Neuronal death in the spinal ganglia of the chick embryo and its reduction by nerve growth factor.

In the spinal ganglia of the chick embryo, two neuronal populations can be distinguished: large, early differentiating ventrolateral (VL) cells and small, late differentiating dorsomedial (DM) cells. It was found that, beginning with stage 25, the DM cells originate from a narrow band of small, immature cells at the medial border of the ganglion, extending to the dorsolateral border. We have designated this band as the inner and outer marginal zone. Neuronal death was investigated in thoracic ganglion 18 and brachial ganglion 15 by counting degenerating cells, separately for the VL and DM populations, at every stage from stage 24 (4% days) to stage 38 (12 days). In both ganglia, separate degeneration periods were found for the VL and DM populations which do not overlap. The peaks of degeneration are: stage 27 (5% days) for the VL population in ganglion 18, stage 30 (6% to 7 days) for VL in ganglion 16, and stage 35 (8% days) for DM in both ganglia. Daily injections of 6 pg of nerve growth factor, (NGF) in 6 to 12 ~1 of 0.9% sterile salt solution into the yolk sac from stage 21 (3% days) to the day of sacrifice resulted in a significant reduction of neuronal death in the VL population of ganglion 18 and in the rescue of practically all VL neurons in ganglion 15 and all DM neurons in both ganglia, which normally would have died. This is the first demonstration of an NGF effect on VL neurons. In the analysis of the multiple effects of nerve growth factor (NGF), the sympathetic ganglia have played the dominant role; the sensory ganglia which are the other target of NGF have attracted much less attention. Yet, if one entertains the notion that NGF might be the natural trophic maintenance agent for these two neuron types (i.e., that NGF is actually produced by their target organs), then one approach to the testing of this hypothesis would be to engage in in viva experiments. For this ’ Dedicated to Dr. Rita Levi-Montalcini. This work was supported by Grant NS05721 from the National Institute of Neurological Communicative Disorders and Stroke, National Institutes of Health and a grant from the Jerry Lewis Neuromuscular Research Center to Washington University School of Medicine. We are grateful to Drs. Bradshaw, Costrini, and Ruben of the Department of Biochemistry, Washington University School of Medicine and Dr. Shooter of the Department of Neurobiology, Stanford University School of Medicine for the generous supply of nerve growth factor. We thank Mr. Cramer Lewis of the Department of Medical Illustrations, Washington University School of Medicine for the photographic work and Ms. Vicki Friedman of the Department of Surgical Illustrations for execution of the line drawings. We also thank Mrs. Helen Gregory for very competent technical assistance and Mrs. Irma Morose for typing the manuscript. ’ Present address: Department of Anatomy, Vanderbilt University School of Medicine, Nashville, TN 37215. 3 Present address: Department of Physiology, University of Pittaburgh School of Medicine, Pittsburgh, PA 15261. undertaking, sensory ganglia would be preferable to sympathetic ganglia, because their early differentiation processes, including cell loss by degeneration, are much better known than those of sympathetic ganglia. Since a direct attack on this problem, for instance an attempt to isolate NGF from specific embryonic tissues seems unrealistic at this time, one has to resort to the marshaling of indirect evidence. The first step was the demonstration that NGF is transported retrogradely and selectively from the hindlimb to lumbar dorsal root ganglia of chick embryos at a stage (10 d of incubation) when the ganglia are known to be responsive to NGF (BrunsoBechtold and Hamburger, 1979). The next step is to test the role of NGF as a trophic maintenance agent in Go. We know that NGF is essential for the survival of dissociated sensory and sympathetic neurons in vitro (LeviMontalcini and Angeletti, 1963; Green, 1977). We have now addressed the question of whether NGF can alleviate normally occurring neuronal death in sensory ganglia in the embryo. In a previous study (Hamburger and Levi-Montalcini, 1949), distinct patterns of distribution of degenerating cells in spinal ganglia had been recognized. First, neuronal death was found to be much more extensive in cervical and thoracic ganglia than in limb-innervating ganglia. This differential accounts, in part, for the conspicuous volume differences between the two groups of ganglia. In addition to this regional pattern, a temporal The Journal of Neuroscience Cell Death in Normal and NGF-treated Spinal Ganglia 61 pattern was observed; massive degeneration occurs in spinal ganglia during the period between 4% and 71/2 d of incubation. Finally, there are two populations of neurons in the spinal ganglia of chick embryos; early differentiating large neurons in a ventrolateral position (VL) and smaller, late differentiating neurons in a dorsomedial position (DM). They show differences in the degeneration pattern. Degenerating cells were seen primarily in the VL population. A corresponding difference was found in limb bud extirpation experiments; a massive degeneration occurred in VL cells, whereas the DM cells seemed to undergo only a severe atrophy. The re-investigation of this phenomenon on a larger scale and extending over a longer period of development necessitates a substantial revision and provides supplementation of these data. Most importantly, the extension of the observations beyond day 8 has revealed a second period of degeneration, affecting the DM neuron population in both brachial and thoracic ganglia. Carr and Simpson (1978) have reported similar observations (see “Discussion”). The new, more detailed data have served as the base line for the study of NGF effects on degeneration. Materials and Methods Two groups of chick embryos (white Leghorn, Spafass) were used in the present study. In the first group, normal embryos were sacrificed and staged according to the Hamburger and Hamilton stage series (Hamburger and Hamilton, 1951). In a second group, the embryos received daily injections of nerve growth factor (NGF) after which they were treated identically to those in the normal series. In each group, five to eight embryos were sacrificed at each stage between stage 24 and stage 38. All embryos were kept in a forced draft humidifier at 37°C and 70% relative humidity. NGF injections. The shells of embryos to receive NGF injections were opened on the 2nd or 3rd day of incubation using a rotary sander. Daily NGF injections were made at approximately the same time each day through this window which was sealed with a sterile coverslip and paraffin after each injection. The NGF injection was made with a lo-p1 Hamilton syringe into the yolk sacs of the embryos beginning at about stage 21. A single dose of NGF consisted of 6 pg of purified NGF in 6 to 12 ~1 of 0.9% sterile saline to which a solution of Millipore-filtered bovine serum albumen (BSA) had been added to produce a concentration of 6 mg/ml of BSA. Histological preparation. The embryos were fixed by immersion in Carnoy’s fluid, embedded in paraffin, and sectioned coronally at 10 pm. Serial sections were mounted and stained with thionine or Pappenheim’s solution (methyl green and pyronin Y). Reconstructions. To identify the individual ganglia used in this investigation, a reconstruction of the lower cervical, brachial, and upper thoracic ganglia was made for each embryo. The reconstruction was made on graph paper by marking the lengths of ganglia 12 through 18 and their relative positions along the rostrocaudal axis. The lengths and positions were drawn to scale by equating each section with a gradation along the length of the reconstruction. To ascertain the identity of the ganglia, the brachial plexus and the emergence of the main brachial nerves were reconstructed. Method of studying degeneration. The major difficulty which one encounters in morphometric studies of embryonic spinal ganglia is the overlap of proliferation and cell death. Hughes (1961) and Prestige (1967), who have studied this phenomenon in anuran embryos, have referred to it as “turnover.” It results in a continuous change in population size, since birth rates and death rates have different (and possibly unrelated) time courses. In a study of neuron degeneration in the spinal ganglia of chick embryos, Carr and Simpson (1978) computed a degeneration index which is the percentage of degenerating cells in the total population. We have not adopted this method for two reasons. First, the total population is heterogeneous; it consists of a mixture of proliferating neuronal and glial precursors as well as neurons and glia in various stages of differentiation. In addition, the ratios of these cell types are changing continuously. Second, the indices are very low (mostly between 2 and 4%). As a result, the curves (their Fig. 5) are flat over long periods, and a visualization and evaluation of the differences between normal and NGF-treated neuron populations would have been very difficult. In the present study, we have obtained satisfactory results by simply counting all frankly degenerating cells in every other section through a ganglion. These counts were made at every stage from stage 25 (4% d) to stage 38 (12 d). To make the project manageable, two representative ganglia were chosen: Ganglion 15 represents limb-innervating ganglia and ganglion 18 represents nonlimb ganglia. The latter is the most rostral ganglion that innervates exclusively thoracic structures since ganglion 17 contributes frequently to the brachial plexus. Identification and counting of degenerating cells. At the magnification used (X 750), the identification of frankly degenerating cells is unequivocal. They can be distinguished clearly from mitotic cells with clumped chromatin. Degenerating cells appear in a variety of forms, ranging from multiple basophilic small spheres in an otherwise intact nucleus and perikaryon to unrecognizable fragments. Frequently, the cell mass is condensed in a deeply stained homogeneous sphere. In other instances, the cells become vacuolated or ballooned. Macrophages containing debris of several cells were encountered infrequently and were counted as two degenerating cells. Further details of degeneration in sensory ganglia are dealt with extensively by Pannese (1976) who also discussed the pertinent literature. Obviously, we miss the incipient stages of degeneration which are revealed only by the electron microscope, but this is not a problem in the present study since the absolute figures are used only for comparison. Cell counts were made separately for the VL and DM populations. Consequently, a major concern was the distinction between degenerating VL and DM cells. The two populations are not always separated by a clear demarcation line. Nevertheless, it was usually possible to identify a degenerating cell as VL or DM on the basis of its topographic position (ventrolateral or dorsomedial). In case of doubt, the size of immediately adjacent cells was used as a criterion, the VL cells being considerably 62 Hamburger et al. Vol. 1, No. 1, Jan. 1981 Figure 1.4 Right thoracic ganglion 18, stage 27 (5 to 5% d), at the peak of the VL degeneration period. Magnification, X 600. larger than the DM cells during the period of our study. In later stages, one has to reckon with degeneration of glia cells. In fact, occasionally, we have encountered very small, solid spheres which might be degenerating glia. They were not counted. Both left and right ganglia 15 and 18 were counted in each embryo. Every other section was counted under oil immersion (X 750) and the totals were multiplied by 2. For each stage, 10 to 16 ganglia were counted both in the normal and in the NGF-treated series. Most counts were made by Viktor Hamburger and some were made by J. Brunso-Bechtold.