Evaluation of antioxidant treatments for the modulation of macrophage function in the context of retinal degeneration

Monocyte isolation and hMDM polarization

Monocyte/hMDM culture was performed as previously described [11]. Patients with nvAMD (n=8, five women, three men, mean age ± standard error of the mean [SEM]: 77.4±4.7 years, range: 66–92) were recruited from the retina clinic of the Department of Ophthalmology at the Hadassah-Hebrew University Medical Center. Criteria for inclusion of patients with nvAMD included age over 55 years, diagnosis of AMD according to the Age-Related Eye Disease Study Research Group (ARED) [20] criteria, and diagnosis of CNV according to fluorescein angiogram and optical coherence tomography. Eyes with neovascular lesions that comprised less than 50% active CNV, subretinal hemorrhage greater than 25% of the lesion size, or the presence of other retinal diseases were excluded from the study. Specifically, eyes with any other potential cause of CNV, such as myopia, trauma, or uveitis, were excluded. Also excluded were patients with a major systemic illness, such as cancer, autoimmune disease, congestive heart failure, or uncontrolled diabetes. All patients signed an informed consent form, and the study was approved by the institutional ethics committee (see the Ethics Declaration). Blood samples (30 ml) were collected from patients with nvAMD in EDTA tubes (BD Bioscience, Franklin Lakes, NJ). Monocytes were isolated from whole blood, differentiated, and activated into M1 (interferon-gamma [IFN-Ɣ] and lipopolysaccharide [LPS]) and M2a (interleukin-4 [IL-4] and IL-13) phenotypes (M1 and M2a hMDMs, respectively), as previously described [5,21-24]. Briefly, peripheral blood mononuclear cells (PBMCs) were separated using a Histopaque-Ficoll density centrifuge according to the manufacturer’s recommendations (Sigma-Aldrich, Munich, Germany). PBMCs (3 × 10⁷ cells/cm²) were suspended in RPMI 1640 medium (Biologic Industries Israel Beit Haemek Ltd, Kibbutz Beit Haemek, Israel), and seeded into six-well plates coated with the amino acid poly-D-lysine, which facilitates the adherence of monocytes. Two hours after incubation in a 37 °C and 5% CO₂ cell culture incubator, the cells were washed with PBS (1X; 120 mM NaCl, 20 mM KCl, 10 mM NaPO₄, 5 mM KPO₄, pH 7.4), and monocytes were cultured for 7 days in RPMI 1640 supplemented with 10% fetal calf serum (FCS), 1% nonessential amino acid, 2 mmol/l L-glutamine, 1 mM sodium pyruvate, 100 units/ml penicillin and 100 µg/ml streptomycin, and 50 ng/ml macrophage colony-stimulating factor (M-CSF, PeproTech, Rocky Hill, NJ). M-CSF was added to the growth medium to induce maturation of the monocytes to hMDMs. Polarization of hMDMs to the M1 (IFN-Ɣ and LPS) phenotype was obtained with the addition of 20 ng/ml IFN-Ɣ (PeproTech) and 100 ng/ml LPS (Sigma-Aldrich) at day 6. To obtain the M2a (IL-4 and IL-13) phenotype, 50 ng/ml IL-13 (PeproTech) and 20 ng/ml IL-4 (PeproTech) were added at day 5 of culture. As M1 (IFN-Ɣ and LPS) macrophages require 24 h for polarization whereas M2a (IL-4 and IL-13) macrophages need 48 h [25], the hMDMs were polarized on different days, so that in vitro experiments would be performed on the same day. The hMDMs were collected with 0.025% trypsin (TRI Reagent; Sigma-Aldrich) which was deactivated with RPMI ‏+ FCS, following three washes with PBS to remove traces of supernatant, so that pure cells remained.

Supplement composition and administration

Four different formulas were tested based on our previous research [19]. The ingredients were diluted in olive oil. For the in vivo studies, five groups of mice were treated with the four different formulas and a control group of olive oil, with gavage, during the 7 days that preceded the light injury and 7 days following it. The formula compositions were as follows: G1: lutein + zeaxanthin (Katra, Karnataka, India; 75 mg/kg); G2: lutein + zeaxanthin (75 mg/kg) and zinc (1.44 mg/kg; Navkar, Maharashtra, India); G3: lutein + zeaxanthin (75 mg/kg), zinc (1.44 mg/kg), Lyc-O-Mato (100 mg/kg; standardized tomato extract containing lycopene [6%], as well as other tomato phytonutrients, like phytoene, phytofluene, tocopherols, and phytosterols; Lycored, Be’er Sheva, Israel), and carnosic acid (100 mg/kg; added as rosemary extract containing 20% CA; Lycored); G4: lutein + zeaxanthin (75 mg/kg), carnosic acid (100 mg/kg), and beta-carotene (100 mg/kg); and G5: olive oil as vehicle control. For the ex vivo retinal explant and choroid sprouting assays, supplements were added to the culture media at the time of induction of macrophage polarization. The concentrations of the different compounds in culture were chosen to mimic serum levels obtained following oral supplementation as follows: G1: lutein + zeaxanthin (1 µM; 0.2 µM); G2: lutein + zeaxanthin (1 µM; 0.2 µM) and zinc (10 µM); G3: lutein + zeaxanthin (1 µM; 0.2 µM), zinc (10 µM), Lyc-O-Mato (2 µM), and carnosic acid (2 µM); G4: lutein + zeaxanthin (1 µM; 0.2 µM), carnosic acid (2 µM), and beta-carotene (2 µM), and G5: vehicle control.

Choroid sprouting assay

An ex vivo angiogenesis assay was performed to evaluate the effect of the different antioxidant treatments on the supernatant of M1 (IFN-Ɣ and LPS)/M2a (IL-4 and IL-13) hMDMs as previously described [26]. Briefly, the supernatant of the two types of polarized hMDMs was collected and kept in −20 °C until used. C57bl/6 4- to 6-week-old mice (n=7) that were treated in accordance with the guidelines of the Association for Research in Vision and Ophthalmology (ARVO) were used. Experiments were conducted with the approval of the institutional animal care ethics committee. Five minutes after ketamine was injected, the animals were checked for responses, and euthanized with cervical dislocation. The eyes were immediately enucleated and kept in ice-cold CO₂-independent medium (cat. 18045–054; Gibco, Paisley, Scotland) containing 100 units/ml penicillin-streptomycin and 1% glutamine before dissection. A choroid-sclera complex from the mice was gently dissected along with the RPE. The complex was cut into 1-mm pieces. Fragments were embedded in 30 µl of growth factor-reduced Matrigel (BD Biosciences, Cat. 354,230) in 24-well plates. The thickness of the Matrigel was approximately 0.4 mm. The plates were then incubated for 10 min in 37 °C, in a 5% CO₂ cell culture incubator without medium to solidify the Matrigel. A mix of 250 µl of medium containing endothelial cell growth medium (ECGM; C-22010, PromoCell, Heidelberg, Germany), 2.5% supplement mix (C-9215, PromoCell), 5% FCS, 100 units/ml penicillin-streptomycin and 1% glutamine, and 250 µl of the hMDM culture supernatant was added to each well. Medium for each well was changed every 3 days. and the cultures were fixed with 4% paraformaldehyde after 8 days. The cultures were viewed with an inverted phase-contrast CKX41 Olympus microscope, and images were photographed with an Olympus DP70 digital camera (Olympus, Tokyo, Japan). ImageJ software was used to quantify the sprouting area. The scale was set to convert pixels to square millimeters. Each picture was converted to 8-bit type to obtain a binary image. The sprouting area was then selected, and measured after the choroid tissue was excluded.

Macrophage coculture with mouse retinal explants

C57bl/6 mouse retinas (n=7) were dissected and incubated with polarized hMDMs (n=7) which were exposed to the different treatments as previously described [12]. Briefly, the different groups of treated hMDMs were harvested, and seeded for a minimum of 2 h on polycarbonate filter into serum-free Dulbecco's Modified Eagle Medium (DMEM; Biologic Industries, Kibbutz Beit Haemek, Israel) medium supplemented with glutamine and penicillin-streptomycin. In parallel, the C57bl/6 mice were anesthetized and then euthanized with cervical dislocation. The eyes were enucleated, and kept in cold serum-free DMEM medium supplemented with glutamine and penicillin-streptomycin. The retinas were gently detached from the choroid tissue, and they were immediately placed on the polycarbonate filter so that hMDMs were in contact with the photoreceptors layer. After 18 h of incubation, the mouse retinas were fixed with methanol, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) labeling was performed according to the manufacturer’s instructions (In Situ Cell Death Detection Kit, TMR red, La Roche, Basel, Switzerland). A Zeiss LSM710 confocal laser scanning system (Carl Zeiss MicroImaging GmbH, Jena, Germany) was used to visualize the stained cells, and ImageJ was used to perform automatic quantification of the cells.

Model of photic retinal injury

Albino (BALB/c) mice underwent genotyping to validate homozygosity to the wild-type crb1, gnat2, and rpe65 genes [27-29]. The process of photic injury was established according to the standard protocol [30], and optimized to validate an approximately 50% reduction in thickness of the outer nuclear layer. Briefly, after 1 h of dark adaptation, 6-week-old albino mice (n=8) that were raised under a normal 12 h:12 h light-dark cycle were exposed to 8,000 lux of white light of fluorescence tubes for a period of 3 h. Light damage was performed according to the appropriate circadian rhythm time period: dilation with Cyclogyl (one drop per eye, Farmaceutica, S.A, Madrid, Spain) and 5% phenylephrine (Fischer Pharmaceutical Labs, Bnei Brak, Israel) at 9:30 AM under red light and the light-level adjustment at 9:45 AM. Light damage was performed between 10 AM and 1 PM. During the light damage, the mice were placed into a cage (maximum two mice per cage) lined with aluminum foil, and temperature under 30 °C was maintained. Immediately after light damage, the mice were returned to the normal 12 h:12 h light-dark cycle.

Electroretinography recording

Full-field electroretinography (ERG) was performed 7 days after exposure to light. ERGs were recorded for the dark-adapted mice by placing an electrode in contact with the cornea and a reference electrode through the tongue. A grounding electrode was placed intramuscularly in the hip area. The eyes were anesthetized with Localin (Fischer Pharmaceutical Labs), and dilated with tropicamide (Fischer Pharmaceutical Labs) and phenylephrine (Fischer Pharmaceutical Labs), and the corneas were kept moist with the application of carboxymethylcellulose (Fischer Pharmaceutical Labs) as needed. All procedures were performed in dim red light or darkness, and the mice were kept warm during the entire procedure. Each mouse was positioned facing the center of a Ganzfeld bowl, ensuring equal, simultaneous illumination of both eyes. ERGs were recorded inside a Faraday cage using the Espion computerized system (Diagnosys Llc, Littleton, MA). Dark-adapted ERG responses to a series of white flashes of increasing intensities (from 0.000006 to 9.6 cd × sec/m²) were recorded with inter-stimulus intervals rising from 10 s for the lowest-intensity flashes to 90 s for the highest-intensity flashes. Light adaptation was accomplished with a background illumination of 30 cd/m². For analysis, the b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave, and the a-wave was measured as the difference in amplitude between the recording at 5 ms and the trough of the negative deflection.

Histological evaluation of oxidative stress and photoreceptors

Seven days after light damage, the mice were euthanized as described above, and the eyes were enucleated and cryosectioned at 10 µm. Anti-4 hydroxynonenal (4-HNE) antibody (10 µg/ml; Abcam (ab46545), Cambridge, UK) was used to detect oxidative stress, and anti-cleaved caspase-3 antibody (1 µg/ml, Cell Signaling Technology (#9661), Danvers, MA) was used to detect apoptosis. The immunohistochemical labeling of tissues was performed as previously described [11]. Briefly, eyes were fixed with 4% paraformaldehyde, and were placed in 30% sucrose overnight at 4 °C. Then the eyes were placed in optimal cutting temperature (OCT) solution (Scigen Scientific, Carson, CA), and placed in blocking solution (10% NGS, 0.01% Triton X in PBS) for 1 h at room temperature, followed by the addition of primary antibody overnight at 4 °C. Secondary antibody (goat anti-rabbit immunoglobulin [IgG] H&L; Alexa Fluor-488, Abcam, Cambridge, UK) was added for 1 h at room temperature. Counterstain with 4′,6-diamidino-2-phenylindole (DAPI) was applied to all sections. Then, the antibody-labeled tissue sections were visualized under a fluorescent microscope using a 488 nm laser for the 4-HNE or caspase-3 antibody-labeled tissues, and a 405 nm laser for DAPI. For the evaluation of the outer nuclear layer (ONL) thickness, the sections were stained with DAPI, and the number of photoreceptor nuclei was counted at different distances from the optic nerve head (ONH).

Statistical analysis

Appropriate statistical tests were applied according to the results of a normalcy test, and the sample distribution and parameters. Biostatistical package InStat (GraphPad Software, San Diego, CA) was used for data analysis. Values over two standard deviations from the average were excluded from statistical analysis, and the mean ± SEM is indicated. Statistical analysis was performed with one-way ANOVA, followed by the Dunnett post-test for comparison among means depending on the experimental design. For the choroidal sprouting assay and the retinal explant assay, a repeated-measure mixed-effect model (in which the random effect was the mouse) was used. An unpaired t test was used when only two groups were compared.

Modulation of the angiogenic effect of macrophages following treatment

M1 and M2a hMDMs from patients with AMD were cultured, and treated with the antioxidant supplements. The culture media was then added to the CSA cultures for 8 days, and the area of the sprouting vessel was calculated. Only treatment of M2a hMDMs with the G3 supplement was associated with a reduced sprouting area (−1.52-fold versus untreated control M2a medium, p=0.05, repeated-measure mixed-effect model; Figure 1B). All other treatments had no effect on the sprouting area (p>0.05, the repeated-measure mixed-effect model; Figure 1B). The angiogenic capacity of the M1 hMDMs was not affected by the different treatments (p>0.05, the repeated-measure mixed-effect model; Figure 1A). This subtype of hMDMs was found to have a stronger angiogenic effect in our previous experiments [11]. Because the antioxidants did not show an effect on this macrophage subtype, we did not conduct in vivo experiments to assess the supplement effect on neovascularization.

Modulation of the neurotoxic effect of macrophages in retinal explants

The cultured hMDMs were treated with the G1 through G4 supplements followed by coculture of the hMDMs with the mouse retinal explant. The apoptosis rate was evaluated with TUNEL labeling after 18 h of coculture. No change in the neurotoxic effect of M1 hMDMs was observed in association with the different treatments compared to the control group (p>0.05, Figure 2A). However, treatment with G1 was associated with an increased level of photoreceptor death compared to treatment with G2, G3, and G4 (2.13-fold, p=0.023, 3.96-fold, p=0.002 and 2.92-fold, p=0.02, respectively, the repeated-measure mixed-effect model, Figure 2A). Regarding the M2a hMDMs, we found that treatment with G2 and G4 was associated with an increased neurotoxic effect compared to the control group (1.37-fold, p=0.047 and 1.48-fold, p=0.004, respectively the repeated-measure mixed-effect model), as well as compared to the G3 treatment group (1.46-fold, p=0.01and 1.58-fold, p=0.001, respectively, the repeated-measure mixed-effect model; Figure 2B).

Effect of antioxidant treatments on photoreceptor survival in the mice model for photic retinal injury

Because the addition of antioxidant treatments modulated the deleterious effect of macrophages on photoreceptors in an in vitro model, we continued our investigation of their impact in an in vivo model which recapitulated some of the features of the atrophic stage of AMD, and where the involvement of the two subtypes of macrophages (M1 and M2a) has already been highlighted [13,31-33]. A mouse model of photic retinal injury was generated to assess the effect of antioxidant treatments on photoreceptor degeneration. Experimental conditions were optimized to obtain the loss of approximately 50% of the photoreceptor and marked retinal oxidative stress 7 days after the light exposure. Photoreceptor loss was evaluated with measurement of the ONL thickness. In addition, recruitment of myeloid cells was observed in the choroid of the light-damaged mice (Figure 3). Mice were treated with the G1 through G4 supplements with gavage during a period of 7 days before light injury and until 7 days following it. Results showed that administration of three of the four supplements (G2, G3, and G4) was associated with markedly reduced oxidative retinal injury according to 4-HNE labeling, compared with that of untreated mice (−2.32-fold, −2.17-fold, and −2.18-fold, respectively; p<0.05 for each of the comparisons, one-way ANOVA, Figure 4A).

We then evaluated whether the reduction of the oxidative injury level in the different groups was sufficient to allow for photoreceptor rescue in this model, using ERG, histology, and immunostaining for activated caspase 3. The immunostaining showed no difference in the number of photoreceptors positive for activated caspase 3 across the G1 through G4 treatment groups and the vehicle group (p>0.05, unpaired t test; Figure 4B). In accordance with those results, histological analysis showed similar ONL thickness in the G1 through G4 treatment groups compared with the vehicle group (Figure 4C), although the G2 and G4 treatment groups showed a trend toward increased damage, reinforcing the results obtained in the retinal explant assay (Figure 4D). Finally, there was an approximatively 40% decrease in retinal function per ERG after light injury, in the vehicle group (−1.78-fold versus control values, p<0.05, one-way ANOVA), and in the G1 (−1.85-fold, p<0.01, one-way ANOVA), G2 (−1, 98-fold, p<0.01, one-way ANOVA), G3 (−1.74-fold, p<0.05, one-way ANOVA), and G4 (−2.08-fold, p<0.01, one-way ANOVA) treatment groups (Figure 4F). However, there was no difference between the ERG values of the mice treated with the G1 through G4 supplements and the vehicle group (Figure 4E,F).