Specific sulphation modifications of heparan sulphate regulate distinct aspects of axon guidance in the developing mouse central nervous system.
MetadataShow full item record
Development of the visual system involves the precise orchestration of neural connections between the retina of the eye, the thalamus (dorsal lateral geniculate nucleus; dLGN) and the superior colliculus (SC). During early development, receptor molecules on the growth cones of retinal ganglion cell (RGC) axons sense molecular guidance cues in the extra cellular matrix (ECM) that define their route and branching behaviour within the visual system. Heparan sulphate proteoglycans (HSPGs) are ECM molecules composed of a core protein and a variable number of disaccharide residues that have been implicated in mediating axon guidance. HSPGs are modified by a number of enzymes that contribute to their structural diversity. Based on this structural diversity; the “heparan sulphate code” hypothesis of Bulow and Hobert (2004) postulated that different HSPG modifications confer different axon navigation responses as the growth cones traverse the local environment. To investigate the roles played by specific modifications of HSPG molecules in the guidance of axons, we examined two lines of mutant mice harbouring mutations in the genes encoding HSPG modifying enzymes, Heparan sulphate-6-O-sulphotransferase-1 (Hs6st1) and Heparan sulphate-2-O-sulphotransferase (Hs2st). These two mutant lines were generated through the use of gene trapping. Previous observations of RGC axon development in the two mutant lines revealed distinct axon guidance errors at the optic chiasm. Loss of Hs6st1 sulphation resulted in RGC axons navigating ectopically into the contralateral eye. Loss of Hs2st sulphation resulted in RGC axons navigating outside the normal boundary of the optic chiasm. Early observations suggested that both Hs2st sulphation and Hs6st1 sulphation have distinct, non-overlapping actions and thus, influence different axon guidance signalling pathways at the optic chiasm. Based on our findings and previous work describing the expression patterns and functions of the chemo-repellent axon guidance molecules, Slit1 and Slit2 at the optic chiasm and their Robo2 in the retina, we formulated the hypothesis of an HSPG sulphation code where Hs2st sulphation is specifically required for Slit1-Robo2 signalling and Hs6st1 sulphation is specifically required for Slit2-Robo2 signalling at the optic chiasm. To further our understanding of the roles Hs2st sulphation and Hs6st1 sulphation have on axon guidance, we looked at a number of key choice points that navigating axons encounter and are known to be influenced by Slit signalling. Further observations of RGC axons at the optic chiasm of Hs2st-/- mutants and Hs6st1-/- mutants showed distinct axon guidance phenotypes, both resulting in statistically significant increases in the width of the optic chiasm at the midline. While Hs6st1 sulphation had no effect on RGC axon navigation within the eye (possibly due to 6-O-sulphation compensation by Hs6st3); the loss of Hs6st1 sulphation at the dLGN resulted in a significant increase in the defasciculation of the optic tract. Observations of other axonal tracts influenced by Slit signalling revealed the importance of Hs2st and Hs6st1 sulphation in aiding callosal axons to successfully traverse the midline in corpus callosum development. Observations of the thalamocortical (TCA)/corticothalamic (CTA) tracts revealed that neither Hs2st sulphation nor Hs6st1 sulphation was required for the development of the mouse TCA tract (the latter may be explained by 6-O-sulphation compensation by Hs6st2). To test whether Hs2st and Hs6st1 enzymes have redundant functions in optic chiasm development, we attempted to create Hs2st-/-/Hs6st1-/- double mutants. A PCR genotyping strategy was developed for the identification of Hs6st1 animals and showed that Hs6st1-/- mutants had high postnatal lethality with only 3% of the offspring surviving to weaning while Hs2st-/-/Hs6st1-/- double mutants all died very early during embryonic development. Observations of Hs2st-/-/Hs6st1+/- mutants and Hs2st+/-/Hs6st1-/- mutants that lacked three of the four Hst alleles showed no differences when compared to single Hst knockouts. Finally, we showed that altered Slit expression at the optic chiasm and Robo expression in the retina could not explain the mutant phenotypes observed in Hs2st-/- mutants and Hs6st1-/- mutants, and therefore we hypothesized that Hs2st sulphation and Hs6st1 sulphation regulate distinct aspects of Slit-Robo signalling at the surface of the navigating axon growth cone.