Progesterone Attenuates Microglial-Driven Retinal Degeneration and Stimulates Protective Fractalkine-CX3CR1 Signaling

Retinitis pigmentosa (RP) is a degenerative disease leading to photoreceptor cell loss. Mouse models of RP, such as the rd10 mouse (B6.CXBl-Pde6brd10/J), have enhanced our understanding of the disease, allowing for development of potential therapeutics. In 2011, our group first demonstrated that the synthetic progesterone analogue ‘Norgestrel’ is neuroprotective in two mouse models of retinal degeneration, including the rd10 mouse. We have since elucidated several mechanisms by which Norgestrel protects stressed photoreceptors, such as upregulating growth factors. This study consequently aimed to further characterize Norgestrel’s neuroprotective effects. Specifically, we sought to investigate the role that microglia might play; for microglial-derived inflammation has been shown to potentiate neurodegeneration. Dams of post-natal day (P) 10 rd10 pups were given a Norgestrel-supplemented diet (80mg/kg). Upon weaning, pups remained on Norgestrel. Tissue was harvested from P15-P50 rd10 mice on control or Norgestrel-supplemented diet. Norgestrel-diet administration provided significant retinal protection out to P40 in rd10 mice. Alterations in microglial activity coincided with significant protection, implicating microglial changes in Norgestrel-induced neuroprotection. Utilizing primary cultures of retinal microglia and 661W photoreceptor-like cells, we show that rd10 microglia drive neuronal cell death. We reveal a novel role of Norgestrel, acting directly on microglia to reduce pro-inflammatory activation and prevent neuronal cell death. Norgestrel effectively suppresses cytokine, chemokine and danger-associated molecular pattern molecule (DAMP) expression in the rd10 retina. Remarkably, Norgestrel upregulates fractalkine-CX3CR1 signaling 1 000-fold at the RNA level, in the rd10 mouse. Fractalkine-CX3CR1 signaling has been shown to protect neurons by regulating retinal microglial activation and migration. Ultimately, these results present Norgestrel as a promising treatment for RP, with dual actions as a neuroprotective and anti-inflammatory agent in the retina.


Introduction
the Biological Services Unit, University College Cork and were humanely euthanized by cervical dislocation.

Norgestrel-supplemented diet
Norgestrel-supplemented and control diets were manufactured by Testdiet (Middlesex, UK). Norgestrel was added to the chow at a concentration of 0.05% (500ppm). Dams of rd10 pups were given a Norgestrel-supplemented diet (LabDiet 5053, custom diet containing D (-) Norgestrel (N2260) (Sigma Aldrich, Missouri, US) when the pups were P10 to allow pups to receive Norgestrel in the milk. This equates to a daily intake of approximately 80mg/kg, assuming a 30g mouse consumes around 5g of food/day. Specialized diet soaked in water was also added to the cage in order for pups to receive Norgestrel directly, as soon as they began to eat. These pups were also given Norgestrel-supplemented diet post-weaning. With control litters, diet changes were replicated with a control diet (LabDiet 5053 control diet). Dams were given regular chow once pups were weaned.

Immunohistochemistry on retinal sections
Eyes were enucleated and fixed at room temperature in 4% paraformaldehyde (PFA) for 1.5 h. Following washes, eyes were cryo-protected in 15% sucrose in 1xPBS for 1 h, 20% sucrose for 1.5 h and 30% sucrose overnight, all at 4°C. Eyes were submerged and frozen in cryochrome (Thermo Scientific, Waltham, US) and sectioned on a cryostat (Leica, Wetzlar, Germany). Sections (7μm) were collected on Superfrost glass slides (Fisher Scientific, Waltham, US) and stored at -80°C. Sections were blocked and permeabilized with 0.1% Triton X and 5% donkey serum in 1xPBS for 30 min and incubated with primary antibody diluted in 5% donkey serum overnight at 4°C. Table 1 lists the details of all primary antibodies used. Following washes, sections were incubated with secondary antibody (Alexa Fluor donkey anti-mouse/rabbit/goat with either a 488 or 594 fluorescent probe; Molecular Probes &/or FITC-PNA (Vector Labs, Burlingame, US.)) with the nuclear stain, Hoechst (Thermo Scientific) for 1 h at room temperature. Eliminating the primary antibody in solution served as a negative control. Sections were mounted using Mowiol (Sigma) with anti-fade agent Dabco (Sigma). Images were taken from central and peripheral regions of the retina. Regions identified as central and peripheral are shown in S1 Fig.

Immunohistochemistry on cells
Microglia and 661W cells cultured on polylysine-coated coverslips in 24 well plates were fixed with 4% PFA for 10 min at room temperature. Cells were blocked and permeabilized in 0.1% Triton X and 5% donkey serum in 1xPBS for 10 min and incubated with primary antibody diluted in 5% donkey serum overnight at 4°C. Table 1 lists the details of all primary antibodies used. Following washes, coverslips were incubated with secondary antibody (Alexa Fluor donkey anti-mouse/rabbit/goat with either a 488 or 594 fluorescent probe) with the nuclear stain, Hoechst (Thermo Scientific) for 1 h at room temperature. Eliminating the primary antibody in solution served as a negative control. Coverslips were mounted on to glass slides using Mowiol.

Microscopy
Retinal sections and cell preparations were viewed using a Leica DM LB2 microscope with Nikon Digital Sight DS-U2 camera, using 10x, 40x and 100x objectives. Images were taken using the software NIS-Elements version 3.0, Nikon, Japan. Immunofluorescence on retinal sections was performed on at least three mice of each group, at each time-point. Immunofluorescence on cell preparations was also performed in triplicate. Confocal micrographs were taken using an Olympus Fluoview FV1000 laser scanning confocal microscope, using 20x and 60x objectives. Images were taken using the software Olympus Fluoview Ver 4.1a and are represented as single slices in the XY plane or maximum intensity projections from acquisition of z-stacks. Identical microscope settings were used when visualizing specific markers across all timepoints and treatments.
Terminal dUTP nick-end labeling (TUNEL) of fragmented DNA DNA strand breaks in retinal cell nuclei and 661W cells were detected by terminal dUTP nickend labeling (TUNEL) on fixed tissue and cells. Frozen retinal sections or coverslips with cells were permeabilzed with 0.1% Triton X for 2 min followed by incubation with terminal deoxynucleotidyl transferase (Promega, Wisconsin, US) and fluorescein-12-dUTP (Roche, Risch-Rotkreuz, Switzerland) according to manufacturer's instructions. Nuclei were counterstained with Hoechst 33342 (1μg/mL) (Sigma). Sections and cells were incubated at 37°C for 1 h in a humidified chamber and following several washes in 1xPBS, were mounted in Mowiol. Sections were viewed under a fluorescence microscope (Leica DM LB2). In co-culture assays, TUNELpositive 661W cells could easily be distinguished from microglia based on nuclear size and phase contrast of cell preparations. For each treatment on mice, at least three animals were used and three fields (x40 magnification) per section of at least three different sections were evaluated. For each treatment on coverslips, 6 coverslips were used per treatment and 50-200 661W cells analyzed per coverslip for TUNEL positivity.

ERG recordings
Dark-adapted (12 h) mice were anaesthetized with an intraperitoneal injection of ketamine (100 mg/kg)-xylazine (4 mg/kg) and maintained on a thermal blanket at 38°C for the entire procedure. 1% tropicamide (Alcon Cusí, Barcelona, Spain) was applied topically on the eyes to dilate the pupils and viscotears (Novartis, Barcelona, Spain) was instilled on the corneas to prevent dehydration and optimize contact with the recording electrodes (DTL fiber with an X-Static silver-coated nylon conductive strand; Sauquoit Industries, Scranton, PA). The reference electrode was a 25-gauge platinum needle inserted under the scalp and a gold electrode, located in the mouth, acted as ground. Animal handling and preparation was done under dim red light and the mice were placed on a Faraday cage where all the recordings were performed in total darkness. Scotopic flash-induced ERG responses were recorded simultaneously from both eyes in response to light stimuli produced with a Ganzfeld illuminator. Light stimuli of 10 ms of duration were presented at 11 different increasing luminances, ranging from −5.2 to 0 log cd s/m2. Three to ten consecutive recordings were averaged for each light presentation. The interval between stimuli was 10 s for dim flashes (−5.2 to −1.4 log cd s/m2) and up to 20 s for the highest luminance's (−0.8 to 0 log cd s/m2). ERG signals were amplified and band-pass filtered (1-1000 Hz, without notch filtering) using a DAM50 data acquisition board (World Precision Instruments, Aston, UK). Stimuli presentation and data acquisition (4 kHz) were performed using a PowerLab system (AD Instruments, Oxfordshire, UK). Recordings were analyzed off-line. The amplitude of the a-wave was measured from the baseline and the amplitude of the b-wave was measured from the trough of the a-wave to the peak of the b-wave.

Culture of cell lines
Experiments were carried out using the mouse 661W cone photoreceptor-derived cell line, generously provided by Dr Muayyad Al-Ubaidi (Department of Cell Biology, University of Oklahoma, Health Sciences Centre, Oklahoma City, OK, USA). This cell line was previously validated by this group [11] through real-time quantitative polymerase chain reaction (rt-qPCR) analysis for cone specific opsins; blue cone opsin (Opn1sw) and red/green opsin (Opn1mw). Cells were cultured in Dulbecco's Modified Eagle's medium (DMEM) (Sigma) supplemented with 10% fetal calf serum (FCS) and 1% penicillin streptomycin (PS) and maintained at 37°C in a humidified 5% CO 2 atmosphere. To analyze the effects of Norgestrel (Sigma) on 661W cells, cells were seeded in to the appropriate culture vessel and allowed to attach overnight. Cells were then washed three times with warmed 1xPBS and complete or serum-free medium supplemented with 20 μM Norgestrel or DMSO vehicle (Sigma) added. After incubation for the indicated times, cells were detached using accutase (Sigma) and, together with their supernatant, centrifuged and washed with 1x PBS to leave a whole cell pellet.
Lysates were centrifuged at 10,000 rpm at 4°C for 30 min. Supernatant was stored at -80°C. Protein concentration was measured using a Bradford assay (Bio-Rad, Hercules, US). 4-15% gradient gels (Bio-Rad) were used for SDS-PAGE and proteins were then transferred to a nitrocellulose membrane. 30μg of protein was loaded per sample. Total protein levels are an accurate way of verifying equal loading [26] and so were analyzed using REVERT total protein stain (LiCor, Lincoln, US) as per manufacturer's instructions and imaged on a LiCor scanner in the 700 channel. Membranes were blocked using Odyssey blocking buffer (LiCor) for 30 min and probed with fractalkine primary antibody (Abcam, Cambridge, UK) in blocking buffer and 0.1% Tween 20 for 4 nights at 4°C. Membranes were washed using TBS-T and probed with rabbit 800 secondary antibody (LiCor). Membranes were washed using TBS and imaged on a LiCor scanner.

Nitrite parameter assay
Cell culture media from rd10 primary microglia treated with 20 μM Norgestrel or DMSO over 24 hr was collected. Media was centrifuged at 1000 rpm for 5 min to remove particulates. Supernatant was stored at -20°C. Samples were analyzed for endogenous nitrite (R&D systems, Minneapolis, US) according to manufacturer's instructions.

Culture of primary microglial cells
A protocol for isolating and culturing retinal microglia was adapted from a previously published protocol [27]. Briefly, retinas were dissected from the eyes of P16 rd10 mice, ensuring minimum contamination with vitreous body and retinal pigment epithelium. Four retinas were pooled, cut into small pieces and incubated for 40 min at 37°C in 1 ml 1xPBS with 1 mg/ ml collagenase type I (Sigma), 0.3 mg/ml DNase I (Roche, Basel, Switzerland), and 0.2 mg/ml hyaluronidase (Sigma). The cell suspension was filtered through a 70-μm cell strainer (Becton Dickinson, Franklin Lakes, US). Cells were washed twice with 10 ml DMEM/10% FCS/1% PS and resuspended in 15ml DMEM/10% FCS/1% PS. To isolate the mononuclear cells, the suspension was gently added to 15ml Ficoll paque premium reagent (GE Life Sciences, Buckinghamshire, UK) and centrifuged for 20 min at 2000 rpm without the brake in a Beckmann GS-6R centrifuge. The interphase was removed carefully and washed twice with 10 ml DMEM/ 10% FCS/1% PS. The cells were cultured for 3 days, on either polylysine-coated coverslips in 24 well plates or T25 cm 2 flasks containing DMEM/10% FCS/1% PS. Culture purity was confirmed by the absence of retinal astrocytes, Müller glia, rods, retinal ganglion, amacrine, horizontal and bipolar cells, as assessed by immunofluorescence (S2 Fig). Although rare, some debris from these cell types was found in culture (S2A Fig, inserts), however all Hoechst-positive cells stained positively for ionized calcium-binding adaptor molecule 1 (Iba1) and so were confirmed to be microglia (S2B Fig). P16 was chosen as the time-point to harvest primary cells. Microglia respond and migrate to the mutated photoreceptors becoming activated by P15 in the rd10 retina compared to C57 control [9]. This time-point therefore represents a time when microglia have responded to the mutated photoreceptors but photoreceptor cell death has not yet taken place. This eliminates the response of microglia to accumulating photoreceptor cell debris.

Culture of 661W cells in microglial-conditioned media
A total of six retinas from three P16 rd10 mice were used to make the primary culture for the conditioned media experiment. Primary microglia were cultured in T25cm 2 flasks for 2 days before treating with Norgestrel or vehicle (DMSO). Cells were exposed to 20 μM Norgestrel in DMSO or vehicle and left for 24 h. Cells were washed with 1xPBS and cultured in fresh DMEM/10% FCS/1% PS for a further 24 h. This conditioned media was collected and added to 661W cells. 661W cells were seeded overnight in 96 well plates (Sarstedt, Numbrecht, Germany) to allow for cell adherence. All cells were then washed three times in 1xPBS and media replaced with either basal microglial media, or conditioned media from treated primary microglial cells (100μl/well). Cells were incubated for 20 h before addition of MTS (3- for a further 4 h.

Co-culture of rd10 microglia with 661W cells
A total of sixteen retinas from eight P16 rd10 mice were used to make the primary culture for the co-culture experiment with 661W cells following treatment with Norgestrel. Primary microglia were cultured on polylysine-coated coverslips in 24 well plates for 1 day before treating with Norgestrel or vehicle (DMSO). Primary cells were exposed to 20 μM Norgestrel or vehicle and left for 24 h. Cells were washed with 1xPBS prior to the addition of healthy or serum-starved 661W cells that had been treated with Norgestrel or vehicle in the preceeding 3 h. 661W cells suspended in fresh DMEM/10% FCS/1% PS were added to the microglia at a density of 25 x 10 3 per well. Co-cultures of microglia and 661W cells were left for 24 h before fixation in 4% PFA. A total of eighteen retinas from nine P16 rd10 mice were used to make the primary culture for the co-culture experiment with 661W cells following treatment with recombinant fractalkine. Primary microglia were cultured on polylysine-coated coverslips in 24 well plates for 1 day before treating with recombinant mouse full-length fractalkine (R&D systems cat. no. 472-FF-025) or vehicle (0.1% BSA in 1xPBS). Primary cells were exposed to 1.2μg/ml fractalkine or vehicle and left for 24 h. Cells were washed with 1xPBS prior to the addition of serum-starved 661W cells. 661W cells suspended in fresh DMEM/10% FCS/1% PS were added to the microglia at a density of 25 x 10 3 per well. Co-cultures of microglia and 661W cells were left for 24 h before fixation in 4% PFA.

Quantification from immunohistochemically-labeled sections and cells
Quantification of outer nuclear layer (ONL) thickness and outer segment length in retinal sections was carried out using ImageJ software. Average ONL thickness was measured by taking measurements from at least 20 sections per mouse. Per section, three distinct measurements were taken and averaged. Average outer segment length was measured by taking measurements from at least four sections per mouse. Per section, three distinct measurements of rod or cone outer segment length were taken and averaged. Three mice were analyzed per group. Fluorescence intensity measurements of CD68, iNOS and MRC1 in primary microglia were performed using ImageJ software, by measuring points of intensity within single cells.

Assessment of cell viability in 661W cells by the MTS assay
Cells seeded overnight (2 x 10 3 per well 96 well plate) were washed three times in 1xPBS before treatment (conditioned microglial media). After 20 h of treatment incubation, 20μl of MTS solution (Promega) was added to each well and incubated for a further 4 h at 37°C. Viable cells in the presence of phenazine methosulfate (PMS) will reduce the MTS solution to form formazan. Detection and quantification of the formazan crystals was carried out with a microplate reader (Molecular Device Corporation, Sunnyvale, US) at 490nm. 490nm readings taken from non-template wells (media and MTS, without cells) were deducted from actual cellular readings. A further reading at 650nm was taken from all wells and deducted from the 490nm readings to account for any cellular debris. The quantity of formazan product as measured by the amount of 490nm absorbance is directly proportional to the number of living cells in culture. Therefore the absorbance of the formazan formed in 'control cell wells' was taken as 100% viability.
Total RNA Isolation and quantitative real-time polymerase chain reaction (rt-qPCR) Total RNA was isolated from whole 661W cells and RNA-later (Qiagen, Hilden, Germany) protected tissue using RNeasy Mini Kit (Qiagen), or from primary microglia using RNeasy Micro Kit (Qiagen) following manufacturer's protocol. Genomic DNA was eliminated in 661W cells/tissue samples using RNase free DNase (Qiagen) and in primary microglia using gDNA eliminator columns (Qiagen) and cDNA was subsequently synthesized using QuantiTect Reverse Transcription Kit (Qiagen). Rt-qPCR was performed using QuantiTect SYBR Green PCR Kit (Qiagen) and 10ng cDNA/well of a 384 well plate (Starstedt AG & Co.) for 661W cells/tissue and 5ng cDNA/well for primary microglia. Plates were run using the Applied Biosystems 7900HT Fast Real-Time PCR System (Life Technologies Ltd., Carlsbad, US) and each set of reactions included both a non-reverse transcription control and a no template sample negative control (data not shown). The protocol consisted of a cycling profile of 30s at 95°C, 60s at 60°C, and 30s at 72°C for 40 cycles. Qiagen QuantiTect Primer assays were used for all genes ( Table 2). Melt curve analysis confirmed single PCR product. Relative changes in gene expression were quantified using the comparative Ct (ΔCt, ΔΔCt) method as described by Livak & Schmittgen [28,29]. The Ct value of the gene of interest was normalized to an average of three endogenous housekeeping genes (Actb, Gapdh and Hprt). This was compared to the normalized control sample-an aged matched control diet rd10 value, representative of all control samples (N = 3) or an equivalent DMSO control (661W cell treatments). Alteration in mRNA expression of genes was defined as fold difference in the expression level in cells/tissue after treatment, relative to that of the control. This is the standard method for presenting rt-qPCR data [28,29].

Statistical analysis
Values in all graphs represent the mean ± standard error of the mean (SEM) and are representative of at least three individual experiments repeated in triplicate. Data were statistically analyzed using Student t-test and two-way ANOVA (Graph Pad, Prism 6) with values of p < 0.05 being considered statistically significant. Values that were more than ± 3 standard deviations

Results
Norgestrel-supplemented diet from P10 results in significant protection of photoreceptors in the rd10 mouse Norgestrel has previously been shown to provide significant neuroprotection to the rd10 retina when administered by intra-peritoneal injection from P18 [10]. We sought to evaluate the protection offered by Norgestrel when using a less invasive method of diet-supplementation. We also recognized a need to administer Norgestrel at an earlier age based on recent findings that the rd10 retina displays significant abnormalities prior to photoreceptor cell loss [9]. Dams of newborn rd10s were given a Norgestrel-supplemented diet when their pups reached post-natal day 10 (P10), and all rd10s continued to receive a Norgestrel-supplemented diet post-weaning. Control rd10 litters received an identical diet without Norgestrel. Norgestrel-supplemented rd10's demonstrated lower levels of TUNEL positivity in the central outer nuclear layer (ONL) at P15 and P20 compared to control (Fig 1A), coinciding with significant preservation of photoreceptors. As expected, control rd10 mice revealed a substantial loss of photoreceptors between P15 and P25 (Fig 1 & S1 Fig) [9]. At P20, ONL thickness in the Norgestrel-fed rd10 mice was comparable to age-matched C57 retina (S1 Fig). Measurement of ONL thickness confirmed significant Norgestrel-induced ONL protection out to P40 in both the central and peripheral retina (Fig 1B). Norgestrel's neuroprotective effects were lost by P50.
Norgestrel-supplemented diet results in significant functional preservation in the rd10 retina In order to determine functionality of the rd10 retina following Norgestrel-diet supplementation, scotopic electroretinogram (ERG) recordings were taken in mice at P20 and P31, where significant preservation of the ONL was observed (Fig 1). ERG results confirmed functional preservation of the rd10 retina (Fig 2A). As previously described [5,6], the amplitudes of the aand b-waves in rd10 mice at P20 were diminished compared to age-matched wild type animals. This decline was significantly attenuated by Norgestrel diet supplementation (Fig 2B). Remarkably, the protective effect of Norgestrel on a-and b-wave amplitudes persisted at P31, where the values corresponding to the b-wave amplitudes were even higher than those of P20 mice fed with control diet.
Norgestrel-supplemented diet results in significant preservation of photoreceptor outer segments in the rd10 mouse In order to assess preservation of photoreceptor outer segments with Norgestrel diet, markers for rod (rhodopsin; red) and cone (PNA; green) outer segments were used on retinal sections. Outer segments were longer and denser in the central retina of Norgestrel treated mice compared to control (Fig 3A and 3B). This morphological change was observed from P15 for rods and from P20 for cones in the central retina (Fig 3A and 3B). Quantification of rod and cone outer segment length confirmed significant preservation out to P25 (Fig 3C).

Microglial response in the rd10 retina is significantly altered with Norgestrel treatment
Previous studies have highlighted a role for microglia in potentiating photoreceptor cell death in the rd10 mouse by driving inflammatory processes [18,19]. We therefore sought to study microglial dynamics in rd10 mice following administration of Norgestrel. Amoeboid microglia (ionized calcium-binding adapter molecule 1; Iba1) were observed surrounding the ONL (Fig  4A), during the peak of cell death (P15 and P20) in the control retina as previously reported [9,18]. Some stained positively for cluster of differentiation 68 (CD68), a marker used to identify phagocytic, activated microglia [30]. Conversely, microglia were ramified in morphology in Norgestrel-fed mice (P15, P20) and showed less CD68 immunoreactivity than control ( Fig  4A). Norgestrel-diet significantly reduced microglial number at P15 and P20 ( Fig 4B) and CD68-positivity at P20 (Fig 4C). The apparent correlation between significant photoreceptor protection and less microglial activation in Norgestrel-fed animals (Fig 4A-4C  . mRNA analysis on whole retina confirmed a Norgestrel-driven, significant decrease in iNOS expression (M1) at P20 and P25, and an increase in MRC1 (M2) at P20 and arginase (M2) expression at P25 and P30 (Fig 4D). High mobility group box 1 protein (HMGB1) and interleukin 1α (IL-1α) belong to the family of danger-associated molecular pattern molecules (DAMPs), expressed and released by damaged cells to trigger inflammation [33-36]. mRNA analysis of DAMP expression showed that Norgestrel induced a significant decrease in HMBG1 and IL-1α at P20 and P25 (Fig 4E), coinciding with increased photoreceptor survival (Fig 1), less M1 and increased M2 microglial activation (Fig 4A-4D).

Norgestrel significantly reduces rd10 microglial-driven, cell contactdependent degeneration of 661W cells
In order to evaluate the contribution of microglia to the course of photoreceptor cell loss in the rd10 retina in more detail, we utilized primary cell cultures of P16 rd10 retinal microglia. Culture purity was confirmed (S2 Fig). Due to the difficulty in isolating pure photoreceptor primary cultures and in order to assess the interaction between microglia and photoreceptors, we utilized the cone photoreceptor-like 661W cell line. Our group has previously used 661Ws as an in vitro model to study how photoreceptors respond to stressful stimuli and to understand how Norgestrel works in providing neuroprotection [11][12][13].
To determine the effects of released factors from rd10 microglia on 661W viability, healthy 661Ws were cultured for 24 h with conditioned medium (CM) from rd10 microglia pre-treated with 20 μM Norgestrel or equivalent DMSO control. 661W cell viability was then assessed using the MTS assay. Whilst CM from DMSO-treated microglia resulted in a~14% increase in 661W cell death, CM from Norgestrel-treated microglia resulted in only a~4% increase compared to untreated control ( Fig 5A). This suggests a novel role for Norgestrel in modulating microglial responses in the rd10 retina.
We hypothesized that direct cell-cell contact may regulate microglial responses [37, 38], for in the early stages of rd10 retinal degeneration, microglia associate closely with cells in the ONL [9,18]. Therefore, a co-culture assay was designed using a combination of rd10 microglia, healthy 661Ws and serum-starved 661Ws. This enabled us to analyze the direct effects of Norgestrel on microglia and photoreceptors in a way that more closely resembled the in vivo state. Prior to co-culture, primary rd10 microglia and 661W cells were treated with 20 μM Norgestrel or equivalent DMSO control (24 h microglia, 3 h for 661Ws). 661W cells were then added to the microglia and incubated for a further 24 h. During this time, microglia appeared to migrate and settle in clusters close to 661W cells (Fig 5B). Subsequent cell viability of co-cultured 661W cells was assessed using the TUNEL assay.
The previous experiment was replicated using serum-starved 661W cells, a previously established method for inducing 661W cellular stress [11,13]. This allowed us to investigate the response of microglia to damaged photoreceptors, in an attempt to mimic photoreceptor stress present in the rd10 mouse retina. 661Ws were serum-starved for 3 h, resulting in~8% cell death as assessed by TUNEL (Fig 5E (ii)). Addition of rd10 primary microglial cells for 24 h increased 661W cell death to~50%, a 42% increase in cell death compared to serum-starved 661W cells cultured in isolation (Fig 5D (ii) and 5E (ii)).
Treatment of rd10 microglia with 20 μM Norgestrel prior to co-culture significantly reduced their ability to damage healthy and serum-starved 661W cells (Fig 5E (i) (ii)), indicating that Norgestrel can work directly on microglia. Interestingly, pre-treatment of 661W cells with 20 μM Norgestrel also resulted in increased 661W protection (Fig 5E (i) (ii)), demonstrating Norgestrel's ability to prime undamaged and stressed photoreceptors against microglialderived harmful stimuli.

Norgestrel modulates pro-and anti-inflammatory states in rd10 microglia
Norgestrel can act directly on rd10 microglia and consequently prevent microglial-driven 661W cell death (Fig 5). Norgestrel mediates its neuroprotective effects through progesterone receptors [11]. mRNA analysis revealed similar expression patterns of progesterone receptor membrane complexes 1 and 2 (PGRMC1, PGRMC2), and membrane progesterone receptors α, β and γ (mPRα, β, γ) in rd10 microglia (Fig 6A). Immunofluorescence confirmed expression of these receptors (Fig 6B). Classical progesterone receptors A and B (PRA/B) were not detected in these cells (Fig 6A and 6B). Primary rd10 microglia were subsequently treated with 20 μM Norgestrel and analyzed by immunofluorescence for M1 and M2 state markers. Norgestrel induced significantly lower levels of CD68 and iNOS (M1) and higher levels of MRC1 (M2) in rd10 microglia compared to control (Fig 6C). Nitric oxide production was assessed indirectly by measuring endogenous nitrite levels in media from primary rd10 microglia treated with Norgestrel or vehicle. Norgestrel significantly lessened endogenous microglial nitrites (Fig 6E). Thus, we hypothesize that Norgestrel dampens pro-inflammatory activity and stimulates anti-inflammatory states in rd10 microglia, resulting in increased protection of photoreceptors.
Norgestrel upregulates neuroprotective fractalkine-CX3CR1 signaling in the rd10 retina . Therefore, we explored Norgestrel's ability to upregulate fractalkine and CX3CR1 gene expression, in the degenerating rd10 retina. Norgestrel treatment of stressed 661W cells significantly upregulated fractalkine mRNA over 24 h (Fig 8A). In the rd10 retina, Norgestrel significantly increased fractalkine levels at P20 by over 1000-fold that of the control ( Fig 8B); coinciding with significant protection of the ONL (Fig 1). Norgestrel also significantly upregulated CX3CR1 gene expression (Fig 8B). Norgestrelinduced fractalkine signaling was dampened at later time points (P25, P30) compared to P20 ( Fig  8B) correlating with photoreceptor loss in Norgestrel-treated mice (Fig 1). Western blotting revealed an increase in 100kDa fractalkine protein at P20 and P25 with Norgestrel ( Fig 8C; red boxes). This isoform is neuroprotective in the retina by dampening harmful microglial processes and consequently improving photoreceptor survival [57,62]. Immunofluorescence on retinal sections also showed an increase in fractalkine with Norgestrel ( Fig 8D). Fractalkine was found predominantly in the ONL and outer plexiform layer (OPL), and staining was increased in all (P15-P40) Norgestrel-fed mice compared to control (Fig 8D). Primary rd10 microglia treated with recombinant full-length fractalkine for 24 h and subsequently co-cultured with serumstarved 661W cells induced significantly less 661W cell death compared to vehicle (Fig 8E and  8F). This suggests that Norgestrel drives an upregulation of fractalkine in the retina, which will reduce microglial-driven neuronal cell death.

Discussion
In this study we demonstrate the protective properties of the synthetic progestin 'Norgestrel' , when administered via non-invasive diet supplementation. Previously shown to protect inflammatory and increased anti-inflammatory markers in rd10 microglia. Scale bar 10μm. (D) Quantification of fluorescence intensity of (D) (i) CD68 and iNOS and (D) (ii) MRC1 in primary rd10 microglial cultures following Norgestrel treatment over 24h. N = 8 mice, n = 4 technical replicates per group. (E) Nitrite concentration in the media of rd10 microglia in vitro treated with Norgestrel or vehicle over 24hr. Norgestrel significantly reduced nitrite release from rd10 microglia. N = 8 biological replicates, n = 2 technical replicates per sample. Results are presented as mean ± SEM (t-test, *p<0.05, ***p<0.005, ****p<0.001).
doi:10.1371/journal.pone.0165197.g006 through an upregulation of growth factors [12,13], this study highlights a novel role for Norgestrel, in acting directly upon microglial cells to dampen damaging inflammatory processes. This is an important finding as recent literature has shown that microglial cells may potentiate retinal disease [18,19,21,44,63,64]. We also provide further mechanistic insight, showing that Norgestrel drives up fractalkine expression in the rd10 retina, which likely acts to further dampen microglial activation.
Our group has previously shown that Norgestrel is neuroprotective in the rd10 retina when administered via intraperitoneal injection from P18 [10]. Since this study, we have demonstrated that significant morphological changes take place in the rd10 retina prior to photoreceptor cell loss [9]. In light of this finding, we sought to administer Norgestrel from an earlier time-point. We also wished to use a less invasive method and avoid the use of the vehicle DMSO, which has been shown to significantly affect cell metabolism in several ways including acting as an anti-inflammatory [65]. Using a supplemented-diet from P10, we have demonstrated increased protection with Norgestrel in the rd10 retina (Fig 1) compared to intra-peritoneal injection from P18 [10].
We hypothesized that rd10 retinal microglia drive targeted neuronal degeneration of viable cells in a cell autonomous manner. In order to substantiate this claim, and subsequently investigate the direct effects of Norgestrel on microglia, we utilized in vitro cultures of 661W cone photoreceptor-like cells and primary rd10 microglia both in isolation and in co-culture. By culturing 661Ws with Norgestrel-treated rd10 microglial conditioned media, we show that Norgestrel dampens damaging inflammatory factor secretion ( Fig 5A). Interestingly, direct microglia-photoreceptor contact induces significantly more cell death of healthy 661Ws than conditioned media alone (compare Fig 5A to 5E (i)). Extrapolating this to the in vivo situation, this suggests that microglia can cause higher levels of cell death when they have reached the ONL and are in direct contact with photoreceptors. These results also confirm our hypothesis, showing that isolated rd10 microglia drive targeted cell death of healthy photoreceptors, as primary C57 microglia did not induce 661W cell death. Inducing~30% and~42% cell death in healthy and serum-starved 661W cells respectively, 661W cells are more vulnerable to damaging microglial-contact when already stressed (compare Fig 5D (i) to 5D (ii)). Importantly, Norgestrel alleviates microglial-driven neuronal death of healthy and stressed 661Ws, by acting directly on both microglia and 661W cells. Of course, microglial-driven cell death is initiated due to effects from the mutation in pde6b. Microglia are therefore stimulated to adopt a damaging phenotype by signals from other cell types in the retina. However, once activated microglia will likely promote further degeneration of viable photoreceptors, exacerbating disease progression.
Norgestrel-diet administration prevented any photoreceptor loss between P15 and P20 in rd10 mice (Fig 1 & S1 Fig). We believe this protection is largely owed to increased levels of Norgestrel-induced fractalkine signaling. Norgestrel upregulated fractalkine expression by over 1000-times that of P20 control (Fig 8). Fractalkine signaling has been shown to prevent cell loss in a number of different disease models [56, 61, [67][68][69][70]. In mouse models of RP, CX3CR1 -/-rd10 mice experience a worsened disease progression [19] and intravitreal injection of recombinant full-length fractalkine improves photoreceptor survival in the rd10 mouse [62]. Indeed we show that rd10 microglia pre-treated with full-length fractalkine cause significantly less 661W cell death in co-culture (Fig 8E and 8F). These studies stand to corroborate the crucial importance of fractalkine-CX3CR1 signaling in retinal neuroprotection and the potential of Norgestrel as a therapeutic for RP.
Of particular interest, full-length fractalkine appears to play a neuroprotective role by dampening harmful microglial activity [57, 62], whereas soluble fractalkine may potentiate disease progression by increasing microglial migration [57]. In support of this, fractalkine cleavage increases prior to neuronal cell death in vitro [71] and loss of membrane-bound fractalkine precedes degeneration in the rd10 retina [50]. Western blots of P20 retinas revealed an increase in full-length (Fig 8C; 100kDa) and a decrease in soluble (Fig 8C; 85kDa) isoforms of fractalkine with Norgestrel.
Ultimately however, Norgestrel's neuroprotective effects are gradually lost from P25, though Norgestrel-driven photoreceptor survival does not lapse entirely until P50 (Fig 1). It is worth noting that the measurable neuroprotective effects observed with Norgestrel are in line with previous studies documenting other types of neuroprotective strategies in the rd10 retina [18,19,62]. Such studies similarly present therapeutic approaches resulting in delayed degeneration rather than complete prevention of photoreceptor loss. This is likely due to the fact that the primary insult remains, the mutation in pde6b, and will therefore persist as a source of cellular stress. Norgestrel-treated mice show a significant loss in fractalkine at the RNA and protein level at P25 compared to P20 (Fig 8B-8D). Taken together with the bulk of evidence supporting fractalkine's neuroprotective properties in the retina, we believe that loss of fractalkine signaling is detrimental to photoreceptors. We hypothesize that such a loss in fractalkine signaling at P25 contributes to microglial infiltration to the ONL (Fig 4A) and subsequent potentiation of cell death (Fig 5).
In conclusion, this study outlines novel and desirable properties of the synthetic progestin 'Norgestrel' , in acting directly on retinal microglia, and indirectly through photoreceptors, to dampen damaging microglial activation in vitro and in vivo. The data presented suggest a critical role for Norgestrel-induced fractalkine signaling in the dampening of microglial activation. Future studies will be aimed at further elucidating the role that fractalkine signaling plays in retinal degeneration and Norgestrel-associated neuroprotection. We believe direct Norgestrel-fed rd10 mice. Fractalkine staining is evident in the outer nuclear layer (ONL) of the retina in Norgestrel-supplemented mice. Hoechst (blue) staining reveals the cell layers in the retina. Scale bar 50μm. N = 3 mice per group, n = 3 sections per mouse. (E) 661W cells were cultured in the absence of serum (serum-starved 661W) over 3 h. 661W cells were then placed in culture with microglia that had been treated with (Microglia Fract.) or without (Microglia) 1.2μg/ml recombinant full-length fractalkine over 24h. Co-cultures were incubated for a further 24h. Apoptosis of 661W cells was detected by TUNEL of DNA strand breaks in cell nuclei. Co-cultures with fractalkine-treated microglia had less TUNEL-positive 661W cells than equivalent vehicle control. Scale bar 30μm. (F) Quantification of TUNEL-positive 661W cells cultured in the absence of serum and following co-incubation with primary rd10 microglial cells pre-treated with recombinant full-length fractalkine or vehicle. N = 9 mice for primary culture, n = 5 technical replicates per group. Results are presented as mean ± SEM (t-test, *p<0.05, ***p<0.005, ****p<0.001).
doi:10.1371/journal.pone.0165197.g008 photoreceptor neuroprotection should remain at the forefront of RP immunotherapy development: targeting mutated photoreceptors would alleviate subsequent photoreceptor-microglial cell crosstalk, minimizing consequential microglial-driven photoreceptor cell loss. Additional anti-inflammatory properties may then further prevent microglial responses shown to exacerbate disease progression.