Cytotoxic, Antimicrobial, Antioxidant Properties and Effects on Cell Migration of Phenolic Compounds of Selected Transylvanian Medicinal Plants

Medicinal plants are widely used in folk medicine but quite often their composition and biological effects are hardly known. Our study aimed to analyze the composition, cytotoxicity, antimicrobial, antioxidant activity and cellular migration effects of Anthyllis vulneraria, Fuchsia magellanica, Fuchsia triphylla and Lysimachia nummularia used in the Romanian ethnomedicine for wounds. Liquid chromatography with mass spectrometry (LC-MS/MS) was used to analyze 50% (v/v) ethanolic and aqueous extracts of the plants’ leaves. Antimicrobial activities were estimated with a standard microdilution method. The antioxidant properties were evaluated by validated chemical cell-free and biological cell-based assays. Cytotoxic effects were performed on mouse fibroblasts and human keratinocytes with a plate reader-based method assessing intracellular adenosine triphosphate (ATP), nucleic acid and protein contents and also by a flow cytometer-based assay detecting apoptotic–necrotic cell populations. Cell migration to cover cell-free areas was visualized by time-lapse phase-contrast microscopy using standard culture inserts. Fuchsia species showed the strongest cytotoxicity and the highest antioxidant and antimicrobial activity. However, their ethanolic extracts facilitated cell migration, most probably due to their various phenolic acid, flavonoid and anthocyanin derivatives. Our data might serve as a basis for further animal experiments to explore the complex action of Fuchsia species in wound healing assays.


Introduction
Nowadays, the investigation of natural extracts from medicinal plants has been increased due to their rich content of bioactive compounds such as polyphenols, vitamins and proteins, which are found in different parts of plants [1]. Phenolic compounds (flavonoids, phenolic acids, anthocyanins, tannins) are secondary metabolites, which play a crucial role in the pharmaceutical sciences, thanks to their extensive biological effects (antimicrobial, antioxidant, anticancer properties). These bioactive substances have diverse basic structures but possess an aromatic ring bearing one or more hydroxyl groups, which can be related to different biological impacts [2,3].
The chromatographic separation was performed on an Agilent 1100 HPLC system equipped with a G1379A degasser, G1312A binary gradient pump, G1329A autosampler, G1316A column thermostat and G1315C diode array detector (DAD) (Agilent Technologies, Waldbronn, Germany). Samples were separated on a Zorbax SB-C18 (Agilent Technologies, Santa Clara, CA, USA) (150 mm length, 3.0 mm i.d., 3.5 µm particle diameter) column, maintained at 25 • C. The mobile phase was composed of 0.3% acetic acid in water (v/v) (A) and methanol (B). The following gradient program was applied, at a flow rate of 0.3 mL/min with the composition of the mobile phase changing from 5% B to 100% B in 30 min, maintaining 100% B for 5 min and returning to 5% B in 1 min. All aqueous solvents were filtered through MF-Millipore (Millipore, Billerica, MA, USA) (0.45 µm, mixed cellulose esters) membrane filters. Chromatograms were acquired at 280 nm. Injection volume was 5 µL. Prior to injection, all samples were filtered through Sartorius (Goettingen, Germany) Minisart RC15 (0.2 µm) syringe filters.

MS Conditions
Mass spectrometric analyses were performed with an Agilent 6410B triple quadrupole equipped with an electrospray ionization source (ESI) (Agilent Technologies, Palo Alto, CA, USA). ESI conditions were as follows: temperature: 350 • C, nebulizer pressure: 40 psi, N 2 drying gas flow rate: 9 L/min, fragmentor voltage: 120 V, capillary voltage: 4000 V, collision energy was changed between 10 eV and 45 eV, depending on the analyzed structure. High purity nitrogen was used as collision gas. Full mass scan spectra were recorded in negative and for anthocyanins in positive ionization mode over the range of m/z 50-1000 Da (scan/s). The MassHunter B.01.03 software was used for data acquisition and qualitative analysis. The microdilution method was performed according to a previously published method with some modifications [30]. In brief, 100 µL of bacterial suspensions (10 5 CFU/mL) in modified RPMI 1640 and 100 µL of diluted aqueous or 50% (v/v) ethanolic leaf extracts in modified RPMI 1640 media were pipetted into each well of sterile 96-well plates. The sterile medium was considered as negative control, the inoculated RPMI 1640 without any treatment was taken as the bacterial growth control, while erythromycin was used as positive control. The final concentration of the ethanolic solvent for the dilution was restricted up to 1.0% v/v in the wells. The absorbance was measured at 595 nm on Multiskan EX 355 (Thermo Electron Corporation, Waltham, Massachusetts, USA) spectrophotometer, after 24 h incubation time at 30 • C. Absorbance values lower than 20% of the bacterial growth controls were considered as MIC 80 . Treatments were carried out with three technical replicates in five independent experiments.

Oxygen Radical Absorbance Capacity (ORAC) Assay
The ORAC test was executed according to the method of Kőszegi et al. without modifications [31]. The method is based on fluorescence quenching of Na 2 -fluorescein oxidized by AAPH. The quenching is delayed by the antioxidants present in the standards/samples. Serial dilutions of Trolox were used as standard. Briefly, into each well of normal 96-well plates 150 µL of working fluorescein solution (400 nM dissolved in 75 mM potassium phosphate buffer, pH 7.5) and 25 µL of blank/standard/plant extract (aqueous/ethanolic) were pipetted and the plates were preincubated for 30 min at 37 • C in the dark. After automated injection of 25 µL of AAPH solution (400 mM dissolved in 75 mM potassium phosphate buffer, pH 7.5) the fluorescence intensities were measured in kinetic mode for 80 min at 37 • C, with excitation and emission wavelengths of 490 and 520 nm, respectively. For the fluorescence measurements, the plate reader (BioTek Synergy HT, Winooski, Vermont, USA) was thermostated at 37 • C. Five independent experiments were done with three technical replicates for each treatment.

Enhanced Chemiluminescence (ECL) Assay
The enhanced chemiluminescence method was performed following our previously published study without modifications [31]. The technique is based on the development of enhanced chemiluminescence (ECL) of luminol in the presence of peroxidase (POD), H 2 O 2 and 4-iodophenol enhancer. The increase of the ECL signal is delayed, depending on the antioxidant capacity of the samples. Briefly, 70 µL of ECL detection reagent (0.15 M boric acid/NaOH, pH 9.6, supplemented Antioxidants 2020, 9, 166 5 of 29 with 0.45 mM luminol and 1.8 mM 4-iodophenol) and 200 µL POD enzyme solution (15 µU/mL) were premixed and kept on ice. Trolox dilutions were used as standard. Into each well of white optical 96-well plates 20 µL Trolox/blank/sample and 270 µL of POD-ECL reagent were added. The reaction was initiated by automated injection of 20 µL ice-cold H 2 O 2 (1.5 mM, in 0.1% citric acid). The chemiluminescence signal was followed for 10 min, using a plate reader (Biotek Synergy HT) in kinetic analysis mode. Five independent experiments were done with three technical replicates for each treatment.
The method is based on the absorbance decrease of DPPH, which is a stable organic radical. The measurement was conducted following the protocol described elsewhere [32,33], with some modifications. Briefly, 50 µL of blank/standard/plant sample dilutions followed by 100 µL of 200 µM DPPH (dissolved in 96% ethanol) and 50 µL of acetate buffer (100 mM, pH 5.5) were pipetted into 96-well general microplates. The absorbance changes were measured at 517 nm by a Perkin Elmer EnSpire Multimode plate reader (Perkin Elmer, Waltham, MA, USA) after 60 min incubation in the dark, at room temperature. The results were compared to serial dilutions of Trolox standard solution.
Five independent experiments were done with three technical replicates for each treatment.

Trolox Equivalent Antioxidant Capacity (TEAC) Assay
The technique is based on the generation of ABTS radical cation (ABTS• + ) through the reaction between ABTS and potassium persulfate (K 2 S 2 O 8 ). The method of Re et al. and Stratil et al. was adapted for the TEAC test with slight modification [34,35]. ABTS• + was produced by reaction of ABTS stock solution (7 mM of ABTS dissolved in distilled water) with 2.45 mM K 2 S 2 O 8 (final concentration) and diluted with PBS (pH 7.4) until the absorbance was 0.70 ± 0.005 at 734 nm. Then 20 µL aliquots of varying concentrations of the leaf extracts (50% ethanolic/aqueous) were allowed to react with 80 µL of ABTS• + (7 mM) and the absorbance readings were recorded at 734 nm by the Perkin Elmer EnSpire Multimode plate reader after 20 min incubation in the dark, at room temperature. Trolox was used as standard. All measurements were carried out in five independent experiments with three technical replicates.

Calculation of Total Antioxidant Capacities (TAC)
For both the ORAC and ECL assays, the results were calculated as Trolox equivalents (TE). In the ORAC method the area under the fluorescence curve (AUC) of the blank was subtracted from that of the standard/sample (netAUC) and a calibration line was calculated for the netAUC of the Trolox standards. In the luminescence technique (ECL) the AUC of the emission curves vs. Trolox standards were used to calculate the calibration line. In both cases the samples' TE values were obtained from the calibration curves which were then multiplied by the dilution factor and expressed as µM TE concentration. Finally, TAC was referred to 1 g of initial dry material for each plant sample.
For the DPPH and TEAC assays, the radical scavenging activity was expressed as IC 50 (the concentration of the plant extract in µg/mL, required to scavenge 50% of DPPH or ABTS reactions), calculated by a linear regression curve made from the scavenging activities vs. amount of extracts of the samples. This means that the lower the IC 50 value of the sample is, the higher antioxidant activity it possesses.
Radical scavenging activity of the leaf extracts in % of the blank was obtained using the following formula: Radical scavenging activity (% inhibition) where A 0 is the absorbance of the blank and A 1 is the absorbance of the sample.

Quantification of Intracellular ROS
Cellular oxidative stress due to the overproduction of reactive oxygen species (ROS) generated by AAPH was measured using the DCFH-DA and the DHR123 methods [36][37][38]. Trolox and quercetin were used as positive controls. The optimal conditions of DCFH-DA and DHR123 assays were as follows on 96-well culture plates: seeding density of 5 × 10 4 cells/mL, cell culture incubation time for overnight. After washing with PBS co-incubation in Hanks' (5.5 mM glucose) with 50 µM DCFH-DA or 10 µM DHR123 and plant extract/quercetin/Trolox on 3T3/HaCaT cell cultures for 1 h was performed. After removal of the treating medium and washing with PBS, 1 mM AAPH oxidant in Hanks' glucose was added. The fluorescence intensity was recorded for 60 min on the Biotek microplate reader at 490/520 nm exc/em wavelengths at 37 • C. The radical scavenging activity was expressed as IC 50 (the concentration of the plant sample (µg/mL), required to scavenge 50% of DCFH or DHR123 fluorescence), calculated by a linear regression analysis of the serial dilutions of the leaf extracts.
The radical scavenging activity was obtained using the following equation: Radical scavenging activity (% inhibition) where AUC 0 is the area under curve values of the blank and AUC 1 is the area under curve values of the sample. Five independent experiments were done with four technical replicates for each treatment.

Plate Reader Cytotoxicity Test
A multiparametric viability assay with one-step extraction was carried out following our previously published study without modifications to investigate the potential toxicity of 50% (v/v) ethanolic and aqueous leaf extracts [28]. A. vulneraria in 500-2500 µg/mL concentrations (ethanolic extracts) and 4000-8000 µg/mL concentrations (aqueous extracts) were tested. F. magellanica and F. triphylla in 50-800 µg/mL concentrations (ethanolic extracts) and 120-1000 µg/mL concentrations (aqueous extracts) were examined. L. nummularia in 250-1500 µg/mL concentrations (ethanolic extracts) and 3000-7000 µg/mL concentrations (aqueous extracts) were investigated. The final concentration of the ethanolic solvent was restricted up to 1.5% v/v in the wells, which concentration does not affect the viability of the cells. Briefly, 3T3 and HaCaT cells were treated with various concentrations of the plant extracts for 24 h, after that ATP was measured from the cell lysates with the bioluminescence method. Nucleic acid content was analyzed with PI staining, while intracellular proteins were quantified after fluorescent derivatization with fluorescamine. All results were expressed as mean ± SD in percentage compared with data obtained for the controls (~100%). Five independent experiments were done with four technical replicates for each treatment and dose response curves were calculated from the measured data. The dose-response curves were obtained after DoseResp fitting by using the OriginLab Pro software (version 2016, OriginLab Corporation, Northampton, MA, USA).

Flow Cytometric Cytotoxicity Test
In the plate reader analysis, we could estimate the cytotoxicity in general, however, using flow cytometry it is possible to reveal the type of potential cell injury induced by the leaf extracts. For this sensitive method we used lower concentrations of the 50% (v/v) ethanolic extracts; A. vulneraria in 50, 100, 200 µg/mL, F. magellanica and F. triphylla in 2.5, 5, 10 µg/mL and L. nummularia in 10, 25, 50 µg/mL concentrations, which were presumably sub-cytotoxic based on the plate reader viability assay. Experiments were carried out in three technical replicates on a BD Canto II cytometer (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Apoptotic, necrotic and late apoptotic cells were measured using Annexin V and 7-aminoactinomycin D (7AAD) staining. Annexin V was conjugated with fluorescein isothiocyanate (FITC) or Pacific Blue (PB), 7AAD was measured on the PerCP channel. Annexin V single positivity marked the apoptotic cells, 7AAD labeled the necrotic cells, the double positive population meant the late apoptotic cells, while the double negative population was live. Since two labels were used in one sample, positivity was defined based on fluorescence minus one (FMO) controls-which were the single-stained ones. Compensation and analysis was carried out in FlowJo v.10 (FlowJo LLC., Ashland, OR, USA).

In Vitro "Wound Healing" Assay
The in vitro migration test was evaluated using culture inserts of 500 µm width (Ibidi GmbH, Gräfelfing, Germany). Briefly, an insert with 2 wells was placed in 24-well sterile culture plates, and then keratinocytes and fibroblasts were seeded into the 2 wells of the culture insert. After cell attachment and production of a monolayer, the culture insert was removed, and cells were incubated for 24 h with different sub-cytotoxic doses of 50% (v/v) ethanolic extracts. A. vulneraria in 50, 100, 200 µg/mL concentrations, F. magellanica and F. triphylla in 2.5, 5, 10 µg/mL concentrations and L. nummularia in 10, 25, 50 µg/mL concentrations were investigated. PDGF-BB was used as positive control at 15 ng/mL concentration. Within the cell-free gap the cell migration was visualized at every 4 h for 24 h by time-lapse imaging in bright field, using phase-contrast microscopy (JuLi Stage Real-Time Cell History Recorder, NanoEnTek, Seoul, Korea). The gap was monitored at an objective magnification of 10×. The closure rate of the open cell-free area was determined by quantifying the micro photo density data obtained for every occasion from the very same loci with ImageJ 1.x processing software (https://imagej.nih.gov/ij/). The closure rate in % was calculated by the following formula: Closure rate(%) = Open area 0.h − Open area x.h Open area 0.h × 100 where "Open area 0.h " is the cell-free area at the beginning of the experiment while "Open area x.h " is the still cell-free space at time points of imaging the samples. Finally, closure rate curves were constructed and the area under curve (AUC) for each treatment and cell line was calculated. The corresponding AUC data were averaged (±SD) and the summarized closure rates for the leaf extracts/PDGF were given in percentage (%) of the untreated controls. Three independent experiments were done with three technical replicates for each treatment.

Statistical Analyses
Where appropriate, data were expressed in % of the control samples, which were assumed to be~100%. In the cytotoxicity plate reader assay correlation coefficients of each tested parameter were given for the dose-response curves. Statistical evaluation was carried out in the antioxidant assays using independent t-test, where the ethanolic and aqueous extracts were compared with each other, furthermore, in the migration assay using one-way ANOVA test, where the control and the sample data of one type of treatment were compared by SPSS software (IBM, SPSS Statistics, version 22, Armonk, NY, USA). In addition, principal component analysis (PCA) was also used to test the differences between the two types of extracts of four selected medicinal plants and four chemical Antioxidants 2020, 9, 166 8 of 29 antioxidant assays. For PCA, prcomp function was used from the stats package within R (R Core Team 2019, version 3.6.1, Vienna, Austria). In all cases, the level of significance was set at p < 0.05.

Qualitative Analysis of Phenolic Compounds in Plant Extracts with LC-DAD-ESI-MS/MS
Aqueous and 50% (v/v) ethanolic extracts of A. vulneraria, F. magellanica, F. triphylla and L. nummularia were studied using LC-DAD-ESI-MS/MS methods, in order to characterize the constituents responsible for the biological actions. Eighty-two gallic acid derivatives, hydroxycinnamic acid derivatives and flavonoid glycosides were detected altogether in the samples; moreover, eight anthocyanins were described in Fuchsia samples. Compounds were tentatively characterized by comparing their chromatographic behaviors, UV spectra and mass spectrometric fragmentation patterns with data from the literature. In order to provide semi-quantitative results regarding the quantities of the constituents, their relative abundance (%) was calculated according to the summarized areas of all the compounds that were detected in the UV (280 nm) chromatogram of the sample. Results are presented in Tables 1 and 2 Flavonol glycosides prevailed in A. vulneraria samples with compounds bearing the aglycone moieties kaempferol (e.g., compounds 23 and 25) and quercetin (e.g., 40), as well as methoxylated aglycone moieties isorhamnetin (e.g., 33) and rhamnocitrin (e.g., 77). In addition, the aqueous extract comprised diverse caffeoyl (e.g., 5, 8), coumaroyl (e.g., 7, 10) and feruloyl acid derivatives (e.g., 20, 22).
The compound analyses data are summarized in dendrograms for the ethanolic and aqueous leaf extracts separately in Figures S9 and S10 (Supplementary Material).

Determination of Minimum Inhibitory Concentration (MIC 80 )
The effects of leaf extracts on Gram-positive and Gram-negative bacteria were determined by CLSI M07-A9 (Vol. 32, No. 2) guidelines (Table 3). For this study, Minimum Inhibitory Concentration (MIC 80 ) values under 100 µg/mL were considered to indicate good antimicrobial activity; from 500 to 100 µg/mL to show moderate antimicrobial activity; from 1000 to 500 µg/mL to point to weak antimicrobial activity; and over 1000 µg/mL to indicate inactivity.
The ethanolic extracts of Fuchsia species were considered to have good antibacterial activities (MIC 80

Total Antioxidant Capacity (TAC) Assays
The antioxidant properties of ethanolic and aqueous extracts of the investigated plants were evaluated by conventional chemical assays such as DPPH, TEAC, ORAC, and ECL methods (Figure 1). The lowest IC 50 values indicate the highest antioxidant effect with DPPH and TEAC scavenging activity tests, while the highest TE/g values mean the strongest antioxidant capacity in case of ORAC and ECL assays. Generally, the ethanolic extracts showed higher antioxidant effect than aqueous extracts. Altogether, the ethanolic and aqueous extracts of Fuchsia species had the strongest antioxidant activity followed by Lysimachia nummularia and Anthyllis vulneraria in all methods.
The results of Principal Component Analysis (PCA) are shown in Figure 2. In the case of ethanolic extracts, a strong positive correlation was observed between the ORAC and ECL methods, while a moderate correlation existed between the data of TEAC and DPPH assays. These relationships were opposite to the results of aqueous extracts because of their stronger correlation between TEAC and DPPH methods than ECL and ORAC tests. The ethanolic extracts of Fuchsia species were more similar to each other than in the case of their aqueous extracts.

Total Antioxidant Capacity (TAC) Assays
The antioxidant properties of ethanolic and aqueous extracts of the investigated plants were evaluated by conventional chemical assays such as DPPH, TEAC, ORAC, and ECL methods ( Figure  1). The lowest IC50 values indicate the highest antioxidant effect with DPPH and TEAC scavenging activity tests, while the highest TE/g values mean the strongest antioxidant capacity in case of ORAC and ECL assays. Generally, the ethanolic extracts showed higher antioxidant effect than aqueous extracts. Altogether, the ethanolic and aqueous extracts of Fuchsia species had the strongest antioxidant activity followed by Lysimachia nummularia and Anthyllis vulneraria in all methods.
The results of Principal Component Analysis (PCA) are shown in Figure 2. In the case of ethanolic extracts, a strong positive correlation was observed between the ORAC and ECL methods, while a moderate correlation existed between the data of TEAC and DPPH assays. These relationships were opposite to the results of aqueous extracts because of their stronger correlation between TEAC and DPPH methods than ECL and ORAC tests. The ethanolic extracts of Fuchsia species were more similar to each other than in the case of their aqueous extracts. values (in µg/mL concentration at 50% inhibition) were calculated in case of DPPH and TEAC methods, while TE/g values (Trolox equivalent in µmol referred to 1 g of initial dry material) were determined in case of ORAC and ECL tests. Mean ± SD of 5 independent experiments, each in 3 replicates. The aqueous and ethanolic extracts were compared with t-probe (** p < 0.01, *** p < 0.001).

Inhibition of Intracellular ROS Production
The oxidation of DCFH and DHR was generated by peroxyl radicals from AAPH in 3T3 and in HaCaT cells [57]. The leaf extracts decreased the fluorescence of DCF and the calculated 50% inhibition values are shown in Figure 3. Ethanolic and aqueous extracts of Fuchsia species had the strongest inhibition in both cell lines, compared with other plant extracts. However, no antioxidant activity was detected for A. vulneraria in the case of 3T3 cells, and only ethanolic extracts of A. vulneraria showed measurable antioxidant property on HaCaT cell culture.
The generation of fluorescence signal from the rhodamine derivative was also decreased by the plant extracts and the calculated 50% inhibition values are shown in Figure 4. Both solvent fractions of Fuchsia spp. exerted the highest inhibition of the fluorescence intensity of rhodamine. Although ethanolic and aqueous extracts of A. vulneraria could be quantified, the aqueous fraction had only weak effectivity (at around the detection limit).

Inhibition of Intracellular ROS Production
The oxidation of DCFH and DHR was generated by peroxyl radicals from AAPH in 3T3 and in HaCaT cells [57]. The leaf extracts decreased the fluorescence of DCF and the calculated 50% inhibition values are shown in Figure 3. Ethanolic and aqueous extracts of Fuchsia species had the strongest inhibition in both cell lines, compared with other plant extracts. However, no antioxidant activity was detected for A. vulneraria in the case of 3T3 cells, and only ethanolic extracts of A. vulneraria showed measurable antioxidant property on HaCaT cell culture.
The generation of fluorescence signal from the rhodamine derivative was also decreased by the plant extracts and the calculated 50% inhibition values are shown in Figure 4. Both solvent fractions of Fuchsia spp. exerted the highest inhibition of the fluorescence intensity of rhodamine. Although ethanolic and aqueous extracts of A. vulneraria could be quantified, the aqueous fraction had only weak effectivity (at around the detection limit).

Inhibition of Intracellular ROS Production
The oxidation of DCFH and DHR was generated by peroxyl radicals from AAPH in 3T3 and in HaCaT cells [57]. The leaf extracts decreased the fluorescence of DCF and the calculated 50% inhibition values are shown in Figure 3. Ethanolic and aqueous extracts of Fuchsia species had the strongest inhibition in both cell lines, compared with other plant extracts. However, no antioxidant activity was detected for A. vulneraria in the case of 3T3 cells, and only ethanolic extracts of A. vulneraria showed measurable antioxidant property on HaCaT cell culture.
The generation of fluorescence signal from the rhodamine derivative was also decreased by the plant extracts and the calculated 50% inhibition values are shown in Figure 4. Both solvent fractions of Fuchsia spp. exerted the highest inhibition of the fluorescence intensity of rhodamine. Although ethanolic and aqueous extracts of A. vulneraria could be quantified, the aqueous fraction had only weak effectivity (at around the detection limit).

Plate Reader Cytotoxicity Tests
Cytotoxicity data obtained for fibroblast (3T3) and keratinocyte cell cultures (HaCaT) are seen in Figure 5. The ethanolic extracts reduced the ATP, cell number and protein contents of both cell lines in a dose dependent manner more effectively than the aqueous extracts. Nevertheless, these differences between the type of solvents were not observed in case of Fuchsia species. Both ethanolic and aqueous extracts of Fuchsias had the same toxic effects in the investigated cell lines. In 3T3 cells, the decrease of ATP, cell number and protein levels were more pronounced than in HaCaT cultures.
Evaluating the results of compound analyses, antimicrobial, antioxidant and cytotoxicity data we decided to use only the ethanolic extracts in the further experiments because of their richer ingredient content and stronger biological effects. Potentially sub-toxic concentrations of the ethanolic extracts (A. vulneraria in 50, 100, 200 µg/mL, F. magellanica and F. triphylla in 2.5, 5, 10 µg/mL and L. nummularia in 10, 25, 50 µg/mL concentrations) were used in all further investigations.

Plate Reader Cytotoxicity Tests
Cytotoxicity data obtained for fibroblast (3T3) and keratinocyte cell cultures (HaCaT) are seen in Figure 5. The ethanolic extracts reduced the ATP, cell number and protein contents of both cell lines in a dose dependent manner more effectively than the aqueous extracts. Nevertheless, these differences between the type of solvents were not observed in case of Fuchsia species. Both ethanolic and aqueous extracts of Fuchsias had the same toxic effects in the investigated cell lines. In 3T3 cells, the decrease of ATP, cell number and protein levels were more pronounced than in HaCaT cultures.
Evaluating the results of compound analyses, antimicrobial, antioxidant and cytotoxicity data we decided to use only the ethanolic extracts in the further experiments because of their richer ingredient content and stronger biological effects. Potentially sub-toxic concentrations of the ethanolic extracts (A. vulneraria in 50, 100, 200 µg/mL, F. magellanica and F. triphylla in 2.5, 5, 10 µg/mL and L. nummularia in 10, 25, 50 µg/mL concentrations) were used in all further investigations. Dose-response curves were created by log10 transformation and nonlinear curve fitting. Correlation coefficients (R 2 ) were calculated for each treatment.

Flow Cytometric Cytotoxicity Test
Our apoptosis-necrosis results for the hypothesized nontoxic concentrations of ethanolic extracts are summarized for 3T3 cells in Table 4 and for HaCaT cells in Table 5. The two different cell lines gave similar results, namely the applied concentrations of the leaf extracts did not cause significant apoptosis/necrosis and most of the cells remained intact (the number of dead/dying cells was under 4% in both cell lines). An example of the gating strategy can be seen in the Supplementary Materials section ( Figure S11). Table 4. Apoptosis-necrosis assay data of 3T3 cells treated with 50% (v/v) ethanolic extracts of A. vulneraria, F. magellanica, F. triphylla and L. nummularia with Annexin V-7AAD flow cytometric method.

Treatment Groups
Annexin V-7AAD Method

In vitro Migration Test
To investigate the impact of the ethanolic leaf extracts on the migration of fibroblasts and keratinocytes a standardized 500 µm cell-free area was created in 24-well culture plates by Teflon inserts. Figure 6 shows the effects of A. vulneraria and L. nummularia extracts while Figure 7 those of F. magellanica and F. triphylla on the migration of 3T3 and HaCaT cells.
The positive control PDGF-BB at 15 ng/mL concentration had a strong significant stimulating effect on cell migration (117.05 ± 6.72% on 3T3 and 115.16 ± 8.26% on HaCaT cells), compared with untreated control cells. Of the tested four plants, A. vulneraria expressed a slight stimulatory effect (107.01 ± 7.35% in 100 µg/mL concentration and 104.54 ± 8.86% in 200 µg/mL concentration) only on HaCaT cells, while enhanced migration and closure rate were observed in F. magellanica and F. triphylla treated cells when compared with untreated cells. Moreover, these extracts reached the response of the positive control, PDGF. Of the two Fuchsia species, F. magellanica had the strongest incentive effects on both cell lines, because the closure rate was 120.26 ± 10.17% on 3T3 cells and 114.61 ± 3.72% on HaCaT cells, respectively, in 2.5 µg/mL concentration. L. nummularia did not show any significant migration effect towards the two cell lines.
On the other hand, it is noteworthy that the less concentrated leaf extracts produced the most efficient stimulation on cell migration when compared with extracts of higher concentrations.

In vitro Migration Test
To investigate the impact of the ethanolic leaf extracts on the migration of fibroblasts and keratinocytes a standardized 500 µm cell-free area was created in 24-well culture plates by Teflon inserts. Figure 6 shows the effects of A. vulneraria and L. nummularia extracts while Figure 7 those of F. magellanica and F. triphylla on the migration of 3T3 and HaCaT cells.
The positive control PDGF-BB at 15 ng/mL concentration had a strong significant stimulating effect on cell migration (117.05 ± 6.72% on 3T3 and 115.16 ± 8.26% on HaCaT cells), compared with untreated control cells. Of the tested four plants, A. vulneraria expressed a slight stimulatory effect (107.01 ± 7.35% in 100 µg/mL concentration and 104.54 ± 8.86% in 200 µg/mL concentration) only on HaCaT cells, while enhanced migration and closure rate were observed in F. magellanica and F. triphylla treated cells when compared with untreated cells. Moreover, these extracts reached the response of the positive control, PDGF. Of the two Fuchsia species, F. magellanica had the strongest incentive effects on both cell lines, because the closure rate was 120.26 ± 10.17% on 3T3 cells and 114.61 ± 3.72% on HaCaT cells, respectively, in 2.5 µg/mL concentration. L. nummularia did not show any significant migration effect towards the two cell lines.
On the other hand, it is noteworthy that the less concentrated leaf extracts produced the most efficient stimulation on cell migration when compared with extracts of higher concentrations.

Discussion
Our investigated medicinal plants are used in Transylvanian folk medicine for the treatment of various skin diseases. Of the tested plants, we could not find scientific studies to evaluate their biological effects, especially on cell migration and proliferation. Although we have some preliminary results of antioxidant effects and phytochemical data of the extracts of Anthyllis vulneraria and Fuchsia species [58], to our best knowledge, the present study is the first to analyze the cytotoxicity and combined effects of the ethanolic extracts in a cellular migration model that mimics wound healing ability.
Bioactive agents of plants may have specific functions on wound healing properties, including antioxidant, antimicrobial, anti-inflammatory effects via migration, proliferation and pro-collagen stimulating actions and they can modulate one or more phases of the wound healing process. For instance, tannins and flavonoids have anti-inflammatory and antibiotic effects [59,60]. Flavonoids and phenolic acids are well-known antioxidant compounds, and the position and degree of hydroxylation are essential to their activity. In flavonoids, O-dihydroxy structure on the B ring, 2,3 double bond in conjugation with a 4-oxofunction on the C ring, the presence of a 3-hydroxyl group on C ring and C3-and C5-OH moieties in a combination with 4-oxofunction on the A and C rings are required for maximum radical scavenging potential [2,61]. In phenolic acids, the presence of a 3-hydroxyl structure (2,3-hydroxybenzoic acid, gallic acid, caffeic acid and caftaric acid) enhances the antioxidant effect [2,62]. Anthocyanins are strong antioxidant compounds and their scavenging activity corresponds with their structural characteristics as well, di-acylated forms possess higher antioxidant activity than the mono-acylated and non-acylated anthocyanins [63]. Sun et al. demonstrated that peonidin and cyanidin-based anthocyanins exert strong antioxidant properties. Furthermore, they verified, that acylated forms have high antibacterial effect [64]. On the other hand, anthocyanins from various berries have a strong anti-inflammatory effect via inhibition of IL-1β, Il-6 and COX-2 gene expression. Moreover, these colorful compounds stimulate the wound healing with increasing the migration of fibroblasts and keratinocytes [65,66].

Discussion
Our investigated medicinal plants are used in Transylvanian folk medicine for the treatment of various skin diseases. Of the tested plants, we could not find scientific studies to evaluate their biological effects, especially on cell migration and proliferation. Although we have some preliminary results of antioxidant effects and phytochemical data of the extracts of Anthyllis vulneraria and Fuchsia species [58], to our best knowledge, the present study is the first to analyze the cytotoxicity and combined effects of the ethanolic extracts in a cellular migration model that mimics wound healing ability.
Bioactive agents of plants may have specific functions on wound healing properties, including antioxidant, antimicrobial, anti-inflammatory effects via migration, proliferation and pro-collagen stimulating actions and they can modulate one or more phases of the wound healing process. For instance, tannins and flavonoids have anti-inflammatory and antibiotic effects [59,60]. Flavonoids and phenolic acids are well-known antioxidant compounds, and the position and degree of hydroxylation are essential to their activity. In flavonoids, O-dihydroxy structure on the B ring, 2,3 double bond in conjugation with a 4-oxofunction on the C ring, the presence of a 3-hydroxyl group on C ring and C3-and C5-OH moieties in a combination with 4-oxofunction on the A and C rings are required for maximum radical scavenging potential [2,61]. In phenolic acids, the presence of a 3-hydroxyl structure (2,3-hydroxybenzoic acid, gallic acid, caffeic acid and caftaric acid) enhances the antioxidant effect [2,62]. Anthocyanins are strong antioxidant compounds and their scavenging activity corresponds with their structural characteristics as well, di-acylated forms possess higher antioxidant activity than the mono-acylated and non-acylated anthocyanins [63]. Sun et al. demonstrated that peonidin and cyanidin-based anthocyanins exert strong antioxidant properties. Furthermore, they verified, that acylated forms have high antibacterial effect [64]. On the other hand, anthocyanins from various berries have a strong anti-inflammatory effect via inhibition of IL-1β, Il-6 and COX-2 gene expression. Moreover, these colorful compounds stimulate the wound healing with increasing the migration of fibroblasts and keratinocytes [65,66].
We found that in A. vulneraria the most abundant amount of flavonol compounds are in glycosidic or methoxylated forms with lower activities than the aglycones [2]. In contrast, Fuchsia species contained, besides common flavonol compounds, various cinnamic acid and benzoic acid derivatives such as caffeic acid, ellagic acid and gallic acid derivatives, kaempferol-and quercetin-galloyl-glycosides. Moreover, the leaves of F. magellanica and F. triphylla contained several anthocyanins, such as cyanidin and peonidin derivatives. In the leaf extracts of L. nummularia, the main flavonoids were myricetin derivatives, which have strong antibacterial effects; however, in spite of this observation we detected weaker activities, compared to the standard antibiotic control and also to Fuchsia species [67]. The main differences between the investigated plants are that Fuchsia species contain high amounts of anthocyanins and hydroxybenzoic acids, which are well-known antioxidants and antimicrobial constituents; moreover, it has been proved, that they have a stimulating effect on cell migration [2,24,67].
Nevertheless, there are some contradictions in the literature regarding cell proliferation and migration effects of phenolic compounds, which act through different signaling pathways to express their effect in various cell lines. The polyphenols of tea suppress the migration of proliferation abilities of tumor cells by inhibiting NF-κB activation and quenching the expression of cyclin D1 [68]. The other compound carnosol inhibits the migration and tumor growth via ROS-dependent proteasome degradation of STAT3 in several breast cancer cell lines [69]. LaFoya et al. investigated that some flavonoids, including resveratrol, apigenin, chrysin, genistein, luteolin, and myricetin, significantly inhibited the endothelial cell migration and proliferation in a statistically significant manner with the moderation of Notch signaling, while quercetin did not regulate these processes [70]. Several studies demonstrated that phenolic acids, such as caffeic acid, coumaric acid and ferulic acid, which belong to hydroxycinnamic acids, have anti-proliferative and apoptotic effects on tumor cells [71][72][73][74]. On the other hand, the caffeic acid phenethyl ester can stimulate wound re-epithelization and enhance the proliferation of keratinocytes [75]. Ly et al. have proved that chlorogenic acid has migration and invasion stimulating effects on trophoblasts through an adenosine monophosphate-activated protein (AMP) kinase-dependent pathway, while not affecting cell proliferation [76]. However, the hydroxybenzoic acids, such as ellagic acid and gallic acid, increase the proliferation and migration of keratinocytes and fibroblasts in a dose-dependent manner [77].
During the past decades, a wide variety of analytical methods has been developed for the measurement of the total antioxidant capacity [78]. Several studies found good linear correlation between the total phenolic and/or flavonoid content and antioxidant capacities of plants. In agreement with other researchers, we found correlations between the TEAC and DPPH assay as well as for ORAC and ECL methods [31,[79][80][81]. On the other hand, these differences between antioxidant methods can be explained by their chemical backgrounds. The underlying mechanism of ECL and ORAC assays is the hydrogen atom transfer (HAT), while TEAC and DPPH tests are based on single electron transfer (SET) reaction [81]. The assessment of total antioxidant capacity by cell-free chemical assays cannot completely reflect the behavior of the complex plant samples in vivo. Therefore, it is important to evaluate the effectiveness of antioxidants in more biologically relevant conditions, such as testing of the compounds in cell-based antioxidant assays [82]. The two used fluorogenic reporter molecules are sensitive to peroxyl radicals with similar oxidative mechanisms, therefore, DHR and DCFH gave closely matching scavenging activities as was expected from literature data [57].
In wound healing assays it is important to differentiate between proliferation and migration because the cells are usually not synchronized. Therefore, some authors use mitomycin C DNA synthesis inhibitor pretreatment to follow the migration solely when extended (48 h) incubation time is applied because it is longer than the generation cycle of the cells [83]. In our tests, we did not apply this compound because our incubation time was only 24 h, and the proliferation ability is minimal within this short period. HaCaT cell culture is an immortalized human keratinocyte cell line, with a doubling time of approximately 24 h (36.2 ± 1.5 h in the early passages (2-8) and 24 ± 0.6 h in the late passages (12)(13)(14)(15)(16) [84], while 3T3 mouse fibroblast cells have a doubling rate of 20-26 h [85].
In our migration tests, nontoxic doses of each extract were applied, where the fibroblast and keratinocyte cells showed more than 90% cell viability by the plate reader assay, and the necrotic (dead) cells were under 4% in the flow cytometric assay. 3T3 fibroblast cell culture was more vulnerable to exposure of the leaf extracts than HaCaT keratinocytes. One possible explanation of the increased sensitivity to potentially cytotoxic plant compounds might be that fibroblasts with the epidermal phenotype are located in the dermis, while keratinocytes are mostly found in the basement membrane zone (epidermis); therefore they should be more resistant against external noxae [86]. Our phytochemical and biological data suggest that Fuchsia species exerted the highest positive effect on cell migration among the four tested plants.
The found antimicrobial IC 50 values of Fuchsia species were lower in the antibacterial assay than IC 50 cytotoxicity values found for the fibroblasts and keratinocytes. Therefore, one may conclude that Fuchsia species extracts may possess effective components for the formulation of new antibacterial and wound healing stimulatory agents.
In our planned further experiments, another powerful approach for the characterization of the mode of action of the test compounds (i.e., polyphenols mediating the pathways of wound healing via modulation of various biomarkers such as growth factors, matrix metalloproteinases and cytokines) could be the measurement of secreted factors from the supernatant or the analysis of intracellular signaling pathways by Western blot and molecular biological methods. These studies are inevitably necessary for identifying the impact of complex plant extracts in wound healing processes [87,88].
Supplementary Materials: The following are available online at http://www.mdpi.com/2076-3921/9/2/166/s1. Figure S1: HPLC-DAD chromatogram of A. vulneraria 50% (v/v) ethanolic extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Table 1. Figure S2: HPLC-DAD chromatogram of A. vulneraria aqueous extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Table 1. Figure S3: HPLC-DAD chromatogram of F. magellanica 50% (v/v) ethanolic extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Tables 1 and 2. Figure S4: HPLC-DAD chromatogram of F. magellanica aqueous extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Tables 1 and 2. Figure S5: HPLC-DAD chromatogram of F. triphylla 50% (v/v) ethanolic extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Tables 1 and 2. Figure S6: HPLC-DAD chromatogram of F. triphylla aqueous extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Tables 1 and 2. Figure S7: HPLC-DAD chromatogram of L. nummularia 50% (v/v) ethanolic extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Table 1. Figure S8: HPLC-DAD chromatogram of L. nummularia aqueous extract, detection wavelength: 280 nm. Numbering of peaks refers to data shown in Table 1. Figure S9: Dendrogram of the summarized compound analysis data of 50% (v/v) ethanolic plant extracts. Figure S10: Dendrogram of the summarized compound analysis data of aqueous plant extracts. Figure S11: Gating strategy of the flow cytometry experiments.