Lack of Paraoxonase 1 Alters Phospholipid Composition, but Not Morphology and Function of the Mouse Retina

Abstract

Purpose.: Biochemical and genetic analyses established a contribution of lipid metabolism to AMD pathology. Paraoxonase 1 (PON1) is an antioxidative protein involved in high density lipoprotein (HDL) function and was found to be associated with AMD. Here, we used Pon1^−/− mice to study the influence of PON1 on retinal physiology and to reveal the potential impact of PON1 on AMD etiology.

Methods.: Laser capture microdissection served to isolate single retinal layers. Retinal function was assessed by ERG. Retinal and RPE morphology were monitored by fundus imaging, fluorescein angiography, light and transmission electron microscopy, and immunofluorescence microscopy. Levels of mRNA and composition of phospholipid species were determined by real-time PCR and LC-MS, respectively.

Results.: Adult (8 weeks old) Pon1^−/− mice displayed normal retinal function and morphology, but their retinas contained reduced amounts of lysophosphatidylcholines (LPCs) compared to controls. Aged (12 months old) Pon1^−/− animals did not show any morphologic or molecular signs of photoreceptor or RPE degeneration, or of accelerated aging. Photoreceptors of Pon1^−/− and control mice were similarly susceptible to light damage.

Conclusions.: Results indicated that PON1 is not essential for normal development, function, ageing, and the defense against light damage of the mouse retina. Reduced levels of LPCs in eyes of Pon1^−/− mice may reflect a decreased activity of phospholipase A2 or altered antioxidative activity in aged eyes.

Introduction

Age-related macular degeneration (AMD) is the leading cause of irreversible blindness and visual disability in the elderly population of industrialized countries. Various environmental risk factors, such as advanced age, cigarette smoking, diet, obesity, atherosclerosis, hypertension, inflammatory disease, as well as genetic predisposition have been implicated in this complex disease. ^1–3 The epidemiological risk factors point toward various molecular mechanisms that might be involved in the etiology of AMD. Among those, cellular and extracellular oxidative stress, along with chronic inflammation are best-known to have a role in disease development and/or progression. ^4–6  

Aging, the strongest risk factor for AMD, is associated with structural and functional changes in Bruch's membrane and the RPE. Drusen, a hallmark of AMD, ⁷ are formed between RPE and Bruch's membrane, and contain polymorphous material. Major components of this material are lipid aggregates containing numerous lipoproteins and cholesterol. ^8,9 These results established a connection between lipid metabolism and AMD etiology, which was confirmed further by genetic studies. Genome-wide association scans identified several genes contributing to high density lipoprotein (HDL) metabolism to be associated with AMD, among them hepatic lipase (LIPC), cholesterylester transfer protein (CEPT), lipoprotein lipase (LPL), and ATP-binding cassette, sub-family A1 (ABCA1). ^10,11 We and others also have found an association of paraoxonase 1 (PON1), a gene involved in HDL function, with advanced AMD in single populations. ^12–14  

Together with PON2 and PON3, PON1 forms a family of lactonases with antioxidant properties, which differ in sites of synthesis and mechanisms of action. ¹⁵ The PON1 gene is mainly expressed in liver ¹⁶ and encodes a secreted enzyme with a broad spectrum of functions that favor the involvement of PON1 in AMD pathology. In particular, PON1 present in serum HDL inhibits low density lipoprotein (LDL) oxidation, or “neutralizes” oxidized LDL by hydrolyzing lipid peroxides. ^17,18 Additionally, PON1 acts in an anti-inflammatory manner by reducing monocyte chemotaxis and adhesion to endothelial cells, ¹⁹ inhibiting monocyte-to-macrophage differentiation, ²⁰ and directly suppressing macrophage proinflammatory responses. ²¹ The antioxidant as well as the anti-inflammatory activities of PON1 suppress the formation of atherosclerotic plaques, pathologic lesions that share several properties with drusen. ^22,23 Whereas PON2 is an ubiquitous intracellular protein that can protect endoplasmic reticulum and mitochondria against reactive oxygen species (ROS)–mediated damage, ^24,25 PON3 acts similarly to PON1, as it is transported by HDL and protects HDL from oxidation. ²⁶  

The genetic studies mentioned above, biochemical analyses reporting lower serum paraoxonase activity in patients with AMD, ²⁷ and reports showing that smoking reduces PON1 activity in serum, ^28,29 are indications for a potential involvement of PON1 in the pathophysiology of AMD. Nevertheless, the role of PON1 in the eye has not been studied to our knowledge. We used Pon1^−/− mice to investigate the role of PON1 in the retina/RPE to elucidate its potential contribution to retinal lesions and eye pathology. 

Materials and Methods

Mice

Animals were treated in accordance with the regulations of the Veterinary Authority of Zurich and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Wild type (WT) 129S6/SvEvTac mice were purchased from Taconics (Eiby, Copenhagen, Denmark). The Pon1^−/− mice (B6.129X1-Pon1^tm1Lus/J) ³⁰ were purchased from Jackson Laboratory (Bar Harbor, ME, USA), and genotyped by PCR using the following primers: common forward primer 5′-CTT GTC CAT CCT CAG CTT GT-3′, WT reverse primer 5′-CCG ATG GTT CTT GTA AAG TGC-3′, mutant reverse primer 5′-CTT GGG TGG AGA GGC TAT TC-3′. The Pon1^−/− mice were backcrossed for 10 generations onto the 129S6/SvEvTac background before analyses. All mice were kept at the animal facility of the University Hospital Zurich in a dark–light cycle (12 hours:12 hours) with 60 lux of light at cage level and a normal chow diet. 

Light Exposure

Six- to 8-week-old mice were dark adapted overnight (16 hours). Pupils were dilated with 1% cyclogyl (Alcon, Cham, Switzerland) and 5% phenylephrine (Ciba Vision, Niederwangen, Switzerland) 30 minutes before exposure to 17,000 lux of white light for 2 hours. After light exposure, mice were kept in darkness until the next day before being returned to cyclic light. Mice were killed at different time points after light offset (N = 3 for each group) and retinas, eyecups, or whole eyeballs were removed. Mice that were dark-adapted, but not exposed to light served as controls. 

Laser Capture Microdissection

Eyes of 8-week-old WT and Pon1^−/− mice were enucleated, immediately embedded in tissue freezing medium (Leica Microsystems Nussloch GmbH, Nussloch, Germany), and frozen in a 2-methylbutane bath cooled in liquid nitrogen. Retinal sections (20 μm) were collected on Arcturus PEN Membrane Glass Slides (Applied Biosystems, Foster City, CA, USA), fixed (5 minutes in acetone), air dried (5 minutes), and dehydrated (30 seconds in 100% ethanol, 5 minutes in xylol). Retinal layers were isolated using an Arcturus XT Laser Capture Microdissection system (Bucher Biotec AG, Basel, Switzerland) and Arcturus CapSure Macro LCM Caps (Applied Biosystems). RNA was isolated using the Arcturus PicoPure RNA Isolation Kit (Applied Biosystems) according to the manufacturer's directions, including a DNase treatment to remove residual genomic DNA. cDNA was synthesized using random hexamer primers (High-Capacity cDNA Reverse Transcription Kit; Applied Biosystems), and further analyzed by real-time PCR as described in the section “RNA Preparation and Semiquantitative Real-Time PCR.” 

Electroretinogram (ERG)

The ERGs were recorded from both eyes simultaneously following published protocols. ^31,32 Briefly, mice were dark-adapted overnight and anesthetized the next day with ketamine (66.7 mg/kg; Ratiopharm GmbH, Ulm, Germany) and xylazine (11.7 mg/kg; Bayer HealthCare, Monheim, Germany). Pupils were dilated with 1% cyclogyl and 5% phenylephrine for 30 minutes before performing single flash ERG recordings under dark-adapted (scotopic) and light-adapted (photopic) conditions. Light adaptation was accomplished with low background illumination starting 5 minutes before photopic recordings. Single white-flash stimulus intensities ranged from −3.7 to 1.9 log cd·s/m² under scotopic and from −0.6 to 2.9 log cd·s/m² under photopic conditions, divided into 12 and 8 steps, respectively. Ten responses per flash intensity were averaged with an interstimulus interval of either 4.95 or 16.95 seconds (for 1.4, 1.9, 2.4, and 2.9 log cd·s/m²). 

Fundus Imaging and Fluorescein Angiography

Pupils were dilated with 1% cyclogyl and 5% phenylephrine. After 30 minutes, mice were anesthetized with ketamine (66.7 mg/kg) and xylazine (11.7 mg/kg), and corneas were moistened with 2% methocel (OmniVision, Puchheim, Germany). Fundi were monitored and photographed using a mouse imaging system (Micron III; Phoenix Research Laboratories, Pleasanton, CA, USA). Fluorescein solution (2% in PBS; AK-FLUOR, Lake Forest, IL, USA) was injected intraperitoneally and eyes were analyzed immediately. 

Light and Transmission Electron Microscopy (TEM)

Eyes were enucleated and fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) at 4°C overnight. Cornea and lens were removed and eyecups cut dorsoventrally through the optic nerve head. Trimmed tissue was washed in cacodylate buffer, incubated in osmium tetroxide for 1 hour at room temperature, dehydrated, and embedded in Epon 812. For light microscopy (Axioplan 2; Carl Zeiss AG, Feldbach, Switzerland) semithin cross-sections (500 nm) were cut and counterstained with toluidine blue. For TEM, ultrathin sections (50 nm) were stained with uranyl acetate and lead citrate, and analyzed using a Philips CM100 transmission electron microscope (Philips, Amsterdam, The Netherlands). 

RPE Flat Mount Preparation and Analysis

Unless stated otherwise, all procedures were conducted at room temperature. Eyes were enucleated and incubated in 2% paraformaldehyde (PFA) in 0.1 M phosphate buffer pH 7.4 (PB, 0.081M Na₂HPO₄, 0.019M NaH₂PO₄ × H₂O) for 5 minutes. Cornea and lens were removed and the remaining tissue was left in PB-salt (PB containing 140 mM NaCl and 2.7 mM KCl) for 20 minutes to allow the retina to separate from the eyecup. The retina was removed gently and the eyecup containing the RPE cut into a “clover-leaf” shape and post-fixed in 4% PFA in PB for 1 hour. Flat mounts were blocked with blocking solution (3% normal goat serum, 0.3% Triton X-100 in PBS) for 1 hour and incubated with the primary antibody anti-β-catenin (1:300 in blocking solution; BD Biosciences, Allschwil, Switzerland) overnight at 4°C. After washing 3 × 10 minutes in PBS, Cy3-conjugated anti-mouse secondary antibody (1:200 in blocking solution; Jackson ImmunoResearch, Suffolk, UK) and Alexa Fluor 488-phalloidin (1.3 U/mL in blocking solution; Applied Biosystems) were applied for 2 hours. After washing, cell nuclei were stained with Hoechst (2 μg/mL in PBS; Sigma-Aldrich, Buchs, Switzerland) for 30 minutes. Flat mounts were washed with PBS and mounted with anti-fade medium (10% Mowiol 4–88 (wt/vol); Calbiochem, San Diego, CA, USA; 25% glycerol (wt/vol); 0.1% 1,4-diazabicyclo[2.2.2]octane in 100 mM Tris, pH 8.5). The RPE sheets were examined using a digitalized fluorescence microscope and an ApoTome module (Axioplan 2; Carl Zeiss AG). 

Morphometric measurements, including eccentricity and form factor were performed with CellProfiler software. ³³ Phalloidin-stained images (magnification, ×20) were illumination corrected and analyzed using the “Tissue Neighbours” pipeline with the background adaptive thresholding method. At least N = 700 RPE cells per group were examined. The RPE cells were counted in 4 quadrants of 176.8 × 235.3 μm (1040 × 1384 pixels), approximately 800 to 900 μm temporal, dorsal, nasal, and ventral of the optic nerve head. Three animals per genotype and condition were analyzed. 

RNA Preparation and Semiquantitative Real-Time PCR

Retinas were removed through a slit in the cornea and placed in an Eppendorf tube. The rest of the eye (eyecup without lens) was isolated separately. All samples were immediately frozen in liquid nitrogen and stored at −80°C. Total RNA from retina and eyecups was prepared using the High Pure RNA Isolation Kit (Roche Diagnostics, Mannheim, Germany) or the RNeasy kit (Qiagen, Hilden, Germany), respectively. RNA isolations included a DNase treatment to digest residual genomic DNA. Equal amounts of RNA were reverse transcribed using oligo(dT) primer and M-MLV reverse transcriptase (Promega, Madison, WI, USA). Real-time PCR with specific primer pairs (Table 1), a polymerase ready mix (LightCycler 480 SYBR Green I Master Mix; Roche Diagnostics), and a thermocycler (LightCycler, Roche Diagnostics) were used to study gene expression. Three animals per time point were analyzed in duplicates. Signals were normalized to β-actin and relative gene expression was calculated using the ΔΔCt method. 

Table 1

 

Primers and Conditions for Real-Time PCR

Gene	Forward 5′–3′	Reverse 5′–3′	Annealing Temperature, °C	Product, bp
Pon1	GCATCTGAAAACCATCACACA	AAGCTCTCAGGTCCAATAGCA	60	72
Pon2	CAGAGGCTCTTCGTGTACCA	ATGTTCTGAATGCGGAGGAC	60	88
Pon3	TTGACCGTTGATCCAGCCAC	GAAGCACAGAGCCGTTGTTC	62	175
Sod1	GAGCAGAAGGCAAGCGGTGA	AGGTCCTGCACTGGTACAGC	62	126
Sod2	GACCTGCCTTACGACTATGG	CTGAAGAGCGACCTGAGTTG	62	168
Hmox1	CCGCCTTCCTGCTCAACATT	GACGAAGTGACGCCATCTGTG	62	99
Abca1	GCGTGAAGCCTGTCATCTAC	CATGAGAGGAGTGATCGACC	62	185
Scarb1	CGCACAGTTGGTGAGATCCT	CACCAGATGGATCCTGCTGA	62	183
Lif	AATGCCACCTGTGCCATACG	CAACTTGGTCTTCTCTGTCCCG	60	216
Edn2	AGACCTCCTCCGAAAGCTG	CTGGCTGTAGCTGGCAAAG	60	64
Fgf2	TGTGTCTATCAAGGGAGTGTGTGC	ACCAACTGGAGTATTTCCGTGACCG	62	158
Stat3	CAAAACCCTCAAGAGCCAAGG	TCACTCACAATGCTTCTCCGC	62	139
Socs3	ATTTCGCTTCGGGACTAGC	AACTTGCTGTGGGTGACCAT	58	126
Mct3	GGCTCAACCCTAAATCCAGA	CTTCGGAGTTTCCTCACCAG	58	75
Gnat1	GAGGATGCTGAGAAGGATGC	TGAATGTTGAGCGTGGTCAT	58	209
Vsx2	CCAGAAGACAGGATACAGGTG	GGCTCCATAGAGACCATACT	60	111
Opn4	CCAGCTTCACAACCAGTCCT	CAGCCTGATGTGCAGATGTC	62	111
Actb	CAACGGCTCCGGCATGTGC	CTCTTGCTCTGGGCCTCG	62	153

Lipid Extraction

For most analyses, lipids were extracted from retinal and eyecup tissue together. We will refer to these samples as “retina/eyecup” samples from here on. In some instances retinas and eyecups were isolated separately and named accordingly. Isolated tissue samples were frozen immediately in liquid nitrogen and stored at −80°C. Retina/eyecup samples were homogenized in lysis buffer (PBS, 0.2% Triton X-100 [vol/vol]) using a Precellys 24 tissue homogenizer (Bertin Technologies, Montigny-le-Bretonneux, France). The protein content was measured using the Bradford assay. Total lipids were extracted from aliquots of homogenized tissue containing 40 μg of protein. Extraction was conducted according to the procedure of Bligh and Dyer ³⁴ in the presence of 200 ng of the internal standards PG 17:0/17:0, PA 14:0/14:0, PE 14:0/14:0, PC 14:0/14:0, PC 24:0/24:0, LPC 17:0, SM d18:1/12:0, and Cer d18:1/17:0 (Avanti Polar Lipids, Alabaster, AL, USA). Samples were mixed with 375 μL of methanol/chloroform (2:1, vol/vol) and vortexed, followed by the addition of 100 μL water and 125 μL chloroform. The mixture was shaken for 15 minutes and centrifuged at 16,100g for 5 minutes at 25°C. The lower phase was collected, 250 μL chloroform added, shaken for 15 minutes, and centrifuged at 16,100g for 5 minutes at 25°C. All lower phases were combined and evaporated to dryness under a stream of nitrogen. Dried material was reconstituted in 200 μL of a mixture of mobile phases A (80%) and B (20%, see below). Then, 10 μL were injected into the liquid chromatography–mass spectrometry (LC-MS) system for phospholipic (PL) analysis. 

Analysis of the Retina/Eyecup Lipidome: LC-MS Instrumentation and Chromatographic Conditions

Analyses of six different PL classes: phosphatidylcholines (PC), sphingomyelins (SM), phosphatidylethanolamines (PE), phosphatidylglycerols (PG), phosphatidic acids (PA) and ceramides/hexosylceramides (Cer/HexCer) were performed by using LC-MS. Details of the method are described in the Supplementary Material. Briefly, total lipid extracts were analyzed on an LC-MS system consisting of a Rheos 2200 pump (Flux Instruments, Reinach, Switzerland), an HTC PAL autosampler (CTC Analytics, Zwingen, Switzerland), and a TSQ Quantum Access triple quadrupole mass analyzer (Thermo Fisher Scientific, Waltham, MA, USA). Separation of lipid extracts was performed on a diol silica-based column (QS Uptisphere 6 OH, 150 × 2.1 mm, 5 μm; Interchim, Montlucon, France). Mobile phase A was a mixture of hexane/isopropanol/water (70:30:2, vol/vol) with 15 mM ammonium formate. Mobile phase B was isopropanol/water (50:2, vol/vol) with 15 mM ammonium formate. The solvent-gradient was 0 to 7 minutes A/B (%) 80/20, 8 to 10 minutes A/B (%) 60/40, 11 to 23 minutes A/B (%) 40/60, and 25 to 30 minutes A/B (%) 80/20 at a flow rate of 0.35 mL/min. The column was maintained at 30°C. Mass spectrometry was performed in positive ionization mode with the following parameters: 4500 V spray voltage, skimmer voltage 2 to 14 V (depending on the scan mode), 250°C capillary temperature, and 10 and 6 (arbitrary units) sheath and auxiliary N₂ gas, respectively. Molecular masses provided by a neutral loss and precursor scan were used to selectively detect specific phospholipids (PLs). A neutral loss of masses m/z 115 and 189 from [M+NH₄]⁺ ions were used for analysis of PA and PG lipids, respectively. A precursor ion scan of m/z 184 specific for phosphocholine-containing lipids was used for PC, SM, and LPC A neutral loss scan of m/z 141 was used for PE and a precursor scanning of m/z 264 was applied for screening of Cer. Acquired data were analyzed using Xcalibur (version 2.0.6, Thermo Fisher Scientific). The structure of the head groups were determined from the type of MS/MS scanning. Molecular species were identified by using LIMSA. ³⁵ All species were assigned to the lipid classes and species with a defined total number of carbon atoms and double bounds in O-linked fatty acids. Data were corrected for isotopic overlap. Quantification was based on calibration curves for seven PL standards including PC 16:0/18:2, SM d18:1/16:0, LPC 16:0, PE 18:0/18:0, Cer d18:1/14:0, PG 16:0/16:16, and PA 14:0/14:0. The calibration curves were constructed by plotting the PL/IS peak area ratios against the nominal concentration of the standards. The concentrations for the individual lipid classes were calculated from linear regressions of the calibration curves. 

Statistical Analysis

Semiquantitative Real-Time PCR and Morphometric Analysis of the RPE Flat Mounts.

Statistical analyses were performed using Prism4 software. All data are given as means ± SD of three animals per group. Statistical differences of means were calculated using 2-way ANOVA followed by a Bonferroni post hoc test. A P value of less than 0.05 was considered significant. 

Lipidomic Analysis.

Statistical analyses were performed using SPSS, version 19 (SPSS Switzerland, Zurich, Switzerland). Normality of the data was determined by using Kolmogorov-Smirnov test. Since not all lipid species had normally distributed values, the nonparametric Mann-Whitney U test was used. The Mann-Whitney U test was applied for comparison between Pon1^−/− and WT mice. Bonferroni correction was applied to adjust the P value for multiple comparisons (37 statistical tests). A P value of 0.001 or below was considered significant. Data are presented as median with lower and upper range values. 

Results

Pon1 is Highly Expressed in the RPE

To evaluate the expression pattern of paraoxonases in the retina we isolated the RPE, the outer nuclear layer (ONL), inner nuclear layer (INL), and ganglion cell layer (GCL) by laser capture microdissection, and analyzed levels of Pon1, Pon2, and Pon3 mRNA by semiquantitative real-time PCR (Fig. 1). Expression of monocarboxylate transporter, member 3 (Mct3; RPE), guanine nucleotide binding protein, alpha transducing activity polypeptide (Gnat1; ONL), visual system homeobox 2 (Vsx2; INL), and melanopsin (Opn4; GCL) were used as layer-specific markers and showed that cross-contamination between collected layers was minimal. In WT mice, Pon1 was highest expressed in the RPE with 267-, 590-, and 651-fold higher levels than in ONL, INL, and GCL, respectively. Similarly, Pon3 was most abundant in the RPE (45-, 43-, 72-fold higher levels than in ONL, INL, and GCL, respectively). The Pon2 content was comparable in the RPE and the INL, approximately 25- and 9-fold higher than in ONL and GCL, respectively. 

Figure 1

 

Pon1 is highly expressed in the RPE. The RPE, ON), IN,) and GCL were isolated from WT (grey bars) and Pon1^−/− mice (black bars) using laser capture microdissection. Relative gene expression was determined by semiquantitative real-time PCR in each individual layer. Values were normalized to β-actin, and expressed relative to levels in the RPE of Pon1^−/− mice, which were set to 1. Shown are mean values ± SD of N = 3. Expression of Mct3 (marker for RPE), Gnat1 (marker for ONL), Vsx2 (marker for INL), and Opn4 (marker for GCL) was determined to monitor contamination between the isolated retinal layers.

Even though Pon1^−/− mice have a neomycin cassette inserted into exon 1 of Pon1 and no paraoxonase activity (Supplementary Fig. S1), ³⁰ the RNA transcript from the disrupted gene can still be detected by RT-PCR. Interestingly, we found differential effects of the neomycin insertion on Pon1 expression in RPE and neuronal retina. Whereas expression of the disrupted Pon1 gene was 33-fold decreased in the RPE, it was 15-fold increased in the INL and 6-fold in the GCL, indicating differential regulation of Pon1 expression in RPE and neuroretina. Levels of Pon2 and Pon3 transcripts were not significantly affected by the lack of functional PON1. 

Absence of PON1 Does Not Compromise Retinal Function and Architecture

To evaluate whether PON1 is important for normal retinal development, ageing, and function, we analyzed 8-week and 1-year-old mice. Pon1^−/− mice showed regular retinal architecture and no signs of RPE or photoreceptor degeneration at either age (Fig. 2A). Similarly, funduscopy and fluorescein angiography revealed no retinal or vascular defects in Pon1^−/− mice (Fig. 2B). Normal scotopic and photopic ERGs with a- and b-wave amplitudes similar to WTs (Fig. 2C) supported the conclusion that Pon1^−/− mice developed a normal and functional retina. 

Figure 2

 

Lack of PON1 does not compromise retinal architecture and function. (A) Retinal morphology of Pon1^−/− and control mice (WT) at 8 weeks and 1 year of age. Shown are representative sections of N = 3. POS, photoreceptor outer segments; PIS, photoreceptor inner segments. Scale bar: 50 μm. (B) Funduscopy (top panels) and fluorescein angiography (bottom panels) of WT and Pon1^−/− mice at 8 weeks of age. Shown are representative images of N = 3. (C) Scotopic and photopic electroretinograms from WT and Pon1^−/− mice at 8 weeks of age. Top panels: representative traces from individual mice. Bottom panels: scotopic a- and b-wave amplitudes, as well as photopic b-wave amplitudes as a function of light intensity. N = 3.

Lack of PON1 Does Not Affect RPE Morphology

Since we found high Pon1 expression in the RPE and PON1 was previously implicated in AMD, ^12–14,27 a disease strongly affecting RPE and Bruch's membrane, we examined these structures in young and old mice. Transmission electron microscopy did not detect any obvious differences in the ultrastructure of RPE and Bruch's membrane between Pon1^−/− and WT mice at any age (Fig. 3A). Also, RPE microvilli and basal infoldings appeared normal and no AMD-like features, such as Bruch's membrane thickening or sub-RPE deposits, ³⁶ were present in any of the analyzed animals up to one year of age. 

Figure 3

 

Lack of PON1 does not affect RPE morphology. The RPE was analyzed in Pon1^−/− and control (WT) mice at 8 weeks and 1 year of age. (A) TEMs of the RPE and adjacent tissues. Shown are representative sections of N = 3 for each panel. N, nucleus; BrM, Bruch's membrane; BI, basal infoldings; MV, apical microvilli; M, melanosomes. Scale bar: 2 μm. (B) RPE flat mounts immunostained for β-catenin (red) and phalloidin (green). Nuclear staining (Hoechst) is visible in blue. Shown are representative microphotographs of N = 3 for each panel. Scale bar: 50 μm. (C) Morphometric analysis of RPE cells. Shown are mean values ± SD of at least N = 700 RPE cells examined per group. A 2-way ANOVA revealed no significant differences in RPE cell shape between WT and Pon1^−/− mice at 8 weeks and 1 year of age (P > 0.05).

To analyze RPE cell morphology on a larger scale, we prepared RPE flat mounts. The RPE monolayer normally is formed by a relatively uniform array of polygonal cells containing one to two centrally positioned nuclei. The regular shape is maintained by adherens and tight junctions, as well as by the actin-myosin cytoskeleton, whose components may serve as markers of epithelial integrity. We monitored the RPE cell boundaries by immunofluorescent staining for β-catenin (adherens junctions) and phalloidin (F-actin), and did not observe any abnormalities in mice lacking PON1 up to one year of age (Fig. 3B). This was supported further by morphometric analyses, which were used to quantitatively assess RPE cell shape. The eccentricity (ECC) and form factor (FF) are two measures of distortion from a circular shape, whose values vary between 0 and 1, with ECC = 0 and FF = 1 for a perfect circle. These factors did not differ significantly between cells in the RPE of Pon1^−/− and age-matched WT mice (Fig. 3C). 

Gene Expression in Pon1 ^−/− Mice

To reveal whether the lack of PON1 influences expression of related genes or of genes encoding known PON1 interaction partners, we analyzed gene expression in retina and eyecup at 8 weeks and 1 year of age. Similar to the LCM data (Fig. 1), expression of Pon1 was lower in the neuronal retina compared to that in the eyecup (Fig. 4). Cross-contamination between RPE and neuronal retina was minimal as shown by the relative expression of Mct3 (RPE) and Gnat1 (retina) in both isolated compartments (Supplementary Table S1). 

Figure 4

 

Expression of PON1-related genes Expression of genes related to PON1 function was determined by semiquantitative real-time PCR in retinas (upper panels) and eyecups (lower panels) of WT (white bars) and Pon1^−/− mice (black bars) at 8 weeks and 1 year of age as indicated. Shown are mean values ± SD of N = 3. Values were normalized to β-actin, and levels in WT mice at 8 weeks were set to 1. A 2-way ANOVA with Bonferroni post hoc analysis was used to test statistical significance. A P value of less than 0.05 was considered significant; **P < 0.01; ***P < 0.001. Shown are statistics for the comparison of expression levels in WT and Pon1^−/− mice at 8 weeks or 1 year of age. See Supplementary Table S1 for statistical analysis of age-dependent expression in WT and Pon1^−/− mice.

In accordance with the LCM experiment, we found significant up-regulation of Pon1 mRNA expression in the retina of 8 week old Pon1^−/− mice and down-regulation in the eyecup. Since lack of the antioxidative PON1 protein may influence the oxidative balance in a tissue, expression of other antioxidative genes may be affected. However, both other members of the paraoxonase family (Pon2 and Pon3), the two superoxide dismutases (Sod1 and Sod2), as well as heme oxygenase 1 (Hmox1) were similarly expressed in control and Pon1^−/− mice (Fig. 4). Remarkably, we detected an age-related moderate decrease in the expression of Sod1, Sod2, Pon3 (retina and eyecup), Pon2 (retina), Pon1, and Hmox1 (eyecup) independently of the Pon1 genotype (Fig. 4, Supplementary Table S1). 

Previous studies showed that the role of PON1 in HDL transport and function involves interaction with partner proteins, such as scavenger receptor class B (SCARB1, also known as SR-BI) and ATP-binding cassette transporter (ABCA1). Whereas SCARB1 facilitates acquisition of PON1 secreted from hepatocytes, ³⁷ interaction between PON1 and ABCA1 transporter enhances HDL-mediated cholesterol efflux from macrophages. ³⁸ Both SCARB1 and ABCA1 were shown previously to be involved in lipid efflux and reverse cholesterol transport in retina and RPE, ³⁹ and this function might depend on interaction with PON1. Nevertheless, expression of Scarb1 was similar in control and Pon1^−/− mice at either age. The expression of Abca1 also was similar in old control and knockout mice, whereas expression in young Pon1^−/− mice was variable in different sets of animals (see also Fig. 5), suggesting the influence of other, yet unknown factors. Further studies are required to investigate a potential functional connection between Pon1 and Abca1 in the retina. 

Figure 5

 

PON1 does not protect against damage by acute light exposure. (A) Representative light micrographs of retinas of 8-week-old control (WT) and Pon1^−/− mice at 5 and 10 days after light exposure. Control mice (dark control, DC) were reared in normal cyclic light conditions (12 hours dark; 12 hours light) and not exposed to light. Arrows: pyknotic nuclei indicating photoreceptor cell death. Arrowheads: macrophages in the subretinal space. Shown are representative sections of N = 3. Scale bar: 50 μm. (B) Immunostaining of RPE flat mounts from 8-week-old WT and Pon1^−/− DC mice and at 10 days after light exposure. Red: β-catenin. Green: phalloidin. Arrowheads: diffuse β-catenin staining in RPE cells of light exposed mice. Scale bar: 50 μm. (C) Gene expression in retinas (upper panels) and eyecups (lower panels) of 8-week-old WT (white bars) and Pon1^−/− (black bars) mice at 12 hours after light exposure or in DC mice. Values were normalized to b-actin and expressed relative to the levels in the retina of unexposed WT animals, which were set to 1. Shown are mean values ± SD of N = 3. A 2-way ANOVA with Bonferroni post hoc analysis was used to test significance. A P value of less than 0.05 was considered significant; *P < 0.05; **P < 0.01; ***P < 0.001. Shown are significantly different expression levels between WT and Pon1^−/− mice in DC or light exposed mice. See Supplementary Table S2 for statistical analysis of differences in gene expression between DC and light-exposed WT and DC and light-exposed Pon1^−/− mice.

Alterations of PL Composition in Retina/Eyecup Samples of Pon1 ^−/− Mice

The retina contains very high levels of polyunsaturated fatty acids (PUFAs). The majority of retinal PUFAs are esterified to PLs, which build up the membranous disks of photoreceptor outer segments. ⁴⁰ Due to their molecular structure and constant exposure to a photo-oxidative environment, ^5,6 retinal PLs are particularly prone to oxidative modifications. Based on the fact that PON1 can reduce lipid peroxidation in LDL and cell membranes, ^17,19 we hypothesized that PON1 might have a similar effect on retinal PLs, and, thus, affect their composition. To test this assumption, we employed a combination of LC and MS for the identification and quantification of retina/eyecup PLs from WT and Pon1^−/− mice. 

The overall PL profile was similar in the retina/eyecup samples of wt and Pon1^−/− mice at the age of 8 weeks. The most abundant PLs were PCs and PEs (Table 2). The most abundant PE species was PE 40:6, which possibly contained polyunsaturated C_22:6n-3 docosahexaenoic (DHA) fatty acid at the sn-1 or sn-2 position. The PC 40:6 was the third most abundant PC species after PC 32:0 and PC 34:1. Interestingly, highly unsaturated PC 44:12 and PE 44:12 species were identified in both mouse strains. They probably represent PC and PE lipids that were esterified with two molecules of DHA at the sn-1 and sn-2 position. The third abundant lipid class was SM. Among lysophosphatidylcholines (LPCs), the most abundant species contained a DHA fatty acid (LPC 22:6). Ceramides/hexosylceramides (Cer/HexCer), PGs, and PAs were found at very low concentrations; therefore, we report only total concentrations of these three lipid classes. 

Table 2

 

PL Species in the Retina/Eyecup Samples From Control (WT) and Pon1^−/− Mice at 8 Weeks of Age

PLs	Content of PL in nmol/mg of Retina and Eyecup Proteins		Mann-Whitney Test, P Values WT 8 w vs. Pon1−/− 8 w
	WT 8 w Median (min;max)	Pon1−/− 8 w Median (min;max)
Total PC	143.5 (92;170.5)	120.5 (54.1;140.3)	0.023
Total SM	20 (7.2;26.8)	17 (7.2;23.7)	0.174
Total LPC*	5.2 (2.1;5.9)*	2.5 (1.2;4.1)*	0.001*
Total PE	26.5 (16.2;29.8)	21.9 (11.4;24.6)	0.013
Total Cer/HexCer	5.6 (2.7;7.6)	6.1 (2.6;9.4)	0.545
Total PG	1.2 (0.2;1.9)	0.9 (0.3;1.5)	0.253
Total PA	2.8 (2.3;3.1)	2.9 (2.5;3.2)	0.326
PC 32:0	27.3 (15.5;31.5)	23.3 (10.9;27.7)	0.070
PC 32:1	2.7 (2.1;3.5)	2.1 (1.1;2.6)	0.010
PC 34:0	6.4 (3.7;8)	5.9 (3;7.6)	0.257
PC 34:1	27.8 (18.8;33.4)	23.4 (10.9;28.9)	0.016
PC 34:2	4.7 (3;5.9)	4.2 (1.7;4.6)	0.096
PC 36:1	14.1 (8;16.2)	11.1 (5.3;13.3)	0.019
PC 36:2	5.6 (4.3;7.2)	4.7 (2.2;5.8)	0.023
PC 36:3	1.5 (1;2)	1.3 (0.4;1.7)	0.070
PC 36:4	6.4 (4.1;7.2)	5.4 (2.6;6.3)	0.059
PC 38:4	7.9 (4.7;10.4)	6.6 (2.8;8.3)	0.013
PC 38:5	2.5 (1.6;2.8)	2.4 (0.9;2.9)	0.450
PC 38:6	12.1 (7.6;15.1)	9.5 (4;12.3)	0.013
PC 40:5	1.7 (0.7;2.3)	1.3 (0.8;1.9)	0.023
PC 40:6	17.7 (12.2;20.4)	13.7 (5.7;15.6)	0.004
PC 40:7	1.4 (1.2;2.4)	1.1 (0.5;1.5)	0.008
PC 44:12	4.9 (3.7;8.3)	3.9 (1.4;5.7)	0.016
SM34:1	5.8 (2.3;7.6)	5.2 (2.4;7.2)	0.174
SM36:1	4.7 (1.9;6.4)	3.9 (1.9;5.5)	0.199
SM38:1	2.4 (0.8;3.3)	2.2 (0.7;2.9)	0.199
SM40:1	2.2 (0.8;3.2)	2 (0.8;2.9)	0.199
SM42:1	2.1 (0.6;2.7)	1.7 (0.5;2.6)	0.131
SM42:2	2.2 (0.8;3.4)	1.8 (0.7;2.7)	0.257
SM42:3	0.4 (0.1;0.5)	0.3 (0.1;0.4)	0.059
LPC 18:0	0.8 (0.6;0.8)	0.7 (0.4;0.9)	0.131
LPC 18:1	1.7 (0.6;2)	0.7 (0.4;1.2)	0.002
LPC 22:6*	2.8 (0.9;3.1)*	1.2 (0.4;2)*	0.001*
PE 34:1	0.8 (0.6;0.9)	0.7 (0.4;0.9)	0.290
PE 36:0	0.3 (0.2;0.4)	0.3 (0.2;0.4)	0.880
PE 36:2	0.9 (0.8;1.2)	0.9 (0.7;1.1)	0.151
PE 36:4	0.6 (0.4;0.8)	0.6 (0.4;0.8)	0.650
PE 38:4	3.3 (1.8;3.7)	2.6 (1.5;3.1)	0.019
PE 38:6	5 (2.8;6.3)	4.2 (2;5.1)	0.008
PE 40:6	12.2 (7.6;13.8)	10.3 (5.1;11.4)	0.013
PE 44:12	2.9 (1.9;3.8)	2.3 (1.1;2.7)	0.013

 The PL analysis was performed by LC-MS. Content of PL classes and species is shown as nmol of PL per 1 mg of retina and eyecup proteins. Total PLs represent sum of molecular species of a given class and are presented for Cer/HexCer, PGs, and PAs Species pattern is displayed for PCs, PEs, SM, and LPCs. Data values are displayed as median with range. N = 10 for both genotypes. Statistical significance was determined by the Mann-Whitney U test. After Bonferroni correction a P value of 0.001 corresponds to a significance level of 0.05. An assignment of PL species includes number of carbon atoms and double bounds in both acyl chains (for example, PC 34:1 consist of 34 carbon atoms and one double bond and can correspond to the following species PC 16:0/18:1 or PC 18:0/16:1, and so forth). SM assignment comprises the number of carbon atoms and double bonds in both sphingoid base and in N-linked fatty acid.

*  Statistically significant results.

The major difference between 8-week-old Pon1^−/− and wt mice was seen for total LPC (P = 0.001) and in particular for LPC 22:6 species (P = 0.001), which were detected at significantly reduced levels in Pon1^−/− mice (Table 2). These differences remained significant after Bonferroni correction for multiple testing. Analysis of PLs in WT retinas and eyecups, which were isolated separately, revealed that LPC 22:6 was the most abundant LPC species in the retina, whereas it was almost absent in the eyecup (data not shown). Some polyunsaturated and monounsaturated PCs as well as polyunsaturated PEs also were decreased in Pon1^−/− mice (P < 0.05). However, these differences missed significance after Bonferroni correction (Table 2). Other PL species also were not significantly different in Pon1^−/− compared to controls. When data for retinal PL were calculated as mole percent (Supplementary Table S4) no significant differences between Pon1^−/− and WT mice were detected. Although the relative abundance of total LPC (P = 0.0032) and LPC 22:6 (P = 0.0015) lipids was markedly reduced in retinas of Pon1^−/− mice, these differences lost significance after correcting for multiple testing. 

Since aging is a major risk factor for the development of AMD, and lipid metabolism is known to be involved in the etiology of this disease, we characterized the PL profiles of 1-year-old WT and Pon1^−/− mice. In contrast to young adults at 8 weeks of age, levels of total LPC and LPC 22:6 species in 1-year-old mice did not differ between the two mice strains (Table 3). Interestingly, age-related changes in the PLs profiles showed different trends in WT and Pon1^−/− mice. While control animals showed a trend for decreased levels of total LPC and LPC 22:6 with age, Pon1^−/− mice revealed almost no age-related changes in the content of those PLs (Table 3). 

Table 3

 

Levels of Total LPCs and LPC 22:6 in the Retina/Eyecup Samples From 8-Week and 1-Year Old WT and Pon1^−/− Mice, and From 8-Week Old Mice Exposed to the Light

PLs	Median (min;max) [nmol/mg]						Mann-Whitney Test, P Values
	WT 8 w	Pon1−/− 8 w	WT 1 y	Pon1−/− 1 y	WT LD	Pon1−/− LD	WT 8 w vs. WT 1 y	Pon1−/− 8 w vs. Pon1−/− 1 y	WT 1 y vs. Pon1−/− 1 y	WT 8 w vs. WT LD	Pon1−/− 8 w vs. Pon1−/− LD	WT LD vs. Pon1−/− LD
Total LPC	5.2 (2.1;5.9)	2.5 (1.2;4.1)	1.9 (1.5;6.8)	3.1 (1.3;5.3)	5 (2.2;6.5)	3.7 (3;8.6)	0.086	0.745	1.000	0.940	0.002	0.059
LPC 22:6	2.8 (0.9;3.1)	1.2 (0.4;2)	0.7 (0.5;3.8)	1.3 (0.4;2.5)	2.6 (1.3;3.2)	1.7 (1.4;4.8)	0.066	0.745	1.000	1.000	0.005	0.049

 The PL analysis was performed by LC-MS. The PL classes and species are shown as nmol of PL per one mg of retina and eyecup proteins. Displayed are median values with range. N = 5 for WT (1y), N = 6 for Pon1^−/− (1y), N = 10 for WT (LD), N = 10 for Pon1^−/− (LD). Statistical significance was determined by the Mann-Whitney U test. After Bonferroni correction a P value of 0.001 corresponds to a significance level of 0.05. An assignment of PL species includes numbers of carbon atoms and double bonds in both acyl chains. Total LPC represent sum of all species of a lysophosphatidylcholines class.

*  Samples were analyzed in a separate LC-MS run.

PON1 Does Not Influence Photoreceptor Degeneration or RPE Survival After Light Exposure

To evaluate whether the lack of the antioxidative PON1 protein influences cell survival under oxidative stress conditions, we exposed WT and Pon1^−/− mice to high levels of white light. ⁴¹ At 5 days after exposure, many photoreceptors were lost and large numbers of pyknotic nuclei were detected in the ONL of WT and Pon1^−/− mice (Fig. 5A). We also observed similar levels of subretinal macrophages in Pon1^−/− and WT mice, although previous reports suggested inhibitory effects of PON1 on monocyte transmigration. ³⁰ At 10 days after light exposure the photoreceptor layer was degenerated further, with only few pyknotic nuclei left in the ONL of both Pon1^−/− and WT mice. Additionally, some RPE cells displayed severe morphologic changes in both mouse lines (Fig. 5B). Enlarged RPE cells with increased cytoplasmic β-catenin localization, as well as a global increase in cytoplasmic phalloidin (F-actin) staining were observed. The number of RPE cells before and after light exposure was comparable in WT and Pon1^−/− mice (data not shown). 

Acute LD not only induces retinal degeneration, but also activates survival pathways to protect retinal cells from cell death. Leukemia inhibitory factor (LIF) is a key cytokine regulating an endogenous rescue pathway that involves activation of the Janus kinase signal transducer and activator of transcription (Jak-STAT) signaling pathway, and expression of protective factors, like endothelin-2 (Edn2) and fibroblast growth factor-2 (Fgf2). ^42,43 In line with the comparable extent of tissue damage (Figs. 5A, 5B), light exposure caused a similar expression pattern of these genes in retinas of WT and Pon1^−/− mice, even though Edn2 and Fgf2 were significantly stronger induced in mice lacking PON1 (Fig. 5C). No differences between the two genotypes were detected in eyecups. Remarkably, upon exposure to the light Pon2, Sod1, and Sod2 expression was reduced in the retina, but not in the eyecup, while Hmox1 was upregulated in both tissues (Fig. 5C, Supplementary Table S2). Additionally, expression of Socs3 was strongly upregulated in the eyecup, suggesting that JAK/STAT signaling also was induced in the RPE. Whether this was a consequence of LIF signaling or whether other factors led to such activation must be established. 

Untreated Pon1^−/− mice had decreased levels of total LPCs and in particular of LPC 22:6 (see above). Since PON1 is an antioxidative enzyme, we tested whether its lack might impact the PL composition under oxidative stress induced by light exposure. At 2 hours after light exposure, PL profiles were comparable in WT and in Pon1^−/− retina/eyecups and showed increased levels of PE 34:1, PE 36:0, and PE 36:4 in both strains (Supplementary Table S3). Furthermore, light exposure increased the levels of total LPC and LPC 22:6 in Pon1^−/− , but not in WT mice (Table 3). However, these differences did not remain statistically significant after correcting for multiple testing. In agreement with a lack of difference, concentrations of potentially DHA containing PC's (PC 38:6, PC 40:6, PC 40:7, and PC 44:12) did not differ between WT and Pon1^−/− mice, neither before nor after light exposure. 

On the basis of mole percent, PE 36:0 levels increased in both strains upon light exposure, while PE 34:1, PE 36:4, PE 38:6, PE 40:6 species were more abundant in WT mice, but not in Pon1^−/− after the treatment. In contrary, PE 44:12 was relatively decreased in Pon1^−/− but not in WT mice. Relative increases of total LPC and LPC 22:6 were observed in Pon1^−/− , but not in WT mice; however, without statistical significance after correction for multiple testing and without any concomitant changes in PC (Supplementary Table S4). 

Discussion

PON1 in the Normal Retina

Since PON1 has been implicated in AMD, ^12–14,27 we assessed whether lack of PON1 would alter retinal physiology in young and old Pon1^−/− mice. We showed that PON1 is not essential for development and maintenance of a normal retinal structure, vasculature, or function. Similarly, RPE cells were morphologically indistinguishable between WT and Pon1^−/− mice, even though the RPE has the highest level of Pon1 expression. Although PON1 has antioxidative activity, no signs of oxidative damage were detected and no compensatory increase in the expression of other antioxidative enzymes was observed. Lack of PON1 also did not influence the susceptibility of retina and RPE to light damage. 

However, retina/eyecup samples of Pon1^−/− mice showed decreased levels of LPCs in general and of LPC 22:6 in particular. The LPCs are molecular species produced by the hydrolysis of PCs through enzymes of the phospholipase A2 (PLA2) superfamily. Previous studies have shown that PON1 has a PLA2-like activity leading to the formation and release of LPCs from macrophages. ^38,44 Thus, in WT mice, PON1 may increase LPC production and modulate the content of PLs in the retina/RPE. In analogy to the stimulatory effect of PON1-generated LPCs on cholesterol efflux from macrophages, ^44,45 LPCs might stimulate reverse lipid transport from RPE regulating lipid content in this cell layer. ⁴⁶  

Since PON1 is a secreted protein that retains its hydrophobic signal peptide, ⁴⁷ it may remain in the membrane of PON1-producing cells, or may reach membranes of neighboring cells. Thus, PON1 synthesized in the RPE might affect photoreceptors as well. The immunohistochemical detection of PON1 in the region of photoreceptor inner segments (and other nonnucleated areas of the retina) ⁴⁸ and our observation that LPC 22:6, which was reduced in Pon1^−/− mice, was very abundant in the retina, but almost absent from the eyecup (data not shown) supports a role for a PLA2-like activity of PON1 in the neuronal retina. 

We detected a general age-related decrease in the expression of antioxidative genes in retinas and especially eyecups of WT mice. Pon1 was among the affected genes and down-regulated about 65%. Remarkably, also levels of total LPCs and of LPC 22:6 were decreased in 1-year-old WT mice, by 63% and 75%, respectively (Table 3). This suggests a correlation between Pon1 expression and LPC levels, and indicates that reduced expression of Pon1 resulted in reduced PON1-related PLA2 activity and, thus, in a diminished PON1-mediated hydrolysis of PCs. Furthermore, increased lipid peroxidation and glycation in the aged eyes might have additionally compromised PON1 activity as it has been observed in metabolic syndrome and type II diabetes in humans. ^49–51  

PON1 in the Light-Exposed Retina

Even though PON1 has antioxidative and anti-inflammatory activities, Pon1^−/− and WT mice were similarly affected by light exposure: photoreceptors were equally susceptible to light induced degeneration; changes in RPE morphology after light exposure were comparable; expression of genes involved in antioxidative defense or the LIF-controlled cell survival pathway was similar; and content of PE 34:1, PE 36:0, and PE 36:4 in nmol/mg of proteins was increased to the same extent. However, levels of total LPC and LPC 22:6 increased only in Pon1^−/− , but not in WT mice after light exposure, suggesting that presence of PON1 prevents generation of LPC upon light exposure. Whether this is a direct effect of PON1 or whether a differential activation of phospholipases A2 in the two mouse strains influences LPC production remains to be explored. 

Several possible mechanisms may explain the lack of striking differences in response to light damage between Pon1^−/− and WT mice. Previous studies showed that PON1 may be inactivated by its own substrates, such as oxidized lipids. ⁵² Thus, generation of a large amount of oxidized molecules by light exposure may substantially reduce PON1 activity in WT mice. Alternatively, PON1 may “neutralize” only a subset of oxidized species produced in LD, such as oxidized PUFAs. ⁵³ Other light-induced oxidation products may remain toxic and participate in the induction of cell death. It also may be of importance that PON1 is a secreted protein ⁴⁷ that may not be able to detoxify intracellular molecules that have been generated by light exposure and that are involved in the degenerative process. 

PON1 in AMD

The condition of AMD is a complex, multifactorial disease. Most existing mouse models for AMD combine advanced age, which is the major risk factor to develop AMD, with genetic defects in key genes involved in AMD pathogenesis. Considering the complex nature of the disease, it may not be surprising that some models require the combination of more than two environmental or genetic risk factors to cause a pathological phenotype. An excellent example is a murine model that combines three known AMD risk factors: advanced age, high fat, and cholesterol-rich diet, and mutations in ApoE. ⁵⁴ Importantly, neither age nor the diet alone was sufficient to elicit changes that mimic human AMD pathology in the transgenic ApoE4 mice. 

The absence of an AMD-like phenotype in eyes of aged Pon1^−/− mice suggests that even long-lasting absence of PON1 is not sufficient to cause detectable pathologic changes, like thickening of Bruch's membrane or sub-RPE deposits typical for AMD. Noteworthy, despite the well-established antiatherogenic properties of PON1, Pon1^−/− mice developed significantly larger atherosclerotic lesions than WT controls only when raised on a high-fat/high-cholesterol, but not on a regular diet. ^30,55 This suggests that additional factors/stimuli may be needed to provoke phenotypes of complex diseases, like atherosclerosis or AMD in Pon1^−/− mice. This hypothesis also is supported by the observation that pathological processes in ApoE^−/− mice were slowed in the presence of a human PON1 transgene, ⁵⁶ but accelerated by PON1 deficiency. ^55,57  

These studies provide evidence that a coexisting chronic metabolic burden may be needed to reveal the role of PON1 in development and/or progression of a complex disease. A long-lasting exposure to low levels of light (in contrast to the short-term exposure to high light levels used in present study), a high-fat diet, or the inclusion of the ApoE-null genetic background might provide such an additional chronic stress to provoke AMD-like lesions in Pon1^−/− mice. 

We postulated that PON1 could be involved in AMD as a modulatory protein and that direct or indirect interactions with additional factors may influence its contribution to disease development and/or progression. Presence or absence of such factors may determine the association of PON1 with AMD in particular populations and may help to explain contradictory results regarding the implication of PON1 in AMD pathology. Nevertheless, the controversy on the role of PON1 in AMD still remains and there is a clear need for further studies to identify the precise function of PON1 in ocular tissues, such as RPE and neuronal retina. Knowledge of these functions seems a prerequisite to understand how the antioxidative and anti-inflammatory properties of PON1 may influence AMD. 

Supplementary Materials

pdf S1 

Acknowledgments

The authors thank Christel Beck, Andrea Gubler, and Coni Imsand for excellent technical assistance, and Marijana Samardzija for helpful discussions. 

Supported by a cooperative project grant by the Zurich Center for Integrative Human Physiology (ZIHP) of the University of Zurich, Zurich, Switzerland, and by a matching fund of the Center for Clinical Research of the University Hospital Zurich, Zurich, Switzerland. 

Disclosure: J. Oczos, None; I. Sutter, None; B. Kloeckener-Gruissem, None; W. Berger, None; M. Riwanto, None; K. Rentsch, None; T. Hornemann, None; A. von Eckardstein, None; C. Grimm, None 

Footnotes

 JO and IS contributed equally to the work presented here and should therefore be regarded as equivalent authors.