The role of the Re-1 silencing transcription factor (Rest) in the injury response of Müller glia in the murine retina

Purpose: In contrast to their robust regenerative abilities in fish, Müller glia have limited ability to regenerate retinal neurons in the mammalian retina. Based on its function as a developmental switch for neurogenesis vs. gliogenesis, we hypothesize that continued expression of REST in Müller glia may block their neurogenic potential of in the mammalian retina. To test this, Rest was conditionally knocked out in Müller glia and changes in cell proliferation after retinal injury were measured. Methods: Triple-transgenic mice carrying homozygous or heterozygous Rest flox, a Müller glia-specific, tamoxifen-inducible Rlbp1-CreERT2 and a Rosa26-tdTomato Cre-reporter or Cre-transgene negative controls received daily intraperitoneal (IP) injections of 4-hydrozytamoxifen (4OHT) from P21 to P25 to induce Cre-expression. Retinal injury was induced with an intraocular injection of N-methyl-D-aspartic acid (NMDA) (0.1M) with contralateral, control eyes injected with PBS. After 2 days, binocular, intravitreal injections of epidermal growth factor (EGF; 1 μg/μl) and 5-ethynyl-2’-deoxyuridine (EdU; 1 μg/μl) were administered to stimulate regenerative response and label S-phase cells respectively. Sections were immunolabeled with anti-glial fibrillary acidic protein (GFAP) or processed for Click-IT detection of EdU. EdU+ cells were counted in frozen retinal sections and scored for co-expression of tdTomato reporter, analyzed using ANOVA and T-tests with Tukey or Bonferroni corrections for multiple comparisons. Results: NMDA followed by EGF increased GFAP expression in Müller glia and increased the numbers of EdU+ cells in the injured retina vs. contralateral controls (p=0.006 and 0.016). In Rlbp1-CreERT2;tdTomato;Rest flox/flox mice, 4OHT did not increase numbers of EdU+ cells in the retinas compared to controls that lacked the Rlbp1-CreERT2 transgene. Unexpectedly, Rlbp1-CreERT2;tdTomato;Rest flox/+ mice had fewer EdU+ cells (p<0.05) in NMDA injured retinas than control or experimental mice homozygous for Rest flox/flox. In all genotypes, very few Edu+ cells were double labeled with tdTomato. Conclusions: Despite induction of a gliotic response, the lack of EdU incorporation in tdTomato-labeled, Rest-knockout cells precluded assessment of direct effects of Rest knockout on injury response. Decreased numbers of EdU+ cells in injured retinas of Rlbp1-CreERT2; tdTomato mice heterozygous, but not homozygous, for Rest flox may reflect the small sample size or a non-cell autonomous effect.