Efferent projections of nucleus locus coeruleus: Morphologic subpopulations have different efferent targets

0306-4522/86$3.08+ 0.00 Pergamon Journals Ltd 0 1986IBRO

Neuroscience Vol. 18, No. 2, pp. 307-319, 1986

Printed in Great Britain

EFFERENT PROJECTIONS OF NUCLEUS LOCUS COERULEUS: MORPHOLOGIC SUBPOPULATIONS HAVE DIFFERENT EFFERENT TARGETS S. E. LOUGHLIN, S. L. Foora* and R. GRZANNA~ Department of Pharmacology, University of California, Irvine, California, *Department of Psychiatry, University of California, San Diego, California and TDepartment of Cell Biology and Anatomy, Johns Hopkins University School of Medicine, Baltimore, Maryland, U.S.A. Abstract--This study quantitatively addresses the hypothesis that there is a systematic relationship between the morphologic characteristics of locus coeruleus neurons and the particular target regions they innervate. Following horseradish peroxidase injections into selected terminal fields, locus coeruleus cell bodies are heavily labeled by retrograde transport so that somata size and shape, and in many cases primary dendritic pattern can be observed. This allows the classification of neurons as one of six cell types: large multipolar cells within ventral locus coeruleus, large multipolar cells in the anterior pole of locus coeruleus, fusiform cells in dorsal LC, posterior pole cells, medium-sized multipolar cells (termed core cells in this report), and small round cells. It was found that while core cells contribute to the innervation of all terminal fields examined, other cell types project to more restricted sets of targets. The contributions of each type to selected efferents are presented in detail. In particular, fusiform cells project to hippocampus and cortex, large multipolar cells in ventral locus coeruleus project to spinal cord and cerebellum, and small round cells in central and anterior locus coeruleus, as well as large multipolar cells in anterior locus coemleus, project to hypothalamus. These results, in conjunction with those described in the preceding report, indicate that locus coeruleus is intrinsically organized with respect to efferent projections with much more specificity than has previously been evident. This high degree of organization is consistent with other recent demonstrations of functional specificity exhibited by locus coeruleus neurons.

described in the preceding” and present reports are an integrated analysis of the intrinsic organization of the locus coeruleus (LC) with respect to its efferent projections. The preceding report demonstrated that LC neurons projecting to particular target regions exhibit characteristic, restricted spatial distributions within LC. ” Since others have shown that there are morphologic subtypes of LC neurons which show preferential spatial distributions within LC,12.*3*3’ these results would lead to the prediction that particular morphologic subtypes might also preferentially innervate certain terminal regions. The present study examines morphologic subpopulations within LC in terms of the contribution of each cell type to the same efferent projections examined in the preceding report.” In the present study, horseradish peroxidase (HRP) injections into each of six different LC terminal fields, and reaction of tissue with tetramethylbenzidine, revealed many cells throughout LC in which dense retrogradely transported label filled the soma and extended into proximal dendrites. This permitted accurate morphologic classification of these neurons, especially since only a fraction of cells were labeled and overlapping profiles seldom obscured each other. Since this method labels only those cells (or a subset of those cells) projecting to specific terminal fields, The studies

Abbreviations: LC, locus coeruleus; HRP, horseradish peroxidase; DBH, dopamine b-hydroxylase.

only the subpopulation of LC neurons contributing to this projection was visible in each preparation. Thus, correlation of LC neuronal morphology with efferent projection targets was possible. Counterstaining of tissue with Nissl stain, as well as careful comparison with analogous sections prepared by staining for the norimmunohistochemical dopamine synthetic epinephrine enzyme /I-hydroxylase (DBH), allowed positive identification of certain LC neuronal subtypes as norepinephrinecontaining. Previous reports addressing the issue of subtype/target relationships are few’4,‘7*19*26 and describe only the predominant morphologic type of LC neuron labeled following various LC terminal field injections. By classifying every labeled LC cell following HRP injections into selected terminal fields, we were able to quantitatively address the following questions: does a relationship exist between the morphology of LC cells and their efferent targets? If so, does one cell type uniquely project to certain terminal field(s), or does a more complex relationship exist? This analysis has provided new information in two ways: (1) it has revealed details of LC cell morphology and has allowed more accurate determination of the distributions of various cell types within LC, and (2) relationships between the morphology of cells and their efferent target have been characterized and quantitatively analyzed. These findings, in conjunction with the previous examination of the spatial distributions of cells with 307

308

S. E. LOUGHLINet ai.

particular efferents, are utilized in the present report to generate a model of LC intrinsic organization with respect to efferent projections. EXPERIME~AL

PROCEDURES

Horseradishperuxidase methods HRP injections and subsequent tissue preparation were conducted, as described in the preceding report.” Briefly, 190-210 g male Sprague-Dawley rats were injected stereotaxically with 0.05~1 of a 30% solution of HRP in Tris buffer, allowed to survive 24-48 h, and perfused with 1% ~rafo~d~yde and 1.25% ~uter~dehyde in phosphate buffer (pII = 7.4) according to Rosene and Mesularr~.~~ Brains were removed, stored in 15% sucrose for 3 h, and 40hm sections were prepared with a freezing microtome. Serial sections through LC and at least every fifth section through the injection site were reacted with tetramethylbenzidine according to Mesulam.” Sections were mounted on subbed slides and counterstained with neutral red. This procedure produced a dense blueblack reaction HRP reaction product which was readily visible under bright-field illumination, and the counterstain allowed simultaneous detection of unlabeled LC cells. The stereotaxic coordinates of injections and diagrams of injection sites are presented in the preceding paper (Ref. 17; Table 1 and Fig. 1). Injections were centered in frontal cortex, dorsal and ventral hippocampus, h~thalamus, cerebellum and spinal cord. Dopamine @-hydroxytasemethods Brains were reacted by the method of Grzanna and colleagues”.‘Z~‘3 utilizing unlabeled antiserum directed against the rat form of the norepinephrine synthetic enzyme DBH. Although DBH is also present in epinephrinecontaining neurons, no phenylethanol~ine N-methyl transferase, the further synthetic enzyme for epinephrlne, has been reported in LC cells. Sagittal 20pm cryostat sections from seven paraformaldehyde perfused brains were washed and incubated with DBH antibody (1: 1000, provided by R. Grzanna) in phosphate-buffered saline for 24 h. Sections were again washed and stained by the immunoperoxidase method. This reaction has been shown to be an extremely sensitive technique for the demonstration of no~pi~h~ne~n~ining ceils and fibers. Analysis For each injection site, five replications were selected and analyzed. Labeled LC cells within each brain were individually rated for density of reaction product and classified as to apparent morphology by two investigators (SEL, SLF). Other reports have described several morphologic cell types in LC including large multipolar cells in ventral LC, fusiform cells in dorsal LC, and large multipolar cells anterior to LC.‘L’3.3i Following extensive examination of HRPlabeled LC neurons, as well as cresyl violet and DBHstained material, LC cells were further subdivided into seven categories based on morphology and distribution within LC. These are described in the Results section of the present report. The number of labeled cells in each category in each LC was computed.

as a result of this procedure: all significant differences between groups still obtained and all nonsignificant differences remained the same. Therefore. all results presented below apply to both manipulations.

The totals for brains receiving

HRP injcc-

tions in each terminal field were compared by one-way analysis of variance. Since variability inherent in HRP histochemistry produces sizeable within-group variance in the total number of cells labeled, and since it was desirable to analyze the relative contributions of each cell type to the populations of neurons projecting to each ternhal field, the

number of labeled cells contributed by each cell type was also computed as a percentage of the total number of labeled LC cells for each brain. These percentages were then compared across injection sites by analysis of variance. All percentages were calculated by both including and excluding the unclassified cells in the totals. No differences occurred

RESULTS

Morphology of locus coeruleus cells labeled by horseradish peroxidase

The following morphologic categories of cells were observed in LC: (1) Large multipolar cells within ventral LC (Fig. 1D): these cells were greater than 25 in diameter, with neither dimension 2fim significantly greater than the other. At least three dendrites arose from the soma and there was no obvious orientation of the cell body along any dimension. These cells tended to be concentrated ventromedially in LC. (2) Large multipolar cells in the anterior pole of LC (Fig. 1C): c&s of similar morphology to (1) which were located outside the boundaries of LC but which corresponded in appearance and location to cells stained for DBH (Fig. 2). These cells were often somewhat elongated with their long axis oriented from anteroventral to posterodorsal. (3) Fusiform cells (Fig. 1B): cells having a length-towidth ratio of at least 2: 1, with the width being at least 12 pm. Virtually no cytoplasm surrounded the nucleus in the short axis, while Nissl substance in the long axis was very dense. A prominent nucleolus of approximately 2 pm was visible in Nissl-stained material. These cells were strongly oriented from anterior to posterior. (4) Posterior pole cells (Fig. 1A): fusiform cells at least 30 pm distant to LC within the superior cerebellar peduncle. These often exhibited an axial ratio of at least 3: 1. (5) Core cells (Fig. 1B): this group consisted of multipolar cells ranging in size from 15 to 20 pm in diameter and having at least three visible dendrites arising from the soma. A large round nucleus contained a prominent nucleolus, and soma shape ranged from slightly elliptical to quite round. These were distributed throughout the main body of LC as well as more sparsely in ventral LC. (6) Small, round cells (Fig. 1C): cells of less than 10pm in diameter which appeared quite round in sagittal sections (Fig. 5). A very pale nucleus of approximately 6 pm contained a 1 pm nucleolus. A very dense, very narrow rim of cytoplasm surrounded the nucleus and several thin, radiating dendrites were visible in HRP-labeled profiles. (7) Unclassified: cells which contained reaction product, but were not labeled heavily enough to permit classiiication. Distributions of morphologic ceN types within locus coeruleus Both DBH and HRP material were utilized to characterize the distributions of these cell types within LC. These results are depicted in Fig. 3 in which the distributions are superimposed (in gray) on a reconstruction (in black) of all LC cells viewed from a sagittal

perspective.

Large

multipolar

cells were

Fig. 1. LC cell types labeled by various HRP injections. (A) A posterior pole cell, embedded within the fibers of the superior cerebellar peduncle, is indicated by the arrow. (B) The small arrow indicates a fusiform cell. The large arrow indicates a medium-sized multipolar cell. (C) The large arrow indicates a large multipolar cell anterior to LC. The small arrows indicate small, round cells in this region. (D) The arrow indicates a large multipolar cell in ventromedial LC. Calibration bar in (A) = 50 km for all panels. Fig. 2. Photomicrographs of sagittal sections showing large multipolar cells in the anterior extension of LC: (A) tissue processed by immunohistochemist~ for DBH, and (B) tissue processed to demonstrate retrogradely transported HRP follo~ng injection into hy~thalamus. Anterior is to the right. Calibration bar = 100 pm. Fig. 3. Distributions of LC cell types. The same sagittally oriented computer reconstruction of the spatial distribution of LC cells is shown in each panel. The gray dots indicate the regions in which each cell type was observed. (1) Large multipolar cells in ventral LC. (2) Large multipolar cells anterior to LC. (3) Fusiform cells. (4) Posterior pole cells. (5) Core cells. (6) Small, round cells. Fig. 5. Small, round cells in LC. (1) Small, round cells are indicated by arrowheads in photo~~ro~aphs of HRP-labeled LCs in lA,lB and IC. These. are all 40 ~1 sag&al sections. Calibration bar = 20 pm. (2) Paraffin-embedded LCs were sectioned in sag&al (A), coronal (B) and horizontal (C) planes and stained with Cresyl Violet. Arrowheads indicate small, round cells. (3) Sag&al sections through LC were processed for DBH immunocytochemistry. Profiles of similar size, shape and distribution to the small, round cells observed in LC are indicated (arrowheads). These may, however, be cell fragments.

309

311

312

313

315

Locus coeruleus cell classes: different efferents

g

E

IOOJ ‘Jt

CORE

100

p 20--

80

E

60

&s 50 s “i’ 40

20.1

-

FUSIFORM

* 19.1

10.3

5 Y D-HPC WPC

*NO

NEOCX

CB

HYP SPCORD

IW ’

D-HPC VHPC

CX

CB

HYP

SP

SIGNIFICANT DIFFERENCES

POSTERIOR

MULTIPOLAR

IOO-

POLE

80-

fi! 60i+

*

as io-

0

a

19.6

-

+

*

40-

2-

1.5 ’

D-HPC V-HPC NEOCX

CB

-

10.6)

1.8 0.0

HYP SPCORD

Quantitative analyses of locus coeruleus cell populations and their efferent projections Data on the frequency of HRP labeling of each cell type after the different injections were q~ntitatively analyzed. The percentage each morphologic subtype contributed to the total number of cells labeled after each injection is shown in Fig. 4. Two major points are evident from visual inspection of these histograms. First, following HRP injections into any terminal region, the largest population of labeled neurons consisted of core cells. Second, each cell type exhibits a profile of preferred targets that differs from that shown by any other cell type. Statistical analysis revealed that the number of core cells labeled did not differ significantly as a function of injection site. Wowever, for all other cell types the absolute (P < 0.05) and relative (P < 0.01) number of labeled neurons did vary significantly as a function of injection site. These significant differences between the preferred efferents of these cell groups suggests that these morphologic cell types constitute sub-

populations within LC with different efferent projections. Some subtypes, e.g. fusiform, exhibit substantial specificity in this regard, while other cell types, e.g. core cells, do not. The differences between cell types were substantiated by further statistical analysis ~euma~Keuls one-way analysis of variance) which revealed significant differences between specific pairs of injection sites (see Fig. 4) for multipolar, fusifotm, and posterior pole cell types. Both spinal cord and cerebeliar injections labeled a greater percentage of multipolar cells within the LC than did injections into hypothalamus (P < 0.01) and ventral hippocampus (P c 0.05). Injections into frontal cortex also labeled a significantly greater percentage of these cells than did injections into hypothalamus (P < 0.05). In contrast to the prominent labeling of the large multipolar cells anterior to LC following hypothalamic injections, very few large multi~lar cells within LC were labeled by these injections. Telencephalic injections (dorsal hippocampus, ventral hippocampus and neocortex) labeled a higher proportion of fusiform cells than did other injections. Posterior pole cells were labeled in significantly higher proportions only by the hip~c~pal injections. For all cell types, no significant differences were observed between injections into the dorsal or ventral hippocampus. Two cell types were labeled by only one injection. Extremely large multipolar cells located anterior to LC and very small, round ceib were labeled only after

316

s.

E.

~OUGHLlN

hypothalamic injections. Since these small cells have not been described in previous studies on LC efferents, they are described in more detail in the next section. Small, round cells

To determine whether the small HRP-labeled profiles observed following injections into hypothalamus were in fact norepinephrine-containing LC cells, Cresyl Violet-stained, 12 pm sections cut in frontal, horizontal and sagittal planes were examined. In Cresyl Violet-stained LCs, small profiles were observed which were of similar size, general appearance, and location within LC to the small HRP profiles (Fig. 5). These did not appear to be larger LC cells which were sectioned such that only a fragment appeared, as they contained prominent, densely staining, small nucleoli within a clear, b8 pm nucleus. To determine whether they might be sectioned in a plane which was misleading as to the size or shape of the cell, sections in the coronal and horizontal planes were examined. Cells of similar size, appearance and distribution within LC were present in all planes of section. The further question existed as to whether these cells were norepinephrine-containing. In sections reacted for immunocytochemical localization of DBH, immunoreactive profiles were observed which were of similar size, shape and distribution within the nucleus to cells observed in Nissl-stained material (see Fig. 5). DISCUSSION

The purpose of the studies described in the preceding” and present reports was to characterize the relationships between morphology, topography and efferent projections of LC neurons. In the preceding report,” accurate, detailed three-dimensional reconstructions were utilized to analyse the spatial distributions of cells giving rise to specific efferent projection was analyzed quantitatively. These studies have yielded more detailed, comprehensive informalationship between morphologic cell type and efferent projection was analysed quantitatively. These studies have yielded more detailed, comprehensive information than was previously available concerning the spatial distribution of morphologic cell types in LC and the preferred efferent targets of each cell type. This new information, combined with the preceding spatial analysis, is utilized later in this section to generate a model of LC organization which integrates these three types of information: (1) the spatial distribution of cells projecting to specific targets; (2) the spatial distribution of morphologic cell types, and (3) the morphologic cell types which project to certain targets. Within LC, no one-to-one relationship exists between cell type and efferent target; that is, no one cell type provides the sole innervation to any one terminal field. This is due to the fact that the large population

et ai.

of core cells projects to all terminal fields examined. The present study demonstrates, however, that the efferents of the other morphologic populations within LC are more specialized. The results of the present studies are summarized below in paragraphs dealing with each major morphologic cell type, its location in LC, and its efferent targets. The results are also discussed in terms of previous reports. Morphologic classes of locus coeruleus cells Anterior multipolar cells. The cells we have called anterior multipolar cells are equivalent to those described by Grzanna and Molliver.‘3 These cells stain densely for DBH (see Fig. 2). Of the injections we performed, only those centered within the hypothalamus labeled these cells. The anterior pole may. however, also project to septum. Previous reports have described cells labeled in “anterior LC” by septal injections.“,” Small cells. Small. round cells have been described with in other reports. Z-UO Following treatment 5-hydroxydopamine, similar ceils did not contain small granular vesicles,29 suggesting that they were not norepinephrine-containing. Very few such cells were observed, however. Other investigators have suggested that similar cells in cat may be intemeurons within LC.u’ We have demonstrated that small cells within LC sustain axons projecting to at least one terminal field, the hypothalamus. A recent report concerning LC projections to the supraaptic nucleus of the hypothalamus did not describe similar cell~.‘~ However, in this previous study, only limited numbers of labeled cells were observed in LC, which may be the reason this particular cell type was not observed. These cells may be “displaced” cells of the parabrachial nuclei, since the parabrachial nuclei have ken reported to project to the hypothalamus.*’ However, these nuclei are separated from LC by the cells of the mesencephalic nucleus of the trigeminal nerve and are not in very close proximity to the regions of LC in which these ceils were observed. They were also not labeled by amygdala injections which did label parabrachial cells (Loughlin, unpublished results). Another possibility is that these are central grey cells; small cells in the central grey are reported to project to the lateral hypothalamus.24 Although we observed profiles of similar size, shape and distribution in DBH-stained sections, it is possible that these were cell fragments. Core cells. Medium-sized multipolar core cells have been observed in LC in Nissl-stained tissue, electron microscopy, DBH immunocytochemistry, and Golgi preparations. UM These contribute a major portion of all LC efferent projections examined in this study (see Fig. 4). The particular efferents of individual core cells are predicted by their location in some cases. That is, only dorsal cells project to hippocampus and only ventral cells project to spinal cord. Other efferents arise from all parts of this group.” Lay et

317

Locus coeruleus cell classes: different e&rents ~1.‘~ have reported “large multipolar cells” labeled following injection of HRP into the ventral hippocampus. These are probably the same cells as those we have called core cells, which are also multipolar. These investigators did not examine cases in which any of the larger cells would be labeled, such as following HRP injection into spinal cord. We therefore believe this apparent discrepancy is simply a difference in terminology. Large multipolar cells. Satoh et aI.% were the first to describe very large multipolar cells in ventral LC projecting to spinal cord. Our results confirm this observation. Fusiform andposterior pole cells. Fusiform LC cells have been described in Golgi preparations as well as in Nissl-stained material.~~3’ Fusiform cells along the dorsal edge of LC have been reported to project to hippocampus. I9 In our tissue, these also are labeled by cortical injections, but not by any others. Cells of similar morphology in the posterior pole are labeled by hippocampal injections and occasionally by cortical injections. Due to the fact that the posterior pole cells are only tenuously contiguous with LC proper, this has been considered a separate group.** Since they differ from other LC cells in morphology, Grzanna and Molliver” also suggest that they should be considered a separate group. We propose from our data that they should be considered a continuation of the same subgroup as the dorsally located fusiform cells since posterior pole cells are similar in morphology and efferents. General discussion Subdivisions of locus coeruleus. The principal findings of this and the previous report can be summarized as follows: (1) LC neurons exhibit a distinct three-dimensional spatial distribution which is demonstrably similar in different individuals; (2) subsets of LC neurons projecting to specific targets exhibit different spatial distributions; (3) particular morphologically defined subpopulations exhibit distinctive spatial distributions within LC, and (4) each morphologic class exhibits a different profile of preferred efferent projections. Both types of analysis are important, since neither spatial distributions nor morphologic class alone can adequately characterize LC organization with respect to efferent projections. in the Spatial gradients are most evident dorsal-ventral dimension and differentiate cells of origin for projections that do not differ in terms of cell type. In a complementary fashion, certain morphologic correlations specify organizational principles with regard to efferent projections that are not evident from spatial analysis. When spatial, morphologic and efferent relationships are integrated, a comprehensive organizational scheme results which is diagrammed in Fig. 6. The specific morphologic/ spatial subdivisions of LC depicted are as follows: (A) Posterior pole cells projecting to the hippocampus. (B) The fusiform cells on the dorsal edge of LC

DORSAL

F -

rp cord.

G-

rp cord, cb

H-

sp cord, cb

cb, cx, hyp

Fig. 6. LC subpopulations. A schematic diagram of a sag&al view of LC indicating proposed subdivisions, with the predominant cell type and efferents for each. Dorsal is up and anterior is to the right. cb, cerebellum; cx, cortex; hpc, hippocampus; hyp, hypothalamus; sp cord, spinal cord.

projecting to hippocampus and cortex. (C) The large multipolar cells anterior to LC projecting to the hypothalamus. (D) The small, round cells in the ventral region of LC and just at the border of the anterior pole of LC which contribute to the projection to hypothalamus. (E) Core cells in approximately the dorsal two thirds of LC projecting to neocortex, hippocampus, hypothalamus and cerebellum, but not to the spinal cord. (F) Core cells in the ventral one third of LC innervating the spinal cord and cerebellum and contributing to the innervation of hypothalamus and neocortex, but not the hippocampus. (G) The large multipolar cells at the ventromedial pole of LC innervating the spinal cord and cerebellum. (A)-(C) These correspond to divisions suggested by Grzanna and Molliver” based on tissue processed for DBH. They proposed no subdivision of the compact region of LC, partially because DBH immunocytochemistry stains this region too densely to reveal the morphology of individual neurons. Division G corresponds to the ventral division proposed by Swanson. 31 (B), (D)-(F) These are further subdivisions of “compact” LC. Although these do in fact constitute defineable anatomical subpopulations of LC neurons, the functional implications of such subdivision of the nucleus is not clear. These studies demonstrate heterogeneity within LC with respect to morphology and preferred efferents. What other characteristics of LC neurons are heterogenous and might be systematically related to one or both of these characteristics? One possibility is that these subdivisions are also physiologically differentiated so that their constituent neurons are active in response to different conditions, and such activity is then communicated to specified subsets of LC target areas. However, studies of the electrophysiological activity of LC cells within compact dorsal LC in rat suggest that the cells fire simultaneously. ‘s**’ That is, multiunit recordings indicate that all units recorded at a given time change

318

S. E.

LOUGHLIN

similarly in discharge frequency. Different patterns of activity however, often exhibited by were, norepinephrine-LC cells “situated near an edge of the nucleus”.’ These “edge” cells may well correspond to the morphologic subpopulations exhibiting more specialized efferents which are described in this report. It is therefore possible that the anatomically defined subpopulations constitute functional entities. Another possibility is that the morphologic subdivisions reflect differences in axonal morphology and conduction properties. For example, axons arising from raphe neurons can be either myelinated or nonmyelinated.4 Myelination is especially evident in the raphe-cortical pathway, and myelinated axons are more frequent in monkey than rat4 This suggests that long-distance pathways may exhibit morphologic and physiologic specificity. Heterogeneity of conduction velocities has been shown among monkey LC neurons,3,28 with more rapid conduction velocities especially evident in LC-to-neocortex projections.’ A related possibility is that clustering of cells with similar efferent projections reflects the sharing of afferent inputs. Another possible consequence of clustering is enhanced local interactions among nearby neurons. For example, there is evidence sug gesting dendrodendritic interactions between LC neurons.9*‘o.‘5 Recent reports have indicated that some LC cells may contain certain neuropeptides. For example,

et

ai.

vasopressin- and corticotropin releasing factor-like immunoreactivity have been observed within LC neurons.5.32 This raises the possibility that the subpopulations defined by differential efferents and mor-

phology may express different neurochemical characteristics. This hypothesis remains to be tested. The subdivisions we have proposed can be conceived of as horizontal “sheets” of cells oriented from anterior to posterior. Axons within LC and LC cell dendritic fields are oriented predominantly along the anterior-posterior axis. 9~10Thus, axodendritic and dendrodendritic interactions influencing the distribution and impact of afferent information may be distributed in the same plane as neurons exhibiting similar efferent trajectories. It is thus possible that the functional organization of LC is far more specific and discrete than has been previously suggested. Such a greater degree of organization is consistent with many recent observations which indicate a high degree of specificity in the anatomic organization of LC efferents and in the postsynaptic physiologic effects of norepinephrine (reviewed in Ref. 8). Acknowledgements-Jim Zalinski provided expert advice on statistical analyses. Nancy Callahan and the Salk Institute Photo Lab assisted in preparing the manuscript. Floyd Bloom kindly provided encouragement and support for these studies.. Supported by PHS?Grant NS 21384 {SLF), MH 40008 (SLF). and NIAAA 06420 (SLF and F. E. Bloom). ~ ’

REFERENCES

1. Aston-Jones

G. and Bloom F. E. (1981) Activity of norepinephrine-containing locus coeruleus neurons in behaving rats anticipates fluctuations in the sleep-waking cycle. J. Neurosci. 1, 876886. 2. Aston-Jones G. and Bloom F. E. (1981) Norepinephrine-containing locus coeruieus neurons in behaving rats exhibit pronounced responses to non-anxious environmental stimuli. J. Neurosci. 1, 887-900. 3. Aston-Jones G., Foote S. L. and Segal M. (1986) Impulse conduction properties of noradrenergic locus coeruleus axons projecting to monkey oerebrocortex. Neuroscience (in press). 4. Axmitia E. and Gannon P. (1983) The ultrastructural localization of serotonin immunoreaetivity in myehnated and unmyehnated axons within the medial forebrain bundle of rat and monkey. J. Neurosci. 3, 20832090. 5. C&e A. R. and van Leeuwen F. W. (1983) Vasopressin-immunoteactive cells in the dorsomedial hypothalamic region, medial amygdaloid nucleus and locus coeruleus of the rat. Cell Tics. Res. W, 23-33. 6. Cintra L., Diaz-Cintra S., Kemper T. and Morgane P. J. (1982) Nucleus locus coeruleus: a morphometric Golgi study in rats of three age groups. Brain Res. 247, 17-28. 7. Foote S. L., Aston-Jones G. and Bloom F. E. (1980) Impulse activity of locus coeruleus neurons in awake rats and monkeys is a function of sensory stimulation and arousal. Proc. Mtn. Acad. Sci. U.S.A. 77, 3033-3037. 8. Foote S. L., Bloom F. E. and Aston-Jones G. (1983) Nucleus locus coetuleus: new evidence of anatomical and physiological speoificity. Physial. Rev. 63, 844-914. 9. Groves P. M. and Wilson C. J. (1980) Fine structure of rat locus coeruleus. J. camp. Neural. 193, 841-852. 10. Groves P. M. and Wilson C. J. (1980) Moooaminergic presynaptic axons and dendrites in rat locus coerukus seen in reconstructions of serial sections. J. camp. Neurol. 193, 853-862. 11. Gtzanoa R., Morrison J. H., Coyle J. T. and Molliver M. E. (1977) The immunohistochemical demonstration of noradrenergic neurons in the rat brain: the use of homologous antiserum to dopamine-beta-hydroxylase. Neurosci. Leff. 4, 127-134. 12. Grxaona R., Molliver M. E. and Coyle J. T. (1978) Visualization of central noradrenergic neurons in thick sections by the unlabelled antibody method: a transmitter-specific Golgi image. Proc. aafn. Acad. Sci. U.S.A. 75, 2502-2506. 13 Gr-xanoa R. and Molliver M. E. (198%) The locus coeruleus in the rat: an immunohistochernical detineation. Neuroscience 5, 21-40. 14. Iijima K. and Ogowa T. (1980) An HRP study on cell types and their regional topography within the locus cocruieus innervating the supraoptic nucleus in the rat. Acfa Hirrochem. 67, 127-138. 15. Koda L. Y., Aston-Jones G. and Bloom F. E. (1980) Small granular vesicles in locus coeruktw may indicate dendritic release of norepiaephrine. Neurosci. Abat. 4, 151.21. 16. Leeer L. and Hernandez N. (I%@) The cat locus coerukus: light and electron mierosoopk study of the neuronal somata. A&. &n&o/. l*, 181-198. 17. Lougbhn S. E., Foote S. L. and Bloom F. E. (1986) Efferent projections of nucleus locus COCN~US: topographic organization of cells of origin demonstrated by three-dimensional reconstruction. Neuroscience 1% 291-306.

Locus coeruleus cell classes: different efferents

319

18. Loy R., Koziell D., Lindsey J. and Moore R. Y. (1980) Noradrenergic innervation of the adult rat hippocampal formation. J. camp. Neurol. 189, 699-710. 19. Mason S. T. and Fibiger H. C. (1979) Regional topography within noradrenergic locus coeruleus as revealed by retrograde transport of horseradish peroxidase. J. camp. Neurol. 187, 703-724. 20. Mesulam M. (1978) Tetramethyl be&dine for horseradish peroxidase neurochemistry: a non-carcinogenic blue reaction product with superior sensitivity for visualizing neural afferents and efferents. J. H&o&m. Cytochem. 26, 106-I 17. 21. Moore R. Y. and Bloom F. E. (1979) Central catecholamine neuron systems: anatomy and physiology of the norepinephrine and epinephrine systems. Ann. Reu. Neurosci. 2, 113-168. 22. Olson L. and Fuxe K. (1971) On the projections from the locus coeruleus noradrenaline neurons: the cerebellar innervation. Bruin Res. 28, 165-171. 23. Rosene D. and Mesulam M. (1978) Fixation variables in horseradish peroxidase neurochemistry. J. Histochem. Cytochem. 26, 28-39.

24. Sakumoto T., Tohyama M., Satoh K., Yimoto V., Kinugasa T., Tanizawa N., Kuraci K. and Shimizu N. (1978) Afferent fiber connections from lower brain stem to hypothalamus studied by the horseradish peroxidase method with special reference to noradrenaline contents. Exp. Brain Res. 31, 81-94. 25. Saper C. B. and Loewy A. D. (1980) Efferent connections of the parabrachial nucleus in the rat. Brain Res. 197,291-317. 26. Satoh K., Tohyama M., Yamamoto K., Sakumoto T. and Shim&u N. (1978) Noradrenaline innervation of the spinal cord studied by the horseradish peroxidase method combined with monoamine oxidase staining. Exp. Bruin Res. 30, 175-186.

21 Segal M. and Landis S. (1974) Afferents to the septal area of the rat studied with the method of retrograde transport of horseradish peroxidase. Brain Res. 82, 263-268. 28. Segal M., Foote S. L. and Aston-Jones G. (1983) Physiological properties of ascending locus coeruleus axons in the squirrel monkey. Bruin Res. 247, 381-387. 29. Shimizu N., Katoh Y., Hida T. and Satoh K. (1979) The fine structural organization of the locus coeruleus in the rat, with reference to noradrenaline contents. Exp. Brain Res. 37, 139-148. 30. Shimizu N., Ohnishi S., Satoh K. and Tohyama M. (1978) Cellular organization of the locus coeruleus in the rat as studied by the Golgi method. Arch. Histol. Jup. 41, 103-112. 31. Swanson L. W. (1976) The locus coeruleus: a cytoarchitectonic, Golgi, and immunohistochemical study in the albino rat. Brain Res. 110, 39-56. 32. Swanson L. W., Sawchenko P. E., Rivier J. and Vale W. W. (1983) Organization of ovine corticotropin-releasing factor immunoreactive cells and fibers in the rat brain: an immunohistochemical study. Neuroendocrinology 26, 165-186. (Accepted 20 October 1985)