Publication Date



Open access

Degree Type


Degree Name

Doctor of Philosophy (PHD)


Molecular Cell and Developmental Biology (Medicine)

Date of Defense


First Committee Member

Pedro J. I. Salas

Second Committee Member

W. Dalton Dietrich

Third Committee Member

Kathryn W. Tosney

Fourth Committee Member

Jeffrey L. Goldberg

Fifth Committee Member

Richard H. Masland


How do neighboring neurons differ? How are they similar? How do they relate to each other? Retinal ganglion cells (RGCs) and amacrine cells are born during the same developmental window, from the same population of progenitor cells, and some amacrine cells even migrate to the same retinal layer where RGCs reside, the ganglion cell layer. Amacrine cells are presynaptic to RGCs, whose axons subsequently project through the optic nerve and carry visual information to the brain. Although the cell biology of RGCs has been studied in some detail, relatively little is known about the cell and molecular biology of amacrine cells. What distinguishes these neighboring cell types and in what ways are they similar? Here I present a series of studies examining amacrine cells and RGCs, with a view towards better understanding the development and cell biology of these neurons. The retina has been long used as a model system to study central nervous system (CNS) development and regeneration: RGCs fail to regrow their axons after injury, and RGC survival is compromised in optic neuropathies such as glaucoma. In contrast, amacrine cells survive even after the loss of their targets, RGCs, in glaucoma and other optic neuropathies. While the signaling for RGC survival in vitro has been widely studied, little is known about the molecular mechanisms that may underlie amacrine cellsâ?? resistance to neurodegeneration. Taking advantage of our unique method to highly purify amacrine cells away from other retinal neighbors and glial cells, I found that amacrine cells can survive at very low densities in culture and that they do not require addition of exogenous trophic factors, unlike RGCs. Interestingly, blocking of MEK1/2 or PI3K signaling pathways significantly impaired survival, suggesting that these intracellular signaling pathways are necessary for amacrine cell survival. Thus, while amacrine cell and RGC survival seem to be regulated through similar signaling pathways, these two cell types have different requirements for exogenous peptide trophic factors Because amacrine cells were able to survive in our low density cultures in serum and peptide trophic-free media, it is possible that amacrine cell survival is regulated by autocrine signaling, or by hormones and/or antioxidants. Thus these retinal interneurons may not depend on target RGCs for peptide trophic support. Amacrine cells are a heterogeneous group of interneurons that modulate retinal signaling of visual information onto RGCs. There are more than 30 subtypes described in the mammalian retina, characterized by myriad morphologies and the secretion of different neurotransmitters. Despite their apparent inability to differentiate axons and dendrites, purified amacrine cells in vitro extended neurites with varied lengths and morphologies, raising the hypothesis that the regulation of these processes has an intrinsic component. Specifically, I asked whether purified amacrine cell subpopulations would extend neurites similarly in vivo and in vitro. Surprisingly, three purified amacrine cell subpopulations recapitulated aspects of their in vivo morphology in vitro, consistent with the existence of intrinsic mechanisms of neurite growth and patterning in the developing retina. Thus, I have demonstrated that there is an intrinsic regulatory component that contributes to the varied morphology of amacrine cell neurites found in vivo. To further characterize differences between amacrine cells and RGCs, I generated a database of amacrine cell gene expression during development and compared it to the transcriptome of RGCs at the same developmental ages. I found ~75% similarity among the genes expressed in RGCs and amacrine cells during development. However, I focused my interest in genes that were differentially regulated because they might underlie amacrine cellsâ?? resistance to neurodegeneration and could help understand the differences in polarity between amacrine cells and RGCs. Comparing the gene expression profiles of these two cell types, I found that RGCs expressed higher levels of the pro-apoptotic molecules Bax and Bad. This raises the interesting hypothesis that amacrine cells may be more resistant to degeneration than RGCs because they do not express as many pro-apoptotic molecules as RGCs do. In addition, I generated a list of polarity-associated candidate genes that are differentially expressed in amacrine cells and RGCs. Together, these data could be combined for therapeutical purposes. Switching dying RGCs to an amacrine cell-like state may help preserve these cells in neurodegenerative diseases like glaucoma. Conversely, regulating polarity genes in amacrine cells might induce changes in their neurite outgrowth ability that could help understand the mechanisms of cell polarization and axon growth, two critical components to achieve CNS regeneration. How might these presynaptic amacrine cells influence their neighboring RGCsâ?? cell biological phenotypes? Previous findings in the laboratory demonstrate that purified RGCs undergo an irreversible loss of their intrinsic axon growth ability during development, and that the process can be signaled by amacrine cells. Thus, amacrine cells are sufficient to signal RGCs to decrease their intrinsic axon growth ability during development. It is not known, however, whether amacrine cells are necessary for this process. I hypothesized that in the absence of amacrine cells, RGCsâ?? axon growth might be dysregulated in vivo. The creation of the Foxn4-/- mouse by the Xiang laboratory allowed me to address this question. Foxn4 is a forkhead transcription factor that is required for amacrine cell genesis during retinal development, and as a result the Foxn4 knockout mice have fewer amacrine cells. I found that in the context of a reduced number of amacrine cells, RGCs projected fewer dendrites to the inner plexiform layer in the retina. In addition, RGCsâ?? axon projection to their target (the superior colliculus) was developmentally delayed, and they failed to penetrate into the retinorecipient layers of the superior colliculus. Finally, Foxn4-/- mice showed disrupted optic nerve architecture, albeit the fluorescence intensity of labeled RGC axons in optic nerve cross- sections was similar among animals. Taken together, these data demonstrate a role for a pre-synaptic partner, amacrine cells, in regulating a neuronâ??s intrinsic axon growth ability and intrinsic patterning. In conclusion, together these data paint a portrait of how two neighboring retinal cell types differ: amacrine cells are resistant to neurodegeneration whereas RGCs express genes that are associated with apoptosis and glaucoma; amacrine cells do not require the presence of exogenous trophic factors to survive in vitro whereas RGCs do require trophic support. Finally, while amacrine cells and RGCs may differ in their mechanisms for establishing axon and dendrite specification, they both exhibit an intrinsic capacity to grow neurites in vitro that recapitulates their phenotype in vivo. These differences in developmental cell biology may point to new approaches to understanding retinal and retino-collicular patterning, neuronal survival and perhaps even optic nerve regeneration.


Retinal Ganglion Cells; Amacrine Cells; Retinal Development