THE JOURNAL OF COMPARATIVE NEUROLOGY 314:478492 (1991)

The Distribution of Neuropeptides in the Dorsomedial Telencephalon of the Pigeon (Columba livia):A Basis for Regional Subdivisions JONATHAN T. ERICHSEN, VERNER P. BINGMAN, AND JOHN R. KREBS Department of Neurobiology and Behavior, SUNY at Stony Brook, Stony Brook, New York 11794 (J.T.E.);Department of Psychology, Bowling Green State University, Bowling Green, Ohio 43403 (V.P.B.);and Edward Grey Institute of Field Ornithology, Department of Zoology, Oxford OX1 3PS, England (J.R.K.)

ABSTRACT The distribution of six neuropeptides [substance P (SP), leucine (leu5-) enkephalin (LENK), vasoactive intestinal polypeptide (VIP), cholecystokinin (CCK), neuropeptide Y (NPY), and somatostatin (SS)]in the dorsomedial telencephalon (hippocampal region) of the pigeon was studied by immunohistochemistry. AH six peptides were found in fibers passing through the septo-hippocampal junction and along the medial wall of the hippocampal region. NPY-, SS-, and VIP-like staining of fibers was seen in the hippocampal commissure. NPY and SS had similar distributions within the hippocampal region, both being most conspicuous in cell bodies, terminals, and fibers of the medial hippocampal region. VIP-positive cells were found in an area dorsal to the SSiNPY cell region. CCK-like immunoreactivity was found in terminal baskets surrounding large cells of a v-shaped structure in the ventromedial hippocampal region. SP- and LENK-like immunoreactivity was found in neuropils in a lateral-dorsal region, the two substances showing similar distributions. This region is thought to lie lateral to the limit of the hippocampal region. Parallels with the distribution of immunoreactivity in' the mammalian hippocampus are used to suggest possible equivalent subdivisions of the avian and mammalian hippocampal regions. Key words: avian hippocampal region, immunohistochemistry, substance P, cholecystokinin, leucine (leu5-)enkephalin, vasoactive intestinal polypeptide, neuropeptide Y, somatostatin

In this and a companion paper (Krebs et al., '911, we use immunohistochemical methods to characterize the boundaries and subdivisions of the pigeon hippocampal region (dorsomedial telencephalon). In addition, we attempt to draw parallels between birds and mammals in the distribution of neuroactive substances in this area and use these parallels to formulate hypotheses about the correlation between different parts of the hippocampal region. Our interest in the avian hippocampal region stems from recent evidence that it may play a similar functional role to the mammalian hippocampus in the processing of spatial memories (Bingman et al., '84, '88a,b; Krebs et al., '89; Sherry and Vaccarino, '89). In a previous paper (Krebs et al., '91), we examined the distribution of a transmitter (serotonin)and three transmitter-related enzymes (choline acetyltransferase, glutamic acid decarboxylase, and tyrosine hydroxylase) in the hippo-

O

1991 WILEY-LISS, INC.

campal region of the pigeon. The results of this study led to the suggestion that the hippocampal region could be divided into at least four areas: a medial fiber tract passing through the septo-hippocampal junction and extending dorsally, a dorsomedial area rich in terminals and neuropil, a ventral v-shaped area of large cells (V), and an area between the arms of the V. A dorsolateral area, characterized by tyrosine hydroxylase-immunoreactive basket-like terminations, was hypothesized to be outside the hippocampal region. The serotonergic, cholinergic, and catecholaminergic fibers of the medial tract suggested that this could be equivalent to the mammalian alveus or fimbria-fornix, but further discussion about possible correspondences with mammals was deferred until the results of the present study could also be considered. Accepted September 4,1991.

479

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON Here we examine the distribution of six neuropeptides: substance P (SP),leucine (leu5-)enkephalin (LENK),vasoactive intestinal polypeptide (VIP), cholecystokinin (CCK), neuropeptide Y (NPY), and somatostatin ( S S ) . Although the exact function of neuropeptides in interneuronal communication is still not fully understood, they have proved to be useful markers to characterize different structures in the central nervous system including components of the mammalian hippocampal formation. Immunoreactive staining patterns must be interpreted with caution because the distribution of some peptides is variable among species (e.g., Gall et al., ’86). Nevertheless, certain generalizations can be made, both about the distribution of peptides within the hippocampal formation and their colocalization with each other and with well-established neurotransmitters, which allow a comparison between birds and mammals. The only previous detailed study of neuropeptides in the dorsomedial telencephalon of the pigeon is the immunohistochemical analysis of the avian wulst by Shimizu and Karten (’90).

A

4 I

3

2

1

I

1

I

B

0

16

I

A 9.00

I

I

I

A 7.50

METHODS The brains of four homing pigeons (Columba liuia) of both sexes were processed for immunohistochemistry. The perfusion, fixation, sectioning, and immunohistochemical methods were identical to those presented in a companion paper (Krebs et al., ’91).

Antibodies The range of primary antisera dilutions used was 15001:2,000. Unless otherwise noted, optimal staining patterns were obtained at a dilution of 1:1,000. All antibodies were diluted in 0.1 M phosphate buffer (pH 7.4) containing 0.3% Triton X-100 (see Krebs et al., ’91). The localization of substance P (SP) was carried out with a monoclonal antibody (IgG from rat) directed against the carboxyl terminal fragment of substance P (MAS035b; Sera-Lab) that has been well characterized (Cuello et al., ’79, ’80; Mai et al., ’86). A monoclonal antibody (I& from mouse) directed against leucine (leu5-) enkephalin (LENK) (MAS083; Sera-Lab) was used. This antibody cross-reacts with met5-enkephalin and some other forms of enkephalin but does not bind to brain areas known to contain dynorphin or beta-endorphin (Cuello, ’84). Somatostatin (SS)like immunoreactivity was localized with a polyclonal antiserum (made in rabbit) (20067; INCSTAR) that has been used in the mammalian hippocampus (e.g., Kohler and Chan-Palay, ’82a) and in other regions in other vertebrates (e.g., Ellis et al., ’83).The antiserum to neuropeptide Y (NPY) (made in rabbit) (RAS7172N; Peninsula) was found to yield the best immunolabeling at a dilution of 1:1,500. Immunoreactivity for the octapeptide form of cholecystokinin (CCK) was localized with a rabbit antiserum (20078; INCSTAR) which has been used in the mammalian hippocampus (e.g., Kohler and Chan-Palay, ’82b; Gall, ’88; Gall

Abbreviations APH CDL

HP HA TSM

area parahippocampalis area corticoidea dorsolateralis hippocampus hyperstriatum accessorium tractus septomesencephalicus

Fig. 1. A schematic of four transverse sections of the left avian hippocampal region indicating the arbitrary areas used for describing the distribution of immunoreactivity (see text). D, dorsal region; DMs, superior part of dorsomedial region; DMi, inferior part of dorsomedial region; VM, ventromedial region. The sections are from the atlas of Karten and Hodos (’67) at A9.00, A7.50, A6.00 and A3.75.

et al., ’87). Vasoactive intestinal polypeptide (VIP)-like immunoreactivity was localized with a rabbit antiserum (7916) supplied by Dr. J.H. Walsh (UCLA). The methods used for mapping the distribution of immunoreactive staining are identical to those described in a companion paper (Krebs et al., ’91).Here we simply identify the major subdivisions used for descriptive purposes and refer the reader to the companion paper for further details and discussion. The major subdivisions used in our description of immunoreactive staining are: ventromedial (VM), including the v-shaped structure of cells that stain heavily for Nissl (i.e., the “V”), dorsomedial (DM), and dorsal (D). As indicated in Figure 1, DM is further subdivided into superior (DMs) and inferior (DMi) regions.

RESULTS Substance P (SP) SP-like immunoreactivity is seen in fibers in the medial fiber tract and passing, in part, through the septohippocampal junction (Fig. 2A). [The medial fiber tract was defined by Krebs et al. (’91) as the fiber tract passing through the septo-hippocampal junction and adjacent to the medial wall, extending onto the dorsal surface where it is called the “dorsal fiber tract.”] In the caudal-most section (A3.751, these fibers are in an oblique plane (Fig.

480

J.T. ERICHSEN ET AL.

Fig. 2. Photomicrographs of peptidergic fibers in the medial fiber tract of the hippocampal region: (A) SP-immunoreactive beaded fiber (A9.00). (B) SP-immunoreactive fibers oriented in an oblique plane a t A3.75 (indicated by arrowheads). (C) SS-immunoreactivefibers (A9.00).

(D) LENK-immunoreactive fiber (A8.50). (E) NPY-immunoreactive fibers (A9.75). (F) VIP-immunoreactive fibers (also in VM; arrowhead) (A8.00). Bar: 50 pm.

2B). The fibers of the medial tract appear to terminate in a restricted area close to the medial wall of DMi (A9.00) or DMs (more caudal portions) (Fig. 3A-D). Additional labeled fibers are seen in A7.50 and A6.00 adjacent to the dorsal wall of the ventricle. In the most rostra1 section (A9.001,SP-like immunoreactivity is found in an area of neuropil lying in D near the boundary with DM. The neuropil, which surrounds small unlabeled cells, varies in density of labeling, being most

densely labeled dorsomedially close to DM and less densely as one moves laterally. In sections A7.50 and A6.00, the labeled neuropil area in D appears to move laterally. In its medial portion, it has a characteristic shape and variation in intensity of staining which gives it a layered appearance (see Fig. 5A). From dorsal to ventral, the staining is dense in the first layer, absent or light in the second layer, and intermediate in the third layer. A second area of labeled neuropil, referred to in the previous paragraph as the

481

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON

B

A

C

I

I

D

I

I

A 6.00

I

1

1

I

A 3.75

\ Fig. 3. (A-D) Schematic representation of the distribution of SPand SS-like immunoreactivity at four transverse sections of the hippocampus, beginning at the rostra1 end and moving caudally (see Fig. 1). SP-positive neuropil is represented by large dots (intense staining) and small dots (light staining); the approximate locations of SP cells are

indicated by stars. Similarly, two levels of intensity of staining for SS are indicated by heavy and light shading with irregular lines. In the case of SS, cells are not indicated, although most are found within the area of shading correspondingto fibers and terminations.

482

J.T. ERICHSEN ET AL.

presumptive site of terminations of medial fibers, lies in DMi most rostrally (A9.00) but otherwise in DMs. Throughout the rostral-caudal extent of the hippocampal region, sparsely scattered labeled small cells are seen (see Fig. 5B). They are rarely found in the areas of dense neuropil but occur primarily in DMi.

Somatostatin (SS) SS-like immunoreactivity is seen in many fibers that pass through the septo-hippocampal junction in the medial fiber tract and extend into the dorsal tract (Fig. 2C). In the caudal-most section (A3.751, the medial fibers are oblique in their orientation. SS is also found in some fibers of the hippocampal commissure (see Fig. 7A). SS-like immunoreactivity is found in the cytoplasm of a polymorphic population of medium to large cells in VM and DMi (Fig. 3A-D) which includes, as with NPY (see below), triangular, multipolar, stellate, and bipolar cells (see Fig. 6A). Some dendrites of the triangular cells appear to be regularly arranged at right angles to the ventricular wall (see Fig. 6B). SS-like labeling is also found in both DMs and DMi in terminations around medium to small-sized cells (see Fig. 6A inset). These are not labeled with NPY (see below).

Leucine enkephalin (LENK) At A9.00-A7.50, LENK-like immunoreactivity is found in a very sparse population of fibers which lies in the medial fiber tract (Fig. 2D) and continues adjacent to the dorsal surface, eventually entering a diffuse area of labeled neuropi1 in D. Adjacent to the medial wall there are short, heavily beaded stained fibers oriented at right angles to the wall. The area of lightly stained neuropil in D is similar in location, shape, and extent to the lateral SP field described above. Like the SP field, it is layered in its medial region in A7.50-A3.75, with an invagination of an unstained area separating two stained areas (Fig. 4A-D). In VM, DMi, and in D along the ventricular surface, there are sparsely scattered labeled terminals and fibers, often completely surrounding unstained cells, especially the large cells of the V. These terminals and fibers are absent from DMs. Very occasionally, lightly stained small to medium-sized multipolar cells are found in VM and DM in A9.00 and A7.50. Similar cells are also found in the V at A7.50 (Fig. 5 0 .

Neuropeptide Y (NPY) The overall distribution of NPY-like immunoreactivity is very similar to that of SS, although the labeling of cells and terminals extends farther laterally. NPY labeling is seen in many fibers which pass through the septo-hippocampal junction and lie in the medial and dorsal fiber tracts (Fig. 2E). Labeling is also found in some of the fibers of the hippocampal commissure (see Fig. 7B). The population of large to medium cells in VM and DMi labeled with NPY-like immunoreactivity includes triangular, multipolar, stellate, and bipolar cells (Fig. 6C). In all but the most rostral section (A9.00, where the distribution is more diffuse), the cells are restricted to an elongated band extending dorsally and laterally from VM and DMi into D. Between the arms of the V itself (in VM), the cells tend to be smaller. The dendrites of the triangular cells are, as with SS, oriented perpendicularly to the wall of the ventricle (Fig. 6D). Throughout the area in which the cells are distributed, there are also labeled terminals and beaded fibers. These labeled terminals extend laterally into D, in a

continuous band in A9.00 and A7.50, but in a discontinuous distribution in A6.00 and A3.75 (Fig. 4A-D).

Vasoactive intestinal polypeptide (VIP) VIP-like immunoreactivity is seen in fibers that pass through the septo-hippocampal junction and in the medial fiber tract (Fig. 2F) as well as in fine beaded fibers scattered throughout VM and DM. Some of these scattered fibers appear to terminate on cell bodies. In the caudal-most section (A3.75), the medial fiber tract is oriented in an oblique plane. VIP-like immunoreactivity is also seen in fibers of the hippocampal commissure (Fig. 7C). In A7.50 and A6.00, there is a small field of scattered terminals in D (Fig. 8A-D). This field is absent inA3.75 but is replaced by a restricted area of terminals close to the dorsal surface of the ventricle in D. In all but the most rostral section, DMs contains a sparse population of medium-sized cells with lightly stained cytoplasm and occasionally with surrounding labeled terminals (Fig. 5D). This cell population lies medial to the field of scattered VIP-positive terminals in D (Fig. 8A-D).

Cholecystokinin (CCK) CCK-like immunoreactivity is seen occasionally in fibers of the medial fiber tract in A7.50. The most conspicuous feature of the distribution of CCK-like immunoreactivity in all except the most caudal section is the occurrence of labeled basket-like terminals surrounding cells of the V (Fig. SA,C-E). In addition, there is a second population of labeled, scattered terminals and occasional baskets in VM and DMi but not in DMs. These are more abundant between the arms of the V than in more dorsal areas. In sections A750 and A6.00, there is a more densely labeled terminal field with larger bouton-like varicosities (Fig. 9B) adjacent to the dorsal surface of the ventricle in D, the lateral portion of which extends dorsally. In the caudal-most section (A 3.79, the medial part of this area of densely stained neuropil extends almost to the medial wall in DMi and laterally into D, but the labeled neuropil is conspicuously absent from DMs.

DISCUSSION The major results of the present study can be summarized as follows: a) The medial and dorsal fiber tracts contain fibers showing SP-, SS-, LENK-, VIP-, CCK-, and NPY-like immunoreactivity. b) NPY-, SS-, and VIP-like staining are found in fibers of the hippocampal commissure. c ) CCK-like immunoreactivity is seen in terminal baskets surrounding the large cells of the V. d) Regions of neuropil exhibiting SP- and LENK-like immunoreactivity have a similar distribution in lateral D. e) VIP and SP are found in neuropil in dorsal DMs. f, NPY and SS show similar distributions, both being found in cell bodies, terminals, and fibers in VM between the arms of the V and extending dorsally and laterally into DMs and medial D. g) VIPpositive cells are scattered in a restricted region in ventral DMs, somewhat dorsal to the area containing SS- and NPY-labeled cells.

Subdivisions of the avian hippocampal region In the study of Krebs et al. ('91), the distribution of serotonin (5HT),tyrosine hydroxylase (TH), choline acetyltransferase ( C U T ) ,and glutarnic acid decarboxylase (GAD) suggested four distinct subdivisions of the hippocampal

483

SUBDIVISIONS OF PIGEON DORSOMEDJAL TELENCEPHALON

8

A I

1

I

I

I

L Fig. 4. (A-D) Schematicrepresentation of the distribution of LENKand NPY-like immunoreactivity. LENK-containing neuropil is indicated at two levels of intensity of staining: large dots for heavier and

I

I

A 7.50

A 9.00

A 6.00

I

A 3.75

small dots for lighter staining. The approximate positions of LENKpositive cells are indicated by stars. NPY details are as for SS in Figure 3 (A-D).

484

J.T. ERICHSEN ET AL. region as determined by the presence or absence of immunoreactive staining: a medial fiber tract, a dorsomedial area of terminals and neuropil, a v-shaped area of large cells staining heavily for Nissl, and a ventral area including VM between the arms of the V and DMi. Krebs et al. also identified a lateral region of D containing basket-like terminations immunoreactive for TH, and they suggested that this might lie beyond the lateral boundary of the hippocampal region. The distributions of the six peptides examined here are consistent with this scheme but also suggest the identification of further subdivisions. On the basis of both the present results and those of Krebs et al., the following six subdivisions of the hippocampal region of the pigeon are proposed (Fig. 10): The medial fiber tract described in Krebs et al. (’91)as containing ChAT-, 5HT-, and TH-labeled fibers, which also contains fibers immunoreactive for SP, CCK, LENK, VIP, NPY, and SS. Whereas Krebs et al. were able to suggest that the cholinergic, serotonergic, and catecholaminergic fibers were probably afferents, it is not possible to identify the peptidergic fibers as afferent or efferent. The v-shaped area of cells which stain heavily for Nissl. Some of these cells are pyramidal cells as revealed by Golgi study (Pisana, ’86; Smith, ’891, and we show here that many of them are surrounded by CCK-labeledbaskets. A region in VM, DMi, ventral DMs, and ventromedial D containing NPY- and SS-positive cells, fibers, and terminations (“the medial area”). A region of DMs, partly within that referred to in (31, defined by the presence of VIP-positive cells (the “VIP-cell area”). The localized area of neuropil in DMs and medial D identified by Krebs et al. as containing ChAT, 5HT, and TH and shown here also to contain SP- and VIP-positive neuropil (the “dorsomedial area”). An area of diverse terminal fields in medial D which contains neuropil immunoreactive for CCK, 5HT, VIP, and NPY. This region lies lateral to area (3) (the “lateral area”) and medial to a seventh region discussed below. Figure 10 also indicates a seventh region, in the lateral portion of D, identified by Krebs et al. as containing TH-positive basket-like terminals around cells and here identified by prominent neuropil staining for SP and LENK. The distributions of these two peptides are almost identical, and they overlap extensively with the region of TH-labeled basket-like terminals. However, the medial boundary of the SP and LENK neuropil has a marked invagination, which is not apparent in the distribution of TH labeling. This area (71, which Krebs et al. suggested might be outside the hippocampal region, will be discussed further in the section below. Shimizu and Karten (’90) do not analyze the medial and ventral portions of the area we refer to as the hippocampal region, but their data for more dorsal and lateral areas (SP,

Fig. 5. Photomicrographs of (A) SP-immunoreactive neuropil in left D (A8.00);(B)SP labeling of cell in DMi (A8.00); (C) LENK labeling of cell in the medial arm of the V (A7.00); (D) VIP-positive cell in DMi (A8.00). Bars, 100 pm (A);25 pm (B-D).

485

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON

Fig. 6. Photomicrographs of (A) SS-labeled cells in DMi (A7.25); inset: SS-labeled cells (indicated by arrowheads) in lateral VM with surrounding heavily labeled bouton-like varicosities (A7.25); (B) SSlabeled cells with oriented dendrites in left DMi (A7.25); ( C ) NPY-

labeled cells in DMi (A7.00); (D) NPY-labeled cells with oriented dendrites in right DMi (A7.25). Bars: 50 Frn (A,C); 100 pm (B,D); 20 pm (inset).

LENK, and TH) appear to correspond with our results (see their Figs. 9, 11,and 5 , panels D and E).

ary that we propose to use for defining the lateral limit of the hippocampal region throughout most of its rostralcaudal extent. In addition. the Dattern of staining in HA displaces the hippocampal region anteriorly between the ventricle and the medial wall, completely occupying this areaat A11.00-A11.50. This immunohistochemically characterized lateral boundary corresponds with the boundary identified by Krebs et al. ('89) in Nissl material of several avian passerines, on the basis of a change in relative cell sizes. However, Shimizu and Karten ('90) call the medial portion of the SP- and LENK-positive neuropil APH (subdivided into M and D). Compare, for example, their Figure 8D and E with our Figure 3A and B. The basis for their identification of this region as APH is unclear, but further work is needed to clarify this issue. Using these boundaries, the hippocampal region first appears at A1l.OO-A11.50 as a small area located ventral to the SP-positive field in HA between the ventricle and the medial wall. At this point, the ventral boundary of the hippocampal region is difficult to define but is characterized by a narrow, virtually cell-free zone, which in more caudal sections merges into the septum. Moving caudally, the hippocampal region enlarges dorsally and then laterally. According to this view, the V of densely Nissl-stained cells

Boundaries of the hippocampal region The focus of our consideration of the boundaries of the hippocampal region will be the lateral and anterior limits which remained unspecified in the atlas of Karten and Hodos ('67). Rostra1 to the V of the hippocampal region (i.e., rostral to A10.00), there is a densely stained, SP-positive neuropil that fills the anterior forebrain region of hyperstriatum accessorium (HA), as identified in the Karten and Hodos ('67) atlas. This pattern of staining persists caudally in a more lateral position and appears to be continuous with the SP-positive neuropil in lateral D through at least A3.50. Whether or not this immunoreactive neuropil can be considered a marker for HA in more caudal sections of the brain, it appears to define an area that is not in the hippocampal region. Therefore, the medial boundary of the SP-positive staining in lateral D might be used as a conservative marker for defining the lateral limit of the hippocampal region. The distributions of the lateral-most LENK- and TH- (see Krebs et al., '91) labeled neuropil correspond well with the location of this SP-containing field. Therefore, taken together, these three substances present a consistent bound-

-

486

J.T. ERICHSEN ET AL.

Fig. 7. Photomicrographs of fibers in the hippocampal commissure that are immunoreactive for (A) SS (A7.25); (B)NPY (A7.00); (C) VIP (A7.25). Bars: 50 pm (A,B); 25 K r n (C).

appears caudal to the first emergence of the hippocampal region.

Comparison with the mammalian hippocampal formation In this section, we use the immunohistochemical evidence from both the present study and from that of Krebs et al. ('91) to consider whether the subdivisions of the pigeon hippocampal region proposed in the previous section

can be associated with any areas of the mammalian hippocampal formation. In making these comparisons, we refer to a "typical mammalian pattern," but it should be borne in mind that the distribution of neuropeptides varies among mammalian species (e.g., Gall et al., '86). Our results suggest that parallels can be drawn between four of our subdivisions and parts of the mammalian hippocampal formation. First, the occurrence of CCK-like immunoreactivity in baskets surrounding the large cells of the V (both in its lateral and medial arms) (area (2) in the scheme above) resembles the findings in mammals where similar CCKlabeled baskets surround pyramidal cells of Ammon's Horn (Nunzi et al., '85). A Golgi study of the cell types of the V in two species of passerine birds (Parus major and Parus palustrzsj (Smith, '89) has shown that this region contains pyramidal cells similar in morphology to mammalian pyramidal cells with oriented dendritic fields perpendicular to the arms of the V (i.e., perpendicular to the ventricular wall in the lateral arm and to the medial wall in the medial arm) (see also Pisana, '86). In mammals, almost all CCK-positive cells are also GABAergic,but only a minority of GABAergic cells contain CCK (Somogyi et al., '84; Kosaka et al., '85). Thus, it is not certain whether the baskets surrounding the cells of the V are GABAergic. Second, the large to medium polymorphic cells immunoreactive for SS and NPY in the medial area (area 3) parallel the situation in the mammalian hilar region (Amaral and Campbell, '86; Kohler et al., '86). These two peptides colocalizein mammals (Kohler et al., '86) and extensively in birds, including the hippocampus (Anderson and Reiner, 'go), so it is almost certain that the same cell population is labeled by both antibodies in our study. The similarities of the NPY- and SS-positive cells, in terms of their morphology and their regional location, support this view. Golgi impregnation reveals that this area contains a polymorphic population of cells, predominantly multipolar spiny cells (Smith, '89). Third, the occurrence of VIP-like immunoreactivity in cells of area (4)parallels the findings of Kohler ('83)for the mammalian granule and superficial hilar region of the dentate gyrus. This result, together with the location of SS-/NPY-positive cells in the medial area, leads to the hypothesis that the medial area and VIP-cell areas correspond, respectively, to the hilar and granule layers of the mammalian dentate gyrus. Neither Nissl staining nor Golgi staining indicates conspicuous layering of cells in the VIP-cell area. Therefore, if the avian granule cell population is contained in this area, as we suggest, the cells may be much less well organized than in mammals. Finally, Krebs et al. ('91) concluded that the medial fiber tract (area l),which appears to contain catecholaminergic, serotonergic, and cholinergic afferents to the hippocampal region, resembles the alvear layer and fimbria-fornix of the mammalian hippocampus. The dorsal continuation of this fiber tract, i.e., the dorsal fiber tract, may contain afferents and efferent connections to other forebrain regions (Casini et al., '86). The medial fiber tract also contains fibers that are immunoreactive for each of the six peptides examined here. Five of the six peptides (i.e., NPY, SS, VIP, LENK, and SP) were found in cells of the hippocampal region, so for these cells the fibers could be either afferent or efferent. In the case of CCK, the absence of labeled cells in the hippocampal region might suggest that the fibers of the medial tract are afferents. However, we cannot exclude the

487

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON

I B

A

I

A 9.00

I

1

I

A 7.50

A 3.75

Fig. 8. (A-D) Schematic representation of the distribution of VIPand CCK-like immunoreactivity in the hippocampal region at four levels. CCK-positive baskets and terminations are indicated by large

dots (intense staining) and small dots (light staining). The general location of VIP-positive cells is indicated by stars, and the area of VIP-immunoreactiveterminals is indicated by irregular stippling.

Figure 9

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON possibility that colchicine pretreatment might reveal CCKlabeled cells. To summarize, we suggest the following hypothesis concerning parallels between the bird and the mammal: the medial fiber tract area corresponds to the alveus (see also Pisana, '86), the V area to Ammon's Horn, and the medial and VIP-cell areas to the hilar and granule layers of the dentate gyrus, respectively. We now turn to a consideration of the remaining subdivisions of the avian hippocampal region. The dorsomedial area (area 5), the localized area of neuropil immunoreactive for ChAT, TH, 5HT, and SP, appears to be the site of the terminations of some of the ascending d e r e n t s . This would be consistent with our hypothesis because this neuropil area is adjacent to, or perhaps part of, our suggested granule cell area in the avian hippocampal region. In the mammalian hippocampal formation, many afferents terminate on the granule cells (Amaral and Campbell, '861, and a parallel situation may exist in birds. As we have already suggested in the previous section, area (7) lies outside (i.e., lateral and anterior to) the area we are defining as the hippocampal region, and we have proposed the SP-containing field of neuropil as a boundary marker. In rostra1 forebrain, this field fills HA, and more caudally, the labeled neuropil moves laterally into CDL (as defined by the atlas of Karten and Hodos, '67). In mammals, the entorhinal cortex is immunoreactive for SP as well as LENK (Amaral and Campbell, '861, suggesting a possible correspondence to this field in birds. However, as mentioned above, the SP-stained neuropil lateral to the avian hippocampal region in more caudal sections appears to be continuous with the label found in HA more rostrally. Further, Shimizu and Karten ('90) call the medial portion of this neuropil APH. Thus, connectivity studies will probably be required to address this issue in more detail. Finally, if our scheme of equivalence with mammals is followed, the lateral area (area 61, which lies between the medial area and area (71, would correspond to the subiculum. The major missing feature of our proposed scheme is the lack of evidence for a mossy fiber system projecting from the granule cells to the pyramidal layer. In some mammalian species, the mossy fibers are immunoreactive for LENK (Gall et al., '86) and/or for dynorphin (StengaardPederson et al., '83; Gall et al., '86), neither of which reveal a mossy fiber system in birds. In the mammalian hippocampus, the mossy fiber system is revealed by Timm's stain for zinc (Geneser-Jensen et al., '741, and a parallel pattern of intense staining of a fiber tract is seen in the medial and dorsomedial cortex of lizards (Olucha et al., '881, suggesting a possible equivalent to the mossy fiber system. In the canary (Serinuscanaria), preliminary results with a modified Timm procedure (Danscher, '81, '82; Danscher and Krebs, unpublished data) show a different pattern of staining. There is more staining in the hippocampal region than in most of the rest of the telencephd o n , but the staining is not localized to a discrete zone within the hippocampal region. The two areas of highest staining intensity within the hippocampal region are i) the

Fig. 9. Photomicrographs of (A) CCK-immunoreactive baskets surrounding large cells in the left medial V: a CCK-positive fiber is also visible in the medial fiher tract (A7.75); (B)CCK-like immunoreactivity in varicosities located in medial D (A8.50); (C-E) CCK-positive basket terminals on cells in the lateral V. Bars: 50 pm (A,B);25 p,m (C-E).

489

medial area, and ii) between the lateral arm of the V and the ventricle wall. Further work is required to establish the sites of origin and termination of these fibers. A further difference between the present results and those reported for mammals is the occurrence of VIP-like immunoreactivity in afferentiefferent fibers passing through the septo-hippocampal junction and in fibers of the hippocampal commissure (Kohler, '83). On the other hand, NPY/SS-like immunoreactivity in commissural fibers resembles results in mammals (Kohler et al., '86).

Comparison with earlier models of the avian hippocampal region Earlier workers (Elliot Smith, '10; Rose, '14; Craigie, '30, '35, '40; Showers, '82) attempted to subdivide the avian hippocampal region on the basis of studies of Nissl-stained material. In comparing the results of these earlier studies with our present hypothesis based on immunohistochemical data, it is worth noting that there may be differences between taxonomic groups of birds (Craigie, '40). Craigie ('40) concluded, on the basis of studies of 24 species of birds (including the pigeon) belonging to 16 orders, that the hippocampal region consists of a medial (hippocampal) cortex corresponding roughly to our V and medial areas and a dorsal (parahippocampal) cortex corresponding to our VIP-cell, dorsomedial and lateral areas. Each of these areas is subdivided by Craigie into about four vertical (ventral-dorsal) regions. The v-shaped structure is divided into a ventral portion where the cells are compressed into a single layer, a slightly more dorsal region with a triangular cross section and two further areas in which the V expands dorsally and laterally. The present immunohistochemical evidence suggests no obvious divisions of the V into these four vertical regions. The parahippocampal region is also subdivided by Craigie into three or four regions on a dorsolateral axis. Although Craigie's divisions were based on the layering and types of cells, they may well correspond to our VIP-cell, dorsomedial and lateral areas. Pisana ('86) provides an extensive review of the theories regarding possible homologies between subdivisions of the avian and mammalian hippocampal regions. Although the details vary considerably, some authors have proposed that the v-shaped structure corresponds to the dentate gyrus (e.g., Showers, '821, while others, noting the presence of pyramidal cells in the V, argue that it corresponds at least in part to Ammon's Horn (e.g., Pisana, '86). The more dorsal region (our VIP-cell, dorsomedial and lateral areas) have been equated with entorhinal cortex (Rose, '14) and with an incipient Ammon's Horn (Elliot Smith, '10; Showers, '82).Thus, our present hypothesis differs at least in part from that of the classical neuroanatomists, although the lack of consensus among earlier workers makes it difficult to draw clear conclusions.

Comparison of birds and reptiles The neuroanatomy of the reptilian cortex has been primarily studied in Squamata (lizards and snakes). The major conclusions of these studies that are pertinent to the present study are as follows: a) the Squamate cortex is divided into four regions (medial, dorsomedial, dorsal, and lateral), each with a three layered organization (Northcutt '67, '81; Ulinski, '75; Bruce and Butler, '84); and, b) the medial cortex and dorsomedial cortex are conjectured to correspond to, respectively, the dentate gyrus and Am-

490

J.T. ERICHSEN ET AL.

1

s

A 9.00

2

A 7.50

C

1 A 6.00

+

A 3.75

Fig. 10. (A-D) A schematic to show the proposed subdivisions of the avian hippocampal region based on immunohistochemicalevidence (see text for further explanation).

-

491

SUBDIVISIONS OF PIGEON DORSOMEDIAL TELENCEPHALON mon’s Horn of the mammalian hippocampus. The evidence for this second point comes from studies of both local connectivity and cell morphology. The medial cortex is the origin of Timm-positive fibers that project to dorsomedial cortex (Olucha et al., ’88), thought to correspond to the mammalian mossy fibers. The characteristic cells of the dorsomedial cortex are spiny pyramidal cells, thought to correspond to pyramidal cells of Ammon’s Horn (Martinez-Guijarro et al., ’84). On the other hand, the medial cortex contains, in its superficial cell layers, three cell types (sparsely spiny horizontal cells, sparsely spiny pyramidal cells, and spiny bitufted cells), thought to be related to granule cells of the dentate gyrus, while the deep cell layer contains large spiny polymorphic neurons, corresponding to multipolar cells of the hilar region (Berbel et al., ’87). If this framework were applied to the avian dorsomedial cortex, it would suggest that the V and medial area correspond to the dentate gyrus, while the VIP-cell, dorsomedial and lateral areas would correspond to Ammon’s Horn, the opposite of the hypothesis we have suggested above. However, the general organization of the Squamate hippocampal region is very different from that of birds, in terms of the layering of cells, the morphology of cell types and the organization of the putative mossy fiber system (see above). Thus, it may not be surprising to see that our hypothesis for the avian hippocampal region does not apply to Squamate reptiles. In phylogenetic terms, the Squamates are not close to the bird line of evolution, whereas the crocodiles (archosaurs) are. In this context, it is interesting to note Crosby’s (’17) conclusion that the organization of the cortex of the alligator corresponds more closely to our model for the bird than either of these to the Squamates. As Pisana (’86) points out, the early authors disagree about whether the v-shaped structure of the crocodilian hippocampus corresponds to the mammalian dentate gyrus or to Ammon’s Horn in much the same way as discussed above for birds.

Conclusion Our present work suggests an hypothesis for the subdivisions of the pigeon hippocampal formation and the relationship between these subdivisions and those of the “typical mammal.” Several approaches can be suggested for a further investigation of the hypothesis: a) analysis of cytoarchitecture in other planes of section; b) further characterization of cell types, in particular pyramidal and granule cells; c) additional immunohistochemical analysis including colocalization studies; and d) studies of intrinsic and extrinsic connections of the putative subdivisions proposed here. The longer term aim of the work is to provide a better understanding of the anatomical organization of the avian hippocampal region, with a view towards identifying more precisely the functional role in memory of different areas within the hippocampal region. For example, future studies might involve the localization of synaptic plasticity, which could form the substrate for memory formation, to specific subdivisions of the avian hippocampus identified here. In particular, the striking observation that species of songbirds, which store and retrieve food using an accurate spatial memory, have an enlarged hippocampal region (Krebs et al., ’89; Sherry et al., ’89) may be further elucidated by a better understanding of avian hippocampal anatomy.

ACKNOWLEDGMENTS We thank Angela K. Levine and Anne F. Bushnell for technical help and the following for financial support: NIH grant EY04587 (JTE), NSF grant BNS8611204 (VPB), and SERC and Royal Society (JRK). Dr. Peter Somogyi was most helpful in discussing the interpretation of the results and in commenting on an earlier draft of the manuscript. Dr. Hugh Perry helped in the early phases of the project and commented on the manuscript. Dr. P. Bagnoli, Dr. C. Gall, Dr. R.O. Kuljis, Dr. R.Y. Moore, and Dr. A. Reiner commented on a draft of the manuscript. Part of the work was done while JRK was a Visiting Professor in the Departments of Ecology and Evolution and Neurobiology and Behavior at SUNY, Stony Brook. Dr. Jeff Levinton and Dr. Lorne Mendell are gratefully acknowledged for their hospitality.

LITERATURE CITED Amaral, D.G., and M.J. Campbell (19861 Transmitter systems in the primate dentate gyms. Human Neurobiol. 5t169-180. Anderson, K.D., and A. Reiner (19901 Distribution and relative abundance of neurons in the pigeon forebrain containing somatostatin, neuropeptide Y, or both. J. Comp. Neurol. 299:261-282. Berbel, P.J., Martinez-Guijarro, and C. Lopez-Garcia (1987) Intrinsic organization of the medial cerebral cortex of the lizard Lacertu pityusensis: A Golgi study. J. Morphol. 194275-286. Bingman, V.P., P. Bagnoli, P. Ioale, and G. Casini (1984) Homingbehavior of pigeons after telencephalic ablations. Brain Behav. Evol. 24:94-108. Bingman, V.P., P. Ioale, G. Casini, and P. Bagnoli (1988a) Unimpaired acquisition of spatial reference memory but impaired homing performance of hippocampal ablated pigeons. Behav. Brain Res. 27:179-187. Bingman, V.P., P. Ioale, G. Casini, and P. Bagnoli (1988b) Hippocampal ablated homing pigeons show a persistent impairment in the time taken to return home. J. Comp. Physiol. A 163: 559-563. Bruce, L.L., and A.B. Butler (1984) Telencephalic connections in lizards. I. Projections to cortex. J. Comp. Neurol. 229:585-601. Casini, G., V.P. Bingman, and P. Bagnoli (19861 Connections of the pigeon dorsomedial forebrain studied with WGA-HRP and ‘H praline. J. Comp. Neurol. 245454-470. Craigie, E.H. (19301 Studies on the brain of the Kiwi (Apteryx australis). J. Comp. Neurol. 19223-358. Craigie, E.H. (1935) The hippocampal and parahippocampal cortex of the Emu (Dromiceius).J. Comp. Neurol. 61:563-592. Craigie, E.H. (19401 The cerebral cortex in paleognathine and neognathine birds. J. Comp. Neurol. 73:179-234. Crosby, E.C. (1917) The forebrain of the alligator, Alligator nississipiensis. J. Comp. Neurol. 27:325-402. Cuello, A.C., G. Galfre, and C. Milstein (1979) Detection of substance P in the central nervous system by a monoclonal antibody. Proc. Natl. Acad. Sci. U.S.A. 76:3532-3536. Cuello, A.C., C. Milstein, and J.V. Priestley (1980) Use of monoclonal antibodies in immunocytochemistry with special reference to the central nervous system. Brain Res. Bull. 5:575-587. Cuello, A.C., C. Milstein, R. Couture, B. Wright, J.V. Priestley, and J. Jarvis (1984) Characterization and immunocytochemical application of monoclonal antibodies against enkephalins. J. Histochem. Cytochem. 32.947957. Danscher, G. (1981) Histochemical demonstration of heavy metals: A revised version of the sulphide-silver method suitable for both light and electron microscopy. Histochemistry 71:l-16. Danscher, G. (1982) Exogenous selenium in the brain: A histochemical technique for light and electron microscopical localization of catalytic selenium bonds. Histochemistry 76281-293. Ellis, J.P., J.M. Sullivan, and M.W. Rana (1983) Somatostatin-like immunoreactivity in the retinae of adult and embryonic chickens. Proc. Soc. Exp. Biol. Med. 172:463-471. Gall, C. (1988) Seizures induce dramatic and distinctly different changes in enkephalin, dynorphin, and cholecystokinin immunoreactivities in mouse hippocampal mossy fibers. J. Neurosci. 8:1852-1862.

492 Gall, C., L.M. Berry, and L.A. Hodgson (1986) Cholecystokinin in the mouse hippocampus: Localization in the mossy fiber and dentate commissural systems. Exp. Brain Res. 624311137. Gall, C., J. Lauterborn, D. Burks, and K. Seroogy (1987) Co-localization of enkephalin and cholecystokinin in discrete areas of rat brain. Brain Res. 403:403408. Geneser-Jensen, F.A., F.-M.S. Haug, and G. Danscher (1974) Distribution of heavy metals in the hippocampal region of the guinea pig: A light microscope study with Timm’s sulphide silver method. Z. Zellforsch. 147:441478. Karten, H.J., and W. Hodos (1967) A Stereotaxic Atlas of the Brain of the Pigeon (Columba liuia). Baltimore: Johns Hopkins Press. Kohler, C. (1983) A morphological analysis of vasoactive intestinal polypeptide (VIP)-like immunoreactive neurons in the area dentata of the rat brain. J. Comp. Neurol. 221:247-262. Kohler, C., and V. Chan-Palay (1982a) Somatostatin-like immunoreactive neurons in the hippocampus: An immunocytochemical study in the rat. Neurosci. Lett. 34:259-264. Kohler, C., and V. Chan-Palay (1982133The distribution of cholecystokininlike immunoreactive neurons and nerve terminals in the retrohippocampal region in the rat and guinea pig. J. Comp. Neural. 210:136-146. Kohler, C., L. Eriksson, S. Davies, and V. Chan-Palay (1986) Neuropeptide-Y innervation of the hippocampal region in the rat and monkey brain. J. Comp. Neurol. 244:384-400. Kosaka, T., K. Kosaka, K. Tateishi, Y. Hamaoka, N. Yanaihara, J.-Y. Wu, and K. Hama (1985) GABAergic neurons containing CCK-8-like and/or VIP-like immunoreactivities in the rat hippocampus and dentate gyrus. J. Comp. Neurol. 239:420-430. Krebs, J.R., J.T. Erichsen, and V.P. Bingman (1991) The distribution of neurotransmitters and neurotransmitter-related enzymes in the dorsomedial forebrain of the pigeon (Columba lauza). J. Comp. Neurol. 314.467477. Krebs, J.R., D.F. Sherry, S.D. Healy, V.H. Ferry, andA.L. Vaccarino (1989) Hippocampal specialization of food-storing birds. Proc. Natl. Acad. Sci. U.S.A. mi388-1392. Mai, J.K., P.H. Stephens,A. Hopf, andA.C. Cuello (1986) Substance P in the human brain. Neuroscience 27:709-739. Martinez-Guijarro, F.J., P.J. Berbel, A. Molowny, and C. Lopez-Garcia (1984) Apical dendritic spines and axonic terminals in the bipyramidal neurons of the dorsomedial cortex of lizards (Laeerta). Anat. Embryol. (Berl.) 170:321-326.

J.T. ERICHSEN ET AL. Northcutt, R.G. (1967) Architectonic studies of the telencephalon of Iguana iguana. J. Comp. Neurol. 130:109-148. Northcutt, R.G. (1981) Evolution of the telencephalon in nonmammals. Annu. Rev. Neurosci. 4r301-350. Nunzi, M.G., A. Gorio, F. Milan, T.F. Freund, P . Somogyi, and A.D. Smith (1985) Cholecystokinin-immunoreactivecells form symmetrical synaptic contacts with pyramidal and nonpyramidal neurons in the hippocampus. J. Comp. Neurol. 237:485-505. Olucha, F., F. Martinez-Garcia, L. Poch, W.K. Schwerdtferger, and C. Lopez-Garcia (1988) Projections from the medial cortex in the brain of lizards: Correlation of anterograde and retrograde transport of horseradish peroxidase with Timm staining. J. Comp. Neurol. 276:469-480. Pisana, M. (1986) Cytoarchitecture and Connectional Organization in the Telencephalic Medial Wall of the Domestic Chick (Gallus domesticus). D. Phil. thesis, University of Sussex. Rose, M. (1914) Uber die cytoarchitektonische Gliederung des Vorderhirns der Voegel. J.F. Psychol. Neurol. 21:278-352. Sherry, D.F., and A.L. Vaccarino (1989) Hippocampus and memory for food cashes in black-capped chickadees. Behav. Neurosci. 103:308-318. Sherry, D.F., A.L. Vaccarino, K. Buckenham, and R.S. Herz (1989) The hippocampal complex of food-storing birds. Brain Behav. Evol. 34:308317. Shimizu, T., and H.J. Karten (1990) Immunohistochemical analysis of the visual wulst of the pigeon (Colurnba liuia). J. Comp. Neurol. 300:346369. Showers, M.J.C. (1982) Telencephalon of birds. In E.C. Crosby and H.N. Schnitzlein (eds): Comparative Correlative Neuroanatomy of the Vertebrate Telencephalon. New York: Macmillan, pp. 218-246. Smith, E.G. (1910) On some problems relating to the evolution of the brain. Lancet 1:Jan 1 (1-61,Jan 15 (147-1531, Jan 22 (221-227). Smith, G.T. (1989) Cell types and organization of the dorsomedial forebrain of the great tit (Parus major) and the marsh tit (Paruspalustris):A Golgi study. Honors thesis, Cornell University. Somogyi, P., A.J. Hodgson, A.D. Smith, M.G. Nunzi, A. Gorio, and J.Y. Wu (1984) Different populations of GABAergic neurons in the visual cortex and hippocampus of cat contain somatostatin- or cholecystokininimmunoreactive material. J. Neurosci. 4:2590-2603. Stengaard-Pederson, K., K. Fredens, and L.-I. Larsson (1983) Comparative localization of enkephalin and cholecystokinin immunoreactivities and heavy metals in the hippocampus. Brain Res. 273r81-96. Ulinski, P.S. (1975) Cortico-septa1 projections in the snakes Natrzz scpedon and Thamnophis sirtalis. J. Comp. Neurol. 164:375-388.