Abstract
Axonal patterning in the neocortex requires that responses to an individual guidance cue vary among neurons in the same location, and within the same neuron over time, to establish the varied and diverse sets of connections seen in the developed brain. The growth cone is the structure at the tip of a growing axon that interprets these guidance cues and translates them into morphological rearrangements. Different brain regions often reuse the same individual guidance cue molecules. Consequently, the sensitivity or type of response the growth cone has to various guidance cues varies over the course of maturation of the neuron, as the growth cone weaves its way into and out of different brain regions in order to reach its target. In addition there are different responses between different parts of the same neuron: the dendritic growth cones respond very differently compared to axonal growth cones to the same guidance cue using the same basic cytoskeletal proteins. How a growth cone responds to a guidance cue is thus a complex cell biological question and the molecular mechanism are not well understood. I propose that the atypical intermediate filament protein nestin, in its role as a kinase modulator, functions in this manner to change responsiveness to the same guidance cue in time and space.
Nestin is expressed highly in neural progenitors and is thus used widely as a progenitor marker. In my work, I discovered a subpopulation of mouse embryonic cortical neurons which transiently express nestin in their axons, but not in their dendrites. Since the prevailing view contends that intermediate filaments are not present in growth cones, intermediate filaments are understudied in developing axons, with very few studies looking into the role of intermediate filaments in axon guidance. Many studies, on the other hand, investigate the role of actin and microtubule dynamics in growth cones. I used high resolution imaging to demonstrate that nestin containing filaments are present in the distal axon and the growth cone and even extend into peripheral regions of the growth cone including filopodia. Nestin expression is thus not restricted to neural progenitors but persists for 2-3 days at lower levels in newborn neurons. Since protein expression level in progenitors is far greater than in axons, without careful and technically challenging analysis axonal nestin may be easily (although incorrectly) dismissed as background staining- particularly in vivo.
This early period of neuronal maturation, in which nestin protein can be detected, is a critical period that establishes the orientation of the axon and its future trajectory. I initially found that nestin-expressing neurons have smaller growth cones, suggesting that nestin affects cytoskeletal dynamics. Nestin, unlike other intermediate filament subtypes, regulates Cdk5 kinase via binding the Cdk5 activator p35. Cdk5 is a critical kinase that operates during early neuronal development, and is part of the downstream signaling machinery of the repulsive axon guidance cue, Sema3a. Sema3a is important for initial axon positioning and organization within the intermediate zone of the developing cortex. I find that nestin selectively facilitates the phosphorylation of the lissencephaly-linked protein doublecortin (DCX) by Cdk5/p35, but surprisingly the phosphorylation of other Cdk5 substrates is not affected by nestin. DCX is a microtubule binding protein, which stabilizes microtubules. Phosphorylation of DCX by Cdk5 decreases microtubule binding and thus induces microtubule instability. DCX and DCX phosphorylation have previously been shown to be important for axon guidance. I uncover that the Cdk5 substrate selectivity imparted by nestin is based on the ability of nestin to specifically interact with DCX, but not with other Cdk5 substrates. In addition, I find that other intermediate filaments do not interact with DCX. Nestin thus creates a selective scaffold for DCX with activated Cdk5/p35 to facilitate robust phosphorylation. Neurons that express nestin are more sensitive to the repulsive effects of Sema3a, and nestin siRNA reduces growth cone sensitivity to Sema3a. A nestin mutant that does not interact with DCX or p35 and does not affect DCX phosphorylation, does not have the same effects on growth cone morphology as WT nestin. Lastly, I use cortical cultures derived from DCX knockout mice to show that the effects of nestin on growth cone morphology and on Sema3a sensitivity are DCX-dependent, thus suggesting a functional role for the specific DCX-nestin complex in neurons. DCX null neurons are more sensitive to Sema3a in a nestin-independent manner. This observation suggests that DCX is operating as a brake or buffer to resist the cytoskeleton destabilizing effects induced by Sema3a. Nestin-enhanced Cdk5 phosphorylation of DCX removes the DCX brake on microtubules and thus DCX dissociates from microtubules in WT neurons.
I propose that nestin changes growth cone behavior by regulating subsets of substrates that are phosphorylated by intracellular kinase signaling downstream of guidance cues in developing neurons. Thus, the transient expression of nestin could allow for temporal and/or spatial modulation of a neuron's response to Sema3a, particularly during early axon guidance. Temporally, at 2-3 days, when nestin expression can no longer be detected in vitro, is the same time period in vivo in which cortico-cortical axons are growing into a Sema3a rich region- and these are insensitive to the repulsive effects of Sema3a. This is consistent with the loss of nestin during that time period. Spatially, nestin’s absence from dendrites may play a role in the lack of sensitivity of dendrites to the repulsive effects of Sema3a.
This is one example of how generalized signaling pathways can have variable signaling outcomes depending on which signaling modifying proteins are present. For example, the absence or presence of signaling modifying proteins, such as nestin, which can alter the subsets of substrates that are phosphorylated, may provide a mechanism how Cdk5 can be involved in both the attractive response of dendrites to Sema3a, and the repulsive response of axons to the same guidance molecule. This raises the question if other intermediate filaments may be playing similar roles. The exquisite developmental and cell type variability of intermediate filament expression could in principle provide unique signaling chemistries depending on which intermediate filaments are expressed. Intermediate filaments may thus have functions beyond mechanically providing a structure or skeleton to cells, but modify signaling networks.
