Insights into antioxidant activities and anti-skin-aging potential of callus extract from Centella asiatica (L.)

Formation of oxidative stress in dermal fibroblasts plays crucial roles in aging processes of skin. The use of phytochemicals that can promote capacity of fibroblasts to combat oxidative stress is an attractive strategy to prevent skin aging and promote skin beauty. Centella asiatica has been used to treat multitude of diseases for centuries. Previous investigations demonstrated that extracts from C. asiatica have a broad range of beneficial activities through their antioxidant activity. Hence, the extract from this medicinal plant could be a great candidate for anti-skin-aging agent. Callus culture offers a powerful platform for sustainable, rapid and large-scale production of phytochemicals to serve extensive demands of pharmaceutical and cosmeceutical industries. Here, we demonstrated the application of callus culture of Centella asiatica to produce bioactive metabolites. The 50% ethanolic extract of callus culture has distinctive features of chemical compositions and biological profiles. Information from HPTLC-DPPH and HPLC analysis suggested that the callus extract comprises distinctive antioxidant compounds, compared with those isolated from authentic plant. Moreover, results from cell culture experiment demonstrated that callus extract possesses promising antioxidant and anti-skin-aging activities. Pre-treatment with callus extract attenuated H2O2-induced-cytotoxicity on human dermal fibroblasts. The results from RT-qPCR clearly suggested that the upregulation of cellular antioxidant enzymes appeared to be major contributor for the protective effects of callus extract against oxidative stress. Moreover, supplementation with callus extract inhibited induction of matrix metalloprotease-9 following H2O2 exposure, suggesting its potential anti-skin-aging activity. Our results demonstrate the potential utility of C. asiatica callus extract as anti-skin-aging agent in cosmeceutical preparations.

www.nature.com/scientificreports/ Preparation of extracts and standards. About 100 mg of ground callus and authentic plant samples were individually mixed with 1 mL of methanol, 95% ethanol, 75% ethanol or 50% ethanol (v/v); extracted for 30 min by using sonication at 25 °C; and then filtered through Whatman Grade 42 paper (Whatman plc; Maidstone, UK). The filtrate was then used for HPTLC and HPLC analysis. For standards (asiaticoside, asiatic acid, madecassoside, madecassic acid, kaempferol, quercetin and rutin), the working solutions were prepared in methanol at final concentration of 1 mg/mL and filtered for further use.
High-performance thin-layer chromatography (HPTLC). The compounds in extracts were separated by using HPTLC technique. Briefly, 5 µL/band of the Centella extracts and 4 µL/band of the standards were applied on the Merck silica gel 60 F 254 glass HPTLC plates (20 × 10 cm) by a semi-automatic sampler Linomat 5 (Camag; Muttenz, Switzerland) using 8 mm bands, 11.4 mm apart, 8 mm from the lower edge, and 20 mm from the left plate edge. The plates were developed in the chamber saturation containing mobile phases. For separation of triterpenoids, the mobile phase was dichloromethane: methanol: water (14:6:1, V/V/V) 16 . For separation of flavonoids, ethyl acetate: methanol: water (100:25:10, V/V/V) was used as mobile phase. The development was done with a migration distance of 70 mm using an automatic development chamber ADC2 (Camag; Muttenz, Switzerland). For observation of triterpenoids, the plate was dipped with sulfuric reagent using Chromatogram Immersion Device III (Camag; Muttenz, Switzerland) and then heated at 100 °C for 3 min on the TLC Plate Heater (Camag; Muttenz,Switzerland). The derivatized plates were visualized under white light using visualizer 2 (Camag; Muttenz, Switzerland). For detection of flavonoids, the chromatogram was heated at 110 °C for 3 min and then was dipped in NP/PEG reagent. The derivatized plates were visualized under UV light at 366 nm using visualizer 2 (Camag; Muttenz, Switzerland).

HPTLC-DPPH bioautographic assay.
In order to estimate the antioxidant activity of tested samples, HPTLC-DPPH was performed as a screening assay. After development, the HPTLC plates were immersed with 0.02% DPPH-methanol solution and incubated in dark for 2 h 17 . The chromatogram was then examined under white light. Generated bands with yellowish color on the purple backgrounds were considered as compounds with antioxidant activities.

DPPH assay.
The radical scavenging activities of tested extracts were evaluated by traditional DPPH assay 18 .
Briefly, 100 µL of DPPH solution (0.008% w/v DPPH in methanol) was mixed with 100 µL of varying concentrations of callus extract in a 96-well microplate and incubated at 25 °C for 30 min. After incubation, the absorbance was recorded at 517 nm using CLARIOStar microplate reader (BMG Labtech; Ortenberg, Germany). 100% methanol was used as a control.
The percentage of radical scavenging activity (%RSA) of sample was calculated using the following formula: OD517 Control = Optical density at 517 nm of 100% methanol; OD517 sample = Optical density at 517 nm of samples. The concentration of extracts that resulted in 50% RSA was estimated from dose-response curve using Graph-Pad Prism version 9.0 (GraphPad Software). HPLC analysis. Seven standards, namely, asiaticoside, asiatic acid, madecassoside, madecassic acid, kaempferol, quercetin and rutin, along with the callus extract were used for the analysis. The analysis was performed using a Shimadzu HPLC (LC-20A) connected with a PDA detector (Shimadzu; Kyoto, Japan). A C-18 column (100 mm × 4.6 mm, 5 µm particle size, Phenomenex, USA) was used. The HPLC analysis was conducted according to the method of Buraphaka and Putalun 19 . The mobile phase used was a gradient of 1% acetic acid (A) and acetonitrile (B) at a flow rate of 1 mL/min with the following linear gradient HPLC solvent program: solvent B was increased from 0 to 25% over 20 min and further increased to 40% over 35 min, followed by further increase to 65% till 40 min and then held at 100% A for 5 min before returning to the initial state. The column temperature was controlled at 30 °C, and chromatograms were recorded at 210 and 320 nm.
Cell culture. Human foreskin fibroblasts, BJ (ATCC ® CRL-2522), were obtained from American Type Culture Collection (ATCC; Manassas, VA, USA). The condition for cell culture was followed the standard protocol as described previously 20 . BJ cells were maintained in MEM supplemented with 1 mM pyruvate, 10% FBS, and 100 U/mL penicillin and 100 µg/mL streptomycin. The cells were cultured at 37 °C in a humidified atmosphere with 5% CO 2 . Centella extract treatments. The 50% ethanolic extracts of Centella were dissolved with DMSO to produce a 1 mg/mL stock solutions. The treatment protocol was slightly modified from our previous in vitro experiments on skin cells 21,22  Measurement of an intracellular ATP. The BJ cells were seeded into multiple culture dishes (60 mm) at a density of 250,000 cells per dish in 4 mL of the MEM. Cells were cultured for 24 h at 37 ºC, 5% CO 2 before exposure to experimental conditions. Following the indicated treatments, treated cells were trypsinized and resuspended with PBS. Levels of intracellular ATP were immediately measured with CellTiter-Glo Luminescent Cell Viability Assay (Promega, Madison, WI, USA; Cat no. G7570) as previously described 24 . Briefly, a cell suspension (50,000 cells in 100 µL of PBS) was added into each well of opaque walled, 96-well white microplate. Subsequently, 100 µL of the CellTiter-Glo reagent was homogenously mixed with the cell suspension at room temperature to induce cell lysis and start the luminescent reaction. After a 10-min incubation, the luminescent signal was measured using the CLARIOStar microplate reader. The generated luminescent signal is directly proportional to the amount of ATP in the sample. To estimate the intracellular level of ATP, standard curves for each experiment were produced with serial dilutions of ATP solutions (0-1000 μM). The amount of ATP in each sample was calculated from the corresponding standard curve and then further transformed to an intracellular concentration using the cell number.  were conducted to evaluate statistical differences between means. All means and standard errors of the mean (SEMs) were calculated from three independent experiments. Data were expressed as the mean ± SEM. Statistical analysis was performed using GraphPad Prism version 9.0.

Antioxidant activities of Centella extracts.
To determine a suitable solvent for extraction process, callus culture and authentic plant of C. asiatica were extracted with series of solvents, including methanol, 95% ethanol, 70% ethanol, and 50% ethanol. Due to their abundance and pharmacological activities, asiaticoside, asiatic acid, madecassoside, and madecassic acid are considered to be major active components of C. asiatica 7,26 .
To examine whether callus extract (CE) and authentic plant extract (APE) contains these triterpenoids, active compounds in extracts were separated with HPTLC approach. All of four centelloids were observed in APE (Lanes 5-8; Fig. 1A); however, none of them were detected in CE (Lanes 9-12; Fig. 1A). To screen active antioxidants in the extracts, HPTLC plates were further derivatized with DPPH reagent. A stable free radical DPPH is a widely used substrate to evaluate free radical scavenging activities of tested compounds. On HPTLC plates, the compounds generating yellow zones against the purple background were considered as antioxidants. The CE and APE exhibited strong free radical scavenging activities. Interestingly, all of tested triterpenoids did not show antioxidant activities ( Fig. 1B; Lanes 1-4). The CE showed significant antioxidant activity in the absence of these standards. These HPTLC-DPPH data suggested that asiaticoside, asiaticoside, asiatic acid, madecassoside, and madecassic acid do not contribute to antioxidant properties of the extracts. Kaempferol, quercetin, and rutin reported in C. asiatica 27,28 are flavonoid compounds with antioxidant activities. It was suspected that these compounds may contribute to the antioxidant activity of the extracts. However, the chromatogram on the HPTLC plate demonstrated that these three flavonoids were not presented in any extracts ( Fig. 2A). The data from HPTLC-DPPH showed that CE and APE exhibited significant antioxidant activities, as evidenced by a great intensity of yellow spots (Fig. 2B). Moreover, Centella extracts exhibited potent radical scavenging properties without the presence of kaempferol, quercetin and rutin (Fig. 2B). Taken together, these results strongly suggested that antioxidant activities of extracts are not due to these three flavonoids but possibly from other flavonoids.
As demonstrated in Figs. 1B and 2B, it was observed that 70% and 50% ethanolic extracts of both authentic plant (Fig. 1B, lanes 7 and 8; Fig. 2B, lanes 6 and 7) and callus (Fig. 1B, lanes 11 and 12; Fig. 2B  www.nature.com/scientificreports/ solvent to be used in extraction of C. asiatica because (1) 50% ethanolic extracts of callus and authentic plant exhibit high antioxidant activity; (2) the amount of organic solvent is less than 70% ethanolic extracts; and (3) ethanol is safer than methanol, then it is practical to develop product for cosemeceutical purposes. Hence, 50% ethanolic extracts of callus and authentic plant were chosen for further experiments of this study. As shown in Figs. 1 and 2, the antioxidant properties of CE were comparable to APE. We then further evaluate total antioxidant activity of these two 50%-ethanolic extracts by determination of IC50 of %RSA with DPPH assay. The IC50 of %RSA represents the concentration of extracts that is required for 50% free radical scavenging activity. We exhibited that CE had approximately 2.5 folds greater ability to neutralize free radicals than APE (Supplementary Fig. 1; CE, 98.6 ± 1.6 µg/mL vs APE, 243.3 ± 14.6 µg/mL). These results suggest that extract from callus culture possesses a strong free radical scavenging ability, which is even greater than from native plant.  According to cell-free based antioxidant assay, CE and APE were selected to further examine their beneficial activities on human dermal fibroblast BJ cells. Firstly, we evaluated potential toxicity of CE and APE on dermal fibroblasts. The data from MTT assay demonstrated that 24-h exposure of CE or APE, up to 60 µg/mL, did not cause toxicity to BJ cells (Fig. 4A, B) (Fig. 4C). Pre-treatment with CE significantly inhibited H 2 O 2 -induced toxicity (Fig. 4C). Interestingly, the protective effects of CE were comparable to APE pre-treatments (Fig. 4D). These data suggested that CE and APE could exert their beneficial effects against H 2 O 2 on dermal fibroblasts via their antioxidant activities.
The imbalance of cellular bioenergetics due to oxidative stress is closely linked to aging processes of several tissues, including skin 30 . Exposure to H 2 O 2 caused an immediate depletion of steady-state levels of ATP in dermal fibroblasts (Fig. 4E). In contrast, treatment with non-toxic concentration of CE (60 µg/mL; 24 h) did not alter intracellular levels of ATP, supporting the safety of CE on the selected concentration on BJ cells (Fig. 4E). Interestingly, pre-treatment with CE significantly inhibited reduction of intracellular ATP pool following H 2 O 2 treatment in dose-dependent fashion (Fig. 4F). The results from these bioenergetics studies were consistent www.nature.com/scientificreports/ with the protective effects of CE against oxidative damage from MTT assay, highlighting the contribution of antioxidant activity of CE on its protective effect against oxidative insults.

Effects of Centella extracts on expression of antioxidant enzymes. The beneficial effects of CE
and APE against oxidative damage as showed in Fig. 4 are possibly due to an induction of capacity of fibroblasts to detoxify reactive oxygen species (ROS). To examine this hypothesis, the mRNA expression of key antioxidant enzymes, including catalase (CAT), glutathione peroxidase 1 (GPx1), superoxide dismutase 1 (SOD1), and superoxide dismutase 2 (SOD2), were measured following 24-h treatments. Our RT-qPCR results demonstrated that CE and APE have different profiles on the induction of antioxidant machinery of fibroblasts; CE induced expression of CAT in dose-dependent fashion (Fig. 5A), while APE enhanced expression of both SOD1 (Fig. 5C) and SOD2 (Fig. 5D). The expression of GPx1 did not alter with either CE or APE treatment (Fig. 5B). These RT-qPCR data suggested that an upregulation of cellular antioxidant enzymes is the principal factor for protective effects of CE and APE. We also observed the effects of Centella extracts on expression of antioxidant enzymes after H 2 O 2 treatment. Exposure to H 2 O 2 leaded to dramatic decrease in GPx1 (Fig. 5B) and SOD1 (Fig. 5C); and significant increase in  www.nature.com/scientificreports/ SOD2 (Fig. 5D). The increase of SOD2 expression is possibly due to adaptive mechanism of fibroblasts following H 2 O 2 treatment. This phenomenon has been reported in several models of aged fibroblasts [31][32][33] . Again, CE and APE demonstrated distinctive effects on induction of antioxidant enzymes following H 2 O 2 exposure. Pre-treatment with CE induced CAT (Fig. 6A), GPx1 (Fig. 6B) and SOD1 expression (Fig. 6C), whereas APE upregulated SOD2 expression following 1-h exposure to H 2 O 2 (Fig. 6D). Altogether, these results clearly exhibited that Centella extracts could prevent H 2 O 2 cytotoxicity by an enhanced capacity of fibroblasts to remove deleterious ROS.  www.nature.com/scientificreports/ Effects of Centella extracts on expression of matrix metalloprotease 9 (MMP-9). Matrix metalloprotease-9 (MMP-9) is the gelatinase enzyme that is responsible for regulation of homeostasis of collagen. Formation of oxidative stress can lead to upregulation of MMP-9 in fibroblasts, which subsequently results in degradation of collagen. The increased breakdown of collagen by MMP-9 appears to be the major factor for aging processes of skin tissue 34,35 . Here, we demonstrated that exposure to H 2 O 2 markedly upregulated the expression level of MMP-9 of dermal fibroblasts (Fig. 7A). Elevation of MMP-9 expression was significantly suppressed by CE and APE pre-treatment (Fig. 7B). These data suggested that CE and APE may possess anti-aging activities via an inhibition of MMP-9 transcription.

Discussions
C. asiatica is a medicinal plant with broad range of pharmacological effects e.g. antioxidant [8][9][10][11] , anti-inflammation 36,37 , wound healing [38][39][40] , neuroprotective 8,9 and memory improvement 41,42 . Due to its high therapeutic potential, the demand for this plant in cosmeceutical and pharmaceutical industries has exceeded the supply from conventional cultivation. Callus culture offers a manufacturing system which ensures the continuous supply of compounds with uniform quality and high yields 13 . Here, we demonstrated that 50% ethanol is the suitable solvent to be used in extraction process for C. asiatica due to (i.) antioxidant activity of extracts; and (ii.) safety from reduced solvent residue. The 50% ethanolic extracts from callus culture has unique chemical profiles and biological activities. Compared to APE, the data from HPTLC (Fig. 1,2) and HPLC (Fig. 3) suggested that (1) CE comprises of unique antioxidant compounds which should be further identified; (2) the types of secondary metabolites in CE are different from APE. Interestingly, results from DPPH assay showed that the total antioxidant activity of CE is higher than APE (Supplementary Fig. 1). The greater antioxidant activity of CE is possibly due to the differences in active components in the extracts. We demonstrated that the major triterpenoids of C. asiatica (asiaticoside, asiaticoside, asiatic acid, madecassoside and madecassic acid) are not responsible for antioxidant activities of CE (Fig. 1). These findings are parallel to previous observation demonstrating that triterpene-free-extract of C. asiatica possesses greater in vitro radical scavenging properties as well as in vivo antioxidant activities than triterpene-enriched-extract. Triterpene-free-fraction of C. asiatica has significantly stronger in free-radical scavenging activities than triterpene-enriched-extract as determined by several in vitro anti-radical assays, including DPPH, ABTS, NORAC, ORAC and NO scavenging assays. Moreover, pre-treatment with triterpene-free-extract of C. asiatica reduced levels of malonaldehyde, an oxidative stress marker, in brain tissues of scopolamine-treated-rat, while triterpene-enriched-fraction did not have these protective effects 43 . These preclinical findings support our results that the centelloids are not principal contributors for antioxidant activity of C. asiatica. Kaempferol, quercetin, and rutin have been reported to be major flavonoids with antioxidant properties of C. asiatica 27,28 . Our HPTLC-DPPH and HPLC results demonstrated that these flavonoids did not present in 50% ethanolic extracts from both callus culture and authentic plant; then these series of flavonoids did not contribute to antioxidant activities of these extracts (Fig. 2). Hence, future studies are required to identify and characterize novel antioxidants presented in these extracts.
The process of skin aging is closely associated with an induction of oxidative stress in dermal fibroblasts 1 . Hence, compound that can prevent oxidative damage in dermal fibroblasts could be a potential candidate to be used as anti-skin-aging in cosmeceutical product. We then further investigated biological activities of Centella extracts against oxidative stress on dermal fibroblasts. Our data from MTT assay as well as bioenergetic study showed that CE significantly inhibited the cytotoxicity of H 2 O 2 on dermal fibroblasts (Fig. 4). The preventive effects of CE were comparable to APE. Moreover, we found that the protective activities of these extracts could be due to an increase in capacity of fibroblasts to eliminate ROS. We used RT-qPCR approach to investigate the alterations of mRNA expression of key antioxidant enzymes in response to treatments. Catalase is the major enzyme involved in the elimination of high fluxes of H 2  ·− -removal enzymes, SOD1 and SOD2 (Fig. 5). In response to H 2 O 2 treatment, pre-exposure to CE induced CAT , GPx1 and SOD1 expression, whereas pretreatment with APE upregulated SOD2 expression (Fig. 6). Upregulation of these antioxidant machineries due to Centella extracts could promote capacity of fibroblasts to eliminate harmful ROS, resulting in suppression of oxidative damage and prevention of cell death. Further studies on protein levels and enzymatic activities of antioxidant machineries following treatments of CE and APE are necessary to better understand the protective mechanisms of these extracts on fibroblasts.
Consistent with our findings, many preclinical studies also demonstrated that the beneficial activity of Centella extracts against oxidative-stress-related-disorders could be due to an augmentation in expression and/or activity of cellular antioxidant enzymes. In rat model of hepatic injury, administration of Centella extract prevented hepatotoxicity through an increased level of catalase, SOD and GPx in liver tissues 46 . In hamster model of hyperlipidemia, supplementation with ethanolic extract of C. asiatica promote hepatic function by an enhanced expression of SOD and GPx in hepatic tissues 47 . In diabetic rat, aqueous extracts from C. asiatica ameliorate hippocampal dysfunction by an induction of catalase, SOD and GPx expression in hippocampus 48 . In stroke model, supplementation with ethanolic extract of C. asiatica prevented brain injury through a restoration of glutathione level and augmentation of catalase, SOD and GPx activities in ischemic rat 49  www.nature.com/scientificreports/ previous observations strongly support our RT-QPCR results which demonstrating that 50% ethanolic extracts of Centella inhibited oxidative damage on human dermal fibroblasts via an upregulation of antioxidant enzymes. The nuclear factor erythroid 2-related factor (Nrf2)/ antioxidant response element (ARE) pathway is an essential signaling network that maintains redox homeostasis for human cells 50 . Nrf2 functions as a redox sensor for oxidative stress. In the presence of oxidative insults, Nrf2 translocates into nucleus and binds to promoter regions of ARE, leading to transcriptional activation of a battery of cytoprotective and detoxification genes, including CAT , GPx and SOD [51][52][53][54] . Activation of Nrf2/ARE pathway by Centella extracts has been observed in preclinical models of oxidative-stress-related disease. A water extract of Centella asiatica can attenuate oxidative stress and activate Nrf2/ARE network in isolated primary neurons and rodent models of pathological cognitive impairment. The activation of this pathway by Centella extract is strongly accompanied by an improvement in neuronal health and cognitive function 8,[55][56][57][58] . Recently, Park and colleagues demonstrated that pre-treatment with Centella extract can prevent the progression of age-related macular degeneration in vitro and in vivo. They clearly showed that the cytoprotective activities against oxidative damage of Centella extract in cell culture and animal experiment is mainly due to the regulation of Nrf2 pathway 59 . Hence, we postulated that Nrf2/ARE pathway could be a potential mediator for the beneficial effects of Centella extracts in our model. Further pharmacological study is required to validate this hypothesis.
In addition to effects on cellular antioxidant enzymes, we also observed activities of Centella extracts on expression of MMP-9. MMP-9 is zinc-containing-gelatinase which plays an important role in degradation of dermal extracellular matrix, especially collagen type IV. Environmental insults, e.g. UV irradiation, tobacco smoke, and pollutants, have been reported to induce expression of MMP-9 of skin cells through formation of oxidative stress 35 . Upregulation of MMP-9 leads to fragmentation of dermal collagen; thereby diminish skin elasticity and integrity; and eventually promote wrinkle and sagging formation of skin 34,60 . Hence, agents with MMP-9 www.nature.com/scientificreports/ inhibitory activities would be an attractive candidate to combat skin aging 61 . Our RT-qPCR data demonstrated that CE and APE significantly inhibited upregulation of MMP9 following H 2 O 2 exposure (Fig. 7), suggesting anti-skin-aging activities of these extracts. Future observations on detailed mechanisms of CE/APE-mediated-MMP9-expression are required to better understand their anti-skin-aging effects. The proposed mechanisms for antioxidant and anti-skin aging activities of CE and APE are summarized in Fig. 8.
In conclusion, our present study provides the potential application of callus culture platform to produce biomass and biochemical from C. asiatica. We demonstrated that the 50% ethanolic extract from callus culture has unique chemical profiles and distinct antioxidant properties. Moreover, this study provides pharmacological evidence to support the use of Centella extracts as anti-skin-aging agents in cosmeceutical and pharmaceutical products.  www.nature.com/scientificreports/