Acetyl-L-carnitine protects adipose-derived stem cells exposed to H2O2 through regulating AMBRA1-related autophagy

The cell activity of adipose-derived stem cells (ADSCs) is affected by the intracellular reactive oxygen species (ROS) and the level of autophagy. Previous studies reveal that acetyl-L-carnitine (ALC) possesses capacities of resisting oxidative stress and regulating autophagy. Activating molecule in Beclin1-regulated autophagy protein 1 (AMBRA1) plays a key role in initiating Beclin1-regulated autophagy. In the present study, we discovered ALC pretreatment (1 mM, 24 h) significantly increased the activity of ADSCs exposed to H2O2 (100 μM, 2 h) in vitro with improved stemness, and reduced the production of intracellular ROS. In addition, we found for the first time that ALC treatment up-regulated autophagy of ADSCs through strengthening the expressions of Beclin1 and AMBRA1 synchronously, which might be involved in the protective effect of ALC.

In addition to the advantages mentioned above, adiposederived stem cells (ADSCs) are easily available from fat and have better compatibility with tissue engineering scaffolds, which make them to be widely applied in regenerative medicine and cell therapy such as cell-assisted lipotransfer (CAL) (Dai et al., 2016;Mizuno et al., 2012). However, the cells are exposed to a lot of oxidative stresses during and after transplantation, which produces a negative impact on cell activity and graft survival. It is vital to explore how oxidative stress damages ADSCs and how to resist it.
L-carnitine (LC) has been shown to retard the aging process of ADSCs through a variety of mechanisms (Farahzadi et al., 2018;Mobarak et al., 2017). It also alleviates the damage induced by H 2 O 2 by regulating reactive oxygen species (ROS) (Shafiei et al., 2020;Ye et al., 2010). As the esterification form of LC, Acetyl-L-carnitine (ALC) possesses functionality similar to LC (Rebouche, 2004) and down-regulates ROS formation (Bodaghi-Namileh et al., 2018). However, there is little literature elaborating whether AMBRA1-related autophagy participates in the procedure. Therefore, we pretreated H 2 O 2 -exposed ADSCs with ALC, detected the intracellular ROS, and investigated the changes of autophagy level and the expressions of related proteins. The autophagy-related proteins assessed in the present study and their functions were listed as below (Tab. 1).

Reagents and catalog numbers
Acquisition and identification of adipose-derived stem cells The adipose tissue was cut into granules and washed with PBS thoroughly. After digested with 0.25% collagenase I at 37°C at a speed of 275 rpm for 40 min, the mixture was filtered with a 100-mesh strainer and a 200-mesh strainer and centrifuged at 1200 rpm for 5 min. Then the pellet was washed once to remove red blood cells and resuspended with complete medium at 5% CO 2 and 37°C. The passage process was carried out when the full confluence was achieved, and the cells at passage 5 were used in this study.
There are two essential requirements for the identification of human mesenchymal stem cells that include the detection of membrane cluster differentiation antigens and multi-lineage differentiation (Fathi et al., 2019a(Fathi et al., , 2019b. Our research group detected the phenotypic characteristics of ADSCs obtained by the same extraction method using flow cytometry, which showed that the stem cell markers, including CD29, CD90, and CD73, except for CD34 and HLA-DR, were positive (He et al., 2019).
Cells (passage 5) prepared for multi-lineage differentiation experiments were first seeded onto a 6-well plate at a density of 4×105 cells/well in complete medium for 2 days at which time the cell confluence reached 60%. For adipogenesis assay, ADSCs were washed thoroughly and incubated in a Human Mesenchymal Stem Adipogenic Kit consisting of insulin, dexamethasone, 3-isobutyl-1methylxanthine, and rosiglitazone for 2 weeks according to the manufacturer's instructions. Osteogenesis assay was similar to adipogenesis assays except that the medium was replaced with Human Mesenchymal Stem Osteogenesis Kit based on dexamethasone, β-glycerophosphate, and Lascorbic acid-2-phosphate. For chondrogenesis assay, cells in per-well were detached to a 15 mL centrifuge tube and  P62 is continuously consumed as a substrate with the autophagy level increased. Lira et al. (2013) aggregated into a sphere by centrifuging for 4 min at 250×g. Then the cell sphere was incubated in Human Mesenchymal Stem Chondrogenesis Kit consisting of dexamethasone, ascorbate, ITS supplement, sodium pyruvate, proline, TGF-β3, and basal medium for 3 weeks according to the manufacturer's instructions, followed with sectioning procedure. After the samples were fixed, Oil red O, Alizarin red, and Alcian blue staining were conducted. The samples were imaged under an Olympus CKX41 microscope (Japan). ADSCs pretreated with ALC (0.1, 0.5, 1, 5 or 10 mM) for 24 h were successfully washed with PBS once, exposed to 100 µM H 2 O 2 for 2 h, washed again, and detected using the CCK-8 test as above to determine the most suitable concentration of ALC (1 mM). The control group and the H 2 O 2 group were established simultaneously. Cell activity was expressed as ratios of all groups to the control group in OD values. In addition, 3-methyladenine (3-MA, 5 mM) was added into the cells for 3 h prior to ALC and H 2 O 2 treatment to study the role of autophagy in the procedure. The CCK-8 test and an analysis of the results were performed as above. The experiment was performed in quadruplicate and repeated three times independently.

5-ethynyl-2'-deoxyuridine (EdU) staining
In order to assess the DNA replication of ADSCs, cells were seeded at 5 × 105 cells/well in a 12-well plate and treated as above. After washed with PBS twice, cells were incubated for 2 h in complete medium containing 50 μM EdU and then fixed using 4% paraformaldehyde, followed with staining using Apollo and Hoechst33342 according to the manufacturer's instructions. Finally, the cells were observed under an ECLIPSE TI fluorescence inverted microscope (Japan) with 550 nm excitation wavelength and 565 nm emission wavelength. This experiment was done in triplicate and repeated independently three times.

ROS fluorescence staining
In order to detect intracellular ROS, cells were cultured and treated as above. After labeled using 10 μM 2',7'dichlorodihydrofluorescein diacetate (DCFH-DA) for 20 min at 37°C according to the manufacturer's instructions, cells were washed twice in PBS and observed under a fluorescence inverted microscope with 484 nm excitation wavelength and 525 nm emission wavelength. The number of replicates and sample size were set in the same way as EdU staining. Co-immunofluorescence (co-IF) After treated as above, cells were fixed with 4% paraformaldehyde, washed with PBS for three times, and sealed with 10% FBS at room temperature for 1 h, followed with incubation in rabbit anti-human primary antibody against AMBRA1 and mouse anti-human primary antibody against Beclin1 at 4°C overnight. Then the samples were washed thoroughly and incubated in goat anti-rabbit (Dylight 594) and goat anti-mouse (Dylight 488) second fluorescent-labeled antibodies for 1 h in dark condition. Afterward, cells were stained with DAPI and observed under a fluorescence microscope (493 nm excitation wavelength and 518 nm emission wavelength for Dylight 488, 591 nm excitation wavelength and 616 nm emission wavelength for Dylight 594, and 359 nm excitation wavelength and 461 nm emission wavelength for DAPI).
Western blot Cells were treated as above. RIPA lysis buffer (Solarbio, CHN) was used to obtain proteins whose concentrations were determined using a BCA kit. Then the samples were boiled and subjected to sodium salt-polyacrylamide gel electrophoresis (SDS-PAGE), followed with Western Blot analysis which was performed according to a previous study (Pagliarini et al., 2012). The primary antibodies against NANOG, SOX2, OCT4, SOD1, CAT, LC3b, p62, Beclin1, caspase3, AMBRA1, PIK3C3, GAPDH and sheep anti-rabbit second antibody were used in this experiment. Finally, the blots were imaged using a ChemiDoc MP imager (Bio-Rad, USA) and analyzed by Image Lab software (Bio-Rad). Co-IF and Western Blot were respectively repeated three times, and samples for each experiment were prepared independently. Determination of the concentration of ALC. ADSCs were pretreated with or without ALC (0.1, 0.5, 1, 5, and 10 mM) for 24 h, treated with 100 µM H 2 O 2 for 2 h, and detected by the CCK-8 test. **p < 0.01, when compared with the control group. #0.01 < p < 0.05, ##p < 0.01, when compared with the H 2 O 2 group (n = 4). (c) Effect of autophagy on the viability of ADSCs exposed to H 2 O 2 . After pretreated with 5 mM 3-MA for 3 h and 1 mM ALC for 24 h successively, cells were exposed to 100 µM H 2 O 2 for 2 h and detected by the CCK-8 test. **p < 0.01, *0.01 < p < 0.05; (n = 4). The results are expressed as the ratios of all groups to the control group in the OD values.

Statistical analysis
The data obtained were expressed as mean ± standard error of the mean (SEM). Statistical analysis was conducted by SPSS 19.0 using one-way ANOVA with the LSD t-test (α = 0.05). A statistical difference with a p-value of < 0.05 was considered significant.

Acquirement and identification of ADSCs
Cells at passage 5 exhibited spindle-shaped morphologies and arranged in swirled (Fig. 1). The results of differentiation induction showed that there were plenty of lipid droplets, bone nodules, and cartilaginous acid mucopolysaccharide under the microscope (Fig. 1), manifesting the ability of multiple-directional differentiation of the extracted cells.
ALC improved the cell activity of ADSCs exposed to H 2 O 2 As seen in the CCK-8 test, the cell activity of ADSCs exposed to H 2 O 2 decreased significantly for 2 h (fell to 63.92 ± 1.75% with 100 μM H 2 O 2 ) and seemed to reach a plateau when the H 2 O 2 concentration is higher than or equal to 150 μM, indicating that the stem cells had been severely damaged (Fig. 2a). Therefore, we treated ADSCs with H 2 O 2 at 100 μM for 2 h in subsequent experiments to stress stem cells. Cell activity recovered with ALC pretreatment for 24 h prior to H 2 O 2 and rebounded significantly to 82.78 ± 2.81% when the concentration of ALC was at 1 mM (Fig. 2b). Preincubating cells with 5 mM 3-MA for 3 h neutralized the effect of 1 mM ALC as the cell activity was reduced to a level (73.33 ± 1.11%) closer to that of the H 2 O 2 group, indicating that 1 mM ALC and 5 mM 3-MA were suitable for the subsequent experiments (Fig. 2c). In addition, we discovered that the protective effect of 5 mM ALC became weaker (75.08 ± 1.07%), and there was no significant difference between the 10 mM ALC group (61.92 ± 1.38%) and the H 2 O 2 group (66.90 ± 1.82%) (Fig. 2b).
In addition, treatment with H 2 O 2 or 3-MA decreased the percentage of EdU-stained ADSCs whose DNA replication rate was higher, while ALC significantly increased it (Figs. 3a and 3b). Moreover, the result of Western Blot showed that the expression of stemness-related proteins (NANOG, SOX2 and OCT4) significantly decreased with H 2 O 2 treatment, increased with ALC pretreatment, and decreased again with ALC plus 3-MA pretreatment (Figs. 3c and 3d).

ALC regulated intracellular ROS induced by H 2 O 2 and increased antioxidant enzymes
The results of ROS staining showed that H 2 O 2 significantly increased the percentage of ROS-positive cells with fluorescent green, while pretreatment with ALC significantly reversed the tendency. However, the content of intracellular ROS increased again when 3-MA was added prior to ALC (Figs. 4a and 4b).
According to the results of Western Blot, H 2 O 2 significantly reduced the expression of SOD1 but not CAT, while both proteins were overexpressed with ALC pretreatment. CAT was expressed weakly with 3-MA pretreatment, whereas SOD1 was expressed with no significant change (Figs. 4c and 4d).
ALC regulated autophagy of ADSCs exposed to H 2 O 2 with the change of AMBRA1 expression As shown in Fig. 5, a higher transformation rate of LC3bI to LC3bII, more expression of PIK3C3, and less expression of p62 occurred when cells were treated with ALC. All of the above were counteracted by 3-MA treatment in advance. In addition, the expression of caspase-3 was significantly reduced with ALC treatment but rose again with 3-MA pretreatment (Fig. 5).
According to the result of Western Blot and co-IF, the expressions of Beclin1 and AMBRA1 not only changed with a consistent trend (both decreased with H 2 O 2 or 3-MA and increased with ALC) (Figs. 6 and 7) but also had highly overlapping distributions in the cells (Fig. 7).

Discussion
It has been found that ALC protects cells from oxidative stress and the resultant cell aging or death, promoting the recovery of diseases in nervous and cardiovascular systems (Cherix et al., 2020;Ferreira and McKenna, 2017;Hagen et al., 2002). In the present study, ALC was proved to reduce the H 2 O 2induced injury and increase the hyperproliferation phenotype of ADSCs in the CCK-8 test and EdU staining, which might be related to the regulated intracellular ROS and autophagy level.
Diverse stresses have different effects on antioxidant enzymes including SOD1 and CAT (Liang et al., 2020;Somasundaram et al., 2019), and the expressions of them are sometimes inconsistent (Djordjevic et al., 2004;Hao and Liu, 2019). We found H 2 O 2 (100 μM, 2 h) significantly inhibited the expression of SOD1 but not CAT, while ALC pretreatment significantly increased the expressions of both enzymes. The optimized SOD/CAT combination can effectively clear up superoxide and hydrogen peroxide, the two of the most important ROS  (Hu and Tirelli, 2012). Correspondingly, we discovered that ALC down-regulated intracellular ROS induced by H 2 O 2 , but such effect was offset by 3-MA pretreatment, accompanied by down-regulated PIK3C3 expression and inhibited autophagy. That means autophagy, which contributes to scavenging ROS (Scherz-Shouval and Elazar, 2011), may be one of the mechanisms by which ALC protects cells against oxidative damage.
There is inconsistency in the effect of H 2 O 2 on autophagy (Duan et al., 2015;Lou et al., 2017), which may be explained by the difference in the experimental models and dosages of H 2 O 2 . Beclin1 was found to be a new substrate of caspase-3 and its cleavage might inactivate autophagy and augment apoptosis (Zhu et al., 2010). We also found that H 2 O 2 treatment (100 µM, 2 h) increased caspase-3 expression and decreased Beclin1 expression, which might account for the FIGURE 5. Expression of autophagy-related proteins. (a) Cells were treated as above. Western Blot was performed to detect the expression of LC3bII/LC3bI, PIK3C3, p62 and caspase-3. (b) The results are expressed as ratios of blot density values of the target proteins to that of GAPDH (Except LC3bII and LC3bI, the ratios between themselves are calculated). Equal loading of proteins is illustrated by GAPDH. **, p < 0.01, *, 0.01 < p < 0.05. n = 3. inhibited autophagy. It is necessary to add animal experiments to confirm these changes.
When cells were starved, AMBRA1 links Beclin1 more tightly and plays an important role in Beclin1-dependent autophagy (Fimia et al., 2007;Gu et al., 2014;Shi et al., 2014). According to the results of Western Blot and coimmunofluorescence in the present study, ALC (1 mM, 24 h) treatment improved the AMBRA1 and Beclin1 expression of H 2 O 2 -exposed ADSCs with fairly consistent increment and distribution, accompanied with raised autophagy level, while treating with PIK3C3 inhibitor 3-MA in advance significantly inhibited autophagy and reversed the changes of AMBRA1 and Beclin1. Nevertheless, further researches are required on the mechanisms by which ALC regulates AMBRA1 and the subsequent Beclin1-dependent autophagy.
In order to avoid the interference caused by the oxidation of H 2 O 2 and the direct ALC-H 2 O 2 reaction, cells were washed thoroughly and gently with PBS before each addition of H 2 O 2 or CCK-8 reagent. Moreover, the tendency that ALC alleviated the damage of H 2 O 2 to ADSCs was stable, although there was a small difference in values of cell activity between the same treating groups in the different CCK-8 tests (Figs. 2a-2c), which might be accounted for by the variability of cell clones. In addition, it is found that the positive effect of ALC was little or even disappeared when its concentration was raised above 5 mM. Whether this is caused by autophagic cell death or by apoptosis needs to be further determined by using 3-MA (to see if the cell viability is improved again) .
In summary, we found that in addition to the adjustment of the balance between ROS and antioxidant enzymes, ALC (1 mM, 24 h) might protect ADSCs against damage induced by H 2 O 2 (100 µM, 2 h) through regulating Beclin1dependent autophagy, which might be related to AMBRA1. Acknowledgement: Sincere thanks to the teachers of the Central Laboratory of Wenzhou Medical University for their help and support.
Availability of Data and Materials: All data generated or analysed during the current study are included in this published article. Conflicts of Interest: The authors declare that they have no conflicts of interest to report regarding the present study.