RESEARCH

While the formation of lipid droplets (LD - image at right) is well studied in other cell types, very little is known about how LDs form and release lipid in the RPE. We are studying these mechanisms, with the future goal of manipulating LDs in a way that affects how much of the RPE's daily lipid load gets secreted (via drusen-forming lipoprotein particles (LP)) versus mitochondrially metabolized (M) (see image at right). We will manipulate the general class of enzymes necessary for LD formation (ACATs and DGATs - see image at right) as well as the enzymes and processes responsible for releasing lipid from LDs (cytosolic lipases and/or lipophagy - see image at right) to study their effects on extracellular secretion versus mitochondrial metabolism of the RPE daily lipid load. Our goal is to manipulate LD formation and lipid release from LD in a manner that promotes mitochondrial metabolism rather than secretion of pathologic drusen-components. 

Autophagy, a universally conserved cellular pathway for degradation of long-lived proteins and most organelles, is dysregulated in AMD. Numerous lines of evidence suggest that enhancing autophagy could alleviate pathology in AMD. In a specialized autophagic process called lipophagy, lipid from lipid droplets (LD at right) can be shuttled via an autophagosome (AP) to the lysosome (Ly), and ultimately imported to the mitochondria for metabolism. Thus, autophagy induction has the potential to shunt intracellular lipid away from secretion (and therefore drusen promotion) towards beneficial mitochondrial metabolism.  However, we've shown that the major mechanism for inducing autophagy, inhibition of the master metabolic regulator mTOR, causes significant disruption to normal RPE function. We have therefore screened and identified numerous small molecules that induce autophagy in the RPE without direct mTOR inhibition. Several of these compounds have the beneficial effect of promoting lipid degradation over lipid secretion in the RPE. We are now testing these compounds in animals for dosing, safety, and toxicity. Our goal is to promote autophagy independent of mTOR inhibition as a mechanism for decreasing RPE secretion of drusen components.

The RPE is a prolific secretor of lipoproteins (blue circles to right), which are generated in the endoplasmic reticulum (ER - see image at right). The major apolipoprotein secreted by the RPE is apolipoprotein E (apoE). ApoE is the most abundant apolipoprotein in drusen deposits below the RPE and it is the only apolipoprotein in reticular pseudodrusen (RPD) deposits above the RPE (see image to right). Thus, whether the RPE secretes apoE above (apical) or below (basolateral) the monolayer likely determines how much drusen versus RPD forms in AMD patients. Interestingly, clinical evidence suggests that patients with RPD have different clinical outcomes than those with just drusen. Thus, the polarity of RPE apoE secretion can be directly tied to patient prognosis in AMD. We have found that several AMD-relevant insults alter which side the RPE secretes apoE to, and we are exploring the mechanisms governing apoE polarized secretion in the RPE. Our goal is to link conditions in the clinic that favor RPD formation with conditions in our cell culture system that favor apical secretion of apoE, since apical secretion should lead to RPD formation. Once we understand conditions contributing to apical apoE secretion, we will probe the mechanism behind this polarized secretion.

RPE mitochondrial dysfunction is known to be a prominent feature of AMD. When this happens, the RPE's physiologic lipid load is theoretically unable to be efficiently degraded. That leaves more lipid to be packaged into lipoprotein particles that are secreted and deposit as pathologic drusen and reticular pseudodrusen (RPD). However, another organelle in the RPE, the peroxisome, is known to degrade particular types of fats. If the peroxisome remains functional even when mitochondria are dysfunctional, can the peroxisome "take over" and degrade the RPE's lipid load? If so, that would decrease secretion of drusen- and RPD-promoting lipoproteins. In addition, the ability to degrade fats would decrease how much glucose the RPE consumes, allowing the glucose to be delivered to photoreceptors where is serves as a critical metabolic fuel. Our goal is to understand how much the peroxisome compensates for mitochondrial dysfunction when it comes to RPE lipid handling. If it does compensate well, our goal is to increase peroxisomal capacity as a therapeutic avenue for dry AMD.

During routine culture of primary human RPE from young donors and old donors, we discovered that old donors are more susceptible to several but not all AMD-relevant stresses. For example, when RPE cultures from young donors are fed  oxidized photoreceptor tips to mimic physiologic phagocytosis (left panel), the tips are efficiently degraded. In contrast, when RPE cultures from old donors are fed the same oxidized photoreceptor outer segment tips, they are poorly degraded, accumulating in intracellular autofluorescent granules (green picture to right). These granules are similar to lipofuscin, an aggregated mass of lipids, proteins, and retinoids that accumulates intracellularly with age in all RPE cells. Intrigued by the dramatic differences in lipofuscin-like particle accumulation between young and old donor cultures (graph at right), we have been cataloguing other phenotypic differences between these cultures. At the same time, we are understanding the transcriptional (and in the future metabolomic) differences between these cultures, and are attempting to tie gene expression or metabolic differences to phenotypic differences we see between the cultures. More generally, we envision that cultures derived from young donors may model RPE resiliency in the face of AMD stressors while cultures derived from older donors may model RPE susceptibility to AMD stressors. Our goal is to use primary human RPE cultures from young versus old donors to model RPE susceptibility in dry AMD in vitro, and decipher mechanisms behind this susceptibility.

With the advent of optical coherence tomography (OCT), we can assess the structure of the retina in AMD patients with 15 micron resolution (!!!). This has enabled a plethora of clinical observations. For example, drusen, which deposit underneath the RPE, often disappear right before there is loss of RPE in the late stages of dry AMD (termed geographic atrophy - GA). At the same time, subretinal drusenoid deposits (SDD - see image at right - also known as reticular pseudodrusen or RPD), which are located above the RPE, can often remain despite the onset of GA. To develop a basic-science explanation for these clinical observations using primary human RPE cultures, we first identified apolipoprotein E (apoE) as the major apolipoprotein in both drusen and SDD (apoE LP to right indicates apolipoprotein-decorated lipoprotein particles). Thus, if the RPE secretes apoE primarily above the monolayer (apically), SDD will form. And if it secretes apoE primarily below the monolayer (basolaterally), drusen will form. When we exposed RPE cultures to sub-lethal oxidative stress, we discovered that our cultures decrease their secretion of apoE basolaterally, but not apically. Thus, unhealthy RPE, as seen in advanced dry AMD, may alter the polarity of apoE secretion to support SDD formation, but not drusen formation. A model of apoE secretion polarity through the stages of dry AMD is depicted to the right. Our goal is to continue translating clinical observations in AMD patients into basic science explanations using our RPE culture system, and we have several projects in this area.