Hyperplasia in the spinal sensory system of the frog. I. Plasticity in the most caudal dorsal root ganglion

Nitric oxide synthase expression and cell changes in dorsal root ganglia and spinal dorsal horn of developing and adultRana esculenta indicate a role of nitric oxide in limb metamorphosis

Metamorphosis of amphibians requires reconfiguration of sensory and locomotor neural networks. In view of such plastic changes and implications of nitric oxide (NO) in neural developmental shaping, we examined via histochemistry and immunohistochemistry its synthetic enzyme nitric oxide synthase (NOS) in dorsal root ganglia (DRGs) and dorsal horn of the developing and adult frog Rana esculenta. In limb DRGs, NOS positivity was first and selectively detected just before limb bud appearance, increased during metamorphosis, and was then down-regulated. In adulthood, NOS was expressed in some DRG neurons at all segmental levels. Similar features were detected in the dorsal horn neuropil. In limb DRGs, cell counts in Nissl-stained sections revealed a twofold increase of differentiated neurons during metamorphosis and an additional twofold increase in adulthood. Perikaryal sizes in limb DRGs did not vary during metamorphosis but increased and were more heterogeneous in the adult frog, probably reflecting adaptation to body size. NOS and cell changes during metamorphosis were much less marked in DRGs at other levels. Carbocyanine tracing documented selective labeling of NOS-expressing hindlimb DRG neurons from the spinal nerve at the time of initiation of hindlimb movements. The findings show that, in limb DRG neurons, NOS parallels cell differentiation and limb development during metamorphosis. The data also provide evidence of NOS expression in DRG cells innervating the hindlimbs when sensorimotor circuits become functionally mature. This study indicates a key role of NO production in the maturation of sensory functions that subserves in amphibians the transition from swimming to tetrapod locomotion.

Excess target-derived neurotrophin-3 alters the segmental innervation of the skin

It is thought that dermatomes are established during development as a result of competition between afferents of neighbouring segments. Mice that overexpress neurotrophins in the skin provide an interesting model to test this hypothesis, as they possess increased numbers of sensory neurons, and display hyperinnervation of the skin. When dermatomal boundaries were mapped in adult mice, it was found that those in nerve growth factor and brain-derived neurotrophic factor overexpressers were indistinguishable from wild-type animals but that overlap between adjacent segments was greatly reduced in neurotrophin-3 (NT-3) overexpressers. However, dermatomes in heterozygous NT-3 knockout mice displayed no more overlap than wild-types. In order to quantify differences across strains, innervation territories of thoracic dorsal cutaneous nerves were mapped and measured in adult mice. Overlap between adjacent dorsal cutaneous nerves was normal in nerve growth factor overexpressing mice, but much reduced in NT-3 overexpressers. However, this restriction was not reflected in the central projection of the dorsal cutaneous nerve, creating a mismatch between peripheral and central projections. Dorsal cutaneous nerve territories were also mapped in neonatal mice aged postnatal day 7-8. In neonates, nerve territories of NT-3 overexpressers overlapped less than wild-types, but in neonates of both strains the amount of overlap was much greater than in the adult. These results indicate that substantial separation of dermatomes occurs postnatally, and that excess NT-3 enhances this process, resulting in more restricted dermatomes. It may exert its effects either by enhancing competition, or by direct effects on the stability and formation of sensory endings in the skin.

Number, size, and regional distribution of motor neurons in the dorsolateral and retrodorsolateral nuclei as a function of sex and neonatal stimulation

Motor neurons were measured in the retrodorsolateral nucleus (RDLN) and the dorsolateral nucleus (DLN) of adult male and female rats that were reared with normal or reduced levels of maternal anogenital stimulation. In contrast with findings for the spinal nucleus of the bulbocavernosus, which is located in the same spinal segments, reduced stimulation had no effect on neuron number in either nucleus. However, several regional and sex differences were observed. Rostrally located neurons were larger in both the RDLN and the DLN; these location effects were greater in females. There was no sex difference in RDLN neuron size, but DLN neurons were larger in females, particularly in the rostral region. Females had significantly more cells in the RDLN, a nucleus previously considered nondimorphic, whereas males had more DLN neurons. Both regional and sex differences may reflect local differences in trophic factors from targets or afferents.

Changes in fast axonal transport in sensory neurons during tadpole metamorphosis

Fast axonal transport of radiolabeled protein was examined in lumbar and tail dorsal root ganglion (DRG) neurons at progressive stages of bullfrog tadpole metamorphosis. Accumulation of [35S]methionine-labeled protein proximal to a lumbar peripheral nerve ligature (at a fixed distance from the DRG) increased as tadpoles advanced from premetamorphosis through prometamorphosis to metamorphic climax. The rate of increase was steeper when expressed as a percentage of protein synthesized in the neurons of origin than when expressed as a percentage of total DRG protein synthesis. Further, the increase was not secondary to a rise in protein synthesis. In contrast, fast axonal transport decreased in DRG neurons of the tail at the onset of metamorphic climax, when tail resorption is initiated. The stage-related increase in protein transport in lumbar nerves is due, at least in part, to an increased rate of transport. As determined from optically detected anterograde organelles in individual lumbar nerve axons, an approximate doubling of the fast transport rate occurred between the premetamorphic stage and metamorphic climax. In addition, the rates of organelle transport in lumbar axons of adult bullfrogs were significantly greater than in corresponding axons of tadpoles at metamorphic climax, further suggesting that organelle velocity is a developmentally regulated parameter of fast axonal transport.

Odor deprivation leads to reduced neurogenesis and reduced neuronal survival in the olfactory bulb of the adult mouse

Neurogenesis persists in the olfactory bulbs of adult mice, with new cells being generated in the proliferative subependymal layer. Our previous work has shown that unilateral odor deprivation through naris closure leads to a net loss of granule neurons in the ipsilateral (odor-deprived) olfactory bulb, while not affecting the contralateral bulb. Here we used several experimental approaches to determine if this loss of neurons results from reduced neurogenesis, reduced neuronal survival, or both. First, bromodeoxyuridine immunohistochemistry was used to determine the number of S-phase cells in the subependymal layer eight weeks after naris closure. Proliferation was reduced within and just caudal to the odor-deprived bulb compared to the open-side (control) bulb. Second, counts of pyknotic nuclei four weeks after naris closure were used to document a higher rate of cell death on the deprived side. Third, 3H-thymidine autoradiography was used to assess differences in granule cell survival on the two sides. Granule cell precursors were labeled by a single injection of 3H-thymidine eight weeks after naris closure, and the number of surviving labeled granule cells assessed four and 16 weeks later. Granule cell survival was significantly reduced within the odor-deprived bulbs. These data indicate that the loss of granule cells which follows odor deprivation is caused, at least in part, by reduced neurogenesis and reduced survival of these adult-generated neurons.

Collateral sprouting in the electrosensory lateral line lobe of weakly electric teleosts (gymnotiformes) following ricin ablation

Sprouted collateral axons were observed in the electrosensory lateral line lobe (ELL) of gymnotiform teleosts (Apteronotus leptorhynchus) following the ablation of the supraorbital branch of the anterior lateral line nerve. Ablation was accomplished by using microinjections of the toxic lectin ricin. Sprouted axons were followed for up to 26 weeks postablation. Ricin exposure severely reduced axonal numbers and the peripheral electroreceptors in the region innervated by these fibers. To visualize sprouted fibers, intact lateral line afferent nerve branches were anterogradely labelled with the neuronal tract tracers horseradish peroxidase or cobalt chloride, or the monoclonal antibody Q26A3. Within the four somatotopically organized ELL segments, sprouted collaterals were first observed two weeks after ricin injection in the medial and centromedial segments, and four weeks postinjection in the centrolateral and lateral segments. Sprouting involved intrasegmental, horizontally directed axons from adjacent nerve branch terminal fields, and mixed intra- and extrasegmental, dorsally directed axons from the ELL deep fiber layer. The sprouting response was robust but variable in its timing, peaking between 6 and 12 weeks. Subsequently, the intrasegmental, horizontally directed fibers were retained but the mixed dorsally directed fibers, including all extrasegmental axons, were retracted. Therefore, this sprouting response appears to consist of a collateral overproduction followed by a selective axonal retraction. In our view, the most likely explanation for this axonal retraction is that the descending inputs from the isthmus and the cerebellum, as well as commissural fibers from the contralateral ELL, maintain established somatotopic relationships by eliminating somatotopically mismatched sprouted collaterals.

Times of origin of brachial sensory neurons are not correlated with neuronal phenotype

The times of origin (birthdays) of sensory and motor neurons that innervate the triceps brachii muscles of the bullfrog (Rana catesbeiana) were determined to learn whether neurons innervating a specific target are generated at a particular developmental time. 3H-thymidine (3H-TdR) was made available continuously throughout a specific developmental period. All neurons that innervated the triceps muscle in juvenile frogs (identified by filling cells retrogradely with HRP) were generated prior to metamorphosis. Triceps motoneurons were all postmitotic by early limb bud stage V. Triceps sensory neurons were generated over a protracted period of larval development, from stage V through early pre-metamorphic stage XV. Most large triceps sensory neurons were generated before the majority of the small cells. However, there was considerable overlap in the times of origin of the two populations; both large and small cells were generated at all stages of sensory neurogenesis. There was thus no strict relationship between sensory soma size and birthdate. Late-generated sensory neurons tended to be located in clusters within ganglia, whereas HRP-filled triceps neurons were not. These 3H-labeled clusters may represent clones of neurons which would indicate that late stage neuroblasts give rise to neurons that supply different peripheral targets. The time course of triceps neuronal generation paralleled that of all other brachial sensory neurons implying that the time of last cell division does not in itself determine either the target a neuron will innervate or the sensory modality to which it will respond.

Myotome and early neurogenesis in chick embryos

The present study was undertaken in order to verify the identification of profiles of presumptive growth cones in vivo. The developing spinal nerves of chick embryos were studied by light and electron microscopy. We traced the onset of efferent and afferent innervation of the myotome in 2- to 4-day-old chick embryos in order to be sure that we were examining the growing tips of axons. In the process of studying these growing axons, we were able to observe some unique relationships of neural tube, myotome, and differentiating spinal nerves. The neural tube tightly abuts the myotome in Hamburger and Hamilton's (HH) stage 14 chick embryos and cytoplasmic projections from the myotome directly abut the neural tube. The first ventral roots could be identified in HH stage 15 embryos and dorsal roots in HH stage 16 embryos, both under 2 1/2 days of age. The advancing spinal nerve courses toward the anterior or cranial half of the myotome, and growth cones directly contact the medial wall of the myotome. The spinal nerves continue to abut tightly the myotome during the succeeding day of embryonic life, and growth cones enter the substance of the myotome by 3 days, or HH stage 19 embryos. These dorsolaterally directed axons will form the dorsal ramus of the spinal nerves and the ventral ramus continues to be contiguous with the myotome. Invasion of the myotome by axons (putative innervation), and thus innervation of myotomal cells in the 3-day chick embryos, was a totally unexpected finding. The myotome and its potential derivatives thus have extensive neural contact by 3 days of embryonic life in the chick. These findings document a parallel differentiation of afferent and efferent elements of the nervous system and confirm previous accounts identifying growth cones in an intact organism. These findings suggest that afferent as well as efferent nerves may have critical roles in the differentiation of the mesodermal as well as ectodermal derivatives.

Peripheral organs control central neurogenesis in the leech

Interactions between developing nerve centres and peripheral targets are known to affect neuronal survival and thus regulate the adult number of neurons in many systems. Here we provide evidence that peripheral tissues can also influence cell numbers by stimulating the production of neurons. In the leech Hirudo medicinalis, there is a population of several hundred neurons that is found only in the two segmental ganglia that innervate the genitalia and which seem to be added gradually during post-embryonic maturation. By monitoring 5-bromo-2'-deoxyuridine incorporation immunohistochemically, we have now determined that these neurons are actually born late in embryogenesis, well after all other central neurons are born and after efferent and afferent projections are established between these ganglia and the periphery. Ablation of the male genitalia early in embryogenesis, or evulsion of the nerves that connect them to the ganglia, prevent the birth of these neurons. However, they fail to appear ectopically when male genitalia are transplanted to other segments, despite innervation by local ganglia. We conclude that the generation of the late-appearing neurons depends on a highly localized signal produced by the male genitalia, to which only the ganglia that normally innervate these organs have the capacity to respond.

Variation and symmetry in the lumbar and thoracic dorsal root ganglion cell populations of newly metamorphosed Xenopus laevis

The sizes of the lumbar and thoracic dorsal root ganglion cell populations in normally developing newly metamorphosed Xenopus laevis were measured in order to determine whether these neuron populations have the same characteristics as the hindlimb motoneuron population (i.e., large individual as well as sibling group differences, striking bilateral symmetry, and a rough correspondence between neuron number and body size that suggests some peripheral control of cell number during normal development (Sperry, J. Comp. Neurol. 264:250-267). Among animals from three sibling groups, the total numbers of thoracic and lumbar ganglion cells are highly variable and symmetrical, although symmetry is not uniformly present at the level of individual ganglion pairs. Significant sibling group differences in neuron number are also present. Metamorphic body size and cell number in the thoracic but not in the lumbar ganglia are significantly correlated. The motoneurons innervating the hindlimbs were also counted and measured in the same animals. While variable as well as symmetrical, motoneuron number and metamorphic body size are correlated in only two of the three sibling groups. Interestingly, the numbers of motoneurons and lumbar ganglion cells, two populations of neurons whose sizes one might predict would be significantly correlated in normally developing animals, are not correlated. The relationship between these observations and currently held views concerning how neuron numbers might be controlled during normal development is discussed.

Peripheral competition in the control of sensory neuron numbers in Xenopus frogs reared with a single bilaterally innervated hindlimb

Sensory neurons were counted in the hind-limb innervating spinal ganglia on both sides of juvenile Xenopus frogs which, as tadpoles, had had one hind limb bud amputated prior to innervation, and a channel made to allow innervation of the remaining limb bud from both sides. The total number of sensory neurons surviving on the two sides approximated the number on one side of normal frogs, the ipsilateral and contralateral numbers being negatively correlated. These effects differ markedly from the effects on motoneuron numbers, suggesting different control mechanisms of cell death in the two neuronal classes.

Reformation of specific synaptic connections by regenerating sensory axons in the spinal cord of the bullfrog

The regrowth of sensory axons into the spinal cord of juvenile bullfrogs was studied after disruption of these fibers in the dorsal root. Within 9 d after the root had been frozen, regenerating sensory axons had reached the spinal cord, as revealed by labeling with horseradish peroxidase. Growth into the spinal cord, however, was much slower. Even several months after denervation, very few fibers had reestablished any of their normal longitudinal projections within the dorsal funiculus. Eventually, however, sensory axons grew across the region and into the dorsal horn. Intracellular recordings from motoneurons revealed that these axons made functional reconnections with spinal neurons. Muscle sensory axons established direct, monosynaptic inputs to motoneurons, whereas cutaneous fibers innervated these neurons polysynaptically. Moreover, sensory afferents from a particular muscle distinguished among different classes of motoneurons, just as in normal frogs. Thus, specific synaptic pathways can be reestablished by regenerating sensory axons if they can reach their appropriate target region within the spinal cord.

The effects of neural crest deletions on the development of sensory innervation patterns in embryonic chick hind limb

Selected lumbosacral dorsal root ganglia (d.r.g.s) were eliminated from chick embryos by removing the parent neural crest, and the dermatomes and axonal projection patterns established in the hind limb by the remaining intact d.r.g.s were studied physiologically and/or anatomically. Dermatomes of intact d.r.g.s expanded into denervated skin regions, partially but never completely replacing the lost innervation; some skin regions consistently remained without apparent innervation. Dermatome expansion was detected in young embryos, soon after skin innervation was established, suggesting that skin sensory axons in operated embryos ignored their usual dermatome borders. The axonal pathways (nerve trunks) normally supplied by axons from the deleted d.r.g.s were missing, and axons from intact d.r.g.s were only rarely found in aberrant pathways. In contrast, the relative distribution of axons from intact d.r.g.s within their usual complement of pathways was altered, with axonal projections shifting toward, but not into the deleted pathways. Shifts in axonal projection patterns were observed in embryos prior to the onset of sensory cell death. Thus, d.r.g. deletions appeared to affect the initial growth of axons into the limb. Together these results suggest that during normal development competitive interactions among axons from neighbouring d.r.g.s play an important role in establishing the borders between dermatomes and in determining the distribution of axons within projection pathways in the limb.

Hyperplasia in the spinal sensory system of the frog. II. Central and peripheral connectivity patterns

Central and peripheral connectivity patterns of hyperplastic dorsal root ganglia (DRGs) in Rana pipiens are examined in order to determine the relative roles of peripheral and central contacts in the production of DRG hyperplasias. The hyperplasias are produced in the intact hindlimb DRGs after the removal in tadpoles and young postmetamorphic frogs of neighboring DRGs (Davis and Constantine-Paton, '83). The peripheral target zones of the hyperplastic DRGs, determined by physiological recordings of sensory receptive fields, are found to undergo a significant degree of expansion relative to controls. Peripheral expansion is most pronounced in caudalmost DRG 10, and this effect occurs in experimental animals operated during larval and postmetamorphic stages. Further, anatomical labelling of peripheral sensory fibers coursing to the hindlimb reveals that the hyperplastic DRG 10 actually contains additional fibers projecting to the denervated regions. The central projection of the hyperplastic DRG 10 does not show corresponding increases in longitudinal arborization after the application of horseradish peroxidase to the appropriate dorsal roots. These observations are made on some of the same experimental animals in which peripheral fields are shown to have vastly expanded. We conclude that the peripheral processes of the hyperplastic DRGs are less rigidly specified than the central terminations, and that it is the periphery which plays the primary role in controlling the cell numbers increases. A second aim of this investigation is to identify whether sexually dimorphic connectivity patterns in normal frogs explain the production of DRG 10 hyperplasias exclusively in male experimental animals (Davis and Constantine-Paton, '83). We apply the same techniques used in our connectivity studies of hyperplastic DRGs to the investigation of connectivity patterns of DRG 10s in normal males and females. No sex-dependent differences in peripheral and central connectivity are found. Thus, since normal male and female frogs possess an equivalent amount of target space for DRG 10, the unique production of hyperplasias in male experimental animals cannot be explained solely on the basis of connectivity. We speculate on what other factors may be involved.