**4. Discussion**

Increasing antibiotic resistance makes the search for alternative antimicrobial compounds of a crucial importance for global health [33]. Failure to defeat fast adapting pathogens without significant damage to host tissues is a key challenge in managemen<sup>t</sup> of chronic infectious-inflammatory disease including periodontitis [34]. Progressive bacteria-driven inflammatory response causes continuous damage on periodontal cells making them more sensitive to harmful effects of antibiotics and antimicrobial chemicals [3,35]. The damage is further exacerbated by the treatment-caused loss of beneficial commensal bacteria [36]. This suggests reconsidering the possibilities of alternative treatment strategies including use of specific pathogen-targeting bacterial strains [37] and plant-derived antibacterials, because such strategies are characterized by lower or no side effects and resistance development risk, as well as complex antiinflammatory and tissue renewal stimulating properties. This study explored antibacterial and antiinflammatory properties of PSRE that is known as potent infection-defeating preparation and PSRE-derived PACN possessing stronger antioxidant and antibacterial properties compared to PSRE [17,19].

Both substances were effective in reducing metabolic activity of the selected strains suggesting a broad range of antibacterial properties. This is in line with previous evidence about various extracts prepared from *P. sidoides* roots. A commercial aqueous-ethanolic extract from *P. sidoides* EPs® 7630 (Umckaloabo®) is reported to inhibit growth of *Streptococcus pyogenes*, *Proteus mirabilis, Staphylococcus aureus*, *Escherichia coli*, *Streptococcus pneumoniae*, *Haemophilus influenza*, *Staphylococcus epidermidis* and some other gram-negative and gram-positive bacterial strains (summarized in [38]). Aqueous-acetone PSRE was efficient in decreasing growth of antibiotic-resistant *S. aureus* strains [39]. The present study for the first time demonstrated the growth-suppressing efficiency of PSRE and PACN on *Aggregatibacter actinomycetemcomitans*, one of the most important gram-negative anaerobic periodontal pathogens [40]. Similarly as in the previously demonstrated case of *P. gingivalis* [19], PACN demonstrated significantly higher toxicity on *A. actinomycetemcomitans*, compared to the effect of PSRE. Fifty μg/mL PACN reduced metabolic activity of *A. actinomycetemcomitans* nearly 10 times more if compared to the untreated control value (Figure 1d). The same concentration had no significant toxicity on other investigated strains. The minimal amount of PACN causing a significant effect on metabolic activity of *E. coli* and *S. aureus* was 80μg/mL, and for *S. epidermidis*the significant toxicity started

from 70 μg/mL PACN. The results indicate that there might be a specific interaction of proanthocyanidins from PSRE with the main pathogenic strains (*P. gingivalis* and *A. actinomycetemcomitans*) responsible for the development of periodontitis. Strain-specific activity of proanthocyanidins was already noticed by other authors. Lacombe and Wu have reviewed the selective pathogen-suppressing and beneficial strain-promoting activity of proanthocyanidins derived from various berries [41]. However, despite many publications reporting a selective activity of natural extracts towards pathogen and non-pathogen strains, it is still not completely clear how this selection occurs [42]. It was shown that cranberry-derived proanthocyanidins are able to interfere with a *<sup>N</sup>*-acylhomoserine lactone-mediated quorum sensing of *Pseudomonas aeruginosa* [43]. Moreover, proanthocyanidins have also been shown to compromise adhesion to host cells by mimicking cell surface signaling [44]. Some authors have proposed the hypothesis that proanthocyanidins might increase bacterial membrane permeability and cause indirect metabolism decrease due to ATP and other intracellular metabolite loss [42,45]. A recent study shows that proanthocyanidins can potentiate antibiotics by acting via bacterial multidrug efflux pumps [46]. Thus, the disturbance in transmembrane transport indeed might be the cause of bacteriotoxicity. However, more studies definitely are required to clarify the mechanism of action of proanthocyanidins against pathogenic bacterial strains.

On the other hand, PSRE was more e fficient than PACN in suppressing both of *Staphylococcus* strains that were investigated in this study suggesting that other than proanthocyanidin fraction compounds were acting against these bacteria. Most likely, the distinct antibacterial activity of PSRE can be ascribed to other phenolic compounds such as coumarins, phenolic acids, flavonols and flavan-3-ols [16].

Bacterial infection simulation in the co-culture "race for the surface" assay revealed that addition of 100 μg/mL of either PSRE or PACN in the medium was effective in preserving viability of human gingival fibroblasts in the presence of both *S. aureus* and *A. actinomycetemcomitans* (Figure 2). Similarly, B-type linked proanthocyanidin-coated surfaces are shown to inhibit bacterial spreading and promote survival of mammalian cells [47]. The mechanism proposed to explain the activity is bacterial attachment and biofilm formation prevention by prodelphinidin-rich proanthocyanidins. In our experimental model, a similar efficiency was achieved by PSRE and PACN solutions, indicating that interaction of soluble compounds with the walls of bacteria also could mediate bacterial adhesion and mammalian cell protection. Accordingly, these results are very promising support to the use of natural extracts as an effective alternative antibacterial compound able to preserve the naïve tissue.

Investigation of gingival tissue protecting properties of PSRE and PACN in the bacterial LPS-mediated inflammation model revealed that both preparations e fficiently prevent necrosis and apoptosis of fibroblast cells. Both substances were more e fficient in decreasing the executing caspase-3 activity compared with the e ffect on apoptosis triggering caspase-8. However, PSRE was less e fficient in suppressing caspase-8 activity than PACN, indicating that the latter had both upstream and downstream targets in the apoptotic cascade. The antiapoptotic activity of proanthocyanidins from grape seeds including decrease in executing caspases-3 and 9 was reported in a rotenone-induced neurotoxicity model of SH-SY5Y cells [48]. However, exposure of human colorectal carcinoma cells HCT-116 to proanthocyanidins from the same source significantly upregulated mRNAs encoding caspase-2, caspase-3 and caspase-9 [49]. Another study reports apoptosis induction in lung cancer cells NCI-H460 via stimulation of caspase-3 and mitochondrial cytochrome c release by gallic acid, one of the important constituents of PSRE [50]. Such controversial data sugges<sup>t</sup> that the e ffect of PSRE and proanthocyanidins on apoptotic signaling pathways is cell type-dependent and they might have opposite e ffects in cancerous and non-cancerous cells as well as in di fferent toxicity models.

Evaluation of pro-inflammatory cytokine secretion and gene expression revealed that PSRE and PACN suppress at least three di fferent inflammatory processes: cytokine secretion (IL-8 from gingival fibroblasts and IL-6 from bone marrow-derived macrophages), inflammatory gene expression (IL-1β, iNOS and COX-2) and macrophage conversion to pro-inflammatory M1 phenotype related to the tissue loss in periodontitis. Downregulation of COX-2 coding mRNA in mononuclear leukocytes and PGE2 release from gingival fibroblasts indicate suppression of the prostaglandin inflammatory pathway. PGE2 is the most prominent in the pathogenesis of periodontitis among prostaglandins [51,52]. PGE2 is involved in the stimulation of inflammatory mediators and MMPs, as well as osteoclast formation via receptor activator of nuclear factor-κB ligand (RANKL) [52,53]. IL-6 and IL-1β also mediate bone resorption via osteoclasts activation [54], and increase in iNOS leads to reactive nitrogen species-mediated apoptosis of gingival fibroblasts [55]. By suppressing these inflammatory pathways, PSRE and PACN are expected to significantly improve condition and survival of periodontal tissues. Similar antiinflammatory activity of PSRE together with *Coptis chinensis* root extract was recently shown in LPS-stimulated RAW 264.7 cells [56]. The extract combination significantly decreased the levels of iNOS, PGE2, TNFα, IL-1β and IL-6 in RAW 264.7 macrophages, and the results were also confirmed in vivo in a paw oedema rat model. Although the study reported lower levels of TNFα secretion from LPS-stimulated RAW 264.7 cells, in our study, we did not observe significant changes on TNFα gene expression in both LPS-stimulated leukocytes and LPS/IFN-γ-stimulated macrophages after PSRE and PACN treatment. Proinflammatory cytokine TNFα plays a critical role not only in inflammatory cell migration, but also in both innate and adaptive immune responses, by up-regulating antigen presentation and the bactericidal activity of phagocytes [57,58]. In periodontitis, TNFα is one of the key signals initiating several signaling pathways leading to chemotaxis of other inflammatory cells, tissue destruction and osteoclast formation [59,60]. The fact that PSRE and PACN had no e ffect on TNFα expression level while suppressing several other related genes indicate the targets of the substances are located either downstream of the TNFα signal or in the TNF-a excluding pathway.

Although antiinflammatory properties of PACN and PSRE revealed in the study were of comparative levels, PACN had stronger efficiency in suppressing caspases and preventing mediator release. Stronger anti-inflammatory activity of PACN might be due to greater amounts of prodelphinidins. These compounds possess higher antioxidant capacity and share certain important structural peculiarities, namely hydroxyl groups in B ring (especially in C4' position and catechol group), hydroxyl groups in the A ring at the C5 and C7 positions [61].
