CD36 plays an important role in the clearance of oxLDL and associated age-dependent sub-retinal deposits.

Age-related macular degeneration (AMD) represents the major cause of vision loss in industrialized nations. Laminar deposits in Bruch's membrane (BM) are among the first prominent histopathologic features, along with drusen formation, and have been found to contain oxidized lipids. Increases in concentrations of oxidized LDL (oxLDL) in plasma are observed with age and high fat high (HFHC) cholesterol diet. CD36 is the principal receptor implicated in uptake of oxLDL, and is expressed in the retinal pigment epithelium (RPE). We determined if CD36 participates in oxLDL uptake in RPE and correspondingly in clearance of sub-retinal deposits. Uptake of oxLDL by RPE in vitro and in vivo was CD36-dependent. CD36 deficiency in mice resulted in age-associated accumulation of oxLDL and sub-retinal BM thickening, despite fed a regular diet. Conversely, treatment of HFHC-fed ApoE null mice with a CD36 agonist, EP80317 (300 μg/kg/day), markedly diminished thickening of BM, and partially preserved (in part) photoreceptor function. In conclusion, our data uncover a new role for CD36 in the clearance of oxidized lipids from BM and in the prevention of age-dependent sub-retinal laminar deposits.


INTRODUCTION
Age-related macular degeneration (AMD) is the leading cause of vision loss among older adults in industrialized nations [1]. One of the first signs of AMD is the presence of drusen and extracellular basal deposits in Bruch's membrane (BM) [2], leading to BM thickening, and ultimately is associated with injury to the retinal pigmented epithelium (RPE), choroid and photoreceptors [3][4][5]. With age, esterified cholesterol associated in lipoprotein-like particles accumulate subretinally in BM deposits [6][7][8][9][10].
Lipids in the RPE mostly originate from spent photoreceptor outer segments, and are exocytosed through the base of the RPE to be eliminated via the choroid [11][12][13]. However, RPE cells also take-up oxidized lipids from their cell base [14,15], suggesting they may be involved in clearing the sub-retinal space from such deposits [16]. Accumulation of lipoproteinand lipid-containing sub-RPE and BM debris implies an imbalance in the accumulation/clearance of these products, leading to cytotoxicity [17,18]. Correspondingly, high fat cholesterol diet (HFHC)-fed mice deficient in low-density lipoprotein (LDL) receptor ligand ApoE (ApoE-/-) and rabbits on fat-enriched diet, feature high LDL and oxidized LDL (oxLDL) plasma levels [19][20][21] and exhibit ultrastructural retinal changes similar to those observed in human AMD [22], including BM thickening and BM lipid deposits [23][24][25]. These findings support the concept that improper clearance of dietary lipids may influence the development of AMD [26].
The RPE and microvascular endothelium contain a number of scavenger receptors involved in uptake and metabolism of oxidized lipids [27,28]. Of the various scavenger receptors CD36 is a principal receptor of oxidized phospholipid [15] expressed at the basolateral side of RPE as well as on microvascular endothelial cells and macrophages [29][30][31][32][33]. Interestingly, a polymorphism of CD36 gene (in the non-coding region [possibly associated with its increased expression]) is protective against AMD [34]. We previously demonstrated that mice with CD36 gene disruption revealed a progressive age-dependent photoreceptor death and choroidal involution [32]. We hereby hypothesized that CD36 participates in oxLDL uptake in RPE and correspondingly in (clearance of) subretinal deposits. Our findings reveal that uptake of oxLDL in RPE is CD36-dependent and that a deficiency in CD36 leads to accumulation of sub-retinal deposits and oxLDL, despite fed a regular diet; conversely, stimulation of CD36 reduces sub-retinal deposits and preserves photoreceptor function.

RESULTS oxLDL uptake by RPE cells is CD36-dependent
Blood levels of oxLDL increase with age [40] and upon ingestion of high cholesterol diet [20], as corroborated ( Figure 1A). Since CD36 clears circulating oxLDL [29], we tested if CD36 deficiency would lead to accumulation of oxLDL in the sub-retinal region. CD36-deficient mice presented abundant sub-retinal oxLDL accumulation ( Figure 1B,C), despite regular diet; oxLDL was minimally detected in comparably raised wild type (WT, CD36+/+) congeners. CD36dependent cellular uptake of oxLDL was confirmed upon systemic administration of fluorescent tagged oxLDL; specific uptake of DiI-oxLDL was seen in the RPE of CD36+/+ mice ( Figure 1D), whereas no fluorescence was detected in RPE of CD36-/-mice in vivo ( Figure 1E). Likewise, exposure of RPE cells from CD36+/+ mice to DiI-oxLDL revealed internalization of the oxLDL, while no internalization was seen in RPE of CD36-/-mice ( Figure 1F,G).

CD36 stimulation prevents thickening of BM and preserves (in part) photoreceptor function
Additional experiments were conducted to determine if CD36 stimulation could prevent thickening of BM in HFHC-fed ApoE-/-mice. For this purpose animals were treated daily either with NaCl or the CD36 agonist EP80317 [36] from 8 to 18 weeks of age. HFHC diet significantly increased the thickness of BM in salinetreated ApoE-/-mice ( Figure 4B,D), as previously reported [24]. Whereas, treatment of ApoE-/-mice with EP80317, which reduces plasma cholesterol levels [36], prevented the accumulation of sub-RPE deposits and thickening of BM ( Figure 4C,D).
Along with BM thickening, ApoE-/-mice also present photoreceptor dysfunction [24,25], as evidenced with the significantly attenuated rod-cone a-wave amplitude of saline-treated ApoE-/-compared to WT mice ( Figure 5A, B). Similarly, the significantly attenuated rod and rod-cone b-waves (-2.7 log cd.sec.m-2 and 0.6 log cd.sec.m-2, respectively) are also suggestive of inner retinal dysfunction in saline-treated ApoE-/compared to WT mice ( Figure 5A, C); treatment of ApoE-/-mice with EP80317 significantly attenuated the loss of the scotopic rod and rod-cone responses ( Figure 5A-C).

DISCUSSION
The molecular mechanisms that lead to lipid accumulation in BM are not well understood. Since lipids including oxidized sterols are believed to be an important component of sub-retinal debris [42,43], and CD36 is a major scavenger receptor of oxidized lipids [29,30] including in RPE [15], we surmised that CD36 modulates sub-retinal deposit formation. We found that oxLDL uptake in RPE is CD36-dependent in vitro and in vivo. We also showed that in animals deficient in CD36 on regular diet, oxLDL accumulated sub-retinally and was associated with BM laminar deposits analogous to those described in ApoE-and LDL receptor-null mice [19,[23][24] fed on high fat high cholesterol diet. Moreover, stimulation of CD36 prevented accumulation of sub-retinal deposits and attenuated photoreceptor malfunction. Findings suggest that insufficiently prompt clearance of dietary lipids is associated with increased circulating levels of oxLDL and sub-retinal build-up of these lipids with other debris; the scavenger receptor CD36 contributes in maintaining integrity of BM.
Bruch's membrane (BM) is a thin layer of connective tissue located between RPE and choroid through which essential molecules for chorio-retinal homeostasis must be transported [44]. With age BM accumulate neutral lipids (triglyceride, esterified cholesterol, fatty acid) [7] and lipoprotein-like particles [6] comprised in laminar deposits, and lead to a decrease in permeability [45,46].
Several receptors are potentially implicated in lipid transport in retina [27,28]. Mice deficient in LDL receptor or its ApoE ligand, display an atherosclerosis phenotype including high circulating levels of oxLDL [21], along with BM thickening, BM lipid deposits and RPE basal deposits [19,[23][24][25]. Although debatable, an association between atherosclerosis and human AMD has also been described [54][55][56]. Our observations complement these findings, by showing that CD36 influences the clearance of oxLDL by RPE (Figures 1  and 3A), affecting (at least partly) sub-retinal structure; along these lines, CD36-null mice accumulated over time sub-retinal oxLDL with deposits (Figures 2 and 3D), while CD36 stimulation prevented the latter (Figure 4).
In summary, deficiency in uptake of oxidized LDL by CD36 expressed on the basolateral side of RPE [32] seems to contribute to age-related BM thickening; conversely, CD36 activation attenuates BM thickening and in parallel preserves visual function. In view of the marked paucity in effective therapeutic modalities for non-proliferative AMD, pharmacological modulation of CD36 activity may be a potential approach for this form of maculopathy.

METHODS
Animals. The ApoE-/-and ApoE/CD36 double deficient (ApoE-/-CD36-/-) mice, obtained as described previously [35], were housed at local animal facilities under 12 hours light-12 hours dark cycles and fed ad libitum with a normal (ND) or a high fat-high cholesterol diet (HFHC) (D12108, cholate, AIN-76A semipurified diet, Research Diets Inc., NewBrunswick, NJ). CD36-/-mice and their controls wild-type littermates (CD36+/+) were reproduced separately under ND unless otherwise indicated. Eight-week-old ApoE-/-mice were treated daily with EP803017 (300µg/Kg) [36] or vehicule (0.9% NaCl) by subcutaneous injections for a period of 10 weeks prior to sacrifice. All mice were sacrificed by carbon dioxide inhalation or by intraperitoneal injection of pentobarbital sodium overdose, prior to ocular enucleation. All experimental procedures were done in accordance with the Institional Animal Ethics Committee and the Canadian Council on Animal Care guidelines for use of experimental animals. Electroretinography. Dark adapted scotopic full-field electroretinograms (ERG) (intensities:-6.3 log cd.sec.m -2 through 0.6 log cd.sec.m -2 ) were obtained from 18week-old ApoE-/-mice on HFHC diet treated or not with EP803017 and WT control-age matched mice as we described in detail [37][38][39]. Amplitudes of ERG awave and b-wave components were measured according to a method previously described [37,38]. Briefly, the amplitude of the a-wave was measured from baseline to trough, and the b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave.
Statistical analysis. Data between groups (other than for ERG) were compared using non-parametric Mann Whitney U-test. For ERG data, 2-way repeated measures ANOVA (P < 0.05) with Bonferroni post-tests were used to compare WT mice to ApoE-/-mice, and to determine the effect of EP80317 treatment on the different parameters of the ERG as the repeated factor and treatment group as the independent factor. All analysis and graphic representations were performed with Prism software (version 4.0c; GraphPad Software); values are represented as mean ± standard error of the mean (SEM). P values were calculated for a confidence interval of 95%; hence P values of less than 0.05 were considered significant.