Earlier we emphasized the need to acknowledge the variability of the glomerular array, which is best appreciated when relative positions of two glomeruli are compared between the two bulbs of the same mouse (Figure 12.1). Self-sorting is expected to be error prone and less precise than guidance by bulb-derived molecular cues. Nevertheless, the conservation of glomerular positions is impressive. How can a self-sorting mechanism result in a glomerular array that is so conserved among non-outbred mice? (Mice of a mixed 129 x C57BL/6 background, which are commonly used in gene-targeting experiments, would qualify as non-outbred.)
We hypothesize that in addition to OR sequence, OR level, and positional cell type, the temporal expression pattern is a critical determinant. Unfortunately and surprisingly, ignorance rules. Very little information is available about the onset of OR expression and the kinetics of the expansion of the population of OSNs expressing a particular OR gene in mouse; moreover, such data have not been related to the dynamics of glomerular development in a comparative fashion among OR genes. No information is available about the first arrival of axons ofOSNs expressing a particular OR gene, and their behavior at the surface ofthe olfactory bulb. No time-lapse imaging system, in vivo or in explants, is available formouse. Aglobal study ofglomerular development indicates that the development of glomeruli in the rostral part ofthe rat olfactory bulb is two to four days ahead of the caudal part, but substantial heterogeneity exists within a microregion (Bailey et al. 1999). P2 glomeruli form in late gestation (Royal and Key 1999), OR37 glomeruli soon after birth (Conzelmann et al. 2001), and M71/ M72 glomeruli a little later (Potter et al. 2001). Thus, the glomerular array develops asynchronously. Axonal wiring could be highly dependent on this asynchronous development; a glomerular array that develops synchronously may be much more difficult to wire.
We speculate that, in addition to OR sequence, OR protein level, and positional cell types, glomerular positions could be the outcome of when, where, and how often a particular OR gene is expressed before and during glomerular formation. A group of early-born and early-arriving axons may self-sort and coalesce into a proto-glomerulus at the first available position in the bulb that is appropriate for that positional cell type (say, extremely dorsally). Axons of another
OR expressed by the same positional cell type may achieve a threshold number for proto-glomerular formation slightly later and coalesce at the next available position in the extreme dorsal domain of the bulb. Axons may coalesce initially into several proto-glomeruli, which may not be homogeneous, and subsequent remodeling may be achieved by death of cells with inappropriately wired axons (Zou et al. 2004). We further suggest that axons may have different propensities to move on the surface of the bulb depending on the expressed ORs. This successive colonization of sites in domains of the olfactory bulb might produce a glomerular array that is characteristic (but not identical) among non- outbred mice. This hypothesis can be tested experimentally by systematically varying the number of OSNs that express a particular OR gene, in heterozygous mice or with transgenes.
Perhaps all that happens in some OR replacements where a novel, ectopic glomerulus is formed is that "when, where, and how often" has been changed. In this respect, ORs would, strictly speaking, not be "axon guidance molecules," but "identity" or "acquaintance molecules." The key to unraveling axonal wiring may lie in understanding the mechanisms that restrict the choice of expression among the repertoire of OR genes, in an OSN at a particular time in development at a particular position in the epithelium.
Was this article helpful?