Competition and the dynamics of axon arbor growth in the cricket

The growth of an identified axonal arborization in the cricket cercal sensory system was studied under conditions that vary the number of neighboring axonal arborizations. The cell studied is one of a small number of neurons that arborize bilaterally and is called the X-neuron. Normally the axonal arborization of X is distributed roughly symmetrically about the midline. Seven days after the birth of this neuron, as soon after its birth as it can be stained, the axonal arbor is more than half the normal size and it exhibits the normal degree of bilaterality. During the remaining 50 days of postembryonic development, the arbor grows to its adult size, maintaining the bilateral distribution. The relationship between the growth of this axonal arbor and the presence or absence of its neighbors was studied by the removal of one of the sensory appendages. The removal of a cercus removes those neighbors near one half of the X-neuron's arbor and causes a dramatic shift in the bilateral distribution of X's axonal processes; after treatment nearly all of the varicosities were found in the deafferented region (Figs. 2, 6). Thus, neuron X responds to the loss of some to its neighbors by removing synaptic material from the area with normal neighbors and inserting additional material in the region with fewer neighbors. This effect is age dependent. Removal of neighbors early in life, during the initial period of synaptogenesis, causes a very rapid change, while similar treatment later in life causes a much slower response (Fig. 7). Thus the flexibility of the neuron is correlated with its growth rate; it is very flexible during its early, rapid growth phase and less flexible when the growth rate slows later in life. This continuing flexibility of the axonal arbor could also be demonstrated by allowing the amputated cercus to regenerate (Figs. 8, 9). Under these conditions the growth of X's arbor returned toward the normal bilateral distribution. However, regeneration of the neighboring arbors never restored the arbor to complete normality because a bias in the structure had already been imposed. In effect, regeneration arrested the change but could not reverse it. In brief, an identified sensory neuron's growth has been demonstrated to depend in part on the presence or absence of neighbors. This dependence extends throughout the life of the animal and fits the definition of a competitive interaction.