Working with GFP in the brain.

A powerful technique currently used widely to track cells in the central nervous system is that of enhanced GFP as a fluorescent marker. Cells can be infected with a GFP-expressing retrovirus (1,2) or taken from GFP transgenic animals and transplanted into a syngeneic host (3). However, this approach may be confounded by considerable autofluorescence within the mammalian central nervous system. It is difficult to distinguish cells labeled with GFP from those that are naturally fluorescent. In the past, the problem with autofluorescence has been so severe that researchers have used nonfluorescent immunohistochemistry with an anti-GFP antibody to track the marked cells (4). This approach solves the problems associated with background fluorescence due to lipofuscin but negates the enormous strength of endogenous fluorescence of GFP to track cell migration and localization. In addition, the potential to co-localize GFP with additional fluorescent markers is no longer possible with such strategies. Our goal here is to discuss a strategy for recognizing when green fluorescence in the central nervous system is due to GFP and not the result of a naturally occurring fluorescent molecule. The cause for most of the naturally occurring fluorescence in the central nervous system is lipofuscin. Lipofuscin is a complex of peroxidized lipids, proteins, and transition metals (5) and is the result of the incomplete breakdown of macromolecules within the lysosomal system. Undegraded material undergoes peroxidation and polymerization to form a heterogeneous complex. Oxidative enzymes catalyze the peroxidation with an increased deposition in cells that are undergoing higher levels of oxidative metabolism. Lipofuscin makes fluorescence microscopy in the central nervous system difficult because it has a very broad excitation and emission spectrum. We have recorded lipofuscin fluorescing in 4′,6-diamidino-2-phenylindole (DAPI), fluorescein isothiocyanate (FITC), rhodamine, and far red channels. This problem is compounded by the fact that lipofuscin also has a very widespread distribution. One study that looked at the topographical distribution of lipofuscin in lemurs found the pigment in the brain stem, neocortex, cerebellum, hypothalamus, basal forebrain, hippocampus, and olfactory bulb (5). In our study with 20-week-old mice, we have found that the mitral cells of the olfactory bulb have the highest depostion of lipofuscin granules, an observation consistent with the high level of oxidative metabolism of these cells (5). Wide-field epifluorescent microscopy offers numerous advantages for visualization of GFP-labeled cells. Greater numbers of tissue sections can be observed in less time and at less cost than with a confocal microscope. A caveat, however, is that GFP and lipofuscin fluorescence is particularly difficult to distinguish when using a widefield microscope. Green fluorescent cells observed using a FITC filter set (excitation 490/20 nm, emission 528/38 nm) are frequently a mixture of lipofuscin-containing cells and GFP-containing cells (Figure 1A). One way to differentiate GFP-containing cells from lipofuscin-containing cells is to use higher magnifications when it is someBenchmarks