Topical application by intravitreal injection is sufficient for G-CSF neuroprotection on retinal ganglion cells. G-CSF was given either systemically by daily subcutaneous injection (20 or 40 μg/kg bodyweight per injection), or by direct application to RGCs via intravitreal injection (500 ng in 2 μl per injection on day 0, 4, 7 and 10 after axotomy; first injection 2 h after axotomy). Leukocyte counts were determined one day before drug administration and after 14 days of treatment (n = 8 per group). Subcutaneous application resulted in the expected leukocytosis, while we could confirm that the intraocular G-CSF injection did not influence leukocyte counts (A). RGC survival was then determined accordingly after subcutaneous or intraocular G-CSF application (B). Both protocols lead to a comparable protective effect on axotomized RGCs in vivo. * p < 0.05 when compared to respective vehicle control; n.s. = not significant; i.o.: intraocular (intravitreal) injection.