*2.6. Effect of Six OOPs on the Proliferation/Viability of Cancer Cells*

The potential anticancer properties of OOPs have been extensively investigated during the last two decades, mainly for tyrosol, hydroxytyrosol and less often with oleocanthal (**1**) or other OOPs. The establishment of the aforementioned new methods for OOPs' isolation has enabled further studies on the biological properties of each one of them on a wide array of cancer cell models. The present study investigated the antiproliferative and/or cytotoxic effect of oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**), ligstroside aglycone (**4a,b**) and the newly identified major phenolic ingredients oleomissional (**6a,b,c**) and oleocanthalic acid (**7**) [34,40] (Figure 1). The six compounds analyzed in this study were isolated as described in the Experimental Section. Their bioactivity was evaluated on sixteen human tumor-derived cell lines, sensitive or resistant to certain chemotherapy agents, from eight different tissue origins. Moreover, the bioactivity of the six OOPs was tested on five non-tumorigenic cell lines (Table S1).

The antiproliferative/cytotoxic effects of the selected OOPs on each cell line were assessed by measuring the cellular ATP levels after 72 h treatments using an ATP-based luminescence assay [73]. This assay, initially developed as a tumor chemosensitivity assay, has shown considerable promise as a general in vitro toxicity assay due to its high sensitivity that allows the detection of a small number of cells. Consequently, it can be applied to both cancer cell lines and primary tissue cells. To evaluate the dose–response effect of each OOP, concentrations of 1–100 μM were used and the bioactivity strength of the six OOPs was compared on the basis of their effective concentration (EC50s; i.e., concentrations that inhibited the proliferation of the cell population by 50% as compared to control cells treated

just with the solvent (i.e., DMSO)). The EC50 values were calculated by nonlinear regression (curve fit) using a sigmoidal dose–response equation (Figures 2A and S2).

**Figure 1.** Structures of the studied OOPs and related compounds: oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**), ligstroside aglycone (**4a,b**), oleomissional (**6a,b,c**), oleokoronal (**7a,b,c**), oleocanthalic acid (**7**), oleocanthadiol (**8**), oleuropein (**9**), oleaceinediol (**10**) and oleomissionadiol (**11a,b**).

The relative antiproliferative/cytotoxic activity of OOPs on cancer cells after 72 h treatment was: oleocanthal (**1**) > oleuropein aglycone (**3a,b**) > ligstroside aglycone (**4a,b**) > oleacein (**2**) > oleomissional (**6a,b,c**) > oleocanthalic acid (**7**) (Figure 2B and Table 1). The calculated EC50 values ranged between 9.1–100 μM (Table 1). The bioactivity of the oleocanthalic acid (**7**), studied here for the first time for its antiproliferative/cytotoxic activity, was initially evaluated systematically in the three breast cancer cell lines using a concentration range of 10–100 μM. In all cases, the viability measured was ≥80%, predicting EC50 values > 100 μM. Since this study was focused on OOPs with EC50 values < 100 μM, which would be promising for future use in cancer prevention or cancer treatment, the range of concentrations tested for all six compounds analyzed was below 100 μM. For this reason, oleocanthalic acid (**7**) was excluded from the rest of this study.

**Figure 2.** OOPs reduce cell numbers/viability of several cancer cell lines. (**A**) The effect of different concentrations of oleocanthal on four cancer cell lines is shown as sigmoidal dose–response curves in two representative panels. Cell numbers/viability was measured using the ATP-based luminescence assay after 72 h treatment. The results shown are from two or three independent experiments performed in triplicates. Data are represented as mean cell count ± SE in each treatment group normalized to the control group (i.e., cells treated only with 0.2% (*v*/*v*) DMSO). (**B**) Effects of different concentrations of oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**), ligstroside aglycone (**4a,b**) and oleomissional (**6a,b,c**) on cell numbers/viability. Bar graphs representing the mean EC50 values for the studied compounds on the cell numbers/viability of a panel of sixteen cancer cell lines from

eight different tissue origins. Mean EC50 values ± SE are from two or three independent experiments performed in triplicate. The EC50 values were calculated using the GraphPad Prism software.

**Table 1.** EC50 values of OOPs (i.e., oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**), ligstroside aglycone (**4a,b**), oleomissional (**6a,b,c**) and oleocanthalic acid (**7**)). Their effect on the proliferation or the viability of cancer and non-tumorigenic transformed cell lines or normal cell lines was determined by the ATP assay. EC50 values were calculated after 72 h treatments for each experiment using GraphPad Prism software and were then used to calculate the average and the SE values. The results presented are from two or three independent experiments performed in triplicate.


Oleomissional (**6a,b,c**), which was also studied here for the first time, presented EC50s > 50 μM for the cell lines from breast cancer (i.e., MDA-MB 231, MCF-7 and SK-BR-3), skin melanoma (i.e., A2058), colon and gastric epithelium cancer (i.e., HT-29, Caco-2, AGS) and cervical cancer (i.e., ME-180, Hela). However, for the MKN-45 gastric cancer, SK-MEL-28 melanoma, liver and pancreas cancer cell lines (i.e., Huh-7, HepG-2 and PANC-1) and the H1437 lung cancer cell line, the EC50 values exhibited by oleomissional (**6a,b,c**), were <50 μM (Table 1). Significant variation was observed in the OOPs' EC50 values for cancer cell lines with different tissue tumor origin but also amongst cell lines of the same tissue origin but with different genetic identities (Figure 3, Table 1).

**Figure 3.** Bioactivity range of OOPs on several cancer cell lines from eight different tissue origins.

With respect to breast cancer, the MDA-MB-231, SK-BR-3 and MCF-7 cell lines were selected as representative for this study. The highly metastatic, triple-negative MDA-MB-231 cells lack the expression of estrogen receptor (ERα)—the target for hormonal therapy—and overexpress c-Met—a breast cancer molecular target of oleocanthal (**1**). The SK-BR-3 cell line overexpresses human epidermal growth factor receptor 2 (HER2) while the MCF-7 cells express ERα and c-Met [28,47,74]. Oleocanthal (**1**) was the most effective OOP on all three breast cancer cell lines (EC50 = 10.5–24.6 μM) and oleuropein aglycone (**3a,b**) followed in potency (EC50 = 17.7–32.2 μM). The MCF-7 cancer cells proved to be the most resistant to oleocanthal (**1**) of all three cell lines, with EC50s > 25 μM.

The cytotoxic effect of oleocanthal (**1**) on breast cancer cells and its potential mechanism of action have been investigated under different conditions in several studies, mostly on MDA-MB 231 and MCF-7 cells. Until now, the reported EC50 values for oleocanthal (**1**) in the three breast cancer cell lines used in this study have varied between 10–18.5 μM for MDA-MB 231 and 18–40 μM for MCF-7 estimated after 48 or 72 h treatment with the OOP in different culture conditions (i.e., serum-free, FBS 0.5% (*v*/*v*), HGF-supplemented media and EGF-supplemented media) [28,47,74,75]. The corresponding EC50 values calculated in this study were close to the lowest values of these ranges. Siddique et al. (2019) determined the EC50 value of oleocanthal (**1**) on SK-BR-3 cells to be 27.3 μM (after 48 h treatment in HGF- and EGF-supplemented media), while in this study, the EC50 value was calculated to be 13 μM [76].

Similar studies on the oleuropein aglycone (**3a,b**) bioactivity are still limited. Menendez et al. (2007) reported that the concentration of oleuropein aglycone (**3a,b**) reducing cell viability by 50% after five days of treatment was 47 μM for the SK-BR-3 cells and >100 μM for the MCF-7 cells [17]. A very recent study by Mazzei et al. (2020) showed that the calculated EC50 values for MCF-7/TR (tamoxifen-resistant) and MDA-MB 231 cells were 70 μM and 53 μM, respectively [27]. These EC50 values are twice to three times higher than the values calculated in the present study (i.e., for MDA-MB 231, EC50 = 24.5 μM; for SK-BR-3, EC50 = 17.7 μM; for MCF-7, EC50 = 32.2 μM). The differences in these results may be due to variations in the experimental conditions or differences in the purity of the compound used.

As for ligstroside aglycone (**4a,b**), Busnena et al. (2013) reported an EC50 value for MDA-MB 231 of approximately 80 μM after 48 h treatment, while in the present study ligstroside aglycone (**4a,b**) was found to be much more effective (EC50 = 31.6 μM) [23]. Moreover, similar to the results presented herein, previous studies have shown that SK-BR-3 cells were sensitive to ligstroside aglycone (**4a,b**) (EC50 = 26 ± 6 μM after 5 day treatment) [53].

With respect to skin cancer, A2058 and SK-MEL-28 melanoma-derived cell lines were included in the present study. The calculated EC50 values for oleocanthal (**1**) were 10.4 μM and 18.4 μM in the SK-MEL-28 and A2058 cells, respectively. These EC50 values are similar to those observed for the breast cancer cell lines. The EC50 values for oleuropein aglycone (**3a,b**) were 15.1 μM and 37.2 μM, respectively, for the two cell lines. Interestingly, SK-MEL-28 cells were more sensitive to all OOPs than the A2058. To our knowledge, no data have been reported until now for the antiproliferative/cytotoxic effect of these two OOPs (**1** and **3a,b**) on the aforementioned cancer cell lines. However, oleocanthal (**1**) has been shown to inhibit cell viability in several human melanoma cell lines, including the A375, 501Mel and G361 cells at low concentrations [21,22]. Moreover, for oleocanthal (**1**) and oleacein (**2**), much higher EC50 values were reported recently on A375 cells (i.e., 67.5 ± 1.9 and 112.9 ± 4.9 μM, respectively) after 72 h treatment [50].

In the hepatic cancer cell lines, oleuropein aglycone (**3a,b**) was almost twice as effective than oleocanthal (**1**), with the Huh-7 being more sensitive to all OOPs than the HepG-2 cells. It is noteworthy that in two previous reports on the antiproliferative/cytotoxic activity of oleocanthal (**1**) against hepatocellular carcinoma [29,48], the calculated EC50 values were similar to those presented in this study; however, in one study, the Huh-7 cells were found to be more resistant than HepG-2 to treatment with oleocanthal (**1**) while in another they responded similarly [29]. Moreover, the pancreatic cancer cell line PANC-1 was almost equally sensitive to oleocanthal (**1**) and oleuropein aglycone (**3a,b**) but more sensitive to the rest of OOPs than the hepatic cancer cell lines.

The Caco-2 colon cancer cells were more sensitive to oleuropein (**3a,b**) and ligstroside (**4a,b**) aglycones than to oleocanthal (**1**) while for the HT-29 colon cancer cells the opposite was observed. It is worth noting that while in this study oleocanthal (**1**) was found to have moderate activity (EC50 = 33.4 μM, after 72 h treatment), others have reported that it had no cytotoxic effect on Caco-2 cells (EC50 > 150 μM) [75]. For the stomach cancer cell lines oleocanthal (**1**) was the most effective OOP, with oleuropein aglycone (**3a,b**) being the second most effective in both cell lines. Moreover, MKN-45 stomach cancer cells were more sensitive to all OOPs than the AGS cells.

The two lung cancer cell lines showed different sensitivities to oleocanthal (**1**) and the two aglycones (**3a,b** and **4a,b**). While oleocanthal (**1**) and oleuropein aglycone (**3a,b**) were the most effective OOPs on H1299 cells, the H1437 cells were more sensitive to the two aglycones (**3a,b** and **4a,b**) with oleocanthal (**1**) following in effectiveness. Moreover, they were more sensitive to the rest of the OOPs than the H1299 cells. Until now, only oleocanthal (**1**) has been shown in one report to inhibit the cell viability of several human lung cancer cell lines, including A549 and NCI-H322M cells [31].

Finally, the ME-180 cervical cancer cells were more sensitive than the Hela to all OOPs with oleocanthal (**1**) the most effective and oleuropein aglycone (**3a,b**) the second. Once more, the EC50 value reported in the present study for the activity of oleocanthal (**1**) on Hela cells was 44.6 μM, while in another study, an EC50 value >150 μM was calculated for oleocanthal [75].

To summarize, for most cancer cell lines tested herein, oleocanthal (**1**) was the most effective OOP in its antiproliferative/cytotoxic effect while oleuropein aglycone (**3a,b**) ranked second. The only exceptions where oleuropein or ligstroside aglycones (**3a,b** and **4a,b**) were more effective than oleocanthal (**1**) were on (a) the two hepatic cell lines, Huh-7 and HepG-2 ((**3a,b**)>(**1**)), and (b) on the H1437 lung cells (both aglycones (**3a,b** and **4a,b**) > (**1**))—results that merit further investigation. A detailed analysis of the bioactivity of the six OOPs highlighted their differential activity on cells of different cancer origin but also on cell lines of the same tissue origin but with different genetic backgrounds. With respect to the EC50 values of the OOPs studied until now, the majority of the EC50 values calculated in this study were either considerably lower or similar to the values already reported.

#### *2.7. Effect of Six OOPs on the Viability of Non-Tumorigenic Human Cells Lines; Selectivity of OOPs' Bioactivity*

The OOPs analyzed above, except oleocanthalic acid (**7**), were also tested for their antiproliferative/cytotoxic effect on non-cancer immortalized or normal human cell lines of different tissue origins. Human mesenchymal stem cells derived from umbilical cord (Wharton's jelly stem cells (WJSCs)) were also used as an alternative cell model [77–79]. Two out of the five cell types (i.e., MRC-5 derived from lung and MCF-10A non-tumorigenic breast epithelial cells) were more sensitive than the cancer cell lines of similar tissue origin in all OOPs tested (Table 1). However, the skin-derived cells (i.e., HaCaT spontaneously transformed immortal keratinocytes and the NHDF Normal Human Dermal Fibroblasts) were either as sensitive as the A2058 melanoma cells (i.e., the HaCaT cells) or more resistant to oleocanthal (**1**) treatment than both the A2058 and SK-MEL-28 melanoma cell lines tested (i.e., the NHDF cells). The WJSCs were found to be as sensitive as NHDF to oleocanthal treatment, but more sensitive to oleacein (**2**) and the two aglycones (**3a,b** and **4a,b**).

To evaluate the anti-cancer potential of a compound, its cytotoxicity against non-tumorigenic cell lines must be determined in order to calculate the selectivity index value (SI). Comparison of the OOPs' selectivity indexes for the cell lines (cancer and non-tumorigenic transformed cells) of the same tissue origin (i.e., the ratio of EC50 for non-tumorigenic cells/EC50 for cancer cells) summarized in Table S3 indicated that the SIs ranged between 0.1–2.8 [80]. Oleocanthal (**1**) and the two aglycones (**3a,b** and **4a,b**) displayed SI > 2 for the melanoma cell line SK-MEL-28 (2.4, 2.8 and 2.1, respectively) while ligstroside aglycone (**4a,b**) showed moderately good selectivity for some breast cancer cell lines, as well (Table S3). According to Weerapreeyakul et al. (2012), a SI value ≥3 is required for classifying a compound as prospectively anti-cancer [81], but others consider SI values >2 as a positive indication for further investigation of a compound's anti-cancer potential [82,83]. The results presented in this study concerning the sensitivity of MCF-10A and HaCaT cells to treatment with OOPs do not correlate with previously reported studies in which the MCF-10A cells were found resistant to treatment with oleocanthal (**1**), ligstroside aglycone (**4a,b**) and oleuropein aglycone (**3a,b**) at concentrations of 40 μM, 50 μM and 150 μM, respectively [23,27,28,30]. Moreover, other studies have demonstrated that the HaCaT cells were resistant to treatment with oleocanthal (**1**) and oleacein (**2**), while EGF-stimulated HaCaT cells were found to be more sensitive to these OOPs [22,50]. Since we obtained the MCF-10A cells directly from the ATCC cell bank and evaluated their sensitivity to OOPs in parallel with the respective cancer cell lines using the same methodology and the same compounds, the results reported herein bear validity. The HaCaT cells were found as sensitive to A2058 melanoma cells. More specifically, they were more sensitive to oleocanthal (**1**) and oleuropein aglycone (**3a,b**) treatment, but the EC50 values of oleacein (**2**), ligstroside aglycone (**4a,b**) and oleomissional (**6a,b,c**) were higher than 50 μM.

Therefore, the above results, although encouraging with respect to the selectivity indexes estimated for oleocanthal (**1**) and the two aglycones (**3a,b** and **4a,b**) for some cancer types, raise questions concerning the validity of using transformed/immortalized cell lines to evaluate the potential use of OOPs or other natural products for anticancer treatments. They may not be representative cell models of normal tissues. Two-dimensional or 3D cell culture systems utilizing human primary cells from different tissues may provide more physiologically relevant information and more predictive data in in vitro assays testing the selectivity of OOPs' anti-cancer effect [84,85].

#### *2.8. Kinetics of Antiproliferative/Cytotoxic Effect of OOPs*

To examine the time dependence of the OOPs' effect on cancer cell lines, treatments were performed for different time lengths and the reductions in cell numbers were compared. The MDA-MB 231 and SK-MEL-28 breast cancer and melanoma cell models, respectively, were treated for 24, 48 and 72 h with OOP concentrations lower, higher or close to the EC50 values of the most active compounds (i.e., oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**) and ligstroside aglycone (**4a,b**)). Cell numbers were assayed using the ATP assay.

As expected, the highest cell numbers in the treated cultures (i.e., the weakest effect of OOPs) were observed at 24 h, while the lowest (i.e., the strongest effect of OOPs) at 72 h treatments. Interestingly, treatments with some OOPs (i.e., ligstroside aglycone (**4a,b**) and oleacein (**2**)) had a similar effect to that observed at 24 h independently of the OOPs concentration used. By contrast, others ((i.e., oleocanthal (**1**) and oleuropein aglycone (**3a,b**)) acted earlier than 48 h, but reached the highest levels of effect at 72 h, when the used concentrations in the treatments were close to or higher than their EC50 values. Apparently, some OOPs exert their antiproliferative/cytotoxic effect earlier than 48 h while others act slower (Figure 4). On the basis of these results, it was decided to determine the EC50 values for all of OOPs tested in all the cell lines at 72 h of treatment.

**Figure 4.** Time- and dose-dependent effect of OOPs on MDA-MB231 and SK-MEL 28 cell lines. Cells were treated with different concentrations (2.5–70 μM) of oleocanthal (**1**), oleacein (**2**), oleuropein aglycone (**3a,b**), ligstroside aglycone (**4a,b**) and oleomissional (**6a,b,c**). Cell viability was measured using the ATP-based luminescence assay at the end of 24, 48 and 72 h treatment. Results from two independent experiments performed in triplicate. Bar graphs represent the mean cell count ± SE in each treatment group normalized to the control group (i.e., cells treated only with 0.2% (*v*/*v*) DMSO). \* *p* < 0.05; \*\* *p* ≤ 0.01; \*\*\* *p* ≤ 0.001 (*t*-test) comparing viability in the treatments for 48 h and 72 h.
