Reagents were from Sigma unless otherwise stated. Antibodies used for the various experiments are detailed in Supplementary Table S1.

Wild-type (C57BL/6 J) and homozygous MerTK^CR mice (Cai et al., 2016) were housed under cyclic 12-h light/12-h dark conditions (light onset at 8.00 AM) and fed ad libitum. Animals were handled according to the Association for Research in Vision and Ophthalmology (ARVO) Statement for the Use of Animals in Ophthalmic and Vision Research and protocols approved by the Charles Darwin Animal Experimentation Ethics Committee from Sorbonne Université and the French Ministry for Education, Higher Studies and Research (APAFIS#1631–2015090415466433, APAFIS#20191–2019040311402311).

Animals were euthanized using CO₂ asphyxiation either at 10.00 AM (time of the phagocytic peak; visual phenotyping cohorts and electron microscopy) or at different times of the light:dark cycle (analysis of the in vivo rhythm of phagocytosis) depending on the experiment. At the time of euthanasia, mouse ages were as follows: 3 months for the retinal structure (section 3.3) and 18 months for the retinal phenotype cohort, except ERGs collected at each age (section 3.5) or as indicated in the other sections. Complete eyes were enucleated gently using scissors to avoid pulling the retina and fixed immediately in 1 mL Davidson fixative (95% ethanol, 100% formalin [saturated aqueous solution of formaldehyde gas, 37–39%], glacial acetic acid in double distilled H₂O) for 30 min–1 h at 4°C. After making a small opening in the cornea to let the fixative enter, eyecups were further fixed in 1 mL Davidson fixative for 3 h at 4°C. The cornea was then dissected and the lens removed while leaving the iris intact, and eyecups were fixed for 3 more hours at +4°C. Samples were further processed overnight on the Spin Tissue Processor STP 120 (Myr, Thermo Scientific), successively with dehydration and paraffin soaking steps. Eyecups were then paraffin-embedded in molds with the eye oriented, and 5-μm sections were cut.

We usually use sections in the optic nerve area for consistency between samples. Paraffin was removed using two 15-min baths of Safesolv (Q-Path), and slides were rehydrated in gradual ethanol baths.

For immunohistochemistry labelings, antibody sites were unmasked in 1X citrate buffer heated at 95°C, and slides were allowed to cool down in the buffer. RPE pigments and endogenous peroxidases were removed in 5% H₂O₂, 1X SSC in deionized formamide under bright light for 15 min. Sections were permeabilized in 0.3% Triton X-100, 1X TBS, and non-specific signals were blocked in 4% BSA in 1X TBS containing 10% donkey serum. Sections were then incubated with primary antibodies overnight at 4°C (Supplementary Table S1). After three washes in 4% BSA in 1X TBS, sections were incubated with appropriate secondary antibodies (AlexaFluor, Invitrogen) for 1 h at room temperature. Cell nuclei were labeled with 1 μg/mL DAPI for 15 min at room temperature, two more washes were made, and slides were mounted using Vectashield (Vector Laboratories). For autofluorescence analysis, only the nuclei were labeled and no primary or secondary antibodies were applied. Images were acquired on an upright Olympus FV1000 laser-scanning confocal microscope equipped with standard PMTs and highly sensitive GaAsP detectors using the Fluoview 2.1c software. Equivalent stacks of images were compiled for each series and further treated equally for signal output levels using NIH ImageJ (version 1.53o).

Sections were stained with Harris hematoxylin (Diapath) for 3 min, staining intensity was adjusted using a mix of 70% ethanol and 0.3% glacial acetic acid (Amresco, VWR), and another staining step was performed with a 2,000:1 eosin (Diapath):glacial acetic acid solution for 30 s. Samples were then dehydrated, and slides were mounted with coverslips in the xylene-substitute Limonene-Mount mounting medium (Interchim). Images were acquired using a pathology slide scanner (Nanozoomer 2.0HT, Hamamatsu) and extracted using the NDP.view2 software at 20X or 40X magnification.

At least three independent sections were treated by immunohistochemistry and stainings, and representative results are shown in the corresponding figures. For in vivo phagocytosis quantification, phagosomes included in the RPE layer were counted in two areas on each side of the optic nerve of sections 1–2 from n = 3–4 independent mice (except wt samples at 6.00 AM, for which n = 2 mice), the numbers were reported to 100 μm retina width and averaged.

Mice of 8- and 22-months of age were euthanized by CO₂ asphyxiation and eyes were enucleated (8-month-old n = 4, 22-month-old n = 3). Globes were fixed in 4% PFA for 45 min, sectioned at the limbus, and the cornea and lens were discarded. RPE/choroid tissues were separated from the retina. RPE/choroids were incubated overnight with a rabbit anti-IBA1 antibody, rat anti-ZO-1, and AlexaFluor647 phalloidin (Supplementary Table S1) in 1X PBS containing 0.1% Triton. Tissues were rinsed and incubated for 2 h with an AlexaFluor488-conjugated donkey anti-rabbit IgG, an AlexaFluor594-conjugated goat anti-rat IgG, and counterstained with Hoechst 33342 in 1X PBS (all 1:1,000, ThermoFisher Scientific). Retinas were incubated overnight with rabbit anti-IBA1 antibody and AlexaFluor 546 peanut agglutinin (PNA) (Supplementary Table S1) in 1X PBS containing 0.1% Triton. Tissues were rinsed and incubated for 2 h with an AlexaFluor 488-conjugated donkey anti-rabbit IgG and counterstained with Hoechst 33342 in 1X PBS (all 1:1,000, ThermoFisher Scientific).

RPE/choroid and retina (photoreceptors side up) tissues were flatmounted, viewed, and photographed with a Leica DM550B fluorescence microscope (Leica Biosystems). MPs were counted on the whole surface of each RPE/choroid and retinal flatmount. Close-up pictures were acquired using an upright Olympus FV1000 confocal microscope coupled with the Fluoview version 2.1c software (Olympus, Rungis, France). Equivalent stacks of images were compiled for each sample using NIH ImageJ (version 1.53o).¹

Electroretinograms (ERGs) were performed on MerTK^CR (n = 7–10) and control (n = 4–7) mice aged 3–12 months and at 18 months (independent cohorts for each age). After an overnight dark adaptation, mice were anesthetized with ketamine 1,000 (80 mg/kg, Axience) and xylazine (8 mg/kg, Rompun, Bayer HealthCare). Body temperature was maintained at 37°C with a heating pad. Pupils were dilated with tropicamide (Mydriaticum 0.5%, Thea) and phenylephrine chloride (Neosynephrine 5%, CSP). The cornea was locally anesthetized with oxybuprocaïne chlorhydrate (Thea) application. Upper and lower lids were retracted to keep the eye open and proptosed. A small gold wire loop electrode was placed in contact with the cornea to record the retinal response to light stimuli through a layer of Lubrithal (Dechra Veterinary Products) for better contact and to avoid eye dryness. A first needle electrode was placed on the head as a reference, and a second one was placed in the lower back to ground the signal. The light stimulus was provided by a white LED in a Ganzfeld stimulator (ColorDome Lab Cradle, Diagnosys LLC). Scotopic ERG recording was performed using a set of four stimulus intensity levels (0.003, 0.03, 0.3, 3, and 10 cds/m²). Responses were averaged over five flash stimulations. Photopic cone ERGs were recorded on a rod-suppressing background light of 20 cd/m², after a 5-min adaption period. A 10 cd/m² stimulus intensity level was used for light-adapted ERGs. Each photopic cone ERG response was averaged over 10 consecutive flashes. Flicker ERGs were recorded at 10 and 20 Hz. Responses were amplified and filtered (1 Hz-low and 300 Hz-high cutoff filters) with a 1-channel DC-/AC-amplifier. A drop of lubricating cream (Lubrithal) was applied on the unrecorded eye during the overall procedure to avoid the development of any corneal opacity, and an application of oxybuprocaine hydrochloride was made on each recorded eye to reduce animal suffering upon waking. Analysis of ERG signals was done using the Espion V6 software.

Optomotor responses (OMRs) to rotating vertical gratings were recorded on the OptoMotry 1.77 system (Cerebral Mechanics, Canada) on the same cohort of animals between the ages of 3 and 12 months to assess the evolution of contrast sensitivity and visual acuity (n = 10 MerTK^CR, 5 males and 5 females, and n = 4 male controls). Each mouse was placed unrestrained on a central pedestal facing four computer monitor screens. Once accustomed to this environment, the mouse is presented with black and white striped patterns that rotate either clockwise or counter-clockwise as determined randomly by the OptoMotry software. Mice track the gratings with reflexive head saccades that are recorded by the overhead video camera. Spatial frequency thresholds between 0.03 and 0.5 cycles/degree were measured by systematically increasing the spatial frequency of the grating up to 100% contrast and rotating at a speed of 2 rpm. All analyses were assessed by the same experimenter, who was unaware of the genotype. OMR analysis was performed using the OptoMotry 1.7.7 software. Changes in visual acuity were quantified by OMR as described (Prusky et al., 2004; Douglas et al., 2005).

To study the retinal structure using fundus photography (FP) and optical coherence tomography (OCT), mouse cohorts (n = 12 MerTK^CR and n = 5 controls for the FP/OCT cohort, n = 9 MerTK^CR and n = 5 controls for the OCT only cohort) were followed between the ages of 3 and 12 months and at 18 months. Mice were first anesthetized under isoflurane combined with air at a rate of 5% for induction and 2% for maintenance. Pupils were dilated using tropicamide and phenylephrine chloride eyedrops as described above. For FP and autofluorescence (AF) detection, eyes were coated with a lubricating cream (see above) to promote contact with the lamp and prevent the eyes from drying out. Each mouse was placed on a platform with a breathing mask to maintain the anesthesia. The lens of the lamp was then placed in front of the eye. Two pictures were taken, the first one under white light and the second one using a fluorescence filter (488 nm) to detect any AF. OCT imaging acquisitions were made using several positions of the eye facing the lens to obtain a video recording of the different retinal layers through the entire depth of the eye. Video recordings were then extracted, and images were processed using the InvivoVueClinicDb.db software v1.4.0.4260 (Bioptigen).

To isolate RPE from 10- to 12-day-old mice for the primary culture, we enucleated eyecups gently to keep perfect contact between RPE and photoreceptors, and successive enzymatic digestion steps were performed as previously described with slight modifications (Nandrot et al., 2004; Farkas et al., 2014). In brief, after removal of the lens, eyes were treated with 1 mg/mL bovine hyaluronidase (Sigma) in Ca²⁺/Mg²⁺-free Hepes-buffered Hanks saline (Gibco) for 45 min at 37°C to allow an easy peeling of the neural retina and to expose the RPE. A second incubation was performed in 1.5 mg/mL trypsin (Difco) in Hepes-buffered Hanks saline for 45 min at 37°C, and patches of RPE were peeled off manually from the Bruch’s membrane and pelleted gently at 1,200 rpm for 2 min. After another short 1.5-min digestion step in 0.25% trypsin–EDTA (Gibco) and centrifugation, purified RPE cells were seeded into serum-coated 384-well plates (Corning). Cells were then grown at 37°C for 4–8 days before further experiments in MEMα medium complemented with 5% FBS (Hyclone, ThermoFisher), 1% N1 supplement, 1% non-essential amino acids (Gibco), 500 μM glutamine (Gibco), 0.25 mg/mL taurine, 0.02 μg/mL hydrocortisone, 0.0130 pg/mL triiodothyronine, and 1% penicillin/streptomycin (Gibco), a medium recipe already published (Maminishkis et al., 2006).

For peritoneal macrophage isolation and culture, 4-month-old mice were euthanized by CO₂ inhalation, and we followed a previously described protocol (Pineda-Torra et al., 2015). Concisely, mouse abdomens were cleaned with 70% ethanol before incision. Then, the abdominal skin was removed to expose the peritoneum. Into the peritoneal cavity, 5 mL of 1% penicillin/streptomycin 1X PBS was added using a 26G needle, and a little abdomen massage/shaking was done before recovering the liquid. This step was repeated twice, and cells were collected into a 15 mL tube and centrifuged at 300 rpm for 10 min. The pellets were resuspended in RPMI medium (Sigma) supplemented with 1% penicillin/streptomycin, and cells were counted in a Malassez counting chamber. Cells were seeded at a density of 20,000–40,000 cells/well for XFp plates and 120,800 cells/well for 96-well plates (see section 2.10), and the medium was changed after allowing live cells to attach for 1 h. Cells were then grown at 37°C for 24–72 h before metabolic flux and mitochondrial activity measurements.

Retinae were taken out of fresh slaughterhouse porcine eyes, and POS were isolated as previously described (Parinot et al., 2014). In brief, enucleation and tissue collections were performed under dim red light, and retinae were collected in a homogenization solution (20% sucrose, 20 mM tris acetate pH 7.2, 2 mM MgCl₂, 10 mM glucose, and 5 mM taurine). Retina suspensions were obtained after full shaking and gauze filtering, separated on continuous 25–60% sucrose gradients (in tris acetate pH 7.2, 10 mM glucose, and 5 mM taurine) and ultracentrifuged at 25,000 rpm for 50 min at 4°C (Beckman SW 32 Ti swinging rotor). Isolated POS observed as orange bands were collected, diluted, and washed in a series of three solutions (20 mM tris acetate pH 7.2, 5 mM taurine; 10% sucrose, 20 mM tris acetate pH 7.2, 5 mM taurine; 10% sucrose, 20 mM tris acetate pH 7.2, 5 mM taurine). During these wash steps, POS pellets were obtained by centrifugation at 5,000 rpm at 4°C for 10 min each (Beckman JA25.50 rotor) and finally resuspended in DMEM with 2.5% sucrose, counted, aliquoted, and stored at −80°C.

For labeling, resuspended POS were incubated with 1 mg/mL fluorescein isothiocyanate (FITC) (Molecular Probes) for 2 h at room temperature (RT) on a rotator in 10% sucrose, 20 mM sodium phosphate pH 7.2, and 5 mM taurine. FITC-POS were then washed, counted, and frozen as described above.

For phagocytosis quantification assays, confluent and polarized RPE cells were challenged for 1.5 h at 37°C in 5% CO₂ with approximately 10 FITC-POS particles per cell in MEMα, as previously described (n = 5) (Nandrot et al., 2004; Farkas et al., 2014). Subsequently, to remove excess POS, cells were washed three times with 1X PBS-CM containing 0.2 mM CaCl₂ and 1 mM MgCl₂. To quench extracellular fluorescence and determine the quantities of internalized POS, some cells were treated with Trypan blue (Gibco) for 10 min, while non-treated wells permit the measurement of total phagocytosis (Finnemann et al., 1997). Binding is evaluated by the following calculation: total phagocytosis minus internalization. After two washes and a chemical fixation step with 4% PFA for 15 min, all cells were treated with 50 mM NH₄Cl to quench free aldehyde groups before blocking 1% BSA in PBS-CM. For labeling tight junctions, RPE cells were incubated with an anti-rabbit ZO-1 (Supplementary Table S1) overnight and then washed three times with 1% BSA in PBS-CM for 10 min. Secondary antibodies anti-rabbit AlexaFluor594 (1:250) and 647-conjugated phalloidin (1:40, FluoProbes) were incubated for 1 h, followed by two more washes in 1% BSA in PBS-CM. Nuclei were specifically stained with Hoechst (1:1,000, Invitrogen).

Finally, an automated fluorescence microscope was used to quantify FITC/Hoechst ratios by counting the number of FITC-labeled POS and Hoechst-stained nuclei per cell for each well (Arrayscan VTI, HCS Studio software [spot detector v4.1], Thermo Scientific). Immunocytochemistry images were acquired using a spinning disk confocal microscope (CQ1, Confocal Quantitative Image Cytometer, CellpathFinder Software, Yokogawa) with a 20X zoom.

After careful dissection out of the eye socket, complete eyeballs from 18-month-old mice were immersed in the fixing solution (1.5% glutaraldehyde, 1% PFA in 0.1 M cacodylate buffer, pH 7.4) for 2 h at room temperature on a rotator (n = 5). For each sample, the cornea and ora serrata were removed, and the eyecup opened in petals. Four different 1-mm³ RPE/choroid pieces were dissected out and further fixed at 4°C until processing by the electron microscopy core facility (Institut de Biologie Paris-Seine, Sorbonne Université, Paris, France). After three 15-min washing steps in a 0.1 M cacodylate buffer, a post-fixation step was proceeded with in a 0.1 M cacodylate buffer containing 1% OsO₄ for 1 to 2 h on ice. Samples were then washed five times for 15 min in ddH₂O. Afterward, tissue pieces were dehydrated using gradual ethanol bath concentrations (50, 70, 95, and 100%) for 15 min each, followed by two 15-min baths in 100% anhydrous acetone. On a rotator, samples were successively impregnated with different acetone:resin ratios and incubation times (1 h at 3:1 ratio, overnight at 1:1, and 1 h at 1:3), and then with pure resin (two solution changes during the day, followed by an overnight bath). Finally, a polymerization step was carried out at 60°C for 48 h. Ultrathin cuts with a length between 500 μm and 1 mm were obtained using an ultramicrotome (UCT, Leica). Pictures were acquired on a JEM-2100HC scanning transmission electron microscope (JEOL) at 6,000–25,000X.

For Seahorse (Agilent) metabolic flux analysis, RPE/choroid and retina samples from 3-month-old mice were separated, and 1-mm RPE/choroid punches (three for two eyes) were immediately collected in individual wells of a FluxPak Seahorse XFp cartridge in the Seahorse XF base medium minimal DMEM without phenol red (Agilent Technologies) already supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose, adjusted to pH7.4 (n = 1 of each genotype in triplicates per experimental plate, with n = 4–7 experiments for tissues samples and n = 3–5 for macrophages). Punches were equilibrated for 1 h at 37°C in an incubator without CO₂. We used mitochondrial inhibitors oligomycin (1.5 μM), FCCP (carbonyl cyanide 4-(trifluoromethoxy)-phenylhydrazone, 0.5 μM), and rotenone/antimycin A (0.5 μM) as detailed in the XFp Cell Mito Stress Test Kit (Agilent Technologies). Oxygen consumption rate (OCR) and ExtraCellular Acidification Rate (ECAR) were automatically calculated and recorded on the machine by the Seahorse Report Generator online software (Agilent Technologies). Modifications to the original program were implemented to meet our specific experiment requirements: each step is separated into four intervals between drug addition, and three measurements are performed for each step in cycles of 3 min mixing/1 min waiting/2 min measuring. The XFp Glycolytic Rate Assay Kit (Agilent Technologies) was used according to the manufacturer’s instructions.

To evaluate the mitochondrial activity, RPE/choroid fractions (n = 4) and peritoneal macrophages primary cultures (n = 6) from wild-type and MerTK^CR mice were incubated in a black 96-well microplate (Corning) for 30 min at 37°C with MitoTracker Red FM and MitoTracker Green FM fluorescent probes (Invitrogen). MitoTracker Red FM stains active mitochondria, whereas MitoTracker Green FM stains total mitochondria. Both fluorescent probes are diluted in the Seahorse XF base medium minimal DMEM without phenol red (Agilent Technologies) supplemented with 1 mM pyruvate, 2 mM glutamine, and 10 mM glucose, and adjusted to pH7.4 to reach a 200-nM concentration. After two washes with this medium, a microplate reader (Spark, Tecan) counts the different fluorescent spots with 0 μs of lag time and 40 μs of integration time. The TR-F mode is used at 37°C (MitoTracker Red FM 581 ± 10 nm excitation and 640 ± 10 nm emission filters; Mitotracker Green FM 490 ± 10 nm excitation and 516 ± 10 nm emission filters). The percentage of active mitochondria on total mitochondria is then calculated.

After all assays, protein levels were determined for each sample to normalize the results using the Bradford method. A second normalization step was performed for each XFp plate measured on the Seahorse against the average starting value measured for wild-type samples. Experiments were repeated 5–7 times for each genotype, averaged, and the most distant outliers were removed.

Statistical significance of results was determined using the unpaired t-test with Welch’s correction or the Holm-Sidak method for multiple comparisons or with 2-way ANOVA with a Sidak correction for multiple comparisons, all providing adjusted p-values, each row being analyzed individually without assuming equal s.d. Significance thresholds were set as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

We first analyzed the overall structure and organization of both retina and RPE in the absence of MerTK activity downregulation. Hematoxylin/eosin stainings clearly show that the MerTK^CR retinae developed normally and did not appear to change with age up to 18 months (Figure 1A). Photoreceptor inner and outer segments were well-aligned facing RPE cells. Rhodopsin, PKCα, blue and cone opsins, and cone arrestin labelings demonstrated that MerTK^CR rods and cones organization is normal (Figures 1B,C).

To evaluate the visual capacities of MerTK^CR mice at the eye and brain levels, we followed up electroretinograms (ERG) and optomotor (OMR) responses in independent cohorts between the ages of 3 and 12 months and at 18 months. Scotopic profiles and measures of a- and b-wave amplitudes (Figure 2A), as well as implicit times (data not shown), were similar between wild-type controls and transgenic MerTK^CR mice. Photopic ERG and oscillating potentials were also equivalent in both strains, and only flickers seemed to be slightly attenuated in MerTK^CR mice (Figure 2B). Optomotor responses did not differ either, and slightly decreased with age in all animals, whatever their genotype (Figure 2C).

To characterize the evolution of each mouse eye fundus with age, we imaged a separate cohort of mice using monthly fundus photography (FP) and optical coherence tomography (OCT) between the ages of 1 and 12 months. In MerTK^CR mice, we observed the appearance of degenerative areas in the dorsal retina close to the optic nerve, with a marked rim and in the middle of which RPE cells became visible (Figure 3A). OCT thickness was reduced in these areas mostly at the outer nuclear layer level, thus confirming photoreceptor cells loss. Interestingly, these areas also displayed autofluorescence AF punctae at the fundus, which seemed to correspond to “bumps” observed between RPE cells and dying photoreceptors on OCT images. Once detected, these degenerative areas did not seem to spread too much in size with time. Indeed, we repeated monthly FP on naive 10-month-old animals, and MerTK^CR mice did develop degenerative areas of similar size and location as younger mice (Supplementary Figure S1A). As another noticeable feature of these degenerative areas, we noticed that the phenotype did not develop at the same pace in all mice. Indeed, males seemed to be affected earlier than females, i.e., with fewer FP repeats, and left eyes showed degeneration before right eyes (Supplementary Figure S1B).

Strikingly, fundus photographs of mice from the ERG and OMR cohorts taken just before sacrifice suggest that no abnormalities were detectable in any control or transgenic mouse. To understand this discrepancy, and after excluding differences in lighting and food regimen, we set out another cohort that we followed only with monthly OCTs to test if the phenotype was due to the repeated fundus photographies. We did not observe any degeneration in these mice, thus suggesting that the phenotype might indeed be related to repeated illumination of their retina with the fundus lamp (Figure 3B).

We analyzed more closely the areas of photoreceptor loss on sections stained with hematoxylin and eosin or after simple nuclei labeling and confocal microscope imaging (Figure 4A). We could clearly see the loss of photoreceptor nuclei that led to the overall thinning of the retina. In contrast, other retinal layers did not seem impacted. Interestingly, degenerated areas showed increased AF inside RPE cells detected on retinal sections. To further characterize the “bumps” observed on OCT scans, we labeled sections using the microglia/macrophage-specific Iba1 and CD11b markers, recognizing monocytes and macrophages. Both proteins were expressed in RPE cells, especially in degenerated areas. Moreover, we detected a signal around the nuclei of cells inserted between RPE cells and degenerated photoreceptors in MerTK^CR retinae. To verify that the presence of microglial cells/macrophages was increased in MerTK^CR mice, we performed RPE/choroid and retina flatmounts at 8 and 22 months of age (Figure 4B). The Iba1 labeling confirms that, in both young and old mice, more microglial cells/macrophages were detected between the RPE and photoreceptors in MerTK^CR mice when compared to wild-type mice, a difference that was significant at 22 months of age.

POS phagocytosis is a multi-step process initiated by the specific recognition of POS (binding) to alphavbeta5 integrin receptors and the subsequent activation of intracellular signaling pathways leading to MerTK activation (Nandrot et al., 2004). MerTK function is required for POS uptake inside the cells, the internalization step of phagocytosis (Feng et al., 2002). As we showed previously, MerTK cleavage appears to regulate its activity, so we assessed the impact of the absence of MerTK cleavage and, thus, the deregulation of MerTK function on RPE phagocytosis (Law et al., 2015). Total phagocytosis, corresponding to the number of POS bound at the cell surface plus the number of POS that have been internalized, of mouse primary RPE cells from MerTK^CR mice was decreased by 35% (± 11%) when compared to wild-type control cells (Figure 5A). When looking more closely at the two subsequent steps of phagocytosis, the reduction was more pronounced for POS binding (−23% ± 14%) than for their internalization (−15% ± 8%, non-significant). However, the internalization percentage compared to total phagocytosis was similar for control (62% ± 14%) and MerTK^CR (72% ± 37%) RPE cell cultures.

One of the cardinal features of POS phagocytosis by RPE cells is the circadian regulation of its activity (LaVail, 1976; Nandrot et al., 2004). We thus proceeded to characterize the daily rhythm of POS phagocytosis in MerTK^CR mice along eight time-points of the light:dark cycle. The overall profile was modified with a peak of activity occurring at light onset (8.00 AM) earlier than in wild-type mice, followed by an immediate significant decrease at the time of the normal phagocytic peak (10.00 AM) (Figure 5B). Interestingly, in the afternoon phagocytosis is usually at its lowest, but in MerTK^CR mice we detected a significant increase of phagocytosis at 16.00 PM. In addition, in these two marked peaks, the overall MerTK^CR profile appears more homogenous during the rest of the light:dark cycle than in control animals. Notably, the total amount of phagocytosis over the quantified 24-h period was equivalent in both strains, amounting to 58.9 ± 1.7 phagosomes per 100 μm retina for MerTK^CR mice versus 58.1 ± 1.7 phagosomes for control mice. Similarly, levels of total MerTK proteins were equal in both strain RPE/choroid and retinal samples (data not shown).

We investigated in more detail the structure and function of MerTK^CR RPE cells. At the ultrastructural level, RPE basal infoldings and apical microvilli seemed normal in MerTK^CR mice (Figure 6A). Similarly, POS appeared well organized and aligned, and with proper contact with RPE microvilli, and phagosomes were present. However, vacuoles of different sizes were visible in older MerTK^CR but not wild-type control mice. Moreover, electron-dense phagolysosomes were present in transgenic mice, suggesting some defective POS processing.

To understand if the observed differences in the phagocytic function and potential cellular stress could be linked to defective metabolism, we then investigated the energetic metabolism of freshly dissected RPE/choroid samples (Figure 6B). We detected a significant decrease in both basal and maximal respiration as well as of ATP production at 3 months that seems to worsen with age. In these samples, no difference was seen in the non-mitochondrial oxygen consumption, thus suggesting that other energy sources might not be affected. As they also express MerTK, we investigated macrophages from the peritoneum, and, in contrast with RPE/choroid samples, they did not show any difference in any factor studied (Figure 6C). However, the analysis of macrophage glycolysis identified a lower basal and significant higher compensatory glycolytic function in mutant peritoneal macrophages (Figure 6D). Interestingly, both MerTK^CR RPE/choroid and macrophages had similar percentages of active mitochondria when compared to wild-type controls (Figure 6E).