Proanthocyanidin Characterization, Antioxidant and Cytotoxic Activities of Three Plants Commonly Used in Traditional Medicine in Costa Rica: Petiveria alliaceae L., Phyllanthus niruri L. and Senna reticulata Willd.

The phenolic composition of aerial parts from Petiveria alliaceae L., Phyllanthus niruri L. and Senna reticulata Willd., species commonly used in Costa Rica as traditional medicines, was studied using UPLC-ESI-TQ-MS on enriched-phenolic extracts. Comparatively, higher values of total phenolic content (TPC), as measured by the Folin-Ciocalteau method, were observed for P. niruri extracts (328.8 gallic acid equivalents/g) than for S. reticulata (79.30 gallic acid equivalents/g) whereas P. alliaceae extract showed the lowest value (13.45 gallic acid equivalents/g). A total of 20 phenolic acids and proanthocyanidins were identified in the extracts, including hydroxybenzoic acids (benzoic, 4-hydroxybenzoic, gallic, prochatechuic, salicylic, syringic and vanillic acids); hydroxycinnamic acids (caffeic, ferulic, and p-coumaric acids); and flavan-3-ols monomers [(+)-catechin and (−)-epicatechin)]. Regarding proanthocyanidin oligomers, five procyanidin dimers (B1, B2, B3, B4, and B5) and one trimer (T2) are reported for the first time in P. niruri, as well as two propelargonidin dimers in S. reticulata. Additionally, P. niruri showed the highest antioxidant DPPH and ORAC values (IC50 of 6.4 μg/mL and 6.5 mmol TE/g respectively), followed by S. reticulata (IC50 of 72.9 μg/mL and 2.68 mmol TE/g respectively) and P. alliaceae extract (IC50 >1000 μg/mL and 1.32 mmol TE/g respectively). Finally, cytotoxicity and selectivity on gastric AGS and colon SW20 adenocarcinoma cell lines were evaluated and the best values were also found for P. niruri (SI = 2.8), followed by S. reticulata (SI = 2.5). Therefore, these results suggest that extracts containing higher proanthocyanidin content also show higher bioactivities. Significant positive correlation was found between TPC and ORAC (R2 = 0.996) as well as between phenolic content as measured by UPLC-DAD and ORAC (R2 = 0.990). These findings show evidence for the first time of the diversity of phenolic acids in P. alliaceae and S. reticulata, and the presence of proanthocyanidins as minor components in latter species. Of particular relevance is the occurrence of proanthocyanidin oligomers in phenolic extracts from P. niruri and their potential bioactivity.


Introduction
Characterization and quantification of secondary metabolites and their bioactivity studies are essential to increase the knowledge on plants with traditional medicinal uses towards efficient and safe utilization.Three plant species of special relevance because of their widespread utilization in traditional medicine in Costa Rica as anti-inflammatories are Petiveria alliaceae, Phyllanthus niruri and Senna reticulata [1].Senna reticulata is a shrub belonging to the Fabaceae family, originally from Mesoamerica and South America (to Brazil), whose leaves and stems are used traditionally for rheumatism and skin conditions [2].Phyllanthus niruri is a shrub native to America and tropical areas of India and China, which belongs to the Phyllanthaceae family, and has diuretic and hepatoprotective properties [3].Finally, Petiveria alliacea, which belongs to the Phytholacaceae family, is distributed from the South of the United States to Brazil and its traditional uses include analgesic, anticoagulant and hypoglycemic properties [4].
Studies have attributed anti-inflammatory and other bioactivities mainly to anthraquinones present in S. reticulate [2], sulfur containing compounds in P. alliaceae [4], and lignans such as phyllanthosides in the case of P. niruri.More recent studies have shown the synergic effect of different families of phenolic compounds in bioactivity studies, for instance showing the important effects of proanthocyanidins on antioxidant and cytotoxic bioactivities [5,6].Regarding phenolic compounds, mainly flavonoids such as kaempferol derivatives have been reported in S. reticulata [2] [7]; quercetin derivatives in P. alliacea [8] and P. niruri [9]; whereas caffeic acid derivatives and ellagitannins have also been reported in P. niruri [10].Flavan-3-ols, including catechin and epicatechin monomers have been reported in P. niruri [11], whereas no detailed studies on proanthocyanidin oligomers have been performed in any of these three species.
Proanthocyanidins are condensed flavan-3-ols that constitute an important group of polyphenols because of their bioactivities, among others, ant-inflammatory, antioxidant and anti-cancer activities [12].Despite the increasing number of studies on phenolics, the characterization of proanthocyanidins remains a complex task because of the need for high-end techniques such as High-Resolution Mass Spectrometry (HRMS).On the other hand, it has been argued that these bioactivities could be mediated by redox interaction, since the regulation on redox homeostasis has been implicated in the control of the transition from cell proliferation to cell differentiation and cell cycle progression in both plants and animals [13].However, the mechanisms and factors that could affect these bioactivities remain to be elucidated [14], suggesting the importance of these studies.
Since proanthocyanidin composition of P. alliaceae, P. niruri and S. reticulata have been scarcely studied and because of findings demonstrating the synergic contribution of proanthocyanidins on plants whose bioactivities were attributed solely to other metabolites [5,15], the objective of this work was to obtain phenolic extracts from these three plant species and to characterize them UPLC-DAD-ESI-TQ-MS.Evaluation of the antioxidant activity through DPPH and ORAC methods, as well as the assessment of cytotoxicity in AGS adenocarcinoma gastric cells, SW620 adenocarcinoma colon cells, and Vero normal cells, was also carried out in the different extracts.

Phenolic Yield and Total Phenolic Contents
The extraction process described in the Materials and Methods section, allowed the phenolic enriched fractions to be obtained, as summarized in Table 1. S. reticulata presented the highest yield (6.53%) whereas P. alliaceae showed the lowest value (5.03%).The total phenolic contents (TPC) shown in Table 1, also resulted in comparatively lower values for P. alliaceae extract (13.45 gallic acid equivalents/g dry extract) than P. niruri extract (328.80 gallic acid equivalents/g dry extract), which exhibited the highest values.These results agree with few reports indicating lower TPC values for an hydroalcoholic extract of P. alliaceae [16], and for an aqueous extract of S. reticulata [17].However, higher TPC values, in the range of 263-270 gallic acid equivalents/g, which are slightly lower than our findings have been reported for ethanolic extracts of P. niruri [10,18].Table 1 also summarizes the total proanthocyanidin (PRO) content for the different extracts.P. niruri showed the highest PRO content (322.23 cyanidin chloride equivalents/g dry extract) whereas no content was found in P. alliaceae.S. reticulata showed intermediate values for both TP and PRO contents.Thus, phenolic content varied according to plant species, the highest values for both TPC and PRO clearly corresponding to P. niruri.

Sample
Extraction Yield (%) 1Total Phenolic Content (TPC) (mg/g) 1 g of extract/g of dry material expressed as %. 2 mg of gallic acid equivalent (GAE)/g extract. 3mg of cyanidin chloride equivalent (CCE)/g extract. 4nd = not detected. 5Different superscript letters in the same column indicate differences are significant at p < 0.05.

Phenolic Profile by UPLC-DAD-ESI-TQ-MS Analysis
Table 2 summarizes the results of UPLC-DAD-ESI-TQ-MS analysis performed in the different extracts, as described in the Materials and Methods section.Also, Figure S1 and Table S1 (Supplementary Materials) show base chromatograms and main MS/MS parameters respectively.Among the 28 phenolic compounds screened, a total of 20 phenolic compounds were characterized and quantified, (benzoic, 4-hydroxybenzoic, gallic, protocatechuic, salicylic, syringic and vanillic acids); hydroxycinnamic acids (caffeic, ferulic, and p-coumaric acids); and flavan-3-ols monomers [(+)-catechin and (−)-epicatechin)], procyanidin dimers, propelargonidin dimers and procyanidin trimers).To our knowledge, our findings report for the first time the presence of proanthocyanidin oligomers (Figure 1) in extracts from P. niruri and S. reticulata.equivalents/g dry extract) than P. niruri extract (328.80 gallic acid equivalents/g dry extract), which exhibited the highest values.These results agree with few reports indicating lower TPC values for an hydroalcoholic extract of P. alliaceae [16], and for an aqueous extract of S. reticulata [17].However, higher TPC values, in the range of 263-270 gallic acid equivalents/g, which are slightly lower than our findings have been reported for ethanolic extracts of P. niruri [10,18].Table 1 also summarizes the total proanthocyanidin (PRO) content for the different extracts.P. niruri showed the highest PRO content (322.23 cyanidin chloride equivalents/g dry extract) whereas no content was found in P. alliaceae.S. reticulata showed intermediate values for both TP and PRO contents.Thus, phenolic content varied according to plant species, the highest values for both TPC and PRO clearly corresponding to P. niruri.  2 summarizes the results of UPLC-DAD-ESI-TQ-MS analysis performed in the different extracts, as described in the Materials and Methods section.Also, Figure S1 and Table S1 (Supplementary Materials) show base chromatograms and main MS/MS parameters respectively.Among the 28 phenolic compounds screened, a total of 20 phenolic compounds were characterized and quantified, (benzoic, 4-hydroxybenzoic, gallic, protocatechuic, salicylic, syringic and vanillic acids); hydroxycinnamic acids (caffeic, ferulic, and p-coumaric acids); and flavan-3-ols monomers [(+)-catechin and (−)-epicatechin)], procyanidin dimers, propelargonidin dimers and procyanidin trimers).To our knowledge, our findings report for the first time the presence of proanthocyanidin oligomers (Figure 1) in extracts from P. niruri and S. reticulata.In fact, five different procyanidin dimers, namely B1, B2, B3, B4 and B5 were found in P. niruri extract, two propelargonidin dimers (with retention times of 5.03 and 5.63 min) were detected in S. reticulata, whereas procyanidin trimer T was determined in P. niruri.On the other hand, flavan-3-ols monomers, namely (+)-catechin and (−)-epicatechin, were found in both P. niruri and S. reticulata, constituting to our knowledge, the first report for both monomers in this latter plant species.In contrast, proanthocyanidins were not detected in P. alliaceae extract.
The total UPLC contents were in agreement with the total phenolics (TPC) as measured by the Folin-Ciocalteau (Tables 1 and 2).For instance, P. niruri showed the higher contents by both methods (TPC of 328.8 mg GAE/g extract and UPLC value of 1909.9 µg/g extract), followed by S. reticulata which exhibited intermediate values (TPC of 72.3 mg GAE/g extract and UPLC value of 708.8 µg/g extract), and finally by P. alliacea (TPC of 13.45 mg GAE/g extract and UPLC value of 473.0 µg/g extract).
Finally, the total proanthocyanidin (PRO) contents was also in agreement with the UPLC findings.For instance, P. niruri showed the highest PRO value by both methods (322.93 mg CCE/g extract and UPLC value of 794.5 µg/g extract), followed by S. reticulata (PRO of 22.35 mg CCE/g extract and UPLC value of 25.6 µg/g extract), whereas no proanthocyanidin content were obtained by either method for P. alliaceae.
Finally, the higher antioxidant values found for P. niruri are in agreement with reports from the literature using DPPH method.For instance, studies on aqueous extracts report IC 50 values of 15.3 µg/mL [3] and 6.85 µg/mL [18], whereas IC 50 values of 9.1 µg/mL and 11.07 µg/mL were reported for methanolic [3] and ethanolic [18] extracts respectively.However, other studies reported lower antioxidant activity for a hydro-alcoholic extract (IC 50 = 32.64 µg/mL) [10].Therefore, our DPPH results for P. niruri are better than previous studies.
The difference in antioxidant values among P. alliaceae, S. reticulata and P. niruri extracts could be attributed to the differences in their phenolic content and distribution.Therefore, in order to investigate if the phenolic composition contributes to the antioxidant activity, a correlation analysis was carried out between DPPH and ORAC values, and the total phenolic contents (TPC, Table 1), as well as with the content by UPLC (Table 3).No correlation was found for DPPH, however, a significant and positive correlation was observed between TPC and ORAC values (R 2 = 0.996) and between UPLC contents and ORAC values (R 2 = 0.990).Therefore, our results are in agreement with previous studies reporting correlation between total polyphenolic contents and ORAC antioxidant activity [17].

Cytotoxicity of P. alliaceae, P. niruri and S. reticulata Extracts
Table 4 summarizes the IC 50 values for the cytotoxicity of P. alliaceae, P. niruri and S. reticulata extracts on AGS human gastric adenocarcinoma, SW620 human colon adenocarcinoma and Vero monkey normal epithelial kidney cell lines.Also, Figure 2 shows dose-response curves.IC 50 values indicate that there is no significant difference (one-way analysis of variance (ANOVA) followed by Tukey's post hoc test) between cytotoxicity values (p < 0.05) against gastric AGS adenocarcinoma cells and SW620 adenocarcinoma cells for P. alliacea and S. reticulata extracts.However, in the case of P. niruri extract, the ANOVA indicates that the cytotoxicity results are dependent on the cancer cell line.208.4 c, * ± 8.9 (SI = 2.4) 202.5 b, * ± 9.1 (SI = 2.5) >500 c,+ 1 Different superscript letters in the same column indicate differences are significant at p < 0.05. 2 Different superscript signs in the same row indicate differences are significant at p < 0.05.
Our results for the cytotoxicity of P. alliacea on both adenocarcinoma cell lines are similar to those reported for an ethanolic extract on Jurkat T cells [23].Other studies indicate variability of results, with IC 50 values ranging from 29 to 36 µg/mL for a C-18 chromatographic fraction of a hydro-alcoholic extract on erythroleukema and melanoma cell lines [24], and breast adenocarcinoma 4T1 cell line [25].In contrast, no cytotoxic effect was observed for a methanolic extract on hepatic adenocarcinoma HepG2 [26].
For S. reticulata extracts, the results for the cytotoxicity are in agreement with a study reporting IC 50 of 232.9 µg/mL for a methanolic extract on KB nasopharyngeal cells.However, no cytotoxicity was found in aqueous extracts [27].Studies on other species of the genus Senna, including a methanolic extract of S. covesii, also reported no activity (IC 50 > 800 µg/mL) on L929 tumor connective tissue cell line, HeLa cervix carcinoma and C3F6 lymphoma [28].
Finally, the cytotoxicity results obtained for the P. niruri extract on AGS tumor gastric cell lines and SW620 tumor colon cell lines, are similar to those found in the literature for methanolic and aqueous extracts on PC-3 prostate cancer cell lines (IC 50 values of 117.7 and 155.0 µg/mL respectively), MeWo skin cancer cell lines (IC 50 values of 153.3 and 193.3 µg/mL respectively) [29], and for an hydro-methanolic extract on MCF-7 human breast carcinoma cell lines (IC 50 of 84.88 µg/mL) [30].Similar to our cell line-dependence findings for P. niruri phenolic extracts, other studies performed on hydro-methanolic extracts of P. amarus and P. virgatus report lower cytotoxicity on Hep G2 hepatic carcinoma (IC 50 > 250 µg/mL) [31].Concerning the selectivity of the cytotoxic activity of extracts against cancerous cell in normal Vero cells, our results indicate significant difference (ANOVA, p < 0.05) between IC50 values for both AGS and SW620 adenocarcinoma cell lines and normal Vero cells.When comparing selectivity index (SI), defined as the ratio of IC50 values of normal (Vero) cells to cancer cells (AGS or SW620), P. alliaceae extract showed the lowest values on both cell lines (SI = 1.4), while P. niruri extract showed the best selectivity result for SW620 colon cancer cells (SI = 2.8), which is in agreement with the selectivity range (SI ≥ 3) reported as promising for further anticancer studies [32,33].In turn, S. reticulata extract exhibited a slightly lower value on SW620 colon cancer cells (SI = 2.5) and on AGS gastric cancer cells (SI = 2.4), followed by P. niruri extract on SW620 cells (SI = 2.2).This variability has been reported in previous studies, showing selective cytotoxicity of a polyphenolic extract Concerning the selectivity of the cytotoxic activity of extracts against cancerous cell in normal Vero cells, our results indicate significant difference (ANOVA, p < 0.05) between IC 50 values for both AGS and SW620 adenocarcinoma cell lines and normal Vero cells.When comparing selectivity index (SI), defined as the ratio of IC 50 values of normal (Vero) cells to cancer cells (AGS or SW620), P. alliaceae extract showed the lowest values on both cell lines (SI = 1.4), while P. niruri extract showed the best selectivity result for SW620 colon cancer cells (SI = 2.8), which is in agreement with the selectivity range (SI ≥ 3) reported as promising for further anticancer studies [32,33].In turn, S. reticulata extract exhibited a slightly lower value on SW620 colon cancer cells (SI = 2.5) and on AGS gastric cancer cells (SI = 2.4), followed by P. niruri extract on SW620 cells (SI = 2.2).This variability has been reported in previous studies, showing selective cytotoxicity of a polyphenolic extract between gastric normal and cancer cells [34], whereas other studies report no selectivity on the cytotoxic effect in normal, as well as in cancer breast and prostate cell lines [35].
In fact, mechanisms related to the effect of polyphenols in cancer cells need to be elucidated, taking into account, for instance, the multiplicity of targets that can be reached [14] and the complexity of factors modulating cancer cell phenotypes [36].However, there is enough evidence to support the potential anticancer effects of proanthocyanidins [12].For instance, previous studies using proanthocyanidins from grape seeds showed a cytotoxic effect in cervix cancer [37], whereas cranberries polyphenols inhibited proliferation in prostate and colon cancer cells [38].Also, other studies indicated the potential of propelargonidin dimers and procyanidin dimers and contents on higher and more selective cytotoxicity in gastric and colon cancer cell lines [6].In addition, a report on another main component of P. niruri extract, gallic acid, indicates that this compound inhibits the growth of human hepatocellular carcinoma cells and induces apoptosis in these cell lines [39].Since polyphenols could work in a synergistic manner, fractioning of S. reticulata and P. niruri extracts would contribute to further elucidate the structure-bioactivity relationship of these plant extracts.

Materials, Reagents and Solvents
Petiveria alliaceae, Phyllanthus niruri and Senna reticulata aerial plant material were acquired from local communities that grow and use the plants as traditional medicine, grouped as an Agricultural Producers Association (AMPALEC) in the Caribbean region.All plants were identified with the support of the Costa Rican National Herbarium and vouchers are deposited there.The material of each plant was rinsed in water and cut into small pieces.Subsequently, it was dried in a stove at 40 • C until completely dry, and after being ground, it was preserved at −5 • C. Reagents such as AAPH, fluorescein, DPPH, and Trypsin-EDTA solution were provided by Sigma-Aldrich (St. Louis, MO, USA), while amphotericin B, penicillin-streptomycin, and Minimum essential Eagle's medium (MEM, 10% fetal bovine serum), were purchased from Life Technologies (Carlsbad, CA, USA).AGS human gastric adenocarcinoma, SW 620 human colorectal adenocarcinoma and Vero monkey normal epithelial kidney cell lines were obtained from American Type Culture Collection (ATCC, Rockville, MD, USA).DMSO was provided by Sigma-Aldrich (St. Louis, MO, USA), while MTBE, chloroform and methanol were purchased from Baker (Center Valley, PA, USA).

Phenolic Extracts from P. alliaceae, P. niruri and S. reticulata
The process followed for obtaining phenolic-enriched extracts was formerly described by our group [40].Briefly, dried material from each plant was first extracted in a mixture of methyl ter-butyl ether (MTBE) and methanol (MeOH) 90:10 (v/v) at 25 • C during 30 min in ultrasound.Afterwards it was extracted for 24 h in order to obtain a non-polar extract of the samples.After filtration, the extraction was repeated once.The extracts were combined and the solvents evaporated in a rotavapor to dryness and subsequently washed with MeOH in order to extract any polyphenols.The residual plant material was extracted with MeOH at 25 • C during 30 min in ultrasound, and then extracted for 24 h.After filtration, the extraction was repeated twice.The three methanol extracts were combined with the previous MeOH washings and were evaporated in a rotavapor to dryness.Finally, the dried extract was washed with hexane, MTBE and chloroform consecutively in order to obtain a phenolic rich-extract.

Total Phenolic Content
The polyphenolic content was determined by a modification of the Folin-Ciocalteu (FC) method [41], whose reagent is composed of a mixture of phosphotungstic and phosphomolybdic acids.Each sample was dissolved in MeOH (0.1% HCl) and combined with 0.5 mL of FC reagent.Afterwards 10 mL of Na 2 CO 3 (7.5%)were added and the volume was completed to 25 mL with water.Blanks were prepared in a similar way but using 0.5 mL of MeOH (0.1% HCl) instead of sample.The mixture was let standing in the dark for 1 h and then absorbance was measured at 750 nm.Values obtained were extrapolated in a gallic acid calibration curve.Total phenolic content was expressed as mg gallic acid equivalents (GAE)/g sample.Analyses were performed in triplicate.

Total Proanthocyanidin Content
The proanthocyanidin content was determined by a modification of the Bate-Smith method, which consists of the cleavage of the C-C interflavanic bond of proanthocyanidins in butanol-HCl through oxidative acid-catalysis [42].Briefly, 0.2 mL of each sample were mixed with 20 mL of butanol/HCl (50:50) and 0.54 mM FeSO 4 .The mixture was incubated at 90 • C for 1 h and after cooling, the volume was completed to 25 mL with butanol-HCl mixture.The absorbance was measured at 550 nm against a blank prepared in a similar way but without heating.The standard used was cyanidin chloride (which served to draw a calibration curve.Results were expressed as mg of cyanidin chloride equivalents (CAE)/g of extract.

DPPH Radical-Scavenging Activity
A solution of 2,2-diphenyl-1-picrylhidrazyl (DPPH) (0.25 mM) was prepared using methanol as solvent.Next, 0.5 mL of this solution were mixed with 1 mL of extract at different concentrations, and incubated at 25 • C in the dark for 30 min.DPPH absorbance was measured at 517 nm.Blanks were prepared for each concentration.The percentage of the radical-scavenging activity of the sample was plotted against its concentration to calculate IC 50 , which is the amount of sample required to reach the 50% radical-scavenging activity.The samples were analyzed in triplicate.

ORAC Antioxidant Activity
Extracts (0.05 g) were mixed with 10 mL of methanol/HCl (1000:1, v/v) and sonicated for 5 min.Afterwards, the mixture was centrifuged and filtered.Fluorescein was used as fluorescence probe [45].The reaction was performed in 75 mM phosphate buffer (pH 7.4) at 37 • C. The final assay

Figure 2 .
Figure 2. Cytotoxicity dose-response curves of extract treatment on tumoral cell lines.(a) Petiveria alliacea, (b) P. niruri, (c) S. reticulate.Results represent the mean ± SE of triplicates of one representative experiment of each cell line.

Figure 2 .
Figure 2. Cytotoxicity dose-response curves of extract treatment on tumoral cell lines.(a) Petiveria alliacea, (b) P. niruri, (c) S. reticulate.Results represent the mean ± SE of triplicates of one representative experiment of each cell line.

Table 1 .
Extraction yield and total phenolic content.

Table 1 .
Extraction yield and total phenolic content.

Table 4 .
Cytotoxicity of extracts to gastric (AGS) and colon (SW620) adenocarcinoma cells as well as to control Vero cells.