LORETA MEDINA, KEITH D. ANDERSON, ELLEN J. KARLE, AND ANTON KEINER. Department of Anatomy and Neurobiology, University of Tennessee, ...
THE JOURNAL OF COMPARATIVE NEUROLOGY 357:408432 (1995)
An Ultrastructural Double-Label Immunohistochemical Study of the Enkephalinergic Input to Dopaminergic Neurons of the Substantia Nigra in Pigeons LORETA MEDINA, KEITH D. ANDERSON, ELLEN J. KARLE, AND ANTON KEINER Department of Anatomy and Neurobiology, University of Tennessee, Memphis, The Health Science Center, Memphis, Tennessee 38163
ABSTRACT Electron microscopic immunohistochemical double-label studies were carried out in pigeons to characterize the ultrastructural organization and postsynaptic targets of enkephalinergic (ENK+ ) striatonigral projection. ENK+ terminals in the substantia nigra were labeled with antileucine-enkephalin antiserum by using peroxidase-ant.iperoxidase methods, and dopaminergic neurons were labeled with anti-tyrosine hydroxylase antiserum by using silver-intensified immunogold methods. ENK+ terminals on dopaminergic neurons were equal in abundance to ENK+ terminals on nondopaminergic neurons, although the former were typically somewhat smaller than the latter (mean size: 0.50 vs. 0.75 pm, respectively). ENK+ terminals were evenly distributed on the cell bodies and dendrites of dopaminergic neurons, and they were evenly distributed on dendrites but rare on perikarya of nondopaminergic neurons. Transection of the basal telencephalic output revealed that 75% of the nigral ENK+ terminals were of basal telencephalic origin. These telencephalic ENK+ terminals included over 80% of those smaller than 0.80 pm on dopaminergic neurons and smaller than 1.0 pm on nondopaminergic neurons, and none greater than this in size. Both telencephalic and the nontelencephalic ENK+ nigral terminals made predominantly symmetric synapses on nigral neurons. Although the basal telencephalic ENK+ terminals uniformly targeted dendrites and perikarya, nontelencephalic ENK+ terminals seemed to avoid perikarya. The results indicate that ENK+ striatonigral neurons in birds may directly influence both dopaminergic and nondopaminergic neurons of the substantia nigra. Based on similar data for substance P-containing striatonigral terminals, the roles of enkephalin and substance P in influencing nigral dopaminergic neurons may differ slightly, as they appear to target preferentially different portions of dopaminergic neurons. The overall results in pigeons are similar to those for ENK+ terminals in the ventral tegmental area in rats, suggesting that the synaptic organization of the ENK+ input to the tegmental dopaminergic cell fields is similar in mammals and birds. i: i ~ rWile?-Liss, , Inc. Indexing terms: basal ganglia, electron microscopy, striatonigral, striatum, comparative neuroanatomy
The substantia nigra plays an important role in control of motor behaviors (Reiner et al., 1984a; Parent, 1986; Albin et al., 1989). In mammals, birds and reptiles, the substantia nigra receives its major input from the striatal part of the basal ganglia, and it projects to the midbrain tectum and thalamus (Reiner et al., 1980,1984a; Chevalier and Deniau, 1990; Chevalier et al.. 1985; Deniau and Chevalier, 1985; Gerfen, 1985; Williams and Faull, 1985; Parent, 1986; Albin et al., 1989; Anderson and Reiner, 1990a,b, 1991; Reiner and Anderson, 1990; Medina and Smeets, 1991; o 1995 WILEY-LISS, INC.
Smith and Bolam, 1991). The striatal input to the substantia nigra is mediated by two major types of neurons: (1) neurons containing tachykinins (substance P and neurokinin A), dynorphins (DYN), and y-aminobutyric acid (GABA); and ( 2 ) neurons containing enkephalins (ENK) and GABA Accepted December 14, 1994. Address reprint requests to LoretaMedina, Ph.D., Department ofAnatomy and Neurobiology, University of Tennessee, 855 Monroe Avenue, Mpmphis, TN, 38163
ENK INPUT TO NIGRAL DA NEURONS IN PIGEONS (Oertel et al., 1983; Aronin et al., 1984; Besson et al., 1986, 1988; Reiner, 1986a; Sugimoto and Mizuno, 1987; Anderson and Reiner, 1990a; Reiner and Anderson, 1990). The vast majority of striatonigral projection neurons (95%) have been shown in rats and pigeons to contain tachykinins, DYN, and GABA (Gerfen and Young, 1988; Reiner and Anderson, 1990; Anderson and Reiner, 1991). Consistent with this, light microscopic (LM) immunohistochemical analysis of the substantia nigra shows that in all amniote species studied the nigra receives a dense innervation by fibers containing substance P (SP), most of which arise from the striatum (Brauth et al., 1983; Reiner et al., 1983; Tnagaki and Parent, 1984; Reiner, 1986a; Haber and Groenewegen, 1989; Anderson and Reiner, 1990a,b, 1991; Reiner and Anderson, 1990). Electron microscopic (EM) immunocytochemical studies in birds and mammals show that the SP-containing terminals in the nigra synapse on dopaminergic (DA) and non-DA neurons (Kawai et al., 1987; Chang, 1988; Mahalik, 1988; Mendez et al., 1989; Bolam and Smith, 1990; Anderson et al., 1991; Smith and Bolam, 1991), thus indicating direct influence of the SP+ striatal neurons on the dopaminergic neurons projecting to striatum and nondopaminergic projecting to thalamus, tectum and locally within the nigra. In contrast to the S P t striatonigral projection system, light microscopic immunohistochemical studies indicate interspecies variation in the prominence of the ENK+ innervation of the substantia nigra. For example, the ENK+ innervation is abundant in the substantia nigra of pigeons, monkeys, and apparently humans; it is moderate in cats; but it is scarce in rats and some reptiles (Haber and Elde, 1982; Khachaturian et al., 1983; Marshall et al., 1983; Inagaki and Parent, 1984; Reiner, 1986a; Sugimoto and Mizuno, 1987; Anderson and Reiner, 1991). Studies involving striatal lesions or surgical disconnection of the basal forebrain from the tegmentum show that the majority of the ENK+ fibers in the substantia nigra in birds and mammals arise from the basal forebrain (Marshall et al., 1983; Reiner, 1986b; Sugimoto and Mizuno, 1987). Retrograde labeling studies show that this ENK+ projection arises from about 5% of the striatonigral projection neurons (Gerfen and Young, 1988; Anderson and Heiner, 1991).Additionally, ENK+ neurons in such limbic striatal territories as the nucleus accumbens and the ventral pallidum contribute ENK+ fibers to the ventral tegmental area and the substantia nigra (Parent, 1986; Harlan et al., 1987; Haber et al., 1990; Anderson and Reiner, 1991; Kalivas et al., 1993; Berendse et al., 1992).Although opiate receptor agonists are known to have potent effects on midbrain DA neurons and to affect locomotor behavior (Clouet and Ratner, 1970; Judson and Goldstein, 1978; Joyce and Iversen, 1979; Alper et al., 1980), the ENK+ projection to tegmental DA neurons has not been studied extensively at the ultrastructural level (Inagaki et al., 1986; Sesack and Pickel, 1992; Liang et al., 1993). To understand better the influence of the striatum an the substantia nigra and on the control of movement, we analyzed in pigeons the relationship between ENK+ terminals and nigral DA and non-DA neurons. We chose pigeons for study for several reasons. First, we wished to study a species with a prominent ENK+ striatonigral system. Second, such studies would enable us to compare the results with our previous ultrastructural study on the SP+ striatonigral system in pigeons (Anderson et al., 1991) and make some inferences about the
409
relative roles of these striatonigral systems in the same species. Third, we would be able to compare our results with those from similar studies in rats (Sesack and Pickel, 1992; Liang et al., 1993), with the rationale that any features found to be common to two species as disparate as pigeons and rats would likely be fundamental features of basal ganglia organization shared by many amniote species, including, for example, humans. In brief, we found that the vast majority of the ENK+ terminals in the avian nigra are of basal forebrain origm, and they equally target DA and non-DA neurons. As considered further in the discussion, our results also show that, although SP+ and ENK+ striatonigral neurons similarly target nigral non-DA neurons, they synapse somewhat differentially on DA neurons.
MATERIALS AND METHODS Animals and tissue fixation Animals. Nine White Carneaux pigeons (Colunbaliuia) were used in this study. Four of these birds were used for EM immunohistochemical double-label studies on the ENK+ input to dopaminergic and nondopaminerglc nigral neurons, one was used for EM immunohistochemical study of the ENK+ input to dopaminergic and nondopaminergic nigral neurons following removal of the forebrain input, to the tegmentum, and four were used for LM immunohistochemical study of the distribution of ENK+ fibers in the tegmentum following removal of forebrain input. EM tissue fixation. The five pigeons used in the EM studies were deeply anesthetized with chloral hydrate (Sigma; 7 g/kg, i.p.) and perfused transcardially according to the procedures described by Anderson et al. (1991) for EM double-label imrnunohistochemistry. The one pigeon used t o study the tegmentum at the EM level following removal of the forebrain input was deeply anesthetized with a combination of ketamine hydrochloride (Ketaset; 67 mgikg, i.p.) and xylazine (Gemini; 6.6 mg/kg, i.p.) 1month prior to transcardial perfusion. Stereotaxic procedures were then used (Karten and Hodos, 1967) to make a unilateral knife cut of the medial forebrain bundle/ansa lenticularis at rostral diencephalic levels, as described by Reiner et al. (1983). Each animal processed for EM was successively transcardially perfused with (1) 20 ml heparin (1,000 unitsiml) in 0.75% saline, (2) 75 ml of 3.5% acrolein and 2.0% paraformaldehyde in 0.1 M sodium phosphate buffer (PB; pH 7.2), and (3) 300 ml of 2.0% paraformaldehyde in 0.1 M PB (pH 7.2). Brains were then removed, and each was cut into six transverse slices (each about 5-mm thick) and postfixed for 2 hours in 2% paraformaldehyde in 0.1 M PB (pH 7.2) at 4°C. The slices were then transferred to 0.1 M PB containing 0.02% sodium azide, and refrigerated a t 4°C overnight. The next day, brain slices were sectioned at 40 +m in the transverse plane on a Vibratome, and sections were collected in 0.1 M PB (pH 7.2). Vibratome sections were then processed for EM preembedding double-label immunohistochemistry, as detailed below. LM tissue fixation. The four pigeons used in the LM studies were deeply anesthetized with a combination of ketamine hydrochloride (Ketaset; 67 mg/kg, i.p.) and xylazine (Gemini; 6.6 mgikg, i.p.), and stereotaxic procedures were used (Karten and Hodos, 1967) to make a unilateral knife cut of the medial forebrain bundleiansa lenticularis at rostral diencephalic levels. These pigeons were allowed to survive 2-3 weeks, and were then perfused with 50 ml of 6% dextran, followed by 300 ml of 4% paraformaldehyde, and 0.1% glutaraldehyde in 0.1 M PB (pH 7.2). The brains of
L. MEDINA ET AL.
410 these pigeons were then removed from the skull and sectioned at 40 km in the transverse plane on a sliding microtome after cryoprotection for at least 24 hours in a 0.1 M PB (pH 7.2) solution containing 20% sucrose, 10% glycerol, and 0.02% sodium azide, as described by Anderson and Reiner (1991). Sliding microtome sections were then processed for LM single-label immunohistochemistry by using antibodies against leucine-enkephalin (LENK), SP, or tyrosine hydroxylase (TH). Our procedures for carrying out single-label immunohistochemistry have been described in detail elsewhere (e.g., Anderson and Reiner, 1990a,b).
Preembedding double-label immunohistochemistry for ENK and TH The double-label procedure used involved combining the peroxidase-antiperoxidase (PAP)technique to label LENK+ terminals and a silver-intensified immunogold (SIG) technique to label TH+ neurons in the nigra (Mahalik, 1988; Chan et al., 1990; Anderson et al., 1991, 1994; Karle et al., 1992, 1994). To carry out the labeling, sections were first pretreated with 1%sodium borohydride in 0.1 M PB for 30 minutes, followed by thorough rinsing ( 5 times, 10 minutes each) in 0.1 M PB. Sections were then incubated for 30 minutes at room temperature in a blocking solution consisting of 400 mg bovine albumin (BA; Sigma), 500 p1 normal goat serum (NGS; Sigma), 250 p1 gelatin (provided by Amersham with the gold-conjugated antiserum), and 7 mg sodium azide in 47 ml 0.01 M phosphate-buffered saline (PBS; pH 7.4). Sections were then washed once in PBS (1 minute) and incubated simultaneously in both primary antisera, rabbit anti-LENK (Incstar, diluted 1:250) and mouse anti-TH (Boehringer-Mannheim, diluted 1:25),for 2 days at 4°C on a rotator. After rinsing in PBS ( 3 times, 10 minutes each), the sections were incubated in a goat anti-rabbit IgG (Jackson ImmunoResearch, diluted 150) for 1hour at room temperature, then washed 3 times in PB, and finally incubated in a rabbit PAP complex (Jackson IrnmunoItesearch, diluted 1 : l O O ) for 1 hour at room temperature. Following washing, the LENK+ structures were visualized by immersion in 50 mlO.05 M cacodylate buffer containing 50 mg 3,3‘-diaminobenzidine tetrahydrochloride (DAB; Sigma) and 175 mg imidazole for 10 minutes, followed by an additional 10-minute incubation in this solution after addition of 200 bl 3% hydrogen peroxide (Anderson and Reiner, 1990a). Sections were then thoroughly washed in PBS (6 times, 10 minutes each) and incubated in a goat anti-mouse IgG conjugated to l-nmdiameter gold particles ( h e r s h a m , diluted 1:50) for 2 hours at room temperature. The sections were then rinsed in PBS and stored overnight at 4°C. The following day, the sections were rinsed in PBS and then 1%sodium acetate. The gold was silver-intensified by using the IntenSE kit (Amersham), incubating for 4, 6, or 8 minutes at room temperature (25°C).
Tissue preparation for electron microscopy Following immunohistochemical labeling, sections were rinsed in 0.1 M cacodylate buffer (CB) and postfixed for 1 hour in 2.0%osmium tetroxide in 0.1 M CB. Sections were then rinsed 3 times in 0.1 M CB, 3 times in 0.1 M PB, and stored in 0.1 M PB containing 0.02% sodium azide overnight. Sections were next dehydrated in a graded series o f alcohols, stained in 1%uranyl acetate in 100% ethanol for 1
hour, rinsed 3 times in 100%ethanol, and finally embedded in Spurr’s resin (Electron Microscopy Sciences) by immersion in the following solutions: ( I ) a 50:50 mixture of Spurr’s resin and 100% ethanol for 4 hours, (2) 100% Spurr’s resin overnight, and (3) fresh 100% Spurr’s resin for 2 hours. Embedded sections were mounted flat on glass slides, covered with plastic coverslips, and cured in an oven for 48-72 hours at 60°C. Pieces of embedded tissue were cut from the substantia nigra and mounted onto blocks, and ultrathin sections were cut from the tissue surface with an ultramicrotome. These sections were mounted onto mesh grids, stained with lead citrate and uranyl acetate (LKB Ultrostainer), and finally viewed and photographed with a JEOL 1200 electron microscope.
Quantitative electron microscopic analysis The EM data presented in this paper are based on examinations of midrostrocaudal levels of the pigeon substantia nigra, as shown in Figure 1. Sections from the tegmentum in each of the four normal pigeons and from the normal tegmentum in the one unilaterally dederented animal were scanned with the electron microscope to collect data on the normal nigra. LENK+ terminals contacting either TH+ or non-TH structures were photographed when encountered. In many cases, more than one terminal was present in the photograph. To ensure that non-TH targets were in fact nondopaminergic, we ( 1 ) only sampled the upper few microns of tissue, in which penetration of immunogold reagents was optimal; and (2) only accepted non-TH+ structures as nondopaminergic when TH + structures were present in the same photographic field (which ensures as much as possible that the non-TH structure did not merely represent a labeling failure). Because the ability to identify reliably postsynaptic targets and synaptic specidizations was critical for our quantification, we concentrated our greatest efforts and collected the greatest number of terminals from those two cases in which the ultrastructural preservation was optimal, i.e., one of the normal pigeons and the normal side of the deafferented pigeon. Note that the results for these two animals were highly similar. The photographs collected were used to (1) measure the size of LENK+ terminals, (2) identify the postsynaptic targets of the LENK+ terminals, (3) determine the synaptic specialization present (if any), (4) quantify the relative frequency of the TH+ versus non-TH targets and the different types of synaptic specializations observed, (5) measure the size of the various TH+ and non-TH dendrites receiving LENK+ input, and (6) categorize LENK+ terminals according to the relative frequency with which they were rich or poor in mitochondria and rich or poor in dense core vesicles. All of this information could not necessarily be determined for each LENK+ terminal photographed. Terminal and dendrite dimensions were measured along the shortest diameter because the long dimension might correspond to the shaft of the axon leading to the terminal or correspond to the dendritic shaft, respectively. Dendrites were classified into three size categories: (1) small: short diameter less than or equal to 0.75 km, (2) medium sized: short diameter between 0.76 and 1.5 pm, and ( 3 ) large: short diameter larger than 1.5 pm. To obtain the data on the normal tegmentum, approximately 2.5 x l o 6 pm2 were sampled across the five animals, wit.h the goal of categorizing LENK+ terminals encountered and photographed according to size, target, type of synaptic contact, and inclusions. To determine specifically the relative abundance of
411
ENK INPZTT TO NIGRAL DA NEURONS IN PIGEONS
meaningful and interpretable data. Note that the total area we sampled greatly exceeds that in other comparable studies in which the investigators chose to perform more limited samplings in each of several animals (Sesack and Pickel, 1992).It is our contention that regional variation in the nigra is as important a concern (and perhaps a greater concern) than interanimal variability. Thus, we believe it is more unambiguous to sample more extensively in a smaller number of animals whose histology is optimal if' the goal is to ascertain the ultrastructural characteristics of a given brain region. To test the statistical significance of differences in the size frequency distribution of LENK+ terminals between different DA and non-DA targets or between normal and deafferented nigra, a Kolmogorov-Smirnov two-sample test at a significance level of P < 0.05 was used (Siege1 and Castellan, 1988).To test the statistical significance of differences in the relative frequency of contact with different categories of postsynaptic target in normal nigra or between normal and deafferented nigra, a chi-square test of independence with multiple comparisons at a significance level of P < 0.05 was used (Miller, 1980).
Fig. 1. Schematic drawing of a transverse section through onc side of the midbrain uf a pigeon shows the area of the substantia nigra that was examined at the electron microscopic level in this study (surrounded by dashed lines). Each dot represent one TH+ perikaryon. Thc midline is shown at the left and dorsal is to the top. SN, substantia nigra: TeO. optic tectum; A V T , ventral tegmental area.
RESULTS Light microscopic observations Many cell bodies stained for TH were observed in pigeon substantia nigra (Fig. 2A), as well as more medially in the ventral tegmental area (AVT),by using PAP single-labeling methods. Our SIG method also yielded extensive labeling of perikarya and proximal and distal dendrites of TH+ neurons. Analysis of adjacent sections stained immunohistochemically for LENK and TH revealed the presence of numerous LENK+ fibers and varicosities in the substantia nigra, as well as in AVT (Fig. 2B,C). The abundance and distribution of LENK+ fibers and terminals in the t.egmenturn was highly similar in all four normal tegmenta examined. On the side of the brain with the unilateral transection of the basal forebrain output bundle, the abundance of LENK+ fibers and varicosities was dramatically decreased in the ipsilateral substantia nigra and AVT (Fig. 2C,D). The extent of LENK+ fiber depletion at the LM level was very similar in all five forebrain deafferented tegmenta examined, and in all cases the residual abundance of LENK+ fibers was less than half of that on the normal side.
the different types of LENK+ terminals per unit area of normal tegmentum, we systematically sampled a defined area of 80,000 pm2 on the normal side of the forebrain deafferented bird. In this sampled region, we counted all LENK+ terminals encountered and used these data as our base frequencies for LENK+ terminals on TH+ structures and LENK+ terminals on non-TH structures per unit area in the normal nigra. The EM data presented in this paper for the forebrain deafferented tegmentum are also based on examinations of midrostrocaudal levels of the pigeon substantia nigra, as shown in Figure 1.Sections from the deafferented tegmentum in the one unilaterally deafferented animal were scanned with the electron microscope. LENK+ terminals Electron microscopic observations contacting either TH+ or non-TH structures were photographed when encountered. In many cases, more than one LENK+ terminals in normal mbstantia nigra. LENK+ terminal was present in the photograph. These photo- terminals were abundant in the substantia nigra of pigeon, graphs were then analyzed and LENK+ terminals catego- with an average of 12.08 LENK+ terminals per 1,000 pm2. rized as for the normal tegmentum. To obtain the data on LENK+ terminals contained densely packed, round or the deafTerented tegmentum, approximately 7.1 x lo5pm2 slightly flattened clear vesicles of various sizes (25-55 nm were sampled, with the goal of categorizing LENK+ termi- in diameter). In many cases, LENK+ terminals contained nals according to size, target, synaptic type, and inclusions. one or more dense core vesicles (95-110 nm in diameter). To estimate specifically the relative abundance of the LENK+ terminals were observed to contact TH+ as well as different types of LENK+ terminals per unit area of non-TH neurons (dendrites and perikarya) in the substandeafferented tegmentum, we systematically sampled a de- t,ia nigra (Figs. 3-61, The frequency of LENK+ terminals fined area of 55,000 +m2 on the deafferented side of the contacting TH+ neurons (5.81contacts per 1,000 pm2)was forebrain-deafferented bird. In this sampled region, we similar to the frequency with which they contacted non-TH counted all LENK+ terminals encountered and used these neurons (6.27 contacts per 1,000 pm21. We noted some data as our base frequencies for LENK+ terminals on TH + tendency for LENK+ terminals to be very abundant on structures and LENK+ terminals on non-TH structures some TH+ perikarya and proximal dendrites and rare or per unit area in deafferented nigra. absent on others. Many of the contacts between LENK+ In our study of normal and forebrain deafferented tegmen- terminals and dendrites or perikarya displayed prominent turn, we chose to sample a large area in our best cases thickened, parallel membranes, with only a slight prebecause we believed that approach best for obtaining andlor postsynaptic density (Figs. 3B; 4B,C, 5A, 6A,B).
412
L. MEDINA ET AL.
Fig. 2. Light microscopic photographs show transverse sections through the midbrain of a medial forebrain bundleiansa lenticularis (FPMIALI hemitransected pigeon that had been irnmunohistochemically labeled with DAB for tyrosine hydroxylase (TH; A) or Ieucinet bodies (A1and IJENK+ enkephalin (LENK; E D ) . Numerous T H ~ cell fihers and terminals (B) are present in thc normal substantia nigra contralateral t o the transection (this side ofthe brain was marked with apin hole; asteriskin thephotographs). Some oftheTH+ perikaryaare noted by small arrows in A tbut the vast majority are not).In contrast to
B, note the lower abundance of LENK+ fibers and terminals in the forebrain deafferented substantia nigra [D). The comparison is most clearly rendered in C, which shows both normal and forebraindeafferented substantia nigra (arrow).Medial is to the right in A and B, and to the left in D. Dorsal is to the top in all four photomicrographs. Asterisks in A-C mark a pinhole placed in thc right tegmentum for orientation. EW, Edinger-Westphal nucleus; IP, interpeduncular nucleus; SN, substantia nigra; TeO, optic tectum; AVT, ventral tegmental area. Scale bars = 200 Frn in A, €3, and D, 1mm in C.
These contacts were therefore considered to be symmetric synapses. Much less frequently,junctions between LENK+ terminals and dendrites (but not perikarya) displayed parallel membranes and a pronounced postsynaptic density, which was in some very rare cases adjoined by postjunctional dense bodies. These junctions were considered to be asymmetric synaptic contacts (Figs. 3C, 6C). The LENK+ terminals observed in the pigeon substantia nigra were typically somewhat rounded or oval in shape, and the size of those we could reliably measure (n = 287) ranged from 0.1 to 1.5 pm (Fig. 7A). We separately examined the size distribution of LENK+ terminals on TH+ and non-TH
dendrites. We found that the distribution of LENK+ terminals on both types of structures was approximately bell shaped, but the distributions were significantly different (as assessed by a Kolmogorov-Smirnov two-sample test). The LENK+ terminals on the TII+ structures were typically smaller (mean & S.E.M. = 0.488 i 0.019 pm) than those on non-TH structures (mean 5 S.E.M. = 0.744 2 0.028 pm). In the following sections, we will separately detail the morphology and synaptology of LENK+ terminals contacting TH+ and non-TH nigral neurons. LENK + terminals contacting T H + neurons. Among the LENK+ terminals whose target structures could be
Fig. 3. A Electron microscopic (EM) photoiiiicrographs show LENK+ terminals (tlLt7) in apposition with a large TH+ dendrite (d) in the normal pigeon substantia nigra. B,C: EM photomicrographs show LENK+ tcrminals (t)making symmetric (B) or asymmetric (C) synaptic contacts with TH+ dendrites (d) in the normal suhstantia nigra. Arrows indicat,e the sites o f synaptic contact. Scale bars = 1 pm in A, 200 nm in B and C.
4 14
Fig. 4. A-C: EM photomicrographs showLENK+ terminals (tl, t4, t5) in apposition with a TH+ perikaryon (p) and its proximal dendrite in the normal substantid nigra. Details of these contacts are shown in B (proximal dendrite) and C (perikaryon). Other LENK+ terminals (t2,
L. -MEDINA ET AL.
t3) are observed in apposition with non-TH dendrites idl, d2 in .4). Arrows in B and C indicate sites of symmetric synaptic contact. Arrowheads in B indicate dense core vesicle. Scale bars = 1 (*m in A, 200 nm in B and C.
ENK INPUT TO NIGRAL DA NEURONS IN PIGEONS
415
Fig. 5. A: EM photomicrograph shows LENK+ terminals (tlLt6'1in apposition with a TH+ perikaryon in the normal substantia nigra of a pigeon. B,C:EM photomicrographs show LENK+ terminals (tl, t2) in apposition with TH+ distal dendrites (dl, dX) in the normal substantia
nigra. A detail uf B is shown in C . Arrows in A and C indicate sites of symmetric synaptic contact. Scale bars = 500 nm in A, 1 km in R, 200 nm in C.
416
Fig. 6 . A EM photomicrograph shows a L E N K i terminal (t2) and a nonlabcled terminal i t l ) contacting a non-TH dendrite ( d l ) in the normal substantia nigra. B,C: EM photomicrographs show LENK+ terminals (t) making symmetric (B) or asymmelric (C) synaptic contact with non-TH dendrites cd) in the normal substantia nigra. Arrows in A,
L. MEDINA ET AL.
€3 and C indicate the sites of synaptic contact. Arrowheads in C indicate dense cnre vesicles. The very fine punctate particles in A and B are sodium phosphate precipitate. Scale bars = 200 nm in h and B, 500 nm in C.
ENK INPUT TO NIGItAL DA NEURONS IN PIGEONS
A
INTACT NlGRA 50
-
1
terminal
B
THneurons non-TH neurans
size
DEAFFERENTED NlGRA 50
-m m
.-K
40
-
1
i
----f
THneurons non-TH nwurons
30Q)
e r
0
C
zol terminal
size
Fig. 7 . Size frequency histograms for the LENK+ tcrminals contact ing TH+ targets or non-TH targets in the normal substantia nigra (A) and in the forebrain deafferented nigra (B). The LENK+ terminals ending on TH+ targets tend to be smaller than those ending on non-TH targets in normal nigra. In contrast, the LENK+ terminals on TH+ targets remaining after removal of forebrain input are indistinguishable in size from the LENK+ terminals ending on non-TH targets that remain after forebrain deafferentation. The abundance of LENK+ terminals on non-TH targets in A4has been corrected for the somewhat lesser total area over which these terminals were collected compared with LENK+ terminals on TH+ structures.
417 identified and measured (n = 267), LENK+ terminals were relatively equally observed to be in apposition with medium TH+ dendrites (36.3%, of LENK+ terminals on TH+ neurons), large TH+ dendrites (22.5% of LENK+ terminals on TH+ neurons), and TH+ perikarya (27.3% of LENK+ terminals on TH+ neurons; Figs. 3-5A, 8A). A chi-square test of independence with multiple comparisons showed that, whereas the abundance of LENK+ terminals on medium TH+ dendrites and TH+ perikarya was not significantly different, LENK+ terminals on medium TH+ dendrites were significantly more abundant than LENK+ terminals on large TH+ dendrites. Less frequently, LENK+ terminals were observed in apposition with small TH+ dendrites (13.9% of LENK+ terminals on TH+ neurons; Fig. 5B,C). This frequency was significantly less than that of LENK+ terminals on either medium TH+ dendrites, large TH+ dendrites, or TH+ perikarya. Although the LENK+ terminals contacting TH + neurons ranged from 0.1 to 1.5 pm in diameter, 89.4% of the terminals whose size could reliably be measured (n = 198) were smaller than 0.81 Fm. The size distribution of the LENK+ terminals showed a peak (possibly a bimodal peak) in the 0.25-0.55 km range (Fig. 7A). Not surprisingly, the mean size for all LENK+ terminals on TH+ structures was near this peak, i.e., 0.488 5 0.019 Fm. Of the LENK+ terminals observed to make apposition with a TH+ target (,n = 267), 57.3% were observed to form synaptic junctions with their TH+ target structures (Figs. 3B,C, 4B,C, 5A, 8B). Of those LENK+ terminals making synaptic contact with a TH+ target, 92.3% made symmetric contacts, and the remainder made asymmetric contacts (Figs. 3C, 8B). Asymmetric contacts were never observed, however, on large dendrites and perikarya (Fig. 8B). However, because our sample of LENK+ terminals making asymmetric synapses was very small (representing less than 7% of the terminals forming synaptic contacts): we cannot rule out the possibility that LENK+ terminals may occasionally form asymmetric contacts on large dendrites and perikarya of TH+ nigral neurons. Furthermore, LENK+ terminals less than 0.4 pm or larger than 0.75 pm in size were never observed to form asymmetric synaptic junctions with TH+ targets. Examination of the morphology of LP:NK+ terminals in the substantia nigra that contacted TH+ neurons iirrespective of terminal size or type of synaptic contact) revealed that the terminals varied in their abundance of dense core vesicles and mitochondria. To help us systematically determine whether or not and in what ways (1) LENK+ terminals ending on TH+ structures differed from those ending on non-TH structures and (2) LENK+ terminals of basal forebrain origin differed from those of nonbasal forebrain origin, we subdivided LENK+ terminals into four types based on their dense core vesicle content (poor vs. rich) and on mitochondrial abundance (poor vs. rich; Table 1). Vesicle type and mitochondrial abundance have been used to describe and classify terminals in various brain regions (Mori et al., 1987; Bolam and Smith, 1990, 1992; Liang et al., 1993; Meredith et al., 1993). Terminals that were considered rich in dense core vesicles (DCR) contained two or more such labeled dense core vesicles, and terminals regarded as poor in dense core vesicles (DCP) contained none. We operationally defined poor in mitochondria (MP) as one or none, and rich in mitochondria (MR) as two or more. In general, the mitochondria-poor terminals contained none, and the rich contained three or more. The
L. MEDINA ET AL.
418
A
RELATIVE FREQUENCY OF LENK+ APPOSITIONS ON DIFFERENT TH+ STRUCTURES IN INTACT NlGRA
DENDRITES L 0.75pm
B
DENDRITES 0.76-1.5am
DENDRITES ~1.5pm
PERIKARYA
RELATIVE FREQUENCY OF SYMMETRIC AND ASYMMETRIC SYNAPTIC CONTACTS OF LENK+ TERMINALS ON DIFFERENT TH+ STRUCTURES IN INTACT NIGRA symmetric synapses
88
asymmetric synapses
69 9% C
UI
.._
60
In
0
a
-
-
40
5
0
m
P
e
20
5
e
n 0
DENDRITES S 0 . 7 5 ~
DENDRITES 0.76-1.5pm
DENDRITES z 1.5pn
PERIKARYA
Fig. 8. A Relative frequency of LENK+ terminals observed in apposition with different types of TH 1~ structures (small,medium, and large dendrites, and perikarya) in the intact substantia nigra of the pigeon, each expressed as a percentage of total LENK+ appositions observed across all TH+ target structures in normal nigra. €3: Fre-
quency of symmetric and asymmetric synaptic contacts made by LENK+ terminals on the four different types of TH+ structures in the intact substantia nigra, expressed as a percentage of all LENK+ appositions observed on that type of TH+ target structure.
relative frequency of these four terminal types was assessed in a sample of 79 LENK+ terminals ending on TH+ structures in the normal nigra of one animal. LENK+ terminals that were poor in dense core vesicles and mitochondria were not significantly different in abundance (35.4% of all LENK+ terminals contacting TH+ structures) from those that were poor in dense core vesicles but rich in mitochondria (39.2% of all LENK+ terminals on TH+ structures; Figs. 3B,C, 5A) by a chi-square test of independence with multiple comparisons. LENK+ terminals contacting TH+ neurons that were rich in labeled dense core vesicles but poor in mitochondria (Fig. 4A,B) made up 7.6% of all LENK+ terminals on TH+ structures. Finally, LENK+ terminals contacting TH+ neurons that were rich in labeled dense core vesicles and rich in mitochon-
dria (Fig. 4A,B) made up 17.7% of all LENK+ terminals on TH+ structures. By chi-square test of independence, the DCR/MP terminals were significantly less abundant than the DCRiMR terminals, and bolh were less abundant than the two types poor in dense core vesicles. Terminals forming asymmetric synapses usually contained only clear vesicles and were rich in mitochondria. Finally, we saw no clear evidence that any of the four types of LENK+ terminals contacting TH+ structures was preferentially of a specific size o r preferentially contacted a specific part of the TH+ neurons (perikarya or large, medium, or small dendrite). LENK + terminals contacting non-TH neurons. LENK+ terminals contacting non-TH structures were most frequently in apposition with medium non-TH dendrites, and slightly less frequently in contact with small and large
ENK INPUT TO NIGRAL L)A NEURONS IN PIGEONS
419
TABLE 1. Relative Frequency of Four Different Types of LENK+ Terminals, as Defined by Dense Core Vesicle and Mitochondria Content, on TH+ nr Non-TH Targets in Normal and Deaffcrcntcd Substantia Nigra (SN)in Pieeon’
all LENK+ terminals contacting non-TH structures) to those that were poor in dense core vesicles but rich in mit.ochondria (23.1% of all LENK+ terminals on non-TH structures; Fig. 6A; Table 1). By a chi-square test of independence with multiple comparisons, this difference between DCPiMP terminals and DCPiMR terminals is not significant at t h e P < 0.05 level. LENK+ terminals contacting non-TH neurons that were rich in labeled dense core vesicles but poor in mitochondria (Fig. 6C) made up 13.8% of all LENK+ terminals on non-TH structures. Finally, LENK+ terminals contacting., non-TH neurons that were rich in labeled dense core vesicles and rich in mitochondria made up 33.8% of all LENK+ terminals on non-TH structures. B~ chi-square test of independence, the DCR/MP terminals were significantly less abundant than the other three terminal types, and the DCR/MR terminals were significantly more abundant than the other three types. LENK+ terminals forming asymmetric synapses with nun-TH targets were typically rich in mitochondria. We saw no clear epjdence that any of the four types of LENK+ terminals contacting non-TH structures was preferentially ofa specific size or preferentially contacted a specific part of the non-TH neurons (perikarya or large, medium, or small dendrite). LENK+ terminals in basal forebrain-deafferentedsubstantia nigra. In the forebrain-deafferented substantia nigra, a mean of 3.47 LENK+ terminals was observed per 1,000 pm2 of nigra examined. By contrast, at a similar level of normal substantia nigra, there was an average of 12.08 LENK+ terminals per 1,000 pm2 of nigral area examined, as noted earlier. Thus, the number of LENK+ terminals in the basal forebrain-deafferented substantia nigra was reduced to 28.7%;of its normal level, indicating that about three-fourths of the LENK+ terminals in the pigeon substantia nigra arise from the basal telencephalon. At the EM level, LENK+ terminals in the deafferented substantia nigra contained densely packed clear vesicles of various sizes (40-60 nm in diameter; Figs. 10, 11).Most LENK+ terminals also contained a few dense core vesicles (90-120 nm in diameter). LENK+ terminals were observed to contact both TH+ and non-TH neurons in the deafferented substantia nigra (Figs. 10, 111, although the frequency was only 1.598 contacts on TH+ neurons per 1,000 pm” and 1.872 contacts on non-TH neurons per 1,000 pm2. As observed in the intact substantia nigra, many contacts between LENK+ terminals and their targets in the deaffercnted substantia nigra displayed prominent, thickened, parallel membranes with only a slight presynaptic and/or postsynaptic density, which is characteristic of a symmetric synaptic junction (Figs. 10-13). Less often, junctions between LENK+ terminals and dendrites (but not perikarya) in the deafferented substantia nigra displayed parallel membranes and a pronounced postsynaptic density. LENK+ terminals contacting TH+ and non-TH neurons ranged between 0.3 and 1.6 pm in diameter, but the majority were larger than 0.75 km (Fig. 7B). LENK + terminals contacting T H + neurons. Among all LENK+ terminals whose postsynaptic contact could be identified in the deafferented nigra (n = 411, LENK+ terminals in the basal forebrain-deafferented nigra were most frequently observed in apposition with medium-sized TH+ dendrites (46.3% of LENK+ terminals on TH+ neurons), and somewhat less frequently with large TH+ dendrites (29.3% of the LENK+ terminals on TH+ neurons; Figs. lOB,C, 11A, 12A). Even less frequently, LENK+ terminals
Y
TH+ targets
T~mu of turminal
w
% Nor.nd
Deafferented
SN
SV
Non-TH targets %, Normal DeaffCrented %
SN
SN
~~
Dense core poor and mitorhondriapmr Dense core poor and mituchondriarich Dense core rich and mitochondriapour Dense core rich and mituchondriarich
35.4 128.’79)
8.3 12,‘24)
29.2 l19!65)
3 6 kli28.1
39.2 131!79)
33.3 18/24)
23.1 115/65l
39 3 iii:28)
12
L3,8(9/65)
45 8 (11:241
33.8 122i1351
7.6 ,8!79)
17.7 114,591
(2,28,
50 fl .14:281
’In this table, terminals are divided into four types based on their coitGit of dense cure vesicles (rich vs. poor) and mitochondrin lrich vs. poor). See text fur further details ofthis classification o f terminals. Percentages with a column sum to 1005{ hecaiiw the relative frequencies are separately expressed per each target t.ype in intact versub deafferented n15a. The results shown indicate t.hat the relative frequency uf the terminal types differs between TIT+ and non-TH s t r u c t u r a in Intact i u g a and t h a t some terminal types are altered in their frequency in the nigra deafferented of basal telencephdic input.
non-TH dendrites (Fig. 9A). For example, 37.9% of LENK+ terminals contacting non-TH structures were in apposition with medium-sized dendrites, whereas 27.6% of LENK+ terminals contacting non-TH structures were in apposition with small dendrites and 27.6% were in apposition with large dendrites (Figs. 6, 9A). By a chi-square test of independence, the abundance of LENK+ terminals was significantly greater on the medium than the small and large dendrites. Finally, LENK+ terminal contacts on non-TH perikarya (6.9% of all LENK+ terminals contacting non-TH structures) were significantly less frequent than on dendrites of non-TH structures. The LENK+ terminals contacting non-TH neurons that could reliably be measured (n 89) ranged in size from 0.2 to 1.5 pm in diameter (Fig. 7A). Nevertheless, 93.3% of these LENK+ terminals were larger than 0.40 pm. The size distribution of the LENK+ terminals showed a prominent peak a t 0.75 pm and a lesser peak at 0.45 pm. Not surprisingly, the mean size for all LENK+ terminals on non-TH structures was at the major peak, i.e., 0.744 f: 0.028 pm. Of the LENK+ terminals observed to make apposition with a non-TH target (n = 87), 69.0% were observed to form synaptic junctions with their non-TH target structures (Figs. 6A-C, 9B). Of those LENK+ terminals making synaptic contact with non-TH structures, 88.3%made symmetric synapses (Fig. 6A,B), and the remainder made asymmetric synapses (Fig. 6C). LENK+ terminals were not observed to make asymmetric synapses on non-TH perikarya, and extremely few were observed to make asymmetric synapses on large non-TH dendrites. Because our sample of LENK+ terminals making asymmetric synapses was very small (representing less than 12% of the terminals forming synaptic contacts), we cannot rule out the possibility that LENK+ terminals may occasionally form asymmetric contacts on perikarya of non-TH nigral neurons. Similarly, LENK+ terminals that measured less than 0.4 pm were never observed to form asymmetric synapses with non-TH neurons. The relative frequency of LENK+ terminals on non-TH structures as classified by dense core vesicle and mitochondrial content was assessed in a sample of 65 terminals from the normal nigra of one animal (Table 1).LENK+ terminals that were poor in dense core vesicles and were poor in mitochondria were relatively equal in abundance (29.2%of
~~
L. MEDINA ET AL.
420
A
RELATIVE FREQUENCY OF LENK+ APPOSITIONS ON DIFFERENT NON-TH STRUCTURES IN INTACT NlGRA
501 n
-
0 .-.-n g n
-I I
L
0
m
L1I L
c
E
P
DENDRITES S 0.75 prn
DENDRITES 0.76-1.5 prn
DENDRITES > 1.5 Irrn
PERIKARYA
RELATIVE FREQUENCY OF SYMMETRIC AND ASYMMETRIC SYNAPTIC CONTACTS OF LENK+ TERMINALS ON DIFFERENT NON-TH STRUCTURES IN INTACT NIGRA !OO
7
DENDRITES 5 0.75 prn
DENDRITES 0.76-1.5 pm
DENDRITES > 1.5 pm
W
symmetric synapses
H
asymmetric synapses
PERIKARYA
Fig. 9. A Relative frequenq of LENK+ terminals observed in apposition with different types of non-TH structures (small, medium, and large dendrites, and perikarya) in the intact suhstantia nigra of the pigeon, each expressed as a percentage of total LENK+ appositions observed across all non-1'H target structures. B: Frequency of symmet-
ric and asymmetric synaptic contacts made by LENK+ terminals on the four different types of non-TH structures in the intact substantia nigra, cxpresscd as a percentage of all LENK+ appositions observed on that type of non-TH target structure.
were observed in apposition with small TH+ dendrites (19.5% of the LENK+ terminals on TH t neurons; Fig. 12A). Finally, the abundance of LENK+ terminals on TI
