*2.2. Omega-3 Supplementation*

Fatty acids (FA) were conjugated to albumin to solubilize them and to replicate the in vivo environment. Cells were seeded with a lipid-free bovine serum albumin (BSA) media in a 2:1 FA/BSA ratio.

Ten different omega-3 supplements were tested grouped in three parts as shown in Table 1: I: EPA and DHA, separately; II: EPA and DHA combined in different proportions and in the form of ethyl esters (EE) or triglycerides (TG); III: DHA-TG combined with DHA phospholipids (PL) from different origins (marine or vegetable). All the supplements were provided by Théa Laboratories (Clermont-Ferrand, France). Details of the composition of each formulation and identification of groups are displayed in Table 1.

#### *2.3. Immunofluorescence Detection of Zonula Occludens*

The effect of all FA groups on the integrity of the epithelial tight junctions was determined by ZO-1 immunofluorescence to determine if any formulations compromised these intercellular unions. One hundred thousand ARPE-19 cells were seeded onto polycarbonate inserts (Corning® Transwell®, Phoenix, AR, USA,) and kept in a culture with 1% DMEM for four weeks to allow epithelial polarization. Supplements were then added to ARPE-19 cells, which were fixed in cold methanol after 24 h. After three washes with 1% PBS, cells were submerged in 1% PBS-BSA for 20 min to block non-specific bonds before incubation with the polyclonal ZO-1 anti-rabbit antibody (1:100; Life Technologies, Gaithersburg, MD, USA) in 1% PBS-BSA at 4 ◦C for 24 h. The cells were washed with 1% PBS (three times of 5 min) and incubated with the secondary antibody goa<sup>t</sup> anti-rabbit Alexa Fluor® 488 (1:250; A11008 Invitrogen, Thermo Fisher, Madrid, Spain) in 1% PBS-BSA for 1 h in the dark at room temperature. TOPRO-3 was used for nuclear staining. Membranes were cut from the transwell insert with a scalpel and placed on a microscope slide and mounted with PBS-Glycerol

1:1. Images were captured in the Z-stack mode with a laser scanning confocal imaging system (Zeiss LSM-510 Meta, Oberkochen, Germany) using a 40× objective.

**Table 1.** Composition of the different groups of n3-PUFA (long-chain n-3 polyunsaturated fatty acids) supplements. All fatty acids (FA) are presented as a percentage of n3-PUFA, the remaining percentage in each group consisted of a non n3-PUFA diluent. EPA: Eicosapentaenoic acid; DHA: Docosahexaenoic acid; TG: Triglycerides; EE: Ethylesters; PL: Phospholipids.


#### *2.4. Cellular Viability and Proliferation Assays*

The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) reduction assay was used to determine cellular viability. Experiments were carried out on 96-well plates seeded with 10,000 ARPE-19 cells per well. Once cells were confluent, a culture medium was changed to 1% FBS and maintained for two days. At this point, cells were exposed to 100 μM of each supplement. After 96 h, cell viability was analyzed using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega, Madison, WI, USA), following the manufacturer's instructions. At 77 h MTT absorbance was determined at 450 and 540 nm with a Sunrise-basic Microplate reader (Tecan, Austria).

To assess the effect of the different formulations under oxidative stress conditions, at 72 h some cells were additionally exposed to 800 uM H2O2.

Cell proliferation was quantified by BrdU incorporation into the ARPE-19 genome using the Calbiochem® BrdU Cell Proliferation Assay (Calbiochem, La Jolla, CA, USA). Between 10,000 to 20,000 cells were seeded onto 96-well plates. After 24 h, 50 μM of the 10 supplements were added to the cells. BrdU was performed 48 h later, according to the manufacturer's protocol. Some cells were also challenged with H2O2 (as described above).

#### *2.5. Reactive Oxygen Species (ROS) Detection*

To measure the production of reactive oxygen species (ROS) generated by the n3-PUFA, the DCF (2-,7--dichlorofluorescein) (H2DCFDA) test was used. Upon being oxidized, DCF emits fluorescent light at a wavelength of 540 nm when excited at 480 nm. The emitted fluorescence is captured and quantified by fluorometry. Once ARPE-19 cells reached confluence, a 10% FBS culture medium was replaced by 1% FBS medium for 24 h. Next, cells were exposed to the different supplements during 24 h. Apigenine was used as a positive control. Following this incubation, H2DCFDA was added to the medium for 30 min; cells were then harvested from the wells and transferred into a black 96-well plate to measure the emitted fluorescence.

#### *2.6. Caspase-3 Immunofluorescence (IF)*

The protein caspase-3 is an essential mediator of the activation of the apoptotic signaling cascade. To detect if the omega-3 formulations activated cell death by apoptosis, this early apoptotic mediator was visualized by confocal immunofluorescence microscopy. For these experiments, 50,000 ARPE-19 cells were cultured on 1 cm diameter cover slips (CB 100RA1, Menzel-Gläser, Braunschweig, Germany) until confluence. Then, cells were treated with a supplement for 24 h and subjected to 800 μM H2O2 for 3 h to activate the caspase-mediated apoptotic cascade (positive control). At the end of the experiment, glass coverslips were fixed and permeabilized in cold methanol. Afterwards, coverslips were washed with 1% PBS and submerged in 1% PBS-BSA for 20 min to block nonspecific bonds. The polyclonal rabbit anti-caspase-3 (G7481, Promega, WI, USA) was used at a 1:100 dilution in 1% PBS-BSA at 4 ◦C for 24 h. The coverslips were then washed with 1% PBS and incubated with an Alexa Fluor ® 488 goa<sup>t</sup> anti-rabbit IgG antibody (A11008, Invitrogen, Carlsbad, CA, USA) diluted in 1% PBS-BSA in the darkness for 1 h. TOPRO-3 was used as a nuclear marker. Next, the coverslips were mounted on microscope slides with PBS-glycerol and gelatin at 1:1 and observed under the confocal fluorescent microscope (LSM 750, Carl Zeiss, Oberkochen, German).

For caspase-3 quantification, each coverslip was divided into eight sectors and one image at 40× was acquired from alternate sectors. A total of four images were acquired from each cover slip and were analyzed using the LSM Zeiss software to compile them into one merged image, and caspase-3 granules per nucleus were quantified using a home-made plugin tool developed for Fiji/ImageJ, an open-source Java-based image analysis software. The plugin has been developed by the Imaging Platform of the CIMA Universidad de Navarra.

#### *2.7. Wound Healing Cell Migration Assay*

A wound healing assay was used to quantify ARPE-19 migration in the presence of n3-PUFA under standard conditions. For these experiments, 150,000 cells were seeded onto 24-well culture plates until confluence. A linear wound was then created in the middle of each well using a 20 μL sterile micropipette tip. Culture media was replaced to eliminate floating cells and debri and the di fferent n3-PUFA formulations were added. Five points on each well were captured every hour for a total of 72 h using an automatic phase contrast inverted microscope equipped with a digital camera (Carl Zeiss, Oberkochen, Germany). Every set of images was analyzed using the Fiji software (a distribution of ImageJ) V1.48q (Fiji Wound Healing Tool by Nathalie Cahuzac, and Virginie Georget, http://dev.mri.cnrs.fr/projects/imagej-macros/wiki/Wound\_Healing\_Tool) to determine the speed of closure.

#### *2.8. Western Blotting for Vascular Endothelial Growth Factor (VEGF)*/*Pigment Epithelium Derived Factor (PEDF) Ratio*

Following 24 h of treatment with the di fferent omega-3 formulations, 5 μg of ARPE cells homogenates were mixed with a Laemmli bu ffer (62.5 mM Tris-HCl, pH 6.8; 2% SDS; 10% glycerol; 0.1% bromophenol blue) and boiled for 5 min. Samples were separated on 12% SDS-PAGE gels and transferred to a nitrocellulose membrane. After blocking with 5% skimmed milk ( *w*/*v*), 0.1% Tween-20 ( *w*/*v*) in TBS for 1 h at room temperature, membranes were exposed to the primary antibodies (0.2 μg/μL, monoclonal anti-VEGF, sc7269, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA and 1:1000 monoclonal anti-PEDF; MAB1059; Millipore, Burlington, MA, USA), at room temperature for 1 h. Membranes where then incubated at room temperature for 1 h with a horseradish peroxidase-conjugated goa<sup>t</sup> anti-mouse antibody (sc2005; 0.4 μg/μL, Santa Cruz Biotechnology Inc., Dallas, TX, USA). Signals were detected with an enhanced chemiluminescence (ECL) kit (ECL Western blotting detection kit, GE Healthcare, Fairfield, CT, USA) and with ImageQuant 400 (GE Healthcare). The relative intensities of the immunoreactive bands were analyzed with Quantity One software (version 4.2.2, Bio-Rad Laboratories, Hercules, CA, USA). The loading was verified by Ponceau S red, and the same blot was stripped and reblotted with an anti-β-actin monoclonal antibody (Sigma-Aldrich, St Louis, Mo, USA) to normalize the VEGF and PEDF levels. Protein levels were used to calculate the VEGF/PEDF ratio.
