation.At four days post injury, FLX-treated mice showed an increase in both levels of Pax7-positive/Ki67positive MuSCs and Pax7-negative/Myogenin-positive MuSCs compared with control mice (Supplemental Figures 1a-b).
These results are consistent with our previous experiments showing that FLX-treated mice showed increased numbers of Pax7 and Myogenin-positive MuSCs number four days post injury (Figures 2be).
Taken together, these results suggest that FLX promotes the proliferation and differentiation of MuSCs during the early phases of muscle regeneration.
The cell-autonomous effects of FLX demonstrated in vivo on MuSC activity are also consistent with the in vitro positive effects of FLX that we previously described on MuSCs, namely stimulation of cell activation with accelerated entry into the cell division cycle, stimulation of cell proliferation with an increased rate of cell division, and stimulation of differentiation capacities (Figures 3c-e, h-k).
Beyond 7 days post injury, it has previously been shown that the proliferation and differentiation capacities of MuSCs declined, suggesting that the myofiber repair process had been completed, giving days, and 10 days after notexin injury, taken from Arnold et al. 1

Figure 1 :
Figure 1: 5-HT levels in sera from FLX-treated and control mice, measured by HPLC (n=3 mice per condition)

Figure 2 :
Figure 2: Muscle regeneration is harmoniously improved only by fluoxetine delivered by a 6-

Figure 1 :Figure 2 :Figure 3 :
Figure 1: Kinetics over time of cell distribution of Pax7/Ki67 immunostaining (top) and Pax7/Myogenin immunostaining (bottom) among SCs on TA sections from control and FLX-treated (6 weeks of treatment) C57Bl6 mice before injury, four days and fourteen days post injury (n=5-7 mice per condition).