/. Embryol. exp. Morph. Vol. 31, 1, pp. 99-121, 1974
Printed in Great Britain
99
Morphogenesis of the trigeminal
mesencephalic nucleus in the hamster: cytogenesis
and neurone death
By KEITH E. ALLEY 1
From the Departments of Anatomy and Oral Anatomy,
Colleges of Medicine and Dentistry, University of Illinois,
Chicago, Illinois
SUMMARY
The chronology of cellular development in the trigeminal mesencephalic nucleus of the
hamster has been studied by light and electron microscopy. Major developmental features
are compressed into a two-week period that spans pre- and postnatal maturation. Special
emphasis is focused on the sequence of developmental events, with the following points
analyzed in this report: cell proliferation, cytological and axonal development, cell death,
and cluster formation.
These centrally located primary sensory neurones seem to arise from a separate germinal
population found in the alar plate of the midbrain and pons. The neurones are remarkable for
their precocious formation and characteristic appearance. In hamster embryos they are first
detected when the neural tube is only a couple of cell layers thick. Prenatally, the neurones
grow rapidly and have doubled in diameter by birth. Axons form shortly after the cells are
formed. This is marked by a sudden upsurge in cytoplasmic tubules and filaments in one end
of the soma. Postnatally, cellular growth continues at a reduced rate. However, shortly after
birth many small spines protrude from the cell body, thus increasing the surface area available
for nutritive exchange. Myelination of the axon commences 6 days after birth.
Cell counts on a series of pre- and postnatal animals detected a considerable overproduction
of neurones in the mesencephalic nucleus. Prior to birth there is a rapid elimination of almost
50 % of these cells. Degenerating cells were identified in both light and electron microscopes.
The temporal relationship between peripheral innervation, cell death and synaptogenesis
suggests a developmental mechanism that provides some flexibility in the formation of synaptic connexions in the mesencephalic nucleus.
Clusters of 2-4 tightly packed neurones are a characteristic feature of the developing and
adult mesencephalic nucleus. Specialized junctions connect adjacent neurones and are thought
to provide the basis for electrotonic coupling within the nucleus. Macula adhaerens plaques
were present in both pre- and postnatal hamsters. However, close appositions were only
identified postnatally.
1
Author's address: Division of Neurobiology and Department of Oral Biology, University
of Iowa, Iowa City, Iowa 52242, U.S.A.
7-2
100
K. E. ALLEY
INTRODUCTION
During the past century the structure of the mesencephalic nucleus of the
trigeminal nerve has been intensively studied in a wide variety of vertebrate
species. From these investigations it has become evident that this nucleus has
many features that are not characteristic of the primary sensory neurones
located in either dorsal root or cranial ganglia. Notable among these special
aspects are their central location (Weinberg, 1928), the presence of synaptic
endings on the soma (Hinrichsen & Larramendi, 1968, 1970; Alley, 1973), and
the possible existence of electrotonic coupling between small clusters of these
neurones (Hinrichsen & Larramendi, 1968, 1970; Baker & Llinas, 1971).
Functionally the mesencephalic nucleus is believed to mediate proprioceptive
input from the jaw apparatus (Corbin & Harrison, 1940; Jerge, 1963). Peripheral processes of these neurones terminate in annulospiral and flower spray
endings in the spindle receptors of the masticatory muscles (Szentagothai, 1948;
Kawamura, Funakoshi, Tsukamoto & Takata, 1959; Jerge, 1963), and in tension receptors located in the periodontal ligament of the teeth (Corbin &
Harrison, 1940; Jerge, 1963; Kidokoro, Kubota, Shuto, Sumino, 1968a, b).
Although both the structure and function of the adult mesencephalic nucleus
are well documented, little is known about the development of these neurones.
Their large size, lack of dendrites, and characteristic location in the midbrain
and pons make them ideal subjects for both quantitative analyses of neuronal
number during ontogenesis and electron microscopic observations of cellular
maturation. Moreover, the dramatic shift of feeding behavior from the rather
stereotyped movements of suckling to the more intricate activity of chewing
suggests a greater role for proprioceptive feedback during mastication. In
postnatal hamsters the timing of these events seems to be correlated with the
maturation of the mesencephalic nucleus.
The purpose of this analysis is twofold: first, it outlines the chronology of
cytogenesis, emphasizing cytoplasmic growth and reorganization, the development of somatic spines, and the formation of the axon. Secondly, the role of
cell death during morphogenesis of the mammalian mesencephalic nucleus is
evaluated by cell counts and histological observation, and the timing of neuronal
death is compared with the onset of synaptogenesis. It was hoped that information on the developmental process might provide additional insight into the
central location of this nucleus.
A previous report of cell death in the mammalian mesencephalic nucleus
was first published in abstract form (Alley & Du Brul, 1972).
MATERIALS AND METHODS
Golden hamsters (Mesocricetus auratus) were procured from Con Olson
Animal Supply of Madison, Wisconsin. All animals were housed in a constant
Morphogenesis of the mesencephalic nucleus
101
temperature, 12-h light-dark environment. A rigorous schedule of timed matings
was employed and at appropriate intervals from 9 days post coitus (p.c.) to
birth the pregnant females were anesthetized with Nembutal and the fetal pups
surgically removed. Postnatal pups were also anesthetized with Nembutal and
collected at 24 h intervals until 15 days of age. Twenty-five-day postnatal (p.n.)
and adult animals were also used. A total of 93 specimens was gathered for
analysis in the light and electron microscopes.
Brains of litter-mates were prepared according to the protocol of Nissl,
Cajal reduced silver, and Golgi (Valverde, 1970) techniques. Material to be
stained with cresylviolet was embedded in paraffin and serially sectioned in
either frontal or sagittal planes. These brains were used to calibrate neuronal
size and to determine cell number during maturation. Further aspects of
somatic and axon developments were examined on the Golgi and reduced silver
preparations. However, considerable difficulty was encountered in impregnating
neurones of the mesencephalic nucleus in fetuses younger than 12 days p.c.
Fixation for electron microscopy was accomplished either by direct immersion or by vascular perfusion. The fixative consisted of a combination of
aldehydes: 2-5% paraformaldehyde, 2-5% glutaraldehyde, 0-001% calcium
chloride, and 0-1 M cacodylate buffer at a final pH of 7-4 (Karnovsky,
1965).
Early fetuses were immersed in cold fixative immediately after severing the
umbilicus, while older fetuses and all postnatal animals were fixed by vascular
perfusion. Each brain was kept overnight in cold fixative. The following day
small pieces of tissue were dissected from the caudal region of the nucleus in the
pons and postfixed in 1 % osmium in 0-1 M cacodylate for two hours. Material
was dehydrated in graded acetone and embedded in Epon-Araldite. Each piece
was oriented so that the sections would be cut in a coronal plane.
Mallory stained sections, 1 /im thick, were used to localize the mesencephalic
nucleus prior to thin sectioning. These thin sections were also of considerable
aid in analyzing cell morphology during development. Furthermore, semi-serial
series of 1 jtim plastic sections through the extent of the mesencephalic nucleus
were prepared for the youngest embryos studied.
Thin sections were cut on a Porter-Blum MT-2 ultratome, double stained
with uranyl acetate and lead citrate and viewed on a Hitachi 7-S electron
microscope.
Serial-sectioned brain stems from 29 animals were used to count the number
of neurones in the developing mesencephalic nucleus. Cell counts were done
according to the selected-section method (Konigsmark, 1970). Every other
section was carefully reviewed and the neurones of the mesencephalic nucleus
counted. Only those cells containing a definite nucleus and nucleolus were
included in the data. Final estimates of the size of the neuronal population were
determined by multiplying the actual cell count by the interval.
Certain difficulties were encountered in identifying accurately the earliest
102
K. E. ALLEY
Fig. 1. Alar plate of the midbrain in 9J days p.c. hamster embryo. Large arrows
indicate the mesencephalic nucleus sandwiched between the mantle layer (top)
and germinal epithelium. Note the mitoses located near the nucleus (Small arrows).
x800.
neurones of the mesencephalic nucleus in the paraffin sections. From 9\ days
p.c. to 10 days p.c. cytoplasmic outlines were indistinct and the intensely stained
germinal cells tended to mask their appearance. Later in the gestation period
care was taken to account for other groups of neurones in the immediate vicinity.
Particular attention was paid to the neurones of the locus coeruleus.
Additional quantitative procedures used in the investigation will be described
in the appropriate sections.
OBSERVATIONS
Initial appearance of the mesencephalic nucleus
The mesencephalic nucleus is conspicuous by its precocious appearance in
the alar plate of the midbrain and pons at 9 \ days p.c. At this stage the neural
tube is characterized by a mitotically active ventricular zone and a thin intermediate zone. The early cells of the mesencephalic nucleus are sandwiched
between these two layers (Fig. 1). On the 1 fim. sections these neurones were
readily distinguished from the adjacent germinal cells by differences in nuclear
and cytoplasmic architecture.
Although the precise primordia of these neurones could not be identified with
certainty, it is interesting to note that in embryos from 9\ days p.c. to \2\ days
p.c. a few mitotic profiles were encountered at some distance from the ventricu-
Morphogenesis of the mesencephalic nucleus
10-5d.p.c.
103
"o
11 d.p.c.
15
15
15
7 10 13 16 19 22 25/im
15 9d.p.n.
7 10 13 16 19 22 25 28/im
7 10 13 16 19 22 25 28/mi
15 11 d.p.n
7 10 13 16 19 22 25 28//m
13d.p.n.
7 10 13 16 19 22 25 28 31 //m
15 15 d.p.n.
1 10 13 16 19 22 25 28 31 34 37 /mi
15 25 d.p.n.
7 10 13 16 19 22 25 28 31 34 37 40/mi
7 10 13 16 19 22 25 28 31 34 37 40 //in
15 Adult
7 10 13 16 19 22 25 28 31 34 3740 43 46 //m
Fig. 2. Histograms of cell diameter in the developing mesencephalic nucleus.
Abscissa - neuronal diameter; ordinate - percentage of cells, d.p.c, Days post
coitus; d.p.n., days post natal.
lar surface in close proximity to the cells of the mesencephalic nucleus (Figs. 1,5).
Even though additional evidence will be required to pin down the origin of
these neurones, the close proximity of the mitotic cells suggests that the
mesencephalic nucleus may arise from an independent primordia within the
neural tube.
Somatic growth and maturation
The timing and rate of cell growth and development were evaluated for both
pre- and postnatal neurones. Indices of growth are portrayed as histograms of
cell diameter (Fig. 2) and as estimates of the surface area and volume of the
104
K. E. ALLEY
Table 1. Quantitation of somatic growth during development of the mesencephalic
nucleus
Age
(days)
10 p.c.
10* p.c.
11 p.c.
12p.c.
13 p.c.
14 p.c.
Newborn
1 p.n.
2 p.n.
3 p.n.
5 p.n.
7 p.n.
9 p.n.
11 p.n.
13 p.n.
15 p.n.
25 p.n.
Adult
Av. diameter
of soma
(/tm)
8-4
10-6
11-4
11 -9
15-6
15-8
17-5
.18-5
191
210
21-2
21-4
20-8
23-3
250
26-8
27-7
341
Surface area
Cytoplasmic
of soma
volume of soma
(/*m2)
(/<m3)
222
353
408
445
764
784
962
1075
1145
1385
1411
1438
1384
1704
1962
2255
2409
3651
310
623
778
858
1985
2030
2757
3315
3580
4836
4990
5128
4187
8180
8567
10176
11243
20569
soma (Table 1). All neuronal measurements were done with a filar micrometer on
the Nissl-stained paraffin sections. The size recorded was the largest diameter of
each profile. At least 100 neurones along the length of the nucleus were measured
at each stage. Data from either frontal or sagittal planes gave comparable
distributions of cell size. It was assumed that this was indicative of a spheroidal
shape for their soma. However, it might also suggest that the neurones lack a
preferential axis of orientation.
Before birth the cells expand very rapidly. At 10 days p.c. each cell averages
less than 9 /an in diameter. Six days later the diameter has doubled as a result
of a tenfold volumetric increase. Postnatally the cell body grows at a much
reduced rate. However, during this period many other factors of importance to
the final maturation of these neurones take place.
Simultaneously with cellular growth there are marked changes in the organization of the cytoplasm which are apparent in both light and electron microscopes.
Early neurones, 9\ days p.c, are irregularly round, oval, or elongated but few
cytoplasmic processes are present. The large nucleus is eccentrically positioned
in the soma and nucleoli are usually flattened against the nuclear envelope. The
remainder of the karyoplasm is granular in appearance and more electron dense
than in the adult (Fig. 3). Many free ribosomes or small polysomal clusters are
evenly distributed in the cytoplasm. In addition, a few strips of granular
endoplasmic reticulum are present, but distinct Nissl bodies were not found.
Morphogenesis of the mesencephalic nucleus
FIGURES
3-5
Fig. 3. Electron micrograph of an immature neurone in the mesencephalic nucleus,
9 | days p.c. x 15000.
Fig. 4. Mesencephalic neurone and process 10J- days p.c. x 8000.
Fig. 5. A cluster of three neurones in the mesencephalic nucleus of an 11 days p.c.
fetus. Note the mitotic cell (asterisk) surrounded by these neurones, x 4300.
105
K. E. ALLEY
FIGURES 6, 7
Fig. 6. Two somatic spines forming palisades adjacent to a small synapse, x 26000.
Fig. 7. An early spine configuration, similar to Fig. 6, in a 5 days p.n. hamster pup.
x 26000.
Mitochondria are also distributed throughout the cytoplasm, while stacks of
Golgi cisternae and vesicles are confined to the perinuclear area. Very few
filaments or tubules could be identified at this stage.
By 10^ days p.c. cell processes are frequently found in the neighborhood of
the mesencephalic nucleus (Fig. 4). The eccentric orientation of the nucleus
confines the cytoplasm to one pole of the cell, and the cell processes arise from
this side of the soma. Ribosomes exist free, as rosettes or polysomal chains, but
there is also an increase in the number of membrane-bound particles. The most
remarkable change is the tremendous upsurge in the number of neurotubules and
filaments at one pole of the neurone.
A major re-ordering of the cytoplasmic organelles begins about 12 days p.c.
(Fig. 5). Most prominent is the increasing basophilia of the peripheral cytoplasm as a result of the condensation of ribosomes (Fig. 5).
Postnatally, free and membrane-bound ribosomes coalesce into distinct
Nissl granules that form a peripheral basophilic ring. Small patches of Golgi
apparatus also migrate into this region, and by 4 days p.n. the cytoplasmic
features of these neurones are generally complete.
Somatic spines
Small thorns protrude from the neurones of the adult mesencephalic nucleus
(Fig. 6). Hinrichsen (1968) suggested that these may be vestiges of the larger
processes sometimes observed during cellular development. However, other
observations on the cerebellum have implicated spine formation with synapto-
107
Morphogenesis of the mesencephalic nucleus
Table 2. Quantitative features of somatic spines on mesencephalic neurones
Age
(days
p.n.)
Perimeter
surveyed
5
7
10
12
25
Adult
1672
1152
755
1014
965
1727
(/im)
Spines
counted
182
128
85
106
76
118
Perimeter
Density Diameter* Perimeter* Calculated increase
of spine
(spines/
of spine number of due to
(jim)
100 /*m) base (/*m)
spines/cell spines (%)
10-9
111
11-3
10-5
7-9
6-8
0-28
0-26
0-27
0-26
0-26
0-27
2-44
2-51
2-24
2-30
2-32
2-35
540
615
690
720
740
920
21
19
20
18
16
14
* Average measurement of 20 spines.
genesis (Larramendi, 1969). In order to determine if either of these was applicable to the mesencephalic nucleus a developmental analysis of these spines was
undertaken.
Prenatally the cell membrane is smooth and devoid of any outpocketings
(Fig. 3, 4, 5). Shortly after birth the cell outline becomes more undulating and
true spines can be recognized 3 or 4 days later, and by 4 or 5 days p.n. the cell
surface forms a smooth boundary punctuated by definite projections (Fig. 7).
The density of spines increases at 5, 7, and 10 days p.n. when the actual surface
density is greater than in the adult (Table 2). Subsequently, the number of
spines per 100/tm of perimeter decreases slightly at 15 days p.n., and by 25
days p.n. the density is approximately equivalent to the adult figure of 6-8
spines per 100 /an of perimeter. This sort of density analysis does not take into
account the continued expansion of the surface membrane, thus a decrease in
surface density could reflect either the gradual growth of the cell surface or the
regression of some spines during maturation. Estimates of the number of thorns
projecting from an average neurone were determined for the mesencephalic
nucleus at 5, 7, 10, 12, 25 days p.n. and in the adult (Table 2). The method
employed was previously used to analyze synaptogenesis in this nucleus and is
described in detail elsewhere (Alley, 1973). Briefly, the calculations include the
following. First, the percentage of the cell perimeter giving off spines was
determined from random cell profiles (spine density x base diameter of spine
4- 100). Subsequently, this percentage was used to compute the actual area of
the surface emitting spines (total surface x percentage of perimeter covered).
Finally, the number of spines per soma was estimated by dividing this portion
of the total surface area by the average cross-sectional area of the spine base
(spinous surface area •*• cross-sectional area of base). According to these
calculations, even though spine density decreases dramatically in the period
from 15 days p.n. to adulthood, the actual number of projections continues to
grow (Table 2). Additional computations that account for the surface added by
108
K. E. ALLEY
Fig. 8. Schematic representation of the adult mesencephalic nucleus portraying the
extended length of the axon from the soma to its bifurcation. Taken from Cajal
(1911).
the spines indicate that the contribution they make to the total area of the soma
declines with age (Table 2).
Although the timing of spine formation and synaptogenesis coincide, they
are only rarely associated in direct synaptic contact. This situation is true
for both adult and developing hamsters. However, it should be noted that
terminals frequently lie near the base of a spine, but the adhesion is on the
soma (Figs. 6, 7).
Development of neuronal polarity
Cajal (1929) and Tennyson (1965) have outlined the morphologic transformations that occur during cellular maturation in spinal and cranial ganglia. These
cells begin as simple apolar neurones, pass through a phase of bipolarity, and
eventually form a stalk from which the central and peripheral processes are
directed.
Neurones of the mesencephalic nucleus emit a single, large process that passes
into the mesencephalic root. Cajal (1911) indicated that in contrast to adult
Morphogenesis of the mesencephalic nucleus
109
Fig. 9. Golgi section and drawing of somatic and axonal structure in the
mesencephalic nucleus 14^- days p.c. x400.
ganglionic neurones, this axon does not bifurcate in the region of the cell body
(Fig. 8), but passes into the pons where many collaterals are given off in the
region of the trigeminal motor nucleus. These provide the basis for the monosynaptic activation of the motoneurones (Szentagothai, 1948; Jerge, 1963;
Kidokoro et aJ. 1968 a).
As a reflexion of this adult configuration young neurones in the mesencephalic
nucleus appear to bypass the true bipolar phases of development. Instead, each
apolar cell may sprout one or more exploratory processes. Eventually one of
these becomes dominant while the others appear to be redundant and are
usually lost later in development. These findings account for the earlier reports
of a decreasing number of multipolar neurones found with advancing maturation of these cells (Pearson, 1949 a, b).
In the hamster this metamorphosis takes place between 9-5 and 14 days p.c.
Neurites first appear at 9-5 or 10 days p.c. when cells with multiple short projections are present. By 12 days p.c. the polarity of most neurones is established,
with the axon emanating from the pole of the cell opposite the nucleus. These
axons have been traced for long distances in 13-, 14- and 15-day fetuses with no
evidence of bifurcation (Fig. 9).
After birth the axon increases greatly in diameter as the axoplasm matures.
Myelin wrappings first appear around the fibers of the mesencephalic root at
6 days p.n.
110
K. E. ALLEY
Table 3. Cell counts in the developing mesencephalic nucleus of the hamster
Age
(days)
Neuronal
count
(right side)
Neuronal
count
(left side)
Total
Adult
Adult
25 p.n.
25 p.n.
15 p.n.
13 p.n.
11 p.n.
9 p.n.
8 p.n.
7 p.n.
6 p.n.
5 p.n.
4 p.n.
3 p.n.
2 p.n.
1 p.n.
1 p.n.
Newborn
Newborn
15 p.c.
14 p.c.
13 p.c.
13 p.c.
121 p.c.
12J- P.c.
12 p.c.
Ill p.c.
11 p.c.
10 p.c.
794
984
850
904
704
916
840
820
950
984
804
832
836
852
974
794
864
844
896
1024
1164
1128
1190
1454
1422
1448
1592
1404
1166
868
920
824
918
738
916
808
800
860
946
824
840
848
894
1072
802
846
856
884
1056
1186
1284
1134
1508
1434
1596
1424
1436
1076
1662
1904
1674
1822
1442
1832
1648
1620
1910
1930
1628
1672
1684
1746
2046
1596
1710
1700
1790
2080
2350
2412
2324
2962
2856
3044
3016
2840
2242
Cell death in the developing mesencephalic nucleus
Preliminary data from a small series of pure-bred dogs suggested that many
more neurones are produced than ultimately reach maturity. Since these initial
observations were confined to a limited number of animals, additional quantitative data were gained from the more extensive series of developing hamsters.
Furthermore, it was thought that the period of cell death would be compacted
into a short span which might provide ample opportunity to observe the
degenerative sequence.
The first age for which reliable quantitative data are available is 10 days p.c.
At this stage a bilateral total of 2242 neurones was counted. The definable
population increases over the next 36 h, reaching an ultimate of over 3000
neurones at 12 days p.c, which is almost twice the adult number (Table 3).
During the final days before birth a precipitous decline in cell number occurs. By
birth the nucleus is almost numerically equivalent to the adult. A comprehensive
Morphogenesis of the mesencephalic nucleus
111
38 -
30
o
=s
-1—•
14
u
//
I
I
II
I
I
13
Prenatal
I
15
I ' ' '
3
I
5
I
I
7
I
I
9
I
I
11
13
15 25 Adult
Postnatal
Fig. 10. Graphic representation of the unilateral cell total in the mesencephalic
nucleus during maturation.
listing of the cell counts is presented in tabular and graphic form in Table 3 and
Fig. 10.
A search for degenerating neurones was carried out on both pre- and postnatal animals. A few degenerating cells were first identified histologically at 13
days p.c. Although only a limited number of dying neurones was observed at
any stage, they were most numerous just before birth (Fig. 11). In addition, a
couple of instances of dying cells were noted in postnatal animals.
In the light microscope moribund cells were characterized by both nuclear
and cytoplasmic changes. The nucleus becomes irregular in shape and stains
more intensely. The cytoplasm exhibits some swelling and premature peripheral
packing of its ribosomes. Later it becomes very basophilic and disrupted.
Attempts to localize degenerating neurones with the electron microscope
were only partially successful. A couple of examples of early nuclear and
cytoplasmic changes were recorded (Fig. 12). The nucleus is more electron dense
and its boundary irregularly involuted. The cytoplasm is full of ribosomes.
Although the initial phases of cell degeneration occur rapidly, later events
must proceed at a more leisurely pace. Postnatal animals show an accumulation
of cellular debris around the viable neurones. The most usual appearance is a
large, round, electron-opaque core surrounded by a granular halo. Frequently
these particles had been engulfed by glial processes (Fig. 13).
Neuronal clusters
Tightly grouped^aggregates of neurones are a normal occurrence along the
length of the mesencephalic nucleus. Recently, Hinrichsen & Larramendi
(1968,1970) identified specialized zona adhaerens and zona occludens junctions
112
K. E. ALLEY
•^3- ^
STO
«• '•'I
JV!
u*«
^T»i
t%5«fc
* •
^> '
?• 1
w:
M
Fig. 1J. Degenerating neurones (arrows) intermingled with normal cells in the
mesencephalic nucleus at 14^ days p.c. MR - mesencephalic root, LC - locus
coeruleus. x 190.
between these closely applied cells in the mouse. Similar junctions were identified in the adult hamster (Fig. 16, 17).
Individual clusters in developing hamsters have been examined under the
electron microscope in order to determine the time of formation of these
specializations. At 9j, 10^ and 11 days p.c. cell membranes of adjacent neurones
may parallel each other for a considerable distance without any definite contacts. Small macula adhaerens junctions were identified 12^- days p.c. when single
or multiple adhesions dot the surface (Fig. 14). Similar kinds of appositions
were also found in newborn and postnatal animals of 5, 7, 12, 15 and 25 days
p.n. (Fig. 15). The overall length of junctions is longer in older animals. Observations on the formation and development of the zona occludens junctions were
not clear. A suitable plane of section through these junctions is only rarely
encountered.
Morphogenesis of the mesencephalic nucleus
113
FIGURES 12, 13
Fig. 12. Low power electron micrograph of early cellular degenerative changes in
a newborn hamster, x 3500.
Fig. 13. Terminal stages of cellular degeneration are represented by glial phagocytosis
of cellular debris, x 9000.
DISCUSSION
Morphogenesis of the trigeminal mesencephalic nucleus in the hamster
occurs during a 2-week period that spans pre- and postnatal development. The
outstanding events of ontogeny are schematically illustrated in Fig. 18. As
illustrated these include: A, the formation and initial cytodifferentiation of
neurones; B, development of cellular polarity and cell loss; C, formation of
intranuclear contacts; D, outpocketing of somatic spines; E, synaptogenesis;
and F, myelination of the mesencephalic root. In the following sections the first
four of these topics will be analyzed in terms of associated developmental
mechanisms.
Neuronalformation and cytological maturation
It has been assumed that neurones of the mesencephalic nucleus develop
from neural crest elements that are either retained in or migrate into the neural
tube after closure (Johnston, 1909; Herrick, 1948; Piatt, 1945; Rogers & Cowan,
1973). Indirect evidence presented here also suggests that these cells arise from
a germinal population other than the ventricular zone of the alar plate. Mitotic
cells were frequently found some distance from the luminal surface in close
proximity to the mesencephalic nucleus. In addition, their presence corresponds
to the time when the number of neurones in the mesencephalic nucleus is
increasing.
8
E M B 31
K. E. ALLEY
FIGURES
14-17
Fig. 14. Two small macula adhaerens junctions (arrows) at the interface of clustered
neurones in the mesencephalic nucleus 12^ days p.c. x 60000.
Fig. 15. Multiple attachment plaques anchor two neurones together in 5-day p.n.
hamster, x 26000.
Fig. 16. High power electron micrograph of zona adhaerens junction in the adult
mesencephalic nucleus, x 140000.
Fig. 17. Region of close apposition between adjacent neurones in the adult nucleus,
x 90000.
By 9j days p.c. some neurones of the mesencephalic nucleus are apparent
in the alar plate of the midbrain and pons. Their conversion from germinal
cells into immature apolar neurones is characterized by cytoplasmic changes
identical to those described for other systems (Lyser, 1964; Tennyson, 1965;
Fisher & Jacobson, 1970). Most notable of these are the rapid increase in
cytoplasmic volume and the accumulation of ribosomes and membranous
Morphogenesis of the mesencephalic nucleus
A
115
B
Fig. 18. Schematic summation of cellular development in the mesencephalic nucleus.
organelles. Neurotubules and filaments are usually sparse and confined to the
perinuclear cytoplasm. A number of authors have implicated proliferation of
neurotubules with the initiation of axonal development (Tennyson, 1965;
Lyser, 1968; Fisher & Jacobson, 1970). In the mesencephalic nucleus proliferation of neurotubules and the formation of the neurites take place soon after
the initial morpho-differentiation of the apolar neurones. By 10^- days p.c.
cell processes are abundant, and in those instances when continuity between
soma and neurite could be determined, a rich meshwork of tubules and filaments was present in the soma.
Neurones in the mesencephalic nucleus do not pass through a true phase of
bipolarity. Instead, each cell sends out a number of processes. Most are very
short and penetrate only a short distance from the cell body, giving the cell a
multipolar appearance (Pearson, 1949 a, b; Hinrichsen & Larramendi, 1969).
Later in development many of the shorter neurites are lost, leaving one main
process to enter the mesencephalic root. Around 13 or 14 days p.c. most neurones have attained their adult configuration. However, a few neurones maintain
these short processes even into adulthood.
As previously noted, Hinrichsen (1968) suggested that the adult somatic
spines might represent vestiges of the lost neurites. However, two facts argue
against this interpretation. First, there is a time gap of almost one week between
the loss of neurites and the onset of spine formation, and during the interim the
cell surface is smooth. Second, the actual number of spines is far in excess of
the number of early processes.
Perhaps a more adequate explanation of spine function can be gleaned from
116
K. E. ALLEY
an analysis of cell growth. It has been well established that the perikaryon is
the metabolic hub of the neurone. The enzyme systems for the production of
cytoplasmic proteins are concentrated in the cell body. Materials are manufactured here and shipped to the far reaches of the axon (Weiss & Hiscoe, 1948;
Droz & Leblond, 1963). Thus it seems crucial that a sufficient surface must be
available for nutrient exchange. Since the surface area of the soma increases
as a square function while the volume expands as a cube function of the cell
diameter, perhaps the tremendous increase of cytoplasmic and axoplasmic
volume during development may outstrip the surface available for exchange.
The accumulation of spines provides a significant increase in surface area
both during postnatal maturation and in the adult. Although their total effect
dwindles with time, they still provide a substantial (14 %) increase of somatic
membrane in the adult.
In other systems the presence of spines has been extensively used as an
index of the presence of synapses. Particular attention has been devoted to the
Purkinje cell (Hillman, 1969; Larramendi, 1969) and the pyramidal cells of
the cerebral cortex (Pappas & Purpura, 1961; Jacobson, 1967; Peters &
Kaiserman-Abramof, 1970). In contrast to these dendritic spines, somal thorns
of the mesencephalic neurones show no special affinity for synaptic endings.
A deeper understanding of spine function has been provided by analyzing
the development of the climbing fiber-Purkinje cell association (Larramendi &
Victor, 1966; Kornguth, Anderson & Scott, 1968). Climbing fibers first contact
spine-like processes on the soma. As these fibers migrate over the soma and on
to the main dendrites the spines regress. Possibly the transitory spines provide
a guidance system for the wandering climbing fibers. The close timing of spine
formation and synaptogenesis in the mesencephalic nucleus initially suggested
that a similar mechanism may be operant here. However, closer inspection of
these two events did not substantiate this notion. From their first appearance the
spines are totally independent of the developing synapses.
Neuronal death as a morphogenetic mechanism
Genetically planned patterns of cell degeneration are an important morphogenetic mechanism in vertebrate and invertebrate development (Gliicksmann,
1951; Lockshin & Williams, 1965; Saunders, 1966; Saunders & Fallon, 1966).
Definite localized arrays of moribund cells occur at selected instances throughout ontogenesis. They help to determine the final size and mold the shape of
organs and organ complexes.
Cell death has also been implicated as a morphogenetic mechanism in the
developing nervous system. Dying cells free the early neural tube from the overlying ectoderm (Gliicksmann, 1951), help shape the optic cup (Silver, 1972),
and determine the final size of various neural centers in the spinal cord and
brain stem (Hughes, 1961; Prestige, 1965, 1967a, b; Cowan & Wenger, 1967;
Rogers & Cowan, 1973). These authors have indicated that the normal matura-
Morphogenesis of the mesencephalic nucleus
111
tion of central-peripheral connectivity is not based on a 1 to 1 relationship, but
in fact many more young neurones are produced than eventually connect with
the periphery. Ultimately, the size of a mature neuronal pool is determined by an
intricate balance between cellular proliferation and degeneration; and the fate
of each neurone is directly dependent on its ability to establish effective contacts
in the periphery. Those unable to do so cannot maintain themselves, and quickly
regress. To date, quantitative analyses on the magnitude of this spontaneous
regression have concentrated on submammalian vertebrates. Hughes (1961)
and Prestige (1965, 1967#, b) counted viable and dying cells in the ventral horn
and spinal ganglia of Xenopus laevis and found that there is a continuous
production and turnover of neurones in these larvae. From a total population
of 10000 neurones only about 1200 reach maturity. In other words, 9 out of 10
cells end up in death. Similar investigations of chick trochlear motoneurone
(Cowan & Wenger, 1967) and mesencephalic neurones (Rogers & Cowan,
1973) have depicted a less severe reduction, averaging 50 and 70 % respectively.
Overproduction of neurones in the mesencephalic nucleus of the hamster
indicates that neuronal death may also play an important role in the ontogenesis
of mammalian brains. Although the degree of cell loss in the hamster is much
less exhaustive than in the amphibian spinal cord, it is in general accord with
the decreases noted in the chick. These differences are at least partially accounted
for by the apparent lack of continuous neuronal production and turnover in
the nervous system of higher vertebrates. In contrast, the proliferative and celldeath phases of development occur sequentially. By 12 days p.c. the mesencephalic nucleus is numerically complete. These quantitative data along with the
apparent lack of degenerating cells prior to this time suggest that no new
neurones are added once cell death commences. Perhaps this is an adaptive
modification to limit unnecessary cell wastage, and thus increases the efficiency
of neurogenesis in the bigger-brained higher vertebrates.
Experimental investigations of limb development in submammalian vertebrates
have begun to unravel the complex central-peripheral relationships that occur
during neurogenesis (Hamburger & Levi-Montalcini, 1949; Prestige, 1961 a,b;
Cowan, Martin & Wenger, 1968). In essence, three periods are evident and
these correspond to an ever-increasing reliance of the nerve cell on the periphery:
1. A proliferative, pre-axonal stage when the neurone is relatively free from
peripheral influence,
2. A saturation period, as the over-abundance of cells send their axons into
the target organ, and
3. A consolidative phase when definite connexions are formed and the survival of selected neurones is guaranteed by virtue of these peripheral contacts.
To date no information is available on either the growth of axons from the
mesencephalic root or the ontogenesis of muscle spindles in the jaw muscles.
However, the chronology of developmental events within the mesencephalic
nucleus suggests that a similar time table may be applicable. The temporal
118
K. E. ALLEY
relationship of axon development and cell death hints that many cell processes
reach the jaw muscles about 12 days p.c.
Of added importance is the timing between cell death and the onset of
synaptogenesis. It would seem to be a futile exercise to form synapses on neurones
that will not survive. In the mesencephalic nucleus cellular degeneration is
generally finished at birth, while the early events indicative of synaptogenesis
were first found four days later (Alley, 1973). This assures that synaptic terminals will contact only neurones that have already made connexions in the periphery and are sure to survive. Aside from the added economy of this sequence,
it also introduces a certain level of plasticity into the development of this system.
For example, any modification of the periphery, such as an increase or decrease
in the muscle mass, will modify the level of cellular degeneration in the mesencephalic nucleus (Piatt, 1946; Rogers & Cowan, 1973). If more muscle is available
to the ingrowing nerve fibers one would expect fewer neurones to die. Furthermore, the effect of peripheral modification would not be limited to the primary
sensory neurones alone. Changes in neuronal number would also effect the
synaptic relationships of those fibers making contact with the cells.
The importance of this sequence and the flexibility it imparts can be best
explained in terms of adaptive evolutionary change, since any modification of
the periphery must be accompanied by concomitant alterations of the associated
neural subsystems. Du Brul (1960) has shown that cell number in the adult
mesencephalic nucleus of a wide variety of vertebrates is correlated with the
size and complexity of the jaw muscles. In a later report (Du Brul, 1967) he
proposed that the target site for evolutionary modification is in the periphery,
and that adaptive changes occurring here somehow impress themselves on the
appropriate neural centers. Evidence presented here on the chronological
sequence of cellular proliferation, axon formation, cell death and synaptogenesis provides a plausible developmental mechanism whereby peripheral
changes can alter the size and synaptic patterns in the relevant neural centers.
Intranuclear cell contacts in the mesencephalic nucleus
Cellular interactions are of fundamental importance in the developing as well
as the mature nervous system. One aspect of membrane adhesions receiving
much attention is their role in morphogenesis. Tight junction and low-resistance
coupling have been found in a number of embryonic systems (Sheridan, 1968;
Trelstad, Revel & Hay, 1966). Loewenstein (1967) and Furshpan & Potter (1968)
have implicated these junctions in the passage of regulatory molecules directly
between cells, which could coordinate the growth and differentiation of cell
groups. Pannese (1968) found macula adhaerens and occludens junctions
between early primary sensory neurones in the dorsal root ganglia of chicks.
However, these appositions are only temporary in nature and are later pried
apart by the infiltration of glial cells. No trace of these junctions is left in the
adult fowl. Identical macula adhaerens junctions were first identified between
Morphogenesis of the mesencephalic nucleus
119
adjacent mesencephalic neurones at 12 days p.c. From this time on they were
frequently found. The earliest adhesions are small, single junctions. Later in
development the occurrence of multiple densities becomes more prevalent and
some of these may extend 7 fim. or more in length. Sheffield & Fischman (1970)
have suggested that these junctions may serve to stabilize intercellular relationships in rapidly growing systems. However, their increasing size with age and
persistence in the adult seem to indicate that they do more than simply anchor
mesencephalic neurones together. Maybe they encircle patches of the cell
surface, isolating these from the surrounding extracellular space and creating a
specialized intercellular environment. This area could be conducive to the
formation of low-resistance electrotonic synapses. A search for zonula occludens junctions in the prenatal hamster was unsuccessful. This may reflect their
small size or labile nature, making identification extremely tedious.
It is interesting to speculate that the membrane specializations could serve a
dual role in the ontogeny of the mesencephalic nucleus. Initially they may be
similar to those observed in other embryonic systems, serving to coordinate
maturation of the nucleus. However, in contrast to spinal ganglia neurones the
junctions are not separated by glia. These persist into adulthood and provide
the basis for electrical coupling in the adult nucleus (Hinrichsen & Larramendi,
1968, 1970; Baker & Llinas, 1971).
The author is grateful to Drs E. L. Du Brul and L. M. H. Larramendi for their continuous
advice and encouragement.
This paper is based on a thesis submitted in partial fulfillment of the requirements for the
degree of Doctor of Philosophy in Anatomy, June 1972. The work was supported by USPHS
training service grant GM-00643.
REFERENCES
K. E. (1973). Quantitative analysis of the synaptogenic period in the trigeminal
mesencephalic nucleus. Anat. Rec. Ill, 49-60.
ALLEY, K. E. & Du BRUL, E. L. (1972). Neuronal death in morphogenesis of the mammalian
mesencephalic nucleus. Anat. Rec. 172, 261.
BAKER, R. & LLINAS, P. (1971). Electrotonic coupling between neurons in the rat mesencephalic nucleus. /. Physiol. Lond. 212, 45-63.
CAJAL, S. R. Y. (1911). Histologie du Systeme Nerveux de VHomme et des Vertebres. Paris:
Maloine.
CAJAL, S. R. Y. (1929). Studies on Vertebrate Neurogenesis, trans. L. Guth, 1960. Springfield:
Charles C. Thomas.
CORBIN, K. B. & HARRISON, F. (1940). Function of the mesencephalic root. /. Neurophysiol.
3, 423-435.
COWAN, W. M. & WENGER, E. (1967). Cell loss in the trochlear nucleus of the chick during
normal development and after radical extirpation of the optic vesicle. /. exp. Zool. 164,
267-280.
COWAN, W. H., MARTIN, A. H. & WENGER, E. (1968). Mitotic patterns in the optic tectum
of the chick during normal development and after early removal of the optic vesicle.
/. exp. Zool. 169, 71-87.
DROZ, B. & LEBLOND, C. P. (1963). Axonal migration of proteins in the central nervous
system and peripheral nerves as shown by radiography. /. comp. Neurol. Ill, 325-346.
ALLEY,
120
K. E. ALLEY
Du BRUL, E. L. (1960). Structural evidence in the brain for a theory of the evolution of
behavior. Perspect. Biol Med. 4, 40-57.
Du BRUL, E. L. (1967). Pattern of genetic control of structure in the evolution of behavior.
Perspect. Biol. Med. 10, 524-539.
FISHER, S. & JACOBSON, M. (1970). Ultrastructural changes during early development of
retinal ganglia cells in Xenopus. Z. Zellforsch. mikrosk. Anat. 104, 165-177.
FURSHPAN, E. J. & POTTER, D. D. (1968). Low resistance junctions between cells in embryos
and tissue culture. In Current Topics in Developmental Biology (eds. A. A. Moscona &
A. Monroy), pp. 95-127. New York: Academic Press.
GLUCKSMANN, A. (1951). Cell death in normal vertebrate ontogeny. Biol. Rev. 26, 59-86.
HAMBURGER, V. & LEVI-MONTALCINI, R. (1949). Proliferation, differentiation and degeneration in the spinal ganglia of the chick embryo under normal and experimental conditions.
J.exp.Zool. 111,457-501.
HERRICK, C. J. (1948). The Brain of the Tiger Salamander. Chicago: University of Chicago
Press.
HILLMAN, D. E. (1969). Morphologic organization of frog cerebellar cortex: a light and
electron microscopic study. J. Neurophysiol. 32, 818-846.
HINRICHSEN, C. (1968). Central Organization of the Mesencephalic Nucleus. Thesis, University
of Illinois.
HINRICHSEN, C. F. L. & LARRAMENDI, L. M. H. (1968). Synapses and cluster formation in the
mouse mesencephalic fifth nucleus. Brain Res. 7, 296-299.
HINRICHSEN, C. F. L. & LARRAMENDI, L. M. H. (1969). The trigeminal mesencephalic nucleus.
I. Light microscopy. Am. J. Anat. 126, 497-506.
HINRICHSEN, C. F. L. & LARRAMENDI, L. M. H. (1970). The trigeminal mesencephalic nucleus.
II. Electron microscopy. Am. J. Anat. 127, 303-320.
HUGHES, A. F. (1961). Cell degeneration in the larval ventral horn of Xenopus laevis. J.
Embryol. exp. Morph. 9, 269-284.
JACOBSON, S. (1967). Dimensions of the dendritic spines in the sensorimotor cortex of the
rat, cat, squirrel, monkey and man. /. comp. Neurol. 129, 49-58.
JERGE, C. (1963). The organization and function of the trigeminal mesencephalic nucleus.
/. Neurophysiol. 26, 379-392.
JOHNSTON, J. B. (1909). The radix mesencephalica trigemini. /. comp. Neurol. 19, 593-644.
KARNOVSKY, M. M. (1965). A formaldehyde-glutaraldehyde fixative of high osmolality for
use in electron microscopy. /. Cell Biol. 27, 137 A.
KAWAMURA, Y., FUNAKOSHI, M., TSUKAMOTO, S. & TAKATA, M. (1959). Brain stem mechanisms controlling jaw muscle tonus of the dog. Jap. J. Physiol. 9, 453-463.
KIDOKORO, Y., KUBOTA, K., SHUTO, S. & SUMINO, R. (1968 a). Reflex organization of cat
masticatory muscles. /. Neurophysiol. 31, 695-708.
KIDOKORO, Y., KUBOTA, K., SHUTO, S. & SUMINO, R. (19686). Possible interneurons of jaw
closing muscles from the inferior dental nerve. /. Neurophysiol. 31, 709-718.
KONIGSMARK, B. W. (1970). Methods for the counting of neurons. In Contemporary Research
Methods in Neuroanatomy (eds. Walle J. H. Nauta & Sven O. E. Ebbesson), pp. 315-340.
New York: Springer-Verlag.
KORNGUTH, S. E., ANDERSON, J. & SCOTT, G. (1968). The development of synaptic contacts
in the cerebellum of Macaca mulatta. J. comp. Neurol. 132, 531-546.
LARRAMENDI, L. M. H. (1969). Analysis of synaptogenesis in the cerebellum of the mouse.
In Neurobiology of Cerebellar Evolution and Development (ed. R. Llinas), pp. 803-843.
Chicago: Institute for Biomedical Research.
LARRAMENDI, L. M. H. & VICTOR, T. (1966). Soma-dendritic gradient of spine resorption in
the Purkinje cell of the cerebellum of the mouse during postnatal development: an electron
microscopic study. Anat. Rec. 154, 373.
LOCKSHIN, R. A. & WILLIAMS, C. M. (1965). Programmed cell death. III. Neural control of
the breakdown of the intersegmental muscles of silkmoths. /. Insect Physiol. 11, 601-610.
LOEWENSTEIN, W. (1967). On the genesis of cellular communication. Devi Biol. 15, 503-520.
LYSER, K. M. (1964). Early differentiation of motor neuroblasts in the chick embryo as
studied by electron microscopy. I. General aspects. Devi Biol. 10, 433-466.
Morphogenesis of the mesencephalic nucleus
LYSER, K.
121
M. (1968). Early differentiation of motor neuroblasts in the chick embryo as studied
by electron microscopy. IL Microtubules and neurofilaments. Devi Biol. 17, 117-142.
PANNESE, E. (1968). Temporary junctions between neuroblasts in the developing spinal
ganglia of domestic fowl. J. Ultrastruct. Res. 21, 233-250.
PAPPAS, G. D. & PURPURA, D. P. (1961). Fine structure of dendrites in the superficial
neocortical neuropil. Expl Neurol. 4, 507-530.
PEARSON, A. A. (1949a). The development and connections of the mesencephalic root of the
trigeminal nerve in man. /. comp. Neurol. 90, 1—46.
PEARSON, A. A. (19496). Further observations on the mesencephalic root of the trigeminal
nerve. J. comp. Neurol. 91, 147-194.
PETERS, A. & KAISERMAN-ABRAMOF, I. (1970). The small pyramidal neurons of the rat cerebral
cortex. The perikaryon, dendrites and spines. Am. J. Anat. 127, 321-356.
PJATT, J. (1945). Origin of the mesencephalic V root cells in Amblystoma. J. comp. Neurol. 82,
35-54.
PIATT, j . (.1946). The influence of the peripheral field on the development of the mesencephalic
V nucleus in Amblystoma. J. exp. Zool. 102, 109-141.
PRESTIGE, M. C. (1965). Cell turnover in the spinal ganglia of Xenopus laevis tadpoles. /.
Embryo!, exp. Morph. 13, 63-72.
PRESTIGE, M. C. (1967a). The control of cell number in the lumbar spinal ganglia during
development of Xenopus laevis tadpoles. J. Embryol. exp. Morph. 17, 453-471.
PRESTIGE, M. C. (1967ft). The control of cell number in the lumbar ventral horns during
development of Xenopus laevis tadpoles. /. Embryol. exp. Morph. 18, 359-387.
ROGERS, L. & COWAN, W. M. (1973). The development of the mesencephalic nucleus of the
trigeminal nerve in the chick. /. comp. Neurol. 147, 291-320.
SAUNDERS, J. (1966). Death in embryonic systems. Science, N.Y. 154, 604-612.
SAUNDERS, J. & FALLON, J. F. (1966). Cell death in morphogenesis. In Major Problems in
Developmental Biology (ed. M. Locke), pp. 289-314. New York: Academic Press.
SHEFFIELD, J. B. & FJSCHMAN, D. A. (1970). Intercellular junctions in the developing neural
retina of the chick embryo. Z. Zellforsch. mikrosk. Anat. 104, 405-418.
SHERIDAN, J. D. (1968). Electrophysiologic evidence for low-resistance intercellular junctions
in the early chick embryo. /. Cell Biol. 37, 650-659.
SILVER, J. (1972). The role of cell death in the developing rat eye. Anat. Rec. 172, 406.
SZENTAGOTHAI, J. (1948). Anatomical considerations of monosynaptic reflex arcs. /. Neurophysiol. 11, 445^54.
TENNYSON, V. M. (1965). Electron microscopic study of the developing neuroblasts of the
dorsal root ganglion of the rabbit embryo. /. comp. Neurol. 124, 267-318.
TRELSTAD, R. L., REVEL, J. P. & HAY, E. D. (1966). Tight junctions between cells in the early
chick embryo as visualized with the electron microscope. /. Cell Biol. 31, C6-C10.
VALVERDE, F. (1970). The Golgi method. A tool for comparative structural analyses. In
Contemporary Research Methods in Neuroanatomy (eds. Walle J. H. Nauta & Sven O. E.
Ebbesson), pp. 12-28. New York: Springer-Verlag.
WEINBERG, E. (1928). The mesencephalic root of the fifth. A comparative anatomical study.
/. comp. Neurol. 46, 249-405.
WEISS, P. & HISCOE, H. B. (1948). Experiments on the mechanism of nerve growth. /. exp.
Zool. 107, 315-396.
{Received 4 April 1973)