Okajjmas Folia Anat. Jpn., 58(4-6) : 419-442, March 1982

Distribution of Innervating the Supraoptic of the Rat with Special Reference to Neighbouring Relay Neurons Using HRP and Golgi Methods

By

KOICHI IIJIMA

Department of Anatomy, Akita University School of Medicine, Akita, Hondo, 010, Japan

-Received for Publication, September 17, 1981-

Key words : Supraoptic nucleus , Relay neurons, Rat, HRP, Rapid Golgi.

Summary : Detailed histochemical studies have been performed on the distribution of the nuclei innervating the supraoptic nucleus (SO) in the whole brain after horseradish peroxidase (HRP) injections through fine glass micropipettes (2-8μm tip diameter) into the SO of 39 Wistar rats. The chromatogenic reaction was carried out according to Mesulam's No. 8 procedure (1976). In the 3 cases, where injected HRP was most exactly restricted to the SO, HRP-labeling was limited to some small neurons located just dorsal to the SO. The other successful HRP injec- tions into the SO resulted in labeling rather few neurons which are sparsely distri- buted in various regions of the brain. The following conclusions were drawn from the analysis of these successful cases. The rapid Golgi method was used to check the morphology and possible functions of these small neurons revealed in the 3 cases . Every nuclei of the and posthypothalamus innervate the SO whereas probably very few neurons of the hypothalamic arcuate and ventromedial nuclei send to it. The afferent monoaminergic fibers arise largely from the B7 and B8, and to a lessor extent from the A6 and A10 cell groups. There is a high possibility that the mesencephalic central gray directly innervates the SO. The septum, diagonal band nucleus and send very few projections, whereas the olfactory tubercle and pyriform cortex may not. It remains unanswered whether the hippocampus projects to it or not. Small non-secretory neurons in and around the SO may be important relay neurons to large neurosecretory ones.

A recent electrophysiological study in studies (Leranth at al., 1975 ; ZAborszky vitro (Haller and Wakerley, 1980) has de- et al., 1975) have estimated two things monstrated that the osmoreceptor is loca- that only one third of terminals in ted separately from the supraoptic nucleus the SO probably originate outside the (SO) of the rat, indicating that function- nucleus, and the propotion of the number ing of the SO is controlled by other nu- of afferent fibers originating in various clei. subcortical structures to that of total aff- Quantitative electron microscopic (EM) erents to the SO. These results were

419 420 K. lijima based on EM obervations on degenerat- by means of the HRP method, and to ing axon terminals in the SO resulting study the merphology and possible role from transection of afferent pathways of of small non-secretory neurons in and various sources in various combinations. around the SO by the rapid Golgi method. However, neither of the studies describe the distribution of neurons innervating Material and Methods the SO at the light microscopic (LM) level. Although Swaab et al. (1975) have 39 female Wistar rats (200 g body we- found that about 10% of the neurons in ight) were used in an HRP study. They the hypothalamic neurosecretory nuclei were anesthetized with sodium pentobar- of the rat without nor vasop- bital (60 mg/kg i.p.) and subjected to ressin according to their immuno-histo- surgical operation in a stereotaxic appa- fluorescence study, the roles of these ratus so that KOnig and Klippel's stereo- cells as well as small neurons just dorsal taxic atlas (1963) could be used for locali- to the SO have not attracted special atten- zation of the SO. A glass micropipette tion hitherto. (2-8 pm tip diameter) filled with HRP It has not been determined by the HRP (Sigma type VI) which was dissolved into and Golgi studies (Millhouse, 1973 ;Bodoky 2% dimethylsulfoxide (DMSO) aqueous and Rethelyi, 1977) whether the hypotha- solution to a final concentration of 40% lamic arcuate (AR) and ventromedial (VM) (W /V) was vertically inserted to a depth nuclei project directly to the SO or not. of 8.5 mm below the cortical surface at Previous studies by the Fink-Heimer me- a point 1.7 mm lateral to the midline and thod and EM did not reveal the direct 6.4 mm anterior to the interaural line. projections to the SO from the olfactory DC currents (2-4 /IA) were passed for 5 tubercle, hippocampus and amygdaloid to 10 min between the micropipette and complex (Leonard and Scott, 1971; Za.bo- the indifferent electrode on the skull rszky et al., 1972). The relationship with the micropipette as anode. Some- between the raphe nuclei and SO is not times a train of square wave pulses (0.5 well known (Zaborszky et al., 1975 ; Pal- sec in diurection, 2-4 1211 in intensity) at kovits and Za.borszky, 1979). Histofluo- a rate of 1/sec was delivered for 10 to rescence, autoradiographic and HRP stu- 20 min instead of DC currents. After a dies demonstrate the presence of a pro- survival time of 48 h, animals were anest- jection from the A6 to the SO (Loizou, hetized with sodium pentobarbital (60 mg 1969 ; Jones and Moore, 1977 ; Iijima and /kg i.p.). Following an intracardial injec- Ogawa, 1980), but it is not clear whether tion of 1 % sodium nitrate solution puls other noradrenergic nuclei such as the 0.1 ml of heparin (1000 units/ml), perfusion Al, A2, A5, A7 and A8 cell groups project was carried out through the heart with to the SO or not, although their efferent a warm Ringer solution followed by a fibers are involved in the medial sequence of 2 phosphate buffered (0.1 M) bunble (MFB) (Lindvall and BjOrklund, aldehyde mixtures ; the 1st containing 1% 1974). paraformaldehyde and 1.25 % glutaral- The HRP method permits us to study dehyde, the 2nd 2 % paraformaldehyde the distribution of all the nuclei innervat- and 2.5 % glutaraldehyde. Thereafter, ing the SO at the LM level. The present the brain was removed from the skull investigation was undertaken to study it and stored overnight in 0.1 M phosphate throughout the whole brain and throw buffer (pH=7.4) containing 5 % sucrose light on these questions mentioned above at 4°C. The whole brain was cut into Neurons Innervating the Rat SO 421

45 ,um serially on a freezing microtome. according to the d!stribution patterns of All the sections were reacted using be injected HRP, and summarized in table 1. nzidine dihydrochloride (BDHC) after a Labeled neurons except for those in raphe slight modification of Mesulam's No 8 nuclei were observed ipsilateral to the procedure (1976). They were counters- injection site in all cases (Diagrams 1-5 ; tamed with a 1 % neutral red aqueous Figs. 1-18, 20-24). solution. Group I. 4 cases (22, 24, 29, 32) were 18 female Wistar rats (200g body included in this group. This group (except weight) were used for the rapid Golgi for case 32) was characterized by the method (Romeis, 1968). The optimal in- finding that injected HRP was completely cubation period was 24 h for small neurons restricted to the SO with some small just dorsal to the SO. Celloidin sections HRP-labeled neurons in nearby tissue were cut into 50 rim. In order to diffe- (Diagram 1G-1; Figs. 1, 2). Photomicro- rentiate them from SO neurosecetory graphs of the injection sites are illustrated neurons, karyometry was carried out by in Figs. 1-7. using a phasecontrast microscope (Nikon, In cases 22 and 24, injected HRP co- DLL x 100) and a screw-micrometer eye- vered mostly the ventral part of the SO piece (Nihonkogaku Co.) on both groups leaving small rostral and dorsal parts of of neurons. The eyepiece allowed a ma- the nucleus free of HRP. In both cases rgin of error less than 0.1 pm. Both the no labeling could be detected throughout largest (A) and shortest (B) axes of each the brain except for the perinclear tissue nucleus were measured. The nuclear of the SO, where small labeled neurons volume was calculated according to the were recognizable (table 1, Diagram 1G-1). formula In case 29, where injected HRP covered the SO almost completely and labeled some V,----47/3(A/2)(B/2)2. small neurons just dorsal to it (Figs. 1, 2), The mean nuclear volume and standard we could not find convincing evidence of deviation were obtained from these data. labeling in any parts of the brain, alt- The t-test was carried out to check the hough a cluster of granule cell-like struc- presence or absence of significant diffe- tures in the hippocampus and a large rence between the 2 groups of neurons. round cell-like one in the linear nucleus (A10 cell group) appeared to be labeled Results with HRP. In these 3 cases, where HRP in- jection into the SO seemed most successful, I. The HRP Method we could detect signs of HRP-labeling The follwing descriptions were limited only in some small neurons located just to the 10 cases where the injection sites dorsal to the SO. were well localized to the principal part In connection with these findings, the of the SO (Leranth et al., 1975) with no results obtained from case 32 were de- (cases 22, 24, 29)(Diagram 1G-1) or negli- sdribed in table 1, where HRP was in- gible (cases 19, 21, 25, 26, 32, 35, 36) jected slightly lateral to the center of Diagrams 1G-2, 1G-3, 1G-4) diffusion of the SO, but contamination of the perinu- HRP in the surrounding tissue and 3 cases clear tissue with HRP was negligible (cases 2, 3, 34) (Diagram 1C) where the (Fig. 3). In this case, some nerve fibers injection sites were outside the SO, al- with a positive beaded apperrance were though in total 39 rats were tested. These observed entering the SO from the sur- 10 useful cascs were cassified into 4 groups rounding hypothalamic tissue (Fig. 3). A Table 1. The distribution of HRP-labeled neurons in the rat brain. LND

The number of + signs indicates the degree of labeling and the — sign indicates the absence of labeling . The * signs indicate successful cases. The LC column numbers indicate the total Dumber of labeled neurons found in the LC. Parenthses in case 29 indicate difficulty in identifying the labeling signs as positive findings. In the 2nd column , the following abbreviations are used. ha ; anterior hypothalamic nucleus ; hl ; lateral hypothalamic nucleus ; pz ; perinuclear tissue just dorsal to the SO ; — and + parenthses indicate the absence and presence of contamination with HRP , respec- tively. In the 4th column, C± signs indicate the presence of minimal diffusion of HRP. Neurons Innervating the Rat SO 423

Diagram 1. Summary diagrams of localization and extent of HRP injection sites are marked on line drawings of frontal section of the rat brain adopted with a little modification from the KOnig and Klippel (1963) atlas. Blackened areas and arrows indicate the injec- tion sites and the track of a glass micropipette, respectively. G-1 through G-4 and C correspond to Groups I through IV and control in the text, respectively. part of them appeared to be coming from was detected in each of the AR, Forel's the MFB. In addition, a few small field H2, and mesencephalic central gray neurons corresponding with "intercalated (CG), but not in the locus coeruleus (LC, neurons" (Swaab et al., 1975) could be A6). discerned from many large neurosecretory Group II. Cases 19 and 36 were included neurons (Fig. 5). A few extrinsic fibers in this group (table 1). This group was were traced very close to the latter, characterized by the findings that HRP indicating that these neurons should be injection into the SO was complete with innervated by other separate nuclei (Haller neglible diffusion to the surrounding tissue and Wakerley, 1980) (Fig. 4). As shown and that the lateral hypothalamic nucleus by table 1, very few neurons were labeled (hl) including a part of axon terminals in preoptic areas. medial septal and di- of the MFB was devoid of HRP deposi- agonal band nuclei. A labeled tion (Diagram 1G-2). No labeled neurons 424 K. Iijima

were found in the LC (table 1). and Fuxe, 1964) and CG (Diagram 2F, In case 36, very few neurons were Fig. 12). moderately labeled in the lateral preoptic In case 19, where the glass micropi- nucleus while a few neurons of the di- pette passed through the anterior hypo- agonal band nucleus showed mild labeling thalamic nucleus before entering the SO, (Fig. 8). In the cortical amygdaloid a very small amount of HRP appeared nucleus a few neurons showed mild to to have diffused out to that nucleus and moderate labeling (Diagram 2B ; Figs. 9, preoptic areas but not to the hl (Diagram 10). Very few neurons were moderately 1G-2). Although they were more sparse, labeled in the Forel's field H2 (Diagram very few labeled neurons in the medial 2C, Fig. 11), ventral premamillary (Diagram basal and the posthypo- 2C) and supramamillary (Diagram 2D, were distributed more widely Fig. 16) nuclei. A mildly labeled neuron than in case 36 (table 1). Very few la- was observed in each of the dorsal raphe beled neurons were recognizable in the nucleus (B7 in Diagram 2F ; cf. DahlstrOm lateral septal nucleus (Diagram 2A), AR

Diagram 2. Summary diagrams of locations of labeled neurons in Group II including cases 19 and 36. The presence of a single or very few labeled neurons is marked with a black triangle and increased labeling with a solid black circle. Neurons Innervating the Rat SO 425

(Diagram 2C ; Figs. 13, 14) and VM (Di- some axon terminals of the MFB, was agram 2B, Fig. 15), (Diagram negligible. The whole SO was completely 2B), posterior hypothalamic (Diagram 2C), covered with injected HRP. premamillary (Diagram 2C) nuclei and In case 25, HRP was actually injected A10 (Diagram 2E, Fig. 17). But, no into the , which resulted in labeled neurons were detected in the covering the adjacent SO completely with diagonal hand nucleus and CG as opposed a slight contamination in the ha (Diagram to case 36 (table 1). 1G-3). The medial septal nucleus showed Group III- Cases 25 and 35 were inclu- very few moderately labeled neurons (Di- ded in this group (table 1). It is to be agram 3A, Fig. 18). In the amygdaloid noted that only 1 or 2 neurons were complex (Diagram 1G-3), a small neuron labeled in the LC (A6). Some small showed moderate labeling. A densely neurons were also labeled just dorsal to labeled neuron was observed in the VM the SO (Diagram 1G-3). Diffusion to the (Diagram 3B), but not in any other medial hl, which might result in contaminating basal hypothalamic nuclei (table 1). Mo-

Diagram 3. Summary diagrams of locations of labeled neurons fn Group III including cases 25 and 35. It is to be noted that labeling appeared in the LC. 426 K. Iijima

derate labeling was detected in very few 1963) (Diagram 1G-4). Although the pe- neurons of the posterior hypothalamic, riuclear tissue just dorsal to the SO dorsal premamillary and supramamillary appeared to be not significantly contami- nuclei (Diagram 3C, D). Mild labeling nated with HRP, positive granules were was recognizable in very few neurons in observable in the perikarya of some neu- the dorsal and median raphe nuclei (B7 rons in the hl, which indicated possible and B8 in Diagram 3E, respectively ; contamination of some axon terminals of DahlstOm and Fexe, 1964). It is to be noted the MFB ending there (case 26). It should that only one moderately labeled mul- be noted that labeling of LC neurons tipolar neuron was found throughout the increased in number significantly in this LC in case 25 (table 1, Diagram 3F, Fig. Group IV compared with the other groups 20). (table 1, Diagram 4H), and that several In case 35, the glass micropipette passed neurons of the AVT were definitely la- through the hl vertically to the SO, but beled (table 1, Diagram 4E). contamination of the ventral zone of it In case 26 (table 1, Diagram 1G-4), very with HRP was not significant (Fig. 1G-3). few neurons were moderately labeled in A few densely labeled neurons were the medial (Diagram 4A). sparsely distributed in the medial and The diagonal band nucleus revealed a lateral preopic areas (Diagram 3A). A few intensely labeled neurons whereas few neurons showed moderate labeling the lateral septal nucleus exhibited very in the diagonal band nucleus, and in both few mildly labeled neurons (Diagram 4A). medial and lateral septal nuclei (Diagram The hippocampus proper showed several 3A). In the cortical amygdaloid nucleus, densely labeled ones (Diagram 4B, Fig. few mildly to moderately labeled neurons 21). In the medial amygdaloid nucleus a were observed (Diagrams 1G-3, 3B). A mildly labeled neuron was detected. moderately labeled neuron was detected However, the cortical amygdaloid nucleus in the Forel's field H2 (Diagram 3C) while showed a collection of several mildly a mildly labeled one was recognizable in labeled neurons (Diagram 4B). In the the zona incerta (Diagram 3B). A mode- AR, a moderately labeled neuron was rately labeled neuron was detected in detected (Diagram 4C), while in the VM, each of the ventral premamillary (Di- very few mildly labeled ones were recog- agram 3C) and supramamillary (Diagram nizable (Diagram 4B). Few neurons in 3D) nuclei. Rather densely labeled neurons the zona incerta (Diagram 4B) and very were sparsely seen in the B7 and B8 few ones in the Forel's field H2 (Diagram (Diagram 3E, Fig. 19). A darkly labeled 4C) showed moderate to intense labeling. neuron was observed in the pontine The supramamillary nucleus showed central gray (table 1, PG in Diagram 3F). several densely labeled neurons (Diagram It is to be noted that 2 fusiform neurons 4D) whereas the ventral premamillary of the LC showed moderate labeling nucleus revealed only a mildly labeled one (table 1, Diagram 3F). (Diagram 4C). The AVT clearly showed Group IV. Cases 21 and 26 were included some densely labeled neurons at the in this group (table 1), where injected hypothalamc-mesencephalic border (Di- HRP slightly contaminated a part of the agram 4E). A few neurons showed in- axon terminals of the MFB ending in tense labeling in the A10 and B8 cell the hl, because The glass micropipette groups (Diagram 4F, G). In case 26, 6 passed through the hl into the SO at the labeled neurons of both fusiform and level of A 5780 pm (Konig and Klippel, multipolar shapes (Swanson, 1976) were Neurons Innervating the Rat SO 427

Diagram 4. Summary diagrams of locations of labeled neurons in Group IV including cases 21 and 26. Note the appearance of labeling in HI (B) and LGB (D) , the further increase in labeling in LC (H) , ac (B) , and sum (D) , and strong labeling in AVT (E) . found throughout the LC (Diagram 4H). shown by table 1, 2 neurons of the LC This number of labeled LC neurons was (Diagram 5H) and several ones of the the maximum number of all the 10 cases AVT Diagram 5F) were labeled. It should from Group I to IV (table 1). A densely be noted that labeling in the septum, labeled multipolar neuron was seen in the diagonal band nucleus, Forel's field FI, and Al. Several densely abeled neurons were raphe nuclei increased the number of la- seen in the lateral geniculate body (Di- beled neurons significantly in this case agram 4D). In both cases 21 and 26, compared with the 10 cases listed in table several moderately labeled neurons were 1, and that few neurons were labeled in also detected in the thalamic regions such the olfactory tubercle (Diagram 5A, Fig. as the periventricular (pvr) and medial 22), PG, marginal nucleus of the superior thalamic (tm) nuclei (table 1, Diagram cerebellar peduncle (MB, Swanson, 1976) 4B). and the reticular formation of the pons Controls. Cases 34, 2 and 3 were adopted (NRC) (table 1, Diagram 5H). as controls. In these cases, HRP was In case 2, injected HRP was centered injected outside the SO (table 1, Diagram in the ventral part of the hl at the level 1C). of A 5780 pm (Konig and Klippel, 1963) In case 34, HRP was injected mainly (Diagram 1C), involving a part of the into the perinuclear tissue just dorsal to lateral preoptic nucleus. 8 neurons of the SO. Diffusion of HRP into the SO fusiform and mulitolar shapes (Swanson, and hl was slight (Diagram 1C). As 1976) were densely labeled in the whole 428 K. lijima

Diagram 5. Summary diagrams of locations of labeled neurons in the control Group including cases 34, 2 and 3. It is to be noted that labeled neurons appeared in the olfactory tu- bercle (TULI in A), MB, NRC, and RM (H) .

LC (table 1, Diagram 5H). In the diagonal nucleus. Raphe nuclei such as the A10 band nucleus some strongly labeled neu- (Diagram 5F), B7 and B8 cell groups (Dia- rons were clearly seen (Diagram 5A, Fig. gram 5G) showed labeling. Above all, 23) whereas in the olfactory tubercle only the B7 showed 1 or 2 labeled neurons in a few moderately labeled ones were almost every section (Diagram 5G). As visible (Diagram 5A). Several densely la- shown in case 34, labeled neurons were ob- beled neurons were observed in the medial servable in the pvr and tm (Diagram 5C). septal nucleus (Dagram 5A), while a few Very few neurons in the PG showed mo- moderately labeled ones were seen in the derate to strong labeling (Diagram 5H). lateral septal nucleus (Diagram 5A, B). In case 3, HRP was injected into the hl at Only very few pyramidal cells of the the level of A 6060 ttm (Konig and Klippel, hippocampus proper and fusiform cells of 1963) involving the dorsal part of the the cortical amygdaloid nucleus were group of axon terminals of the MFB in mildly to moderately labeled (Diagram it (Diagram 1C). Contamination of them 5C). As shown in table 1 and Diagram ending there resulted in significant in- 5C-E, labeled neurons were detected in crease in the number of labeled neurons every area of the medial basal hypo- in the LC (table 1, Diagram 5H, Fig. 24) thalamus, subthalamus and posthypotha- as well as in the olfactory tubercle Di- lamus except for the paraventricular agram 5A, cf. Fig. 23), the B7 (Diagram Neurons Innervating the Rat SO 429

5G) and the AVT (Diagram 5F) whereas Discussion it reduced significantly labeling in the septum and diagonal hand nucleus, so At first sight, the essentially negative that it was difficult to find labeled neurons result in case 22, 24 and 29, where inje- in them (table 1). In fact, 22 densely cted HRP was most completely restricted labeled neurons were clearly visible in to the SO. would seem to be not recon- the whole LC (table 1, Fig. 24). An in- ciled with our previous HRP study (Iijima tense reaction was observed also in the and Ogawa, 1980) and to support the lateral preoptic nucleus (Diagram 5B). concept that SO neurons have a direct However, no labeled neurons could be osmosensitivity (Lenk, 1979). However , detected in the amygdala and hippo- at least three factors are probably respon- campus (table 1). In regard to the thal- sible for this negative result. First, the amic regions (pvr and tm), there was no paucity of extrinsic axon terminals in the difference between control cases 2 and SO (Leranth et al., 1975) is very likely 3. Labeled neurons were observed in to make it difficult to label the neurons case 3 just as seen in case 2 (table 1, in their origins (cf. Jones, 1975). Second , Diagram 5C). it is technically very difficult to cover II. The Rapid Golgi Method the whole SO with injected HRP while The perinuclear tissue just dorsal to avoiding diffusion of it outside the SO. the SO revealed a fine nerve plexus in Third, small neurons of various shapes which some small neurons of various around the SO were definitely labeled shapes were darkly stained, whereas the with HRP (Fig. 2) in these cases, indica- SO was lightly stained. LBut, in the SO ting that they would innervate SO neur- several afferent fibers appeared to termi- ons and may play an important role as nate with many fine collaterals (Figs. 6, relay ones (Compare Fig. 8 with Fig. 23). 7). In order to restrict HRP to the SO By means of a screw-micrometer eye- more precisely, much finer glass micro- piece, the following _results were obtained pipettes (2-8 pm tip diameter) were used from the karyometry on the small neurons in this study than those (50-70 pm tip and SO neurons in case 29. The mean diameter) in our previous one (Iijima and nuclear volume of the small neurons just Ogawa, 1980). Thus, the previous results dorsal to the SO is 144. 76+109. 28 pms apparently correspond with those in Group (n=72). The mean nuclear volume of III in the present study, but failed to label SO neurons in case 29 is 311. 01+184. 86 the small neurons as shown by Group pm' (n=69). Thus, t=6. 48***(p<0. 005). I. In cases 22 and 24, rostral and dorsal The difference between the 2 groups of parts of the SO were not covered neurons is significant in regard to the with HRP. In case 29, though the whole nuclear volume. In comparing small im- SO appeared to be completely covered pregnated neurons by the Rapid Golgi with it, significant variations of the in- method, in size, shape and location with tensity of staining observed in SO neurons those labeled with HRP around the SO, in every section (Fig.2) indicate the high the former is certain to be identical to possibility that the concentration of in- the latter. jected HRP might not be enough to label the nuclei projecting into the SO. Ftr- thermore, present findings in case 32 clearly show the presence of some afferent fibers entering the SO (Fig. 3), which 430 K. lijima seems to be consistent with ',a recent are very likely to play an important role study by Swanson et al. (1981). The as relay neurons interconnecting various presence of projections from the LC to origins such as the septum, diagonal band the SO has been already demonstrated by nucleus, Forel's field H2 and dorsal raphe the histofluorescence and autoradiographic nucleus with the SO, but not the LC and methods (loizou, 1969 ; Jones and Moore, medial basal nuclei. 1977). In view of the present findings to- Sakumoto et al. (1978) report marked gether with these previous reports, it seems increase of the number of labeled LC reasonable to conclude that the negative neurons after HRP injection into the hl. resnit in Group I except for labeling signs It is known that some neurons of the hl in and around the SO should be unreliable are relay ones on which the axons of the and indicate merely the paucity of affe- MFB are partly teminated (Palkovits and rents entering the SO (Figs., 6, 7). This ZAborszky, 1979). Therefore, it seems conclusion gains support by physiological important to note that moderate to in- studies in vivo and in vitro reporting tense labeling in the LC must be con- that a separate osmoreceptor is required sidered to be due to contamination of for the activation of SO neurons during the hi with HRP after injection into the osmotic stimulation (Anderson et al., 1975 ; SO. In Group IV, the lateral geniculate Van Gemert et al., 1975 ; Haller and Wa- body and thalamic regions (pvr, tm) also kerley, 1980). Moreover, in case 32 (Group revealed labeling, indicating the presence I) the labeling signs in certain sites except of significant diffusion of HRP outside for those in and around the SO were the SO. Therefore, it seems safe to di- not evident enough to be evaluated. sregard the results in Group IV. It is to be noted that small HRP-labeled It seems noteworthy that labeling could neurons just dorsal to the SO in Group be detecred sparsely but constantly in I (cases 22, 24, 29) corresponded with every area of the subthalamus and po- small neurons of various shapes embedded sthypothalamus in Groups II and III. in the fine nerve plexus whose axons Although Woods et al. (1969) concluded appeared going toward neurosecretory the presence of direct projections from neurons in Golgi preparations in this them to the rabbit SO with physiological study. With regard to these neurons, and degeneration methods, Zaborszky and Swaab et al. (1975) found "non-secretory Makara (1979) limited the possible origins intercalated neurons" in the the rat SO of the axons in SO to premamillary indirectly by the immunohistofluorescene nuclei and VM in the rat. The present method. Although they estimated the results indicate that all nuclei of both number of these cells at about 10 % of the subthalamus and posthypothalamus all SO neurons, their morphology was send axons to the rat SO as well as the not described. A recent EM study failed rabbit SO. to identify them in the SO (Leranth et It is interesting to note that although al., 1975) because of paucity of these ZAborszky et al. (1975) previded the as- cells within it. Their functions are not smption with the EM that a substantial well known. The present study indicates part of total afferents to the SO might they are present around the SO rather arise from the hypothalamus, in the pre- than within it, that they are undoubtedly sent study very few labeled neurons different from neurosecretory ones (p < were detected in the AR (case 19) and O. 005), and that on the basis of the VM (cases 19 and 25). Some authors results obtained from control cases, they have reported with the EM that both Neurons Innervating the Rat SO 431

nuclei may project to the rat SO (ZAbor- The positive result in Group II (Fig. 12) szky et al., 1975 ; ZAborszky and Makara, indicates that this area has a high pos- 1979) whereas others have limited their sibility of innervating the SO directly. lateral projections to the hl with the It is important to note the marked Golgi anp HRP methods (Millhouse, 1973; differences between the diagonal band Bodoky and Rethelyi, 1977). As both nucleus and the olfactory tubercle shown nuclei have rostral projections to the by the present experiments (table 1). anterior hypothalamic nucleus (ha) (Mil- Although Zaborszky et al. (1975) have lhouse, 1973 ; Harris and Sanghera, 1974), emphasized that the olfactory tubercle the positive results in cases 19 and and a further rostra! pyriform cortex 25 may not be convincing evidence must directly innervate the SO, the posi- owing to a slight contamination in the tive result was obtained from control ha, but this suggests strongly that both cases only (table 1). On the contrary, neurons in the AR and VM probably send the diagonal band nucleus showed labeling much fewer axons than estimated before in most cases (able 1). The present re- (Zaborszky et al., 1975) into the SO. sults indicate that the diagonal band It is to be noted that certain raphe nucleus projects directly to the SO whe- nuclei such as the B7, B8 and A10 reas the olfactory tubercle and the pyri- (DahlstrOm and Fuxe,1964) showed label- form cortex do not. This conclusion ing fairly constantly. Zaborszky et al . gains support with the Fink-Heimer, (1975) conclude that a considerable propor- Nauta and Nauta-Gygax methods (Powell tion of afferents to the SO must arise et al., 1965 ; Palkovits and Zaborszky, from the brain stem, without indicating 1979) their possible origins. According to bio- In the present study, the septum show- chemical and degeneration studies (Palko- ed labeling in almost all cases (table vits et al., 1977), most of serotonergic 1). The present positive results support fibers in the hypothalamus originate in the traditional concept based on the LM the B7, but their exact course in passing and EM lesion studies that the septum through there is still unknown (Palkovits must have direct projections to the rat and Zaborszky, 1979). With regard to SO (Szentagothai et al., 1968 ; Tangapre- noradrenergic fibers to the SO, Iijima and gasson et al., 1974; Zaborszky et al., Ogawa (1980) have reported with HRP 1975), although the number of axons that a few LC neurons project to the entering the SO must be very few, as ipsilateral SO of the rat. The present indicated by the positive cases 32, 19, 25, findings together with these previous 35). reports suggest strongly that mono- Although Zborszky et al. 1975) argue aminergic fibers entering the SO arise the presence of direct projections from largely from the B7 and B8, and to a the amygdala and hippocampus to the rat lassor extent from the A6 and A10 in SO, this is only partly supported by the the rat. present results, because the amygdala Previously, the mesencephalic central showed labeling in cases 36, 25 and 35, gray was pointed out by an electrophy- whereas the positive result in the hippo- siological study as one of the most active campus was limited to the unreliable sites related to functioning of the rabbit case 26 (the hl was contaminated ; table SO with which it was probably connected 1) and control cases only. The problem in directly by way of the subthalamus whether the hippocampus innervates the and posthypothalamas (Woods et al., 1969). SO directly or not remains to be clarified. 432 K. Iijima

author with his excellent techniques. Acknowledgements The author is also thankful to Dr. K, Dr. Tetsuro Ogawa, professor of Physio- Imai, Miss. Y. Hayami and Mr. N. Kojima logy Departmeht, has encouraged the for their technical assistance.

Abbreviations

A10: Linear nucleus SM : Medullary stria AR : Hypothalamic SO: Supraoptic nucleus AVT : Ventral tegmental area TCC : Trunk of the corpus callosum B7: Dorsal raphe nucleus TO: Optic tract B8: Median raphe nucleus TUL : Cortical part of the olfactory CA : Anterior commissure tubercle CAI: Internal capsule TULI : Inner part of the olfactory tubercle CC : Cerebral crus VM : Hypothalamic ventromedial nucleus CG : Mesencephalic central gray Vm : Motor nucleus of the trigeminal nerve F: Fornix ZI : Zona incerta FOP : Optic fascicle am : Medial amygdaloid nucleus FH : Hippocampal fimbria ac : Cortical amygdaloid nucleus FLM : Medial longitudinal fascicle cp : Caudate-putamen complex FMT : Mamillothalamic fascicle ha : Anterior hypothalamic nucleus GP : Globus pallidus hdm : Hypothalamic dorsomedial nucleus H : Forel's field H2 hl : Lateral hypothalamic nucleus HI : Hippocampus proper hp : Posterior hypothalamic nucleus LC : Locus coeruleus lh : Lateral habenular nucleus LGB : Lateral geniculate body pd : Dorsal premamillary nucleus LM : pol : Lateral preoptic nucleus MB : Marginal nucleus of the superior pom : Medial preoptic nucleus cerebellar peduncle pv : Ventral premamillary nucleus MFB : pvr : Periventricular thalamic nucleus MT : Mesencephalic tr. nucleus of the r: Red nucleus trigeminal nerve sl : Lateral septal nucleus NRC : Reticular formation of the pons sm : Medial septal nucleus Pcs : Superior cerebellar peduncle sum : Supramamillary nucleus PG : Pontine central gray td : Diagonal band nucleus PVN : Paraventricular nucleus tm : Medial thalamic nucleus RM : Large raphe nucleus tv : Ventral thalamic nucleus Neurons Innervating the Rat SO 433

References and distribution of the amygdalofugal pathways in the rat : An experimental 1) Anderson, B., Leksell, L. G., Lishajko, neuroanatomical study. J. comp. Neurol., F. : Perturbations in fluid balance in- 141 : 313-340, 1971. duced by medially placed forebrain le- 12) Lerânth, Cs., nborszky, L., Marton, J. sions. Brain Res., 99: 261-275, 1975. and Palkovits, M. : Quantitative studies 2) Bodoky, M. and Rethelyi, K.: Dendritic on the supraoptic nucleus in the rat. I. arborization and axon trajectory of neu- Synaptic organization. Exp. Brain Res., rons in the hypothalamic arcuate nucleus 22: 509-523, 1975. of the rat. Exp. Brain Res., 28 : 543-555, 13) Lindvall, 0. and BjOrklund, A. : The 1977. organization of the ascending catechol- 3) DahlistrOm, A. and Fuxe, K.: Evidence amine neuron system in the rat brain as for the existence of monoamine-containing revealed by the glyoxylic acid fluores- neurons in the central nervous system. cence method. Acta physiol. scand., 92: I. Demonstration of monoamines in the (Suppl. 412) , 1-48, 1974. cell bodies of brain stem neurons. Acta 14) Loizou, L. A.: Projections of the nucleus Physiol. scand., 62 : (Suppl. 232) , 1-55,1964. locus coeruleus in the albino rat. Brain 4) Haller, E. W. and Wakerley, J. B.: Elec- Res., 15: 563-566, 1969. • trophysiological studies of paraventricular 15) Mesulam, M. M.: The blue reaction and supraoptic neurons recorded in vitro product in horseradish peroxidase neuro- from slices of rat hypothalamus. J. histochemistry : incubation parameters Physiol., 302 : 347-362, 1980. and visibility. J. Histochem. Cytochem., 5) Harris, M. C. and Sanghera, M.: Projec- 24: 1273-1280, 1976. tion of medial basal hypothalamic neu- 16) Millhouse, 0. E.: The organization of tones to the preoptic anterior hypotha- the ventromedial nucleus. Brain Res., lamic areas and the paraventricular 55: 71-87, 1973. nucleus in the rat. Brain Res., 81 : 401- 17) Palkovits, M., Saavedra, J. M., Jacobowitz, 411, 1974. D. M., Kizer, J. S., Ziborszky, L. and 6) Iijima, K. and Ogawa, T.: An HRP Brownstein, M. J.: Serotonergic innerva- study on cell types and their regional tion of the forebrain : Effects of lesions topography within the locus coeruleus on serotonin and tryptophan hydroxylase innervating the supraoptic nucleus of the levels. Brain Res., 130: 121-134, 1977. rat. Acta histochem., 67: 127-138, 1980. 18) Palkovits, M. and nborszky, L.: Neural 7) Jones, B. E.: Possible determinants of connections of the hypothalamus. In : the degree of retrograde neuronal label- Anatomy of the Hypothalamus (Morgane, ing with horseradish peroxidase. Brain P. J. and Punksepp, J. eds.) New York Res., 85 : 249-253, 1975. and Basel, Marcel Dekkar, 1979. 8) Jones, B. E. and Moore, R. Y.: Ascend- 19) Powell, T. P. S., Cowan, W. M. and ing projections of the locus coeruleus in Raisman, G.: The central olfactory con- the rat. II. Autoradiographic study. nexions. J. Anat., 99: 791-813, 1965. Brain Res., 127: 23-53, 1977. 20) Romeis, B.: Rasche Methode von Golgi. 9) Konig, J. F. R. and Klippel, R. A.: In In : Mikroskopische Technik. Miinchen "The Rat Brain . A Stereotaxic Atlas of und Wien : R. Oldenbourg Verlag, 1968. ' the Forebrain and Lower Parts of the 21) Sakumoto, T., Tohyama, M., Satoh, K., Forebrain and Lower Parts of the Brain Kimoto, Y., Kinugasa, T., Tanizawa, 0., Stem". Baltimore, Williams and Wilkins, Kuraki, K. and Shimizu, N.: Afferent 1963. fiber connections from lower brain stem 10) Leng, G.: Responses of supraoptic neu- to hypothalamus studied by the horsera- rones to local hypertonic saline in the dish peroxidase method with special rat. J. Physiol., 296: 62-63, 1979. reference to noradrenaline innervation. 11) Leonard, C. M. and Scott, J. W.: Origin Exp. Brain Res., 31 : 81-94, 1978. 434 K. Iijima

22) Swaab, D.F., Nijeveldt, F. and Pool, C. W. : ture du noyau supraoptique. Ann. Endo- Distribution of oxytocin and crinol. (Saris) ., 35 : 149-152, 1974. in the rat supraoptic and paraventricular 27) Van Gemert, M., Miller, M., Carey, R. J. nucleus. J. Endocr., 67: 461-462, 1975. and Moses, A. M.: Polyuria and impaired 23) Swanson, L. W.: The locus coeruleus : ADH release following medial preoptic a cytoarchitectonic, Golgi and immunohis- lesioning in the rat. Amer. J. Physiol., tochemical study in the albino rat. Brain 228: 1293-1297, 1975. Res., 110: 39-56, 1976. 28) Woods, W. H., Holland, R. C. and Powell, 24) Swanson, L. W., Sawchenko, P. E., Berod, F. W.: Connections of cerebral structure A., Hartman, B. K., Helle, K. B. and functioning in neurohypophysial Vanorden, D. E.: An immunohistochemi- release. Brain Res., 12: 26-46, 1969. cal study of the organization of catechol- 29) nborszky, L., Lerânth, Cs., Makara, aminergic cells and terminal fields in the G. B. and Palkovits, M.: Quantitative paraventricular and supraoptic nuclei of studies on the supraoptic nucleus in the the hypothalamus. J. comp. Neurol., 196 : rat. II. Afferent fiber connections. Exp. 271-285, 1981. Brain Res., 22: 525-540, 1975. 25) SzentAgothai, J., Flerk6, B., Mess, B. 30) Zaborszky, L., Leranth, Cs. and Palko- and Halgsz, B.: Afferent hypothalamic vits, M.: Faserdegeneration in Hypo- pathways and their mode of termination thalamus und im limbischen Systeme nach in the different cellular masses. In: konventionellen (dorsomedianen) Penetra- Hypothalamic Control of the Anterior tionen. Anat. Anz. Erg., 130: 595-600, Pituitary. pp. 56-62, Budapest : Akade- 1972. miai Kiad6, 1968. 81) ZAborszky, L. and Makara, G. B.: Intra- 26) Tangapregasson, A. M., Tangapregasson, hyptothalamic connections : An electron M. I., Soulailac, A. and Soulailac, M. L.: microscopic study in the rat. Exp. Brain Effects lesions septales sur l'ultrastruc- Res., 34 : 201-215, 1979. Neurons Innervating the Rat SO 435

PLATES 436 K. lijima

Explanation of Figures

Plate I

Figs. 1-5. Photomicrographs show the sections in Group I where HRP was injected within the SO.

Fig. 1. A low-power photomicrograph shows the principal part of the SO (Lerânth et al., 1975) adjacent to the optic tract (TO) and the perinuclear tissue just dorsal to the SO (pz) . A part of the pz is surrounded by arrows. The lines indicate the inner limit of the pz, which is adjacent to the ventral part of the lateral hypothalamic nucleus (hl). Only small cells of the pz are labeled. Case 29. x 150.

Fig. 2. The same part of the pz as indicated by arrows in Fig. 1 is magnified ( x 375) in Fig. 2. The SO is almost completely covered with HRP. The perinuclear tissue just dorsal to it (pz) reveals some densely labeled small neurons. Case 29. x 375.

Figs. 3-5. The sections of case 32 are illustrated at medium ( x 300) and high ( x 1,500) magni- fications.

Fig. 3. Some afferent fibers (arrows) with a positively stained beaded appearnce enter the SO adjacent to the optic tract (TO). x 300.

Fig. 4. An afferent fiber with a positive beaded appearance (arrows) appears to terminate very close to a secretory neuron (N) showing strong activity nearby a blood capillary (b) x 1,500.

Fig. 5. This photomicrograph shows 2 types of SO neurons with strong activity lying close to each other at high magnification. The large one (N) is secretory whereas the smaller multipolar one (ic) is to be non-secretory intercalated in nature. x 1,500.

Figs. 6-7. Medium ( x 375) and high -power ( x 600) photomicrographs show the section corres- ponding with those of Figs. 1 and 2, by the Rapid Golgi Method.

Fig. 6. Afferent fibers (arrows) appear to terminate within the lightly stained SO where many fine collaterals arise from these fibers. The perinuclear tissue surrounding the SO a-id TO (pz) reveals the fine nerve plexus in which some small neurons of various shapes (n) are embedded, indicating relay functions. x 375.

Fig. 7. The same neuron as indicated by line=n in Fig. 6 is shown at a higher magnification. The nucleus (line N) and nucleolus (line n) are unstained whereas the perikaryon and (d) are darkly stained. The processes are extending into the surrounding tissue for some distance. x 600. 437 Plate I

K. lijima 438 K. lijima

Plate II

Fig. 8. A low-power photomicrograph shows the diagonal band nucleus (td) of case 36, where a labeled neuron (arrow) nearby a blood capillary (line b) is recognizable. x 150. This neuron is shown at a higher magnification in the upper left insertion (N) . x 600.

Fig. 9. A high-power photomicrograph shows a part of the cortical amygdaloid nucleus (AC) of case 36, where a few neurons (arrows) exhibit mild to moderate labeling. x 600.

Fig. 10. The upper neuron (arrow) in Fig. 9 is shown at a higher magnification. AC, the cortical amygdaloid nucleus. x 1,050.

Fig. 11. A moderately labeled neuron of the Forel's field H2(H) of case 36 is shown at a higher magnification. Many positive granules are packed in a cell process as well as in the peri- karyon of this neuron (line N) . H, Forel's field H2 x 1,500.

Fig. 12. A high-power photomicrograph exhibits a mildly labeled fusiform neuron (line N) located in the mesencephalic central gray (CG) of case 36. x 1,500.

Fig. 13. A low-power photomicrograph shows the location of a labeled neuron (arrow) of the hypothalamic arcuate nucleus (AR) . It is located at the ventrolateral part of the AR surrounding the third ventricle (V) . x 150.

Fig. 14. A high-power photomicrograph shows the same neuron (N) as in Fig. 13. The peri- karyon is densely labeled, and located nearby a blood capillary (b) with mildly positive endothelial cells (arrows e) . x 1,050.

Fig. 15. A high-power photomicrograph exhibits an intensely labeled fusiform neuron (N) of the ventromedial nucleus (VM) of case 19. x 1,500.

Fig. 16. A moderately labeled neuron (arrow) of the supramamillary nucleus (sum) of case 36 is recognizable dorsal to the mammillothalamic fascicle (FMT) . Several blood capillaries (b) are clearly seen. x 150.

Fig. 17. An intensely labeled neuron (arrow) of the A10 cell group in case 19 is observable just dorsal to the interpeduncular nucleus (IP) . x 600. 439 Plate II

K. lijima 440 K. Iijimr

Plate III

Fig. 18. A medium-power photomicrograph shows moderately (upper, line N) and mildly (lower, line N) labeled fusiform neurons in the medial septal nucleus (sm) of case 25. x 300.

Fig. 19. An intensely labeled neuron (N) is located in the dorsal raphe nucleus (dr) of case 35, which is mediodorsal to the medial longitudinal fascicles (FLM) on both sides. x 300.

Fig. 20. A moderately labeled multipolar LC neuron is shown at a medium magnification (arrow) . The dorsal division of the LC is surrounded by the mesencephalic tract nucleus of the trigeminal nerve (MT) and the superior cerebellar peduncle (Pcs). Blood capillaries (b) are seen in the LC. case 35. x 300.

Fig. 21. A high-power photomicrograph shows dense labeling in a pyramidal cell (arrow P) as well as in a cluster of granule cells (white arrows), of the hippocampus proper of case 26. x 600.

Figs. 22-24. All the figures exhibit the results obtained from the control cases where HRP was injected outside the SO.

Fig. 22. An intensely labeled multipolar neuron (arrow) of the olfactory tubercle (TULI) is seen inside the cortical part of it (TUL) in case 34. The letter b indicates a blood capil- lary. x 150.

Fig. 23. Some intensely labeled neurons of various shapes are seen in the diagonal band nucleus of case 2. x 150.

Fig. 24. A low-power photomicrograph shows that 4 intensely labeled neurons of fusiform and multipolar shapes (white arrows) are located in the dorsal division (d) but not in the ventral division (v) of the LC (indicated by arrows) which is surrounded by the mesence- phalic tract nucleus of the trigeminal nerve (MT), superior cerebellar peduncle (Pcs) pontine central gray (PG) and the fourth ventricle (V) . Case 3. x 150. .0 441 Plate III

K. lijima