The Transcription Factor Math5 Is Not Required For Specification Of A Subset Of Retinal Ganglion Cells

Abstract

Purpose: Retinal ganglion cells (RGCs) are the sole conduits of light information to the brain for both image and non-image forming functions. During embryonic development, RGCs arise from retinal progenitors via expression of transcription factors at discrete developmental timepoints. One such transcription factor, Math5 (Atoh7), is a basic helix loop helix transcription factor that was thought to be required for RGC specification. In support of this, Math5 mutant mice lack an optic nerve and chiasm, and have greater than 80% reduction in RGC number. Interestingly, we have previously reported that the melanopsin-expressing, intrinsically photosensitive (ip)RGCs continue to be generated in the retina after Math5 is downregulated (through embryonic day 18.5). This led us to hypothesize that a subset of ipRGCs might arise independently of Math5.

Methods: We utilized Math5^Cre and Bax^-/- mouse lines as well as Cre-dependent reporter lines to analyze RGC development and lineage.

Results: Lineage tracing indicated that, surprisingly, only half of melanopsin immunopositive cells express Math5 during development. However, Math5^-/- animals lack 90% of ipRGCs. To resolve this apparent inconsistency, we sought to determine whether Math5 negative ipRGCs die secondarily in Math5 mutant animals. To test this, we generated double mutant animals that lack Math5 as well as the proapoptotic factor Bax (Bax^-/-). In double mutant animals (Math5^-/-; Bax^-/-), we observe a 6-fold greater proportion of ipRGCs (~60%) remaining relative to Math5-/- animals. We next examined whether conventional RGCs that express the transcription factor Brn3a (expressed in 70% of RGCs) also die secondarily in Math5^-/- animals. We found a substantial rescue of Brn3a positive RGCs in Math5^-/-; Bax^-/- retinas. Neurobiotin labeling confirmed these surviving Brn3a positive cells project axons to the optic disc.

Conclusions: These results indicate that a subset of both ipRGCs and conventional RGCs do not require Math5 for specification. The greater loss of ipRGCs and RGCs in Math5^-/- animals compared to Math5^-/-; Bax^-/- animals, suggests that Math5 negative RGC survival is dependent on Math5 positive RGCs. This secondary cell death could be due to lack of trophic support since there is a severely retarded optic nerve in Math5 mutant animals.

Keywords: 698 retinal development • 629 optic nerve • 531 ganglion cells