The termination of forelimb nerves in the feline cuneate nucleus demonstrated by the transganglionic transport method

209

Brain Research, 248 (1982) 209-222 Elsevier Biomedical Press

The Termination of Forelimb Nerves in the Feline Cuneate Nucleus Demonstrated by the Transganglionic Transport Method GUNNAR NYBERG and ANDERS BLOMQVIST*

Department of Anatomy, University of Uppsala, Box 571, S-751 23 Uppsala (Sweden) (Accepted March 1lth, 1982)

Key words: transganglionictransport - - horseradish peroxidase - - lectin conjugate - - forelimb nerves - cuneate nucleus - - somatotopy - - modality segregation

The projection of forelimb nerves to the cuneate nucleus was studied in the cat by the transganglionictransport method. The cut ends of the median, ulnar, musculocutaneous, medial cutaneous, lateral brachial and antebrachial cutaneous nerves and of the superficial and deep branches of the radial nerve were exposed to horseradish peroxidase (HRP) or to HRP conjugated to wheat germ agglutinin. Nerves innervatingthe skin terminated in a somatotopical pattern on the cell clusters in the middle region of the cuneate nucleus. Afferents from the paw occupied the largest area and were situated dorsally, with the ulnar part represented medially and the radial part laterally. The palmar side of the digits seemed to be represented superficial to the dorsal side. The projection from the arm was split into a ventromedial and a ventrolateral area. Superimposed on this somatotopy, a reversed termination pattern was also present. Thus medially projecting nerves also hada small separate projection to the lateral part of the nucleus and vice versa. The rostral region of the nucleus was organized in a similar way except for the rostral pole where the somatotopy was lost. The caudal region differed from the middle one in that it appeared to lack representation of the upper and lower arm. The deep branch of the radial nerve terminated in the middle-ventral, 'reticular' region of the cuneate nucleus, with a sparse projection also to the ventral parts of the rostral and caudal regions, includingthe base of the dorsal horn. Also the musculocutaneous, median and ulnar nerves, but not the pure cutaneous nerves, had projections to these areas, indicating a modality segregation in the cuneate nucleus. The rostral pole of the nucleus, however, appeared to constitute an area of overlap between projections from deep and superficial receptors. INTRODUCTION

intimately connected with the t e r m i n a t i o n p a t t e r n of the peripheral nerves, b u t the lack of a suitable

The gracile a n d cuneate nuclei, collectively k n o w n as the dorsal c o l u m n nuclei ( D C N ) , are the first relay in the dorsal c o l u m n - m e d i a l lemniscus pathway, which is one of the m a j o r channels for somesthetic i n f o r m a t i o n to the t h a l a m u s a n d ultimately to the s o m a t o s e n s o r y cortex. The m a i n i n p u t to the D C N comes t h r o u g h p r i m a r y afferent fibers, which are the central processes of the dorsal root ganglia, the distal ones c o n s t i t u t i n g the peripheral nerves. The D C N are characterized by a high degree of spatial a n d t e m p o r a l resolution for tactile and, for the forelimb, kinesthetic i n f o r m a t i o n (for refs., see Brown a n d Gordona). It m a y be assumed that the f u n c t i o n a l properties o f the D C N n e u r o n s are

* To whom all correspondence should be addressed. 0006-8993/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press

n e u r o a n a t o m i c a l m e t h o d to trace c o n n e c t i o n s across the spinal ganglia has m e a n t that it has only been possible so far to study the projection of the dorsal roots (see e.g. refs. 15, 17, 19, 21 a n d 35). However, the recently i n t r o d u c e d m e t h o d of utilizing transganglionic t r a n s p o r t of horseradish peroxidase 16,25 has provided a tool for accurately determ i n i n g the central projection of peripheral nerves. The capacity of this m e t h o d is d e m o n s t r a t e d by the present study, in which, despite considerable distances - - in some cases almost half a meter - - it was possible to trace the projection of cat forelimb nerves to their t e r m i n a l sites in the medulla oblongata. The results provide a n u m b e r of new pieces

210 of information on the organization of the primary afferent input to the cuneate nucleus, and also give some tentative data on that of the external cuneate nucleus. MATERIAL AND METHODS This report is based on experiments on 16 adult cats. The central projection of 8 forelimb nerves (see below) was studied. The nerves were dissected free with the epineurium and accompanying blood vessels intact, and then transected. The proximal stump was dipped into a small cup filled with horseradish peroxidase (HRP) or H R P conjugated to wheat germ agglutinin ( W G A - H R P ) and was exposed for 1 h. The W G A - H R P was prepared as described by Gonatas et al. 12 and Brushart and Mesulam 1°. Great care was taken to avoid leakage of tracer into surrounding tissue, and in none of the experiments reported here did such leakage occur. Following exposure, the nerve stump was gently wiped so as to remove superfluous tracer and then wrapped in a small sheet of plastic (ParafilmR), which was carefully sealed around the stump. The plastic was then fixed in position with a tissue adhesive (HistoacrylR). In 5 cats only one forelimb was subjected to surgery, and in the other 11 both forelimbs. In 10 of the latter cases the same nerve was investigated in both forelimbs, but in one cat (no. 163, see Table I) different nerves were studied on the two sides. Nerves investigated The superficial branch o f the radial nerve, (studied in 3 cats), was sectioned in the proximal part of the lower forelimb. The deep branch o f the radial nerve (two cats) was sectioned just proximal to its entry into the extensor muscles. In two cats the ulnar nerve was divided where it passes over the medial epicondyle of the humerus. In a third experiment a small branch, the palmar digital nerve for the medial side of the fifth digit (n. digitalis palmaris proprius medialis), was studied. In one experiment the median nerve was cut just proximal to the cubital fossa. In two other cats the nerve was sectioned at the wrist, immediately distal to the carpal ligament, and thus after the branches to the antebrachial muscles and the cutaneous branch to the palm had been given off. The musculocutaneous

nerve (3 cats) was sectioned at the middle of the upper arm after the branches to the coracobrachial and biceps muscles had been given off but proximal to the branches to the brachial muscle and elbow joint. The lateral antebrachial cutaneous nerve (one cat) - - the continuation of the musculocutaneous nerve - - was cut just distal to its emergence under the biceps tendon. The medial cutaneous nerve (one cat) was sectioned where it becomes subcutaneous in the distal part of the upper arm. The lateral brachial cutaneous nerve (one cat), the skin branch of the axillary nerve, was sectioned where it emerges from the quadrangular space of the axilla. A survey of the different experiments is presented in Table I. Following survival periods varying between 2 and 3 days (Table I) the cats were fixed by transcardial perfusion with 2.5% glutaraldehyde in phosphate buffer. Fixation was terminated by perfusion with chilled (4 °C) buffer 34, and after removal the brain and upper half of the spinal cord (including the spinal ganglia) were rinsed overnight in buffer containing 30 % (w/w) sucrose. The medulla oblongata and spinal cord were cut transversely into 40/zm-thick sections. Two adjacent sections out of every 10 were incubated for H R P histochemistry with tetramethyl benzidine (TMB) 24. One of the sections was stained with neutral red and the other was left unstained. The spinal ganglia from segments C4 to T2 were cut longitudinally at 40 #m; every section was incubated with TMB and then mounted unstained. Peroxidase activity was charted with an electronic pantograph. RESULTS Following exposure of the cut ends of the investigated nerves to H R P or W G A - H R P , peroxidase activity was found in the spinal ganglia, in the spinal cord and in the main and external cuneate nuclei. The retrograde labeling of ganglion cells, indicating the segmental origin of the nerves, is shown in Table II. The terminationin the spinal cord was outside the frame of the present study and will thus not be reported here. The projection to the main and external cuneate nuclei The labeling in the main and external cuneate

211 TABLE I

Survey of the experiments Case no.

Nerve

Site of exposure

Side of exposure

Tracer

Survival period (days)

151 163 191 193 195 180 207 216 163 178 182 179 189 200 203 201 211

Superficial branch of radial nerve Superficial branch of radial nerve Superficial branch of radial nerve Deep branch of radial nerve Deep branch of radial nerve Ulnar nerve Ulnar nerve Medial palmar digital nerve Median nerve Median nerve Median nerve Musculocutaneous nerve Musculocutaneous nerve Musculocutaneous nerve Lateral antebrachial cutaneous nerve Medial cutaneous nerve Lateral brachial cutaneous nerve

Proximally in lower arm Proximally in lower arm Proximally in lower arm Proximally in lower arm Proximally in lower arm Elbow Elbow Metacarpus Carpal ligament Carpal ligament Upper arm Middle of upper arm Middle of upper arm Middle of upper arm Elbow Distally in upper arm Shoulder

right left right left right right right right left right right right right right right right right

33 ~ 33 ~ 20 ~ 20~ 20 ~ 33 ~ 20 ~ 20 ~ 33 ~ 33 ~ 33 ~ 33 ~ 20~ 20 ~ 20 ~ 20 ~ 20~

2 2 3 3 3 3 3 3 2 3 2.5 3 3 3 3 3 3

÷ left ÷ + + ÷

left left left left

+ left

+ + -÷

left left left left

HRP HRP HRP HRP HRP HRP WGA-HRP WGA-HRP HRP HRP HRP HRP HRP HRP WGA-HRP WGA-HRP WGA-HRP

nuclei was very consistent from one experiment on

(see Table I), the contralateral m e d u l l a was always

the same nerve to another. Labeling of the same

free from peroxidase activity.

nerve in b o t h forelimbs thus resulted in virtually identical t e r m i n a t i o n patterns o n either side in the b r a i n stem, b u t there was n o significant difference

I n 5 cats W G A - H R P was used (see Table I). A consistent impression was o b t a i n e d that this tracer

between different cats either. The intensity of the labeling varied, however, a n d it was f o u n d that 3 days of survival was generally superior to two days (cf. Arvidsson a n d GobelS). I n the experiments where only one of the forelimbs was operated o n TABLE II

Retrograde labeling of ganglion cells Nerve

Labeled ganglia

Superficial branch of the radial nerve Deep branch of the radial nerve* Ulnar nerve Median nerve Musculocutaneous nerve Lateral antebrachial cutaneous nerve Medial cutaneous nerve* * Lateral brachial cutaneous nerve

C6-T1 C7-T1 C8-T2 C6-TI C6-C7 C6-C7 T 1-T2 C6-C7

* In one experiment (cat 193) a few labeled cells were also found in the C6 ganglion. ** One or two cells were labeled bilaterally in each of ganglia C6-C8

p r o d u c e d more intense labeling t h a n H R P . However, in the experiments on the u l n a r nerve, where H R P was used in one cat a n d W G A - H R P in a n o t h e r (see Table I), there was no difference in the distribution o f peroxidase activity in the b r a i n stem.

The superficial branch o f the radial nerve. Peroxidase-positive structures suggestive of t e r m i n a l labeling were seen along the whole extent of the cuneate nucleus (Fig. 1). The most intense labeling was f o u n d in the middle region (situated from the obex a n d 4 m m caudally; levels 2 - 4 in Fig. 1), where there was a dense projection to the dorsal half of the nucleus, covering the central two-thirds of its mediolateral extent. The labeling was entirely confined to the cell clusters characteristic for this part of' the cuneate nucleus19, 21. As shown in Fig. 2, the peroxidase granules covered the clusters, whereas the intercluster areas were almost free from peroxidase activity. The labeling did n o t involve the dorsal-most clusters (Fig. 1, level 3) a n d also spared the outer m a r g i n a l zone, which consists of fusiform n e u r o n s oriented along the circumference of the nucleus. The absence of projection to the m a r g i n a l zone was a

212 S U PE R F I C I A L BRANCH OF THE RADIAL N.

DEEP B R A N C H OF THE RADIAL N.

©

(,_(-

(..i~,~:~ ~'

:

~.

GN

SPIN V

J

SPIN V

i

Fig. 1. Drawings of transverse sections through the lower brainstem and the uppermost part of the spinal cord, showing the extent of peroxidase activity (dots) following exposure of the superficial and deep branches of the radial nerve to H R P (for details, see text). Sections 1-6 represent progressively more rostral sections. The distance from the obex ( = section 4) is 6.8 m m for section 1, 3.6 mm for section 2, 2.0 m m for section 3, 1.2 for section 5 and 2.4 m m for section 6. Abbreviations: AP, area postrema; ECN, external cuneate nucleus; GN, gracile nucleus; LCN, lateral cervical nucleus; SPIN V, spinal trigeminal nucleus; TS, solitary tract; X, dorsal motor nucleus of the vagus; XII, nucleus of the hypoglossal nerve; Z, nucleus Z.

213

Fig. 2. Photomicrograph showing two anterogradely labeled cell clusters in the cuneate nucleus after exposure of the superficial branch of the radial nerve to HRP. The section has been lightly stained with neutral red. x 150. consistent finding for all the investigated nerves. No labeling was present in the ventral, 'reticular' part 19, zl of the middle region. In addition to the centrally located main terminal field, a small projection area was also seen close to the border to the gracile nucleus at the level of the obex (Fig. 1, level 4). In the rostral region the cell clusters are gradually replaced by more irregularly arranged neurons giving the cytoarchitecture a reticular appearanceS9, 21. At least in the caudal part of this region, however, the same patch-like termination pattern was present as in the middle region (Fig. 1, level 5), although these rostral patches did not correspond to any cytoarchitectonically discernible cell groups. Initially the labeled area had about the same location as in the middle region, but near the rostral end of the nucleus it moved laterally towards the border against the external cuneate nucleus (Fig. 1, level 6). In the caudal region (Fig. 1, level 1) the labeling was very sparse and mostly only a few peroxidase granules per section were seen. No labeling was found within the external cuneate nucleus. Exposure of the deep branch of the radial nerve to H R P yielded much fainter labeling in the cuneate nucleus than did exposure of the superficial branch. The peroxidase activity was entirely confined to the ventral part of the nucleus (Fig. 1). In the middle

region (Fig. 1, levels 2-4) it was found exclusively in the ventral, 'reticular' zone, with no labeling in the dorsal, 'cell nest' part, and consequently the terminal areas of the deep and superficial branches of the radial nerve appeared to be completely separated. Also in the rostral region of the nucleus the projection of the deep branch seemed to be separated from that of the superficial branch (Fig. 1, levels 5-6). However, since there was no cytoarchitectonically distinguishable border between the two terminal areas, overlap could not be excluded, but if there was any, it was insignificant. In the caudal region very sparse labeling was found at the base of the nucleus (Fig. 1, level 1). The caudal-most pole appeared to be unlabeled, but a few peroxidase granules were seen in lamina VI of the dorsal horn, just ventral to the border against the cuneate nucleus. Fairly intense labeling was present in the medial part of the external cuneate nucleus (Fig. 1). The ulnar nerve projected to the whole rostrocaudal extent of the cuneate nucleus. At all levels except for the rostral pole it terminated in the medial part of the nucleus (Fig. 3); however, in addition a small laterally located projection, extending rostrally from a site 2 mm caudal to the obex, was also observed (Fig. 3, levels 3-5, and Fig. 4b). The labeling was most intense in the middle region (Fig. 3, levels 2-4), where there was a dense termination

214 ULNAR N.

MEDIAN N.

~':[~,..""

:...

~

.... : x~"

...?

.'~'.'-..

-

.5" :' " '

I x

I

Fig. 3. Extent of transganglionic labeling in the lower brainstem and uppermost part of the spinal cord after exposure of the median and ulnar nerves to HRP and W G A - H R P respectively. Sections 1-6 represent the same rostrocaudal levels as in Fig. 1. For details, see text.

215

Fig. 4. Photomicrographs showing the termination of the ulnar nerve in the cuneate nucleus. Dorsal upwards, medial to the left. × 62. a: this shows the patch-like terminal pattern in the middle part of the nucleus. Each patch corresponds to a cell cluster, b: this shows a small separate projection (arrow) to the lateral part of the nucleus in addition to the projection to the main terminal area in the medial part. on the cell clusters (Fig. 4a), including the most dorsal ones (cf. the superficial branch of the radial nerve). A sparse projection to the cells in the ventral, 'reticular' part of the middle region was also seen. The labeling in the rostral region had a similar distribution to that in the middle part of the nucleus (Fig. 3, level 5). However, near the upper end it suddenly spread out and covered most of the area of the cuneate nucleus (Fig. 3, level 6). In the caudal region (Fig. 3, level 1) there was sparse labeling in the medial part o f the nucleus. A few peroxidase granules were present in lamina VI. The external cuneate nucleus was labeled in its medial corner (Fig. 3, levels 4-6). The medial palmar digital nerve projected to the dorsomedial part of the area occupied by the ulnar nerve, but in a few sections it also had a small lateral terminal field. In the rostral pole of the nucleus a diffuse widespread projection was seen. There was no labeling in the ventral, 'reticular' region, nor in lamina VI. The external cuneate nucleus was also devoid of labeling. Projection from the median nerve was observed in the cuneate nucleus from about 7 m m caudal to the obex to about 2.5 m m rostral to the obex; the rostral and caudal extremes were thus unlabeled. The labeling after exposure of the median nerve in the axillary fossa to H R P is shown in Fig. 3. The most intense

peroxidase activity was seen in the middle region; here the superficial clusters in the dorsal and dorsolateral parts of the nucleus were labeled (cf. the superficial branch of the radial nerve). In addition, labeling was found ventrally and involved both cell clusters and the reticular zone, but in the latter it was weak. The terminal area in the caudal region occupied a mid-dorsal position; in addition there was sparse labeling in lamina VI. Similarly, rostral to the obex the labeling was located mid-dorsally, but in some sections a small ventromedial terminal area was also seen (Fig. 3, level 5). The external cuneate nucleus was labeled in its medial corner. The main difference observed after H R P exposure of the median nerve at the carpal ligament was a much sparser ventral projection. The labeling in the external cuneate nucleus was also much reduced. The musculocutaneous nerve (Fig. 5). Labeling was present in the cuneate nucleus from about 6 m m caudal to the obex, but was not distinct until 2 m m more rostrally. The caudal pole of the nucleus was thus unlabeled. There was, however, a projection to lamina VI; although weak, this was fully discernible in all sections, being situated dorsally in the lamina close to the border against the cuneate nucleus (Fig. 5, level 1). In the middle region, where the labeling was most

216 LATERAL A N T E B R A C H I A L CUTANEOUS N.

M U S C U L O C U T A N E O U S N.

O ~__.'~.__

_ ~_--~J~'~ O

i L

r i t

Fig. 5. Extent of transganglionic labeling in the lower brainstem and uppermost part of the spinal cord after exposure of the musculocutaneous and lateral antebrachial cutaneous nerves to HRP and W G A - H R P respectively. Sections 1-6 represent the same rostrocaudal levels as in Fig. 1. For details, see text.

217 LATERAL BRACHIAL CUTANEOUS N.

MEDIAL CUTANEOUS N.

0 z--s~--_~

~_.,.___.,-c~0

"x

",'.

~.:..V

I

Fig. 6. Extent of transganglionic labeling in the lower brainstem and uppermost part of the spinal cord after exposure of the medial cutaneous and lateral brachial cutaneous nerves to W G A - H R P . Sections 1-6 represent the same rostrocaudal levels as in Fig. 1. For details, see text.

218

a

b

,

Fig. 7. Dark-field microphotographs showing anterograde labeling after exposure of the medial cutaneous nerve to WGA-HRP. The photograph in (a) demonstrates the very restricted but dense projection to the middle region of the cuneate nucleus, whereas (b) shows the widespread labeling in the rostral pole. x 85.

prominent, the projection was split into two terminal areas, one situated laterally and the other medially (Fig. 5, levels 3-4). The lateral area, which was largest, displayed quite dense labeling, whereas in some sections the small medial area was represented only by a few peroxidase granules. Both these terminal areas appeared to be located on cell clusters; in a few sections (in case 200) they were connected by a thin band of H R P granules suggestive of fiber labeling. There was virtually no labeling in the ventral, 'reticular' zone. Rostral to the obex labeling was present in the venttal and ventrolateral parts of the cuneate nucleus (Fig. 5, levels 5-6). In the external cuneate nucleus a transverse band of peroxidase-positive granules, located lateral to the projection from the ulnar, median and radial nerves (cf. Fig. 5 with Figs. 1 and 3), was seen. The lateral antebrachial cutaneous nerve, the terminal skin branch of the musculocutaneous nerve, had a very similar projection to the latter (Fig. 5). However, no labeling was present below 3 m m caudal to the obex, and there was no projection to lamina VI or to the external cuneate nucleus. The termination of the medial cutaneous nerve was found rostrally from a site 3 m m caudal to the obex (Fig. 6). Thus similarly to the lateral antebrachial cutaneous nerve, it did not project to the caudal region of the cuneate nucleus. As shown in Fig. 6, the medial cutaneous nerve had two terminal fields; the main one was located in the ventromedial part of the nucleus, but there was also a small projection to

the ventrolateral part. H R P granules, suggestive of fiber labeling and connecting the two terminal areas, were observed in a few sections from the level of the obex. Near the rostral end of the nucleus the projection was more diffuse, and at the rostral pole the labeling was dispersed all over the nucleus. The difference in appearance between the circumscribed termination in the middle region and the widespread one in the uppermost part is depicted in Fig. 7a and b. The projection from the lateral brachial cutaneous nerve was situated slightly ventral to that from the lateral antebrachial cutaneous nerve (cf. Figs. 5 and 6). It showed a picture that was almost the reverse of and partly overlapped that of the medial cutaneous nerve (Fig. 6); a dense circumscribed termination was thus found in the ventrolateral part of the cuneate nucleus, but in addition a small patch of peroxidase activity was also present ventromedially (Fig. 6, levels 3-5). In some sections labeled fibers were seen near the dorsal surface of the cuneate nucleus, appearing to connect the two terminal areas. A diffuse widespread projection was found in the rostral-most part of the nucleus (Fig. 6, level 6). DISCUSSION On the basis of the modality of their afferent impulses, the nerves investigated in this study can be divided into 3 different groups. Four of the nerves (the superficial branch of the radial nerve, the

219 medial cutaneous and the lateral brachial and antebrachial cutaneous nerves) carry impulses exclusively from receptors in the skin, and one nerve (the deep branch of the radial nerve) transmits proprioceptive stimuli only. The ulnar, median and musculocutaneous nerves, finally, are 'mixed' nerves carrying afferent impulses from both kinds of receptors. The results of the study show that apart from some regional differences, to be discussed later, the pure cutaneous nerves project to the dorsal part of the cuneate nucleus, and the deep branch of the radial nerve to the ventral part, including the adjoining lamina VI. The mixed nerves terminate in both areas. This termination pattern was most apparent in the middle region. Its dorsal part consists of clusters of round cells with bushy dendrites densely intermingled, whereas the base of the nucleus contains smaller, multipolar neurons with long, sparsely ramified radiating dendrites, giving it a reticular appearance 19,21. The cutaneous nerves terminated exclusively on the cell clusters, while the projection from the deep branch of the radial nerve was confined to the ventral, 'reticular' region; the same appears also to be the case for all forelimb muscle projection (Nyberg and Blomqvist, study in progress), which is in accordance with the observation of a proprioceptive input to the deep part of the nucleus in physiological studies28,a2, 33. The present findings thus suggest a modality segregation in the middle part of the cuneate nucleus. Such a segregation of impulses from different structures is present in the ventroposterior nucleus of the thalamus22, 2s and in the somatosensory cortex2Z,29,37; in the first somatosensory cortical region the body surface and the deep body structures are thus represented in different cytoarchitectonic areas23,29, 37. It may seem somewhat puzzling that despite a demonstrable projection to the external cuneate nucleus, the musculocutaneous nerve appeared to lack a terminal area in the ventral, 'reticular' region. The branches to both the biceps and coracobrachial muscles were given off proximal to the site of exposure (see Materials and Methods), and it is probable that the cuneate projection from the braehial muscle (and the elbow joint nerve, if any) was thus too sparse to be discernible. The projection of the median nerve to the ventral part of the cuneate nucleus was very slight considering the extensive mus-

cle innervation by this nerve. However, the intensity of the labeling in the experiment where the median nerve was cut proximal to its muscle branches (case 182) was quite weak, and it is therefore possible that the deep projection might have been underestimated. Physiological investigations on single cells in the D C N have shown that the exteroceptive components have a somatotopic organization20, z6. The present study demonstrates that this somatotopy is based upon down by the pattern of termination of the primary afferent fibers. The results are in accordance with the finding of a re-sorting of the fibers in the dorsal columns from a mediolaterally arranged dermatomal pattern of organization in the lower segments of the cord towards a somatotopic one in its rostral part 30,38. The picture of the forearm representation in the cuneate nucleus that emerges from the present resuits is in excellent agreement with observations in previous physiological studies 28. Thus the largest palt of the nucleus is concerned with impulses from the paw, and the findings indicate that in the middle region the plantar surfaces of the digits are represented superficial to the dorsal ones (cf. the median nerve and the superficial branch of the radial nerve, Figs. 1 and 3). The projection of the arm is split so that the radial half is represented ventrolaterally and the ulnar half ventromedially. The upper arm appears to project ventral to the lower arm (cf. the lateral brachial and antebrachial cutaneous nerves, respectively). The deep region is concerned with proprioceptive input. Superimposed on this picture, however, there seems also to be a reversed projection pattern. This is clearly seen for the projections of the small cutaneous nerves of the upper and lower arm. Thus, the lateral brachial and antebrachial cutaneous nerves, in addition to their main terminal fields ventrolaterally in the cuneate nucleus, also had small ventromedial projections, which seemed to overlap with the main terminal area of the medial cutaneous nerve and vice versa (Figs. 5 and 6). As shown in Fig. 3, there was a similar relationship between the projections from the ulnar and median nerves - - the ulnar nerve had a separate lateral terminal field (see also Fig. 4b) within the main terminal area of the median nerve (found also for the small branch to the

220 fifth digit) and the median nerve had a projection to the main terminal field of the ulnar nerve. A reversed projection was also found for the superficial branch of the radial nerve (see level 4 in Fig. 1, where the superficial branch of the radial nerve has a separate medial terminal field). The significance of this projection pattern is unclear. Since the terminal areas of the reversed projections are so widely separated from the main, somatotopically located ones, not only for the nerves from the upper and lower arm but also for those innervating the distal part of the limb, it seems less probable that they correspond to the overlapping receptive fields of the peripheral nerves, especially since this overlap is quite limited on the hand and in particular on the digits (see e.g. Sunderland36; note the lateral terminal field of the small ulnar branch to the medial side of the fifth digit). Another possibility may be that the reversed projections are concerned with the mechanism of afferent inhibition, which is considered to be the basis for contrast enhancement in the dorsal column - - medial lemniscus system (see Mountcastle27). Studies on the D C N have suggested that the inhibitory action is mediated by interneurons 1,2. In accordance with the physiological data, Golgi studies on the gracile nucleus have revealed the presence of small neurons whose axons ramify after a short distance and end in fine branches within the nucleus s, and retrograde tracing experiments have shown the presence of a population of small unlabeled neurons among the diencephalic projecting neurons in the cell clusters 6,7. Since recurrent axon collaterals appear to be of no importance for the afferent inhibition in the D C N 13, the most probable mechanism is a feed-forward one, where afferent fibers terminate either on excitatory interneurons that exert presynaptic inhibition, or on inhibitory interneurons exerting postsynaptic inhibition. In the cuneate nucleus both mechanisms appear to be involved 3. A tempting hypothesis is thus that the reversed projections, via interneurons, have an afferent inhibitory function. However, also in this connection, the large distances between the main, somatotopically located terminal areas and the reversed ones offer a considerable problem; thus the peripheral inhibitory areas in almost all cases are continuous with or in the vicinity o f the excitatory receptive fields 4.

The somatotopic organization of' the primary afferents fibers was most apparent in the cluster region, but was also maintained in most of the rostral part of the nucleus. This is of interest in view of the finding that the receptive fields of the DCN neurons are much larger in the rostral than in the middle region 14. The present study indicates that such differences in the size of the receptive fields seem to be due not to an asomatotopia in the rostral region but rather to differences in the architecture of the neurons; thus, the dendrites of the cluster cells are short and bushy 21, giving the cells an idiodendritic appearance characteristic of neurons in relay nuclei concerned with a high degree of spatial resolution 3~, whereas those of the cells in the rostral region are long and sparsely branching 21, resembling the isodendritic or reticular cell type of Ram6n-Moliner and Nauta3L The organization of the primary afferent input to the rostral-most part of the cuneate nucleus differs greatly from that to the other regions. The rostral pole was thus characterized by an asomatotopia, which was particularly apparent for some of the cutaneous nerves from the arm, and by an overlap between the terminations of the cutaneous nerves and that of the deep branch of the radial nerve, suggesting a convergence between input from deep and superficial receptors. Physiological studies have shown such a convergence between hind limb group I afferents and low threshold cutaneous afferents to this part of the cuneate nucleus TM, and our results thus indicate that the same is also the case for the forelimb projection. The projection to the caudal region was in general quite sparse, but appeared to be somatotopically organized. However, two particular features should be noted. Firstly, the cutaneous nerves from the upper and lower arm seemed to have no representation in the caudal region, which thus appears to be concerned mainly with impulses from the paw. Secondly, when proceeding caudally the deep projection was located progressively more ventrally, so that at the caudal pole it was situated not in the cuneate nucleus but in lamina VI. It is of interest, therefore, that physiological studies have repo~ted that group I afferents activate cells in a narrow zone in the base of the dorsal horn of the rostral cervical cord immediately caudal to the main cuneate nucle-

221 us az,33. M a n y of the cells in the rostral part of

vely more caudally. The findings in the present study

l a m i n a VI become labeled following H R P injection into the t h a l a m u s (Blomqvist, u n p u b l i s h e d results)

a n d observations o n the projections f r o m individual

a n d they m a y thus be analogous in f u n c t i o n to the n e u r o n s in the ventral part o f the cuneate nucleus.

forelimb muscles (Nyberg a n d Blomqvist, in progress) p o i n t to a similar o r g a n i z a t i o n of the external cuneate nucleus in the cat.

The projection to the external cuneate nucleus was limited to its medial part. I n an electrophysiological study in the rat, Campbell et al. 11 d e m o n s t r a ted a s o m a t o t o p y in the external cuneate nucleus,

ACKNOWLEDGEMENTS We t h a n k Mr. A n d e r s A h l a n d e r , Mrs. Siv Beijer

pole, a r m a n d shoulder muscles more c a u d o m e -

a n d Mrs. I n g m a r i e Olsson for skilfull technical assistance. This study was supported by grants from the

dially a n d forearm a n d h a n d muscles progressi-

Swedish Medical Research Council (Project 02710).

REFERENCES

27 (1979) 728-734. 13 Gordon, G., The concept of relay nuclei. In A. Iggo (Ed.), Handbook of Sensory Physiology, VoL 11, Somatosensory System, Springer-Verlag, Berlin, 1973, pp. 137-150. 14 Gordon, G. and Jukes, M. G. M., Dual organization of the exteroceptive components of the cat's gracile nucleus, J. Physiol. (Lond.), 173 (1964) 263-290. 15 Glees, P. and Soler, J., Fibre content of the posterior column and synaptic connections of nucleus gracilis, Z. Zellforsch. 36 (1951) 381MO0. 16 Grant, G., Arvidsson, J., Robertson, B. and Ygge J., Transganglionic transport of horseradish peroxidase in primary sensory neurons, Neurosci. Lett., 12 (1979) 23-28. 17 Hand, P. J., Lumbosacral dorsal root termination in the nucleus gracilis of the cat. Some observations on terminal degeneration in other medullary sensory nuclei, J. comp. Neurol., 126 (1966) 137-156. 18 Johansson, H. and Silfvenius, H., Connexions from large, ipsilateral hind limb muscle and skin afferents to the rostral main cuneate nucleus and to the nucleus X region in the cat, J. PhysioL (Lond.), 265 (1977) 395-428. 19 Keller, J. H. and Hand, P. J., Dorsal root projections to nucleus cuneatus of the cat, Brain Research, 20 (1970) 1-17. 20 Kruger, L., Siminoff, R. and Witkovsky, P., Single neuron analysis of dorsal column nuclei and spinal nucleus of trigeminal in cat, J. Neurophysiol., 24 (1961) 333-349. 21 Kuypers, H. G. J. M. and Tuerk, J. D., The distribution of the cortical fibres within the nuclei cuneatus and gracilis in the cat, J. 4nat. (Lond.), 98 (1964) 143-162. 22 Loe, P. R., Whitsel, B. L., Dreyer, D. A. and Metz, C. B., Body representation in ventrobasal thalamus of macaque: a single-unit analysis, J. NeurophysioL, 40 (1977) 1339-1355. 23 Merzenich, M. M., Kaas, J. H., Sur, M. and Lin, C.-S., Double representation of the body surface within cytoarchitectonic areas 3b and 1 in 'SI' in the owl monkey (Actus trivirgatus), J. comp. Neurol., 181 (1978) 41-74. 24 Mesulam, M.-M., Tetramethyl benzidine for horseradish peroxidase neurohistochemistry: a non-carcinogenic blue reaction-product with. superior sensitivity for visualizing neural afferents and efferents, J. Histochem. Chytochem.,

with the neck muscles represented in its rostrolateral

1 Andersen, P., Eccles, J. C., Oshima, T. and Schmidt, R. F., Mechanisms of synaptic transmission in the cuneate nucleus, J. NeurophysioL, 27 (1964) 1096-1116. 2 Andersen, P., Eccles, J. C., Schmidt, R. F. and Yokota, T., Identification of relay ceils and interneurons in the cuneate nucleus, J. NeurophysioL, 27 (1964) 1080-1095. 3 Andersen, P., Etholm, B. and Gordon, G., Presynaptic and postsynaptic inhibition elicited in the cat's dorsal column nuclei by mechanical stimulation of skin, J. Physiol. (Lond.), 210 (1970) 433-455. 4 Aoki, M., Afferent inhibition on various types of cat's cuneate neurons induced by dynamic and steady tactile stimuli, Brain Research, 221 (1981) 257-269. 5 Arvidsson, J. and Gobel, S., An HRP study of the central projections of primary trigeminal neurons which innervate tooth pulps in the cat, Brain Research, 210 (1981) 1-16. 6 Blomqvist, A., Gracilo-diencephalic relay ceils: a quantitative study in the cat using retrograde transport of horseradish peroxidase, J. comp. Neurol., 193 (1980) 1097-1125. 7 Blomqvist, A., Morphometric synaptology of gracilodiencephalic lelay cells: an electron microscopic study in the cat using retrograde transport of horseradish peroxidase, J. NeurocytoL, 10 (1981) 709-724. 8 Blomqvist, A. and Westman, J., Interneurons and initial axon collaterals in the feline gracile nucleus demonstrated with the rapid Golgi technique, Brain Research, 111 (1976) 407-410. 9 Brown, A. G. and Gordon, G., Subcortical mechanisms concerned in somatic sensation, Brit. Med. Bull., 33 (1977) 121-128. 10 Brushart, T. and Mesulam, M.-M., Transganglionic demonstration of central sensory projections from skin and muscle with HRP-lectin conjugates, Neurosci. Lett., 17 (1980) 1-6. 11 CampbeU, S. K., Parker, T. D. and Welker, W., Somatotopic organization of the external cuneate nucleus in albino rats, Brain Research, 77 (1974) 1-23. 12 Gonatas, N. K., Harper, C., Mizutani, T. and Gonatas, J. O., Superior sensitivity of conjugates of horseradish peroxidase with wheat germ agglutinin for studies of retrograde axonal transport, J. Histochem. Cytochem.,

222 26 (1978) 106-117. 25 Mesulam, M.-M. and Brushart, T. M., Transganglionic and anterograde transport of horseradish peroxidase across dorsal root ganglia: a tetramethylbenzidine method for tracing central sensory connections of muscles and peripheral nerves, Neuroscienee, 4 (1979) 1107-1117. 26 Millar, J. and Basbaum, A. I., Topography of the projection of the body surface of the cat to cuneate and gracile nuclei, Exp. Neurol., 49 (1975) 281-290. 27 Mountcastle, V. B., Neural mechanisms in somesthesis. In V. B. Mountcastle (Ed.), Medical Physiology, Vol. 1, C. V. Mosby, St. Louis, 1980, pp. 348-390. 28 Poggio, G. F. and Mountcastle, V. B., The functional properties of ventrobasal thalamic neurons studied in unanesthetized monkeys, J. Neurophysiol., 26 (1963) 775-806. 29 Powell, T. P. S. and Mountcastle, V. B., Some aspects of the functional organization of the postcentral gyrus of the monkey: a correlation of findings obtained in a single unit analysis with cytoarchitecture, Bull Johns Hopk. Hosp., 105 (1959) 133-162. 30 Pubols, L. M. and Pubols, B. H., Modality composition and functional characteristics of dorsal column mechanoreceptive afferent fibers innervating the raccoon's forepaw, J. Neurophysiol., 36 (1973) 1023-1037. 31 Ram6n-Moliner, E.andNauta, W. J. H., The isodendritic

32

33

34

35

36 37

38

core of the brain stem, J. comp. Neurol., 126 (1966) 311-335. Ros6n, I., Afferent connexions to group I activated cells in the main cuneate nucleus of the cat, J. PhysioL (Lond.), 205 (1969) 209-236. Ros6n, I., Localization in caudal brain stem and cervical spinal cord of neurones activated from forelimb group I afferents in the cat, Brain Research, 16 (1969) 55-71. Rosene, D. L. and Mesulam, M.-M., Fixation variables in horseradish peroxidase neurohistochemistry. I. The effects of fixation time and perfusion procedures upon enzyme activity, J. Histochem. Cytochem., 26 (1978) 28-39. Rustioni, A. and Macchi, G., Distribution of dorsal root fibers in the medulla oblongata of the cat, J. comp. Neurol., 134 (1968) 113-126. Sunderland, S., Nerves and Nerve Injuries, Livingstone, Edinburgh, 1968, 1161 pp. Werner, G. and Whitsel, B. L., Functional organization of the somatosensory cortex. In A. Iggo (Ed.), Handbook of Sensory Physiology, Vol. II, Somatosensory System, Springer-Verlag, Berlin, 1973, pp. 621-700. Whitsel, B. L., PetrucellJ, L. M., Sapiro, G. and Ha, H., Fiber sorting in the fasciculus gracilis of squirrel monkeys, Exp. NeuroL, 29 (1970) 227--242.