Novel liquid-liquid extraction and self-emulsion methods for simplified isolation of extra-virgin olive oil phenolics with emphasis on (-)-oleocanthal and its oral anti-breast cancer activity

Currently, there is a wealth of epidemiological evidence demonstrating the reduced risk of Mediterranean populations to certain chronic diseases including atherosclerosis, cardiovascular disease, and particular types of cancer, in addition to extended life expectancy compared to other geographical populations [1, 2]. Different statistical and experimental studies revealed a strong correlation between these consistent long-term health-promoting properties and the regular consumption of extra-virgin olive oil (EVOO), which is a major component of the Mediterranean diet [1, 2]. Besides its unsaturated fatty acid content, EVOO contains minor bioactive phenolic ingredients such as: a) simple phenols, including tyrosol and hyroxytyrosol, phenolic acids, and flavonoids, b) lignans, including (+)-pinoresinol (PR) and (+)-acetoxypinoresinol (APR, Fig 1), and c) secoiridoids that are phenolic monoterpenes containing substituted pyran core and represented by S-(-)-oleocanthal (OC) (Fig 1), (-)-hydroxyoleocanthal (HOC), S-ligstroside aglycone (LA), and S-oleuropein aglycone (OA) [1–3].

Among various naturally occurring secoiridoids in EVOO, OC (Fig 1) became renowned for its diverse health-benefiting properties against several ailments [4]. OC was first isolated from EVOO by Beauchamp group with a potent nonsteroidal anti-inflammatory activity due to its cyclooxygenases inhibitory activity [5]. Furthermore, OC has been reported as a potent antioxidant, a neuroprotectant that alters the structure, function, and enhance the clearance of the amyloid-β and reduce Tau hyperphosphorylation associated with the debilitating effects of Alzheimer’s disease [6–9]. OC also inhibited macrophage inflammatory protein-1α (MIP 1α) dysregulated in multiple myeloma [6–9]. The anti-inflammatory activity of OC is mediated, at least in part, through the inhibition of IL-6 expression and secretion and 5-lipoxygenase [6–9]. In addition, the antimicrobial activity of OC against Helicobacter pylori, along with its high stability in gastric juice strongly confirmed the OC’s capacity to protect against gastric ulcers and some gastric cancers [10]. On the other hand, OC has been proven to mediate its anticancer effects through the modulation of HGF-c-Met axis and its downstream signaling pathways [11, 12]. Consistently, in vitro studies have demonstrated effective inhibition of the proliferation, migration, and invasion of c-Met-dependent human breast and prostate cancer cells, upon OC treatment [11, 12]. In addition, OC markedly reduced tumor volume in triple negative breast cancer (TNBC) mouse model, validating OC lead rank [12]. In support of its health benefits, OC has no toxic effects on the viability and growth of the non-tumorigenic human mammary epithelial cells [13]. Meanwhile, OC caused a marked in vitro downregulation of mTOR phosphorylation in TNBC cells, most likely as a downstream effect due to c-Met inhibition [13]. OC inhibited the viability and COX-2 expression in HT-29 colon cancer cells via the activation of AMPK-mediated pathways, plausibly explaining, in part, its preventive and chemotherapeutic potential against colon cancer [14]. OC also inhibited hepatocellular carcinoma (HCC) cells migration and invasion in vitro and prevented its lung metastasis in vivo [15]. OC inhibited HCC epithelial-mesenchymal transition (EMT) by downregulating Twist, a direct STAT3 target. Recently, OC modulated estrogen receptor expression in luminal breast cancer in vitro and in vivo and synergized with tamoxifen treatment [16].

Hydroxyoleocanthal (oleacein, HOC, Fig 1) showed antiproliferative, antioxidant, ACE and 5-lipoxygenase inhibitory activities [17–19]. Ligstroside aglycone (LA, Fig 1) induced apoptotic cell death in HER2-dependent breast cancer and showed moderate in vitro cytotoxicity against a panel of human cancer cell lines [20]. Oleuropein aglycone (OA, Fig 1) identified as the most in vitro potent EVOO phenolic against HER2-dependent breast cancer cell viability by selectively triggering high levels of apoptotic cell death and suppressing lipogenic enzyme fatty acid synthase expression [21, 22].

Lignans including (+)-pinorsinol (PR) and (+)-acetoxypinoresinol (APR, Fig 1) are the second phenolic group in EVOO [23]. PR showed anti-inflammatory activity via inhibition of NO, PGE2, COX-2, TNFα, IL1β, IL-6, and NF-κβ [23, 24]. PR also showed antifungal activity against various Candida and Fusarium species by inhibiting their trichothecene biosynthesis, antioxidant activity via ROS scavenging and reducing LDL oxidation, hypoglycemic activity via inhibition of α-glucosidase, and in vitro neuroprotective effects [23, 24]. Anticancer effects of PR against breast cancer cells proved via reduction of fatty acid synthase (FASN), reducing HER2 expression and activation in HER2-dependent cells [23, 24]. PR inhibited the proliferation of TNBC, prostate, melanoma, ovarian, colon, and leukemia cells in vitro [23, 24]. APR exclusively occurs in olive with very limited number of bioactivity studies [23].

Together, this compelling data supported the hypothesis that secoiridoids, specifically OC, and lignans correlate with the positive health outcomes associated with EVOO intake, rendering the discovery of new simplified methods for their isolation appealing. Such reliable isolation method will make EVOO phenolics readily available in large quantities, which will facilitate their future use as pure or combined dietary supplements for improved human health applications.

Currently there are several methods for detection and quantification of EVOO phenolics, including OC, such as direct measurement using q¹H NMR [3], capillary electrophoresis [25], as well as reversed-phase HPLC methods [26]. Available isolation methods require the use of reversed-phase stationary phases and sophisticated technologies such as HPLC, high performance countercurrent chromatography (HPCCC) [27], as well as several liquid-liquid extraction methods using different organic solvents [11, 28]. In addition, Garcia et al proposed the use of deep eutectic solvents, such as choline chloride in various mixing ratios with sugars, as an alternative to methanol for extraction of EVOO phenolics [29]. On the other hand, Wen et al invented a method for extracting OC from EVOO with high yield using modified starch and performing subcritical liquid extraction with a mixture of organic solvents and spray-drying [30]. The majority of the current methods are focusing on EVOO phenolics/OC detection and quantitation for analytical purposes rather than large-scale isolations. The available methods still have serious drawbacks including the need for sophisticated instruments, massive consumption of organic solvents, high cost, tedious, and time-consuming. Importantly, the majority of EVOO phenolics isolation techniques have poor yields, which hinders research, clinical studies, and commercialization efforts that will require large quantities. OC was previously extracted from EVOO using methanol followed by washing with n-hexanes and successive purification of the dried extract using C-18 RP-preparative HPLC [11, 28, 31, 32]. The overall yield was poor due to significant OC loss by C-3, and possibly C-1, acetalization and the process was time-consuming and impractical.

Therefore, the main objective of this study was to develop reliable EVOO phenolics purification techniques useful for practical large-scale use while still time and cost-effective as well as environmentally friendly [33]. Purified OC was emulsified in water and assessed orally against breast cancer in nude mouse xenograft model.

The isolation of minor bioactive phenolic compounds in EVOO is challenging, time and cost consuming, need significant organic solvent amounts, require sophisticated technologies and reversed phase use, and poor yield due to potential acetalization of secoiridoid aldehydes [3, 27]. Therefore, two simple but effective isolation methods have been developed. It has been well-documented that methanol use for EVOO secoiridoid purification induces immediate and reversible aldehyde acetalization interaction leading to significant yield loss [3]. In order to eliminate the possibility of secoiridoids loss due to the expected acetalization, methanol has been replaced by CH₃CN in the EVOO liquid-liquid extraction procedure [3]. It is well documented that CH₃CN does not react with any analyte when used as a solvent for extraction of EVOO [3]. EVOO has been extracted three consecutive times with CH₃CN for efficient phenolic contents extraction (Fig 2). An ultra-freezing protocol has been then applied to the CH₃CN fraction via storing at -80 ⁰C for 90–150 min, three consecutive cycles (Fig 2). The major fatty content of the CH₃CN extract solidified and formed a precipitate under ultra-freezing conditions, while minor phenolics remained soluble in CH₃CN liquid form. Therefore, enormous fatty acid content amount has been filtered out and eliminated by simple filtration (Fig 2). A clear oily solution containing phenolics was collected after each ultra-freezing cycle (Fig 2) with significant reduction in fatty contents evidenced by ¹H NMR and HPLC analysis. A final Sephadex LH20 column chromatography purification step described earlier was only needed to timely and cost-effectively obtain pure EVOO phenolics in high purity and yield.

Though the acetonitrile-ultra-freezing method is efficient for extracting EVOO phenolics, it still has some limitation. The capacity of space in the -80 ⁰C facility is limited and not suitable for large-scale isolation. The massive organic solvent use is not environmentally friendly and the exhausted EVOO following CH₃CN extraction can’t be reused. EVOO phenolics can dissolve in polar as well as non-polar solvents and thus considered amphiphilic. For example, OC specifically underwent a spontaneous interaction with D₂O forming the more polar acetal/ monohydrate, by water addition to the C-3 aldehyde, which was characterized by ¹H NMR data [3]. Therefore, EVOO phenolics can reversibly interact with water in a similar manner to form more polar water-soluble acetal forms. Recovery rate of original aldehydic form will be high if the contact with water maintained as short as possible, preferably less than 12 h [3]. Since water is readily available, cost-effective, as well as environment friendly solvent, it has been hypothesized that water can be the appropriate solvent for large-scale extraction of EVOO phenolics. The water-EVOO extraction proposed to selectively dissolves phenolics, leaving crude fatty contents undissolved. Accordingly, EVOO was extracted three consecutive times with water via a simple liquid-liquid extraction protocol and the water fractions were collected (Fig 3). As expected, once EVOO mixed with water, it formed an emulsion thus, the process can be considered as a self-emulsifying extraction method (Fig 3). A quantitative study carried out to determine the amount of OC and other major EVOO phenolics in the resultant emulsion as well as in EVOO before and after water extraction using HPLC (Fig 4, Tables 1 and 2). OC and other phenolics were mostly shifted to the water layer, with negligible amount of fatty acids, dropping out the majority of fats in the oil layer (Fig 4). After three water extraction cycles, only 5–10% of the major EVOO phenolic OC content remained in EVOO compared to the fresh EVOO batch and nearly 90–95% of OC has been switched towards water layer as determined by HPLC (Fig 4). The residual amount of OC left in EVOO can be carried over by additional water extraction cycles. Furthermore, the resultant emulsion was subjected to freeze drying for 24–48 h to identify its ingredients using TLC and q¹HNMR analysis (Fig 4). The water extracts afforded EVOO phenolics, of which OC was the most major, representing ~80–85% of the extract content, along with ~15% monounsaturated fats and the rest of other minor EVOO phenolics, as suggested by TLC, q-¹HNMR and confirmed by HPLC analysis (Fig 4). In addition, the percentage recovery of OC using the water self-emulsifying method has been calculated and found to be around 90–100% after the fifth extraction round (Table 1). These results strongly confirm the efficiency of water capacity to extract EVOO phenolics. This novel self-emulsifying extraction method has a significant environmentally friendly advantage over all previously reported methods where organic solvents have been used. The resulting EVOO after water extraction can be reused for food without pungency and bitterness, soap, or cosmetics industry [27, 29, 30].

The self-emulsifying extraction method required multiple freeze-drying hours to evaporate water (Fig 3), which is time-consuming and may slightly affect EVOO phenolics yield. To bypass this drawback and make the water extraction method more efficient, wet packed resin adsorbent beads have been used as freeze-drying alternative to remove water and improve purity (S2 Fig). Three different types of entrapping resins were assessed; SP-70, Amberlite XAD-7, and Diaion HP-20. The EVOO phenolics-containing emulsion passed over a column packed with each resin type in water and the eluent was collected after using water as a mobile phase (S2 Fig). This eluted water was extracted with CH₂Cl₂, dried, and subjected to ¹HNMR analysis, which revealed no phenolics content (S2 Fig). Thus, the three resin types fully entrapped OC and other phenolics when eluted with water (S2 Fig). Interestingly, the water fractions contained only unsaturated fatty glycerides evidenced by their corresponding ¹HNMR spectra (S2 Fig). The resin-packed columns were finally washed with acetone to elute the entrapped phenolics and the acetone fractions were subjected to ¹HNMR analysis, which confirmed its phenolics content (S2 Fig, Table 1). Acetone is a non-halogenated, biodegradable, volatile, water miscible, environment friendly solvent, and will not acetalize OC and aldehydic secoiridoids, unlike alcohols. The SP-70 was the most efficient resin to entrap OC and phenolics, followed by Amberlite XAD-7, and finally the Diaion HP-20 resin (S2 Fig). Only a single Sephadex LH20 size exclusion chromatographic column was needed to obtain EVOO phenolics in high purity. The order of elution was lignans: PR, APR, followed by LA, and OC. Elution of hydroxytyrosol-based secoiridoids required the use of increasing amounts of EtOAc in CH₂Cl₂ to elute OA, HOC, and finally TY/HT.

In an attempt to validate the new extraction method for OC and phenolics from EVOO it was repeated three times to confirm its reproducibility. The same extraction protocol was used for each of the seven EVOO samples from different sources. For each sample, three water extractions were used. In addition, a corn oil sample has been used as a negative control to evaluate the selectivity of the extraction protocol (Table 1). Quantitative-¹HNMR as well as HPLC analysis have been used for detection and quantification of the major EVOO phenolics among different tested samples (Fig 4, Table 1). The proposed extraction protocol was efficiently and reproducibly able to extract OC from EVOO samples of different sources including Greece, Italy, and Florida. The Florida Olive Systems EVOO Sample-4 as well as corn oil sample did not demonstrate any appreciable quantities of OC (Table 1). Samples containing <80 mg OC/L EVOO will usually require 3–4 extractions, evidenced by HPLC analysis of each water extract, but OC-rich EVOO samples (>80 mg OC/ L EVOO) will need additional water extractions for complete OC extraction.

Upon water extraction, OC rich water extracts were subjected to further characterization of their emulsion properties including particle size, zeta potential, and polydispersity index (Fig 5). Particle size is a function of volume-weight distribution. Results showed that the water extract of different EVOO samples in solution have excellent nano-system volume-weight distribution of 2.87 nm. The particle size similarity among the water extract of multiple EVOO samples from different sources clearly confirmed the self-emulsifying potential of EVOO phenolics in aqueous solutions.

To confirm the stability of the formed nano-emulsion, the zeta potential of each water extract sample has been determined (Fig 5). Zeta potential serves as a reliable indicator of particle charge and emulsion stability. It reflects the potential difference between the dispersion medium and the stationary layer of fluid attached to the dispersed particle. Nearly all samples had zeta potential >40 mV, suggesting that the self-emulsified phenolics in water is a highly stable nano-emulsion.

In order to evaluate the heterogeneity and correctness of the particle size measurements, the polydispersity index (P.I.) has been determined for each water extract sample (Fig 5). The water extract showed an average P.I. of 0.3, suggesting the uniformity of the measured particle size and implying the lower susceptibility of the nano-emulsion to dispersion (Fig 5). Overall, these results confirm the ability of OC rich water extracts obtained from different EVOO samples to form stable nano-emulsions in aqueous solutions, which might be facilitated by the remained traces of oleic acid triglycerides.

It was interesting to note that different tested samples were able to form stable nano-emulsions with similar particle size, zeta potential and polydispersity index values, suggesting the reproducibility of the OC extraction protocol and eliminating the possibility of any EVOO source-specific results for this method (Fig 5).

The stability of nano-emulsions from different EVOO sources were evaluated by HPLC analysis before and after storing each sample for one month at either RT or 4 ⁰C based on its OC content. In addition, particle size, zeta potential, and polydispersity index have been determined for each nano-emulsion batch after one month to validate the HPLC results (Fig 5). The OC-rich nano-emulsions tend to degrade within one-month storage at RT (Table 3), with significant drop in zeta potential, along with a concomitant increase in particle size and polydispersity index in all investigated EVOO samples (Fig 5). On the contrary, the refrigerated OC nano-emulsion samples at 4 ⁰C showed better stability profile than those stored at RT, with no significant changes in zeta potential, particle size, and polydispersity index, compared to their respective initial values before storage (Fig 5). HPLC analysis showed slow degradation in the refrigerated nano-emulsion samples compared to samples stored at RT (Table 3). Overall, this preliminary stability study suggests that OC/phenolics nano-emulsions are unstable at RT and progressively degrade over time at RT while their stability improved by storage at 4 ⁰C.

Sephadex-LH20-purified OC was emulsified in water and assessed for ability to inhibit the proliferation of key human BC cell lines using the MTT assay. The cell lines have been selected to represent different BC phenotypes. The TNBC cells MDA-MB-231 and MDA-MB-468 lack the expression of estrogen receptor (ERα), an important target for hormonal therapy, while they overexpress c-Met, a well-established receptor tyrosine kinase that mediate BC survival and motility and validated as the molecular target of OC in BC [11, 12, 16, 34]. In contrast, MCF-7 and BT-474 BC cells express ERα and c-Met, however, the latter is expressed at relatively lower levels compared to the TNBC cells [12, 16, 34]. The antiproliferative effects of different doses of pure OC emulsion in water on the growth of MDA-MB-231, MDA-MB-468, MCF-7, and BT-474 BC cell lines after 72 h culture period are shown in Fig 6. Treatment with pure OC water emulsion dose-dependently inhibited the proliferation of tested BC cell lines (Fig 6). The IC₅₀ values for OC nano-emulsion were 19.4, 25.2, 24.4 and 20.1 μM in MDA-MB-231, MDA-MB-468, BT-474, and MCF-7 cells, respectively (Fig 6). Similarly, effects of OC emulsion on the viability of the human immortalized non-tumorigenic mammary epithelial cells MCF10A was studied to assess its selectivity (Fig 6E). Results showed that treatments with 0–40 μM OC had no effect on MCF10A cell viability compared to vehicle-treated control (Fig 6E). Progressive MCF10A cells growth inhibition was observed only at OC doses higher than 60 μM. These results further suggest the good selectivity of therapeutic OC concentrations toward BC cells at which the non-tumorigenic cells are minimally affected.

In our previous report, intraperitoneal administration of 5 mg/kg of pure OC, 3X per week showed 65% tumor growth inhibition of the MDA-MB-231 xenograft in female athymic nude mice [12]. OC potency is significantly better when used in early rather than late treatment modes [12]. In the current study oral in vivo antitumor efficacy of OC assessed using MDA-MB-231 human BC cells. Freshly prepared OC water emulsion was administered in preventive/early treatment mode orally at 10 mg/kg every day, starting seven days before tumor cell inoculation and continued for 5 weeks. Tumor progression was followed by direct measurement of tumor volume starting 14 days after orthotopic xenografting. Early treatment of OC showed delayed in on set of tumor development in compared to vehicle treated control group (Fig 7). A significant reduction in both tumor volume and tumor weight observed with the OC water emulsion-treated group, compared to the vehicle-treated control group (Fig 7). The 10 mg/kg oral daily OC water emulsion treatments significantly suppressed the MDA-MB-231 tumor growth by 90% on the final day of study, compared to the vehicle-treated control group, without negatively affecting the treated mice’s body weight or their behavior (Fig 7). It is worth noting the significant in vivo activity of OC, unlike its modest in vitro activity, which may suggest possible metabolic bioactivation [12, 15, 16]. This is the first report of oral anti-BC activity of OC. These results indicate that OC emulsion in water demonstrate a robust oral antitumor efficacy in a clinically relevant mouse model, which may also imply good pharmacodynamics effects and tolerability.

Study results present novel, simplified, environmental friendly, and cost-effective extraction and purification of EVOO phenolics (Figs 3–5), which will motivate future preclinical and clinical investigations of individual olive phenolics and expand their therapeutic applications. The most bioactive EVOO phenolic OC maintained its in vitro and oral in vivo anticancer activity as freshly prepared nano-emulsion (Figs 6 & 7) suggesting potential future use of this method to develop proper EVOO phenolic formulations for therapeutic and dietary applications.