Eleostearic Phospholipids as Probes to Evaluate Antioxidants Efficiency against Liposomes Oxidation

Regardless of the applications: therapeutic vehicle or membrane model to mimic complex biological systems; it is of a great importance to develop simplified, reproducible and rapid model assays allowing for a relevant assessment of the liposomal membrane oxidation and therefore antioxidant activity of selected molecules. Here, we describe a new and highthroughput assay that we called “Vesicle Conjugated Autoxidizable Triene (VesiCAT)”. It is based on specific UV absorbance spectral properties of a new phospholipid probe, synthesized with natural conjugated eleostearic acid extracted from Tung oil. The VesiCAT assay has been developed with two different radical generators (2,2’-azobis(2-amidinopropane) dihydrochloride; AAPH and 2,2'-azobis(2,4-dimethylvaleronitrile); AMVN), producing a V er si on p os tp rin t Comment citer ce document : Durand, E., Delavault, A., Bourlieu, C., Lecomte, J., Barea, B., Figueroa Espinoza, M.-C., Decker, E. A., Salaun, F. M., Kergourlay, Villeneuve, P. (2017). Eleostearic Phospholipids as Probes to Evaluate Antioxidants Efficiency against Liposomes Oxidation. Chemistry and Physics of Lipids, 209, 19-28. DOI : 10.1016/j.chemphyslip.2017.10.006 3 constant flux of oxidant species, either in membrane or in aqueous phase. This method appears very efficient in assessing the effect of various pure antioxidant molecules in their ability to preserve liposomes from oxidative degradation. In addition, the AAPHand AMVN-induced oxidations offer the possibility of extracting different but complementary information with respect to the antioxidants efficacy.


INTRODUCTION
Liposomes consist in spherical phospholipids vesicles which stabilize an aqueous core from the external medium 1 . Over the past decades, several commercial applications of liposomes have been developed to deliver bioactive molecules or therapeutic agents into their target zone of activity within organisms (e.g., cells, tissues) while improving clinical efficacy and limiting side effect linked to non-target delivery 2 . Modulation of liposome structure is easy since both the alkyl chains or polar heads of phospholipids can be tailored to modify vesicle size, charge but also membrane physical state or fluidity 3 . Liposomes present different physical structures that depend on their chemical composition and their method of preparation 4 . Multilamellar vesicles (MLV) contain several bilayers surrounding each other whereas unilamellar liposomes are made of a single bilayer. These latter can be distinguished as small unilamellar vesicles (SUV; diameter <100 nm) or large unilamellar vesicles (LUV; diameter >100 nm).
Multi-vesicular vesicles (MVV) correspond to smaller vesicles trapped into a large vesicle.
In the context of their use as bioactive delivery vehicles, liposomes exhibit drawbacks owing to their relative poor physical or chemical stability 5 . For example, depending on environmental conditions (e.g., temperature, ionic strength, pH), they may undergo 4 aggregation phenomena 6 . Similarly, liposomes are prone to chemical degradation mainly via oxidation of their fatty acid constitutive moieties. Despite the relative simplicity of liposomal lipids, oxidation mechanisms remain complex because they depend on the oxidation inducer, the antioxidant's partitioning and activity, but also on both the composition and physical properties of the liposomes 7 . For example, the size and number of layers in the liposome impact its stability 3 . Apparently, LUV maintain their structural integrity when exposed to reactive oxygen species (ROS) generated outside the bilayer (in aqueous media). LUV are more sensitive to oxidation than MLV and they avoid the problem of partial accessibility of the oxidation inducer to external lipid bilayer as in MLV 7 . The oxidation of liposomes can be limited using the same strategies that are employed for the protection of classical oils (made of triacylglycerols). Such strategies involve the limitation of high temperature, light or oxygen exposure, as well as the use of exogenous antioxidants that can act by various mechanisms (e.g., metal chelators or radical scavengers) 8 .
In addition to their therapeutic use, liposomes have been used extensively as models for in vitro lipid oxidation studies 9 . The involvement of oxidants in several pathological disorders, including cancer, diabetes, cardiovascular diseases, chronic inflammatory disease, postischaemic organ injury, neurodegenerative disorders, and xenobiotic/drug toxicity has been widely documented [10][11][12] . No matter how oxidation is involved in tissue injury in human disease (origin of the disease or simply produced during the development of the damage), the mimetic bilayer structure of liposome represents an interesting tool to investigate the antioxidant potential of a compound, as well as drug-membrane interaction.
In a complex system where liposomes are involved, either in food, cosmetic or pharmaceutical formulations, the efficiency of these antioxidants would be governed, not only by their chemical reactivity, but also by their interactions with other components and their localization in the concerned systems. In particular, the most efficient antioxidants will be the unclear how this probe in its free fatty acid form could affect antioxidant effect due to its undefined membrane anchoring or because it may alter physical properties of membranes.
C11-BODIPY 581/591 (4,4-difluoro-5-(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-indacene-3-undecanoic acid), initially developed by Naguib (1998) 18 , is another fluorescent probe that, once incorporated in liposomes, is extensively used as oxidizable substrate to evaluate lipid membrane peroxidation. This fluorescent fatty acid analog, which the BODIPY core is connected to a phenyl moiety via a conjugated diene, displays bright red fluorescence. This substrate is highly oxidizable by peroxyl radicals (ROO . ) and its oxidation leads to gradual extinction of the fluorescent signal. However, Huang et al (2002) 19 showed that the C11-BODIPY 581/591 probe could undergo photobleaching and lose 30% of its fluorescence in the absence of AMVN. This suggests that, like cis-parinaric acid, C11-BODIPY 581/591 is photosensitive and should thus be taken into account. More recently, other probes corresponding to BODIPY conjugates of α-tocopherol were also used to evaluate the activity of antioxidants in liposomes 20,21 . The mechanism of these probes relies on the reaction of the In 2008, our group has developed a new method to evaluate antioxidants capacity in oil-inwater emulsion, using Tung oil as oxidizable substrate which is particularly rich in trieleostearin 24 . Due to its conjugated triene, eleostearic acid exhibits a unique UV absorbance spectrum that is very convenient to estimate its oxidation rate by UV spectrophometry equipment (micro-plate reader). This method (CAT assay) is now used to assess the potential of natural antioxidants or plant extracts 25,26 , phenolipids 27-29 , synthetic antioxidants 30 or essential oils 31 . More recently, this method was also adapted to the use of a lipophilic azo initiator, namely AMVN to compare the behavior of hydrophilic and lipophilic antioxidants 32 .
Herein, we propose a further version of this eleostearic-based emulsion assay, adapted to vesicle suspension. For this, an eleostearic phospholipid probe was synthesized, and its concentration in artificial membrane suspension was fine tuned in order to visualize its natural absorbance using microplate reader. Then, the conditions were established with the aim at promptly and efficiently probe the membrane oxidation by simply following the eleostearic phospholipid absorbance decay. Finally, this new method called "Vesicle Conjugated Autoxidizable Triene (VesiCAT)", was tested with addition of diverse antioxidants. Either in aqueous or membrane region, linear and reproducible responses over the concentration range were observed, with respect to liposomes induced peroxidation and antioxidant efficacy. were all purchased from Sigma-Aldrich (USA).

Synthesis of 1,2-α-Eleostearoyl-sn-glycero-3-phosphocholine (DEPC) adapted from 36 :
To a solution of sn-Glycero-3-phosphocholine (GPC) (0.100 g, 0.4 mmol), pure α-eleostearic acid (0.556 g, 2 mmol) was added, in alcohol-free and anhydrous CHCl3 (6 mL To normalize data, the raw absorbance signal was transformed in relative absorbance according to the Equation 1 Relative absorbance = Abst / Abs0 (1) Where Abst and Abs0 are absorbances measured at times t and 0 min, respectively. It is worth mentioning that if the measurement is not rapid enough after initiating the oxidation, the Abs0 for the blank (without antioxidant) may be lower than the sample containing the antioxidant.
In this case, to normalize Abs0, the experimental Abs0 of blank can be artificially replaced with the Abs0 of samples in Equation (1). The area under curve (AUC) corresponding to relative absorbance decay was then calculated as follows The net protection area provided by an antioxidant sample was then calculated using the difference between the AUC in the presence of an antioxidant sample (AUCSample) and the AUC of the blank (AUCControl), the latter consisting of the same mixture without antioxidant.
Trolox was used as a calibrator for antioxidant capacity measurements. Thus, the antioxidant capacity of a sample relative to Trolox (VesiCAT value) is given as: For this, a two-steps procedure ( Figure.   We also studied the natural oxidation of the DEPC in a liposomal system in the absence of azo initiator (data not shown). However, we observed that oxidation kinetics were much slower (~days) than artificially induced oxidation (~hours), which appears to be incompatible with high-throughput purpose. That is why the use of azo initiators was chosen. Indeed, despite their artificiality, they are easy to use, enabling a constant, fast, and temperaturecontrolled rate of peroxidation. Two different radical generators (namely AAPH (hydrophilic) and AMVN (lipophilic)) were used with the aim at comparing water-soluble and interface membrane radical initiation. We first adjusted the main parameters (e.g., oxidizing conditions, liposome preparation, DEPC concentration), and then validated the method by using some phenolic antioxidants.

VesiCAT with water radical initiation (AAPH)
The method was developed with the simplest liposomal system, made with non-oxidizable phospholipids having different alkyl chain lengths, namely DLPC and DPPC. For this, two different large unilamellar vesicles, prepared with a blend of DEPC/DLPC and DEPC/DPPC, were made to evaluate how membrane structure can affect the kinetics of oxidation and antioxidant response. Liposomes suspensions were oxidized at 34.5 ± 0.5 °C (liquid temperature in wells) with a constant flux of radical initiators generated in the aqueous phase by thermo degradation of AAPH. For this, the AAPH concentration was fixed at 2 mM, which represents a good compromise between the oxidation rate and the substrate/azo initiator ratio.  Figures. 3 and 4). However, for methods utilizing AUC, there is a clear starting point and a clear endpoint, and its calculation exploits both inhibition time and degree of oxidation, thus reflecting the different reaction kinetics. From those reasons, we believe that assays using AUC provide global information, whereas other approaches may give more specific data. However, kinetic parameters such as lag phase duration and initial rate can be also evaluated to get more insight on the oxidation mechanisms.
In the AAPH-induced oxidations, the net protection AUC of the reference (AUCTrolox -AUCControl) versus concentration, allowed for a perfect linear relationship (R 2 ≥ 0.997), with values equivalent to 9.8-fold ± 0.4 the Trolox concentration in DEPC/DLPC ( Figure. 3) and 8.3-fold ± 0.5 the Trolox concentration in DEPC/DPPC membranes ( Figure. 4). In addition, various antioxidants at different concentrations were tested in the different membrane composition ( Figures. 3 and 4).

VesiCAT with membrane radical initiation (AMVN)
The same approach was performed using liposoluble AMVN as azo initiator. AMVN is a synthetic azo-compound that dissociates to form C-centered free radicals and then peroxyl radicals by oxygen reaction, in the hydrophobic phospholipid bilayer. Unlike water-based initiators, AMVN must be added before vesicles formation. For that reason, the VesiCAT method was described with DLPC membranes, where the lower liquid-crystalline phase transition temperature than DPPC allowed preparing of LUV at low temperatures (20° C) without triggering AMVN radical production and initiation of fatty acid oxidation. Liposome  Figure. 7). Through calculation of the AUC80, good linear relationships (R 2 ≥ 0.98) were established between net areas at 80% of DEPC oxidation (AUC80_Antioxidant-AUC80_Control) and antioxidant concentrations ( Figure. 8). Unlike the AAPH-induced oxidation, Trolox appeared to be the best antioxidant. Contrariwise, chlorogenic acid showed the poorest antioxidant efficacy, whereas it was one of the best ones with AAPH assays.

VesiCAT assays: comparison of the results
It appears that the order of effectiveness of the AAPH-induced oxidation assays is gallic acid < Trolox < quercetin ~ chlorogenic acid < rosmarinic acid, whereas it is chlorogenic acid < gallic acid < quercetin < rosmarinic acid < Trolox for AMVN-induced oxidations. When the equation (3)   . In addition, Trolox seems to have the best mobility and/or distribution in membrane, since it showed a low antioxidant efficacy in the AAPH assays, but the highest net protection in AMVN assay. Conversely, chlorogenic acid, which is quite a good radical