Muscarinic receptors modulate Nerve Growth Factor production in rat Schwann-like adipose-derived stem cells and in Schwann cells

Regenerative capability of the peripheral nervous system after injury is enhanced by Schwann cells (SCs) producing several growth factors. The clinical use of SCs in nerve regeneration strategies is hindered by the necessity of removing a healthy nerve to obtain the therapeutic cells. Adipose-derived stem cells (ASCs) can be chemically differentiated towards a SC-like phenotype (dASCs), and represent a promising alternative to SCs. Their physiology can be further modulated pharmacologically by targeting receptors for neurotransmitters such as acetylcholine (ACh). In this study, we compare the ability of rat dASCs and native SCs to produce NGF in vitro. We also evaluate the ability of muscarinic receptors, in particular the M2 subtype, to modulate NGF production and maturation from the precursor (proNGF) to the mature (mNGF) form. For the first time, we demonstrate that dASCs produce higher basal levels of proNGF and mature NGF compared to SCs. Moreover, muscarinic receptor activation, and in particular M2 subtype stimulation, modulates NGF production and maturation in both SCs and dASCs. Indeed, both cell types express both proNGF A and B isoforms, as well as mNGF. After M2 receptor stimulation, proNGF-B (25 kDa), which is involved in apoptotic processes, is strongly reduced at transcript and protein level. Thus, we demonstrate that dASCs possess a stronger neurotrophic potential compared to SCs. ACh, via M2 muscarinic receptors, contributes to the modulation and maturation of NGF, improving the regenerative properties of dASCs.


Results
cholinergic modulation of nGf expression. Firstly, we investigated the ability of muscarinic agonists to modulate NGF expression after 24 h of treatment. NGF transcript levels were significantly decreased following APE treatments in both dASCs and SCs (Fig. 1A,D), compared to untreated controls, whereas muscarine was able to reduce NGF gene expression only in SCs (Fig. 1D).
Gene expression analysis showed that APE exposure was able to upregulate proNGF-A in dASCs compared to untreated controls (Fig. 1B), while a significant decrease was observed in SCs (Fig. 1E). On the other hand, the non-selective agonist muscarine upregulated proNGF-A transcript only in dASCs (Fig. 1B), but its protein levels remained still unmodified (Fig. 1G).
proNGF-B isoform expression was significantly downregulated after APE treatment (Fig. 1C, F), whereas muscarine treatment was able to significantly decrease its expression only in dASCs (Fig. 1C).
Following 48 h of exposure to 100 μM of APE or muscarine, rat dASCs did not display any variation of proNGF-A protein but showed lower levels of 25 kDa proNGF-B, compared to the untreated cells (Fig. 1G,H).
In SCs, the protein levels of proNGF-A remained unmodified whereas the 25 kDa proNGF-B isoform was significantly downregulated following both treatments (Fig. 1I, L).
ProNGF isoforms typically bind low-affinity nerve growth factor receptor p75NTR 7,36 . qPCR analyses demonstrated that p75NTR was significantly decreased in dASCs after both muscarinic agonists treatments ( Fig. 2A), but it was upregulated after muscarine treatment in SCs (Fig. 2B). Immunocytochemistry analysis confirmed the expression of p75NTR receptor in both cell types (Fig. 2C), however the expression of the receptor was not altered after muscarinic agonists treatments in SCs cells (Fig. 2C). In dASCs, the expression of p75NTR receptor was increased upon APE and muscarine treatment (Fig. 2C). Densitometric analysis of p75NTR immunostaining demonstrated an increased of p75NTR expression in dASCs upon both muscarinic agonist treatments in dASCs (Fig. 2D). Conversely, p75NTR expression was unaltered in SCs upon muscarinic challenge (Fig. 2C, D). dASCs produce and release higher basal levels of NGF. Several studies described how dASCs produce neurotrophic factors increasing the regenerative properties of these cells 30,37 .
Comparing dASCs with native SCs, we showed that higher levels of proNGF were detectable, by ELISA, after 24 and 48 h of cultures in both dASCs lysates and in the supernatants (Fig. 3A,C). The baseline concentration of proNGF was 2059 ± 256.8 pg/mg in dASCs lysates, whereas in SCs we observed a mean of production of 473.7 ± 68.69 pg/mg after 24 h of culture, four times lower than dASCs (Fig. 3A, 24 h). After 48 h dASCs produced six times more than SCs, with an average of 2124 ± 239.2 pg/mg in dASCs compared to 337.4 ± 45 pg/mg in SCs (Fig. 3A, 48h). dASCs released more proNGF in the culture media with a concentration of 736 ± 99.76 pg/mg vs 308 ± 78.59 pg/mg in SCs in the first 24 h after plating (Fig. 3C, 24h). After 48h, proNGF concentration was 897 ± 104.9 pg/mg in dASCs media, five times more than SCs, where it was 176.4 ± 87.68 pg/mg (Fig. 3C, 48h).
The average of concentration after 24 h was 23.43 ± 0.32 pg/mg for dASCs and an average production of 94.95 ± 19.54 pg/mg in SCs, a four-fold difference. Forty-eight hours after plating, mNGF concentration was 14.27 ± 1.7 pg/mg in dASCs whereas it was 129.2 ± 13.83 pg/mg in SCs, 9 times more than dASCs (Fig. 3B).
The results obtained suggest that proNGF production and release ( Fig. 3A and C), as well as NGF maturation (Fig. 3D) seem more efficient in dASCs than in SCs. cholinergic modulation of pronGf and mnGf production. By ELISA assays, we evaluated the amount of proNGF and mNGF production in dASCs (Fig. 4) and SCs (Fig. 5) upon cholinergic agonists stimulation.
concentrations were not affected by muscarinic agonists in the first 24 h of stimulation, while a significant upregulation after 48 h of muscarine exposure was observed (Fig. 4D).
Using an approach similar to that followed for dASCs, NGF production, maturation and release were evaluated in native SCs (Fig. 5). proNGF basal (controls) level production significantly decreased after 48 h of cell culture (Fig. 5A). After 24 h of APE treatment, proNGF concentration was significantly increased in cell lysates ( Fig. 5A), but it was decreased at control levels after 48 h (Fig. 5A). Similarly, muscarine increased proNGF concentration in lysates after 24 h of treatment, while a significant decrease was observed after 48 h (Fig. 5A). In SCs lysates, NGF maturation was unaltered after 24 h and 48h of exposure to APE and muscarine (Fig. 5B).
Similarly to what observed in cell lysates, in the culture media we observed a significant decrease of the basal proNGF released between 24 and 48 h of culture (Fig. 5C). Twenty-four hour of APE exposure increased proNGF release (Fig. 5C), while a significant decreased in concentration was observed following 48 h (Fig. 5C). There was no significant modulation of proNGF concentration in the cell media after non-selective challenge with muscarine.

Discussion
A relationship between ACh and NGF in the CNS has been previously reported. Indeed, NGF is important to regulate septo-hippocampal physiology and there are potential therapeutic uses in neurodegenerative disorders involving a cholinergic deficit 39 . Moreover, NGF was reported to attenuate cholinergic deficits in a rat model of traumatic brain injury 40 , and NGF gene transfer into aged animals can improve neuronal depolarization induced by ACh release from hippocampal synaptic terminals 41 . Conversely, ACh can regulate neurotrophin metabolism; indeed, the cholinergic system is implicated in the modulation of hippocampal NGF and BDNF mRNA expression, i.e. the injection of muscarinic agonist, pilocarpine, increases hippocampal BDNF and NGF mRNAs in postnatal and adult rats 42 .

Scientific RepoRtS |
(2020) 10:7159 | https://doi.org/10.1038/s41598-020-63645-w www.nature.com/scientificreports www.nature.com/scientificreports/ In the present work we demonstrate for the first time that the pharmacological modulation of muscarinic receptors regulates the production and secretion of NGF also in peripheral nervous system, and in particular in SCs. Moreover, we also demonstrate the ability of dASCs to produce higher levels of NGF compared to SCs, supporting the ability of dASCs in promoting peripheral nerve regeneration.
This work follows our previous study, focused on the effects mediated by M2 muscarinic receptors in rat SC-like adipose-derived stem cells 33 ; however, herein we have also analysed the effects of all other muscarinic receptors in the modulation of NGF production and release, highlighting the following points: cholinergic regulation of pro-nGf isoforms. M2 selective activation with APE treatment results in a significant decrease of proNGF-B gene and protein expression in both dASCs and SCs; whilst proNGF-A gene expression is significantly downregulated in SCs and it is significantly upregulated in dASCs after both APE and muscarine treatments. Distinct proNGF isoforms are expressed in mouse tissues, resulting from alternative splicing and/or activation of different promoters 5,43,44 ; as recently reported, proNGF-A has a pro-survival and differentiative effects in vitro 6 , whereas proNGF-B isoform acts as an apoptotic signal 7 . The longer variant proNGF-A (32-34 kDa) and the shorter variant proNGF-B (25-27 kDa) are produced in both CNS and PNS [43][44][45][46] . Our results suggest that muscarinic challenge promotes a proNGF-mediated neuroreparative response in dASCs, whilst cholinergic stimulation could improve neuronal survival and axon regeneration via proNGF-A/proNGF-B ratio modulation in glial cells.
Effects on the expression of the low affinity p75 receptor. ProNGF preferentially binds the low-affinity p75 neurotrophic receptor (P75NTR) 47 . After APE exposure, p75NTR gene expression decreases in dASCs but remains unchanged in SCs; whilst muscarine stimulation decreases p75NTR transcript levels in dASCs but promotes its expression in SCs. Interestingly immunocytochemistry analysis demonstrates an increased expression of p75NTR protein upon both muscarinic agonists treatment but only in dASCs. Considering recent evidence suggesting that several miRNA are involved in the regulation of p75NTR in different physiological and pathological conditions 48,49 , the differential expressions of gene and protein in SCs following APE and muscarine stimulation may indicate p75NTR post-transcriptional regulation. These data, together with the reported p75NTR-mediated pro-survival effect of proNGF-A 6 , support the idea that dASCs could positively promote the neuro-reparative process after M2 receptor challenge. dASCs produce and release higher levels of pro-and mNGF compared to SCs. The comparison between proNGF and mNGF content in dASCs and SCs point at the former cell type as a suitable and perhaps preferable regenerative tool in the treatment of peripheral nerve lesions. To investigate proNGF release and www.nature.com/scientificreports www.nature.com/scientificreports/ maturation, ELISA was employed with two different antibodies to discriminate proNGF and mNGF 50 . Our data clearly indicate that dASCs produce and release higher levels of proNGF and mNGF than SCs in basal condition. In light of previous reports, demonstrating the role of NGF in the process of peripheral nerve regeneration 1,11-13 , our findings suggest that dASCs could supply a larger amount of the neurotrophic factor, if used instead of native SCs as nerve-repair promoting tool. It is worth noticing that higher levels of mNGF in dASCs may be produced both after proNGF release and proteolytic cleavage in extracellular space and after proNGF intracellular processing by furin 4 .Though this matter deserves more specific investigation, our data suggest that the extracellular conversion of proNGF into mNGF is prevalent in the dASCs phenotype. www.nature.com/scientificreports www.nature.com/scientificreports/ Cholinergic stimulation differently modulates NGF production in dASCs and SCs. All muscarinic receptor activation by muscarine should generate a balance on their effect on NGF metabolism, however it is evident that muscarine and APE treatments show similar effects, suggesting that M2 receptor may be the main subtype involved in NGF metabolism both in dASCs and SCs. The difference between the two different ligands was especially observed in mNGF levels that is higher in dASCs lysates after 48 h of APE exposure and upon 48 h of muscarine treatment in cell media. ELISA data also show a relative decrease of proNGF and mNGF concentration after 48 h of treatment, more evident in SCs. These results suggest that in SCs, despite having lower NGF levels than dASCs, muscarinic stimulation may contribute to trigger proNGF/mNGF production more efficiently. On the other hand, the higher basal levels of proNGF and mNGF observed in dASCs could explain the delayed response to muscarinic stimulation in this cell type. This hypothesis is supported by the evidence that in dASCs a decrease in proNGF followed by an increase in mNGF cellular content is observed 48 h after APE stimulation, indicating that, in these cells, the M2 stimulation promotes proNGF maturation. Conversely, the intracellular proNGF increases upon only 24 h of cholinergic treatment in SCs, suggesting that proNGF and mNGF may be previously accumulated and then released after cholinergic stimulation in SCs. Overall, these data indicate that SCs may be less reactive than dASCs in generating an NGF-based neuro-reparative process in response to muscarinic activation.

NGF maturation and degradation machinery is differently expressed and modulated in dASCs
and Scs. ProNGF can be converted to mNGF in the extracellular environment and the tissue plasminogen activator (tPA) is involved in the proNGF to mNGF cleavage 4 . The tissue plasminogen activator tPA is a key enzyme involved in several biological processes, including vascular and tissue remodelling, tumour progression 51 and nervous system pathophysiology 52,53 . Tissue PA is a secreted protease that converts plasminogen into the active protease plasmin, which in turn participates in the cleavage of proNGF into mNGF 4 . In the PNS, tPA activity is elevated during axonal outgrowth and tPA is secreted by primary neurons and cultured SCs 54,55 , and its activity is particularly concentrated in growth cones 54 . Interestingly, the tPA system is potently induced also after peripheral nerve injury in mice 56,57 . As described above, peripheral nerve regeneration acts on the capacity of axons to regrow in a very intricate environment, composed by altered ECM, myelin and axon debris and infiltrating inflammatory cells. In these hostile conditions axonal growth seems hampered, thus regenerating neurites secrete tPA to degrade cell-cell and cell-matrix adhesion molecules. Accordingly, tPA mRNAs are strongly and early upregulated after peripheral nerve injury in Schwann cells and neurons, and their activity increases at the injury site up to 7 days post-crush 58 .
We found that tPA gene expression and activity are upregulated after 24 h of muscarinic stimulation both in dASCs and SCs. The increase of tPA activity is significant in both dASCs and SCs and this may indicate that the cholinergic stimulation may efficiently promote the proNGF maturation. In SCs, the increase of tPA activity is less evident, supporting the previous hypothesis that cholinergic stimulation in these cells promotes the proNGF and mNGF previously accumulated in the cells. In addition, tPA can convert plasminogen to plasmin, and plasmin, in turn, can work indirectly through the activation of matrix metalloproteinases (MMPs) to complement its degrading activities 59 . Accordingly, MMP9 is mainly involved in mNGF degradation 4 .
Although our data demonstrate that APE or muscarine treatments do not regulate the MMPs activity, we found a higher activity of MMP9 in dASCs than SCs. This suggests that the increased pro-and mNGF levels in dASCs supernatants indicate the possible NGF maturation and degradation in extracellular space by tPA and MMP9 activity respectively. MMP2 gene expression is significantly upregulated by cholinergic stimulation in dASCs but activity levels remain unchanged. Considering the role of MMP2 in the extracellular matrix remodelling and in the modulation of pro-inflammatory cytokines (i.e. TNFα) 38 , the presence of MMP2 activity in dASCs and in SCs may contribute to the regenerative property of these cells.

conclusions
Our data demonstrate that muscarinic receptors promote the NGF production and maturation in SCs and dASCs. The higher levels of basal NGF and the increase of proNGF and mNGF production upon cholinergic stimulation, accompanied by an efficient increase of tPA activity and higher levels of MMP9, suggest that dASCs are efficiently capable to produce NGF and that this production can be modulated by cholinergic stimulation. In the perspective of regenerative medicine, this feature is fundamental, suggesting the use of these cells as a reservoir of neurotrophic factors to accelerate the regenerative process. Moreover, it is known that NGF, acting through neuronal TrkA, strongly regulates myelination of DRG axons by both SCs and oligodendrocytes 60 . In accordance with our previous work on SCs 23 and dASCs 33 , NGF release could act in an autocrine way, via trkA and p75NTR receptors, improving SCs activity during nerve regeneration. This article postulates the first evidence of the involvement of muscarinic receptors, and in particular M2 receptor, in the regulation of nerve growth factor expression and release in the PNS. Although further analyses are needed to fully elucidate the role of ACh in modulating peripheral nerve regeneration, cholinergic receptor stimulation may represent, together with dASCs, a promising and clinically applicable pharmacological intervention.  Table 1. Data were normalized with 18 S housekeeping gene and the ΔΔCt method was used to determine the fold changes in the gene expression, as compared to control.

Methods
immunocytochemistry. After fixation with 4% Paraformaldehyde (PFA) for 20 min at RT, cells were incubated with 0.2% TritonX-100 for 30 min at RT. Then cells were washed twice with phosphate buffer Saline (PBS) and treated with block solution (PBS 0.1% TritonX-100 and 10% normal donkey serum (NDS)) for 1 h at RT. The next step was an incubation with primary antibody, rabbit anti-p75 NGF receptor antibody (Abcam, UK) at 4 °C overnight. The day after, cells were washed with PBS three times for 10 min and they were incubated with secondary antibody (Donkey anti-rabbit IgG (H + L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor 568, Life Technologies, UK) in 0.1% TritonX-100, 0.1% (w/v) BSA, 0.1% (w/v) Sodium Azide in PBS for 1 h at RT. Afterwards, plates were washed 3 times with PBS and slides were mounted with Vectashield mounting 7.9; 150 mM sodium chloride; 1% Nonidet P-40; 0.1 mM EDTA; 0.1 mM EGTA). Total protein content was evaluated by DCTM Protein Assay (Biorad, Italy). Twenty micrograms of proteins were resolved in 8-12% SDS-PAGE, blotted onto nitrocellulose membrane and processed for Western blot according to what has been previously reported 50 . Used antibodies are detailed in Table 2. Densitometry was performed on scanned immunoblot images using ImageJ gel analysis tool (National Institutes of Health, NIH, 469 Bethesda, MD, USA). The optical density (OD) of each protein band was normalized against the OD of the β-actin band. The results from at least three independent experiments were averaged and the standard error of the mean (SEM) was calculated.
Enzyme-linked immunosorbent assay (ELISA). proNGF and mNGF content in cell lysates and conditioned media was measured by specific ELISA, as previously described 50 . Briefly, capture antibody was the same for both the ELISA assays (Table 2) and was incubated overnight at room temperature (RT). Unbound antibody was removed by washing the plate once with washing buffer (0.5% (v/v) Tween-20 in PBS). After blocking 1 h at RT with 1% (w/v) BSA in PBS, the plate was rinsed with washing buffer and samples or standard curves were added to the wells and incubated for 2 h at RT. The microwells were then rinsed three times and incubated with specific detection antibodies dissolved in blocking buffer for 2 h, at RT. After three washes to remove unbound detection antibody, HRP-conjugated antibody, diluted in blocking buffer, was added and incubated for 1 h at RT. To visualize antibody reactivity, the chromogenic substrate 3' ,3' ,5' ,5'-tetramethylbenzidine (TMB, cat. T8768, Sigma-Aldrich, UK) was used and colour development was stopped by adding 1 N HCl. The colorimetric reaction was measured in absorbance mode at 450 nm by a Multiskan EX ELISA reader (Thermo Fisher Scientific Laboratory).

Zymography for matrix metalloproteinases (MMPs) detection and analysis of PAs activities.
Gelatinolytic activity of conditioned media was assayed as previously described 65 . Aliquots of conditioned media were analysed on 7% sodium dodecyl sulfate polyacrylamide gel electrophoresis SDS-PAGE containing 0.1% gelatin under non-reducing conditions. Following electrophoresis, gels were washed twice in 2.5% TritonX-100 for 30 min at RT to remove SDS and then in water for 30 minutes. The gels were incubated at 37 °C overnight in substrate buffer, stained with 0.5% Coomassie Brilliant Blue R250 and distained in 30% methanol and 10% glacial acetic acid (vol⁄vol). To analyse PA activity, aliquots of conditioned media were separated by 10% SDS-PAGE under non-reducing conditions 66 . After electrophoresis, gel was washed in 2.5% TritonX-100 for 30 minutes at RT to remove SDS and then in water for 30 minutes. The TritonX-100 washed gel was placed on a casein-agar-plasminogen underlay, as previously described 67 . All the bands were plasminogen dependent. The gels were photographed and the densitometric analysis was performed using ImageJ software (National Institutes