Gallic acid (≥ 97%), syringic acid (≥ 98%), quercetin hydrate (≥ 95%), oleuropein (≥ 97%), tyrosol (98%) were purchased from Sigma-Aldrich (Steinheim, Germany). Rutin (≥ 90%) was obtained from Wako Pure Chemical Industries, Ltd. (Osaka, Japan), 3-hydroxytyrosol (≥ 98%) from TCI America (OR, USA). Reagents 2,2-diphenyl-1-picrylhydrazyl (DPPH), 2,4,6-tris(2-pyridyl)-1,3,5-triazine (TPTZ) from Sigma-Aldrich, iron (III) chloride (FeCl3), aluminum chloride (AlCl3), Folin-Ciocalteu’s phenol reagent from Loba Chemie (Mumbai, India). 6-Hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox) was provided by Acros Organics (NJ, USA). All solvents used were HPLC grade and provided by VWR International (Darmstadt, Germany). Sunflower oil was obtained from a local market in Sidi Thabet (Ariana, Tunisia).

Olive leaves were obtained from “Chétoui” variety olive groves in Raf-Raf during February 2020. Raf-Raf is in the North-East region of Tunis (latitude of 37°11’ N; longitude of 10°11’ E) at an altitude of 136 m above sea level. A representative sample of old leaves (1 kg) was randomly collected from ten individuals (1 kg/batch). The olive leaves were washed thoroughly with tap water and then dried for 48 h at 40 °C in a Venticell oven (MMM Medcenter Einrichtungen GmbH, Munich, Germany). The dried leaves were ground using a Polymix MFC 90D mill (Kinematica, Couëron, France) to a fine powder and passed through a sieve of 1mm mesh size.

Powdered olive leaves (1 g) were dispersed in 20 mL of a 50% ethanol-water solution (v/v) at a solid/liquid ratio of 1:20 (w/v). Aqueous ethanol was chosen as the extracting solvent due to its efficient recovery of bioactive compounds from olive leaves (Cifá et al., 2018). The suspension was sonicated for 20 min using an ultrasonic probe (Sonics & Materials Inc., Kolkata, India) at 20 kHz with a power of 500 W. During sonication, the dispersion was held in a thermostatic bath at 20 °C. The resulting suspension was centrifuged for 15 min at 2000 rpm. The supernatant was filtered using a 0.45 μm syringe filter and concentrated in a rotary evaporator (Schwabach, Germany). The obtained OLE was freeze-dried (Christ-Alpha, Osterode, Germany) and stored in amber glass bottles at -20 °C until use. The extraction yield of OLE was expressed as % of dry olive leaf weight.

Three different concentrations of OLE (0.1, 0.25, and 0.5%, w/v) were mixed with 25 mL sunflower oil. The sunflower oil initially showed (at T0, time corresponding to its production) 0.59 mg KOH·kg^-1 acid value, 2.1 meq O₂·kg^-1 peroxide value, conjugated dienes (K₂₃₂) and conjugated trienes (K₂₇₀) at 0.93 and 0.3, respectively. It was a linoleic acid-rich oil with 8.2% C16:0; 3.4% C18:0, 27.6% C18:1 and 59.3% C18:2 as main fatty acids. It also contained α-tocopherol and carotenoids with average values of 39.6 mg·100g^-1 oil and 0.34 mg·100g^-1 oil, respectively (data not shown). The different mixtures were homogenized for 20 min at 13500 rpm with an Ultra-Turrax T25 (IKA, Labortechnik GmbH, Munich, Germany). Then, the obtained samples of sunflower oil incorporating OLE were incubated in a forced air circulation oven (Venticell) to accelerate their oxidation at 70 °C for 1 week according to the Schaal oven test. The oxidative stability of treated sunflower oil was evaluated in terms of quality indices and fatty acid composition. Sunflower oil without OLE was used as a control.

Acid value (AV), peroxide value (PV) and ultraviolet spectrophotometric indices at 232 nm (K₂₃₂ value) and 270 nm (K₂₇₀ value) were determined according to the standard protocols of the association of official analytical chemists (AOAC, 2003).

Fatty acid methyl esters (FAME) of sunflower oil fatty acids were obtained by cold transesterification in alkaline conditions. Briefly, 0.1 g of sample dissolved in 2 mL of n-hexane were mixed with 0.5 mL of sodium methoxide in methanol (3 %), followed by methanolic H2SO4 (1 N). The hexane layer containing fatty acid methyl esters (FAMES) was washed with 10% NaCl solution and concentrated under a nitrogen steam before analysis. FAME analysis was carried out on a Hewlett-Packard HP-6890 gas chromatograph (Agilent Technologies, Palo Alto, CA, USA) equipped with a flame ionization detector (FID) and polar TR-FAMEs (Thermo Fisher Scientific Inc., Bordeaux, France) capillary column (60 m × 0.25 mm, 0.25 µm film thickness). The column temperature was initially held at 100 °C for 5 min, raised to 240 °C (4 °C·min^-1), and then kept isothermal for 15 min. The temperatures of the injector and detector were held at 240 °C and 260 °C, respectively. Fatty acids were identified by comparison of their retention times with those of standard FAME mixtures injected in the same conditions (Supelco, 37 Component FAME Mix, PA, USA). The area of each fatty acid peak was integrated by Agilent. ChemStation software and the result is expressed as a percentage of total peak areas.

The Iodine value (IV) and oxidative susceptibility (OS) were determined from the percentages of fatty acids using the following equations (Torres and Maestri, 2006):

All experiments were performed in triplicate and results are reported as mean ± standard deviation (SD). One-way ANOVA with Tukey’s HSD (Honestly Significant Difference) test was used to assess significant differences between means with a significance level of p < 0.05. Data were treated using SYSTAT (Systat Software, San Jose, CA).

Olive leaves, which can be obtained during olive tree pruning or harvesting as well as from olive mills, are a low-cost source of active compounds. An OLE extraction yield of 25 ± 1.7% was obtained using UAE for 20 min at room temperature and a solid/liquid ratio of 1/20 (w/v). Under similar conditions (same extraction procedure (UAE), 50% ethanol and 1:20 solid/liquid ratio) an average extract yield ranging from 16.2-20.12% was recorded by Şahin and Şamli (2013) for olive leaves from an unidentified variety. The observed discrepancy might be attributed to varietal differences as has been reported by Martín-García et al. (2022). Variable extraction yields depending on olive variety and extraction procedure have also been described for Tunisian olives (Taâmalli et al., 2012). Extraction yields ranging from 5.2 to 22.4% were obtained for the varieties Jarboui and Chemchali, using supercritical fluid extraction and pressurized liquid extraction, respectively. Regarding the variety Chétoui, a maximum extract yield (19.5 %) was achieved with pressurized liquid extraction using ethanol as extracting solvent and a temperature of 150 °C (Taâmalli et al., 2012). At this point, we can postulate that UAE (operating at low temperature) was more suitable for the extraction of bioactive compounds from olive leaves without possible thermal alterations, a fact that could justify the orientation of most of the research teams to the use of UAE as a method of choice for extracting useful nutraceuticals from olive leaves (Cifá et al., 2018). Indeed, the acoustic cavitation generated by ultrasound treatment induces cuticle erosion and fragmentation of leaf surface protrusion (hair), the polyphenol-rich structures of olive leaves lead to an increased release of phenolic compounds.

The TPC, TFC and antioxidant activities of OLE are summarized in Table 1. In this study, The TPC of 50% (v/v) ethanol OLE obtained by UAE was 135 mg GAE·g^-1 extract (Table 1). TPC values ranging from 73 to 144 mg GAE·g^-1 extract have been reported for OLE obtained with 70% (v/v) ethanol from eight Tunisian olive leaf varieties (Ben Salah et al., 2012). This value is higher than that reported by Edziri et al. (2019) for the methanol OLE of the “Chétoui” cultivar (47.5 mg GAE·g^-1 extract) obtained by regular maceration. Higher TPC was achieved with MAE by Mohammadi et al. (2016) for “Mission” variety (206.8 mg GAE·g^-1 extract) but using methanol as extraction solvent. The TFC value for the “Chétoui” variety OLE is highly variable and fluctuates between 7.29 - 94 catechin equivalent CE·g^-1 extract (Edziri et al., 2019; Ben Salah et al., 2012). The difference between the obtained value (Table 1) could be attributed to the type of standard used for calibration (CE vs. QE in this study) and to the known limits of this method (Shraim et al., 2021). In fact, Shraim et al. (2021) showed significant variation in the AlCl₃-flavonoid complex response according to the flavonoid type used and the composition of the extract, leading to an under or an overestimation of this parameter.

The profiling of phenolic compounds by HPLC-DAD resulted in the identification of 7 compounds by comparison of retention times and UV-visible spectra with those of standards analyzed under the same conditions (Figure 1). These compounds have been previously reported in the literature (Edziri et al., 2019; Quirantes-Piné et al., 2013) and can be grouped, as shown in Table 1, into simple phenols (hydroxytyrosol, tyrosol), secoiridoids (oleuropein), flavonols (quercetin; rutin) and phenolic acids (caffeic acid; p-coumaric acid). The OLE obtained by UAE is rich in secoiridoids with oleuropein being the major phenolic compound (87.9% of total identified compounds) (Figure 1). This agrees with studies that demonstrated that oleuropein is the main phenolic compound in ethanolic extracts from olive leaves; while hydroxytyrosol and phenolic acids are the most prevalent phenolic compounds in aqueous extracts (Quirantes-Piné et al., 2013). The oleuropein content (32.56 mg·g^-1 freeze-dried OLE) is in the range of the oleuropein contents reported for OLE obtained by maceration in ethanol:water (70:30 v/v) (27.3 mg oleuropein·g^-1 extract) and those obtained by UAE using aqueous ethanol as a solvent (35.6 - 39.2 mg oleuropein·g^-1 extract) (Cifá et al., 2018). Other phenolic compounds were identified, such as caffeic acid (3.0%), hydroxytyrosol (2.4%), tyrosol (2.2%), rutin (2%), p-coumaric acid (1.4%) and quercetin (1.2%). However, due to the lack of analytical standards for confirmation, other known compounds previously identified in olive leaves: luteolin, apigenin, verbascoside and ligstroside derivatives were not assigned (Quirantes-Piné et al., 2013). The differences observed in the phenolic composition of OLE may be related to many factors including variety, cultivar, geographic origin, solvent type and polarity, solvent-solid ratio, extraction method and parameters (Ghasemi et al., 2018).

The antioxidant activity of OLE was evaluated by three complementary tests. As shown in Table 1, average values of 1.01, 1.4 and 0.96 mmol TE·mg^-1 for freeze-dried extract have been observed for DPPH, FRAP and H₂O₂ assays, respectively. The ability of OLE to scavenge DPPH radicals and reduce Fe³⁺ to Fe²⁺ suggests that its antioxidant activity involves a single electron transfer (SET) mechanism (Ivanova et al., 2020). Moreover, OLE also has an effective H₂O₂ scavenging activity. Being a hydrogen atom transfer-based (HAT) mechanism (Ivanova et al., 2020), the result of the H₂O₂ assay indicates that OLE might act as a hydrogen donor. Collectively, the antioxidant activity of OLE could be attributed to the simultaneous presence of electron and hydrogen donors.

The antioxidant activity was found to be correlated with TPC and TFC values (Miliauskas et al., 2004). The phenolic compounds of OLE, mainly oleuropein, may directly contribute to its antioxidant capacity. In a comparative study among four Tunisian olive varieties, the OLE from “Chétoui” was the most effective DPPH radical scavenger (Edziri et al., 2019). Moreover, the ability of OLE to scavenge H₂O₂ and reduce Fe³⁺ has also been evidenced (Martín-García et al., 2022), confirming our findings. The antioxidant activity of OLE appears to be intimately linked to the presence of putative antioxidants which have strong electron and hydrogen donating ability. Among the identified components, some have received particular attention because of their well-known antioxidant properties, including oleuropein, hydroxytyrosol, rutin, gallic and caffeic acids, tyrosol, elenolic acid, luteolin, apigenin and their glycosylated forms (Taâmalli et al., 2012; Tarchoune et al., 2019; Martín-García et al., 2022). In addition to phenolic components, other fat-soluble components such as pigments (carotenoids and chlorophylls) and tocopherols have also been reported as strong radical scavengers and reducing agents, thereby conferring a potent antioxidant activity to OLE (Tarchoune et al., 2019). It should be noted that the overall antioxidant activity resulted from synergistic interactions between the active ingredients present in OLE (Lee et al., 2009).

In this study, the ability of OLE to improve the oxidative stability of sunflower oil was assessed through the determination of some analytical indices and fatty acid composition of sunflower oil. At the beginning of the experiments (before thermal treatment), the acid value, peroxide value, K₂₃₂ and K₂₇₀ of sunflower oil were 1.16 ± 0.07 mg KOH·g^-1, 11.67 ± 0.58 meq O₂·Kg^-1, 2.14 ± 0.18 and 1.28 ± 0.24, respectively (Table 2). The thermal treatment of sunflower oil for 7 days resulted in a 2.9-fold increase in acid value compared to the untreated oil sample at the start of the experiment. This may be attributed to extended lipid peroxidation leading to excessive production of hydroperoxide. The incorporation of OLE significantly decreased (p < 0.05) the acid value with 0.25 and 0.5% concentrations being the most effective in reducing the content of free fatty acids. The ability of OLE to delay the thermal-induced production of free radicals and lipid peroxidation could be due to its antioxidant capacity (Şahin et al., 2017). So, the presence of putative antioxidants in the OLE (e.g. tyrosol, hydroxytyrosol, oleuropein, quercetin, and rutin) could effectively inhibit the formation of free radicals at the initial stage of lipid peroxidation, break the chain reaction during the propagation stage, and/or scavenge the free radicals in the oil by donating a hydrogen atom (Blasi and Cossignani, 2020). Peroxide value (PV) as a key parameter reflecting the early stage of lipid oxidation was determined in all sunflower oil samples. As shown in Table 2, the PV of thermally treated-control oil was increased significantly owing to an extended autoxidation of the oil and its subsequent increase in hydroperoxides. In contrast, the incorporation of OLE reduced peroxide formation and delayed the thermal oxidation of sunflower oil in the early stage, thereby delaying the onset of oil autoxidation. Sunflower oil incorporated with 0.25 and 0.5% OLE showed the lowest PV value among the enriched oil samples. As the PV is closely related to the polyunsaturated fatty acid (PUFA) content, it was clear that the decrease in the PV in sunflower oil incorporated with 0.25 and 0.5% OLE was mainly due to the extended degradation of hydroperoxides to secondary oxidation products such as volatile aldehydes and ketones (Benkhoud et al., 2022). The specific extinction coefficients are indicative of the presence of conjugated dienes (CD) and trienes (CT) (K₂₃₂ and K₂₇₀, respectively). CD as a measure of the oxidative state of the oil is considered a reliable indicator of the effectiveness of antioxidants to prevent lipid oxidation. The results from Table 2 show that the highest CD contents were found in the thermally treated control owing to their high content in polyunsaturated fatty acids, namely linoleic. In contrast, samples incorporated with OLE showed CD values close to the non-treated control, suggesting the protective function of OLE against thermal-induced lipid peroxidation. The CT contents in all samples did not differ significantly (p > 0.05), indicating a similar rate of oxidation. This may be due to the inhibition of the production of CT through the dehydration of CD hydroperoxides and/or the absence of secondary oxidation products (Zhang et al., 2020). Regardless of thermal treatment, Iodine values (IV) were similar in all sunflower oil samples, suggesting that this parameter could not be considered a reliable parameter for estimating the rate of oxidation of sunflower oil. Moreover, the implication of native antioxidants like tocopherols could, at least in part, explain the steady oxidation state of the untreated control and the oil samples incorporated with OLE.

The fatty acid profiles of control untreated and thermally treated sunflower oil samples are presented in Table 2. PUFA made up the highest contribution (53.6-60.6%) with linoleic (53.2-56.2%) and linolenic (0.17-0.46%) acids as the main components. The monounsaturated fatty acids (MUFA) represent the second-most abundant class with percentage contributions ranging from 27.4 to 32.6%. The MUFA fraction was dominated by oleic (29.1-32.1%), followed by palmitoleic (0.63-1.05%) and gadoleic (0.08-0.65%) acids. The saturated fatty acids (SFA) made up the lowest contribution (12.0-13.74%) and the main fatty acids were palmitic (8.39-9.49%), followed by stearic (3.29-4.05%) and behenic (0.40-0.73%). Despite their high nutritional quality, the oil samples were particularly prone to oxidation as revealed by their high oxidative susceptibility (OS) (2465-2594) in comparison with other edible oils like olive oil (453-922) (Benkhoud et al., 2022). At this point, the incorporation of 0.25 or 0.5% OLE confers better resistance to thermal oxidation of sunflower oil and improves its oxidative stability.

The protective effect of OLE against the thermal oxidation of sunflower oil may be attributed to the presence of putative antioxidants such as polyphenols (Samli et al., 2020) and fat-soluble antioxidants, including tocopherols (Jaski et al., 2022), carotenoids and chlorophylls (Şahin et al., 2017; Sousa et al., 2022). In a previous study, it has been reported that the enrichment of olive oil with OLE containing 272 ppm oleuropein resulted in a 2.5-fold increase in the free-radical scavenging ability, thus resulting in enhanced oil stability (Şahin et al., 2017). In another study, hydroxytyrosol was found to be the most effective in inhibiting the formation of conjugated hydroperoxides in bulk fish oil and fish oil-in-water emulsions at a concentration of 100 ppm, mainly due to its electron-donating and ferrous-chelating properties (Pazos et al., 2008). Using antioxidant activity guided-fractionation, Lee et al. (2009) showed that the protective effect of OLE against lipid oxidation was linked to oleuropein, caffeic acid, vanillin, rutin and catechin. More recently, Sousa et al. (2022) showed that the incorporation of chlorophyll-rich extract from Crithmum maritimum markedly enhanced the oxidative stability of sunflower oil. Another point to be considered is that the incorporation of OLE into sunflower oil might not only improve its oxidative stability through its direct action on retarding the lipid peroxidation process, and/or scavenging free radicals, but also indirectly through the protection of α-tocopherol and carotenoids, the natural antioxidants of sunflower oil from thermal degradation empowering thus its oxidative stability. Further studies are needed to confirm this assumption.