**4. Discussion**

The aim of the present study was to assess for the first time the antioxidant effects in endothelial cells of the polyphenolic extracts obtained from the fruits of three wild growing plant species, *R. canina*, *R. sempervirens* and *P. coccinea*. It should be noted that *R. canina* is a well-studied plant species [33], but there is only one study for the polyphenolic composition and the antioxidant activity of *R. semprevirens* fruit extracts [34] and there are only few reports for *P. coccinea* fruit extracts [23].

Initially, the polyphenolic composition of the extracts was analyzed. *R. canina* extract exhibited the highest TPC among the three tested extracts. So, as expected *R. canina* had also the highest TFC. However, although the *R. sempervirens* extract had higher TPC than *P. coccinea*, the TFC of the former was lower than that of the latter. This was probably due to the high content of the *R. sempervirens* extract in polyphenols other than flavonoids. Other studies showed that the TPC values of *R. canina* extracts were from 50 to 500 mg/g dw, and so our results were within this range [35,36]. There has been so far only one study on the polyphenols of *R. sempervirens*, which reported 57.9 mg GAE/g dw for TPC and 0.47 mg CE/g dw for TFC, while our values were 267.67 mg GAE/g dw and 65.78 mg CE/g dw, respectively [34]. Nadpal et al., [34] have used different solvents (i.e., methanol and water) for isolating polyphenolic ectracts from *R. sempervirens.* Their results showed that water extraction yielded higher TPC and TFC values compared to methanol extraction [34]. Moreover, another study demonstrated that the TPC values were different between *R. canina* extracts isolated from plants grown in different locations [36]. In a previous study, it was also shown that ethanol extracts of *R. canina* had higher TPC values than water extracts [36]. Bhave et al., [37] have also shown that the content of biologically active compounds in Rosa species depended on specific genotypes. In addition, extracts from more ripened fruits of *R. canina* have been reported to have higher TPC compared to less ripened fruits [38]. Thus, the differences between our results and those from other studies could be atrributed to different factors such as differences in the analytical methods, different extraction methods, genetic and environmental factors and the maturity stage of fruits and harvesting time [37–39].

The HPLC analysis showed that the *R. canina* extract was especially rich in the polyphenols hyperoside, astragalin, rutin, (+)-catechin and (−)-epicatechin as well as in the terpenoid ursolic acid and a poly-hydroxylated organic acid organic acid, which is the quinic acid. The presence of these compounds have also been reported in previous studies [33,40]. Moreover, several other polyphenols have been identified in trace amounts in *R. canina* extracts such as ellagic acid, salicylic acid, vanillic acid, ferulic acid and caffeic acid (not found by us) [33]. The *R. sempervirens* extract had higher amounts of quinic acid, (+)-catechin, (−)-epicatechin, astragalin, and hyperoside. Like our findings, Nadpal et al. (2018) [34] have also identified gallic acid, quercitrin, quercetin, hyperoside and (+)-catechin in *R. sempervirens* extract, although not at the same concentrations as our samples. However, Nadpal et al. (2018) [34] have also reported ellagic acid, protocatechuic acid (not found in our extract), ferulic acid and kaempferol-3-*O*-glucoside. In general, our study is the first that identified quinic acid, (−)-epicatechin, rutin, astragalin, eriodictyol, and genistein in *R. sempervirens* extract. The *P. coccinea* extract contained higher concentrations of hyperoside, rutin, (−)-epicatechin, (+)-catechin, astragalin, vanillin, syringic acid and chlorogenic acid. As expected, the polyphenolic profiles of the two extracts from the *Rosa* genus (i.e., *R. canina* and *R. sempervirens*) had more similaritities with each other than with the extract from *P. coccinea* of *Pyracantha* genus.

All three extracts exhibited strong free radical scavenging activity in DPPH and ABTS•<sup>+</sup> assays. The highest potency of the *R. canina* extract in both assays was in accordance with its highest TPC value. Some studies have reported IC50 values in DPPH assay for *R. canina* extracts, which were similar to ours, but other studies have found values that were significantly different from ours [18,40]. Previous studies have also shown that polyphenolics were the most important compounds of *R. canina* extracts for the DPPH radical scavenging activity [40]. *R. sempervirens* extract having higher TPC than *P. coccinea* extract also exhibited greater free radical scavenging activity. Nadpal et al. (2018) [34] have reported 28 μg/mL as IC50 value in DPPH assay for a *R. sempervirens* extract, while our value was 130 μg/mL. The di fferences between our values and those from other studies are probably, as mentioned above, due to di fferent factors which lead to di fferences in the polyphenolic composition of the extracts.

Moreover, all three extracts exerted protective activity against ROO•-induced DNA damage. Like in DPPH and ABTS•<sup>+</sup> assays, the potency order in this assay followed the order of polyphenolic concentration, that is, *R. canina* > *R. sempervirens* > *P. coccinea*. Therefore, once again, the polyphenolic concentration seemed to play a crucial role in the protective activity from ROO•-induced DNA damage. Among the polyphenols identified in the extracts, (+)-catechin, (−)-epicatechin, rutin, vanillin, astragalin, phloridzin and gallic acid have been reported to scavenge ROO• radical [41–47]. As far as we know, protective activity of *R. canina*, *R. sempervirens* and *P. coccinea* extracts from free radical-induced DNA damage was examined for the first time in this study. Thus, based on these results, extracts from the tested plant species may be used for protection against diseases caused by ROS-induced DNA damage.

Since the tested extracts showed strong free radical scavenging activity, their antioxidant e ffects were also investigated at noncytotoxic concentrations in endothelial cells. Thus, the extracts' ability to increase GSH, one of the most important antioxidant molecules within cells, was assessed by flow cytometry [48]. In agreemen<sup>t</sup> with the free radical scavenging assays, *R. canina* extract exhibited the greatest capacity to increase GSH in the cells. However, it should be noted that when humans consumed rose hip powder from *R. canina*, there were no e ffects on the activity of enzymes related to GSH metabolism in erythrocytes [49]. Treatment of cells with *P. coccinea* extract also increased significantly GSH levels in EA.hy926 cells. However, *R. sempervirens* extract treatment had no e ffect on GSH levels. This finding was intriguing, since *R. sempervirens* extract contained more polyphenols than *P. coccinea* extract, while *R. sempervirens* extract's polyphenolic profile was quite similar to that of *R. canina*. This contradiction could be explained by examining the polyphenols that were found at a greater concentration in *P. coccinea* extract than in *R. sempervirens*. For example, hyperoside and rutin found at higher concentration in *R. canina* and *P. coccinea* extracts compared to *R. sempervirens* have been reported to increase GSH levels in endothelial cells [50,51]. The observed increase in GSH levels induced by *R. canina* and *P. coccinea* extract treatment is important, since GSH apart from its antioxidant role is a crucial regulator of cell signaling in endothelial cells [52,53].

Regarding extracts' e ffects on ROS levels in endothelial cells, only *R. canina* extract treatment exerted a significant decrease. This result was in accordance with *R. canina* extract's highest antioxidant potency exhibited in all the other assays. Moreover, *R. canina* extract-induced decrease in ROS levels was attributed, at least in part, to extract's capacity to increase antioxidant defense mechanisms such as GSH. Moreover, a *R. canina* extract has been shown to inhibit an H2O2-induced increase in ROS in colon cancer cells [54]. Furthermore, some polyphenols such as (+)-catechin, (−)-epicatechin and ursolic acid identified at higher concentrations in *R. canina* than in the other two tested extracts have been demonstrated to decrease ROS in endothelial EA.hy926 and human umbilical vein endothelial cells (HUVEC) cells [55–57].

In conclusion, the results of the present study provided new information concerning the polyphenolic composition of the tested extracts, especially those of *R. sempervirens* and *P. coccinea* fruit extracts. Moreover, all three tested extracts were demonstrated for the first time to protect against ROS-induced DNA damage, which thus suggests their possible use for prevention of relative diseases. In addition, the extracts from *R. canina* and *P. coccinea* were shown to increase GSH levels, the most important antioxidant molecule in endothelial cells. This finding suggests that these extracts may be used for the development of food supplements or biofunctional foods preventing diseases caused by oxidative damage to endothelium such as cardiovascular. Interestingly, previous *in vivo* studies have reported that administration of rose hip powder to humans or experimental animals could reduce the risk for cardiovascular diseases [22,58]. Of course, further studies are needed in order to investigate further the molecular mechanisms through which these protective activities are exerted. Moreover, since the environmental variability between di fferent years may a ffect the chemical composition of the

plant extracts and consequently their bioactivities, it should also be examined how the tested activities are changing from one year to the next.

**Author Contributions:** Conceptualization, D.S., S.A.H., D.K.; methodology, D.S., S.A.H., D.K.; supervision, D.S., S.A.H., D.K.; carried out the experiments, E.K. (Eleana Kokka), A.A., K.C., E.K. (Efthalia Kerasioti), C.D., E.N.T., A.P., S.D.K., I.K.; data curation, E.K. (Efthalia Kerasioti), A.P., A.A., D.S., S.A.H.; writing, D.S., S.A.H., K.C., E.K. (Eleana Kokka), C.D., I.K.; funding acquisition, D.S., S.A.H., D.K.; resources, D.S., S.A.H., D.K.; K.C., E.K. (Eleana Kokka), C.D., have contributed equally to the study.

**Funding:** The work was funded in part by the "Toxicology" MSc program (grant no.: 5835) in the Department of Biochemistry and Biotechnology at the University of Thessaly.

**Conflicts of Interest:** The authors declare no conflict of interest.
