Brain Research, 538 (1991) 231-245 Elsevier
231
BRES 16236
The projection of locus coeruleus neurons to the spinal cord in the rat determined by anterograde tracing combined with immunocytochemistry Frank M. Clark and Herbert K. Proudfit The University of Illinois at Chicago, Department of Pharmacology, Chicago, IL 60680 (U.S.A.) (Accepted 7 August 1990) Key words: Dopamine-fl-hydroxylase; Fluoro-Gold; Lesion; Norepinephrine; Phaseolus vulgaris leucoagglutinin
Pontospinal noradrenergic neurons located in the A5, A7 and locus coeruleus/subcoeruleus (LC/SC) nuclei are the major source of the noradrenergic innervation of the spinal cord. However, the specific terminations of spinally-projecting noradrenergic neurons located in these nuclei have not been clearly defined. The purpose of the experiments described in this report was to more precisely define the spinal terminations of neurons located in the LC/SC using the anterograde tracer phaseolus vulgaris-leucoagglutinin in combination with dopamine-fl-hydroxylase (DflH) immunocytochemistry. In addition, the spinal cord regions in which LC/SC neurons terminate was assessed by measuring the reduction in the density of DflH-immunoreactive axon terminals in specific spinal cord regions after a unilateral electrolytic lesion that included LC/SC neurons. The results of these experiments indicate that the axons of LC neurons are located primarily in the ipsilateral ventral funiculus and terminate most heavily in the medial part of laminae VII and VIII, the motoneuron pool of lamina IX, and lamina X. LC neurons provide a moderately dense innervation of the ventral part of the dorsal horn, but only a very sparse innervation of the superficial dorsal horn. The SC projects ipsilaterally in the ventrolateral funiculus and terminates diffusely in the intermediate and ventral laminae of the spinal cord. INTRODUCTION The innervation of the spinal cord by bulbospinal noradrenergic neurons is extensive, but mainly concentrated in the superficial dorsal horn, the ventral horn motoneuron pool and lamina X of all spinal cord segments and the intermediolateral cell column of thoracic and sacral segments 1°'32'38. Since there are no known catecholamine cell bodies in the spinal cord, it is presumed that this innervation arises from one or more of the brainstem catecholamine cell groups ( A 1 - A 1 2 ) first described by Dahlstr6m and Fuxe 7. The most compelling evidence indicates that only the pontine catecholamine cell groups (A5, A7 and nucleus locus coeruleus (A6)) have spinal projections 36-39. The innervation of the spinal cord by noradrenergic neurons located in the nucleus locus coeruleus (LC) and subcoeruleus (SC) has been studied extensively using retrograde tracers 12'13'15'19'2°'22'23'29'31'33'36-39, electrolytic lesions 1' 5,6,18,25,30
, anterograde transport of tritiated amino acids 16'26'35 and electrophysiologically by antidromic activation of spinally-projecting LC/SC neurons 12. However, none of these reports has definitively described the location of axonal terminations in the spinal cord. The
purpose of the present studies was to more accurately define the location of noradrenergic terminations in the spinal cord that arise from neurons in the LC and SC. To this end, the anterograde tracer, phaseolus vulgarisleucoagglutinin ( P H A - L ) , was used to directly determine the spinal cord projections from LC/SC in detail. In addition, fluorescence immunocytochemistry was used to detect the presence of dopamine-fl-hydroxylase (DflH) within PHA-L-labeled axon terminals. The presence of this enzyme was assumed to be a specific marker for neurons and axons that contain norepinephrine 14'34. A second approach was also used to determine the spinal projections of noradrenergic neurons located in the LC. These experiments determined the reduction in the density of DflH-labeled terminals in dorsal and ventral horns following a complete unilateral LC/SC lesion. Preliminary results of t h e s e experiments have been reported previously 4. MATERIALS AND METHODS Animal preparation All experiments were performed using adult female SpragueDawley derived rats (Sasco Inc., Omaha, NE), weighing 320-420 g. Animals were anesthetized with ketamine (100 mg/kg) and immo-
Correspondence: H.K. Proudfit, Department of Pharmacology (mc 868), University of Illinois at Chicago, P.O. Box 6998, Chicago, IL 60680, U.S.A. 0006-8993/91/$03.50 © 1991 Elsevier Science Publishers B.V. (Biomedical Division)
232 bilized in a stereotaxic frame. PHA-L injections and LC/SC lesions were made using stereotaxically-guided micropipets or metal electrodes, respectively.
Anterograde tracing experiments Multiple (3-5) unilateral iontophoretic injections of PHA-L (Vector Laboratories) were made into the left LC using the following stereotaxic coordinates: AP, -2.1; V, +2.5 and L, +1.5 mm with respect to the interaural line. The incisor bar was set at +5.0 mm. Iontophoretic current (5-10/aA) and pipette tip diameters (10-15/am) used to deliver PHA-L were similar to those used by Gerfen and Sawchenko 1~. Animals were maintained for 30 days, at which time they were perfused intracardially with 200 ml cold saline, 300 ml of cold 4% phosphate-buffered paraformaldehyde (pH = 7.4), followed by 200 ml of 10% sucrose in 0.1 M phosphate buffer. Brains and spinal cords were removed and stored at 4 °C for at least 3 days in a solution containing 30% sucrose in 0.1 M phosphate buffer. Brainstem and spinal cord tissue was blocked and 40-/am transverse sections were cut on a cryostat microtome and processed for PHA-L and DflH immunocytochemistry. All immunocytochemistry was done on free-floating sections. Sections were washed twice in 187 mM phosphate-buffered saline (PBS; pH 7.4) between each antibody incubation unless otherwise stated. Sections were cut, washed with cold PBS, treated with 2.5% H20 2 in methanol to inactivate endogenous peroxidases3, washed 4 times with PBS and incubated for 48 h at 4 °C in pooled goat anti-PHA-L (Vector) and rabbit anti-DflH (Eugene Tech) antibody solutions which contained 0.5% Triton X-100. Both antibodies were diluted 1:1000. Following the 2-day incubation, the tissue was warmed to room temperature and washed in PBS. Sections were incubated for 30 min in a solution containing pooled donkey anti-goat immunoglobulins (1:80 dilution; Chemicon), swine anti-rabbit immunoglobulins (1:80 dilution; Accurate) and 0.5% Triton X-100. The sections were washed in PBS, then placed in a solution containing goat peroxidase-antiperoxidase (PAP) complex (Chemicon) diluted 1:180 for 30 min. Sections were washed with PBS again and incubated for 4 min in a solution which contained 0.05% filtered diaminobenzidene (DAB) and 0.005% H20 2 in PBS. Sections were then washed in PBS and incubated in rabbit PAP (Cappel) diluted 1:180 for 30 min. Sections were then washed in PBS and incubated for 4 min in a 50 mM Tris-HCl buffer (pH = 8) which contained filtered DAB (0.013%) and nickel ammonium sulfate (0.67%) to which 0.005% H202 was addecl. Sections were then washed, mounted on gelatin-coated slides and coverslipped with Permount. This processing produced brown PHA-L staining and blue-black DflH-ir somata and terminals. Some of the spinal cord sections were processed for PHA-L- and DflH immunocytochemistry using fluorescent secondary antibodies. These sections were incubated as above in pooled primary antibody solutions for 48 h at 4 °C, washed and then incubated for 30 min in donkey anti-goat immunoglobulins conjugated with fluorescein isothiocyanate (FITC; Chemicon). Sections were then washed twice and incubated in sheep anti-rabbit immunoglobulins conjugated with rhodamine (Cappel). Both antibodies were diluted 1:80. Sections were washed twice, then mounted on gelatin-coated slides and coverslipped with DPX (BDH Chemicals). To determine the percentage of PHA-L-labeled axon terminals that were also DflH-ir, 10 sections from the lumbar spinal cord were systematically examined for PHA-L-labeled axon terminals under an FITC filter. When a PHA-L-labeled axon terminal was encountered, the field was viewed under a TRITC filter to determine if the axon terminal was also DflH-ir. The number of single- and double-labeled axons in these 10 sections were counted. The location of PHA-L injection sites was assessed using camera lucida drawings made of brainstem sections which included the injection site and the location of D/~H-immunoreactive (DflH-ir) neurons in the LC. The distribution of PHA-L-labeled axons and terminals in the spinal cord was determined by making composite drawings of the location of all labeled axons which appeared in 12 sequential sections taken from cervical, thoracic, or lumbar segments. Axons in the spinal cord which contained both PHA-L and
DflH-ir were assumed to have their origin in the LC or SC.
Retrograde tracing studies Rats were anesthetized with ketamine (100 mg/kg) and a laminectomy was performed to expose the dorsal aspect of the lumbar spinal cord. The dura was cut, retracted and a 30-gauge stainless steel injection cannula filled with a 2% Fluoro-Gold (Fluorochrome, Inc.) solution was inserted into the spinal cord. Three to five injections of Fluoro-Gold, each in a volume of 0.5/al were made at different depths in the spinal cord using a 10-/al microsyringe. Ten days after surgery rats were intracardially perfused as described above. Frozen sections were cut from brainstem and spinal cord blocks and washed in cold PBS, then incubated for 20 min in a solution of 2.5% H202 in methanol. Spinal cord sections were washed 4 times in PBS, then mounted on slides and coverslipped with DPX. Brainstem sections were washed 4 times and then incubated in mouse anti-tyrosine hydroxylase antibody (1:500, Incstar) overnight at 4 °C. Sections were warmed to room temperature, washed twice in PBS, then incubated in goat anti-mouse immunoglobulins conjugated with rhodamine (Cappel) for 30 min. Sections were then mounted on slides and coverslipped with DPX. All fluorescence microscopy was done using a Zeiss epifluorescence microscope. Spinal cord injection sites were examined and drawn using an X-Y plotter coupled to the microscope stage. Brainstem sections were examined and retrogradely-labeled neurons, TH-immunoreactive neurons and double-labeled neurons which contained both Fluoro-Gold and TH-immunoreactivity were plotted at the appropriate locations on the brainstem sections using an X-Y plotter.
Lesion experiments Unilateral LC/SC lesions were made using Teflon-coated stainless steel electrodes (0.013" o.d.; A-M Systems) to pass anodal current of 2 mA for 5 s. Fourteen or 40 days after surgery, animals were intracardially perfused as described above. Brainstem and spinal cord sections were cut, washed in cold PBS, incubated in 2.5% H20 2 in methanol for 20 min, washed 4 times and then incubated at 4 °C for 48 h in a solution which contained rabbit anti-DflH antibody (Eugene Tech) diluted 1:1000 in 0.5% Triton X-100. Sections were then washed twice and incubated in goat anti-rabbit immunoglobulins (Atlantic) diluted 1:80, then washed again and incubated in rabbit PAP complex (Cappel) diluted 1:180 for 30 min. Sections were again washed and then incubated for 6 rain in a solution which contained 0.05% filtered DAB and 0.005% H20 2 in PBS for 6 min. Finally, the sections were washed, mounted on gelatin-coated slides and coverslipped with Permount. The effect of LC/SC lesions on the number of DflH-ir axons in the spinal cord was assessed by quantifying the density of these axons on the ipsilateral and contralateral sides of each section using an image analysis system (JAVA, Jandel Scientific). Twenty sections taken from the lumbar spinal cord segments from each of 5 rats in which the LC and SC were completely lesioned were analyzed. The DflH-ir axons were visualized under darkfield illumination in sections processed using PAP antisera directed against the primary anti-DflH antibody. The image was digitized and the density of D~H-ir axons in selected areas of the gray matter was determined using computer-assisted densitometry. The areas of the spinal cord analyzed were the dorsal horn (approximately laminae I-IV) or the ventral horn (approximately laminae VI-IX). The accuracy of the image analysis was assessed by comparing the axon densities determined using the imaging system with that determined by manually counting the number of axon segments in the same area. Fig. 1 illustrates that the ratio of the density of labeled axons in the dorsal horn on the lesioned side to that on the side contralateral to the lesion was approximately the same using both methods of analysis.
Statistical analysis The effect of LC lesions on the density of DflH-ir axons in the
233
spinal cord was determined by comparing the density on the ipsilateral side with that on the contralateral side using Student's paired t-test. Such comparisons were made for both the dorsal and ventral horns.
RESULTS
Anterograde tracer experiments The anterograde tracer PHA-L
was iontophoretically
injected into the area of the LC and SC, and the spinal
A
--
i
¸
•
~
D
E
B f
I
C
F
500/.
m
I
120 100 80 60 40 20 0 Manual
Computer
Contralateral Ipallat~al Fig. 1. Comparison between manual and computer-assisted image analysis of the effects of LC/SC lesions on DflI-I-ir axons and axon terminals in the spinal cord. A: darkfield photomicrograph of the ipsilateral side of a lumbar spinal cord section stained for DflH-ir after a unilateral LC/SC lesion. B: enhanced image of the ipsilateral side. C: camera lueida drawing of DflH-ir axon segments and varicosities in the same section. D: darkfield photomicrograph of the side contralateral to the lesion. E: enhanced image of the side contralateral to the lesion. F: bar graph: the number of DflH-ir varicosities counted manually from the drawing in (C) are represented by the pair of bars on the left. The values derived from the image analysis are represented by the bars on the right. The values for the ipsilateral side are expressed as a percent of the values determined for the contralateral side.
234 cord terminations of neurons in the injected areas were determined. Fig. 2 shows the location of a representative
ventral caudal aspect of the locus coeruleus where the spinally-projecting LC neurons are located (Fig. 2). A
P H A - L injection site with respect to the spinallyprojecting LC neurons in the caudal and ventral aspect of
fourth P H A - L injection (Fig. 4A) was centered in the SC, but did not include DflH-ir n e u r o n s in the main body
the LC. Figs. 3 (A, B and C) and 4 (A) illustrate the
of the LC. Additional injections of P H A - L in two other rats were located ventral (Fig. 4B) and dorsal (Fig. 4C)
results of 4 experiments in which P H A - L injections encompassed DflH-ir neurons in the LC or SC. Three of these injections (Fig. 3A, B and C) were centered in the
A
Following P H A - L injections which included DflH-ir
PHA-L '
B
to the DflH-ir LC neurons.
C
TH
D
FLUORO-GOLD
3
CRESYLVIOLET F
Fig. 2. The location of a PHA-L injection in the ventral LC that encompassed some of the spinally-projecting noradrenergic neurons in this area. Panel A shows the location of the PHA-L deposit in the ventral LC. The arrows point to the 3 separate PHA-L deposits. Panel B shows the region of the LC stained with Cresyl violet. Panels C and D illustrate the location of spinally projecting LC neurons in a different animal. Panel C shows the noradrenergic neurons in the LC stained for TH-ir. Panel D shows the location of neurons in the LC that were retrogradely-labeled by Fluoro-Gold injected into the spinal cord. This panel illustrates the location of coeruleospinal neurons in the ventral LC. The arrows indicate the only Fluoro-Gold-labeled neurons that did not also contain TH-ir. Bar = 200 #m.
235
A
6'
C
Fig. 3. The location of PHA-L injection sites that included the nucleus locus coeruleus and the distribution of PHA-L-labeled axons and terminals in the spinal cord. Panels A, B and C illustrate 3 different PHA-L injection sites. An enlargement of the injection sites is shown to the right of each brainstem section, and camera lucida drawings of the distribution of PHA-L-iabeled terminals in the lumbar segments of the spinal cord are shown on the far right. The extent of PHA-L injection sites is drawn in black and the location of DflI-I-ir neurons in the locus coeruleus is represented by the stippled region. The brainstem sections in the illustration are those in which maximal PHA-L-labeling was seen. The size of the injection site was taken to be the size of the dense core of PHA-L labeling. Some diffuse PHA-L labeling was seen outside the central core of the injection sites. Each of the spinal cord drawings is a composite of 12 sections.
236
A
/
---k- t
B
C
J Fig. 4. The location of PHA-L injection sites in the dorsal pontine tegmentum and the distribution of PHA-L-iabeled axons and terminals in the spinal cord. A: PHA-L injection that encompassed DflH-ir neurons (f'flled circles) in the nucleus subcoeruleus. B: PHA-L injection that was located ventral to DflH-ir neurons in the LC and SC. C: PHA-L injection that was located dorsal to D/5'l-I-ir neurons in the LC. See the legend for Fig. 3 for details.
Fig. 5. Distribution of PHA-L-labeled axons and terminals in different levels of the spinal cord following an injection of PHA-L in the locus coeruleus. A: location of the PHA-L Injection site. B: enlargement of the injection site. The solid black area represents the extent of the PHA-L injection site and the stippled area represents the location of D~H-ir neurons in the locus coeruleus. C: photomicrograph of a thoracic spinal cord section under darkfield illumination showing the distribution of PHA-L-labeled axons and terminals. D: distribution of PHA-L-iabeled axons in cervical spinal cord segments. Each drawing is a composite of 12 contiguous sections. E: distribution of PHA-L-labeled axons in thoracic spinal cord segments. F: distribution of PHA-L-labeled axons in lumbar spinal cord segments.
237
A
D
B
E
C
F
i~
238
A
B
C
Fig. 6. Brightfieid photomicrographs of PHA-L-labeled axons and terminal segments in lamina VIII of the spinal cord following a PHA-L injection in the LC. Note the fine diameter of the axons and the location of the varicosities along the length of the axon. Bar = 40/~m.
neurons in the caudal part of the ventral LC, PHAL-labeled axons were seen to course through the medial ventral funiculus. The terminations of these axons were found mainly on the side ipsilateral to the injection site in the medial aspect of the ventral horn (laminae VIII and IX) and intermediate zone (lamina VII) and in lamina X near the central canal (Fig 3A, B and C). There was a moderate projection to the motoneuron pool of lamina IX, but there were very few PHA-L-labeled axons in the dorsal horn (laminae I-V). However, low to moderate labeling was present in the ventromedial part of the dorsal horn (lamina VI). Only sparse labeling was
found on the side contralateral to the injection site in the three cases illustrated in Fig. 3. The distribution of labeled axons and terminals in the spinal cord following injection of P H A - L in the SC was somewhat different from that produced by similar injections in the LC (Fig. 4A). The labeled axons in the lumbar spinal cord coursed through the ipsilateral ventrolaterai funiculus and terminated ipsilaterally in both the medial and lateral part of the ventral horn including the motoneuron pool of lamina IX. In addition, there was sparse labeling in the dorsal horn but the density of these terminals appeared to be higher than that produced by
239
A
C
B
D
PHA-L
DBH
Fig. 7. Examples of dopamine-fl-hydroxylase-immunoreactive axons in the ventral horn that were anterogradely labeled by a PHA-L injection in the LC. Panel A illustrates an axon in lamina X that was anterogradely labeled by PHA-L injected into the ventral LC and panel B shows that the same axon also contains DflH-ir. Panel C illustrates two other anterogradely-labeled axons located in lamina VIII of the ventral horn (arrows). Panel D shows that these two axons also contain DflH-ir. Panel D also illustrates a number of DflH-ir axons that were not anterogradely labeled from the LC. Bar = 40/zm.
L C injections. There was also some sparse labeling in the contralateral intermediate zone. Injection of P H A - L into sites ventral (Fig. 4B) or dorsal (Fig. 4C) to the LC/SC complex resulted in very little labeling in the spinal cord. These two P H A - L injections did not encompass any DflH-ir neurons in either the LC or SC. A more detailed analysis of the projections labeled by the P H A - L injection shown in Fig. 3A is illustrated in Fig. 5. Injections of P H A - L in the LC resulted in consistent labeling at all levels of the spinal cord. The labeling in the cervical segments of the spinal cord was predominantly ipsilateral and was more dense than that in the lumbar segments. However, the distribution of ipsilateral labeling was similar in all 3 spinal cord levels. In addition, there was moderate contralateral labeling in both the lateral aspect of the dorsal horn and the medial aspect of the ventral horn in the cervical segments. Such contralateral labeling was not present in either the thoracic or lumbar segments. Finally, very few labeled
axons or terminals were seen in either the intermediolateral cell column or the dorsal horn of the thoracic segments after P H A - L injections in either the LC or SC. The axons in the spinal cord that were labeled by a P H A - L injection in the LC were of fine diameter with varicosities regularly spaced along the length of the preterminal axons (Figs. 6 a n d 7). Branching of axons was infrequent and there was no apparent pattern to the orientation of the axons and terminal segments in the gray matter. Several cases were examined to ascertain whether the PHA-L-labeled axons and terminals in the spinal cord also contained DflH-ir. These double-labeling experiments were done on spinal cord sections taken from the same animal illustrated in F i g . 5. Nearly 80% of PHA-L-labeled axons and terminals also contained DflHir, which indicates that the majority of PHA-L-labeled axons and terminals in the spinal cord contain catecholamines. Examples of double-labeled axons in the spinal
240
A 0.0
B
-0.25
C
-0.50
D
Fig. 8. The location of neurons in the pontine tegmentum that were retrogradely labeled following bilateral injections of Fluoro-Gold into the lumbar segments of the spinal cord. The location of retrogradely-labeled neurons and those that contained TH-ir are plotted on coronal sections (A, B and C). Neurons labeled with Fluoro-Gold are represented by crosses, TH-ir somata are represented by open circles and TH-ir somata that were also labeled with Fluoro-Goid are represented by filled circles. All of the labeled neurons that were identified in each section have been plotted on their respective sections. The Fluoro-Gold injection site is shown in panel D. The numbers to the right of each section are the anterior-posterior stereotaxic coordinates for that section. cord which contain both P H A - L label and DflH-ir are shown in Fig. 7. In addition, the distribution of doublelabeled axon terminals determined using secondary antibodies conjugated to fluorophores was no different from the distribution of PHA-L-labeled axon terminals processed using peroxidase antiperoxidase.
Retrograde tracing experiments It is possible that spinally-projecting non-catechol-
amine-containing n e u r o n s in or near the LC may also transport P H A - L to the spinal cord. To estimate the n u m b e r and location of these neurons, large bilateral injections of the retrograde tracer F | u o r o - G o l d were made into the lumbar spinal cord. Fig. 8 shows the location of neurons labeled with retrogradely-transported Fluoro-Gold, those labeled with tyrosine hydroxylaseimmunoreactivity (TH-ir) and those double-labeled with both Fluoro-Goid and TH-ir. At the level of the caudal
241
Fig. 9. Photomicrograph of a unilateral electrolytic lesion that included the entire LC/SC complex. Bar = 1.0 mm. 6o
! 4o
60 50 40 30 20
le ~
10
~"
0
5O 40
10
5O 4O 3O
20 10
2O 10
ri
0
Ventral Horn
5O 40
~
Dorsal Horn
Conb'alateral
Ip~l * p
231
BRES 16236
The projection of locus coeruleus neurons to the spinal cord in the rat determined by anterograde tracing combined with immunocytochemistry Frank M. Clark and Herbert K. Proudfit The University of Illinois at Chicago, Department of Pharmacology, Chicago, IL 60680 (U.S.A.) (Accepted 7 August 1990) Key words: Dopamine-fl-hydroxylase; Fluoro-Gold; Lesion; Norepinephrine; Phaseolus vulgaris leucoagglutinin
Pontospinal noradrenergic neurons located in the A5, A7 and locus coeruleus/subcoeruleus (LC/SC) nuclei are the major source of the noradrenergic innervation of the spinal cord. However, the specific terminations of spinally-projecting noradrenergic neurons located in these nuclei have not been clearly defined. The purpose of the experiments described in this report was to more precisely define the spinal terminations of neurons located in the LC/SC using the anterograde tracer phaseolus vulgaris-leucoagglutinin in combination with dopamine-fl-hydroxylase (DflH) immunocytochemistry. In addition, the spinal cord regions in which LC/SC neurons terminate was assessed by measuring the reduction in the density of DflH-immunoreactive axon terminals in specific spinal cord regions after a unilateral electrolytic lesion that included LC/SC neurons. The results of these experiments indicate that the axons of LC neurons are located primarily in the ipsilateral ventral funiculus and terminate most heavily in the medial part of laminae VII and VIII, the motoneuron pool of lamina IX, and lamina X. LC neurons provide a moderately dense innervation of the ventral part of the dorsal horn, but only a very sparse innervation of the superficial dorsal horn. The SC projects ipsilaterally in the ventrolateral funiculus and terminates diffusely in the intermediate and ventral laminae of the spinal cord. INTRODUCTION The innervation of the spinal cord by bulbospinal noradrenergic neurons is extensive, but mainly concentrated in the superficial dorsal horn, the ventral horn motoneuron pool and lamina X of all spinal cord segments and the intermediolateral cell column of thoracic and sacral segments 1°'32'38. Since there are no known catecholamine cell bodies in the spinal cord, it is presumed that this innervation arises from one or more of the brainstem catecholamine cell groups ( A 1 - A 1 2 ) first described by Dahlstr6m and Fuxe 7. The most compelling evidence indicates that only the pontine catecholamine cell groups (A5, A7 and nucleus locus coeruleus (A6)) have spinal projections 36-39. The innervation of the spinal cord by noradrenergic neurons located in the nucleus locus coeruleus (LC) and subcoeruleus (SC) has been studied extensively using retrograde tracers 12'13'15'19'2°'22'23'29'31'33'36-39, electrolytic lesions 1' 5,6,18,25,30
, anterograde transport of tritiated amino acids 16'26'35 and electrophysiologically by antidromic activation of spinally-projecting LC/SC neurons 12. However, none of these reports has definitively described the location of axonal terminations in the spinal cord. The
purpose of the present studies was to more accurately define the location of noradrenergic terminations in the spinal cord that arise from neurons in the LC and SC. To this end, the anterograde tracer, phaseolus vulgarisleucoagglutinin ( P H A - L ) , was used to directly determine the spinal cord projections from LC/SC in detail. In addition, fluorescence immunocytochemistry was used to detect the presence of dopamine-fl-hydroxylase (DflH) within PHA-L-labeled axon terminals. The presence of this enzyme was assumed to be a specific marker for neurons and axons that contain norepinephrine 14'34. A second approach was also used to determine the spinal projections of noradrenergic neurons located in the LC. These experiments determined the reduction in the density of DflH-labeled terminals in dorsal and ventral horns following a complete unilateral LC/SC lesion. Preliminary results of t h e s e experiments have been reported previously 4. MATERIALS AND METHODS Animal preparation All experiments were performed using adult female SpragueDawley derived rats (Sasco Inc., Omaha, NE), weighing 320-420 g. Animals were anesthetized with ketamine (100 mg/kg) and immo-
Correspondence: H.K. Proudfit, Department of Pharmacology (mc 868), University of Illinois at Chicago, P.O. Box 6998, Chicago, IL 60680, U.S.A. 0006-8993/91/$03.50 © 1991 Elsevier Science Publishers B.V. (Biomedical Division)
232 bilized in a stereotaxic frame. PHA-L injections and LC/SC lesions were made using stereotaxically-guided micropipets or metal electrodes, respectively.
Anterograde tracing experiments Multiple (3-5) unilateral iontophoretic injections of PHA-L (Vector Laboratories) were made into the left LC using the following stereotaxic coordinates: AP, -2.1; V, +2.5 and L, +1.5 mm with respect to the interaural line. The incisor bar was set at +5.0 mm. Iontophoretic current (5-10/aA) and pipette tip diameters (10-15/am) used to deliver PHA-L were similar to those used by Gerfen and Sawchenko 1~. Animals were maintained for 30 days, at which time they were perfused intracardially with 200 ml cold saline, 300 ml of cold 4% phosphate-buffered paraformaldehyde (pH = 7.4), followed by 200 ml of 10% sucrose in 0.1 M phosphate buffer. Brains and spinal cords were removed and stored at 4 °C for at least 3 days in a solution containing 30% sucrose in 0.1 M phosphate buffer. Brainstem and spinal cord tissue was blocked and 40-/am transverse sections were cut on a cryostat microtome and processed for PHA-L and DflH immunocytochemistry. All immunocytochemistry was done on free-floating sections. Sections were washed twice in 187 mM phosphate-buffered saline (PBS; pH 7.4) between each antibody incubation unless otherwise stated. Sections were cut, washed with cold PBS, treated with 2.5% H20 2 in methanol to inactivate endogenous peroxidases3, washed 4 times with PBS and incubated for 48 h at 4 °C in pooled goat anti-PHA-L (Vector) and rabbit anti-DflH (Eugene Tech) antibody solutions which contained 0.5% Triton X-100. Both antibodies were diluted 1:1000. Following the 2-day incubation, the tissue was warmed to room temperature and washed in PBS. Sections were incubated for 30 min in a solution containing pooled donkey anti-goat immunoglobulins (1:80 dilution; Chemicon), swine anti-rabbit immunoglobulins (1:80 dilution; Accurate) and 0.5% Triton X-100. The sections were washed in PBS, then placed in a solution containing goat peroxidase-antiperoxidase (PAP) complex (Chemicon) diluted 1:180 for 30 min. Sections were washed with PBS again and incubated for 4 min in a solution which contained 0.05% filtered diaminobenzidene (DAB) and 0.005% H20 2 in PBS. Sections were then washed in PBS and incubated in rabbit PAP (Cappel) diluted 1:180 for 30 min. Sections were then washed in PBS and incubated for 4 min in a 50 mM Tris-HCl buffer (pH = 8) which contained filtered DAB (0.013%) and nickel ammonium sulfate (0.67%) to which 0.005% H202 was addecl. Sections were then washed, mounted on gelatin-coated slides and coverslipped with Permount. This processing produced brown PHA-L staining and blue-black DflH-ir somata and terminals. Some of the spinal cord sections were processed for PHA-L- and DflH immunocytochemistry using fluorescent secondary antibodies. These sections were incubated as above in pooled primary antibody solutions for 48 h at 4 °C, washed and then incubated for 30 min in donkey anti-goat immunoglobulins conjugated with fluorescein isothiocyanate (FITC; Chemicon). Sections were then washed twice and incubated in sheep anti-rabbit immunoglobulins conjugated with rhodamine (Cappel). Both antibodies were diluted 1:80. Sections were washed twice, then mounted on gelatin-coated slides and coverslipped with DPX (BDH Chemicals). To determine the percentage of PHA-L-labeled axon terminals that were also DflH-ir, 10 sections from the lumbar spinal cord were systematically examined for PHA-L-labeled axon terminals under an FITC filter. When a PHA-L-labeled axon terminal was encountered, the field was viewed under a TRITC filter to determine if the axon terminal was also DflH-ir. The number of single- and double-labeled axons in these 10 sections were counted. The location of PHA-L injection sites was assessed using camera lucida drawings made of brainstem sections which included the injection site and the location of D/~H-immunoreactive (DflH-ir) neurons in the LC. The distribution of PHA-L-labeled axons and terminals in the spinal cord was determined by making composite drawings of the location of all labeled axons which appeared in 12 sequential sections taken from cervical, thoracic, or lumbar segments. Axons in the spinal cord which contained both PHA-L and
DflH-ir were assumed to have their origin in the LC or SC.
Retrograde tracing studies Rats were anesthetized with ketamine (100 mg/kg) and a laminectomy was performed to expose the dorsal aspect of the lumbar spinal cord. The dura was cut, retracted and a 30-gauge stainless steel injection cannula filled with a 2% Fluoro-Gold (Fluorochrome, Inc.) solution was inserted into the spinal cord. Three to five injections of Fluoro-Gold, each in a volume of 0.5/al were made at different depths in the spinal cord using a 10-/al microsyringe. Ten days after surgery rats were intracardially perfused as described above. Frozen sections were cut from brainstem and spinal cord blocks and washed in cold PBS, then incubated for 20 min in a solution of 2.5% H202 in methanol. Spinal cord sections were washed 4 times in PBS, then mounted on slides and coverslipped with DPX. Brainstem sections were washed 4 times and then incubated in mouse anti-tyrosine hydroxylase antibody (1:500, Incstar) overnight at 4 °C. Sections were warmed to room temperature, washed twice in PBS, then incubated in goat anti-mouse immunoglobulins conjugated with rhodamine (Cappel) for 30 min. Sections were then mounted on slides and coverslipped with DPX. All fluorescence microscopy was done using a Zeiss epifluorescence microscope. Spinal cord injection sites were examined and drawn using an X-Y plotter coupled to the microscope stage. Brainstem sections were examined and retrogradely-labeled neurons, TH-immunoreactive neurons and double-labeled neurons which contained both Fluoro-Gold and TH-immunoreactivity were plotted at the appropriate locations on the brainstem sections using an X-Y plotter.
Lesion experiments Unilateral LC/SC lesions were made using Teflon-coated stainless steel electrodes (0.013" o.d.; A-M Systems) to pass anodal current of 2 mA for 5 s. Fourteen or 40 days after surgery, animals were intracardially perfused as described above. Brainstem and spinal cord sections were cut, washed in cold PBS, incubated in 2.5% H20 2 in methanol for 20 min, washed 4 times and then incubated at 4 °C for 48 h in a solution which contained rabbit anti-DflH antibody (Eugene Tech) diluted 1:1000 in 0.5% Triton X-100. Sections were then washed twice and incubated in goat anti-rabbit immunoglobulins (Atlantic) diluted 1:80, then washed again and incubated in rabbit PAP complex (Cappel) diluted 1:180 for 30 min. Sections were again washed and then incubated for 6 rain in a solution which contained 0.05% filtered DAB and 0.005% H20 2 in PBS for 6 min. Finally, the sections were washed, mounted on gelatin-coated slides and coverslipped with Permount. The effect of LC/SC lesions on the number of DflH-ir axons in the spinal cord was assessed by quantifying the density of these axons on the ipsilateral and contralateral sides of each section using an image analysis system (JAVA, Jandel Scientific). Twenty sections taken from the lumbar spinal cord segments from each of 5 rats in which the LC and SC were completely lesioned were analyzed. The DflH-ir axons were visualized under darkfield illumination in sections processed using PAP antisera directed against the primary anti-DflH antibody. The image was digitized and the density of D~H-ir axons in selected areas of the gray matter was determined using computer-assisted densitometry. The areas of the spinal cord analyzed were the dorsal horn (approximately laminae I-IV) or the ventral horn (approximately laminae VI-IX). The accuracy of the image analysis was assessed by comparing the axon densities determined using the imaging system with that determined by manually counting the number of axon segments in the same area. Fig. 1 illustrates that the ratio of the density of labeled axons in the dorsal horn on the lesioned side to that on the side contralateral to the lesion was approximately the same using both methods of analysis.
Statistical analysis The effect of LC lesions on the density of DflH-ir axons in the
233
spinal cord was determined by comparing the density on the ipsilateral side with that on the contralateral side using Student's paired t-test. Such comparisons were made for both the dorsal and ventral horns.
RESULTS
Anterograde tracer experiments The anterograde tracer PHA-L
was iontophoretically
injected into the area of the LC and SC, and the spinal
A
--
i
¸
•
~
D
E
B f
I
C
F
500/.
m
I
120 100 80 60 40 20 0 Manual
Computer
Contralateral Ipallat~al Fig. 1. Comparison between manual and computer-assisted image analysis of the effects of LC/SC lesions on DflI-I-ir axons and axon terminals in the spinal cord. A: darkfield photomicrograph of the ipsilateral side of a lumbar spinal cord section stained for DflH-ir after a unilateral LC/SC lesion. B: enhanced image of the ipsilateral side. C: camera lueida drawing of DflH-ir axon segments and varicosities in the same section. D: darkfield photomicrograph of the side contralateral to the lesion. E: enhanced image of the side contralateral to the lesion. F: bar graph: the number of DflH-ir varicosities counted manually from the drawing in (C) are represented by the pair of bars on the left. The values derived from the image analysis are represented by the bars on the right. The values for the ipsilateral side are expressed as a percent of the values determined for the contralateral side.
234 cord terminations of neurons in the injected areas were determined. Fig. 2 shows the location of a representative
ventral caudal aspect of the locus coeruleus where the spinally-projecting LC neurons are located (Fig. 2). A
P H A - L injection site with respect to the spinallyprojecting LC neurons in the caudal and ventral aspect of
fourth P H A - L injection (Fig. 4A) was centered in the SC, but did not include DflH-ir n e u r o n s in the main body
the LC. Figs. 3 (A, B and C) and 4 (A) illustrate the
of the LC. Additional injections of P H A - L in two other rats were located ventral (Fig. 4B) and dorsal (Fig. 4C)
results of 4 experiments in which P H A - L injections encompassed DflH-ir neurons in the LC or SC. Three of these injections (Fig. 3A, B and C) were centered in the
A
Following P H A - L injections which included DflH-ir
PHA-L '
B
to the DflH-ir LC neurons.
C
TH
D
FLUORO-GOLD
3
CRESYLVIOLET F
Fig. 2. The location of a PHA-L injection in the ventral LC that encompassed some of the spinally-projecting noradrenergic neurons in this area. Panel A shows the location of the PHA-L deposit in the ventral LC. The arrows point to the 3 separate PHA-L deposits. Panel B shows the region of the LC stained with Cresyl violet. Panels C and D illustrate the location of spinally projecting LC neurons in a different animal. Panel C shows the noradrenergic neurons in the LC stained for TH-ir. Panel D shows the location of neurons in the LC that were retrogradely-labeled by Fluoro-Gold injected into the spinal cord. This panel illustrates the location of coeruleospinal neurons in the ventral LC. The arrows indicate the only Fluoro-Gold-labeled neurons that did not also contain TH-ir. Bar = 200 #m.
235
A
6'
C
Fig. 3. The location of PHA-L injection sites that included the nucleus locus coeruleus and the distribution of PHA-L-labeled axons and terminals in the spinal cord. Panels A, B and C illustrate 3 different PHA-L injection sites. An enlargement of the injection sites is shown to the right of each brainstem section, and camera lucida drawings of the distribution of PHA-L-iabeled terminals in the lumbar segments of the spinal cord are shown on the far right. The extent of PHA-L injection sites is drawn in black and the location of DflI-I-ir neurons in the locus coeruleus is represented by the stippled region. The brainstem sections in the illustration are those in which maximal PHA-L-labeling was seen. The size of the injection site was taken to be the size of the dense core of PHA-L labeling. Some diffuse PHA-L labeling was seen outside the central core of the injection sites. Each of the spinal cord drawings is a composite of 12 sections.
236
A
/
---k- t
B
C
J Fig. 4. The location of PHA-L injection sites in the dorsal pontine tegmentum and the distribution of PHA-L-iabeled axons and terminals in the spinal cord. A: PHA-L injection that encompassed DflH-ir neurons (f'flled circles) in the nucleus subcoeruleus. B: PHA-L injection that was located ventral to DflH-ir neurons in the LC and SC. C: PHA-L injection that was located dorsal to D/5'l-I-ir neurons in the LC. See the legend for Fig. 3 for details.
Fig. 5. Distribution of PHA-L-labeled axons and terminals in different levels of the spinal cord following an injection of PHA-L in the locus coeruleus. A: location of the PHA-L Injection site. B: enlargement of the injection site. The solid black area represents the extent of the PHA-L injection site and the stippled area represents the location of D~H-ir neurons in the locus coeruleus. C: photomicrograph of a thoracic spinal cord section under darkfield illumination showing the distribution of PHA-L-labeled axons and terminals. D: distribution of PHA-L-iabeled axons in cervical spinal cord segments. Each drawing is a composite of 12 contiguous sections. E: distribution of PHA-L-labeled axons in thoracic spinal cord segments. F: distribution of PHA-L-labeled axons in lumbar spinal cord segments.
237
A
D
B
E
C
F
i~
238
A
B
C
Fig. 6. Brightfieid photomicrographs of PHA-L-labeled axons and terminal segments in lamina VIII of the spinal cord following a PHA-L injection in the LC. Note the fine diameter of the axons and the location of the varicosities along the length of the axon. Bar = 40/~m.
neurons in the caudal part of the ventral LC, PHAL-labeled axons were seen to course through the medial ventral funiculus. The terminations of these axons were found mainly on the side ipsilateral to the injection site in the medial aspect of the ventral horn (laminae VIII and IX) and intermediate zone (lamina VII) and in lamina X near the central canal (Fig 3A, B and C). There was a moderate projection to the motoneuron pool of lamina IX, but there were very few PHA-L-labeled axons in the dorsal horn (laminae I-V). However, low to moderate labeling was present in the ventromedial part of the dorsal horn (lamina VI). Only sparse labeling was
found on the side contralateral to the injection site in the three cases illustrated in Fig. 3. The distribution of labeled axons and terminals in the spinal cord following injection of P H A - L in the SC was somewhat different from that produced by similar injections in the LC (Fig. 4A). The labeled axons in the lumbar spinal cord coursed through the ipsilateral ventrolaterai funiculus and terminated ipsilaterally in both the medial and lateral part of the ventral horn including the motoneuron pool of lamina IX. In addition, there was sparse labeling in the dorsal horn but the density of these terminals appeared to be higher than that produced by
239
A
C
B
D
PHA-L
DBH
Fig. 7. Examples of dopamine-fl-hydroxylase-immunoreactive axons in the ventral horn that were anterogradely labeled by a PHA-L injection in the LC. Panel A illustrates an axon in lamina X that was anterogradely labeled by PHA-L injected into the ventral LC and panel B shows that the same axon also contains DflH-ir. Panel C illustrates two other anterogradely-labeled axons located in lamina VIII of the ventral horn (arrows). Panel D shows that these two axons also contain DflH-ir. Panel D also illustrates a number of DflH-ir axons that were not anterogradely labeled from the LC. Bar = 40/zm.
L C injections. There was also some sparse labeling in the contralateral intermediate zone. Injection of P H A - L into sites ventral (Fig. 4B) or dorsal (Fig. 4C) to the LC/SC complex resulted in very little labeling in the spinal cord. These two P H A - L injections did not encompass any DflH-ir neurons in either the LC or SC. A more detailed analysis of the projections labeled by the P H A - L injection shown in Fig. 3A is illustrated in Fig. 5. Injections of P H A - L in the LC resulted in consistent labeling at all levels of the spinal cord. The labeling in the cervical segments of the spinal cord was predominantly ipsilateral and was more dense than that in the lumbar segments. However, the distribution of ipsilateral labeling was similar in all 3 spinal cord levels. In addition, there was moderate contralateral labeling in both the lateral aspect of the dorsal horn and the medial aspect of the ventral horn in the cervical segments. Such contralateral labeling was not present in either the thoracic or lumbar segments. Finally, very few labeled
axons or terminals were seen in either the intermediolateral cell column or the dorsal horn of the thoracic segments after P H A - L injections in either the LC or SC. The axons in the spinal cord that were labeled by a P H A - L injection in the LC were of fine diameter with varicosities regularly spaced along the length of the preterminal axons (Figs. 6 a n d 7). Branching of axons was infrequent and there was no apparent pattern to the orientation of the axons and terminal segments in the gray matter. Several cases were examined to ascertain whether the PHA-L-labeled axons and terminals in the spinal cord also contained DflH-ir. These double-labeling experiments were done on spinal cord sections taken from the same animal illustrated in F i g . 5. Nearly 80% of PHA-L-labeled axons and terminals also contained DflHir, which indicates that the majority of PHA-L-labeled axons and terminals in the spinal cord contain catecholamines. Examples of double-labeled axons in the spinal
240
A 0.0
B
-0.25
C
-0.50
D
Fig. 8. The location of neurons in the pontine tegmentum that were retrogradely labeled following bilateral injections of Fluoro-Gold into the lumbar segments of the spinal cord. The location of retrogradely-labeled neurons and those that contained TH-ir are plotted on coronal sections (A, B and C). Neurons labeled with Fluoro-Gold are represented by crosses, TH-ir somata are represented by open circles and TH-ir somata that were also labeled with Fluoro-Goid are represented by filled circles. All of the labeled neurons that were identified in each section have been plotted on their respective sections. The Fluoro-Gold injection site is shown in panel D. The numbers to the right of each section are the anterior-posterior stereotaxic coordinates for that section. cord which contain both P H A - L label and DflH-ir are shown in Fig. 7. In addition, the distribution of doublelabeled axon terminals determined using secondary antibodies conjugated to fluorophores was no different from the distribution of PHA-L-labeled axon terminals processed using peroxidase antiperoxidase.
Retrograde tracing experiments It is possible that spinally-projecting non-catechol-
amine-containing n e u r o n s in or near the LC may also transport P H A - L to the spinal cord. To estimate the n u m b e r and location of these neurons, large bilateral injections of the retrograde tracer F | u o r o - G o l d were made into the lumbar spinal cord. Fig. 8 shows the location of neurons labeled with retrogradely-transported Fluoro-Gold, those labeled with tyrosine hydroxylaseimmunoreactivity (TH-ir) and those double-labeled with both Fluoro-Goid and TH-ir. At the level of the caudal
241
Fig. 9. Photomicrograph of a unilateral electrolytic lesion that included the entire LC/SC complex. Bar = 1.0 mm. 6o
! 4o
60 50 40 30 20
le ~
10
~"
0
5O 40
10
5O 4O 3O
20 10
2O 10
ri
0
Ventral Horn
5O 40
~
Dorsal Horn
Conb'alateral
Ip~l * p
