Pink-eyed dystrophic Royal College of Surgeons (RCS) rats (rdy/rdy-p) originally obtained from the National Center for Research Resources (NIH, Bethesda, MD) and Sprague–Dawley (SD) wild-type (WT) albino rats originally obtained from Charles River Labs (Wilmington, MA) were raised in a 12 h light/12 h dark light cycle with standard food and water ad libitum and bred to yield litters for RPE isolation. Animal experimentation was conducted according to the ARVO guidelines for the Use of Animals in Ophthalmic and Vision Research and reviewed and approved by the Institutional Animal Care and Use Committee of Fordham University.

ARPE-19 cells were obtained from ATCC and maintained in DMEM/F-12 containing 10% fetal bovine serum (FBS). Primary bovine retinal pigment epithelial cells (RPE) were generated by Dr. Huajun Yan and maintained in RtEBM (Lonza, United States) supplemented with RtEGM SingleQuots Supplement Pack (Lonza, United States). Both cell lines were incubated at 37°C with 5% CO₂. ARPE-19 cells and primary bovine RPE cells were seeded in 12-well plates and grown to full confluency before experiments.

ARPE-19 cells were electroporated with three unique siRNA specific for human MERTK, GAS6, and MFGE8 (Table 1, Integrated DNA Technologies (IDT), United States), and cells electroporated without siRNA was used as a negative control, using Neon™ Transfection System (Thermo Fisher, United States) according to the manufacturer’s protocol. Experiments were carried out 5 days after the transfection procedure. The effectiveness of the MERTK, GAS6, and MFGE8 knockdown was examined using quantitative RT-PCR (qRT-PCR).

Rat primary RPE cells were isolated from 4–9 days-old SD and RCS rats as described previously (Mao and Finnemann, 2012). Mixed sex litters were used. Briefly, the lens and vitreous body were removed from freshly enucleated eyes, and eyecups were dissected following sequential hyaluronidase and trypsin enzymatic treatments. Sheets of RPE were manually collected and re-trypsinized for 90 s. Cells were cultured in DMEM with 10% FBS for 4 days before experiments.

Total RNA was extracted from cells using RNeasy Plus mini kits (Qiagen). RNA concentration was measured by a Qubit fluorescence assay (Thermo Fisher, United States). A total 1 μg of extracted RNA was reverse transcribed using SuperScript IV VILO Master Mix kit (Thermo Fisher, United States), according to manufacturer’s instructions. cDNA was diluted 1:10 in nuclease free water prior to qPCR analysis using a CFX384 real-time PCR system (Bio-Rad, United States). Reaction mixtures containing 5 μL PowerUp SYBR Green Master Mix (Applied Biosystems, United States), 4 μL cDNA, 1 μL primer mix (final concentration of 200 nM per primer) were prepared in Hard-Shell PCR 384-well plates (Bio-Rad, United States). Amplification conditions were as follows: 50°C for 2 min, followed by 40 cycles at 95°C for 15 s and 60°C for 60 s. Melting curve analysis was performed from 65°C–95°C. All samples were run in triplicate and mean Ct (cycle threshold) values were used for further analysis.

Cell pellets were resuspended in 200 μL ddH₂O and transferred into a clean glass test tube. 1 μg 6Z, 9Z, 12Z, 15Z, 18Z-heneicosapentaenoic acid (FA 21:5, Cayman, United States) was added as internal standard. Lipids were extracted by the Bligh–Dyer method followed by hydrolysis and extraction of total fatty acids. Briefly, 750 μL ice-cold 1:2 (v/v) CHCl₃:MeOH was added directly to the sample and the sample was vortexed. Then, 250 μL CHCl₃ was added and mixed well. 250 μL ddH₂O was added and vortexed well to produce a two-phase system. After centrifuging at 3,000 rpm for 3 min, the bottom phase was collected and evaporated under a constant stream of nitrogen. To recover the total fatty acids, 720 μL acetonitrile and 10 μL hydrochloric acid were added to the dried film and the sample was kept at 95°C for 1 h. Then, 1 mL of hexane was added to the sample and vortexed. The upper phase containing total fatty acids was transferred to a new tube and evaporated under nitrogen. The total fatty acids were dissolved in acetonitrile-isopropanol (50:50, v/v) for further LC-MS analysis (Figure 1A).

Tissue samples were homogenized in ddH₂O. 1 μg FA 21:5 was added as internal standard. Lipids were extracted by the Bligh–Dyer method, followed by hydrolysis and extraction of total fatty acids as described above.

Separation of VLC-PUFAs was achieved on an Acquity UPLC^® BEH C18 column (1.7 μm, 2.1 mm × 100 mm, Waters Corporation) using a mobile phase consisting of 1% 1 mol/L ammonium acetate and 0.1% formic acid in water (A) acetonitrile/isopropanol (1:1, v/v), 1% 1 mol/L ammonium acetate, and 0.1% formic acid (B) at a flow rate of 0.4 mL/min, with the following linear gradient: 0–2 min, 35–80% B, 2–7 min, 80–100% B, 7–14 min, 100% B, 14–15, 100–35% B.

For VLC-PUFAs analyses, 14Z, 17Z, 20Z, 23Z, 26Z, 29Z-dotriacontahexaenoic acid (32:6, Cayman, United States) and the extracted mixed VLC-PUFAs from bovine retina was used as VLC-PUFA standards, since most commercial VLC-PUFA standards are not available, and retina contains the highest levels of VLC-PUFAs of any mammalian tissue analyzed thus far (Aveldaño, 1987; Liu et al., 2013; Gorusupudi and Rallabandi, 2021). The identification of each VLC-PUFA in retina samples was conducted extracting a chromatographic peak for each of the corresponding [M-H]⁻ ions using 5 ppm tolerance in full scan mode. Then fragments of the different VLC-PUFAs were studied through a separated parallel reaction monitoring (PRM) method for further identification of the compounds.

The established retention times were used for identification of LC-PUFAs and VLC-PUFAs in phagocytosed POS (Table 2). For quantification, the Q Exactive MS (Thermo Fisher Scientific, Waltham, MA) was operated in a full MS scan mode (resolution 70,000) in negative mode with a scan range of m/z 250–800, and the VLC-PUFAs of interest were quantified by extracting a chromatographic peak for the corresponding [M-H]⁻ ions using 5 ppm tolerance. The comparison was made by normalizing each peak area with the internal standard fatty acid 21:5.

RPE cells were seeded at equal density on 12-well plates to have at least triplicate samples for each condition. Then, the RPE cells were cultured in the cell culture incubator until they were fully confluent.

Bovine POS (InVision BioResources) was counted and resuspended in cell culture medium. POS aliquots were stored at −80°C. For phagocytosis assays, cell number was determined, and appropriate POS concentration (POS/cell) was added. RPE cells were incubated with POS in the cell culture incubator for the appropriate time. Phagocytosis was terminated by washing 3 times with PBS.

To detect internalized POS only by VLC-PUFA-based assay, trypsin containing EDTA (0.25%) was added to RPE cells and incubated in the cell culture incubator until the cells were transparent, but still attached. Trypsin-EDTA, along with most of the surface-bound POS, was removed and the cells were then harvested and washed with PBS for 3 times. The cell pellet was kept at −20°C for further analysis.

For quantification of phagocytosed POS, the lipids were extracted and hydrolyzed. VLC-PUFAs 32:6, 34:4, 34:5, 34:6, 36:6 were analyzed as the markers of internalized POS following the procedures mentioned above.

Immunofluorescence quantification of POS phagocytosis was performed following a previously established protocol (Mao and Finnemann, 2012). Briefly, confluent ARPE-19 monolayers grown on coverslips were incubated with 5, 10, or 20 FITC-conjugated POS/cell for 2 and 5 h. Washed cells were briefly fixed in 4% PFA in PBS for 20 min and the remaining fixative was quenched with NH₄Cl. Cells were blocked in 1% BSA for 10 min and incubated with mouse-rhodopsin antibody (clone 1D4, a kind gift from Dr. David Salom) diluted 1:1000 in 1% BSA for 25 min. After washing 3 times in PBS, cells were incubated with goat anti-mouse AlexaFluor-647 secondary antibody diluted 1:5000 in 1% BSA for 1 h. After washing 3 times in PBS, nuclei were stained with DAPI, and coverslips were mounted with Fluoromount G. Images were acquired using Keyence All-in-One Fluorescence Microscope (BZ-X800, Osaka, Japan) at 100× magnification, and the number of green particles was counted and normalized to the number of cell nuclei using single extract colocalization in Keyence Hybrid Cell Count software.

Opsin immunoblotting was performed following a previously established protocol (Mao and Finnemann, 2012). Confluent ARPE-19 monolayers were incubated with 10 POS/cell for 2 h. Triplicate samples were designated for detection of either total (bound plus internal) POS or only internalized POS. Samples designated for total POS detection were washed with PBS-CM (PBS with calcium and magnesium), while only internalized POS samples were incubated with PBS-EDTA for 10 min. Total protein was extracted with ice-cold lysis buffer (20 mM HEPES pH 7.9, 420 mM NaCl, 3 mM MgCl₂, 10% glycerol with freshly supplemented 5 mM DTT and protease inhibitor cocktail), subjected to SDS-PAGE electrophoresis, transferred onto a PVDF membrane, and blocked with casein in PBS for 1 h. The membrane was incubated with mouse-rhodopsin (clone 1D4) antibody (1:1000) overnight at 4°C. Following 3 washes with PBST, the membrane was incubated with HRP-conjugated secondary antibodies for 1 h. The bands were visualized by a chemiluminescent reaction and imaged using the ChemiDoc Imaging System. Densitometric analysis of obtained bands was performed by FIJI, and the data were normalized to total protein stained by Ponceau S.

POS phagocytosis assays were adapted from published protocols (Mazzoni et al., 2019). Confluent, primary rat wild-type or MERTK-deficient RCS RPE cells in wells of a 48-well plate were challenged with purified POS (~10 POS/cell) in the presence of 1 μg/mL recombinant mouse MFG-E8 and 2 μg/mL human protein S in DMEM at 37°C for 45 min, followed by removal of excess POS, rinse with DMEM, and incubation in DMEM with 2 μg/mL protein S for 1 h to promote further POS internalization. Cells were washed three times with ice-cold PBS, trypsinized for 5 min at 37°C to remove bound POS, resuspended in PBS and pelleted at 500 x g for 5 min at 4°C. Pellets were resuspended and washed twice with PBS. After the final rinse, 10% of the cell suspension was separated for DNA analysis and 3% was separated for western blot protein analysis. Remaining cells were pelleted, and dry cell pellets were stored at −80°C until further western blot or lipid analysis.

For immunoblotting, cell suspension aliquots were lysed in RIPA buffer supplemented with protease inhibitor cocktail. Proteins were separated by reducing SDS-PAGE on 4–20% gradient polyacrylamide gels, transferred to nitrocellulose membranes, and blocked in 10% non-fat milk powder in TBS. Blots were incubated with rhodopsin [clone B6-30, a kind gift from Dr. Paul Hargrave (Adamus et al., 1991)] and alpha-tubulin (Cell Signaling CAT #2125) primary antibodies, and incubated with horseradish peroxidase-conjugated secondary antibodies followed by enhanced chemiluminescence digital detection using a KwikQuant Imager. Bands were quantified by densitometry using ImageJ software.

To obtain the PUFA composition from different samples, we developed a liquid chromatography-mass spectrometer (LC-MS) based method to establish retention times of LC-PUFAs and VLC-PUFAs by using standard VLC-PUFAs, 22:6 (DHA), 24:5 (n − 3), 32:6 (n − 3), 34:5 (n − 3) and 34:6 (n − 3) and mixed VLC-PUFAs extracted from bovine retina as VLC-PUFA standards (see Methods). LC-PUFAs and VLC-PUFAs in samples were extracted using the Bligh–Dyer method and detected using the established retention times in full scan (m/z 250–800) MS under negative mode (Figure 1A and Table 2). The identification of each VLC-PUFA (e.g., from bovine retina samples) was achieved through extraction of a chromatographic peak for each of the corresponding [M-H]⁻ ions (Figure 1B) in full scan (FS) mode. Then, a separate parallel reaction monitoring (PRM) method was also used to obtain the fragments of different VLC-PUFA peaks for identification. The product ion spectra of the precursor ion [M-H]⁻ of 32:6 and 34:5 at m/z 467.3893 and m/z 497.4368 showed the product ions at m/z 423.4004 and m/z 453.4459 by loss of 44 Da at 35% collision energy (CE), while peak clusters with a mass difference of 14 Da corresponding to CH2 with increased CE. A similar fragmentation pattern was observed from DHA standard (Supplementary Figures S1A,B). The linear range and reproducibility of the FS mode-based quantification were further evaluated by using 32:6 and fatty acid extract, respectively. For 32:6, the linear range was from 50 nmol/L to 1,000 nmol/L with a correlation coefficient of 0.994, and the limit of quantification was 5 nmol/L (Supplementary Figure S1C). The standard deviations of the peak area of the 5 VLC-PUFAs of interest in bovine retina extract were below 15% (n = 5) (Supplementary Figure S1D). Bovine and mouse eyecups, as well as RPE cells isolated directly from eyecups, exhibited a high abundance of VLC-PUFAs ranging from 32:4 to 36:6, and the composition resembled that of retina tissue and POS (Figure 1C; Supplementary Figure S2A). On the contrary, cultured RPE cells, such as the widely used ARPE-19 cell line, primary bovine RPE cells, and human induced pluripotent stem cell iPSCs-derived RPE have minimal amounts of endogenous VLC-PUFAs (Figure 1D; Supplementary Figure S2B). This suggests that VLC-PUFAs in isolated RPE cells most probably originated from phagocytosed POS in vivo (Figure 2). Therefore, the deficiency of VLC-PUFAs in cultured RPE cells suggested the feasibility of using the abundance of VLC-PUFAs after incubation with photoreceptor outer segments as a measure of phagocytosis activity in RPE cells.

To evaluate the feasibility of the VLC-PUFA-based method for phagocytosis quantification, we followed the workflow presented in Figure 3A. Briefly, ARPE-19 and primary bovine RPE cells were plated at equal density and grown to confluency in 12 well plates. Then, the ARPE-19 and primary bovine RPE cells were exposed to 20 POS/cell for 5 h and 0.5 h, respectively, following previously published protocols (Mao and Finnemann, 2012). To terminate phagocytosis, cells were washed 3 times with PBS, trypsinized to remove free and bound POS (Westenskow et al., 2012) and collected for analysis. Total fatty acids were extracted and quantified using LC-MS method described above. To assess phagocytic activity, the abundance of VLC-PUFAs 32:6, 34:4, 34:5, 34:6, 36:6 was quantified as markers of internalized POS. All VLC-PUFAs present in POS were detectable in the cells that were challenged with the cargo, i.e., in both ARPE-19 cells and primary bovine RPE cells. An apparent increase was observed for PUFAs longer than 30:4, otherwise not detectable in these cells (Figures 3B,C). We therefore concluded that the lipidomic method is able to detect phagocytosed POS in RPE cells.

Next, we compared our VLC-PUFA-based assay with the current standard fluorescent phagocytosis assay in ARPE19 cells (Mazzoni et al., 2019). To do so, ARPE-19 cells were challenged with fluorescently labeled POS (FITC-POS) using an optimized concentration of POS (20 POS/cell) and incubated for 2 and 5 h. Following standard protocol (Mao and Finnemann, 2012), after fixation and without permeabilization, the cells were briefly stained with an antibody against the C-terminal nine amino acids of bovine rhodopsin (1D4). Internalized FITC-POS appear as green particles on fluorescence microscopy, whereas surface-bound POS appear yellow, due to the overlapping FITC- and secondary antibody (red)-derived fluorescence signals (Figure 4A). Fluorescence-based quantification showed that the number of engulfed POS increased from 21.01 ± 10.49 to 32.12 ± 8.280 (~53% increase) from 2 to 5 h when incubated with 20 POS/cell (Figures 4B,C). Then we assessed whether the new method is applicable for the detection of the dynamic property of phagocytosis. ARPE-19 cells were challenged with POS with varying POS concentrations (0, 10, 20, 40 POS/cell), and POS uptake by the cells was quantified by the abundance of marker VLC-PUFAs. The levels of 32:6, 34:4, 34:5, 34:6, 36:6, which were not detected in unchallenged ARPE-19 cells, showed gradual, significant increases of VLC-PUFA content after incubation with increasing amounts of POS per cell (Figure 4D; Supplementary Figures S3A–C). Similarly, primary bovine RPE cells also showed a significant increase of VLC-PUFAs with increasing amounts of POS per cell and with increasing incubation time (Figures 4E,F).

Next, using the VLC-PUFA-based approach, we attempted to quantify the amount of the individual VLC-PUFAs per POS and per cell after incubation with POS. The result suggested that there is 1.20E−15 mol, 2.01E−15 mol and 9.86E−16 mol 32:6, 34:6, 36:6 per POS, respectively. After incubation with 20 POS/cell for 5 h, the amount of 32:6, 34:6, and 36:6 in ARPE19 cells increased 3.80E−16 mol, 5.85E−16 mol and 3.83E−16 mol/cell, respectively (Figure 4G). This data shows the level of precision that can be attained when VLC-PUFA based method of quantification is applied.

We further evaluated whether the VLC-PUFA-based assay could detect impaired RPE phagocytosis secondary to knockdown of genes known to be key molecules in the binding and engulfment of POS, namely the engulfment receptor MERTK, its ligand GAS6, and the αvβ5 integrin ligand MFG-E8 (Hall et al., 2001; Finnemann and Nandrot, 2006; Nandrot et al., 2007; Law et al., 2015). We performed MERTK/GAS6/MFG-E8 triple knockdown in ARPE19 cells and detected 50, 90, and 60% knockdown of MERTK, GAS6, and MFG-E8 mRNAs, respectively (Figure 5A). Next, wildtype and triple knockdown (triple KD) ARPE-19 cells were exposed to 10 POS/cell, and the VLC-PUFA content was measured after 2 h. The quantification of marker VLC-PUFAs 32:6, 34:4, 34:5, 34:6 and 36:6 decreased 84.82, 76.63, 94.70, 76.50, 77.40%, respectively, (82.01% ± 7.90%) in triple KD cells when compared to the control group (Figure 5B). The opsin immunoblot-based method indicated a decrease of ~91% of rhodopsin content (1D4) from internalized POS in triple KD cells (Figures 5C,D). This decrease was consistent with the result obtained by the VLC-PUFA-based assay, indicating the validity of the proposed method.

To test the potential usage of the proposed assay as a high throughput screening method, the phagocytosis assay was performed in 48-well plates and assessed for reproducibility of VLC-PUFA detection. In brief, 0.03 × 10⁶ cells were plated, grown to confluency, and incubated with 20 POS/cell for 2 h. Total fatty acids were extracted, and the reproducibility of 5 independent experiments was analyzed using the method described above. Coefficients of variation for FA 34:4, 34:6 and 36:6 were 16.65, 16.80 and 19.75%, respectively (Figure 6A).

Finally, we used our newly developed assay to compare the phagocytic capacity of primary RPE cells prepared from WT and mutant rats with known phagocytosis competence and deficiency, respectively. The Royal College of Surgeons (RCS) rat displays a defect in RPE phagocytosis as a result of a mutation in the Mertk gene (Gal et al., 2000). It is a common model used to study the development of cellular therapies for retinal diseases. Therefore, we tested our proposed assay on primary RPE cells cultured from WT and RCS rats. Following POS challenge, total fatty acids were extracted using the method described above and quantified using mass-spectrometry. Compared to WT RPE cells, phagocytosis of RCS RPE cells on 48-well plates was reduced 86.4% by 34:4 based quantification (Figure 6B). This result was in line with earlier publications (Finnemann et al., 1997) and with western blot analyses performed in parallel (93.9%) (Figure 6C).