Abstract
Purpose :
Confocal laser scanning microscopy (CLSM) is commonly employed to image sections of retinal tissue and can be useful in comparing fluorescence from different retinal layers following immunohistochemical staining. If such quantification is to be meaningful, it is important to understand the relationship between measured fluorescence and image acquisition parameters. Here we sought to define the effect of gain on recorded signal in mouse retinas induced to express green fluorescent protein (GFP), and to correlate this with protein expression levels.
Methods :
Bilateral subretinal injections of an adeno-associated viral vector driving GFP expression from a strong, ubiquitous promoter (AAV2/2-CAG-GFP) were delivered to 7 week old C57BL/6J mice (n=12). Five weeks post-injection, 12 eyes were fixed and sectioned to a thickness of 18μm for imaging using a Zeiss Laser Scanning Microscope (LSM710) and ZEN 2009 software with standardised acquisition parameters. Serial images were acquired of the same sections using an Argon 488nm laser at multiple gain values (representing the high voltage setting of the photomultiplier). For every gain setting, mean grey values (GV) for each retinal layer were calculated using ImageJ software. For the other 12 eyes, GFP levels normalised to β-actin were quantified for the retina and remaining eye cup separately by Western blot analysis.
Results :
On CSLM, GFP was found to be confined to the retinal pigment epithelium (RPE) and photoreceptor layer (PRL), being more highly expressed in the former. The effect of gain on GV was profoundly non-linear, following an exponential relationship. After logarithmic transformation this could be described by a linear function for all retinal layers (R2= 0.9998, p=8x10-17 for RPE; R2= 0.9911, p=1.54x10-10 for PRL). Mean GV ratio RPE:PRL on confocal sections was 5.31±0.84 for a low gain value (550), 6.40±1.16 for moderate gain (600), and 6.70±1.35 for high gain (650). Mean ratio of GFP in eye cups compared with retinas by Western blot was 6.13±2.0.
Conclusions :
Quantification of fluorescence in CSLM images reflects expression levels of fluorescent proteins within ocular tissues as measured by Western blot analysis. Given the exponential relationship between acquisition gain and measured fluorescence, GV ratios between retinal layers within a single image are likely to be underestimated at lower levels of gain and overestimated at higher levels.
This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.