The Organization of the Facial Nucleus of the Brush-tailed Possum (Trichosurus vufpecuta) JAN PROVIS School of Anatomy, University of New South Wales, P. 0.Box I, Kensington, New South Wales 2033, Australia
The facial nucleus of the brush-tailed possum has been studied ABSTRACT using Nissl staining and the horseradish peroxidase (HRP) retrograde tracing technique. In Nissl stained sections the nucleus is seen to comprise five distinct subnuclei. Injections of HRP into individual facial muscle groups have shown that these subnuclei reflect the peripheral innervation pattern of efferents from this nucleus. Although in most cases, injection of HRP into a single facial muscle group resulted in the labelling of neurons in more than one facial subnucleus, the following subnuclei were most completely labelled subsequent to intramuscular injection of HRP: the dorsal intermediate subnucleus was labelled with HRP reaction product following injection of m. auricularis anterior; the middle intermediate subnucleus was labelled following injection of the muscle underlying the malar vibrissae; the ventral intermediate subnucleus was labelled following injection of the m. mentalis; the medial subnucleus was labelled following injection of the m. auricularis posterior; the lateral subnuleus was labelled following injection of the m. nasolabialis with HRP. In general there is a mediolateral representation in the facial nucleus of neurons innervating facial muscle groups which are found in anteroposterior succession along the head of the animal. Muscle groups found in dorsoventral succession on the animal are represented dorsoventrally in the facial nucleus. The facial nucleus of mammals has been shown in a number of studies to be made up of a number of distinct subnuclei. Most of these investigations have been carried out on the cat -notable are those of Papez ('27) and Courville ('66).These authors arrived at similar conclusions: firstly, the subnuclei apparent in histological preparations represent discrete sources of fibers for individual facial nerve rami; secondly, rami innervating oral muscles originate in the lateral parts of the nucleus, rami innervating posterior auricular and dermal neck muscles originate more medially, while cells interposed between these two sites innervate the muscles of the anterior auricular region. Similar conclusions were drawn by Vraa-Jensen ('42) following experiments carried out in the dog. For a more complete discussion of these results the reader is referred to Courville ('66). J. COMP. NEUR., 172: 177-188.
Two studies of the facial nucleus have been carried out in the marsupial opossum (Didelphis sp.). Although the earlier of these two investigations (Papez, '27) was inconclusive, a recent study has provided a fairly complete picture of its organization (Dom et al., '73). Following lack of success with the Gudden method (Which was successfully employed b y Courville, '66, in his study on the cat), Dom et al. used a histochemical procedure to demonstrate changes in cholinesterase activity following transection of individual facial rami. Although the authors did not emphasize subdivision of the nucleus as seen in Nisslstained sections, subnuclei were demonstrated according to the origins of fibers comprising the various facial rami. These subnuclei are oriented with respect to one another in a manner resembling that described for the cat (fig. 4).
177
178
JAN PROWS
In the present study of the organization of the facial nucleus of the brush-tailed possum (Trichosurus vulpecula) the horseradish peroxidase (HRP) retrograde tracing technique was chiefly used because it offers many advantages over the use of axon reaction and cholinesterase techniques. It was originally demonstrated by Kristensson and Olsson ('71) that HRP could be taken up by motor neuron terminals and transported to the cell body. La Vail and La Vail ('721 subsequently demonstrated that axon terminals located within the CNS similarly were capable of incorporating and transporting the enzyme in a retrograde direction (See also La Vail et al., '731. It has also been shown by other workers (De Vito, Clausing and Smith, '74) that where peripheral nerves have been severed and the cut end exposed to HRP, uptake and retrograde flow of the enzyme will take place. In addition to detailing the general organization of the possum facial nucleus for comparative purposes, it was of interest to determine whether or not there is any specialization with respect to the innervation of vibrissae. It has been reported that the vibrissae are represented by somatosensory specializations, referred to as cortical barrels, in the mouse (Woolsey and van der Loos, '701, rat (Welker, '711 and brushtailed possum (Weller, '721, and it is inferred that this phenomenon may be correlated with the use of vibrissae as specialized tactile receptors, through whisking. It is still not known, however, whether or not these remarkable sensory specializations are paralleled by a comparable organization of the neuronal pool supplying the muscles innervating the vibrissae. MATERIALS AND METHODS
Preliminary information regarding the branching patterns of the facial nerve in the brush-tailed possum was gathered from dissection of formalin fixed specimens and by nerve stimulation (0.5 V, 2-second period) in six animals which had been anesthetized for other operations. A great deal of variation was observed but an addi-
tional fresh dissection (with nerve branch stimulation) was useful in determining the basic branching pattern. This animal was dissected as extensively as possible under anesthesia so that the muscles supplied by each nerve branch could be determined. Photographs were then taken, and the animal perfused transcardially with normal saline followed by a solution of 1%paraformaldehyde, 1.25% glutaraldehyde in 0.1 M phosphate buffer. The dissection was completed after perfusion because removal of the parotid gland proved difficult during live dissection because of its rich vascular supply. On this same animal before and after perfusion) a study was made of the dermal muscles innervated by the facial nerve. Observations had been made in preliminary studies regarding the presence or absence of the various dermal muscle groups and their degree of development, but the drawing presented here is primarily based on information gained from the one animal used for dissections under anesthetic. The facial motor nucleus of eight animals was examined in 30 or 40 pm frozen sections taken in parasagittal, coronal or horizontal planes. Nissl stains (thionin or cresyl violet) were used for the most part, but one series of coronal sections was examined where alternate sections were Nissl stained (thionin technique) and fiber stained (Weil's method). Examination of sections in both planes revealed distinct groups of facial subnuclei and the pattern of organization of these subnuclei relative to each other did not vary from animal to animal. A count of cells in the facial subnuclei was carried out using one out of every eight sections in a series of ninety 33 pm serial coronal sections through the nucleus. The sections were cut on a freezing microtome and mounted and stained with thionin, as described previously. Using a microscope fitted with a drawing apparatus, maps of the various subnuclei were drawn. In each case, only cells which displayed a prominent nucleolus were included in the map. The mapped cells were
179
POSSUM FACIAL NUCLEUS
then counted and an estimate made of the number of cells in each subnucleus and in the entire facial nucleus. A sample of four sections for each subnucleus was also selected from the eleven slides used for the cell count. From the sample of sections all cells which had previously been mapped into subnuclei for the cell count, were measured using an eyepiece graticule ( x 12.5) and objective lens ( x 25). The recorded cell size was the average of the largest and smallest diameters for each cell. In order to determine the relationship of the functional subdivisions of the facial nucleus to the observed subnuclei, two kinds of experiments were carried out. Although other similar studies have most frequently employed young animals the relative scarcity of possum pouch young, combined with the reported successes with the HRP technique on mature animals (see for instance La Vail et al., '73;Jones and Leavitt, '74) seemed to justify the use of adult animals in all experiments. In an initial series of experiments, brush-tailed possums were anesthetized with amylobarbitone sodium and the facial rani exposed, severed and a 25% solution of type I1 HRP applied to the cut end. Because of the failure of this first method, a second series of experiments were carried out. In the second series, 100pl injections of 25% HRP (type I1 Sigma) in saline were placed in the dermal muscle groups of the face. Overall, nine sites were injected in seven animals, five animals receiving ipsilateral injections and two animals receiving bilateral injections into different facial muscles. In all cases the injections were placed using a 1 cc syringe. For injections of vibrissal areas the needle was introduced subcutaneously at the anterior end of the vibrissae rows, and the HRP solution injected fist into the posterior end of the rows and finally to the area closest to the point of penetration. In other cases (except stapedius) a midline incision was made and the dermal muscle exposed as completely as was possible without risking damage to its neurovascular supply. The HRP was thus delivered directly into dif-
ferent parts of each muscle by a series of small injections. In the case of stapedius 100 p1 of HRP solution was introduced to the middle ear through the tympanic membrane in order to expose that muscle to the enzyme. Following survival period of three to four days the animals were perfused transcardially (as described above) and the brains then left in the animal overnight in fixative. The follwing day, the brains were removed and placed in fixative saturated with sucrose until they sank (1to 2 days). They were then sliced on a freezing microtome at 33 pm and sections collected in 0.1 M phosphate buffer (pH 7.2). Within 24 hours of sectioning the material was subjected to the histochemical method for demonstration of peroxidase activity (Graham and Karnovsky, '66). The incubation medium contained 0.05% 3,3'diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide in Tris-HC1 buffer (pH 7.6).Incubation periods ranged from 15 to 30 minutes. The only difference observed in the staining as a result of this variation in reaction time was that longer periods tended to cause brown discoloration of the sections and made examination of the tissue more difficult. Sections were mounted in a warm solution of 0.75% gelatin and 40% alcohol in distilled water, and allowed to dry. One series of sections in each animal was lightly counterstained with cresyl violet or thionin while an adjacent series was left unstained. RESULTS
The facial nerve and muscles The facial nerve and main muscle groups are shown in a drawing of a dissection in figure 1. Although some variation is present from animal to animal, the illustration shows the basic pattern of facial nerve branching. A common variation worth noting is that the ramus zygomaticus is often present as two or three smaller branches which supply the same facial muscles as are otherwise supplied by a single branch. Similarly, the rami buccolabiales may be di-
180
JAN PROWS
m. retractor palp.
Fig. 1 Drawing of the facial nerve and its peripheral rami together with a representation of the main facial muscles. Except for m. auricularis anterior and m. auricularis posterior, the muscles are completely dermal. These two auricular muscles insert into the antero-medial and postero-lateral aspects of the auricle, respectively.
vided into from two to four major branches. The m. platysma is poorly developed in the adult brush-tailed possum, and comprises only a thin layer of anteroposteriorly directed fibers found on the ventro-lateral aspect of the neck. For convenience, neither the m. platysma nor the m. sphincter colli profundus is represented in figure 1. The latter muscle comprises fibers which pass from side to side on the ventral surface of the neck as far caudal as the clavicle. Although extensive, the muscle is quite thin except for an auricular thickening (shown in fig. 1) which inserts into the dermis o f t h e inferolateral part of the auricle. The other facial muscles (shown in the diagram) form discrete, localized layers of dermal muscle which in two cases (the m. auriculares anterior and posterior) have a boney attachment. The m. nasolabialis is completely dermal and it underlies and controls the movement of the largest group of facial vibrissae, the mystacial vibrissae. Another smaller group of vibrissae, the malar vibrissae, lies about 2 cm posterior to the mystacial group. These vibrissae are controlled by a smaller, separate layer of dermal muscle fibres. The m. mentalis
comprises anteroposteriorly directed fibres on the inferior surface of the mandible.
Facial motor nucleus In the brush-tailed possum the facial motor nucleus extends for a distance of 2.9 mm throughout the medulla and its average width is about 2.0 mm. The nucleus presents five distinct subgroups of cells, the two largest being the lateral and medial subnuclei. Between these two lies an area in which three further cell groups can be distinguished - the dorsal intermediate, middle intermediate and ventral intermediate (fig. 2). A sixth group of cells has been consistently observed to lie dorsal to the medial group. However, the cells are much larger than any others observed in facial subnuclei and since no evidence has been found in this study to relate them to any part of the facial motor nucleus, they are not considered to comprise a sixth subnucleus. This cell group will subsequently be referred to as the suprafacial nucleus. Not all subnuclei extend for the full rostrocaudal length of the facial nucleus. In the most caudal sections only medial and lateral subnuclei are present, while in the
POSSUM FACIAL NUCLEUS
181
Fig. 2 Coronal section through a normal facial nucleus, stained with thionin. The five subnuclei are labelled. Oat, lateral; dors.int.,dorsal intermediate;mid.int.,middle intermediate;vent.int.,ventral intermediate; med., medial).
most rostra1 sections the medial, dorsal intermediate and lateral subnuclei are the only ones present. The results of a count of cells comprising the five facial subnuclei are presented in table 1, along with mean cell size (and standard deviation) for each subnucleus. The range of cell sizes for each subnucleus was 15-40 p. There is no obvious difference in distribution of cell sizes between subnuclei, as is reflected in the comparable average cell sizes and similar standard deviation values.
transection. This was also true of animals which, in preliminary studies, survived up to twelve days following ramus transection. On the other hand, adequate retrograde transport of HRP as judged by subsequent labelling of cells of facial subnuclei occurred where solutions of HRP had been injected into the facial muscle groups. In some cases. labelled cells took on a dark golden brown coloration (that of the HRP reaction product) which was sufficiently distinctive to enable the identification of such cells in unstained sections, using a x 10 or x 4 objective lens. Other cells HRP labelling of facial were identified as being labelled only after subnuclei close examination with x 25 or x 40 obAs mentioned previously, no HRP labell- jective lens. These two kinds of labelling ing of somata was observed in the first were frequently found in the same facial series of experiments in which HRP was nucleus coronal section, and will subseapplied to the cut end of the nerve. Simi- quently be referred to as heavy and faint larly there was no observed axon reaction labelling. The results of the facial injecin the somata as a consequence of ramus tions are presented in table 2 and examples
182
JAN PROVIS
TABLE 1
The number and average sizes of cells in the facial subnuclei Subnc1cleus
Lateral Medial Ventral intermediate Middle intermediate Dorsal intermediate
n, no. of cells per subnucleus in eleven sections
N, estimated total in subnucleus '
Percentage of cells measured in subnucleus
Average size (pin)
128 353 40 86 46
1047 2888 327 704 376
5.1 4.2 4.8 6.5 5.0
25i 5 22+ 6 24k 3 23i 4 26+ 5
' Estimated total number of neurons in facial nucleus: 5,342 E = 653)where N = n tions, S, = number of sections in which cells were counted). of labelled cells are shown in figure 3. When the injections were placed in the dermal muscles controlling the movement of both malar and mystacial vibrissae (see animal 6, table 2) heavy labelling was observed throughout the lateral subnucleus. Some labelling was also present in cells of the middle intermediate and dorsal intermediate subnuclei, but was consistently fainter than that in the lateral subnucleus. Injections confined to the muscle underlying the mystacial vibrissae resulted in
& (St = total number ofsec-
heavy labelling of cells in the lateral subnucleus while only a few faintly labelled cells were seen in the dorsal intermediate and middle intermediate subnuclei. When only the muscle underlying the malar vibrissae was injected, faintly labelled cells were found throughout the intermediate subnucleus and a few labelled cells were observed in the dorsal parts of the lateral subnucleus and in the caudal dorsal intermediate subnucleus. After injections of the m. auricularis posterior the majority of
Fig. 3 HRP labelled cells from the dorsal intermediate nucleus of animal 1.
183
POSSUM FACIAL NUCLEUS TABLE 2
The degree of labelling in facial subnuclei following HRP injections of facial muscles Labelled subnucleus Animal Side
Injection site
1
L
2
R L R
6
L
7
R L M. auric.ant.
8
R L R
M. auric. ant.
L
M.mentalis
9
M.nasolab. M.auric.post. M.nasolab. Malar vibrissae
I
Malar vibrissae
Lateral
Dorsal Int.
Mddle Int.
+++
+
+
+++
++
++
+
+
Key.
L
Stapedius
R
Platysma
Medial
Comments
++
Injection placed too laterally
+
Label in rostral parts of nucleus only
++
R 10
Ventral Int.
+++
Injection unsuitable-see text Injection unsuitable-see text
+++, heavy labelling; ++, faint labelling of most cells; +, faint labelling at a few cells
cells in the medial subnucleus (particularly in more dorsal parts) were labelled. The injections of m. auricularis anterior resulted in labelling of cells in the rostral half of the dorsal intermediate nucleus. The m. mentalis was also injected with HRP solution and the somata which supply this muscle were demonstrated to be located in the ventral intermediate subnucleus. An attempt was made at a later date to quantify these results in terms of percentages of labelled cells in each subnucleus. However, it was found that over this period of about eight months, fading of both the Nissl stain and the HRP reaction product was so pronounced that counting of cells was impossible in most cases. It was possible, however, to count the labelled cells present in the lateral, dorsal and middle intermediate subnuclei following injection of the m. nasolabialis (animal 1, table 2).In this case, fading of the Nissl stain prevented an accurate assessment of the total number of cells in the subnuclei of this animal, and the labelling percentages calcu-
lated are therefore based on the estimated totals previously recorded in table 1. This data is presented in table 3. These percentages do not include cells (also located in these subnuclei) which innervate the malar vibrissae or anterior auricular muscle. The figures presented in table 3 therefore indicate that a large proportion of cells innervating the injection site are subsequently labelled. Since the possibility of fading of TABLE 3
The number and location of cells labelled following injection of M. nasolabialis (mystacial uibrissael in animal 1 Subnucleus
Labelled cells counted Estimated total of labelled cells (EJ Percentage of labelled cells Et/N (table 1)
Lat.
Dors. int.
Mid. int.
Totals
115
18
27
159
713
111
167
992
68
30
24
47
184
JAN PROWS
A. CAT (Papez)
00 C . DOG (Vraa-Jensen)
B. CAT
(Courville)
w D. RAT (Papez)
E. OPOSSUM (Dometal.)
lateral
medial
Fig. 4 Diagram compiled from material gathered from various sources (as indicated) showing subnucleation and origins of motor fibers as described by the authors according to the key. A, r. auricularis caudalis; B, r. buccolabialis superior; C, ramus colli; SD, branch to stapedius and ramus profundus; M, buccolabialis inferior and r. marginalis mandibulae; T, r. temporalis; Z, r. zygomatico-orbitalis. The names in parenthesis refer to Papez (‘27), Vraa-Jensen (‘42),Courville (‘661, and Dom et al. (‘731.
reaction product in some cells of these subnuclei cannot be excluded, the percentage of labelled cells may be an underestimate - a fact which further supports the con-
tention that most or all facial motor neurons related to a specific muscle are labelled when that muscle is injected with HRP.
POSSUM FACIAL NUCLEUS
The organization of the facial nucleus of the possum as indicated by the HRP retrograde tracing technique is represented in figure 4 along with that determined by other workers in the cat (Papez, '27; Courville, '661, rat (Papez, '271, dog (Vraa-Jensen,'42) and opossum (Dom et al., '73) Attempts to observe the nerve cells which give rise to motor fibers for platysma, sphincter colli profundus and the posterior belly of digastric and stapedius proved unsuccessful. This may be due to one of several factors. Firstly, the innervation of these relatively minor facial muscles may be so sparse that only a small number of cells would be labelled subsequent to HRP injection of the muscle. Thus, labelled cells could be missed in scans or not present in the sample of sections reacted for demonstration of peroxidase activity. Secondly, it is possible that at least some facial nerve fibers do not retain the capacity either to take up and/or transport the enzyme (La Vail et al., '73). If such fibers formed a major component of any branch then labelling of its facial subnucleus would not occur. Thirdly, in the case of stapedius it is probable that due to technical difficulties with the injection procedure the HRP was not deposited close enough to the muscle. a
DISCUSSION
Previous studies on the functional organization of the mammalian facial nucleus have almost all been based on the results of transection of individual facial nerve rami (fig. 4).In the present study an attempt has been made to directly relate separate muscle groups to the subnuclei of the facial nucleus by using the HRP retrograde tracing technique. One obvious advantage of this latter approach in which facial muscle groups were injected is that it avoids the complication introduced by the supply of more than one muscle group by a single facial ramus. For example, stimulation showed that the ramus zygomaticus in the possum supplies both the mystacial and malar vibrissal muscle groups. Intramuscu-
185
lar injection of HRP enables one to determine the origin of the motor supply to a functional unit - a single muscle group. However, one difficulty does result from the otherwise highly suitable HRP technique adopted for the present study; the difficulty lies in the comparison of the present results with those derived from previous ramus transection studies. It is in fact impossible to make precise comparisons between the two types of studies because of the often comple relationship between facial rami and the facial muscle groups. For these reasons some allowance must be made by the reader for interpretation of the relationship between subnuclei and the peripheral rami and muscles in the different animals studied by other authors. Clearly, the overall organization of the facial subnuclei in the possum is similar to that found in other mammals studied (Papez, '27; Vraa-Jensen, '42; Szentagothai, '48;Courville, '66; Dom et al., '73). For example, the subnuclei related to the muscles of the vibrissae are laterally and dorsolaterally placed, whereas those cells related to the posterior ear are found in the medial subnucleus. Figure 4 shows that there is some variation within this basic plan, but as a general statement it seems that such differences as are apparent between the result of this and previous studies can be accounted for by the different methods used. In addition, there is some variation in the general orientation of the nucleus. The whole facial nucleus is horizontally oriented in coronal sections in marsupials whereas the long axis of the nucleus lies at about 45" to the horizontal in carnivores. Following this general introduction to the results of this study, each of the facial muscle groups and the possible relations to the facial subnuclei will be considered. Horseradish peroxidase injections of the m. auricularis anterior and muscles of the malar and mystacial vibrissae reveal that the cells which give rise to the rami zygomatico-orbitalis and temporalis and r. buccolabialis (fig.1) are confined to the dorsal and middle intermediate, and lateral sub-
186
JAN PROVIS
nuclei. While injections of the m. auricularis anterior clearly indicated that the axons of cells in the dorsal intermediate nucleus in its rostral parts fully comprised the ramus temporalis, the restricted origins for the other two rami of the upper facial muscles was not immediately clear. Injections of the malar vibrissae (innervated by r. zygomatico-orbitalis) resulted in labelling of the dorsal intermediate (caudal sections) and middle intermediate subnuclei along with dorsal parts of the lateral subnucleus. Injections of the mystacial vibrissae (innervated by the r. zygomatico-orbitalis along with the r. buccolabialis superior) resulted in similar labelling in the dorsal and middle intermediate subnuclei, along with labelling of the entire lateral subnucleus. Figure 4 shows cells labelled after malar vibrissal injections as only giving rise to the r. zygomatico-orbitalis. The additional labelling present in the lateral subnucleus following mystacial injections is represented in figure 4 as being the cells whose axons comprise r. buccolabialis superior. The alternative representation would be to present all of the lateral subnucleus, along with the dorsal and middle intermediate subnuclei as comprising cells whose axons belong to both zygomatic and buccolabial rami.
The significance of embryonic origins, and overlap in the facial nucleus A great deal of our knowledge of facial musculature is due to the work of E. Huber. In a monograph (Huber, '31) he presented an account of the phylogenetic development of vertebrate facial musculature. Huber considered that the musculature of the hyoid arch (i.e., the embryonic facial musculature) differentiated into three embryonic layers, the sphincter colli superficialis, the platysma, and t h e sphincter colli profundus. The latter two undifferentiated muscle masses are those which Huber considered to give rise to the facial musculature of the adult, whereas the sphincter colli superficialisis present as a vestigal muscle in the ventral neck region
of the adult members of only some species. In some cases remnants of the undifferentiated platysma layer and sphincter colli profundus layer are also persistent in the adult. However, the major contribution of these muscle layers is made in a developmental sense, so that nearly all of the important adult facial muscles are the derivatives of the embryonic platysma and sphincter colli profundus layers while a less important role is taken by their remnants, such as the adult platysma muscle (see Vraa-Jensen, '42; pp. 29-37 for a more complete discussion of this point). The sphincter colli profundus layer is said to give rise to the orbicularis oris, the dermal muscles surrounding the eye and the anterior auricular muscles. The platysma layer is represented only to a limited degree in the brush-tailed possum. Posteriorly the platysma muscle proper is replaced by one of its derivatives, the m. auricularis posterior. The m. mentalis is another possible derivative of the plat ysma layer (fig. 1). Vraa-Jensen, in his study of the dog ('42) found that muscles described by Huber as having different embryonic origins (that is, those derived from sphincter colli profundus, and those derived from platysma) had discrete origins of innervations from within the facial nucleus. Those muscles derived from the embryonic sphincter colli profundus were innervated by the lateral part of the nucleus, and those derived from platysma, from the medial. A similar result can be seen in the possum where rami that originate laterally in the nucleus innervate derivatives of the sphincter colli profundus. The only platysma derivative, the m. auricularis posterior, on the other hand, has been observed to be supplied by cells from the medial subnucleus. On such a basis one would expect the deep facial muscles Istapedim, posterior belly of digastric and the m. stylohoidius) - also platysma derivatives (Huber, '311, to occupy other parts of the medial subnucleus. However, this is yet to be demonstrated in the possum, although such a proposal is apparently upheld in the dog (fig.4C) and opossum (fig. 4).
POSSUM FACIAL NUCLEUS
Vraa-Jensen ('42) found that within facial muscle groups of similar embryonic origins, those which were distinctly separate had similarly distinct subnuclei from which their innervations were derived. Injections of facial muscle groups in the possum have indicated, at least in the present study, that this is not the case. It has shown that the cells which innervate the mystacial vibrissal muscle overlap with those which supply the malar vibrissal muscle. A similar situation is seen in the case of the m. auricularis anterior, although there does seem to be some rostrocaudal separation in the dorsal intermediate subnucleus between the origin of innervation for the m. auricularis anterior and the origin of motor fibers to the vibrissae. The case of the m. mentalis presents some ambiguity. The longitudinal orientation of its fibers suggests that it is derived from platysma, however it is partly innervated by the inferior buccolabial ramus, whose main area of supply is the orbicularis oris, a derivative of the embryonic sphincter colli profundus. Neither this muscle nor its main source of innervation the r. marginalis mandibulae, are described by Huber ('31)but the r. marginalis mandibulae is present in the opossum (Dom et al., '73). In this study the fibers from the two nerves which innervate it are derived from cells entirely within the ventral intermediate subnucleus, and there is no evidence of overlap with any of the other subnuclei. The innervation of malar and mystacial
187
vibrissae is derived from the lateral subnucleus and the middle and dorsal intermediate subnuclei. In these subnuclei there is no apparent enlargement of cells or increase in cell density which might reflect an augmented innervation of the muscles related to whisking of the vibrissae and thus the apparently important tactile function of the facial vibrissae of the possum is not reflected in any way in the amount of facial nucleus devoted to their motor innervation. The results of this study are presented in figure 5, using facial muscle groups rather than facial nerve rami as the basis of organisation in the facial nucleus. By defining the most heavily labelled cells as those which exclusively or most densely innervate the muscle group injected (Jones and Leavitt, '74) it has been possible to clarify the organisational representation of figure 4F. The large area shown as the origin of axons comprising r. zygomatico-orbitalis can be seen in figure 5 to have an emphasis in different parts for different muscle groups. The emphasis falls in such a way that the subnucleation pattern from lateral to medial represents the groups of facial muscles from nose to occiput. Similarly, passing dorsoventrally through the nucleus, the animals facial musculature is also represented dorsoventrally. However, the consistent overlap of areas labelled following injection of the vibrissal areas and the anterior auricular muscles (sphincter colli profundus-layer derivatives) cannot
LAT.
1 mm Fig, 5 The facial nucleus of the brush-tailed possum (coronal) represented according to the sources of innervation for the facial muscle groups.
188
JAN PROVIS
be ignored. Since the facial muscles of the possum each form discrete dermal bands, it is unlikely that spread of enzyme from the injected muscle to another is the cause of overlap. The fact that no overlap of labelled subnuclei occurred following injection of facial muscle groups derived from the different embryological layers suggests that functional specialization within the nucleus is complete for muscles of different embryological origin. The persistence of overlap in the subnuclei labelled following injection of the sphincter colli profundus layer derivatives implies that for the motorneurons innervating muscles of a single layer, segregation within the nucleus is incomplete. The presence of varied labelling intensity in these subnuclei following a single muscle injection implies that at least some segregation has, however, occurred. It is on this basis that the functional association between the dorsal and middle intermediate, and lateral subnuclei, and the facial muscles was made in figure 5. Although it is clear that subnucleationon a functional basis is present in the facial nuclei of mammals, the functions of these subnuclei are not mutually exclusive in the possum. The only apparent form in which exclusive subnucleation is present is that which occurs in relation to the embryonic subdivisions of the muscle of the hyoid arch -the undifferentiated platysma and sphincter colli profundus, which subsequently gives rise to the facial musculature of the adult.
Service for their co-operation in obtaining animals. LITERATURE CITED
Courville, J. 1966 The nucleus of the facial nerve; the relation between cellular groups and peripheral branches of the nerve. Brain Res., 1; 338-354. De Vito, J. L., K. W. Clausing and D. A. Smith 1974 Uptake and transport of horseradish peroxidase by cut end of vagus nerve. Brain Res., 82; 269-271. Dom, R., W. Falls and G. F. Martin 1973 The motor nucleus of the facial nerve in the opossum Didelphis marsupidis oirginiuna)Its organisation and connections. J. Comp. Neur., 152: 373-402. Graham, R. C., and M. J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Huber, E. 1931 The Evolution of Facial Musculature and Facial Expression. John Hopkins Press, Baltimore. Jones, E. G., and R. Y. Leavitt 1974 Retrograde axonal transport and the demonstration of nonspecific projections to the cerebral cortex and striatum from thalamic intralaminar nuclei in the rat, cat and monkey. J. Comp. Neur., 154: 349-378. Kristensson, K., and Y. Olsson 1971 Retrograde axonal transport of protein. Brain Res., 29: 363-365. La Vail, J. H., and M. M. La Vail 1972 Retrograde axonal transport in the central nervous system. Science, 176; 1416-1417. La Vail, J. H., K. R. Winston and A. Tish 1973 A method based on retrograde intra-axonal transport of protein for identification of cell bodies of origin of axons terminating within the CNS. Brain Res., 58: 470-477. Papez, J. W. 1927 Subdivisions of the facial nucleus. J. a m p . Neur., 43: 159-191. Szentagothai, J. 1948 The representation of facial and scalp muscles in the facial nucleus. J. Comp. Neur., 88: 207-220. Vraa-Jensen, G. F. 1942 The Motor Nucleus of the Facial Nerve. With a Survey of the Efferent Innervation of the Facial Nucleus. Ejnar Munksgaard, ACKNOWLEDGMENTS Copenhagen. C. 1971 Microelectrode delineation of I wish to thank Doctor Charles Watson Welker, fine grain somatotopic organisation of SMI cerebral for his advice and encouragement, both in neocortex in albino rat. Brain Res., 26; 259-275. execution of experimental procedures and Weller, W. L. 1972 Barrels in somatic sensory neocortex of the marsupial Trichosurus oulpecula in the preparation of the manuscript. Also (brush-tailed possum). Brain Res., 43: 11-24. Mr. Brian Freeman for his interest and adT. A., and H. van der Loos 1970 The strucvice and Mr. Des Shimeld for assistance Woolsey, tural organisation of layer IV in the somatosensory with the anesthetics and operations. 1 am region 61) of mouse cerebral cortex. Brain Res., 17; grateful to The National Parks and Wildlife 205-242.
The facial nucleus of the brush-tailed possum has been studied ABSTRACT using Nissl staining and the horseradish peroxidase (HRP) retrograde tracing technique. In Nissl stained sections the nucleus is seen to comprise five distinct subnuclei. Injections of HRP into individual facial muscle groups have shown that these subnuclei reflect the peripheral innervation pattern of efferents from this nucleus. Although in most cases, injection of HRP into a single facial muscle group resulted in the labelling of neurons in more than one facial subnucleus, the following subnuclei were most completely labelled subsequent to intramuscular injection of HRP: the dorsal intermediate subnucleus was labelled with HRP reaction product following injection of m. auricularis anterior; the middle intermediate subnucleus was labelled following injection of the muscle underlying the malar vibrissae; the ventral intermediate subnucleus was labelled following injection of the m. mentalis; the medial subnucleus was labelled following injection of the m. auricularis posterior; the lateral subnuleus was labelled following injection of the m. nasolabialis with HRP. In general there is a mediolateral representation in the facial nucleus of neurons innervating facial muscle groups which are found in anteroposterior succession along the head of the animal. Muscle groups found in dorsoventral succession on the animal are represented dorsoventrally in the facial nucleus. The facial nucleus of mammals has been shown in a number of studies to be made up of a number of distinct subnuclei. Most of these investigations have been carried out on the cat -notable are those of Papez ('27) and Courville ('66).These authors arrived at similar conclusions: firstly, the subnuclei apparent in histological preparations represent discrete sources of fibers for individual facial nerve rami; secondly, rami innervating oral muscles originate in the lateral parts of the nucleus, rami innervating posterior auricular and dermal neck muscles originate more medially, while cells interposed between these two sites innervate the muscles of the anterior auricular region. Similar conclusions were drawn by Vraa-Jensen ('42) following experiments carried out in the dog. For a more complete discussion of these results the reader is referred to Courville ('66). J. COMP. NEUR., 172: 177-188.
Two studies of the facial nucleus have been carried out in the marsupial opossum (Didelphis sp.). Although the earlier of these two investigations (Papez, '27) was inconclusive, a recent study has provided a fairly complete picture of its organization (Dom et al., '73). Following lack of success with the Gudden method (Which was successfully employed b y Courville, '66, in his study on the cat), Dom et al. used a histochemical procedure to demonstrate changes in cholinesterase activity following transection of individual facial rami. Although the authors did not emphasize subdivision of the nucleus as seen in Nisslstained sections, subnuclei were demonstrated according to the origins of fibers comprising the various facial rami. These subnuclei are oriented with respect to one another in a manner resembling that described for the cat (fig. 4).
177
178
JAN PROWS
In the present study of the organization of the facial nucleus of the brush-tailed possum (Trichosurus vulpecula) the horseradish peroxidase (HRP) retrograde tracing technique was chiefly used because it offers many advantages over the use of axon reaction and cholinesterase techniques. It was originally demonstrated by Kristensson and Olsson ('71) that HRP could be taken up by motor neuron terminals and transported to the cell body. La Vail and La Vail ('721 subsequently demonstrated that axon terminals located within the CNS similarly were capable of incorporating and transporting the enzyme in a retrograde direction (See also La Vail et al., '731. It has also been shown by other workers (De Vito, Clausing and Smith, '74) that where peripheral nerves have been severed and the cut end exposed to HRP, uptake and retrograde flow of the enzyme will take place. In addition to detailing the general organization of the possum facial nucleus for comparative purposes, it was of interest to determine whether or not there is any specialization with respect to the innervation of vibrissae. It has been reported that the vibrissae are represented by somatosensory specializations, referred to as cortical barrels, in the mouse (Woolsey and van der Loos, '701, rat (Welker, '711 and brushtailed possum (Weller, '721, and it is inferred that this phenomenon may be correlated with the use of vibrissae as specialized tactile receptors, through whisking. It is still not known, however, whether or not these remarkable sensory specializations are paralleled by a comparable organization of the neuronal pool supplying the muscles innervating the vibrissae. MATERIALS AND METHODS
Preliminary information regarding the branching patterns of the facial nerve in the brush-tailed possum was gathered from dissection of formalin fixed specimens and by nerve stimulation (0.5 V, 2-second period) in six animals which had been anesthetized for other operations. A great deal of variation was observed but an addi-
tional fresh dissection (with nerve branch stimulation) was useful in determining the basic branching pattern. This animal was dissected as extensively as possible under anesthesia so that the muscles supplied by each nerve branch could be determined. Photographs were then taken, and the animal perfused transcardially with normal saline followed by a solution of 1%paraformaldehyde, 1.25% glutaraldehyde in 0.1 M phosphate buffer. The dissection was completed after perfusion because removal of the parotid gland proved difficult during live dissection because of its rich vascular supply. On this same animal before and after perfusion) a study was made of the dermal muscles innervated by the facial nerve. Observations had been made in preliminary studies regarding the presence or absence of the various dermal muscle groups and their degree of development, but the drawing presented here is primarily based on information gained from the one animal used for dissections under anesthetic. The facial motor nucleus of eight animals was examined in 30 or 40 pm frozen sections taken in parasagittal, coronal or horizontal planes. Nissl stains (thionin or cresyl violet) were used for the most part, but one series of coronal sections was examined where alternate sections were Nissl stained (thionin technique) and fiber stained (Weil's method). Examination of sections in both planes revealed distinct groups of facial subnuclei and the pattern of organization of these subnuclei relative to each other did not vary from animal to animal. A count of cells in the facial subnuclei was carried out using one out of every eight sections in a series of ninety 33 pm serial coronal sections through the nucleus. The sections were cut on a freezing microtome and mounted and stained with thionin, as described previously. Using a microscope fitted with a drawing apparatus, maps of the various subnuclei were drawn. In each case, only cells which displayed a prominent nucleolus were included in the map. The mapped cells were
179
POSSUM FACIAL NUCLEUS
then counted and an estimate made of the number of cells in each subnucleus and in the entire facial nucleus. A sample of four sections for each subnucleus was also selected from the eleven slides used for the cell count. From the sample of sections all cells which had previously been mapped into subnuclei for the cell count, were measured using an eyepiece graticule ( x 12.5) and objective lens ( x 25). The recorded cell size was the average of the largest and smallest diameters for each cell. In order to determine the relationship of the functional subdivisions of the facial nucleus to the observed subnuclei, two kinds of experiments were carried out. Although other similar studies have most frequently employed young animals the relative scarcity of possum pouch young, combined with the reported successes with the HRP technique on mature animals (see for instance La Vail et al., '73;Jones and Leavitt, '74) seemed to justify the use of adult animals in all experiments. In an initial series of experiments, brush-tailed possums were anesthetized with amylobarbitone sodium and the facial rani exposed, severed and a 25% solution of type I1 HRP applied to the cut end. Because of the failure of this first method, a second series of experiments were carried out. In the second series, 100pl injections of 25% HRP (type I1 Sigma) in saline were placed in the dermal muscle groups of the face. Overall, nine sites were injected in seven animals, five animals receiving ipsilateral injections and two animals receiving bilateral injections into different facial muscles. In all cases the injections were placed using a 1 cc syringe. For injections of vibrissal areas the needle was introduced subcutaneously at the anterior end of the vibrissae rows, and the HRP solution injected fist into the posterior end of the rows and finally to the area closest to the point of penetration. In other cases (except stapedius) a midline incision was made and the dermal muscle exposed as completely as was possible without risking damage to its neurovascular supply. The HRP was thus delivered directly into dif-
ferent parts of each muscle by a series of small injections. In the case of stapedius 100 p1 of HRP solution was introduced to the middle ear through the tympanic membrane in order to expose that muscle to the enzyme. Following survival period of three to four days the animals were perfused transcardially (as described above) and the brains then left in the animal overnight in fixative. The follwing day, the brains were removed and placed in fixative saturated with sucrose until they sank (1to 2 days). They were then sliced on a freezing microtome at 33 pm and sections collected in 0.1 M phosphate buffer (pH 7.2). Within 24 hours of sectioning the material was subjected to the histochemical method for demonstration of peroxidase activity (Graham and Karnovsky, '66). The incubation medium contained 0.05% 3,3'diaminobenzidine tetrahydrochloride and 0.01% hydrogen peroxide in Tris-HC1 buffer (pH 7.6).Incubation periods ranged from 15 to 30 minutes. The only difference observed in the staining as a result of this variation in reaction time was that longer periods tended to cause brown discoloration of the sections and made examination of the tissue more difficult. Sections were mounted in a warm solution of 0.75% gelatin and 40% alcohol in distilled water, and allowed to dry. One series of sections in each animal was lightly counterstained with cresyl violet or thionin while an adjacent series was left unstained. RESULTS
The facial nerve and muscles The facial nerve and main muscle groups are shown in a drawing of a dissection in figure 1. Although some variation is present from animal to animal, the illustration shows the basic pattern of facial nerve branching. A common variation worth noting is that the ramus zygomaticus is often present as two or three smaller branches which supply the same facial muscles as are otherwise supplied by a single branch. Similarly, the rami buccolabiales may be di-
180
JAN PROWS
m. retractor palp.
Fig. 1 Drawing of the facial nerve and its peripheral rami together with a representation of the main facial muscles. Except for m. auricularis anterior and m. auricularis posterior, the muscles are completely dermal. These two auricular muscles insert into the antero-medial and postero-lateral aspects of the auricle, respectively.
vided into from two to four major branches. The m. platysma is poorly developed in the adult brush-tailed possum, and comprises only a thin layer of anteroposteriorly directed fibers found on the ventro-lateral aspect of the neck. For convenience, neither the m. platysma nor the m. sphincter colli profundus is represented in figure 1. The latter muscle comprises fibers which pass from side to side on the ventral surface of the neck as far caudal as the clavicle. Although extensive, the muscle is quite thin except for an auricular thickening (shown in fig. 1) which inserts into the dermis o f t h e inferolateral part of the auricle. The other facial muscles (shown in the diagram) form discrete, localized layers of dermal muscle which in two cases (the m. auriculares anterior and posterior) have a boney attachment. The m. nasolabialis is completely dermal and it underlies and controls the movement of the largest group of facial vibrissae, the mystacial vibrissae. Another smaller group of vibrissae, the malar vibrissae, lies about 2 cm posterior to the mystacial group. These vibrissae are controlled by a smaller, separate layer of dermal muscle fibres. The m. mentalis
comprises anteroposteriorly directed fibres on the inferior surface of the mandible.
Facial motor nucleus In the brush-tailed possum the facial motor nucleus extends for a distance of 2.9 mm throughout the medulla and its average width is about 2.0 mm. The nucleus presents five distinct subgroups of cells, the two largest being the lateral and medial subnuclei. Between these two lies an area in which three further cell groups can be distinguished - the dorsal intermediate, middle intermediate and ventral intermediate (fig. 2). A sixth group of cells has been consistently observed to lie dorsal to the medial group. However, the cells are much larger than any others observed in facial subnuclei and since no evidence has been found in this study to relate them to any part of the facial motor nucleus, they are not considered to comprise a sixth subnucleus. This cell group will subsequently be referred to as the suprafacial nucleus. Not all subnuclei extend for the full rostrocaudal length of the facial nucleus. In the most caudal sections only medial and lateral subnuclei are present, while in the
POSSUM FACIAL NUCLEUS
181
Fig. 2 Coronal section through a normal facial nucleus, stained with thionin. The five subnuclei are labelled. Oat, lateral; dors.int.,dorsal intermediate;mid.int.,middle intermediate;vent.int.,ventral intermediate; med., medial).
most rostra1 sections the medial, dorsal intermediate and lateral subnuclei are the only ones present. The results of a count of cells comprising the five facial subnuclei are presented in table 1, along with mean cell size (and standard deviation) for each subnucleus. The range of cell sizes for each subnucleus was 15-40 p. There is no obvious difference in distribution of cell sizes between subnuclei, as is reflected in the comparable average cell sizes and similar standard deviation values.
transection. This was also true of animals which, in preliminary studies, survived up to twelve days following ramus transection. On the other hand, adequate retrograde transport of HRP as judged by subsequent labelling of cells of facial subnuclei occurred where solutions of HRP had been injected into the facial muscle groups. In some cases. labelled cells took on a dark golden brown coloration (that of the HRP reaction product) which was sufficiently distinctive to enable the identification of such cells in unstained sections, using a x 10 or x 4 objective lens. Other cells HRP labelling of facial were identified as being labelled only after subnuclei close examination with x 25 or x 40 obAs mentioned previously, no HRP labell- jective lens. These two kinds of labelling ing of somata was observed in the first were frequently found in the same facial series of experiments in which HRP was nucleus coronal section, and will subseapplied to the cut end of the nerve. Simi- quently be referred to as heavy and faint larly there was no observed axon reaction labelling. The results of the facial injecin the somata as a consequence of ramus tions are presented in table 2 and examples
182
JAN PROVIS
TABLE 1
The number and average sizes of cells in the facial subnuclei Subnc1cleus
Lateral Medial Ventral intermediate Middle intermediate Dorsal intermediate
n, no. of cells per subnucleus in eleven sections
N, estimated total in subnucleus '
Percentage of cells measured in subnucleus
Average size (pin)
128 353 40 86 46
1047 2888 327 704 376
5.1 4.2 4.8 6.5 5.0
25i 5 22+ 6 24k 3 23i 4 26+ 5
' Estimated total number of neurons in facial nucleus: 5,342 E = 653)where N = n tions, S, = number of sections in which cells were counted). of labelled cells are shown in figure 3. When the injections were placed in the dermal muscles controlling the movement of both malar and mystacial vibrissae (see animal 6, table 2) heavy labelling was observed throughout the lateral subnucleus. Some labelling was also present in cells of the middle intermediate and dorsal intermediate subnuclei, but was consistently fainter than that in the lateral subnucleus. Injections confined to the muscle underlying the mystacial vibrissae resulted in
& (St = total number ofsec-
heavy labelling of cells in the lateral subnucleus while only a few faintly labelled cells were seen in the dorsal intermediate and middle intermediate subnuclei. When only the muscle underlying the malar vibrissae was injected, faintly labelled cells were found throughout the intermediate subnucleus and a few labelled cells were observed in the dorsal parts of the lateral subnucleus and in the caudal dorsal intermediate subnucleus. After injections of the m. auricularis posterior the majority of
Fig. 3 HRP labelled cells from the dorsal intermediate nucleus of animal 1.
183
POSSUM FACIAL NUCLEUS TABLE 2
The degree of labelling in facial subnuclei following HRP injections of facial muscles Labelled subnucleus Animal Side
Injection site
1
L
2
R L R
6
L
7
R L M. auric.ant.
8
R L R
M. auric. ant.
L
M.mentalis
9
M.nasolab. M.auric.post. M.nasolab. Malar vibrissae
I
Malar vibrissae
Lateral
Dorsal Int.
Mddle Int.
+++
+
+
+++
++
++
+
+
Key.
L
Stapedius
R
Platysma
Medial
Comments
++
Injection placed too laterally
+
Label in rostral parts of nucleus only
++
R 10
Ventral Int.
+++
Injection unsuitable-see text Injection unsuitable-see text
+++, heavy labelling; ++, faint labelling of most cells; +, faint labelling at a few cells
cells in the medial subnucleus (particularly in more dorsal parts) were labelled. The injections of m. auricularis anterior resulted in labelling of cells in the rostral half of the dorsal intermediate nucleus. The m. mentalis was also injected with HRP solution and the somata which supply this muscle were demonstrated to be located in the ventral intermediate subnucleus. An attempt was made at a later date to quantify these results in terms of percentages of labelled cells in each subnucleus. However, it was found that over this period of about eight months, fading of both the Nissl stain and the HRP reaction product was so pronounced that counting of cells was impossible in most cases. It was possible, however, to count the labelled cells present in the lateral, dorsal and middle intermediate subnuclei following injection of the m. nasolabialis (animal 1, table 2).In this case, fading of the Nissl stain prevented an accurate assessment of the total number of cells in the subnuclei of this animal, and the labelling percentages calcu-
lated are therefore based on the estimated totals previously recorded in table 1. This data is presented in table 3. These percentages do not include cells (also located in these subnuclei) which innervate the malar vibrissae or anterior auricular muscle. The figures presented in table 3 therefore indicate that a large proportion of cells innervating the injection site are subsequently labelled. Since the possibility of fading of TABLE 3
The number and location of cells labelled following injection of M. nasolabialis (mystacial uibrissael in animal 1 Subnucleus
Labelled cells counted Estimated total of labelled cells (EJ Percentage of labelled cells Et/N (table 1)
Lat.
Dors. int.
Mid. int.
Totals
115
18
27
159
713
111
167
992
68
30
24
47
184
JAN PROWS
A. CAT (Papez)
00 C . DOG (Vraa-Jensen)
B. CAT
(Courville)
w D. RAT (Papez)
E. OPOSSUM (Dometal.)
lateral
medial
Fig. 4 Diagram compiled from material gathered from various sources (as indicated) showing subnucleation and origins of motor fibers as described by the authors according to the key. A, r. auricularis caudalis; B, r. buccolabialis superior; C, ramus colli; SD, branch to stapedius and ramus profundus; M, buccolabialis inferior and r. marginalis mandibulae; T, r. temporalis; Z, r. zygomatico-orbitalis. The names in parenthesis refer to Papez (‘27), Vraa-Jensen (‘42),Courville (‘661, and Dom et al. (‘731.
reaction product in some cells of these subnuclei cannot be excluded, the percentage of labelled cells may be an underestimate - a fact which further supports the con-
tention that most or all facial motor neurons related to a specific muscle are labelled when that muscle is injected with HRP.
POSSUM FACIAL NUCLEUS
The organization of the facial nucleus of the possum as indicated by the HRP retrograde tracing technique is represented in figure 4 along with that determined by other workers in the cat (Papez, '27; Courville, '661, rat (Papez, '271, dog (Vraa-Jensen,'42) and opossum (Dom et al., '73) Attempts to observe the nerve cells which give rise to motor fibers for platysma, sphincter colli profundus and the posterior belly of digastric and stapedius proved unsuccessful. This may be due to one of several factors. Firstly, the innervation of these relatively minor facial muscles may be so sparse that only a small number of cells would be labelled subsequent to HRP injection of the muscle. Thus, labelled cells could be missed in scans or not present in the sample of sections reacted for demonstration of peroxidase activity. Secondly, it is possible that at least some facial nerve fibers do not retain the capacity either to take up and/or transport the enzyme (La Vail et al., '73). If such fibers formed a major component of any branch then labelling of its facial subnucleus would not occur. Thirdly, in the case of stapedius it is probable that due to technical difficulties with the injection procedure the HRP was not deposited close enough to the muscle. a
DISCUSSION
Previous studies on the functional organization of the mammalian facial nucleus have almost all been based on the results of transection of individual facial nerve rami (fig. 4).In the present study an attempt has been made to directly relate separate muscle groups to the subnuclei of the facial nucleus by using the HRP retrograde tracing technique. One obvious advantage of this latter approach in which facial muscle groups were injected is that it avoids the complication introduced by the supply of more than one muscle group by a single facial ramus. For example, stimulation showed that the ramus zygomaticus in the possum supplies both the mystacial and malar vibrissal muscle groups. Intramuscu-
185
lar injection of HRP enables one to determine the origin of the motor supply to a functional unit - a single muscle group. However, one difficulty does result from the otherwise highly suitable HRP technique adopted for the present study; the difficulty lies in the comparison of the present results with those derived from previous ramus transection studies. It is in fact impossible to make precise comparisons between the two types of studies because of the often comple relationship between facial rami and the facial muscle groups. For these reasons some allowance must be made by the reader for interpretation of the relationship between subnuclei and the peripheral rami and muscles in the different animals studied by other authors. Clearly, the overall organization of the facial subnuclei in the possum is similar to that found in other mammals studied (Papez, '27; Vraa-Jensen, '42; Szentagothai, '48;Courville, '66; Dom et al., '73). For example, the subnuclei related to the muscles of the vibrissae are laterally and dorsolaterally placed, whereas those cells related to the posterior ear are found in the medial subnucleus. Figure 4 shows that there is some variation within this basic plan, but as a general statement it seems that such differences as are apparent between the result of this and previous studies can be accounted for by the different methods used. In addition, there is some variation in the general orientation of the nucleus. The whole facial nucleus is horizontally oriented in coronal sections in marsupials whereas the long axis of the nucleus lies at about 45" to the horizontal in carnivores. Following this general introduction to the results of this study, each of the facial muscle groups and the possible relations to the facial subnuclei will be considered. Horseradish peroxidase injections of the m. auricularis anterior and muscles of the malar and mystacial vibrissae reveal that the cells which give rise to the rami zygomatico-orbitalis and temporalis and r. buccolabialis (fig.1) are confined to the dorsal and middle intermediate, and lateral sub-
186
JAN PROVIS
nuclei. While injections of the m. auricularis anterior clearly indicated that the axons of cells in the dorsal intermediate nucleus in its rostral parts fully comprised the ramus temporalis, the restricted origins for the other two rami of the upper facial muscles was not immediately clear. Injections of the malar vibrissae (innervated by r. zygomatico-orbitalis) resulted in labelling of the dorsal intermediate (caudal sections) and middle intermediate subnuclei along with dorsal parts of the lateral subnucleus. Injections of the mystacial vibrissae (innervated by the r. zygomatico-orbitalis along with the r. buccolabialis superior) resulted in similar labelling in the dorsal and middle intermediate subnuclei, along with labelling of the entire lateral subnucleus. Figure 4 shows cells labelled after malar vibrissal injections as only giving rise to the r. zygomatico-orbitalis. The additional labelling present in the lateral subnucleus following mystacial injections is represented in figure 4 as being the cells whose axons comprise r. buccolabialis superior. The alternative representation would be to present all of the lateral subnucleus, along with the dorsal and middle intermediate subnuclei as comprising cells whose axons belong to both zygomatic and buccolabial rami.
The significance of embryonic origins, and overlap in the facial nucleus A great deal of our knowledge of facial musculature is due to the work of E. Huber. In a monograph (Huber, '31) he presented an account of the phylogenetic development of vertebrate facial musculature. Huber considered that the musculature of the hyoid arch (i.e., the embryonic facial musculature) differentiated into three embryonic layers, the sphincter colli superficialis, the platysma, and t h e sphincter colli profundus. The latter two undifferentiated muscle masses are those which Huber considered to give rise to the facial musculature of the adult, whereas the sphincter colli superficialisis present as a vestigal muscle in the ventral neck region
of the adult members of only some species. In some cases remnants of the undifferentiated platysma layer and sphincter colli profundus layer are also persistent in the adult. However, the major contribution of these muscle layers is made in a developmental sense, so that nearly all of the important adult facial muscles are the derivatives of the embryonic platysma and sphincter colli profundus layers while a less important role is taken by their remnants, such as the adult platysma muscle (see Vraa-Jensen, '42; pp. 29-37 for a more complete discussion of this point). The sphincter colli profundus layer is said to give rise to the orbicularis oris, the dermal muscles surrounding the eye and the anterior auricular muscles. The platysma layer is represented only to a limited degree in the brush-tailed possum. Posteriorly the platysma muscle proper is replaced by one of its derivatives, the m. auricularis posterior. The m. mentalis is another possible derivative of the plat ysma layer (fig. 1). Vraa-Jensen, in his study of the dog ('42) found that muscles described by Huber as having different embryonic origins (that is, those derived from sphincter colli profundus, and those derived from platysma) had discrete origins of innervations from within the facial nucleus. Those muscles derived from the embryonic sphincter colli profundus were innervated by the lateral part of the nucleus, and those derived from platysma, from the medial. A similar result can be seen in the possum where rami that originate laterally in the nucleus innervate derivatives of the sphincter colli profundus. The only platysma derivative, the m. auricularis posterior, on the other hand, has been observed to be supplied by cells from the medial subnucleus. On such a basis one would expect the deep facial muscles Istapedim, posterior belly of digastric and the m. stylohoidius) - also platysma derivatives (Huber, '311, to occupy other parts of the medial subnucleus. However, this is yet to be demonstrated in the possum, although such a proposal is apparently upheld in the dog (fig.4C) and opossum (fig. 4).
POSSUM FACIAL NUCLEUS
Vraa-Jensen ('42) found that within facial muscle groups of similar embryonic origins, those which were distinctly separate had similarly distinct subnuclei from which their innervations were derived. Injections of facial muscle groups in the possum have indicated, at least in the present study, that this is not the case. It has shown that the cells which innervate the mystacial vibrissal muscle overlap with those which supply the malar vibrissal muscle. A similar situation is seen in the case of the m. auricularis anterior, although there does seem to be some rostrocaudal separation in the dorsal intermediate subnucleus between the origin of innervation for the m. auricularis anterior and the origin of motor fibers to the vibrissae. The case of the m. mentalis presents some ambiguity. The longitudinal orientation of its fibers suggests that it is derived from platysma, however it is partly innervated by the inferior buccolabial ramus, whose main area of supply is the orbicularis oris, a derivative of the embryonic sphincter colli profundus. Neither this muscle nor its main source of innervation the r. marginalis mandibulae, are described by Huber ('31)but the r. marginalis mandibulae is present in the opossum (Dom et al., '73). In this study the fibers from the two nerves which innervate it are derived from cells entirely within the ventral intermediate subnucleus, and there is no evidence of overlap with any of the other subnuclei. The innervation of malar and mystacial
187
vibrissae is derived from the lateral subnucleus and the middle and dorsal intermediate subnuclei. In these subnuclei there is no apparent enlargement of cells or increase in cell density which might reflect an augmented innervation of the muscles related to whisking of the vibrissae and thus the apparently important tactile function of the facial vibrissae of the possum is not reflected in any way in the amount of facial nucleus devoted to their motor innervation. The results of this study are presented in figure 5, using facial muscle groups rather than facial nerve rami as the basis of organisation in the facial nucleus. By defining the most heavily labelled cells as those which exclusively or most densely innervate the muscle group injected (Jones and Leavitt, '74) it has been possible to clarify the organisational representation of figure 4F. The large area shown as the origin of axons comprising r. zygomatico-orbitalis can be seen in figure 5 to have an emphasis in different parts for different muscle groups. The emphasis falls in such a way that the subnucleation pattern from lateral to medial represents the groups of facial muscles from nose to occiput. Similarly, passing dorsoventrally through the nucleus, the animals facial musculature is also represented dorsoventrally. However, the consistent overlap of areas labelled following injection of the vibrissal areas and the anterior auricular muscles (sphincter colli profundus-layer derivatives) cannot
LAT.
1 mm Fig, 5 The facial nucleus of the brush-tailed possum (coronal) represented according to the sources of innervation for the facial muscle groups.
188
JAN PROVIS
be ignored. Since the facial muscles of the possum each form discrete dermal bands, it is unlikely that spread of enzyme from the injected muscle to another is the cause of overlap. The fact that no overlap of labelled subnuclei occurred following injection of facial muscle groups derived from the different embryological layers suggests that functional specialization within the nucleus is complete for muscles of different embryological origin. The persistence of overlap in the subnuclei labelled following injection of the sphincter colli profundus layer derivatives implies that for the motorneurons innervating muscles of a single layer, segregation within the nucleus is incomplete. The presence of varied labelling intensity in these subnuclei following a single muscle injection implies that at least some segregation has, however, occurred. It is on this basis that the functional association between the dorsal and middle intermediate, and lateral subnuclei, and the facial muscles was made in figure 5. Although it is clear that subnucleationon a functional basis is present in the facial nuclei of mammals, the functions of these subnuclei are not mutually exclusive in the possum. The only apparent form in which exclusive subnucleation is present is that which occurs in relation to the embryonic subdivisions of the muscle of the hyoid arch -the undifferentiated platysma and sphincter colli profundus, which subsequently gives rise to the facial musculature of the adult.
Service for their co-operation in obtaining animals. LITERATURE CITED
Courville, J. 1966 The nucleus of the facial nerve; the relation between cellular groups and peripheral branches of the nerve. Brain Res., 1; 338-354. De Vito, J. L., K. W. Clausing and D. A. Smith 1974 Uptake and transport of horseradish peroxidase by cut end of vagus nerve. Brain Res., 82; 269-271. Dom, R., W. Falls and G. F. Martin 1973 The motor nucleus of the facial nerve in the opossum Didelphis marsupidis oirginiuna)Its organisation and connections. J. Comp. Neur., 152: 373-402. Graham, R. C., and M. J. Karnovsky 1966 The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney: ultrastructural cytochemistry by a new technique. J. Histochem. Cytochem., 14: 291-302. Huber, E. 1931 The Evolution of Facial Musculature and Facial Expression. John Hopkins Press, Baltimore. Jones, E. G., and R. Y. Leavitt 1974 Retrograde axonal transport and the demonstration of nonspecific projections to the cerebral cortex and striatum from thalamic intralaminar nuclei in the rat, cat and monkey. J. Comp. Neur., 154: 349-378. Kristensson, K., and Y. Olsson 1971 Retrograde axonal transport of protein. Brain Res., 29: 363-365. La Vail, J. H., and M. M. La Vail 1972 Retrograde axonal transport in the central nervous system. Science, 176; 1416-1417. La Vail, J. H., K. R. Winston and A. Tish 1973 A method based on retrograde intra-axonal transport of protein for identification of cell bodies of origin of axons terminating within the CNS. Brain Res., 58: 470-477. Papez, J. W. 1927 Subdivisions of the facial nucleus. J. a m p . Neur., 43: 159-191. Szentagothai, J. 1948 The representation of facial and scalp muscles in the facial nucleus. J. Comp. Neur., 88: 207-220. Vraa-Jensen, G. F. 1942 The Motor Nucleus of the Facial Nerve. With a Survey of the Efferent Innervation of the Facial Nucleus. Ejnar Munksgaard, ACKNOWLEDGMENTS Copenhagen. C. 1971 Microelectrode delineation of I wish to thank Doctor Charles Watson Welker, fine grain somatotopic organisation of SMI cerebral for his advice and encouragement, both in neocortex in albino rat. Brain Res., 26; 259-275. execution of experimental procedures and Weller, W. L. 1972 Barrels in somatic sensory neocortex of the marsupial Trichosurus oulpecula in the preparation of the manuscript. Also (brush-tailed possum). Brain Res., 43: 11-24. Mr. Brian Freeman for his interest and adT. A., and H. van der Loos 1970 The strucvice and Mr. Des Shimeld for assistance Woolsey, tural organisation of layer IV in the somatosensory with the anesthetics and operations. 1 am region 61) of mouse cerebral cortex. Brain Res., 17; grateful to The National Parks and Wildlife 205-242.
