The formation of glomerular structure in the OB is a highly regulated process. Here, we report the expression of Shh in the glomeruli during early postnatal stages. Shh is able to regulate olfactory sensory axon terminal branching. When Shh function is blocked during early postnatal stages, olfactory axons were seen accumulating outside the glomerular structure, suggesting that Shh is required for promoting olfactory axon entry into this region.
We showed in this report that shh mRNA is expressed by mitral and tufted cells in the OB. Shh protein, most likely transported through mitral/tufted cell dendrites, was detected in the early appearing protoglomeruli during development and persisted in mature glomeruli into adulthood. However, contributions from the other cell types cannot be completely ruled out because we also detected low levels of shh transcripts in the astrocytes or juxtaglomerular neurons in the glomerular layer. Cultured OSNs, after exposing to Shh protein, exhibit more axonal branches than the controls. These data, taken together, suggest that Shh is involved in growth and branching of olfactory axon terminals and their entry into OB glomeruli.
Shh Is an Important Signaling Protein in Neural Development
Hedgehog function is not restricted to nearby targets by the cholesterol modification tying it to cell membranes (Porter et al.,1996), but is also involved in long-range target activation. For example, the observation of long cytoplasmic processes (cytonemes) produced by target cells makes it plausible for Hh-N to act by means of direct cell–cell interaction even when the cell bodies are separated by substantial distances (Ramirez-Weber and Kornberg,1999). In addition, the discovery of Drosophila tout-velu suggests that transcellular trafficking of cholesterol-modified Hh-N may occur and probably involves proteoglycans (The et al.,1999). Hedgehog can also be transported through axons to signal over distances. In the Drosophila visual system, photoreceptors in the eye imaginal disc transmit Hh through their axons to reach the target ganglion, the lamina. Hedgehog secreted by photoreceptor axons triggers neurogenesis to control the number of target neurons (Huang and Kunes,1996,1998). Our observations in the olfactory system suggest that Shh protein is transported by means of mitral and tufted cell dendrites to the glomerulus where it interacts with olfactory axon terminals. Further investigation is needed to confirm this hypothesis.
Shh plays multiple roles during neural development (Marti and Bovolenta,2002). Shh is known to influence the behavior of chick retinal ganglion cell growth cones. Ectopic expression of shh along the visual pathway alters the trajectory of the retinal ganglion cell axons. In addition, providing Shh-N coated beads suppresses neurite outgrowth from retinal explants. It appears that the Shh-mediated growth cone arrest is dependent on the intracellular level of cyclic AMP (cAMP). The presence of Shh-N in vitro induces a decrease in the level of cAMP inside the growth cone. In this study we have shown that the Shh seems to promote the growth and the branching of the olfactory sensory axons. Whether or not this phenomenon is also mediated through changes in cAMP levels in the growth cone remains to be tested. Changes in cAMP or cGMP signaling pathways can elicit opposite effects in growth cone behavior (Song et al.,1998). It is conceivable that the olfactory sensory axons react to the Shh signal in a manner opposite that of the retinal ganglion cells.
Shh Function in Glomerular Formation
Glomerular formation has been studied in both invertebrates and vertebrates, but some of the molecular signals involved in this process are not clear. Most investigators agree that glomerular formation is initiated by an interaction between OSN axons and the OB (Oland et al.,1988; Valverde et al.,1992; Oland and Tolbert,1998; Bailey et al.,1999; Treloar et al.,1999). Morphological changes during glomerular initiation have been investigated in detail in rats. From the reports of both Treloar et al. (1999) and Bailey et al. (1999), OSN axons first coalesce into protoglomeruli at E19. The presence of glial cells is required in early glomerular formation (Oland et al.,1988). The protoglomeruli develop into mature glomeruli in early postnatal life in rodents. The formation of glomeruli may follow a rostrocaudal gradient within the OB. In our experiments, the temporal and spatial expression of Shh correlates with this possible developmental sequences of glomerular formation. Multiple cell types are involved in the formation of glomeruli during postnatal development (Bailey et al.,1999; Treloar et al.,1999; Blanchart et al.,2006). Although Shh-N peptide regulated OSN axon growth behavior in vitro, blocking Shh function postnatally for 5 days did not appear to alter the size of the glomeruli in the olfactory bulb. During postnatal development, in addition to the ingrowth of OSN axons, several developmental events are progressing simultaneously. Developmental events contributing to the formation of glomeruli include olfactory axon terminal arborization within the glomeruli, ingrowth of immature olfactory sensory axons, maturation of innervated olfactory axons (from GAP43-positive to OMP-positive), ingrowth of periglomerular neuron dendrites, and elaboration of mitral cell dendritic tufts within the glomeruli (Halász and Greer,1993; Malun and Brunjes,1996; Klenoff and Greer,1998; Lee et al.,2008). Therefore, short-term decreased GAP43-positive axon ingrowth may not necessarily alter the sizes of the existing glomeruli. It is reasonable to speculate that, if Shh function is blocked for a longer period of time, glomerular size may be altered.
Mitral and tufted cells are not considered essential for the initiation of glomerular formation. In moths, no obvious disruption was observed in the initiation of glomerular formation when mitral-like cells were surgically removed during development. (Oland and Tolbert,1998). This appears to be true in mammals as well. In Tbr-1 knockout mice, the mitral and tufted cells fail to develop in the OB. Even in the absence of mitral and tufted cells, olfactory axons expressing the same odorant receptor are able to converge and enter the bulb at their “appropriate” positions (Bulfone et al.,1998). Although these studies provide evidence that mitral and tufted cells are dispensable for the initiation of OSN axon convergence and protoglomerular formation, they do not rule out the possibility that the mitral/tufted cells plays a role to assist the target recognition and initiation of synapse formation.
Mitral and tufted cells express semaphorin 3A (Sema 3A), which has been shown to have a chemorepellent function (i.e., it inhibits axonal extension and induces growth cone collapse; Giger et al.,1998). The receptor of Sema 3A, neuropilin-1, is expressed by primary olfactory neurons (Kawamaki et al.,1996). Kobayashi et al. (1997) have done in vitro experiments to demonstrate that chick OSN axons collapse when Sema 3A is present in the culture. These findings strongly support the notion that the mitral/tufted cells participate in influencing the growth behavior of OSN axons in the OB.
This study provides evidence that Shh is part of the molecular complex that regulates OSN axon growth into the glomerulus. When Shh function is blocked between P0 and P5, we observed severely reduced ingrowth by immature olfactory axons into glomeruli. Between P0 and P5, olfactory axons from different maturation stages normally can be found in the glomeruli. Some will have entered the glomeruli already and will be essentially mature (i.e., they had entered the glomerulus before P0 and had stopped growing); these would be OMP-positive and GAP43-negative. Axons from more recently generated neurons will have just entered the glomerulus and would still be GAP43-positive and OMP-negative. And some, from even more recently generated neurons, will still be growing toward their target glomerulus and would be GAP43-positive and OMP-negative. This finding that inhibiting Shh function results in an accumulation of immature OSN axons in the nerve layer and decreases the number of OSN terminal branches in the glomeruli, combined with our knowledge in the expression of growth inhibitory molecules, together argue that Shh participates in gating axon entry into glomeruli either directly, as an attractant, or indirectly, by disinhibiting the repellent.
In addition to Sema 3A and Shh, several other molecules may participate in regulation of growth and synapse formation in the OB. For example, some may act to prevent OSN axon growth beyond the glomerular layer (Gonzalez and Silver,1994; Kafitz and Greer,1998). These growth regulating signals could be produced by mitral/tufted cells, juxtaglomerular neurons, and/or glial cells. Axonal entry into glomeruli and branching may involve complex integration of both positive and negative signals. Our studies suggest that Shh may serve as one of the signals in this process.
A characteristic behavior of OSN axons when they enter a glomerulus is to form branches. Target-derived factors play important roles in axon branching and terminal arborization in multiple systems (Diamond et al.,1992; Cohen-Cory and Fraser,1995; Wang et al.,1999). Our in vitro evidence supports the notion that Shh is one of the signals involved in OSN axon branching. When OSNs were treated with Shh-N in vitro, we observed a significant increase in the numbers of axon branch points. It is not yet clear why the OSN axons failed to enter the glomeruli when Shh function was blocked. It is possible that Shh provides the general target recognition signal to help OSN axons innervate the glomerular region. It is also possible that Shh serves as a growth promoting signal to the OSN axons. We did observe a slight increase in OSN axon elongation when Shh peptide is provided in vitro. Therefore, stalling of the immature OSN axons in the nerve layer could be caused by disturbance of the delicate balance between inhibiting and promoting cues in the glomeruli for OSN axons. When one of the promoting cues was missing, a negative growth environment was created, and OSN axons remained within the nerve layer. Further study is needed to investigate these hypotheses.