Collected subretinal fluid was cytocentrifuged (GP centrifuge; Beckman, Fullerton, CA) onto a glass slide coated with poly-l-lysine (P4832; Sigma-Aldrich, St. Louis, MO) at the bottom of a 24-well multiwell plate (BD Falcon, Bedford, MA) at 1500 rpm for 10 minutes. Supernatant was removed. Subsequently, the cells were incubated at 37°C in RPMI 1640 medium containing 10% FBS. After incubation for 3 hours, the cells were washed three times with PBS and fixed in acetone/methanol (2:3) for 10 minutes at −20°C. After blocking nonspecific protein in 0.1% BSA in PBS for 30 minutes at room temperature, primary antibody CD68 (ab845; Abcam, Cambridge, UK) was applied for 3 hours at room temperature, and secondary antibody (Alexa Fluor 647; Invitrogen, Carlsbad, CA) was applied for 30 minutes at room temperature. The negative control was without secondary antibody. Immunofluorescence-labeled and differential interference images were obtained (LSM510 META; Carl Zeiss Meditec, Dublin, CA). The emission fingerprint of fluorescence was also recorded using the lambda META scanning mode, in which fluorescence in a 10-nm width is recorded by a polychromatic 8-channel detector that allows fast acquisition of lambda stacks. Fluorescence was excited by an Ar+ laser for 458, 488, and 514 nm; an HeNe laser for 532 nm excitation; and an HeNe laser for 633-nm excitation through an appropriate dichroic mirror. To compare the emission fingerprint, the detector gain and excitation light intensity were set constant for a set of experiments. The fluorescence intensity profile was exported as text and used to analyze the peak, by using the local maximum method in the peak-finding module of the software (Origin 8; OriginLab Corp., Northampton, MA).