Functional somatotopic organisation of motoneurons supplying the rabbit masseter muscle

THE JOURNAL OF COMPARATIVE NEUROLOGY 3643279-289 (1996)

Functional Somatotopic Organisation of Motoneurons Supplying

the Rabbit Masseter Muscle

W.A. WEIJS Department of Functional Anatomy, Academic Center for Dentistry Amsterdam (ACTA),

1105 M Amsterdam, The Netherlands

ABSTRACT The localisation within the trigeminal motor nucleus of motoneurons supplying different

regions of the rabbit masseter muscle was investigated to test the hypothesis that muscle regions with different motor tasks are controlled from different subregions of the motor nucleus. Motoneurons were labeled retrogradely with horseradish peroxidase, applied surgi- cally to small sections of the masseter in 22 animals, and also by applying this tracer to the cut masseteric nerve. After sacrifice, the labeled muscle sections were mapped. The distribution of labeled motoneurons within the nucleus was described and compared for the muscle regions. The motoneurons for the masseter muscle are confined to the dorsal and lateral sections of the motor nucleus, along its full rostrocaudal extent. Within this subnucleus, the motoneurons for the superficial masseter occupy the dorsolateral portion, the motoneurons for the deep masseter the dorsomedial portion. The anatomical and functional subdivision of the deep masseter into an anterior and posterior portion appeared to be matched by a separation of the motoneurons for these portions in the rostrocaudal direction along the nucleus. The separation of the motoneurons for the anterior and posterior deep masseter is not complete; the territories in the motor nucleus overlap each other for about 50%. The well-established differentiation in motor tasks between the masseter portions during feeding is thus clearly reflected in a separation of motoneurons, making possible differentiation of descending or afferent input to the separate regions in the nucleus. 1!19(i Wiley-Lish, Inc.

Indexing terms: somatotopg, compartments, HRP, motor control, trigeminal motoneurons

During mastication, the mammalian jaw muscles do not simply fire in an alternation of openers and closers; the jaw closers show an elaborate differentiation of motor behaviour (for review, see Weijs, 1993). This is reflected in the rather fixed, somatotopic fashion in which the trigeminal motor nucleus is organised. The motoneurons for the various jaw closers, i.e., the masseter, temporalis, medial pterygoid, and lateral pterygoid muscles, occupy character- istic regions in the dorsolateral section of the nucleus (for review, see Travers, 1985).

The masseter muscle itself is multilayered and not active as a single unit. The fleshy layers between aponeuroses (tendon sheets) will be referred to in this paper as compartments. Groups of interconnected compartments with simi- lar fiber orientations will be called muscle regions. In man (McMillan and Hannam, 1991), pig (Herring et al., 1989), and rabbit (Weijs et al., 1993), the masseter motor unit territories are very small and confined to single compartments. They make possible the differences in motor activity that have been shown to be present between superficial and deep and anterior and posterior muscle regions (rabbit:

Weijs and Dantuma, 1981; pig: Herring and Wineski, 1986; man: Blanksma et al., 1992; McMillan and Hannam, 1991). Apparently, the masseter motoneurons are segregated into a number of pools that can be differentially activated to execute various tasks .

In this paper, the hypothesis is tested that these motoneuron pools are also spatially separated within the masseter area of the trigeminal motor nucleus. A functional organisation of motoneurons was recently proposed for the hypoglossal motor nucleus (Sokoloff and Deacon, 1992) and some leg muscles (see, e.g., Weeks and English, 1985) and could facilitate the organisation of a differential descending or afferent input to these motoneurons. To investigate the degree of topological organisation, neurons innervating different compartments of the masseter were labeled with a retrograde neuronal tracer, horseradish peroxidase (HRP),

Accepted July 13, 1995. Address reprint requests to Prof. Dr. W.A. Weijs, Department of Func-

tional M o r p h o l o ~ , P.O. Box 80.157,3508 TD Utrecht, The Netherlands.

c 1996 N'ILEY-LISS, INC.

280 W.A. WEIJS

selectively applied in small quantities. After a short survival period, labeled muscle sections and labeled motoneurons in the brainstem were mapped. The results show that, within the masseteric subnucleus of the trigeminal motor nucleus, a clear somatotopic gradient could be discerned, corresponding to the functional subdivision of the muscle during mastication.

MATERIALS AND METHODS Materials

Experiments were conducted on 26 young adult male rabbits (New Zealand White), ranging in weight between 2.0 and 2.5 kg. In the last 8 hours of the 30 hour survival period, most of these animals were also used for recording masseter unit properties by means of stimulation of trigeminal motoneurons on the side, contralateral to the side on which HRP was applied. The procedures for these experiments did not interfere (except for the anaesthesia) with any of the structures involved in the uptake and transport of HRP described in this paper.

Surgery For implantation of HRP, the animals were anaesthe-

tised with 10 mg fluanison and 0.315 mg fentanylcitrate per kilogram body weight. A small incision of the skin overlying the masseter exposed it partially. In the first experiments, small quantities (0.1-0.5 p1) of an aqueous solution of HRP (10%) were injected into the muscle with a Hamilton syringe, but it appeared that HRP spread rather widely over the muscle. In the subsequent experiments, a tiny piece ( < 0.5 mm diameter) of gelatin foam soaked in HRP (Sigma type VI) or wheat germ agglutinin WGA-HRP (Sigma) was inserted via a hollow needle and a plunger (outer diameter 1.0 mm). This resulted in less spreading of label across the muscle. The applications were aimed to label small sections of each of the muscle regions listed below. After a survival period of 24-30 hours, the animals were killed by intrave- nous administration of an overdose of pentobarbital. They were perfused via the left ventricle with phosphate-buffered saline (pH 7.4) at body temperature, followed by chilled fixative containing 1% paraformaldehyde, 1.25% glutaralde- hyde, buffered with phosphate at pH 7.4. After 1 hour of fixation, the animals were perfused with a chilled phosphate- buffered solution containing 10% sucrose. The brains and masseter muscle plus zygomatic arch were removed and stored overnight in the sucrose solution.

Transverse (axial) serial cryostat sections of 40 pm were cut from the brainstem from the level of the mesencephalic nucleus throughout the facial motor nucleus. The sections were stained for HRP using the tetramethylbenzidine procedure (Mesulam, 1978) and counterstained with neu- tral red. Sections were checked for the presence of HRP- labeled neurons using both bright- and darkfield microscopy. The labeling was usually heavy (Fig. 2A,B) but was sometimes light. Lightly labeled neurons were always in- cluded in this study as long as it could be ascertained that the HRP was located inside the cell. To verify the success of labeling the muscle, horizontal 10 pm cryostat sections of the masseter muscle were cut at every 1 mm. The muscle fibers and extracellular space were checked for the presence of HRP particles using bright- and darkfield illumination.

To describe the shape of the motor nucleus and the territory of the masseter motoneurons, we attempted to

label the maximum number of masseter motoneurons in three animals. In one rabbit, eight pieces of HRP were placed, evenly spaced, over the muscle, and the motor nucleus was reconstructed from sagittal serial sections. In two other animals, the masseteric nerve was exposed under anaesthesia and soaked for 2 hours in a 10% HRP solution, and reconstructions were made from transverse and sagittal serial sections from one animal each.

Microscopy The position of the labeled brainstem neurons in the

approximately 50 transverse serial sections containing the trigeminal motor nucleus was marked in a single coordinate system with the help of a camera lucida, mounted on a microscope. First, under low magnification, the plot was rotated until its y-axis coincided with the median plane of the slide. Then, under higher magnification, the stage of the microscope was translated until the center of the motor nucleus coincided with the origin of the coordinate system. For this purpose, the plot had been provided with concen- tric circles so that the motor nucleus could be centered on the plot.

The trigeminal motor nucleus is a more-or-less cylindri- cal structure, so the position of the labeled neurons in all transverse sections was compressed into a single cross section. The rostrocaudal position of the neurons was thereby marked, so that it could be checked whether the distribution of neurons in the transverse plane changed from rostra1 to caudal. Because the long axis of the nucleus makes an angle of about 30" with the long axis of the brainstem, the compressed plots are slightly distorted (see Discussion).

The sections of the muscles were first traced at low magnification with the help of a photographic enlarger and then studied under the microscope at intermediate magnification. The areas containing HRP particles in the extracellular and/or intracellular space were marked in the trac- ings.

Masseter anatomy For a detailed account of the regional and compartmental

subdivisions, see Weijs and Dantuma (1981) and Weijs et al. (1987); the following description and Figure 1 are restricted to five muscle regions. The superficial masseter (MSS) originates from the anterior crest of the zygomatic arch and diverges to the rim of the angular process. The masseter pars reflexa (MSPR) originates from a long tendon, arising from the anterior crest, and inserts to the medial side of the angular margin of the mandible. The posterior superficial masseter (MSS4) originates posterior to MSS and attaches to the posterior end of the angular process. The middle masseter (not illustrated) lies deep to the superficial masseter, originates from the ventral margin of the zygomatic arch, and attaches to the lateral side of the mandible. The deep masseter is larger and wider than the middle masseter and originates from the medial side of the zygomatic arch. The anterior deep masseter (MPAN) lies anterior to the entrance point of the masseteric nerve and is largely vertically oriented. The posterior deep masseter (MPPO) lies posterior to this nerve and emerges to the surface, posterior to the superficial masseter. Its fibers pull the jaw backward.

MOTONEURONS OF RABBIT MASSETER 881

r'

C

MPPO

MSS4

MSS

MSPR

C

MPPO

. MSPR

Fig. 1. Superficial (top) and deep (bottom) dissection of the rabbit left masseter muscle. Anterior is to the left. The top diagram shows the superficial masseter (MSS) fibres converging to a tendon sheet, at- tached to the anterior end of the zygomatic arch. Behind MSS is the posterior superficial masseter (MSS4). The origin of the posterior deep masseter (MPPO) emerges from behind MSS4. The bottom diagram shows the anterior (MPAN) and posterior (MPPO) masseter in lateral view, after removal of the superficial and middle masseter. In both diagrams the pars reflexa of the superficial masseter (MSPR) is shown. c, Mandibular condyle; m, mandible; z, zygomatic arch.

Statistical analysis The accuracy of the measurements of neuron position in

the transverse plane was determined as follows. The plot- ting procedure for each of six labeled neurons in different animals was repeated five times, and neuron coordinates were determined with the X-Y tablet. The measurement error was calculated as the standard deviation of the five measurements. For the six neurons, the SD varied from 14 to 54 km. The pooled SD was 44 and 41 pm, for the dorsoventral and lateromedial axes, respectively. This is less than 4% of the size of the masseteric subnucleus.

Differences between the distribution of motoneurons in the transverse plane were tested by considering their dorsoventral and mediolateral coordinates as variables. Hotelling T2 and Mahalanobis D2 were used for bivariate estimates of distance, to test the null hypothesis that the distributions of the neurons of the masseter regions in the transverse plane were the same. Differences in distribution of neurons along the rostrocaudal axis were investigated using ,y2 tests. First, the null hypothesis was tested that the distribution along the rostrocaudal axis of the masseteric neurons was equal to that of all trigeminal motoneurons. Second, the null hypotheses were tested that the motoneurons of each of the five masseter regions had the same

distributions along the rostrocaudal axis as the entire population of trigeminal motoneurons.

RESULTS General findings

Figure 2A,B,D shows examples of labeled motoneurons in the trigeminal motor nucleus resulting from tracer applications to compartments of the masseter. The labeled neurons occupy the dorsal and dorsolateral one-half of the motor nucleus. The perikaryon, and also the dendrites, were normally heavily labeled, in both HRP and WGA- HRP experiments. Figure 2C shows that, after the applica- tion of HRP to the cut end of the masseteric nerve, many neurons in the dorsal section of the nucleus take up HRP. Labeled axons frequently could be seen lateral to the motor nucleus.

In 22 animals the tracer had remained confined to a small area of the masseter muscle (see, e.g., Fig. 3). In these animals 10-123 (mean 40 ? 6 SD) neurons, all situated in the trigeminal motor nucleus, were found to be labeled. No other labeled cells were found within the brainstem. In four other animals, HRP had spread too widely throughout the muscle to be useful; two of these animals had been injected with dissolved HRP. Tendon sheets proved to be effective barriers preventing the spread of HRP. These sheets show small gaps for vessels, through which HRP was sometimes found to have penetrated. Within the 24-30 hour survival period, the HRP rarely spread completely within a compartment. In 16 experiments (summarized in Fig. 4), the labeled area of the masseter muscle was found to be limited to the target area, which was either the superficial (MSS), the posterior superficial (MSS41, the anterior deep (MPAN), or the posterior deep (MPPO) masseter. In none of the experiments was the middle masseter labeled exclusively. In the other experiments (data not shown in Fig. 4), the label was found in two of the above-mentioned areas (three cases) or had been applied to the pars reflexa (MSPR), where it was found exclusively (three cases).

Distribution of motoneurons in the transverse plane

Figure 4 shows the distribution of labeled motoneurons and the labeled muscle regions in the above-mentioned 16 experiments. The cross section of the muscle where the largest territory was occupied by HRP is represented in the Figure 4. In none of the experiments did the area occupied by the labeled motoneurons change in a consistent fashion, in going through the subsequent sections, i.e., from rostra1 to caudal. Therefore, the data from individual animals were collected in a single cross section (see Materials and Meth- ods).

For the superficial masseter (Fig. 4A), the labeled neurons occupied almost exclusively the dorsolateral sector of the cross-sectional area. Very different areas of the superficial masseter were labeled, but the neurons innervating these areas were all located in the same general area of the nucleus.

The motoneurons supplying the posterior deep masseter (MPPO) differ greatly in position from the previous ones: They occupy the dorsomedial sector and the central portion of the ventromedial sector of the motor nucleus. There is

282 W.A. WEIJS

Fig. 2. A Transverse brainstem section, showing most of the trigeminal motor nucleus. Lateral is to the left, dorsal to the top. Three HRP-labeled neurons (arrows) are situated in the dorsal section of the nucleus, after labeling of the superficial masseter (MSS). B: As in A labeled motoneurons now in the dorsomedial section, after labeling of

the posterior deep masseter (MPPO), as shown in Figure 3. C : As in A, but only part of the motor nucleus is shown. Large amounts of HRP in dorsally located motoneurons after labeling of the masseteric nerve. D: Detail of an HRP-labeled motoneuron. Scale bars = 0.1 mm in A-C, 25 IJ-m in D.

only slight overlap between the neurons innervating the posterior deep and superficial masseter. The labeled neurons in one experiment (squares in Fig. 4) have a more

central location in the nucleus than the neurons in the other two experiments, but this difference does not corre- late with a difference in labeled muscle section.

MOTONEURONS OF RABBIT MASSETEK

Fig. 3. Horizontal section of rabbit masseter, below the zygomatic arch, showing the HRP-labeled muscle area, stained darkly. The diagram at right shows that the HRP is confined mostly to the lateral portion nf the posterior deep masseter (MPPO). The solid black

The anterior deep masseter iMPAN) was labeled in four experiments, and in each of them the majority of its fibers was involved. There is little difference in neuronal labeling between the four experiments. Although the area of the nucleus representing the anterior deep masseter shows a large overlap with that of the posterior deep masseter, it can differ from the latter by including the central area of the dorsolateral sector. This, however, was seen in only two of four experiments.

The posterior superficial masseter iMSS4) appears to be innervated by motoneurons in a position exactly intermediate between the posterior deep and superficial masseter. The labeled areas in four experiments were small and differed somewhat from one another. Nevertheless, the territories the neurons occupy within the motor nucleus are largely the same. Note that the size of these territories

P

283

MPPO

MSS

MSS4

MSM

HRP

A

structures in the diagram are the aponeuroses; the patterns indicate the muscle reaons. The anterior deep masseter (MPAN) is not shown; it is situated dorsal to this section. A, anterior; L, lateral; P, posterior; M, medial.

differs markedly but is unrelated to the size of the labeled area.

Figure 5 summarises the data by providing the mean position of the motoneurons in the 16 experiments represented in Figure 4 and, in addition, in one extra experiment involving the MPPO and three experiments involving the pars reflexa (MSPR). In the latter, most of the pars reflexa appeared to be labeled, whereas no HRP was found in the rest of the masseter. The neurons innervating MSPR share most of their territory with those innervating the superficial masseter but differ slightly in average location. In using Hotelling T2 tests, it appeared that the distributions of the motoneurons (pooled for the different experiments) of the MSS, MSS4, MPAN, MPPO, and MSPR regions were statistically significantly different from each other ( P < 0.0001). I t appeared that, for the total group of 22 experi-

284 W.A. WEIJS

D B D

MSS

v M O O H D

c D

L

I I

I MPAN

V M 0 . H

M

Fig. 4. Transversal plane plots of the trigeminal motor nucleus, showing distribution of labeled motoneurons. The circles that can be inscribed in the squares presented match the cross section of the motor nucleus. The results of 16 experiments are shown separately, grouped according to masseter regions. A MSS (superficial masseter). B: MSS4 (posterior superficial masseter). C : MPAN (anterior deep masseter). D: MPPO (posterior deep masseter). The horizontal muscle cross sections of the corresponding experiments are shown below the plots. The

N fl

"

MSS4

D

v moo^

D

M

v m o a

selected sections are the ones with the maximal HRP-labeled area (hatched). The sections are orientated as indicated in the lower right corner. The continuous contours refer to the free muscular borders, the dashed contours to muscular insertion to the mandible, the lines inside the muscle to sectioned tendon sheets. Different symbols refer to muscle maps and neuron positions in particular animals. L, lateral; D, dorsal; M, medial; V, ventral; A, anterior; P, posterior.


0 +

0 .L

O%

0

L

superficial (MSS4) masseter motoneurons did not differ significantly from those of all trigeminal motoneurons. However, the distributions of both deep masseters did differ significantly ( x 2 test, 11 df, P < 0.001 for MPPO, P < 0.01 for MPAN). Figure 8 shows that the motoneurons innervating the posterior deep masseter are situated relatively rostrally in the motor nucleus. Those of the anterior deep masseter are more caudally situated. The histograms of MPPO and MPAN do overlap, but only by 50%.

+ ... e x

X

+ X

“ M

0 MSS + MSS4 0 MSPR

MPAN x MPPO

V Fig. 5. Average position of neurons in 20 experiments in a trans-

verse section of the trigeminal motor nucleus for the superficial (MSS), posterior superficial (MSS4), anterior deep (MPAN), and posterior deep (MPPO) masseter and the masseteric pars reflexa (MSPR). Note the difference in scale from Figure 3; the more central part of the nucleus is represented.

ments, there was no statistically significant correlation between the size of the labeled area in the muscle cross section and either the number of labeled neurons or the size of the area the labeled neurons occupy within the cross section of the motor nucleus.

Rostrocaudal gradient in motoneuron distribution

The distribution within the trigeminal motor nucleus of the motoneurons innervating the masseter muscle is represented in Figures 6 and 7, which show the three subdivisions of the nucleus (a , @, y, according to Meessen and Olszewski, 1949). The a subnucleus is a mediocaudally pointing strand of neurons, linking the trigeminal and facial motor nuclei. The p subnucleus forms the ventromedial portion and the y subnucleus the dorsal and lateral portion of the motor nucleus. Furthermore, “cell group k” containing small neurons (see Donga et al., 1992) is shown. The masseteric motoneurons are distributed over the y-portion of the nucleus, and there is no clear gradient in numbers of labeled neurons from rostral to caudal.

The rostrocaudal distribution of all trigeminal motoneurons in the p and y subnuclei was determined in transverse sections by counting large perikarya with nuclei (assumed to be trigeminal motoneurons) in two control animals (no label applied); 1,770 and 2,220 neurons were counted, respectively. The peak number of neurons per cross section is situated in the rostral one-half of the nucleus (Fig. 8). The rostrocaudal distribution of the 600 HRP masseteric motoneurons labeled in our experiments (Fig. 8) did not differ significantly from the one representing all (trigeminal) y and p motoneurons ( x 2 test, P < 0.01, 11 degrees of freedom). Again using a x2, it was checked whether the rostrocaudal distributions of motoneurons innervating particular areas of the masseter muscle differed from the

DISCUSSION Technical aspects

The procedure for determination of neuron position had an error of 40-45 pm, less than 4% of the size of the masseteric area of the motor nucleus and much smaller than the magnitude of differences between the mean locations of the neurons for the different masseteric regions. The distortion produced by superimposing the serial transverse sections is larger. Neurons located predominantly in the caudal section of the nucleus would attain a position somewhat more dorsally than neurons in the rostral section. However, the neurons of the subnuclei for MSS, MSS4, and MSPR and of the deep masseter are distributed in the same way along the rostrocaudal axis, such that their relative positions in the transverse plots remain unaffected. The MPAN and MPPO neurons do differ in their rostrocaudal distribution. Therefore, the observed small differences in their dorsoventral position might be attributable to the reconstruction procedure and were not further considered.

The method of establishing the location of the motor neuron pool of a muscle by injecting it with relatively large amounts of liquid HRP has been criticized by Richmond et al. (1978). They found that fascial planes did not sufficiently prevent the spread of HRP into neighbouring muscles. On the other hand, Landmesser (1978), Ruigrok and Crowe (1984), and Haase and Hrycyshyn (1986) demonstrated in their experiments that HRP injected in small quantities usually did not pass the fascial envelope of muscles. Following Werf et al. (1993), we introduced HRP or WGA-HRP into the muscle in the form of one tiny piece of gelatin loaded with solidified HRP. This prevents the uncontrolled spreading of HRP during pressure injection. No labeled neurons were found in the facial motor nucleus, despite the fact that facial muscles are overlying the masseter and had to be damaged. This suggests that the label did not spread outside the masseter muscle.

However, checking the final intramuscular distribution of label has remained vital, insofar as the amount of spread of HRP turned out to be variable. Given the distribution pattern of HRP in the muscle, it was obvious that internal tendon sheets acted as barriers against spreading. Because these sheets are close together, the exact compartment into which the HRP implant is injected cannot be controlled externally. Checking of labeled muscle area was often omitted in previous studies but has been performed, for instance, by Herring (1992) and for tongue muscles by Sokoloff and Deacon (1992). The number of labeled neurons varied rather greatly from experiment to experiment, even in cases in which amounts of labeled muscle tissue were comparable (see, e.g., Fig. 4C, MPAN). No satisfactory

286 W.A. WEIJS

V

Fig. 6. A Lateral view of trigeminal nucleus, based on sagittal sections of the brainstem of an animal in which the masseteric nerve was labeled. B-F HRP-containing neurons (dots) are shown in five combined sections lmedial (B) to lateral (F)] of the nucleus. The a, p,

and y portions of the nucleus have been indicated, and cell group “k” (see text) is shaded. R, rostral; C, caudal; D, dorsal; V, ventral. Scale bar = 1 mm.

Fig. 7. A: Rostra1 view of trigeminal nucleus based on transverse sections showing HRP-containing neurons after labeling of the masseteric nerve. The facial nerve is situated just behind the nucleus. Sections B-E go from rostral to caudal. For further explanation, see legend to Figure 6.

explanation can be given. Possibly the release of HRP from the pellets and final concentration of HRP were influenced by the position of the pellet relative to tendon sheets and fascia.

At the time of sacrifice HRP has spread beyond the limits from which it could have been transported to the brain. It probably did not spread very far, because in the rat HRP is transported from the digastric muscle to the brain within 6


Trigeminal and Masseter

o.2 - I i I

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 rostral caudal

0.3

u)

g 0.2 a, C L

C .-

0.1 E U.

0

\n

0

2 4 6 8 10 12 rostral caudal

Fig. 8. Rostrocaudal distribution of motoneurons in the p and y portions of the trigeminal motor nucleus. The rostrocaudal length of the nucleus (mean 2.2 mm; 50-60 microscopic sections) was subdivided in 12 equidistant intervals. The top left diagram shows that the distribution of all (unlabeled) trigeminal motoneurons (heavily shaded) does not differ from that of the masseteric neurons (lightly shaded).

hours (Kemplay and Cavanagh, 1983). If, in a worst case, HRP keeps spreading from the injection focus with a constant speed for 30 hours, only HRP in the outer %, or 20%, rim of the diameter of the labeled muscle area would have had no chance to reach the brain. For this reason, situations in which different muscle regions are represented by different regions in the motor nucleus should be carefully interpreted. However, it quickly became clear that as long as marked muscle compartments are within one of the following regions-MSS, MSS4, MPAN, MPPO- labeled regions of the motor nucleus were identical. This lack of somatotopia within regions is convincing insofar as the area from which HRP has reached the motoneurons must be even smaller than the HRP-stained area in the muscle.

Gordon et al. (1991) rightly argue that fluorescent labeling of muscle nerve branches is superior to direct intramuscular HRP labeling. Uncontrolled spreading of label is prevented, and double-labeling can be used to contrast muscle regions in single individuals, circumventing interin- dividual variation. The technique is more sensitive and better able to pick up slight differences in neuronal distribu-

MPPO and MPAN 0.2 1

u)

e

'ij 0.1 a,

C ._ c

m L;

n U

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 rostral caudal

0.3

u) S ; 0.2 a, c L

C ._ p 0.1 I;

0

MPAN (4 exp)

R

0 2 4 6 8 10 12 rostral caudal

The top right diagram shows a great difference in rostrocaudal distribution of posterior deep (heavily shaded) and anterior deep (lightly shaded) masseter motoneurons. The experiments on which these diagrams are based are represented individually for the posterior deep masseter (bottom left) and anterior deep masseter (bottom right).

tion (Gordon and Richmond, 1991). However, the branch- ing pattern of the masseteric nerve is complex and variable (Herringet al., 1989), and so-called neuromuscular compartments, as defined by their innervation by separate primary nerve divisions (English and Letbetter, 19821, cannot readily be discerned. Therefore, we have chosen the method of direct comparison of position of motoneurons in the brainstem and the region of the muscle these neurons innervate. Our method was obviously inappropriate for labeling cell group k neurons (Donga et al., 1992). The functions of these small neurons (motor, parasympathetic) remain to be clari- fied.

Somatotopic arrangement of motoneurons There are few reports of separation of motoneurons that

innervate distinct subregions of jaw muscles. For the rat, Rokx et al. (1985) describe separate motor areas for superficial and deep masseter, the deep masseter motor area being located between the motor areas of the superficial masseter and temporalis muscle. Studying the rabbit, Matsuda et al. (1978) found the deep masseter motoneurons, medial to the

288 W.A. WEIJS

ones of the superficial masseter area, intermingling with those of the temporalis. Our findings confirm and expand these data, showing that in the rabbit masseter at least three regions exist whose neurons show < 50% overlap: The superficial masseter motoneurons are more laterally located than those of the deep masseter, whereas within the deep masseter the anterior portion is represented more caudally in the nucleus than the posterior portion, again with an overlap of about 50%. Furthermore, our data show that the average locations of two anatomically discrete regions of the superficial masseter, i.e., the posterior superficial masseter and the pars reflexa, differ slightly from the location of the rest of the superficial masseter.

Therefore, the motoneurons of the rabbit masseter are arranged according to muscle region, but this somatopic arrangement is less distinct than the arrangement of motoneurons for different muscles. This idea is confirmed by experiments conducted by Herring (1992) on pig masseter. She found a rostrocaudal gradient in representation of motoneurons, innervating rostrocaudal regions of the muscle, with a very large overlap. In other complex muscles, a somatotopic distribution of motoneurons according to muscle region can sometimes be seen, for instance, in the dog pectoralis muscle (Krogh and Towns, 1986) and in the cat gastrocnemius (Weeks and English, 1985).

Our findings seem to be contradicted by preliminary results from Widmer et al. (19951, who double labeled different primary nerve branches of the rabbit masseter with fluorescent tracer and found no difference in localiza- tion in the brainstem of neurons reaching the muscle via different primary branches. A full explanation must await publication of their final data. We suspect that the difference may be due either to the incomplete separation of neurons innervating different regions demonstrated here or, alternatively, to the primary nerve branches containing axons for different muscle regions. For instance, more distal branches may reach both the superficial and the deep anterior portions of the masseter. These regions have different representations in the masseter area of the motor nucleus.

Functional significance of motoneuron separation

The partial spatial separation of motoneurons into “subnuclei” innervating superficial, anterior deep, and posterior deep masseter found in the present study is closely paral- leled by different masticatory motor patterns of these subregions. The superficial masseter, including its pars reflexa shows no anteroposterior differences in timing of electromyographic (EMG) bursts but differs markedly from the deep masseter. In the deep masseter, motor behaviours of the anterior and posterior portions are clearly different (Weijs and Dantuma, 1981; Weijs et al., 1989). There was no sharp transition in activity pattern from the superficial to deep portion but, rather, a gradual change. The motor units of the masseter muscle occupy very small territories and usually remain confined within compartmental borders (Weijs et al., 1993; Kwa et al., 19951, making possible the fine differentiation of motor behaviour. The observed subdivision in function and innervation corresponds to differences in muscle spindle and fiber type composition. From superficial to deep and from posterior to anterior, there is a marked increase in the amount of slow muscle fibers (Bredman et al., 1990) and muscle spindles (Bredman et al., 1991). The fiber type and spindle distribution, small size of

the motor unit territories, and differential motor behaviour in the masseter muscle are found not only in rabbits but in most mammals (for review, see Weijs, 1993).

The question of whether the spatial separation of motoneurons innervating different functional subregions of a muscle makes possible a differentiated input has yet to be answered. For spinal motoneurons, a rostrocaudal separation between subnuclei might lead to a differentiation of afferent input (for discussion, see Iliya and Dum, 1984; Weeks and English, 1985). Even if statistically significant differences in distribution of motoneurons are present, the overlap between subnuclei is often large (see the references cited above and Gordon and Richmond 1990,1991; Gordon et al., 19911, and it is difficult to envisage how such a partial separation can lead to totally different behaviour of motoneuron pools. Gordon and Richmond (19911, for instance, reject the idea that motoneuron position is a critical factor underlying the specialized actions of different neck muscles.

For the jaw muscles and also for the tongue muscles (Sokoloff and Deacon, 19921, a complex spatial separation is seen in the trigeminal and hypoglossal motor nuclei, corresponding to a functional arrangement. On the other hand, the dendritic trees of cat masseter motoneurons have been proved to be quite large and able to accept input in a large subvolume of the trigeminal motor nucleus (Shigenaga et al., 1988). The question of whether the topographical arrangement of trigeminal subnuclei is therefore merely a consequence of a developmental process, unrelated to muscle function, or determines at least part of the observed gradual differences in motor behaviour in the masseter muscle cannot be solved until further information is avail- able on the distribution of input to these motoneurons.

ACKNOWLEDGMENTS I am very grateful to Dr. The0 van Eijden and Dr. Frits

Jiich for their valuable comments on the manuscript; to Hans Korfage, Leo van Ruijven, Henny Rolleman, and Peter Brugman for their technical assistance; and to Scho- lastica Kwa, Rob Minkels, and Stan Turkawski for their help with the experiments.

LITERATURE CITED Blanksma, N.G., T.M.G.J. van Eijden, and W.A. Weijs (1992) Electromyo-

graphic heterogeneity in the human masseter muscle. J. Dent. Res. 71:47-52.

Bredman, J.J., W.A. Weijs, A.F.M. Moorman, and P. Brugman (1990) Histochemical and functional fibre typing of the rabbit masseter muscle. J. Anat. 168:31-47

Bredman, J.J., W.A. Weijs, and P. Brugman (1991) Relationships between spindle density, muscle architecture, and fibre type composition in different parts of the rabbit masseter. Eur. J. Morphol. 29:297-307.

Donga, R., R. Dubuc, A. Kolta, and J.P. Lund (1992) Evidence that the masticatory muscles receive a direct innervation from cell group k in the rabbit. Neuroscience 49:951-961.

English, A.W., and W.D. Letbetter (1982) A histochemical analysis of identified compartments of cat lateral gastrocnemius muscle. Anat. Rec. 204:123-130.

Gordon, D.C., and F.J.R. Richmond (1990) Topography in the phrenic motoneuron nucleus demonstrated by retrograde multiple-labelling techniques. J. Comp. Neurol. 292424-434.

Gordon, D.C., and F.J.R. Richmond (1991) Distribution of motoneurons supplying dorsal suboccipital and intervertebral muscles in the cat neck. J. Comp. Neurol. 304:343-356.

Gordon, D.C., G.E. Loeb, and F.J.R. Richmond (1991) Distribution of motoneurons supplying cat sartorius and tensor fasciae latae, demon-


strated by retrograde multiple-labelling methods. J. Comp. Neurol. ,304: 35 7-3 7 2 .

Haase, P., and A.W. Hrycyshyn (1986) On the diffusion of horseradish peroxidase into muscles and the “spurious” labeling of motoneurons. Exp. Neurol. 91:399-403.

Herring, S.W. ( 19921 Muscles of mastication: Architecture and functional organization. In Z. Davidovitch (ed.): The Biological Mechanisms of Tooth Movement and Craniofacial Adaptation. Columbus, OH: Ohio State University Press, pp. 541-548.

Herring, S.W., and L.E. Wineski (1986) Development of the masseter muscle and oral behavior in the pig. J . Exp. Zool. 237:191-207.

Herring, S.W., L.E. Wineski, and F.C. Anapol(1989) Neural organization of the masseter muscle in the pig. J . Comp. Neurol. 280:563-576.

Iliya, A.R.. and R.P. Dum (1984) Somatotopic relations between the motor nucleus and its innervated muscle fibers in the cat tibialis anterior. Exp. Neurol. 86.272-292.

Kemplay, S., and J.B. Cavanagh (1983) Bilateral innervation of the anterior digastric muscle by trigeminal motor neurons. J . Anat. 136:417-423.

Krogh, J .E., and L.C. Towns (1986) Is there morphologxal separation between the spinal cord motor nuclei which innervate the heads of a multiheaded muscle? Brain Res. 369:331-335.

Kwa, S.H.S., W.A. Weijs, and P.J.W. Juch ( 1995) Contraction characteristics and myosin heavy chain composition of rabbit masseter motor units. J . Neurophysiol. 73:538-549.

Landmesser, L. ( 1978) The distribution of motoneurones supplying chick hind limbs. J. Physiol. 284:371-389.

Matsuda, K., M. Uemura, M. Kume, R. Matsushima, and N. Mizuno (1978) Topographical representation of masticatory muscles in the motor trigeminal nucleus in the rabbit: An HRP study. Neurosci. Lett. 8:1-4.

McMillan. A.S., and A.G. Hannam (1991) Motor-unit territory in the human masseter muscle. Arch. Oral Biol. 36:435-441.

Meessen, H., and J. Olszewski (1949) A Cytoarchitectonic Atlas of the Rhombencephalon of the Rabbit. New York: S. Karger.

Mesulam, M.-M. ( 1978) Principles of horseradish peroxidase neurohistochem- istry and their applications for tracing neural pathways-axonal transport enzyme histochemistry and light microscopic analysis. In M.-M. Mesu- lam led. ): Tracing Neural Connections. New York: John Wiley and Sons, pp. 1-151.

Richmond, F.J.R., D.A. Scott, and V.C. Abrahams (1978) Distribution of motoneurones to the neck muscles, biventer cervicis, splenius complexus in the cat. J. Comp. Neurol. IXI:451-464.

Rokx, J.T.M., P.J.W. Juch. and J.D. van Willigen (1985) On the bilateral innervation of masticatory muscles: A study with retrograde tracers. J . Anat. 140.237-243.

Ruigrok, T.J.H., and A. Crowe (1984) The organization of motoneurons in the turtle lumbar spinal cord. J. Comp. Neurol. 228:24-37.

Shigenaga, Y., A. Yoshida, K. Tsuru, Y. Mitsuhiro, K. Otani, and C.Q. Can (1988) Physiological and morphological characteristics of cat motoneurons-Intracellular injection of HRP. Brain Res. 461:238-256.

Sokoloff, A.J., and T.W. Deacon (19921 Musculotopic organization of the hypoglossal nucleus in the cynomolgus monkey, Mucuca fusciculans. J . Comp. Neurol. 324:81-93.

Travers, J.B. (1985) Organization and projections of the orofacial motor nuclei. In G. Paxinos i d ) : The Rat Nervous System. Sydney: Academic Press, pp. 11 1-128.

Weeks, O.I., and A.W. English (1985) Compartmentalization of the cat lateral gastrocnemius motor nucleus. J . Comp. Neurol. 235255-267.

Weijs, W.A. (1993) Evolutionary approach of masticatory motor patterns in mammals. In V.L. Bels, M. Chardon, and P. Vandewalle (eds.): Biome- chanics of Feeding in Vertebrates IComp. Environ. Physiol. 181. Berlin: Springer, pp. 281-320.

Weijs, W.A., and R. Dantuma (1981) Functional Anatomy of the masticatory apparatus in the rabbit t0ryctolugu.s cuniculus L.). Neth. J. Zool. 31:99-147.

Weijs, W.A., P. Brugman, and E.M. Klok (1987) The growth of the skull and jaw muscles and its functional consequences in the New Zealand rabbit (0ryctolugu.s cuniculusl. J. Morphol. 194:143-161.

Weijs, W.A., P. Brugman, and C.A. Grimbergen (1989) Jaw movements and muscle activity during mastication in growing rabbits. Anat. Rec. 224:407W 16.

Weijs, W.A., P.J.W. Juch, S.H.S. Kwa, andJ.A.M. Korfage (19931 Motorunit territories and fiber types in rabbit masseter muscle. J . Dent. Res. 72: 1491-1498.

Werf, v.d.F., B. Baljet, M. Prins, A. Timmerman, and J.A. Otto (1993) Innervation of the superior tarsal (Muller’s) muscle in the cynomolgus monkey: A retrograde tracing study. Invest. Ophthalmol. Vis. Sci. 34:2333-2340.

Widmer, C.G., M. Saad, J .P. Lund, R. Dubuc, and A.W. English ( 1995) Rabbit masseter muscle partitioning is not reflected in the spatial organization of its motoneurons. In T. Morimoto, T. Matsuya, and K. Takada (eds.): Brain and Oral Function. Amsterdam: Elsevier Science, pp. 19-27.

Documents

Functional somatotopic organisation of motoneurons supplying the rabbit masseter muscle