Willshaw, David

Price, David

Down, Matthew Paul

Engineering and Physical Sciences Research Council (EPSRC)

2012-03-27T13:40:04Z

2012-03-27T13:40:04Z

2012-06-25

Healthy pre-natal development of the mammalian visual system requires that retinal ganglion cell (RGC) axons navigate a precise path to their targets in the thalamus and superior colliculus by making a precise series of turns determined by the complex interactions between growth cone and extracellular environment. One important choice point for RGC axons is the crossing of the midline at the optic chiasm, where ipsilateral/ contralateral sorting takes place. In this thesis a novel image analysis method using steerable filters for quantifying the gross orientation and turning of axons from a static image (such as from DiI filled axons) is presented. This method was applied to understanding Slit dependent axon guidance at the mouse optic chiasm. It was possible to quantify the differences at the chiasm between the wildtype and various classes of mutants involving heterzygous or homozygous knockout of the Slit1 and the Slit2 genes. Assessment was in terms of the spatial distributions in axon density and axon orientation as derived from DiI labeled RGCs originating from one eye. The animals were assessed at embryonic day 13.5. To my knowledge this is the first quantification of its kind in the field of axon guidance. It was found that there were strong statistical differences from wildtype in both the Slit1-/-;Slit2-/- and Slit1+/+;Slit2-/- knockouts in terms of both axon density and axon orientation across large extents of the chiasm. In both these knockouts it was found that the changes in axon density were localised to the anterior region of the chiasm, but the changes in axon orientation were spread across almost the entire extent of the chiasm. No other combination of the Slit1 and Slit2 knockouts for which embryos could be generated showed significant differences from wildtype in terms of spatial changes in axon density or axon orientation. No embryos were generated for the Slit1+/-;Slit2-/- combination. No changes in the spatial distribution of axon density or axon orientation were found between the Slit1-/-;Slit2-/- and Slit1+/+;Slit2-/- knockouts, suggesting that in terms of these two quantities, the two phenotypes are indistinguishable. This evidence suggests that the role of Slit2 is more important than the role of Slit1 at the optic chiasm in terms iii of axon guidance. In addition, the gradients of mRNA expression of Slit1 and Slit2 were quantified using in situ hybridisation, and these data were used to compare the mRNA gradients with the orientation and turning of axons in both the wildtype and Slit1/Slit2 knockout chiasms. Although this provided a powerful visualisation tool, no simple mathematical relationship was found between the mRNA gradient of Slit1 or Slit2 and the orientation or turning of axons at the optic chiasm. These approaches now provide an important suite of methods for spatial analysis of axon tracts and molecular gradients in axon guidance.

http://hdl.handle.net/1842/5840

en

The University of Edinburgh

computational

axon guidance

slit

Universal quantitative method for studying axon guidance and its application to Slit-dependent axon guidance at the developing mouse optic chiasm

Thesis or Dissertation

Doctoral

PhD Doctor of Philosophy