Proliferative and Degenerative Events in the Early Development of Chick Dorsal Root Ganglia I. NORMAL DEVELOPMENT VIRGINIA MCMILLAN CARR ' AND SIDNEY B. SIMPSON, JR. Department of Biological Sciences, Northwestern Uniuersity. Euanston, Zllinois 60201

ABSTRACT Development of the chick dorsal root ganglia was examined in 4.5- t o 9.5-day embryos. Tritiated thymidine (3H-TdR)and autoradiography was used t o analyze proliferative activity and the Feulgen procedure to analyze degenerative activity in ganglia 12-17. Proliferative activity was found to be elevated through 4.5 days of incubation when as many as 14%of the ganglionic cells become labelled following a one-hour exposure to 3H-TdR. By 6.5 to 7.5 days proliferative activity decreases to 2-4%in the lateroventral (LV) regions and to approximately 1%in the mediodorsal (MD) regions of the ganglia. However, there appears to be increased proliferative activity by the end of the experimental period a t 9.5 days. Birthdate studies demonstrate that large-scale neuronal production occurs between 4.5 and 6.5 days in the LV regions and between 4.5 and 7.5 days in the MD regions. After those times ganglionic proliferative activity must be largely nonneuronal in nature. This nonneuronal proliferation is greater in LV than in MD regions and in brachial than in nonbrachial ganglia. Degenerative activity was found to be absent from the ganglia until after 4.5 days of incubation. It then increases rapidly, and by 5.5 days 5%of the LV cells in nonbrachial ganglia are degenerating. Degenerative activity then declines but is still present a t 9.5 days. In contrast t o results of an earlier study (Hamburger and Levi-Montalcini, '491, degenerative activity was also found in the LV region of brachial ganglia and the MD regions of brachial and nonbrachial ganglia. The pattern of LV degenerative activity in brachial ganglia is similar to that in nonbrachial ganglia, but the level of activity is lower. In the MD regions degenerative activity increases throughout the experimental period, and by 9.5 days as many as 4%of the MD cells are degenerating. One feature of development of many parts of the nervous system is the occurrence of two opposing processes: cellular proliferation which leads to the production of large numbers of neurons and massive cellular degeneration which results in the loss of many of these same neurons. These two processes ultimately control the final number of neurons of a neural center. In the dorsal root ganglia of chick embryos these processes were first examined by Hamburger and Levi-Montalcini ('49) in a well-known study involving cell and mitotic counts as well as estimates of relative amounts of degenerative activity. This study was carried out in normal embryos and embryos subjected t o unilateral limb bud amJ. COMP. NEUR. (1978) 182: 727-740

putation or addition early in development. Hamburger and Levi-Montalcini concluded that normally proliferation reached a peak of activity at five to six days, then declined, and was no longer present in the ganglia by nine days of incubation. This proliferative activity was greater in limb than nonlimb ganglia. Degenerative activity was reported to commence between four to five days of incubation, rapidl y to reach a peak at five to six days, and t o be absent after seven days. This degenerative activity was concluded to occur only in nonlimb ganglia where it was reported to be confined to the lateroventral (LV) regions. Mediodorsal ' Present address Department of Biology, University of Chicago, 1103 East 57th Street, Chicago, Illinois 60637

727

728

VIRGINIA MCMILLAN CARR AND SIDNEY B. SIMPSON, JR.

(MD) regions of nonlimb ganglia showed no degenerating cells per se but only a slowly progressing atrophy later in development. A similar atrophy was reported in LV regions of nonlimb ganglia, also only a t later stages of development. Since the time of their publication, these observations have been generally taken to define the sequence of proliferative and degenerative events in the chick dorsal root ganglia. However, i t seemed as though a re-examination of these events was desirable. Hamburger and Levi-Montalcini's ('49) conclusions regarding proliferative activity were based on absolute numbers of mitotic cells and thus made no allowance for different sized ganglia. The study had also been carried out on a rather small population of ganglia. Furthermore, the technique of autoradiography using radioactively labelled DNA precursors is now known to provide a more accurate approximation of proliferative activity in a structure than do mitotic counts (e.g., Jacobson, '70; Sidman, '70). Finally, although Hamburger and LeviMontalcini ('49) had been the first to demonstrate massive cellular degeneration as a normal process in neural development, their conclusions had been based on subjective estimates of relative amounts of degenerative activity. Consequently, we have undertaken a rigorous re-examination of these events in early development of the chick dorsal root ganglia using tritiated thymidine and autoradiography to examine proliferative activity and the Feulgen procedure to examine degenerative activity. The period we examined was primarily t h a t in which Hamburger and LeviMontalcini ('49) reported all proliferative and degenerative events occurred, 4.5 to 9.5 days of incubation. Our results are presented in t h i s and t h e following paper (Carr and Simpson, '78). Although our findings corroborate some of those of Hamburger and LeviMontalcini ('491, they fail to corroborate others. A preliminary report of our findings has already appeared in abstract form elsewhere (Carr and Simpson, '75).

tor and swabbed with ethanol. A small hole was made in the blunt end of each egg with a sterile probe to allow the blastoderm to settle; and a square, approximately 1cm on a side, was cut in the shell of each egg with a small sanding disc fitted on a hand drill. All subsequent procedures were carried out in a sterile hood. The square flap of shell was carefully removed, and, after being wetted by one or two drops of sterile Dulbecco's Phosphate Buffered Saline (PBS; GIBCO, Grand Island, New York), the exposed vitelline membrane was removed from over the embryo. All labelling and degeneration index data presented in this paper come from the control, left sides of embryos that had been subjected to unilateral limb bud removal or addition on the right side at the brachial level at 2.5 days of incubation (stages 15 and 16, Hamburger and Hamilton, '51). The effects of those procedures on ganglionic development are discussed in the following paper (Carr and Simpson, '78). Following the operative procedures one drop of antibiotic-antimycotic (GIBCO) containing streptomycin (10,000 pg/ml), penicillin (10,000 u/ml) and fungizone (25 pglml) was placed over the embryo. The window in the shell was then tightly sealed with transparent tape and the egg returned to the incubator to continue development. At 4.5 to 9.5 days of incubation the embryos were exposed to [meth~l-~HI-thymidine (3H-TdR;sp. act. > 15 Ci mM, 500 pCi/ml, Amersham-Searle, Arlington Hgts., Illinois), as described below, for one hour and fixed. The 3H-TdR, diluted in PBS to 125 gCi/ml, was delivered to embryos of less than 9.5 days of incubation by injection onto the blastoderm or chorioallantois. The amounts delivered were: 4.5-day embryos, 12.5 pCi; 5.5- to 7.5day embryos, 25 &I; and 8.5-day embryos, 50 pCi. However, 9.5-day embryos received 25 pCi undiluted 3H-TdRby direct injection into the chorioallantoic vein because by this age the embryos could no longer pick up sufficient amounts of 3H-TdR from the blastoderm to give a n acceptable autoradiographic label. To avoid diurnal variations in incorporation of the labelled material, 3H-TdRwas always deMATERIALS AND METHODS livered between 10 A.M. and noon. Labelling and degeneration Fixation was in 10% (v/v) formalin in PBS. index studies The region from approximately the level of White Leghorn chick embryos (L.Sharp Co., ganglia 10 to that of ganglia 18 was isolated Glen Ellyn, Illinois) were incubated at 37-38' from the rest of the body, left in the fixative in a forced draft incubator. At the appropriate for two days, dehydrated in a graded series of time the eggs were removed from the incuba- ethanol, cleared in xylene, embedded in Para-

CHICK DORSAL ROOT GANGLIA DEVELOPMENT

plast (m.p. 56-57", Sherwood Medical Industries, St. Louis, Missouri), and sectioned a t 7 pm. Unincorporated 3H-TdR was removed from the mounted serial sections by exposure to trichloroacetic acid (TCA, 5% w/v) for five minutes. The sections were then stained using the Feulgen procedure as modified by Deitch e t al. ('68). The stained slides were dipped in NTB 3 emulsion (Eastman Kodak, Rochester, New York) diluted 1:3 (v/v) with distilled water, dried at room temperature for several hours, and stored in the dark a t 4" in small slide boxes containing Drierite@. Exposure time varied from five days to three weeks. The autoradiographs were then developed in D-19 photographic developer (Eastman Kodak) for five minutes a t 18". Following the autoradiographic procedures, the sections were counterstained with Fast Green FCS. Controls for the autoradiography consisted of sectioned material incubated with TCA a t 90" for 30 minutes to extract all nucleic acids from the sections or incubated with DNase at 37" for 24 hours to remove all DNA (Schultze, '69). No autoradiographic label was found in either case; thus, the labelling observed in our autoradiographs must be due entirely to 3HTdR incorporation into DNA. Labelling and degeneration indices were determined for dorsal root ganglia 12 through 17 in embryos sacrificed one hour after 3H-TdR delivery. A cross section through a ganglion was positioned on a microscope so that the area circumscribed by a 10 x 10 mm2 ocular grid a t 1,000 x under oil immersion included either the dorsal-most area of the MD region or the ventral-most area of the LV region. All cells in the circumscribed area with the exception of those comprising blood vessels or those along fiber tracts were counted. Thus, the population counted was heterogeneous in nature, including both neuronal and nonneuronal elements as well as their precursor cells. Each count included 100 to 300 cells, depending on the age of the embryo and the region (LV or MD) of the ganglion examined. The number of these cells that were labelled with 3H-TdR was also counted and expressed as a percentage of the total. This percentage was determined for each of three cross sections, the sections being separated by a t least 35 pm and representing the anterior, middle, and posterior of the ganglion. The average of these percentages was taken to be the labelling index for that region, LV or MD, of the ganglion. Individual labelling indices from 1 2 to 30 gan-

729

glia in four to ten embryos were averaged to give the mean labelling index for a set of ganglia. In a similar manner the percentages of cells degenerating were determined for the three sections and averaged to give the degeneration indices for the LV and MD regions of each ganglion. Separate determinations were made for the LV and MD regions of all embryos examined with the exception of 4.5-day embryos. Because of the small size of the 4.5-day dorsal root ganglia and the similarities in appearance of the LV and MD regions, counts a t this age were made in the center of each ganglion section and included both LV and MD cells. No effort was made to combine the separate data from the limb bud amputation (Exp. 1) and addition studies (Exp. 2) (Carr and Simpson, '78) and the data are tabulated separately throughout this paper. Although there are some differences between the two sets of data, the patterns of the labelling or degeneration indices are the same in both. The differences that are observed result most likely from carrying out the two studies a t different times of the year: the amputation experiments from November to April, the additions from April to August. The identity of the ganglia was determined from the location of the brachial plexus, which is formed from spinal nerves 1 3 to 16. In these studies ganglia 14 to 16 were taken t o be brachial ganglia. Ganglion 13 was taken to be a nonbrachial ganglion despite its contribution to the brachial plexus. This assignment was made on the basis of labelling and degenerative responses of ganglion 13 to limb bud removal (Carr and Simpson, '78). Birthdate determinations Two studies were undertaken to determine the birthdates of ganglionic cells. In neither study were any operative procedures performed prior to delivery of 3H-TdR. In a cumulative labelling study 3H-TdRwas delivered repeatedly to embryos for several days prior to fixation a t either 6.5 or 8.5 days of incubation to ensure adequate labelling of all proliferating cells. The 3H-TdR was given three times a t 6-hour intervals on the first day of exposure and once each day thereafter until fixation. Of the embryos fixed a t 6.5 days, those beginning labelling a t 2.5 days received doses of 10, 10, and 5 pCi 3H-TdR on the first day and 10, 5, and 6.25 pCi, respectively, on the subsequent days; those beginning a t 3.5

730

VIRGINIA MCMILLAN CARR AND SIDNEY B. SIMPSON, JR.

and 4.5 days received 12.5 pCi three times on the first day and once on each subsequent day; and those beginning a t 5.5 days received three treatments of 25 pCi each. All of the embryos fixed a t 8.5 days were given 25 pCi 3H-TdR three times on the first day and once on each subsequent day. Histological and autoradiographic procedures were as described for the labelling index studies. Sectioned material was examined for the percentage of ganglionic cells that were unlabelled. Such cells were assumed to have withdrawn from the proliferative cycle prior to the time of 3H-TdR delivery. One embryo was examined for each exposure period. In another series of experiments embryos were exposed to a single dose of 3H-TdR between 4.5 and 8.5 days of incubation. However, in this experiment, unlike the labelling index experiments, embryos were allowed t o develop until 12.5 days of incubation. The usual histological and autoradiographic procedures were followed except that sectioning was at 9 rather than 7 pm and that hemotoxylin and eosin replaced the Feulgen reaction and Fast Green FCS counterstain. Sectioned material was examined for the percentages of neurons and nonneurons that were labelled in brachial and nonbrachial ganglia. Two animals were examined for each exposure period. Relative ganglionic sizes To compare the relaiive sizes of brachial and nonbrachial ganglia, every third section of a ganglion was projected onto paper and drawn with the aid of a microscope fitted with a drawing tube. Magnification factors were the same for all ganglia. All tracings for a ganglion were then cut out and weighed. The total weights of these paper cutouts provided an estimate of the relative sizes of the various ganglia. RESULTS

The histogenesis of the dorsal root ganglia has been summarized by Pannese (‘74). Here it is necessary only t o re-emphasize the differences between the LV and MD regions of the ganglia. These differences first become apparent around 5 to 5.5 days of incubation when LV neurons begin to take on the characteristics of the “bipolar” or “primitive” neuroblasts of Tennyson (’65) or Pannese (’68). Mediodorsal neurons do not begin these processes until around 7.5 days. By 9.5 days our preparations show LV neurons to be quite

large with faintly staining nuclei which have diameters of up to 12 pm. The MD neurons, however, still appear less mature. Although their nuclei have begun to take on a reticular appearance seen in LV neuronal nuclei at 8.5 days, they remain smaller and more darkly staining. The LV region also shows an earlier appearance of nonneuronal cells. Nonneuronal cells are first observed in the LV region at 6.5 days of incubation. By 9.5 days they are quite numerous and have begun to surround the neuronal cell bodies. In MD region, however, nonneurons are not readily apparent until 9.5 days. Proliferative activity Labelling indices Four and one-half days The mean values 2 the standard error of the mean (s.e.m.1 for the labelling indices of ganglia of 4.5-day embryos are given in table 1. The means were determined for the labelling indices of all brachial ganglia (14-16)as a group and for those of all nonbrachial ganglia (12, 13, and 17) as another group. The results show that a t 4.5 days of incubation the labelling indices are quite high, 12-14%, in all ganglia. No differences between brachial and nonbrachial ganglia are apparent. Lateroventral region By 5.5 days, when separate tabulations were first made for LV and MD regions, the labelling indices of the LV region had declined to 7-9% (9%,Exp. 1; 7-8%,Exp. 2; fig. 1). These values decline further, to approximately 4%, by 6.5 days. This is true in both brachial and nonbrachial ganglia. After 6.5 days, however, pronounced differences in LV labelling indices become apparent between brachial and nonbrachial ganglia. In brachial ganglia the labelling indices rise slowly from approximately 4%a t 7.5 days, to 5.5%by 9.5 days. In the nonTABLE 1

Mean labelling fLI) and degeneration fDI) indices 2 standard error ofthe mean ofganglia in 4.5-day embryos LI

DI

Nonbrachial

Exp. 1 Exp. 2

14.121.5(10) 12.920.7(12)

0.420.1 (12) 0.5rt0.1 (12)

Brachial

Exp. 1 Exp. 2

14.1?1.3(15) 12.2e0.4112)

0.320.1 (15) 0.520.1 (12)

Ganglia

I

Number of ganglia examined.

731

CHICK DORSAL ROOT GANGLIA DEVELOPMENT a. LV 14

b.

-

-

12-

-.-

10

5

p I6

=

4-

2-

MD

\

\-: -*.

--.-e

I

,p-

-----

’\

~

I

I

-

L-----

- _--

i

I

-

1412-

-

10-

“,

.c

1 6

---

m

‘*---

2I

I

i

1

I

I

1

developing neurons, neuroblasts, and glioblasts. To do so a t 6.5 days is quite difficult in Feulgen-stained material and would be a very subjective process. The only cells specifically categorized were “obvious nonneurons.” These Mediodorsal region are small cells with triangular, darkly stained In contrast, the MD labelling indices show nuclei of approximately 5 by 2 by 2 pm which little difference between brachial and nonbra- first appear in our material at 6.5 days of inchial ganglia (fig. 1).All MD labelling indices cubation. Table 2 shows that with only small excepfall from high levels of 11-12%(Exp. 1)or 9.5% (Exp. 2) at 5.5 days to low values of approx- tion cells present in the dorsal root ganglia a t imately 1%at 7.5 days. There may be a slight 6.5 days have remained in the proliferative cytendency to increase again by 9.5 days, espe- cle through 3.5 days of incubation: less than 1%of the 6.5-day ganglionic cells were not lacially in brachial ganglia. belled by an exposure to 3H-TdRbeginning at Birthdate studies 2.5 or 3.5 days. (It is because of such small To define the events responsible for these la- numbers of unlabelled cells that only unlabelling index patterns, further studies were belled cells were actually counted in the emundertaken to establish the birthdates of cells bryos in which exposure t o 3H-TdRwas begun prior to 4.5 days of incubation.) Between 3.5 included in the labelling index studies. The results of the 6.5-day cumulative label- and 4.5 days the large-scale withdrawal of ling studies are given in table 2. In these stud- cells from the proliferative cycle begins. This ies no effort was made to distinguish between is especially apparent in the LV regions where brachial ganglia, however, the labelling indices continue to fall sharply to approximately 2%a t 7.5 days. These nonbrachial values than also rise, to approximately 4% by 9.5 days.

732

VIRGINIA MCMILLAN CARR AND SIDNEY B. SIMPSON, JR. TABLE 2

Cumulative labelling studies Exposure period

Ganglia and region

Number sections examined

Total number cells counted

Number cells unlabelled

idaysi

2.5-6.5

LV MD

3.5-6.5

LV MD

4.5-6.5

LV MD

5.5-6.5

LV MD

1

Percentage cells unlabelled

Percentage total cells that are unlabelled nonneurons

-

31 9 4 1

-

0 0 0 0

-

33

-

0

41

-

5

-

0

Cervical Brachial Cervical Brachial

23 27

4,526 2,542

644 504

14.2 19.8

0

23 27

3,851 2,488

112 58

2.9 2.3

0

Cervical Brachial Cervical Brachial

17 32 17 32

3,874 3,781 3,318 4,047

1,813 2,013 605 1,050

46.8 53.2 18.2 26.0

2.4 3.9 0.3 0.2

-2

Cervical Brachial Cervical Brachial

21 13 21 13

Cervical Brachial Cervical Brachial

-3

-

-

41 -

3

0

0

One emhryo per exposure period. Total cell number not counted for 2.5-to 6 5-OT 3.5- to 6.5-day exposure periods. See text. Cervical reeon of embryo not included in sections

approximately 14%of the cells in cervical and 20% in brachial ganglia remain unlabelled when 3H-TdR exposure is begun a t 4.5 days. Mediodorsal cells also begin t o cease proliferative activities at this time, but many fewer have done so than in the LV regions. These trends continue over the next day. By 5.5 days approximately 50%of the LV and 20%of the MD cells present at 6.5 days have ceased proliferating. Again, slightly greater proportions of LV and MD cells appear to have withdrawn from the mitotic cycle in brachial than in cervical ganglia. The data also indicate that it is between 4.5 and 5.5 days of incubation when the first obviously nonneuronal cells cease proliferating. These nonneuronal cells constitute a rather small proportion of the unlabelled 6.5-day cells (2.5%for the cervical and 4% for brachial LV population and less than 0.5%for MD populations). However, as development proceeds this proportion of unlabelled cells that are nonneurons continues to increase, as is apparent in the cumulative labelling studies in which the embryos were sacrificed at 8.5 days (fig. 2 ) . Again, the proportion of unlabelled cells is slightly higher in LV than in MD regions and in brachial than in cervical ganglia. The fact that the first nonneurons do not

C

307

DAY Fig. 2 Cumulative labelling studies in which embryos were sacrificed at 8.5 days of incubation. Percentages of cells counted in ganglia of 8.5-day embryos that have remained unlabelled nonneuronal cells during continuous exposure to 3H-TdRbeginning on 5.5, 6.5, or 7.5 days of incubation. One embryo was examined for each exposure period, 0-0, LV region of cervical ganglia; 0 - - - - - 0, LV region of brachial ganglia; 0-0, MD region of cervical ganglia: 0 - - - - - 0 , MD region of brachial ganglia.

withdraw from the proliferative cycle until after 4.5 days makes it appear quite likely that the earliest unlabelled cells observed in these studies are neurons, which are born between 3.5 and 4.5 days of incubation in both LV and MD regions. In the second set of birthdate determina-

CHICK DORSAL ROOT GANGLIA DEVELOPMENT

733

a.

b. 5 07

Fig. 3 Percentage of ganglionic neurons (a) or nonneurons (b) labelled in 12.5-day embryos given a single dose of 3H-TdRat 4.5, 5.5, 6.5, 7.5, or 8.5 days of incubation. Two embryos were examined for each exposure period. Note that in the later time periods a few neurons continued to become labelled: approximately 0.5%in the LV and MD regions of brachial ganglia between 7.5 and 12.5 days and 0.1 and 0.3%in LV regions of nonbrachial and brachial ganglia, respectively, between 8.5 and 12.5 days. 0-0, LV region of nonbrachial ganglia; 0 - - - - C ,LV region of brachial ganglia; 0-0, MD region of nonbrachial ganglia; 0 - - - - 0 , MD region of brachial ganglia.

tions embryos given a single dose of 3H-TdRat 4.5 to 8.5 days were allowed to continue their development to 12.5 days of incubation. By this age neuronal maturation is well underway in the ganglia so that neuronal and nonneuronal cells can be readily distinguished. A single rather than repeated doses of 'H-TdR was used to avoid unnecessary damage to the embryos. Because delivery of 3H-TdRafter 8.5 days requires intravenous injection, repeated injections until 12.5 days would have led to serious impairment of the chorioallantoic circulation. Such repeated injections would also have eventually led to accumulation of fairly large amounts of labelled material and, therefore, to an increased risk of radiation effects in the embryos. Autoradiographs were examined for the percentages of neuronal and nonneuronal cells that were labelled. The results (fig. 3a) indicate that proliferation of neuronal precursor cells has ceased almost entirely by 6.5 days in the LV regions of brachial ganglia and slightly later in the LV regions of nonbrachial ganglia. In the MD regions this proliferation has essentially ceased by 7.5 days. However, in each embryo labelled after 6.5 or 7.5 days there continued to be a few neurons, primarily in the LV regions of brachial ganglia, that had incorporated 3H-TdR. These may represent a very small population of cells, 0.6% or less, that remains in the proliferative cycle to a later time. Proliferative activity of nonneuronal cells

was found in these studies to remain elevated through 4.5 and 5.5 days (fig. 3b). From this time through 7.5 days there was a slight reduction of cells incorporating label, especially in the MD regions of the ganglia. However, a rise in labelling begins between 7.5 and 8.5 days, the last age a t which 3H-TdRwas administered. This rise is more pronounced in the LV than MD regions and in brachial than nonbrachial ganglia. Degenerative activity Characteristics of degeneration Cells categorized as degenerating included both pycnotic cells and macrophages. When stained by the Feulgen procedure, typical pycnotic cells showed one or two large Feulgenpositive chromatin globules (fig. 4a). Occasionally, other cells were observed which, instead of a single large globule, appeared t o have several smaller ones dispersed within the nucleus (fig. 4c). There is little doubt that such pycnotic cells were in fact degenerating. When we examined such cells by electron microscopy, they were identical in appearance t o the degenerating neurons described by Birks and Welson ('71), O'Connor and Wyttenbach ('74), and Pannese ('74). The pycnotic cells appeared to be ingested by macrophages (fig. 4d). Many large macrophages could be seen in each ganglion at the height of the degenerative activity. These macrophages were quite lobulated and filled with Feulgen-positive globules of various

734

VIRGINIA MCMILLAN CARR AND SIDNEY B. SIMPSON. JR.

Fig. 4 a region. X b c d parent. X

Micrographs of degenerating cells in dorsal root ganglia. Degenerating cells with large chromatin globules (g) and with rim of condensed chromatin (r),LV 1,300. Degenerating cell in MD region. x 1,700. Degenerating cell containing several small chromatin globules. X 1,700. Macrophage containing remnants of several degenerating cells. Macrophage nucleus (arrow) is ap1.700.

sizes, which were most likely the remains of ingested degenerating cells. Often, especially a t the height of degenerative activity, the macrophages were so engorged that their large nuclei appeared to be squeezed between several debris-filled lobules. Since it is impossible to determine how many degenerating cells may have been ingested by a single macrophage, each debris-laden macrophage was counted as a single degenerating cell. Thus, our counts may, if anything, underestimate the true degenerative activity in the ganglia. Occasionally a degenerating cell was seen having no chromatin globules a t all but rather a thin Feulgen-positive rim around all or most of the nucleus (fig. 4a). Such rimmed nuclei represent an earlier stage of degeneration than the globular structures (Glucksmann, '51; Bellairs, '61; Scharrer, '66; O'Connor and Wyttenbach, '74). Their more infrequent occurrence than the globular structures probably reflects a shorter duration of the rimmed than the pycnotic stage of degeneration. It should be noted that degenerating cells in the LV and MD regions were identical in ap-

pearance in our preparations (cf. figs. 4a and b). This observation stands a t variance with those of Levi-Montalcini and Levi ('43)and Hamburger and Levi-Montalcini ('49) who found no evidence of degeneration in the MD region but, rather, only a slowly progressing neuronal atrophy after seven days. The primary difference we observe between LV and MD regions is that the MD regions tend to have a predominance of the degenerating cells containing one to two large Feulgen-positive globules whereas the LV regions have the degenerating cells containing several small globules as well. Macrophages also occur with lower frequency in the MD than in the LV regions, a t least a t the stages examined in' our studies. Degeneration indices Four and one-half days At 4.5 days (table 1) the degeneration indices are quite low, 0.5%or less, in all ganglia.

Lateroventral region In the LV region degenerative activity is

735

CHICK DORSAL ROOT GANGLIA DEVELOPMENT a. L V 14

b. M D

-

12-

-m 10

z

8-

1 I 14

I

1

I

9

10

-

12-

R

c

10

-

8

-

U

z

m

64-

2-

5

6

7

8

9

10

5

6

7

8

DAYS

Fig. 5 Mean degeneration indices of LV and MD regions of nonbrachial and brachial ganglia from 5.5- to 9.5-day embryos. Each data point represents the average of values from 14 to 26 ganglia. 0-0, Exp. 1; 0 -----0, Exp. 2.

well under way by 5.5 days (fig. 5a). In nonbrachial ganglia 5.5 days marks a peak in this activity, and the degeneration indices are 7.4% and 5.0% in Exp. 1 and Exp. 2, respectively. The degeneration indices fall sharply from this peak. In Exp. 1 they are 2.8%a t 6.5 days, where they level off with only slight further decline. In Exp. 2 the degeneration indices fall continuously from 5.5 days. By 9.5 days of incubation they have declined to 0.7%. In brachial ganglia, while degelrerative activity is much lower than in nonbrachial ganglia, it is still quite noticeable. Small peaks of activity are seen in the degeneration indices, 2.4%a t 5.5 days in Exp. 1 and 2.3%a t 6.5 days in Exp. 2. The degeneration indices decline from these values to 0.7-0.8% by 9.5 days in both experiments.

erally higher in nonbrachial ganglia. By 9.5 days, MD degeneration indices have risen to 4.0% (Exp. 1) and 2.8% (Exp. 2) in nonbrachial ganglia and to 2.5% (Exp. 1)and 1.8%(Exp. 2) in brachial ganglia.

Relative sizes of the ganglia The relative sizes of ganglia 12, 15, and 17 are given in table 3. These data show that differences in the sizes of the various ganglia are established as early as 4.5 days of incubation. At this time ganglion 15 is already twice as TABLE 3

Mean weight (gms) 2 s.e.m. ofpaper cutouts of projections of ganglia ' Age

Ganglion 12

Ganglion 15

Ganglion 17

0.11* 0.01 0.36k0.02 0.68'0.03 0.89k0.04

0.22' 0.03 0.72*0.07

0.16' 0.06 0.47'0.02

1.70'0.14

0.791k0.05

3.02'0.22

1.17k0.37

(days)

Mediodorsal region In the MD region (fig. 5b) degeneration indices appear to rise slowly over the entire period examined in both and nonbrachial ganglia. However, the values are gen-

4.5

5.5

7.5 9.5

' Mean values based on data from three embryos except 4.5-day ganglion 15 (4 embryos) and 4.5-day ganglion 17 (2 embryos).

736

VIRGINlA MCMILLAN CARR A N D SIDNEY B. SIMPSON, JR

large as ganglion 1 2 while ganglia 17 is of intermediate size. Hamburger and Levi-Montalcini ('49) also observed size differences between cervical and brachial ganglia, as early as four days. Since 4.5 days is prior to the onset of any degenerative or proliferative differences between ganglia, such size differences must indicate that the original amounts of neural crest material which condensed to form the ganglia varied depending on the location of the ganglia. Thus, regional difference in ganglion size may be a function of the original quantity of crest material as well as differential proliferative and degenerative activities. Also apparent in the data is the rapidity with which changes in the degenerative and proliferative activity, coupled with cellular growth, are reflected in the sizes of all three ganglia. DISCUSSION

Our study represents the first autoradiographic analysis of cell proliferation as well as the first detailed quantitative examination of cell degeneration in the early development of the chick dorsal root ganglia. The labelling indices demonstrate that proliferative activity is quite elevated through 4.5 to 5.5 days in the brachial ganglia as well as in nonbrachial ganglia 12,13, and 17. In the LV regions this activity declines precipitously until 6.5 days in brachial and 7.5 days in nonbrachial ganglia. In both brachial and nonbrachial ganglia it increases again after 7.5 days. In the MD regions the proliferative activity falls to very low levels by 7.5 days. However, there may be a slight rise by 9.5 days, especially in brachial ganglia. These findings differ somewhat from those of the earlier study of Hamburger and LeviMontalcini ('49) in which proliferative activity was reported to rise continually to a peak at five to six days of incubation. However, Hamburger and Levi-Montalcini's mitotic activities represented the absolute numbers of mitotic cells in a ganglion while our labelling indices represent the percentage of all cells counted within a ganglion that are labelled and presumably part of the proliferative population. Thus, the discrepancies between the two studies are most likely due to the different methods of expressing the data. A study by Yates ('61) makes this clear. Yates found that the absolute numbers of mitotic figures in the chick dorsal root ganglia did indeed peak a t five to six days of incubation. However, when

the data were expressed as the ratio of mitotic to nonmitotic cells, a ratio not unlike our labelling index, Yates found that proliferative activity actually remained high through four days and then declined continuously. These seemingly different results depending on the method of data presentation reflect the fact that until the onset of degenerative activity at 4.5 to 5.5 days the ganglia are continually increasing in cell number. Because many of the new cells are still mitotically active, the absolute number of mitotic cells in a ganglion will also increase until substantial numbers of cells begin to cease proliferating. Thepercentage of total cells that are proliferating, however, need not change until this time. Our birthdate studies indicate t h a t this largescale cessation of proliferation occurs from 4.5 to 6.5 or 7.5 days of incubation, the same time as the decline in the numbers of mitotic figures reported by Hamburger and Levi-Montalcini ('49). A similar explanation can be used to clarify t h e discrepancy between Hamburger and Levi-Montalcini's ('49) report of greater praliferative activity in limb than nonlimb ganglia and ours of no difference in labelling indices between them: although the limb ganglia would have a larger number of mitotic cells than the smaller nonlimb ganglia, the percentage of the total cells proliferating in each would be the same. A more serious discrepancy between our results and those of Hamburger and Levi-Montalcini ('49) lies in our finding of moderate amounts of labelling a t 9.5 days, which is after the time when proliferative activity was reported to have ceased in the earlier paper. Not only do our data fail to show such a cessation of proliferative activity, but they may actually indicate a slight increase in this activity from 7.5 days on (fig. 1).Similar observations to ours were made by Yates ('61), who reported the presence of mitotic figures through 12 days of incubation, the oldest age he examined. Identification of the proliferative populations represented by our labelling indices is provided by the birthdate studies. These studies demonstrate that, for the most part, largescale neuronal production occurs between 4.5 and 6.5 days in the LV and between 4.5 and 7.5 days in t h e MD regions of the ganglia. Prior to 6.5 or 7.5 days, therefore, the labelling indices must represent the proliferation of both neuronal and nonneuronal cell precursors. After

CHICK DORSAL ROOT GANGLIA DEVELOPMENT

these times the labelling indices must be almost entirely nonneuronal in nature. The birthdate data also suggest that the marked declines in the labelling indices after 4.5 or 5.5 days as well as the abrupt terminations of these declines are reflections of the cessation of neuronal production. The timing of these events corresponds exactly. Even in the nonbrachial ganglia, where cessation of large-scale LV neuronal production occurs slightly later than in brachial ganglia (fig. 3a), the decline of the nonbrachial LV labelling indices appears to be more gradual than in the brachial ganglia, and the inflection points in the labelling index graphs do not occur until 7.5 days. These findings corroborate conclusions of Yates ('61) made on the basis of relative amounts of RNA in various ganglionic cell types. Yates found t h a t cells which stain intensely for RNA, and were, therefore, assumed to be neuroblasts, cease proliferation by seven or eight days of incubation. All proliferating cells observed by Yates a t later times stained much less intensely for RNA and, therefore, were assumed to be supporting elements. The large-scale production of the larger LV neurons before that of the smaller MD neurons is an observation that is in agreement with observations made in other animals and other neural systems. Lawson et al. ('74) reported a similar pattern of neuronal production in the dorsal root ganglia of rat embryos. Using pulse labelling, they found that large neurons, presumably corresponding to the LV cells of the chick, were produced between 10 and 15 days and small ones between 11and 17 days of gestation. The same phenomenon has also been reported for the mesencephalic V nucleus (Rogers and Cowan, '73) and the optic tectum (LaVail and Cowan, '71) of chicks. Other systems are described in Rogers and Cowan ('73). While our birthdate studies demonstrate the cessation of large-scale neuronal production by 6.5 to 7.5 days of incubation, they also reveal the possibility of continued production of a small number of neuronal precursors a t later times, especially in LV regions of the ganglia. Scott's ('77) recent findings of increased numbers of neurons in 8- to 12-day chick dorsal root ganglia cultured in the presence of high potassium concentrations may reflect the presence of such proliferating cells. However, the small numbers of labelled cells we observe would suggest that these cells

737

probably do not constitute a major source of ganglionic neurons. The patterns of nonneuronal proliferative activity observed in our studies correlate well with the quantitative relationships between satellite cell number and neuronal volume and metabolic function found in dorsal root ganglia by Pannese ('60, '64) and co-workers (Pannese e t al., '75). Both our labelling indices and our birthdate studies carried to 12.5 days show that after the end of large-scale neuron production a t 6.5 to 7.5 days, proliferation of nonneuronal cells is higher in LV than in MD regions. Presumably, this differential nonneuronal proliferative activity is a response to the earlier production, greater maturity, and larger size of LV than MD neurons. The cumulative labelling studies reflect this relationship also. In those studies the percentage of unlabelled cells that are nonneurons is greater in the LV than MD regions. The amount of nonneuronal proliferative activity appears to be greater in brachial than nonbrachial ganglia also, although the difference is less than i t is between LV and MD regions. This relationship can be seen in each of our measurements of proliferative activity. It probably reflects a nonneuronal response to the greater neuron numbers in brachial than nonbrachial ganglia that results from differential neuronal degenerative activities in the two sets of ganglia. Degenerative activity is shown by our degeneration indices to be absent from the chick dorsal root ganglia until after 4.5 days of incubation. This is in agreement with the original observation of Hamburger and Levi-Montalcini ('49) that the onset of degenerative activity occurs between four and five days. The degeneration indices also confirm the earlier report that once underway this degenerative activity rapidly increases to a peak a t five to six days of incubation. Our degeneration study does not confirm Hamburger and Levi-Montalcini's ('49) observation t h a t such activity is confined to the LV regions of nonbrachial ganglia. Our findings make clear the presence of degenerating cells in both the LV regions of brachial ganglia and in the MD regions of brachial and nonbrachial ganglia. Chu-Wang and Oppenheim ('78) have recently reported the presence of degenerating cells in chick lumbar ganglia as well. The finding of measurable degenerative activity in brachial ganglia indicates that these ganglia, like those a t nonbrachial levels, must

738

VIRGINIA MCMILLAN CARR AND SIDNEY B. SIMPSON, JR. D a y of Incubation

2 I

3 -

4 I

~

6

5 I

'

I

7 ~

I

9

8 '

II

'' l

i

Neural crest condensation' "Indifferent cells" ** X t .

Primitive

neuroblasts

* I nte rmediote

neuroblasts " * *

Large-scale

LV neuron production

Large-scale

MD neuron production

Nonneuronal cell

production

Neuronal growth"' Degenerotion

LV degeneration peok 'Werton('63) "Pannese('680) *.I

Levi-Montolcini and Levi ('43)

Fig. 6 Chronology of the major events in the early development of chick dorsal root ganglia.

undergo some overproduction of neurons that necessitates cell loss for final adjustment of neuronal numbers to peripheral requirements. Given the lower degeneration indices we find in brachial ganglia, the degree of overproduction must obviously be much lower in these than in nonbrachial ganglia. The import of Hamburger and Levi-Montalcini's ('49) observations, however, was that such overproduction did not occur a t all. Nevertheless, cellular degeneration has now been reported to occur in limb level ganglia of other animals (Prestige, '65; Hughes, '73), findings which support the observations in the chick. Our results also disagree with Hamburger and Levi-Montalcini's ('49) report that degenerating cells are no longer present in the ganglia after seven days. Degenerative activity is quite apparent in our material a t later times. In fact, in the MD regions, the level of such activity seems to be in the process of increasing as late as 9.5 days of incubation. Finally, we can also not confirm the earlier report of atrophy in the MD regions a t later stages of development. This, however, may be related t o the different staining methods used in the two studies and remains to be examined thoroughly. The results of our labelling and degeneration studies can now be incorporated along with the findings of earlier investigators to

provide a revised chronology of the major events in early development of the chick dorsal root ganglia (fig. 6). We feel this chronology is a more accurate reflection of developmental events than has existed heretofore. Unfortunately, our studies concluded a t 9.5 days under the assumption that the major proliferative and degenerative events had ceased by that time. This has now been shown to be an erroneous assumption, and examination of events at ages past 9.5 days must now be undertaken. ACKNOWLEDGMENTS

The authors wish to express their appreciation to Mrs. Nan Simpson for her assistance with some of the statistical evaluation of data. The authors are also grateful to Doctors V. Hamburger and M. Hollyday for reviewing the text and for their helpful comments and suggestions pertaining thereto. This work was supported by NINDS Grant 5 R01 NS 0960 and NIH Biomedical Sciences Support Grant SOS RR07028. LITERATURE CITED Bellairs, R. 1961 Cell death in chick embryos as studied by electron microscopy. J . Anat. (London), 95: 54-60. Birks, R. I., and P. R. Weldon 1971 Formation of crystalline ribosomal arrays in cultured chick embryo dorsal root ganglia. J. Anat. (London), 109: 143-156. Carr, V. McM., and S. B. Simpson, Jr. 1975 Peripheral effects on early development of chick spinal ganglia. Neuroscience Abstracts, 1: 749.

CHICK DORSAL ROOT GANGLIA DEVELOPMENT

739

1964 Number and structure of perisomatic 1978 Proliferative and degenerative events in t he satellite cells of spinal ganglia under normal conditions early development of chick dorsal root ganglia. 11. Reor during axon regeneration and neuronal hypertrophy. sponses to altered peripheral fields. J. Comp. Neur., 182: Z. Zellforsch., 63: 568-592. 741-756. 1968 Developmental changes of the endoplasmic Chu-Wang, I.-W., and R. W. Oppenheim 1978 Cell death of reticulum and ribosomes in nerve cells of the spinal motoneurons in the chick embryo spinal cord. I. A light ganglia of the domestic fowl. J. Comp. Neur., 132: and electron microscopic study of naturally occurring and 331-364. induced cell loss during development. J. Comp. Neur., 1974 The histogenesis of the spinal ganglia. Adv. 177: 33-58. Anat. Embryol. Cell Biol., 47 (5): 1-97. Deitch, A. D., D. Wagner and R. M. Richart 1968 CondiPannese, E., R. Ventura and R. Bianchi 1975 Quantitative tions influencing t h e intensity of t he Feulgen reaction. relationships between nerve and satellite cells in spinal J. Histochem. Cytochem., 16: 371-379. ganglia: a n electron microscopical study. 11. Reptiles. J. Gliicksmann, A. 1951 Cell death in normal vertebrate onComp. Neur., 160: 463-476. togeny. Biol. Rev., 26: 59-86. Prestige, M. C. 1965 Cell turnover in the spinal ganglia of Hamburger, V., and H. L. Hamilton 1951 A series of normal Xenopus laeuis tadpoles. J. Embryol. Exp. Morph., 13: stages in t h e development of the chick embryo. J. Morph., 63-72. 88: 49-92. Rogers, L. A., and W. M. Cowan 1973 The development of Hamburger, V., and R. Levi-Montalcini 1949 Proliferation, the mesencephalic nucleus of the trigeminal nerve in the differentiation, and degeneration of the spinal ganglia of chick. J. Comp. Neur., 147: 291-320. the chick embryo under normal and experimental condiScharrer, B. 1966 Ultrastructural study of regressing protions. J. Exp. Zool., 111: 457-501. thoracic glands of blattarian insects. Z. Zellforsch., 69: Hughes, A. 1973 The development of dorsal root ganglia 1-21. and ventral horns in t he opossum. A quantitative study. Scbultze, B. 1969 Autoradiograpby at the cellular level. In: J. Embryol. Exp. Morph., 30: 359-376. Physical Techniques in Biological Research. Vol. 111,Part Jacobson, M. 1970 Developmental Neurobiology. Holt B. A. W. Pollister, ed. Academic Press, New York, Rinehart and Winston, New York, p. 270. pp. 28-29. LaVail, J.,and W. M. Cowan 1971 The development of the Scott, B. S. 1977 The effect of elevated potassium on the chick optic tectum. 11. Autoradiographic studies. Brain time course of neuron survival in cultures of dissociated Res., 28: 421-441. dorsal root ganglia. J. Cell. Physiol., 92: 305-316. Lawson, S.N.,K. W. T. Caddy and T. J. Biscoe 1974 DevelSidman, R. L. 1970 Autoradiographic methods and princiopment of rat dorsal root ganglion neurons. Studies of cell ples for study of the nervous system with thymidine-HJ. birthdays and changes in mean cell diameter. Cell. Tiss. In: Contemporary Research Methods in Neuroanatomy. Res., 153: 399-413. W. J. H. Nauta and S. 0. E. Ebbesson, eds. Springer-VerLevi-Montalcini, R., and G. Levi 1943 Recherches quanlag, New York, pp. 252-274. titatives sur la marche du processus de differentiation Tennyson, V. M. 1965 Electron microscopic study of the dedes neurons dan les ganglions spinaux de l’embryon de velopingneuroblast of the dorsal root ganglion of the rabpoulet. Arch. Biol., 54: 198-206. bit embryo. J. Comp. Neur., 224: 267-318. OConnor, T. M., and C. R. Wyttenbach 1974 Cell death in t h e embryonic chick spinal cord. J. Cell Biol., 60: 448-459. Weston, J. A. 1963 A radiographic analysis of the migration and localization of trunk neural crest cells in the Pannese, E. 1960 Observations on the morphology, subchick. Devel. Biol., 6: 279-310. microscoDic str u c tu r e and biological of -DroDerties . Yates. R. D. 1961 A studv of cell division in chick embrvsatellite cells (S.C.) in sensory ganglia of mammals. onic ganglion. J. Exp. Neur., 147: 167-181. Z. Zellforsch., 52: 567-597.