Clinical trials for age-related macular degeneration (AMD) depend on imaging outcomes to measure success. Strong autofluorescence (AF) signal from the retinal pigment epithelium (RPE) has high potential for noninvasive, spatially and molecularly precise early detection, and longitudinal follow-up. Subcellular signal sources of RPE AF are lipofuscin (LF) granules, many also containing melanin (melanolipofuscin, ML). New concepts of human RPE cell biology make fulfilling the potential of fundus AF more possible than ever. Each adult human perifoveal RPE cell has >1400 organelles that generate reflectivity for optical coherence tomography (OCT). Half are LF or ML that also generate fundus AF (FAF) signal to blue-green exciting light. Melanosomes (M) preferentially localize to apical processes and may contribute their own FAF signal. The congruent topographies of FAF and photoreceptors has been confirmed and detailed, with low signal in the fovea and strong signal at a ring of high rod density encircling fovea and optic nerve head. At this ring is also an abundance of well-studied bisretinoid fluorophore A2E. By OCT, hyperreflective foci conferring high progression risk include RPE anteriorly migrating into the retina. A far-reaching new finding from the previous project period is that foveal RPE is dominated by ML, thus imparting a specific molecular signature to RPE at this cone-rich site. Consistent with regional molecular differences in RPE, our imaging mass spectrometry (IMS) studies revealed lipid signals specific to RPE in central macula. Unresolved questions are whether multiple fluorophores localize to each organelle, what organelle emits FAF signal with near-infrared (NIR) excitation, and what molecules are responsible for macular AF. We hypothesize that major fluorophores driving macular FAF signal are bisretinoids localized to specific organelle subtypes, which in turn are regionally distributed in accordance with the distribution of cones and rods. Drs. Curcio, Ach, and Schey, members of the multidisciplinary Hyperspectral Retinal Autofluorescence Team (HYRAFT) with expertise in AMD pathology, clinical imaging and microscopy, and analytic chemistry, respectively, propose studies in human donor eyes to address these knowledge gaps. In RPE-choroid flat mounts and tissue cross-sections of aged normal (N=20) and AMD (N=25) eyes subject to ex vivo OCT, Aim 1 will fortify a regional and morphologic basis of FAF imaging using tissue cross-sections to use imaging mass spectrometry to identify lipid signals including fluorophores. In these tissues Aim 2 will enumerate and determine emission spectra of LF, ML, and F in 3- dimensions using high-resolution structured illumination microscopy and a NIR-sensitive camera. Aim 3 will use discontinuous sucrose gradients to isolate RPE organelles in pooled maculas and peripheries (80 normal eyes, 40 donors) to extract fluorophores, separate them with fluorescent thin layer chromatography, and identify fluorophores using liquid chromatography – tandem mass spectrometry. Results will directly translate to molecularly informed clinical FAF imaging by many technologies, in AMD and other retinal disorders.

NARRATIVE Strong autofluorescence signal from the retinal pigment epithelium (RPE) has high potential for noninvasive, spatially and molecularly precise early detection, and longitudinal follow-up for AMD clinical trials. Using human donor eyes, we will use high-resolution microscopy and imaging mass spectrometry to localize RPE organelles emitting autofluorescence signal in regions rich in cone or rod photoreceptors, as well as areas in high-risk AMD pathology. With analytic chemistry we will identify the responsible molecules.