Brain Research, 168 (1979) 493-511 © Elsevier/North-HollandBiomedicalPress
493
MIDBRAIN PROJECTIONS TO THE TRIGEMINAL, FACIAL AND HYPOGLOSSAL NUCLEI IN THE OPOSSUM. A STUDY USING AXONAL TRANSPORT TECHNIQUES
W. M. PANNETON and G. F. MARTIN Department of Anatomy, The Ohio State University, College of Medicine, Columbus, Ohio 43210 (U.S.A.)
(Accepted September 14th, 1978)
SUMMARY It has been proposed (see Berntson and Micco 6 for review) that circuits intrinsic to the midbrain play an important role in the elaboration and control of behaviors involving the motor nuclei of the trigeminal, facial and hypoglossal nerves (e.g. defense, threat, attack); but because of technical problems, it has been difficult to analyze their organization. Using the horseradish peroxidase technique we have localized those midbrain neurons which project to each of the above nuclei and by using the autoradiographic method we have plotted the intranuclear distribution of their axons. Using both techniques, we have seen that mesencephalic projections to oralfacial motor nuclei strongly favor the nucleus of the facial nerve. Cells ventral to the cerebral aqueduct, including the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch and the rostral oculomotor nucleus provide major midbrain-facial projections in the opossum. Their axons terminate densely and bilaterally within areas innervating auricular muscles and to a lesser extent, the platysma sheet tT. The projection to the caudal auricular area of the facial complex is particularly dense. Neurons within and dorsal to the red nucleus project to regions of the contralateral facial nucleus reported to supply buccolabial, zygomatic and cervical musculature 17. There is also a minor tectal projection to the facial nucleus. Direct projections to the hypoglossal nuclei also arise within the periaqueductal gray and interstitial nucleus, but if such regions influence the motor trigeminal nucleus, it is mainly by way of dendrites that extend outside the nucleus or by at least one synaptic delay. The mesencephalic nucleus of the trigeminal nerve, however, projects strongly to the motor trigeminal nucleus. These data are discussed in light of their possible functional significance.
494 INTRODUCTION The motor nuclei of the trigeminal, facial and hypoglossal nerves are focal points upon which brain stem networks converge to direct the oral-facial components of defense, threat, attack, grooming, eating, and vocalization. Although these behaviors can be produced by limbic and hypothalamic stimulation (see Mogenson and Huang 44 for data on placental animals and Roberts et a l ) 3 for data on the opossum), they may also be generated by stimulation of the midbrain, even in decerebrate animals1,4,7,9, 37. The major goals of the present study were to identify midbrain areas which project to the trigeminal, facial and hypoglossal nuclei by the horseradish peroxidase method, and to reveal the distribution of axons from each of them within the nuclei in question. Because of technical problems, it has been difficult to analyze the organization of midbrain projections to cranial nerve nuclei. Although degeneration studies have provided considerable data10,11,16, the cutting of axons originating from areas outside the lesion makes interpretation of the results difficult. For that reason the autoradiographic technique has been employed to advantage18,19, 49 and certain patterns have begun to emerge. Since the specific peripheral innervation of neurons within the trigeminal (cat3,41,57,59, rat 41, dog47), facial (catl~,~°, 52,~6, dogSO,61, rat40, ao, possum 51, and opossumlV,50), and to some extent, hypoglossal (many species 2, rat36, 4~) nuclei has been reported, tentative statements about the functional significance of such circuits can be made. Our experimental model, the North American opossum, is generally considered to be a prototypic mammal and it is our hypothesis that the brain stem connections described herein are present in most, if not all, mammals. Unless otherwise specified the terminology used is taken from the opossum atlas by OswaldoCruz and Rocha-Miranda 48. MATERIALS AND METHODS The horseradish peroxidase technique was used to tentatively identify those neurons that project to the trigeminal, facial and hypoglossal nuclei. For these studies 12 animals were anesthetized with intraperitoneal injections of sodium pentobarbital and prepared for stereotaxic surgery. Because small injections were desired, the enzyme was deposited iontophoretically24 using a Grass S-8 power source in series with a microampmeter. A platinum wire lead from the ampmeter was inserted into a micropipette containing a 6 ~o HRP solution. The micropipette was made by placing a few fiberglass strands into a capillary tube and then pulling in to a tip diameter of 20 #m. The necessary stereotaxic coordinates were extrapolated from the atlas of Oswaldo-Cruz and Rocha-Miranda as, but prior to each injection the area at the tip of the micropipette was stimulated to check its location. This was particularly important for the injections of the facial nucleus, where attempts were made to center the pipette within specific divisions of the nucleus. The injections were made by using a direct current of 6 #A for 15-20 min. After survival times of 24-27 h the animals were reanesthetized and thoracotomized for intracardiac perfusion. Each animal was perfused with a solution of 0.5 ~ procaine hydrochloride in physiologic saline, followed immediately by the fixative solution (1 ~o
495 paraformaldehyde, 1.25 % glutaraldehyde and 1% sucrose in 0.1 M phosphate buffer, pH 7.4). After removal, the brains were post-fixed for 6-10 h, then transferred to a 30 % sucrose-phosphate buffer solution (4 °C) for 24-48 h. Frozen sections were cut at 40 #m and incubated in a cold sodium sulfate solution prior to the histochemical reaction. Although diaminobenzidine and hydrogen peroxide were used to visualize the HRP in several cases, a modification of the de Olmos' O-dianisidine technique 15was utilized for most of them. The 'green reaction' using O-dianisidine requires a reaction mixture of 0.2% gelatin, 0.05 M acetate buffer, 5/zM magnesium sulfate, sodium nitroprusside and O-dianisidine. Hydrogen peroxide is added immediately before the tissues are incubated. After a 20-25 min reaction time, the tissues are washed twice (5 rain/wash) in a 4.5% sodium nitroprusside solution. Low temperatures are maintained during the reaction and washings by performing the procedures over an ice bath. The reacted sections were mounted on chrom-alum coated slides and stained with neutral red. The presence of HRP-positive perikarya was noted after systematic scanning of all sections under both bright- and dark-field illumination. The locations and numbers of the labeled neurons, as well as the extent of the injection sites were recorded. Brains with no injections as well as brains with injections which missed the nuclei being studied were also reacted by the O-dianisidine method. The results obtained from such cases indicate that the HRP labeling observed after injections of desired areas was not artifactual. Twenty-five opossums were used for autoradiographic analysis. Although some of the cases had been prepared for other studies, for many of them the injection was targeted so as to center over neurons comparable to those labeled in the HRP cases. Each animal was injected with 0.1--0.3 /~1 (100--400/zCi//zl) of [ZH]leucine using a Hamilton syringe (31-gauge needle) mounted to a stereotaxically guided microdrive. Following survival times of 8-10 days, the animals were anesthetized and perfused with a 10~o formalin solution. Frozen sections were mounted, coated with diluted Kodak NTB-2 liquid emulsion and refrigerated for 4-8 weeks. The slides were subsequently developed with D-19 high contrast developer, stained through the emulsion with cresyl violet, and coverslipped. Sections were examined by dark-field optics and the results plotted. All of the injection sites were drawn as they appear through dark-field optics, although the effective part of each injection may be less extensive. For the following account clusters of silver grains over axonal bundles are interpreted as indicating labeled axons which course to distal sites, whereas randomly scattered grains over a particular nucleus are considered to reveal terminal labeling. RESULTS
Results of HRP injections into the trigeminal motor nucleus Motor trigeminal injections of HRP were attempted in 3 animals, but only the most successful case will be described. In that animal, the tip of the micropipette was located in the ventral, caudal one-third of the nucleus, where elevation of the mandible was produced by electrical stimulation prior to injection. Although the dense core of
496
497 reaction product was limited mainly to the motor trigeminal nucleus, sections reacted by a modified de Olmos' method revealed some spread of the enzyme into adjacent sensory trigeminal and reticular nuclei (Fig. 1A). As expected, heavily labeled neurons were present in the mesencephalic nucleus of the trigeminal nerve (Fig. 2A, C, P-584), particularly within more rostral and dorsal parts of the nucleus. A few neurons were also reactive within the superior colliculus, the interstitial tegmentum, the pretectal complex, the rostral oculomotor nucleus, the periaqueductal gray, the interstitial nucleus of Cajal and the substantia nigra. Although the autoradiographic material shows that some of the above areas project in close proximity to the motor trigeminal nuclei (see autoradiographic results), it does not provide evidence that they contribute to terminals within them.
Results of HRP injections into the facial nucleus Iontophoretic injections of H R P were centered within either lateral or medial subdivisions of the facial nucleus. Although the greatest concentration of H R P was limited to the facial neuropil, sections reacted with O-dianisidine revealed evidence for spread to closely adjacent areas (Fig. IB, C). In alternate sections reacted with diaminobenzidine, the injection sites were smaller and limited to the facial nucleus. After all facial injections, reactive neurons were abundant in areas just ventral to the cerebral aqueduct (Fig. 1E, F, Fig. 3), i.e. the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch, and the rostral oculomotor complex. The periaqueductal gray cells were labeled on both sides and were located just dorsal and rostral to the oculomotor complex. Those contralateral to the injection were slightly more dorsal than their ipsilateral counterparts. Cells of the interstitial nucleus of Cajal were labeled bilaterally after most injections. The majority of reactive neurons within the rostral part of the oculomotor nucleus were located on the side of the injection, whereas most of those within the nucleus of Darkschewitsch were present on the contralateral side, especially after lateral injections. Other regions of the mesencephalon also contained a few HRP-positive neurons after facial injections. One such area was the superior colliculus where reactive neurons were found within the strata griseum intermedius and profundum, particularly ipsilateral to medial injections. Neurons situated along the dorsal border of the
Fig. 1. A-D: light-field photomicrographs of iontophoretic HRP injections into the trigeminal motor nucleus, P-584 (A) ; medial areas of the facial nucleus, P-567 (B) ; lateral areas of the facial nucleus, P-566 (C); and the hypoglossal nucleus, P-591 (D). All of the HRP cases illustrated were reacted according to a modification of the de Olmos1~method using O-dianisidine. E and F: light-field photomicrographs showing labeled neurons ventral to the cerebral aqueduct after an injection into the medial facial nucleus (P-567). The closed arrow in E points to a reactive neuron in the ventral lceriaqueductal gray and F shows the same cell at higher magnification (closed arrow). The olzen arrow in E points to a labeled cell of the rostral oculomotor nucleus while F shows the same cell at higher magnification (open arrow). The asterisk in F is centered among labeled neurons of the ventral gray shown at higher power in the insert. G : dark-field photomicrograph of silver grains in the emulsion over medial subdivisions of the ipsilateral facial nucleus after the injection of [3H]leucineshown in the insert (P-417). All autoradiographic cases were exposed for 4 weeks. Abbreviations for this and subsequent figures are listed at the end of the discussion.
498
m
Y/--'-x
/
Fig. 2. Drawings of sections through the midbrain from rostral (A) to caudal (D) to show the location of labeled neurons following HRP injections into the trigeminal motor nucleus (P-584, left side) and the hypoglossal nucleus (P-591, right side). Note the retrograde fill of neurons within both the mesencephalic nucleus of the trigeminal nerve and the lateral periaqueductal gray in P-584, and within the interstitial nucleus of Cajal and ventral gray in P-591. The inserts are drawings through the injections. In each the central core is black while the stippled area depicts light spread of the enzyme (compare with Fig. 1).
499
\\,
V,,,,/
I,
"
,
f Y0
~. ~" c
~"~ )
," )
~.
lib
Fig. 3. Drawings ofsections through the midbrain from rostral (A) to caudal (D) to show location of labeled neurons following HRP injections into the medial facial nucleus (P-567, left side) and lateral facial nucleus (P-566, right side). Reactive neurons are especially numerous in the area ventral to the cerebral aqueduct, including the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch, and the rostral oculomotor nucleus.
500 contralateral red nucleus were reactive after both medial and lateral injections. A number of labeled neurons were present in lateral parts of the ipsilateral substantia nigra after medial injections and an occasional neuron was labeled within the nucleus of the posterior commissure, the pretectal complex, the lateral and dorsal periaqueductal gray, as well as within the interstitial tegmentum.
Results of HRP injections into the hypoglossal nucleus HRP injections into the hypoglossal nucleus were attempted in 3 animals. Due to the extensive rostral-caudal length of the nucleus, the best results were achieved in case P-591 (Fig. 1D, Fig. 2) where 3 separate penetrations of the micropipette were made over a distance of 3 ram. Some spread of the enzyme was noted within the nucleus reticularis ventralis, the nucleus reticularis dorsalis, the nucleus intercalatus of Staderini, as well as within medial parts of the opposite hypoglossal nucleus following processing by a modified de Olmos' technique aS. The injections described above resulted in light labeling of small, stellate neurons within the lateral periaqueductal gray of the mi.dbrain. Such neurons were present bilaterally and were most numerous in rostral sections. The interstitial nucleus of Cajal also contained labeled neurons, bilaterally, although most of them were located on the side of the injection. A single cell of the mesencephalic nucleus was labeled as were a few within the ventral periaqueductal gray just dorsal and rostral to the oculomotor complex.
Midbrain projections to oral-facial nuclei, autoradiographic results Tritiated leucine was deposited within midbrain regions comparable to those containing reactive neurons in the HRP experiments, as well as within areas devoid of such neurons. Since the area ventral to the cerebral aqueduct was labeled to some degree after HRP placements within all of the nuclei in question, several cases were prepared with injections in that area. In case P-417 the injection included the ventral periaqueductal gray, the interstitial nucleus of Cajal, parts of the nucleus of Darkschewitsch, the rostral oculomotor nucleus and the dorsomedial tip of the red nucleus. Axonal labeling extended caudally and surrounded the trigeminal motor nuclei (Fig. 4A, P-417), bilaterally, and a few labeled fibers coursed through the nucleus on the contralateral side. Clumps of grains were found also over the medial longitudinal fasciculus, especially on the ipsilateral side and over the contralateral rubrobulbarrubrospinal tracts (black arrow, Fig. 4A, P-417). At the level of the facial nucleus (Fig. 4B, P-417), labeled axons were present within the same tracts as well as within the gigantocellular reticular formation, especially on the side of the placement. In addition, however, terminal label (Fig. 1G) was present over medial facial areas, bilaterally, particularly over the dorsomedial regions which have been reported to supply caudal and rostral auricular muscles 17. Some label was also present over the ventromedial area thought to innervate the cervical platysma muscle xT. Although there was a sharp decrease in grains over intermediate portions of the nucleus, counts above background were evident over the ventromedial part of the lateral division, ipsilaterally (the area which supplies zygomatic muscles17), as well as over the dorsal aspect of
501
\ ' \ . ~ oee •
Ira
,,
•
c//
.%-0-
P-418
,dl ql alp'
Fig. 4. Drawings of sections through the brain stem at levels of the trigeminal (A), facial (B), and hypoglossal (C) motor nuclei that show the distribution of label resulting from injections of laH]leucine. P-417, left side, shows the results produced by an injection ventral to the cerebral aqueduct that included the area labeled after HRP injections into oral-facial motor nuclei. A composite of the HRP results is shown in the insert at the left (open circles, P-566; closed circles, P-567; open squares, P-594; closed squares, P-591). P-418, right side, illustrates the results produced by an injections that included the red nucleus and the interstitial area just dorsal to it. Arrows point to labeled axons in the crossed rubrobulbar-rubrospinal tracts in both P-417 and P-418.
the lateral division on the contralateral side. The latter area contributes to buccolabial rami in the opossum 17 and has been described as specifically innervating the orbicularis oculi muscle in the cat 52. The labeled axons which entered the lateral division contralaterally appeared to course within the rubrobulbar-rubrospinal tracts. At medullary levels labeled axons were still present within the medial longitudinal fasciculus, ipsilaterally, and the rubral tracts contralateraUy. Light terminal label was seen over the hypoglossal nuclei of both sides and also over the reticular neuropil just ventral to them. The labeling over these areas was most abundant ipsilaterally.
502 The injection in P-418 (Fig. 4) covered the interstitial nucleus of Cajal and the red nucleus but missed the periaqueductal gray. As might be expected, the medial longitudinal fasciculus was labeled ipsilaterally, and the crossed rubral tracts were extensively labeled (arrows, Fig. 4, P-418). In support of the H R P results, few labeled axons arborized within the motor nuclei of the trigeminal nerves, although some coursed around and through them, particularly on the side opposite the placement (Fig. 4A, P-418). At more caudal levels (Fig. 4B, P-418) the grain pattern over the ipsilateral facial nucleus was similar to that observed in P-417, although it was less dense (same survival and exposure times). On the contralateral side, however, there was an even distribution of grains over the lateral division, as well as sparse label in the medial part of the nucleus. The differences between the facial labeling in the two animals can be appreciated best by comparing section B in each case (Fig. 4). More caudally (Fig. 4C, P-418), silver grains were abundant over both the ipsilateral medial longitudinal fasciculus and the contralateral rubral tracts. Labeled axons could be followed from the latter bundle to a region just lateral to the tractus solitarius and deep to the nucleus cuneatus, Along this route, a dispersion of grains was present just lateral to the nucleus ambiguus and within the nucleus reticularis dorsalis, as well as over fibers which turned medially into the reticular formation (Fig. 4C, P-418). The latter axons appeared to end ventral to the nucleus alaris and ventrolateral to the nucleus hypoglossi. On the side of the placement labeled axons left the medial longitudinal fasciculus to terminate ventral to the hypoglossal nucleus. In another case, P-469 (Fig. 5, left side), the injection was slightly rostral and lateral to that in P-418 and included both the red nucleus and the tegmentum dorsal and lateral to it. As in P-418, both the ipsilateral medial longitudinal fasciculus and the crossed rubral tracts were labeled throughout the brain stem. At the level of the trigeminal motor nuclei (Fig. 5A, P-469), the pattern of label was generally similar to that in the other cases. However, the label over caudal parabrachial areas and the ventral trigeminal sensory nucleus was heavier and labeled axons were seen in ventromedial areas of the ipsilateral reticular formation. More caudally (Fig. 5B, P469), there was little label over the ipsilateral facial nucleus, but grain counts were above background over several areas of the same nucleus on the opposite side. Such areas included the ventromedial region which supplies the platysma muscle and that part of the lateral facial nucleus which innervates buccolabial muscles 17. At medullary levels (Fig. 5C, P-469), sparse label was seen over both the hypoglossal nuclei and the nuclei reticularis ventralis, mostly ipsilaterally, and rather heavy label was located just lateral to the nucleus ambiguus, contralateraily. The latter axons issued from the crossed rubral tracts. Labeled axons also distributed to lateral parts of nucleus tractus solitarius and to the nucleus reticularis dorsalis. Since the HRP results gave only scant indication of projections from either the interstitial tegmentum or cuneiform area to bulbar motor nuclei, injections of [aH]leucine were made into those regions in several cases to test the negative results. Case P-505 serves as an example of several and is plotted in Fig. 5. Although labeled axons were present within many brain stem areas there was little evidence of transport to the facial nuclei (Fig. 5B, P-505). Some labeling was present over the ipsilateral
503
P-469
tl
-
I
~•0,;. ~i:)•~;!~~
TrMo /
/
T~Y
/ ~
'--p~
.
.
.
-.
.
" ....~((~ ~" ~ ~ o !.".... ~'~'~'i~" . .
. m,
/
•
....
:
•.•
.
•
:~s) %, /:.
" n .:. .........
,.
!
.
,.." •
J •
C
Am
Fig. 5. Illustrations of sections comparable to those in Fig. 4 showing the transport of labeled protein after selected midbrain injections. P-469, left side, shows the results of a placement in the red nucleus and the tegmentum dorsal and lateral to it. P-505, right side, illustrates the results obtained after an injection of [3H]leucine into interstitial and cuneiform areas of the opossum midbrain. Contrast sections B and C in Fig. 4 and 5 and note the shifting pattern of label, especially in the facial nucleus, as the injection sites are displaced more laterally.
trigeminal and hypoglossal nuclei (Fig. 5A, C, P-505), but comparison of different cases suggested that it was a result of marker incorporation by neurons within the mesencephalic nucleus of the trigeminal nerve and that part of the lateral periaqueductal gray labeled after hypoglossal injections of HRP (see P-591, Fig. 2). Several cases are available with [aH]leucine deposits within the superior colliculus and adjacent regions. One such case, P-488 (Fig. 6) was subjected to a double injection in order to label large areas of the superior colliculus and part of the caudal pretectal complex. Terminal label was present within the ipsilateral trigeminal nucleus (Fig. 6A), as might be expected from inclusion of the mesencephalic nucleus at the injection site, and there was evidence for a sparse projection to the ventromedial facial nucleus on the ipsilateral side (Fig. 6B).
('
,
493
MIDBRAIN PROJECTIONS TO THE TRIGEMINAL, FACIAL AND HYPOGLOSSAL NUCLEI IN THE OPOSSUM. A STUDY USING AXONAL TRANSPORT TECHNIQUES
W. M. PANNETON and G. F. MARTIN Department of Anatomy, The Ohio State University, College of Medicine, Columbus, Ohio 43210 (U.S.A.)
(Accepted September 14th, 1978)
SUMMARY It has been proposed (see Berntson and Micco 6 for review) that circuits intrinsic to the midbrain play an important role in the elaboration and control of behaviors involving the motor nuclei of the trigeminal, facial and hypoglossal nerves (e.g. defense, threat, attack); but because of technical problems, it has been difficult to analyze their organization. Using the horseradish peroxidase technique we have localized those midbrain neurons which project to each of the above nuclei and by using the autoradiographic method we have plotted the intranuclear distribution of their axons. Using both techniques, we have seen that mesencephalic projections to oralfacial motor nuclei strongly favor the nucleus of the facial nerve. Cells ventral to the cerebral aqueduct, including the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch and the rostral oculomotor nucleus provide major midbrain-facial projections in the opossum. Their axons terminate densely and bilaterally within areas innervating auricular muscles and to a lesser extent, the platysma sheet tT. The projection to the caudal auricular area of the facial complex is particularly dense. Neurons within and dorsal to the red nucleus project to regions of the contralateral facial nucleus reported to supply buccolabial, zygomatic and cervical musculature 17. There is also a minor tectal projection to the facial nucleus. Direct projections to the hypoglossal nuclei also arise within the periaqueductal gray and interstitial nucleus, but if such regions influence the motor trigeminal nucleus, it is mainly by way of dendrites that extend outside the nucleus or by at least one synaptic delay. The mesencephalic nucleus of the trigeminal nerve, however, projects strongly to the motor trigeminal nucleus. These data are discussed in light of their possible functional significance.
494 INTRODUCTION The motor nuclei of the trigeminal, facial and hypoglossal nerves are focal points upon which brain stem networks converge to direct the oral-facial components of defense, threat, attack, grooming, eating, and vocalization. Although these behaviors can be produced by limbic and hypothalamic stimulation (see Mogenson and Huang 44 for data on placental animals and Roberts et a l ) 3 for data on the opossum), they may also be generated by stimulation of the midbrain, even in decerebrate animals1,4,7,9, 37. The major goals of the present study were to identify midbrain areas which project to the trigeminal, facial and hypoglossal nuclei by the horseradish peroxidase method, and to reveal the distribution of axons from each of them within the nuclei in question. Because of technical problems, it has been difficult to analyze the organization of midbrain projections to cranial nerve nuclei. Although degeneration studies have provided considerable data10,11,16, the cutting of axons originating from areas outside the lesion makes interpretation of the results difficult. For that reason the autoradiographic technique has been employed to advantage18,19, 49 and certain patterns have begun to emerge. Since the specific peripheral innervation of neurons within the trigeminal (cat3,41,57,59, rat 41, dog47), facial (catl~,~°, 52,~6, dogSO,61, rat40, ao, possum 51, and opossumlV,50), and to some extent, hypoglossal (many species 2, rat36, 4~) nuclei has been reported, tentative statements about the functional significance of such circuits can be made. Our experimental model, the North American opossum, is generally considered to be a prototypic mammal and it is our hypothesis that the brain stem connections described herein are present in most, if not all, mammals. Unless otherwise specified the terminology used is taken from the opossum atlas by OswaldoCruz and Rocha-Miranda 48. MATERIALS AND METHODS The horseradish peroxidase technique was used to tentatively identify those neurons that project to the trigeminal, facial and hypoglossal nuclei. For these studies 12 animals were anesthetized with intraperitoneal injections of sodium pentobarbital and prepared for stereotaxic surgery. Because small injections were desired, the enzyme was deposited iontophoretically24 using a Grass S-8 power source in series with a microampmeter. A platinum wire lead from the ampmeter was inserted into a micropipette containing a 6 ~o HRP solution. The micropipette was made by placing a few fiberglass strands into a capillary tube and then pulling in to a tip diameter of 20 #m. The necessary stereotaxic coordinates were extrapolated from the atlas of Oswaldo-Cruz and Rocha-Miranda as, but prior to each injection the area at the tip of the micropipette was stimulated to check its location. This was particularly important for the injections of the facial nucleus, where attempts were made to center the pipette within specific divisions of the nucleus. The injections were made by using a direct current of 6 #A for 15-20 min. After survival times of 24-27 h the animals were reanesthetized and thoracotomized for intracardiac perfusion. Each animal was perfused with a solution of 0.5 ~ procaine hydrochloride in physiologic saline, followed immediately by the fixative solution (1 ~o
495 paraformaldehyde, 1.25 % glutaraldehyde and 1% sucrose in 0.1 M phosphate buffer, pH 7.4). After removal, the brains were post-fixed for 6-10 h, then transferred to a 30 % sucrose-phosphate buffer solution (4 °C) for 24-48 h. Frozen sections were cut at 40 #m and incubated in a cold sodium sulfate solution prior to the histochemical reaction. Although diaminobenzidine and hydrogen peroxide were used to visualize the HRP in several cases, a modification of the de Olmos' O-dianisidine technique 15was utilized for most of them. The 'green reaction' using O-dianisidine requires a reaction mixture of 0.2% gelatin, 0.05 M acetate buffer, 5/zM magnesium sulfate, sodium nitroprusside and O-dianisidine. Hydrogen peroxide is added immediately before the tissues are incubated. After a 20-25 min reaction time, the tissues are washed twice (5 rain/wash) in a 4.5% sodium nitroprusside solution. Low temperatures are maintained during the reaction and washings by performing the procedures over an ice bath. The reacted sections were mounted on chrom-alum coated slides and stained with neutral red. The presence of HRP-positive perikarya was noted after systematic scanning of all sections under both bright- and dark-field illumination. The locations and numbers of the labeled neurons, as well as the extent of the injection sites were recorded. Brains with no injections as well as brains with injections which missed the nuclei being studied were also reacted by the O-dianisidine method. The results obtained from such cases indicate that the HRP labeling observed after injections of desired areas was not artifactual. Twenty-five opossums were used for autoradiographic analysis. Although some of the cases had been prepared for other studies, for many of them the injection was targeted so as to center over neurons comparable to those labeled in the HRP cases. Each animal was injected with 0.1--0.3 /~1 (100--400/zCi//zl) of [ZH]leucine using a Hamilton syringe (31-gauge needle) mounted to a stereotaxically guided microdrive. Following survival times of 8-10 days, the animals were anesthetized and perfused with a 10~o formalin solution. Frozen sections were mounted, coated with diluted Kodak NTB-2 liquid emulsion and refrigerated for 4-8 weeks. The slides were subsequently developed with D-19 high contrast developer, stained through the emulsion with cresyl violet, and coverslipped. Sections were examined by dark-field optics and the results plotted. All of the injection sites were drawn as they appear through dark-field optics, although the effective part of each injection may be less extensive. For the following account clusters of silver grains over axonal bundles are interpreted as indicating labeled axons which course to distal sites, whereas randomly scattered grains over a particular nucleus are considered to reveal terminal labeling. RESULTS
Results of HRP injections into the trigeminal motor nucleus Motor trigeminal injections of HRP were attempted in 3 animals, but only the most successful case will be described. In that animal, the tip of the micropipette was located in the ventral, caudal one-third of the nucleus, where elevation of the mandible was produced by electrical stimulation prior to injection. Although the dense core of
496
497 reaction product was limited mainly to the motor trigeminal nucleus, sections reacted by a modified de Olmos' method revealed some spread of the enzyme into adjacent sensory trigeminal and reticular nuclei (Fig. 1A). As expected, heavily labeled neurons were present in the mesencephalic nucleus of the trigeminal nerve (Fig. 2A, C, P-584), particularly within more rostral and dorsal parts of the nucleus. A few neurons were also reactive within the superior colliculus, the interstitial tegmentum, the pretectal complex, the rostral oculomotor nucleus, the periaqueductal gray, the interstitial nucleus of Cajal and the substantia nigra. Although the autoradiographic material shows that some of the above areas project in close proximity to the motor trigeminal nuclei (see autoradiographic results), it does not provide evidence that they contribute to terminals within them.
Results of HRP injections into the facial nucleus Iontophoretic injections of H R P were centered within either lateral or medial subdivisions of the facial nucleus. Although the greatest concentration of H R P was limited to the facial neuropil, sections reacted with O-dianisidine revealed evidence for spread to closely adjacent areas (Fig. IB, C). In alternate sections reacted with diaminobenzidine, the injection sites were smaller and limited to the facial nucleus. After all facial injections, reactive neurons were abundant in areas just ventral to the cerebral aqueduct (Fig. 1E, F, Fig. 3), i.e. the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch, and the rostral oculomotor complex. The periaqueductal gray cells were labeled on both sides and were located just dorsal and rostral to the oculomotor complex. Those contralateral to the injection were slightly more dorsal than their ipsilateral counterparts. Cells of the interstitial nucleus of Cajal were labeled bilaterally after most injections. The majority of reactive neurons within the rostral part of the oculomotor nucleus were located on the side of the injection, whereas most of those within the nucleus of Darkschewitsch were present on the contralateral side, especially after lateral injections. Other regions of the mesencephalon also contained a few HRP-positive neurons after facial injections. One such area was the superior colliculus where reactive neurons were found within the strata griseum intermedius and profundum, particularly ipsilateral to medial injections. Neurons situated along the dorsal border of the
Fig. 1. A-D: light-field photomicrographs of iontophoretic HRP injections into the trigeminal motor nucleus, P-584 (A) ; medial areas of the facial nucleus, P-567 (B) ; lateral areas of the facial nucleus, P-566 (C); and the hypoglossal nucleus, P-591 (D). All of the HRP cases illustrated were reacted according to a modification of the de Olmos1~method using O-dianisidine. E and F: light-field photomicrographs showing labeled neurons ventral to the cerebral aqueduct after an injection into the medial facial nucleus (P-567). The closed arrow in E points to a reactive neuron in the ventral lceriaqueductal gray and F shows the same cell at higher magnification (closed arrow). The olzen arrow in E points to a labeled cell of the rostral oculomotor nucleus while F shows the same cell at higher magnification (open arrow). The asterisk in F is centered among labeled neurons of the ventral gray shown at higher power in the insert. G : dark-field photomicrograph of silver grains in the emulsion over medial subdivisions of the ipsilateral facial nucleus after the injection of [3H]leucineshown in the insert (P-417). All autoradiographic cases were exposed for 4 weeks. Abbreviations for this and subsequent figures are listed at the end of the discussion.
498
m
Y/--'-x
/
Fig. 2. Drawings of sections through the midbrain from rostral (A) to caudal (D) to show the location of labeled neurons following HRP injections into the trigeminal motor nucleus (P-584, left side) and the hypoglossal nucleus (P-591, right side). Note the retrograde fill of neurons within both the mesencephalic nucleus of the trigeminal nerve and the lateral periaqueductal gray in P-584, and within the interstitial nucleus of Cajal and ventral gray in P-591. The inserts are drawings through the injections. In each the central core is black while the stippled area depicts light spread of the enzyme (compare with Fig. 1).
499
\\,
V,,,,/
I,
"
,
f Y0
~. ~" c
~"~ )
," )
~.
lib
Fig. 3. Drawings ofsections through the midbrain from rostral (A) to caudal (D) to show location of labeled neurons following HRP injections into the medial facial nucleus (P-567, left side) and lateral facial nucleus (P-566, right side). Reactive neurons are especially numerous in the area ventral to the cerebral aqueduct, including the ventral periaqueductal gray, the interstitial nucleus of Cajal, the nucleus of Darkschewitsch, and the rostral oculomotor nucleus.
500 contralateral red nucleus were reactive after both medial and lateral injections. A number of labeled neurons were present in lateral parts of the ipsilateral substantia nigra after medial injections and an occasional neuron was labeled within the nucleus of the posterior commissure, the pretectal complex, the lateral and dorsal periaqueductal gray, as well as within the interstitial tegmentum.
Results of HRP injections into the hypoglossal nucleus HRP injections into the hypoglossal nucleus were attempted in 3 animals. Due to the extensive rostral-caudal length of the nucleus, the best results were achieved in case P-591 (Fig. 1D, Fig. 2) where 3 separate penetrations of the micropipette were made over a distance of 3 ram. Some spread of the enzyme was noted within the nucleus reticularis ventralis, the nucleus reticularis dorsalis, the nucleus intercalatus of Staderini, as well as within medial parts of the opposite hypoglossal nucleus following processing by a modified de Olmos' technique aS. The injections described above resulted in light labeling of small, stellate neurons within the lateral periaqueductal gray of the mi.dbrain. Such neurons were present bilaterally and were most numerous in rostral sections. The interstitial nucleus of Cajal also contained labeled neurons, bilaterally, although most of them were located on the side of the injection. A single cell of the mesencephalic nucleus was labeled as were a few within the ventral periaqueductal gray just dorsal and rostral to the oculomotor complex.
Midbrain projections to oral-facial nuclei, autoradiographic results Tritiated leucine was deposited within midbrain regions comparable to those containing reactive neurons in the HRP experiments, as well as within areas devoid of such neurons. Since the area ventral to the cerebral aqueduct was labeled to some degree after HRP placements within all of the nuclei in question, several cases were prepared with injections in that area. In case P-417 the injection included the ventral periaqueductal gray, the interstitial nucleus of Cajal, parts of the nucleus of Darkschewitsch, the rostral oculomotor nucleus and the dorsomedial tip of the red nucleus. Axonal labeling extended caudally and surrounded the trigeminal motor nuclei (Fig. 4A, P-417), bilaterally, and a few labeled fibers coursed through the nucleus on the contralateral side. Clumps of grains were found also over the medial longitudinal fasciculus, especially on the ipsilateral side and over the contralateral rubrobulbarrubrospinal tracts (black arrow, Fig. 4A, P-417). At the level of the facial nucleus (Fig. 4B, P-417), labeled axons were present within the same tracts as well as within the gigantocellular reticular formation, especially on the side of the placement. In addition, however, terminal label (Fig. 1G) was present over medial facial areas, bilaterally, particularly over the dorsomedial regions which have been reported to supply caudal and rostral auricular muscles 17. Some label was also present over the ventromedial area thought to innervate the cervical platysma muscle xT. Although there was a sharp decrease in grains over intermediate portions of the nucleus, counts above background were evident over the ventromedial part of the lateral division, ipsilaterally (the area which supplies zygomatic muscles17), as well as over the dorsal aspect of
501
\ ' \ . ~ oee •
Ira
,,
•
c//
.%-0-
P-418
,dl ql alp'
Fig. 4. Drawings of sections through the brain stem at levels of the trigeminal (A), facial (B), and hypoglossal (C) motor nuclei that show the distribution of label resulting from injections of laH]leucine. P-417, left side, shows the results produced by an injection ventral to the cerebral aqueduct that included the area labeled after HRP injections into oral-facial motor nuclei. A composite of the HRP results is shown in the insert at the left (open circles, P-566; closed circles, P-567; open squares, P-594; closed squares, P-591). P-418, right side, illustrates the results produced by an injections that included the red nucleus and the interstitial area just dorsal to it. Arrows point to labeled axons in the crossed rubrobulbar-rubrospinal tracts in both P-417 and P-418.
the lateral division on the contralateral side. The latter area contributes to buccolabial rami in the opossum 17 and has been described as specifically innervating the orbicularis oculi muscle in the cat 52. The labeled axons which entered the lateral division contralaterally appeared to course within the rubrobulbar-rubrospinal tracts. At medullary levels labeled axons were still present within the medial longitudinal fasciculus, ipsilaterally, and the rubral tracts contralateraUy. Light terminal label was seen over the hypoglossal nuclei of both sides and also over the reticular neuropil just ventral to them. The labeling over these areas was most abundant ipsilaterally.
502 The injection in P-418 (Fig. 4) covered the interstitial nucleus of Cajal and the red nucleus but missed the periaqueductal gray. As might be expected, the medial longitudinal fasciculus was labeled ipsilaterally, and the crossed rubral tracts were extensively labeled (arrows, Fig. 4, P-418). In support of the H R P results, few labeled axons arborized within the motor nuclei of the trigeminal nerves, although some coursed around and through them, particularly on the side opposite the placement (Fig. 4A, P-418). At more caudal levels (Fig. 4B, P-418) the grain pattern over the ipsilateral facial nucleus was similar to that observed in P-417, although it was less dense (same survival and exposure times). On the contralateral side, however, there was an even distribution of grains over the lateral division, as well as sparse label in the medial part of the nucleus. The differences between the facial labeling in the two animals can be appreciated best by comparing section B in each case (Fig. 4). More caudally (Fig. 4C, P-418), silver grains were abundant over both the ipsilateral medial longitudinal fasciculus and the contralateral rubral tracts. Labeled axons could be followed from the latter bundle to a region just lateral to the tractus solitarius and deep to the nucleus cuneatus, Along this route, a dispersion of grains was present just lateral to the nucleus ambiguus and within the nucleus reticularis dorsalis, as well as over fibers which turned medially into the reticular formation (Fig. 4C, P-418). The latter axons appeared to end ventral to the nucleus alaris and ventrolateral to the nucleus hypoglossi. On the side of the placement labeled axons left the medial longitudinal fasciculus to terminate ventral to the hypoglossal nucleus. In another case, P-469 (Fig. 5, left side), the injection was slightly rostral and lateral to that in P-418 and included both the red nucleus and the tegmentum dorsal and lateral to it. As in P-418, both the ipsilateral medial longitudinal fasciculus and the crossed rubral tracts were labeled throughout the brain stem. At the level of the trigeminal motor nuclei (Fig. 5A, P-469), the pattern of label was generally similar to that in the other cases. However, the label over caudal parabrachial areas and the ventral trigeminal sensory nucleus was heavier and labeled axons were seen in ventromedial areas of the ipsilateral reticular formation. More caudally (Fig. 5B, P469), there was little label over the ipsilateral facial nucleus, but grain counts were above background over several areas of the same nucleus on the opposite side. Such areas included the ventromedial region which supplies the platysma muscle and that part of the lateral facial nucleus which innervates buccolabial muscles 17. At medullary levels (Fig. 5C, P-469), sparse label was seen over both the hypoglossal nuclei and the nuclei reticularis ventralis, mostly ipsilaterally, and rather heavy label was located just lateral to the nucleus ambiguus, contralateraily. The latter axons issued from the crossed rubral tracts. Labeled axons also distributed to lateral parts of nucleus tractus solitarius and to the nucleus reticularis dorsalis. Since the HRP results gave only scant indication of projections from either the interstitial tegmentum or cuneiform area to bulbar motor nuclei, injections of [aH]leucine were made into those regions in several cases to test the negative results. Case P-505 serves as an example of several and is plotted in Fig. 5. Although labeled axons were present within many brain stem areas there was little evidence of transport to the facial nuclei (Fig. 5B, P-505). Some labeling was present over the ipsilateral
503
P-469
tl
-
I
~•0,;. ~i:)•~;!~~
TrMo /
/
T~Y
/ ~
'--p~
.
.
.
-.
.
" ....~((~ ~" ~ ~ o !.".... ~'~'~'i~" . .
. m,
/
•
....
:
•.•
.
•
:~s) %, /:.
" n .:. .........
,.
!
.
,.." •
J •
C
Am
Fig. 5. Illustrations of sections comparable to those in Fig. 4 showing the transport of labeled protein after selected midbrain injections. P-469, left side, shows the results of a placement in the red nucleus and the tegmentum dorsal and lateral to it. P-505, right side, illustrates the results obtained after an injection of [3H]leucine into interstitial and cuneiform areas of the opossum midbrain. Contrast sections B and C in Fig. 4 and 5 and note the shifting pattern of label, especially in the facial nucleus, as the injection sites are displaced more laterally.
trigeminal and hypoglossal nuclei (Fig. 5A, C, P-505), but comparison of different cases suggested that it was a result of marker incorporation by neurons within the mesencephalic nucleus of the trigeminal nerve and that part of the lateral periaqueductal gray labeled after hypoglossal injections of HRP (see P-591, Fig. 2). Several cases are available with [aH]leucine deposits within the superior colliculus and adjacent regions. One such case, P-488 (Fig. 6) was subjected to a double injection in order to label large areas of the superior colliculus and part of the caudal pretectal complex. Terminal label was present within the ipsilateral trigeminal nucleus (Fig. 6A), as might be expected from inclusion of the mesencephalic nucleus at the injection site, and there was evidence for a sparse projection to the ventromedial facial nucleus on the ipsilateral side (Fig. 6B).
('
,
