University of Groningen

The basic neural circuitry for sexual behavior van der Horst, Veronica Gerarda Johanna Maria

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 1996

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): van der Horst, V. G. J. M. (1996). The basic neural circuitry for sexual behavior: Pathways and plasticity. [S.n.].

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 10-10-2021 The basic neural circuitry for sexual behavior; Pathways and plasticity ISBN 90 367 0670 x

Drukkerij: Print Partners Ipskamp, Enschede RIJKSUNIVERSITEIT GRONINGEN

The basic neural circuitry for sexual behavior; Pathways and plasticity

Proefschrift

ter verkrijging van het doctoraat in de

Medische Wetenschappen

aan de Rijksuniversiteit Groningen

op gezag van de

Rector Magnificus, dr. F. van der Woude,

in het openbaar te verdedigen op

woensdag 25 september 1996

des namiddags te 2.45 uur

door

Veronica Gerarda Johanna Maria van der Horst

geboren op 21 mei 1969

te Heeze Promotor:

Prof. dr. G. Holstege Promotion Committtee:

Prof. dr. D. Kernell University of Groningen, Groningen

Prof. dr. D.W. Pfaff Rockefeller University, New York

Prof. dr. H.J. Ralston III University California San Francisco, San Francisco

Prof. dr. J. Voogd Erasmus University, Rotterdam

Paranymphs:

Dr. D.D. Ralston

Drs. M.C. van der Horst Run hard, be strong, think big Contents

General introduction 9

Chapter 1 Distinct cell groups in the lumbosacral cord of the cat project to different areas in the periaqueductal gray 17

Chapter 2 The organization of lumbosacral motoneuronal cell groups innervating hindlimb, pelvic floor and axial muscles in the cat 43

Chapter 3 Caudal medullary pathways to lumbosacral motoneuronal cell groups in the cat: evidence for direct projections possibly representing the final common pathway for lordosis 79

Chapter 4 Nucleus retroambiguus projections to lumbosacral moto- neuronal cell groups in the male cat, possibly representing the final common pathway for mounting behavior 99

Chapter 5 Estrogen induces axonal outgrowth in the nucleus retro- ambiguus-lumbosacral motoneuronal pathway in the adult female cat 113

General discussion 129

References 145

Abbreviations 158

Summary 159

Samenvatting 160

Dankwoord 163

List of publications 164

Curriculum vitae 165 General introduction

General introduction

9 General introduction General introduction

The somatic and emotional motor systems Motoneurons All motor activities, for example running, Motor system reaching, jumping, writing, talking, drinking, and Voluntary motor Emotional motor mating, require the activation of specific sets of striated system system and smooth muscles. The striated muscles are innervated Lateral Medial Lateral Medial by somatic motoneurons, whereas the smooth muscles eye, neck, specific axial and proximal emotional are innervated by sympathetic or parasympathetic body movements behaviors preganglionic motoneurons. Each muscles, or even independent gain setting systems movements of including triggering muscle compartment, is innervated by its own group of the extremities mechanisms of rhythmical and other motoneurons. Somatic motoneuronal cell groups spinal reflexes innervating muscles of the head are located in the brain Basic system (premotor interneurons) stem, those of the remaining parts of the body in the ventral horn of the spinal cord (lamina IX of Rexed, 1954; see Fig. 1). The axial muscles of the neck and back are innervated by motoneurons in the medial part of the ventral horn throughout the ventral horn, whereas Motoneurons those innervating the muscles of the extremities are located in the lateral ventral horn of the cervical and Figure 2 Schematic overview of the three subdivisions of the lumbosacral enlargements. Sympathetic preganglionics motor system (from Holstege, 1996). are present in the lateral horn of the thoracic and upper lumbar cord (see Cabot, 1996 for review), whereas The basic premotor interneuronal system The in- parasympathetic preganglionics are located in certain dividual components of the voluntary or emotional motor brain stem nuclei, and in the sacral cord (see Holstege, system reach the motoneurons via direct projections, or 1996 for review). All somatic and autonomic moto- indirectly via interneurons (see Fig. 2). Propriospinal neuronal cell groups are controlled by other structures premotor interneurons are located in the intermediate in the central nervous system (CNS), and together they zone of the spinal cord (Rexed’s laminae V to VIII; see form the “motor system” (Holstege, 1991; Fig. 2). Fig. 1). Premotor interneurons for the somatic motoneurons in the brain stem are located in the caudal pontine and medullary lateral tegmental field, which can be considered as the rostral extent of the spinal Rexed's laminae I to X intermediate zone (Holstege et al., 1977). However, premotor interneurons are not necessarily located in

II I close proximity to their target motoneurons, but can also III be found at significant distances. Examples are the C2 Dorsal horn IV interneurons projecting to C8 (see Holstege, 1988), and V the pontine M-region neurons projecting to bladder X VI preganglionic motoneurons in the sacral cord (see Blok

VII and Holstege, 1996). Together, all these premotor interneuronal cell groups constitute the “basic motor IX system” (Fig. 2; Holstege, 1991; 1996). VIII Ventral horn IX The voluntary motor system Part of our behavior is first sacral under control of the so-called “voluntary motor system” segment (S1) IX (Fig. 2). Examples are goal directed motor activities such as reaching and writing. The lateral component of the Figure 1 Schematic drawing of a transverse section of the voluntary motor system specifically involves the first sacral segment of the spinal cord, showing the laminar supraspinal control of goal directed movements of subdivisions in the gray matter according to Rexed (1954). especially the distal limb muscles, but also the orofacial

11 General introduction

Dorsal Midbrain Rostral Caudal periaqueductal gray Ventral

Nucleus Lumbosacral retroambiguus motoneurons

L4-S1 Dorsal

Ventral

Figure 3 Schematic representation of the central nervous system in sagittal view (at the top) , and transverse view (at the bottom. The location of the periaqueductal gray, the nucleus retroambiguus, and motoneuronal cell groups in the lumbosacral segments of the spinal cord is indicated in gray.

and tongue muscles during speech. It consists of the The emotional motor system Not all behavioral patterns cortico- and rubrospinal tracts, which act on laterally are voluntary. Crying, laughing and fear reactions for located motoneurons either directly, or via interneurons example are not mediated by the voluntary motor of the basic motor system (see Kuypers, 1981; Holstege, system, but by other structures in the CNS, which have 1996). been defined as the “emotional motor system” (Holstege, Postural control, involving axial and proximal muscles, 1991). Similar to the somatic motor system, the is controlled by cortico-, interstitio-, tecto-, vestibulo-, emotional motor system can be divided into a medial and reticulospinal tracts. Apart from the medial and a lateral component, corresponding to their location corticospinal tract, these pathways originate from the in the CNS. The lateral component is involved in specific medial tegmentum of the rostral midbrain, caudal pons, motor activities such as vocalization (=production of and medulla. They form the medial part of the voluntary sound; Holstege, 1989; Zhang, 1992; Davis et al., 1996), motor system, and act on medially located motoneurons, micturition (Holstege et al., 1986; Blok and Holstege, but mostly via the premotor interneurons of the basic 1996), bloodpressure control (Lovick, 1993; 1996), and motor system (see Kuypers, 1981; Holstege, 1996). mating (Pfaff et al., 1994). These activities are controlled

Limbic system

Pedunculopontine Periaqueductal gray and cuneiform nuclei (PAG)

Ventral 1/3 of caudal pontine Barrington's subretrofacial nucleus and medullary medial tegmentum nucleus nucleus retroambiguus

sympathetic sensory neurons motoneurons and parasympathetic sympathetic motoneurons of the motoneurons of the spinal cord T1-T2 C4-T8 preganglionics in the premotor interneurons preganglionics preganglionics larynx, iliopsoas, laminae intermediolateral lamina X in the dorsal horn in the in the in the pharynx, adductor longus, VIII and medial VII cell column intermediolateral ventral horn and sacral cord intermediolateral soft palate, and hamstring, cell column intermediate zone cell column expiratory muscles pelvic floor, and axial muscles

general level nociception locomotion micturition cardiovascular vocalization Present thesis defensive pupil dilatation ? ? of sympathetic control changes posture ? activity

Figure 4 Schematic overview of the descending projections from the PAG to different regions in the caudal brain stem and spinal cord and their possible functional roles.

12 General introduction by distinct projections to interneurons of the basic motor and gainsetting mechanisms (Holstege and Kuypers, system. For example, vocalization is produced by 1982; White et al., 1996; Holstege, 1996; Mason and projections from the midbrain periaqueductal gray to Leung, 1996). premotor interneurons in the nucleus retroambiguus in To integrate these motor activities, the PAG receives the medulla, which in turn project to specific groups of information from structures in the forebrain, such as motoneurons innervating the “vocalization” muscles from several nuclei of the hypothalamus, bed nucleus (Holstege, 1989; Zhang, 1992; Davis et al., 1996). of the stria terminalis, and amygdala (see Holstege, In contrast, the medial component consists of pathways 1991), but also from the spinal cord (Wiberg and which terminate diffusely and acts on motoneurons not Blomqvist, 1984; Yezierski, 198 ; VanderHorst and only via premotor interneurons, but also directly. It has Holstege, 1992; VanderHorst et al., in press; Chapter a more general function, such as modulating the 1). excitability of motor- and sensory neurons (see Holstege, 1996; White et al., 1996; Mason and Leung). In this Scope of the thesis system cell groups in the ventral part of the ponto- The main topic of the present tracing study deals with medullary tegmentum are of importance, which project the question how the PAG controls specific emotional to the dorsal (nucleus raphe magnus and adjoining behavior. Using retro- and anterograde tracing reticular formation) or ventral horn (nucleus raphe techniques, pathways involved in such behavior have pallidus and adjacent area) throughout the spinal cord been identified. For this technique, a tracer is injected (Holstege and Kuypers, 1982; White et al., 1996; in a certain structure in the CNS (Fig. 5). From the Holstege, 1996; Mason and Leung, 1996). These injection site, the tracer is taken up by axon terminals structures cause that motoneurons are more easily and transported from the axons back into their cell body. activated by the voluntary motor system during Thus, retrograde transport results in labeled cells. It is threatening events. also possible that the tracer is taken up by cell bodies and transported to their axon terminals. This is so-called The central role of the periaqueductal gray in the anterograde transport results in labeled fibers and emotional motor system terminals. After transport of the tracer, labeled fibers An important structure in the emotional motor system which are derived from cells within the injection site, is the midbrain periaqueductal gray (PAG). The PAG and cells which project with their fibers to the injection consists of neurons surrounding the aqueduct of Silvius, site can be visualized light- or electronmicroscopically. a pipeline for cerebrospinal fluid connecting the third At the beginning of the study, the major line of research and fourth ventricles (Fig. 3). It controls basic reactions, concerned the control of the micturition reflex. such as aggressive and defensive behaviors (Bandler, 1991) which are important for the survival of the individual, as well as reproductive behavior which is Dendrites crucial for survival of the species (Sakuma and Pfaff, Neuron 1979a,b). Each of these reactions consists of a combination of specific motor activities and levelsetting mechanisms. For example, aggressive behavior Terminal comprises specific components, such as vocalization and changes in bloodpressure, as well as a general increase in the pain threshold (Bandler, 1991). It has been demonstrated that different parts of the PAG are involved in the control of these components (Bandler, 1991; Bandler and Shipley, 1994). The ventrolateral PAG regulates quiesecence, immobility, hypotension, Retrograde Anterograde bradycardia, and opiod dependant analgesia, whereas transport transport the lateral PAG controls jumping, hypertension, tachycardia, and non-opioid analgesia. Axon In order to reach the motoneurons involved in each of these reactions, the PAG makes use of premotor interneuronal cellgroups in the brainstem or spinal cord (Fig. 4). For example the nucleus retroambiguus in the caudal medulla forms the relay for vocalization, the M- region in the pons for micturition (Holstege, 1986; Blok and Holstege, 1996), the retrofacial nucleus or rostral ventrolateral medulla for changes in blood pressure Figure 5 Schematic drawing showing transport of an injected (Carrive and Bandler, 1991; Lovick, 1993), and the neuronal tracer (in gray) in the retrograde and anterograde ventral part of the pontomedullary tegmentum for level- direction.

13 General introduction Stimulation in the PAG has been demonstrated to elicit Case 2267 micturition, which indicates that it is involved in the motor output of micturition (Skultety, 1963). This raised OT the question whether the PAG receives direct input from lumbosacral segments of the spinal cord, at which level CGMp information from the bladder enters the CNS via the ML pelvic nerve. Since no detailed reports were present III BIC describing direct projections from the lumbosacral cord to the PAG, a series of experiments was started which OT CS resulted in the data presented in the first chapter. They show that lumbosacral-PAG projecting cells can be CGMp divided into 3 groups, each projecting differently to the III BIC PAG. One of these groups appeared to precisely match ML the location of incoming fibers of the pelvic nerve. Since CS this nerve not only conveys information from the bladder, but also input from the rectum, vagina, and cervix, this CGMp III lumbosacral-PAG pathway is thought to be also BIC important for reproductive behavior. ML Another intriguing question concerned an unclear CS projection from the nucleus retroambiguus (NRA; see Fig. 3) to the lumbosacral cord. The NRA is a group of interneurons at the transition between the brain stem BIC and the spinal cord and is known to be involved in MLF expiration, vomiting, straining, and vocalization (for CS review see VanderHorst and Holstege, 1996). Retro- and anterograde tracing experiments have shown that it receives a distinct projection from the PAG (Holstege, BIC 1989; VanderHorst and Holstege, 1996; see Figs. 6 and MLF

7). The NRA interneurons in turn are known to project CS to motoneuronal cell groups of larynx and pharynx

muscles in the brainstem (Holstege, 1989), and of the IC internal intercostal and abdominal wall muscles in the thoracic and upper lumbar cord (Holstege and Kuypers, IV 1982; Feldman et al., 1985; Holstege, 1989; Miller et al., 1989). These specific PAG-NRA-motoneuronal

projections form the final common pathway for IC vocalization (production of sound), an example of

specific emotional behavior (Holstege, 1989; Zhang et CUN MLF al., 1992). Remarkably, the NRA appeared also to project BC to the ventral horn of the lumbosacral enlargment (Fig. 2), containing motoneurons of hindlimb muscles which are not involved in expiration or vocalization. In order P to determine the funtion of this projection, it was necessary to identify which motoneuronal cell groups were targeted by the NRA. For this purpose a detailed and complete overview of the location of motoneuronal G inj. cell groups in the lumbosacral cord was necessary. A NRA map of the motoneurons innervating the hindlimb, axial and pelvic floor muscles is presented in Chapter 2. After G precisely reconstructing the NRA-lumbosacral pathway inj. Vsp and its termination pattern in the female cat (Chapter P 3), a distinct set of motoneuronal cell groups was identified receiving NRA fibers. Functionally, the muscles innervated by this set are not specifically Figure 6 Schematic drawings of retrogradely labeled cells in involved in jumping or running, but very well fitted the the midbrain after a small injection with WGA-HRP in the receptive posture of female cats during mating. This NRA. Note the circumscribed groups of retrogradely labeled raised the question whether the NRA-lumbosacral neurons in the lateral PAG which are present bilaterally (from pathway is also present in the male cat. Chapter 4 shows VanderHorst and Holstege, 1996).

14 General introduction that the NRA also terminates in distinct motoneuronal that estrogen induces axonal sprouting of the NRA- cell groups in the male, but with a few but significant lumbosacral pathway in adult female cats. This finding differences in comparison to the female, suggesting a not only is important for our understanding of the CNS role for this pathway in mounting behavior. control of sexual behavior, but also shows that growth Cats only display reproductive behavior when sufficient still can take place in the adult CNS, even at the level of sex steroids are present. This implicates that these the spinal cord. hormones affect pathways in the central nervous system The general discussion brings the results of all 5 chapters involved in reproductive behavior. In female cats, the into perspective, presenting a scheme for the sensory levels the sex steroid estrogen determine whether the and motor pathways in the spinal cord and brain stem animals are receptive. In Chapter 5 it is demonstrated involved in mating behavior.

PAG CS CS

IC IC BIC IV III ML ML ML PC P P P L3-hemisection A B L7

Figure 7 Photomicrographs of the medulla oblongata caudal to the obex. The NRA is indicated by white arrows. On the left, a dense anterograde projection is present in NRA bilaterally after an injection with WGA-HRP in the lateral PAG (for injection site see drawing at the top). On the right, a bilateral injection with WGA-HRP had been made in the L6 and L7 spinal segments combined with a hemisection in the left L3 segment (for injection site see drawing at the top). The NRA ipsilateral to the hemisection contains numerous retrogradely labeled neurons.

15

lumbosacral-PAG projections

Chapter 1

Perineum, Vagina, Cervix

17 Chapter 1

18 lumbosacral-PAG projections Distinct cell groups in the lumbosacral cord of the cat project to different areas in the periaqueductal gray

ABSTRACT The periaqueductal gray (PAG) is involved in aggressive and defensive behavior, micturition and lordosis. Especially for the latter two functions, PAG afferents from the lumbosacral cord are of vital importance because, in addition to information regarding homeostasis and thermoregulation, they convey information from the pelvic viscera and sex organs. In the present retro- and anterograde tracing study, the projection patterns of different lumbosacral cell groups in the PAG were determined. In the retrograde study, wheatgerm agglutinin-horseradish peroxidase (WGA-HRP) injections were made in the PAG and/or adjacent tegmentum, and in the anterograde study, WGA-HRP was injected in different lumbosacral segments. The results revealed that lumbosacral-PAG neurons could be divided into three groups. The first and largest group was present in lumbar 7- sacral 3 segments (L7-S3) and consisted of small, oval and fusiform neurons. It extended from the dorsolateral part of lamina I in L7, along the lateral part of the dorsal horn in S1 into lamina V of S2. In the lateral part of S2, some of its neurons formed clusters with intervals of ±230 µm. The location of the first group overlapped extensively with the termination area of pelvic and pudendal afferents. The main midbrain target of the first group was the medial part of the lateral PAG. The second group consisted of small to large multipolar neurons in laminae VIII and medial VII of caudal L6, L7, and rostral S1. This group projected strongly to a distinct region in the lateral part of the lateral PAG and the laterally adjacent tegmentum. About 10% of the labeled neurons did not fit in the two groups. They were evenly distributed throughout lumbar 4- coccygeal 3 segments (L4-Co3) and consisted of large multipolar lamina V neurons and small lamina I neurons which projected diffusely to the lateral and dorsal PAG. The large lamina V neurons also targeted the laterally adjacent tegmentum. The possible involvement of the lumbosacral-PAG projections in micturition, lordosis, and defensive and aggressive behavior is discussed.

INTRODUCTION The midbrain periaqueductal gray (PAG) is known to of survival behavior, the PAG receives descending input play an important role in emotional behavior. Examples from more rostrally located limbic structures, e.g., the are threat display, fight, flight, immobility, micturition anterior, lateral, and ventromedial hypothalamus, medial and lordosis, all of which can be elicited by stimulating preoptic area, central nucleus of the amygdala and the the PAG (Abrahams et al., 1960; Skultety, 1963; Sakuma bed nucleus of the stria terminalis (see Hopkins and and Pfaff, 1979a; Zhang et al., 1990b). In order to Holstege, 1978; Saper et al., 1976, 1978, 1979a,b; Risold produce such integrated behaviors, the PAG controls et al., 1994; Canteras et al., 1994). In addition, nociception and autonomic and somatic motor systems somatosensory and visceral input from brainstem and (see Abrahams et al., 1960; Besson and Chaouch, 1987; spinal cord also reaches the PAG. Especially the Holstege, 1989; Bandler et al., 1991; Carrive and lumbosacral projections to the PAG are of interest, Bandler, 1991; Lovick, 1993; Zhang et al., 1994). In because they convey information from the urogenital respect to the motor output, the PAG projects to the tract, relevant for the micturition (Noto et al.,1991) and autonomic and somatic motoneurons via interneuronal lordosis reflex (for review see Pfaff et al., 1994). In rat, cell groups in the pons, medulla oblongata and spinal cat, and monkey, lumbosacral projections to the cord (Holstege; 1991). For each motor component, a mesencephalon have been studied by using retrograde separate pathway seems to exist. For example, the PAG (cat: Wiberg and Blomqvist, 1984; monkey: Wiberg et mediates blood pressure changes via the rostral al., 1987; Zhang et al., 1990a; rat: Menetrey et al., 1982; ventrolateral medulla (RVLM, see Lovick, 1993), Harrmann et al., 1988; Yezierski and Mendez, 1991), vocalization and presumably lordosis via the nucleus and anterograde tracing techniques (cat, rat, monkey: retroambiguus (Holstege, 1989; Zhang et al., 1992; Yezierski, 1988 ; cat: Wiberg and Blomqvist, 1984; VanderHorst and Holstege, 1995), and micturition monkey: Wiberg et al. 1987) as well as physiological (Skultety, 1959; Gjone, 1966), possibly via the pontine methods (Yezierski and Schwartz, 1986; Hylden et al., micturition center (Blok and Holstege, 1994). 1986b). Recent findings demonstrated that lower lumbar In order to “trigger” the various behavioral components and sacral segments contain more PAG-projecting

19 Chapter 1

neurons than segments in the thoracic cord or in the Photomicrographs of representative sections were taken cervical enlargement (VanderHorst and Holstege, 1992). with a Zeiss Axioskop under polarized light-darkfield However, a specific study on the ascending projections illumination. from the different lumbar and sacral segments to the PAG does not exist. The present study shows that there Quantification of retrogradely labeled neurons are two distinct lumbosacral-PAG cell groups, which Labeled profiles were counted to find out whether probably serve different functions. neurons in different laminae of the various lumbosacral segments project to different areas of the PAG and adjacent tegmentum. It was not the aim of the present MATERIALS AND METHODS study to determine the absolute number of lumbosacral neurons projecting to different areas in the midbrain. In all cases except for case 2187, ipsi- and contralaterally Retrograde tracing study located retrogradely labeled profiles were counted in Surgical and histological procedures every fifth section of the L4-Co3 spinal segments. The In 19 adult cats (weight 3.0-5.5 kg) stereotaxic injections volume of the spinal gray matter of the L4-S1 segments of 10-200 nl 2.5-5% wheatgerm-agglutinin-horseradish is much larger than of the S2-Co3 segments. In order to peroxidase (WGA-HRP) were made in the PAG or in get a good impression of the density of PAG-projecting the tegmentum ventral and lateral to it. The injections neurons in each segment, the absolute number of counted were made with a Hamilton microsyringe or with a glass profiles in each segment was corrected for its volume pipette using a picopump (World Precision Instruments of gray matter. This was done as follows: 1. the total PV830). length of the respective segments was determined (the The surgical procedures, pre- and postoperative care, total number of one out of five consecutive sections x 5 and the handling and housing of the animals occurred x 40 µm); 2. in each segment the surface area of the according to the protocols approved by the Faculty of gray matter was determined by measuring it in one or Medicine of the University of Groningen. The animals two representative sections using a computer program; were anesthetized with an initial dose of ketamin 3. the volume of gray matter in each segment was (Nimatek; 10 mg/kg, i.m.) and xylazine hydrochloride determined by multiplying the length with the surface (Sedamun; 2 mg/kg, i.m.) and subsequently artificially area. This volume was devided by the volume of a ventilated under mixed halothane - nitrous oxide standard segment (T3) of the same case, resulting in the anesthesia. During surgery ECG and temperature were relative volume of gray matter; 4. finally, for each monitored. segment the relative number of labeled neurons was After 3 days survival time, the animals were initially determined by dividing the absolute number of anesthetized with ketamine (Nimatek) and xylazine retrogradely labeled neurons in each segment by its hydrochloride (Sedamun) followed by an intraperitoneal relative volume. injection of 6 ml 6% pentobarbital sodium (Nembutal). To study the rostrocaudal distribution of neurons in Subsequently, they were transcardially perfused with 2 different parts of the L6-Co3 gray matter projecting to liters of heparinized saline at 37 °C, followed by 2 liters different parts of the PAG and adjacent tegmentum, in of fixative containing 2% glutaraldehyde, 1% para- five cases (2279, 2282, 2300, 2316, and 2338) in every formaldehyde, and 4% sucrose in 0.1 M phosphate buffer 12 alternate consecutive sections labeled profiles were (pH 7.2-7.4; at room temperature). The brain was cut counted. Because lamina I gradually extends into lamina stereotactically at level A 10, brain and spinal cords were V at levels L7-S2, it was not possible to precisely removed, postfixed for 1 to 2 hours, and stored overnight determine the number of labeled profiles in each of these in 25% sucrose buffered phosphate at 4 °C. In all cases laminae. Therefore, the spinal gray matter was divided except for 2187, the brainstem and all spinal segments artificially into 4 quadrants: dorsomedial (DM), were cut on a freezing microtome into 40 µm serial, dorsolateral (DL), ventromedial (VL), and ventrolateral transverse sections. In case 2187, the L4-Co3 segments (VL; for examples see A, B, and C of Fig. 10) using a were cut horizontally. vertical and a horizontal line. In the individual sections, Of 4 cases (2155, 2159, 2161, and 2167), 1 out of 5 the vertical line forms the midline of the smallest part sections were processed, of 11 cases (2227, 2239, 2248, of the neck of the dorsal horn, whereas the horizontal 2250, 2254, 2259, 2262, 2300, 2316, 2336 and 2338) line is located at the dorsal border of the central canal. every second section, and of 4 other cases (2179, 2182, Since the division of the lumbosacral cord into segments 2187, and 2190) each section was processed using the varies considerably between individual cases (Sherring- tetramethylbenzidine (TMB) procedure according to ton, 1892; Romanes, 1951; VanderHorst and Holstege, Mesulam (1982). To determine the extent of the injection in preparation), for a more exact analysis, the sites, sections of the midbrain were processed with rostrocaudal levels in the different cases were matched diaminobenzidine (DAB). In all cases the distribution at level 0. Level 0 was defined as the level at which the of retrogradely labeled neurons in the L4-Co3 segments ventrolateral part of the ventral horn rapidly decreases was microscopically examined with a Zeiss Axioskop. in size, i.e. the ventrolateral border of the ventral horn

20 lumbosacral-PAG projections follows a straight line (B in Fig. 10). Rostrocaudally, L5 (case 2309), L6 (cases 2249 and 2278), L7 (2295), the sections extending from 10 mm rostral to 15 mm L7-S1 (case 2261), S1-S2 (case 2265), S2-S3 (2302), caudal to level 0 were examined. and S1-S3 segments (2329). After laminectomy, multiple penetrations were made with a glass pipette using a Plottings picopump. The injections were made on the left side in To obtain an overview of the location of PAG-projecting cases 2249, 2261, 2265, and 2278, and on both sides in neurons, in cases 2179, 2182, 2239, 2300, 2316, and cases 2295, 2302, 2309, and 2329. In order to 2338, retrogradely labeled neurons in the L4-Co3 differentiate between ipsi- or contralaterally ascending segments were plotted into drawings. Each of these pathways, in cases 2278, 2295, 2302, 2309 and 2329, drawings represented 12 consecutive alternate sections. the injections were preceded by a left sided hemisection In order to find out where in the white matter the ascending lumbosacral-PAG fibers were located, in case a few segments rostral to the injection. The hemisections 2161 of the C8, T7, L5, L7, S1, and S2 segments one or were made by suction. All further surgical and two sections were plotted for labeled fibers. All plottings histological procedures were similar to the procedures were made using a drawing tube, a digitizer and a of the retrograde study. Macintosh computer. The distribution of anterogradely labeled fibers was microscopically examined with a Zeiss Axioskop under combination of polarized light and darkfield condensor Anterograde tracing study and photomicrographs of representative sections were In order to determine the precise projection areas of the taken. Drawings of the distribution of anterogradely PAG-projecting neurons in the different lumbosacral labeled fibers in the PAG and adjacent tegmentum were segments, in 8 adult cats (5 females and 3 males; 3.1- made by means of a drawing tube and scanned into a 4.5 kg) 2-6 µl WGA-HRP (2.5-5%) was injected in the Macintosh computer.

Group A. 2155 2161 2179 2182 2187 2190 2300 A4.3

A3.4

A2.5

A1.6

A0.7

P0.2

P0.9

P1.5

P2.2

Figure 1 Schematic drawings of WGA-HRP injection sites involving the lateral and ventrolateral parts of the caudal PAG (group A). On the left are indicated the rostrocaudal levels according to Berman (1968).

21 Chapter 1

Group B. lateral rostral PAG tegmentum adjacent to PAG ventrolateral intermediate PAG and adjacent tegmentum 2159 2167 2336 2338 2239 2248 2250 A4.3

A3.4

A2.5

A1.6

A0.7

P0.2

P0.9

P1.5

P2.2

Group B.

dorsal and dorsolateral PAG 2227 2254 2259 2262 2316 A4.3

A3.4

A2.5

A1.6

A0.7

P0.2

P0.9

P1.5

P2.2

Figure 2 Schematic drawings of WGA-HRP injection sites in group B, involving the rostral or dorsal PAG, the ventrolateral part of the intermediate PAG, and the adjacent tegmentum.

22 lumbosacral-PAG projections

RESULTS (compare cases 2161 and 2155 with case 2300; Fig. 4). However, despite the fact that the injection site in the lateral PAG in case 2300 was only a small fraction of Retrograde tracing study that in the case with the largest injection (2161), the Location of injections total number of retrogradely labeled profiles in case 2300 The cases were divided into two groups (A and B). The still was 44% of that in case 2161. It indicates that the injections in the cases of group A involved the lateral lateral PAG is a main target of the lumbosacral-PAG and ventrolateral part of the caudal PAG (2155, 2161, projection. 2179, 2182, 2187, 2190, and 2300; Fig. 1). The injections In all cases of group B, with injections in the dorsal and in the cases of group B (Fig. 2) were made in the dorsal rostral PAG and adjacent tegmentum (Fig. 2), a much (2227, 2254, 2259, 2262, and 2316) and rostral PAG smaller number of labeled neurons was present (Fig. (2159), in the ventrolateral intermediate PAG ( 2239, 5). They were more evenly distributed over the different 2248, and 2250), as well as in the tegmentum lateral lumbosacral segments than the labeled profiles in group (2336 and 2338) and ventrolateral to the PAG (2167). A. An indication of what we have considered the core of the injection site is shown as the black area in figure 3 Location and morphology of the spino-PAG (case 2339). This core produces the great majority of neurons in the L4-Co3 segments the retrogradely labeled neurons. The location of labeled lumbosacral PAG neurons of group A will be described using case 2300 as an example, because its injection was almost completely restricted to the PAG (Fig. 1).

Labeled neurons in Lissauer’s tract and lamina I These Aq ICICIC neurons were found in the L4-Co1 segments, but not in Co2 and Co3 (Fig. 6). They were most numerous at the levels caudal L6 to caudal S2 and formed a continuous column. Rostrally in this column, labeled lamina I neurons were predominantly present dorso-laterally, at the apex of the dorsal horn (Fig. 6: arrows in L7a-g, S1a-d; Fig. 7a). From rostral to caudal in the S1-S3 segments, the column of labeled neurons in lamina I extended from the apex of the dorsal horn into lamina V, forming a cell bridge along the lateral edge of the dorsal horn (Rexed, 1954; Fig. 6: arrows in S1e-S3c; Figure 3 Photomicrograph showing an example of an injection Figs. 7b, c, d). In the L6 and L7 segments, an occasional site (case 2239). The black area represents the core of the labeled neuron was present in the white matter at the injection as indicated in figure 2. IC= Inferior colliculus; Aq= entrance of dorsal root fibers (Fig. 6 L7d and e). Aqueduct. Bar represents 500 µm. Labeled neurons in Lissauer’s tract and lamina I were small and fusiform or oval shaped (Table 1). Horizontal Number of labeled neurons in the L4-Co3 sections (case 2187) showed that at levels L7-rostral S1, segments the labeled cell bodies and dendrites were oriented In all cases of group A (injections in the lateral and rostrocaudally along the axis of the fibers in Lissauer’s ventrolateral part of the caudal PAG), labeled neurons tract. In caudal S1 and in S2 they were oriented in a were present bilaterally, but with a strong contralateral dorsolateral-ventromedial direction along the lateral preponderance (Fig. 4). To quantatively compare the edge of the dorsal horn (Fig. 6: S1 e-f and S2 a-c; Fig. different PAG afferent inputs from the various levels of 7c). the lower lumbosacral cord, in all cases of group A except 2187 which was cut horizontally, for each of the Labeled neurons in the lateral laminae V and VII In L4-Co3 segments the number of retrogradely labeled the sacral cord, many labeled neurons were found in profiles in every fifth transverse section was counted these laminae (Fig. 6 S2b-S3c). Their shape and size (Fig. 4, left). Retrogradely labeled profiles appeared to (Table I) was similar to the neurons in the superficial be most numerous in the L6-S2 segments. In all 6 cases, dorsal horn i.e. they were relatively small and mainly the density of retrogradely labeled neurons, i.e. the fusiform or oval shaped, although some were triangular number of labeled neurons relative to the volume of gray or multipolar. From a sample of 115 labeled cells in this matter of each segment (Fig. 4, right; see Materials and area (cases 2300 and 2161) from which the orientation Methods) was highest in the S2 segment. could be determined, many neurons were oriented in a The number of retrogradely labeled neurons was dorsolateral-ventromedial (42%) or mediolateral (39%) proportional with the extent of the injection sites direction (Fig. 7 B-D). They were located at the transition

23 Chapter 1

Group A. Absolute Relative

350 250 300 2155total: 1294 2155 200 250 = ipsi ipsi: 21% 200 150 # neurons contra: 79% 150 = contra 100 100 50 50

0 0 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 2161 total: 1522 2161 200 250 200 ipsi: 21% 150 # neurons contra: 79% 150 100 100 50 50 0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 2179total: 804 2179 200 250 200 ipsi: 15% 150 # neurons 150 contra: 85% 100 100 50 50 0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 2182 total: 789 200 2182 250 200 ipsi: 21% 150 # neurons

150 contra: 79% 100 100 50 50 0 0 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

250 350 300 2190 total: 1004 200 2190 250 150 200 ipsi: 22% # neurons 150 contra: 78% 100 100 50 50

0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 total: 666 2300 200 2300 250

200 ipsi: 20% 150

# neurons contra: 80% 150 100 100 50 50

0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 level level Figure 4 On the left, histograms showing the number of retrogradely labeled profiles in the L4-Co3 segments after injections involving the lateral and ventrolateral part of the caudal PAG. In each segment, the labeled profiles in every fifth section were counted. “Total” refers to the total number of labeled profiles observed in the L4-Co3 segments of the respective cases. Note that in all cases labeled profiles are most numerous in the L6-S2 segments. Histograms on the right show the number of labeled profiles relative to the volume of gray matter of the respective segments of each case. Note that PAG-projecting neurons are relatively most numerous in segment S2.

24 lumbosacral-PAG projections

Group B. Absolute Relative

350 250

300 2159total: 151 200 2159 250 = ipsi 150 200 ipsi: 11% # neurons contra: 89% 150 100 = contra 100 50 50

0 0 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3

350 250 300 total: 201 2167200 2167 250 200 ipsi: 15% 150 # neurons contra: 85% 150 100 100 50 50

0 0 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3 L4 L5 L6 L7 S1 S2 S3 Co1 Co 2 Co 3

350 250 300 2316 2316 total: 198 200 250 200 150

# neurons ipsi: 27%

150 contra: 73% 100 100 50 50 0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 2336 total: 79 200 250 2336 200 150 # neurons ipsi: 24% 150 contra: 76% 100 100 50 50 0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3

350 250 300 2338 total: 191 200 2338 250 200 150

# neurons ipsi: 23%

150 contra: 77% 100 100 50 50 0 0 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 L 4 L 5 L 6 L 7 S1 S2 S3 Co1 Co2 Co3 level level Figure 5 Histograms showing the distribution of retrogradely labeled profiles in the L4-Co3 segments as in figure 3, but after injections in the rostral and dorsal PAG or adjacent tegmentum. Note that the number of labeled profiles in these cases is much lower and more evenly distributed over the different segments than in the cases of group A. area between lamina I and lateral V and especially (7%) were oriented medio-ventrolaterally with dendrites numerous in caudal S1-rostral S2. Their dendrites extending medially and ventrolaterally (Fig. 7 G, H). extended medially into the dorsal gray commissure (as The orientation of 17% of the labeled neurons of this far as 243 µm) and dorsolaterally along the lateral edge sample was not clear. of the dorsal horn (as far as 252 µm; Figs. 7 C-E). Other In the case that was cut horizontally (2187) and in the labeled cell bodies located laterally in laminae V-VII cases in which all consecutive transverse sections were were oriented dorsoventrally (12%). Their dendrites collected (2179, 2182, and 2190), the sacro-PAG neurons extended along the lateral edge of the dorsal and ventral formed small clusters at distinct rostrocaudal intervals. horn (Fig. 7 F). At the levels S2-S3 some labeled neurons This was especially clear for the neurons in the most

25 Chapter 1

Case 2300

L4

L5

L6

L7a

L7b

L7c

L7d L7e

L7f

L7g

S1b S1a

S1c

S1d S2a S1e S2b S1f S2c S2d S3c S2e S3d S3a Co 1 S3b Co 2 Co 3

Figure 6 Schematic drawings showing the location of retrogradely labeled neurons in case 2300 (group A) with an injection involving the medial part of the lateral PAG. Each drawing represents 12 alternate sections. From the L7-S3 segments, all drawings are shown, but from segments L4-L6 and Co1-Co3, only one representative drawing was selected. The arrows indicate a large group of labeled neurons located in the lateral parts of lamina I in the L7-S1 segments, gradually extending into the lateral parts of laminae V and VII at the level S2-S3.

26 lumbosacral-PAG projections

A B

C D

EF

GH

Figure 7 Darkfield-polarized light photomicrographs of retrogradely labeled neurons (case 2300) in segments L7 (A), S1 (B and C), S2 (D,E,F, and G) and S3 (H). Note that there is a rostrocaudal shift in the location of labeled cells from the apex of the dorsal horn (A), along the lateral edge of the dorsal horn (B), into lamina V (C, D, E, F, G, and H) and lamina VII (F, G). Note further the few labeled cells dorsal to the central canal (F and G) and in the white matter adjoining lamina VII (E and G). Bar represents 200 µm.

27 Chapter 1

lateral part of the intermediate zone at the caudal S2 Labeled neurons in medial lamina V and VII and lamina level (Figs. 7 D and E). In this area, in every 4-5 X In all L4-Co3 segments, very few large sized consecutive sections of cases 2179, 2182, and 2190, multipolar neurons were found in this region close to hardly any labeled neurons were found, whereas each the central canal. Their dendrites extended bilaterally of the subsequent 2-5 sections contained small groups into the lateral intermediate zone. of labeled cells (6-18). In case 2179 a total of 4 and in In the sacral segments, some small sized labeled neurons cases 2182 and 2190 5 of such clusters were observed. were located dorsal or dorsolateral to the central canal In case 2187, four such clusters were observed over a (Table I; Fig. 7f, g). length of about 850 µm with intervals between 220-300 µm, measured from centre to centre (Fig. 8 S2). The Distribution of neurons projecting to different size of the clusters was 50-200 µm, rostrocaudally. The areas in the PAG and adjacent tegmentum labeled neurons in the cell bridge just lateral to the dorsal The laminar distribution of labeled neurons as described horn in S1 and rostral S2 (Fig. 8 S1) also formed clusters, above for case 2300 was also found in the other cases of although not as prominent as in the lateral laminae V group A (injections in the lateral and ventrolateral part and VII of S2-S3. Each of the S1-rostral S2 clusters of the caudal PAG), but with one particular quantitative extended rostrocaudally over a length of 240-440 µm difference (Fig. 10). Large numbers of labeled profiles and they were located at intervals of 600-800 µm, were found in lamina I of the L6-rostral S1 segments measured from centre to centre. In the dorsolateral part and in laminae V and VII of the sacral cord, but when of lamina I in the L6-L7 segments, labeled neurons did the injection site extended into the laterally adjacent not form clusters. tegmentum and deep layers of the superior colliculus Large (Table 1), multipolar neurons were found in the (case 2179), labeled neurons became more numerous lateral part of lamina V, not only in S1-S3, but throughout in the medial part of the L6-S1 intermediate zone (28% the segments L4-Co3 (Figs. 7H arrow, and 9 left). Their in case 2179 compared to 13% and 10% in cases 2182 dendrites extended into the deeper layers of the dorsal and 2300, respectively). These findings suggested that horn, the dorsolateral funiculus, the intermediate zone neurons in laminae VIII and medial VII of L6-S1 project of the ventral horn, the dorsal gray commissure and predominantly to the area lateral to the PAG. along the lateral edge of the dorsal horn. They were never found in groups and horizontal sections showed that Tegmentum laterally adjoining the PAG This hypothesis these single neurons were present at regular rostrocaudal was tested by making injections limited to the intervals of 583±163 µm (n=27). tegmentum laterally adjoining the PAG (cases 2336 and 2338 of group B; Figs. 10 and 11). Indeed, labeled Labeled neurons in the transition area of laminae VII neurons were present in laminae VIII and medial VII in and VIII These neurons were most numerous in the L6, L7 and rostral S1 (65% of all labeled neurons in caudal part of the enlargement i.e. between caudal L6 case 2338; see Figs. 5 and 10). In addition, throughout and caudal S1 (Fig. 6: L7 d-S2b; Fig. 9 right). They had L4-Co3 some large sized (Table I) labeled neurons were a stronger contralateral predominance (89%) than the present in laminae V (30%) and X (5%), but the labeled neurons in laminae I and V (77%). Their cell numerous labeled profiles in lamina I, and the lateral bodies were multipolar, with multiple dendrites laminae V and VII of the sacral cord, as present in after extending into all directions. Most of them were injections in the lateral and ventrolateral PAG, were relatively large (Table 1), but some smaller (largest absent in this case (2338, Figs. 10 and 11). Labeled diameter 25 µm) or giant neurons (largest diameter 89 neurons in case 2336 were distributed similar to case µm) were also observed. 2338, but were not as numerous (Fig. 5).

Table I Size of dorsal midbrain-projecting neurons in different regions of the lumbosacral cord

Lamina I Lateral lamina V Lamina VIII-medial VII Lamina X

Case long axis short axis long axis short axis long axis short axis long axis short axis L5-L7 2179 22.4±7.8* 11.0±2.5 (n=37) 51.7±15.8 26.7± 6.4 (n=18) 47.0±14.7 28.2± 8.9 (n=49) - - 2300 23.5±8.2 11.4±3.2 (n=42) 52.1±18.6 32.3±12.8 (n=11) 47.1±19.3 31.0±13.3 (n=31) - - 2316 21.3±9.6 12.5±3.4 (n=17) 53.6±15.4 34.0± 9.2 (n=15) - - - - 2338 - - 56.7±19.4 26.0± 8.4 (n=6) 46.1±14.2 29.9±11.2 (n=19) 62.9±18.3 30.5±8.0 (n=11)

S1-S2 2179 42.7±11.2 13.9±2.5 (n=8) 35.6± 9.7 14.7± 3.4 (n=82) - - 32.1± 3.5 16.9±2.5 (n=8) 2300 37.7±14.1 12.2±3.1 (n=22) 38.4±13.9 14.7± 3.2 (n=45) - - 32.2± 4.6 22.1±3.1 (n=15) 2316 ------2338 ------

* diameter (µm) ±S.D. - indicates that in the respective area less than 5 labeled neurons were found.

28 lumbosacral-PAG projections

S1 S2 Figure 8 Darkfield-polarized light photographs showing horizontal sections of the sacral cord of case 2187. On the left, section through the dorsal horn of S1. The arrow indicates a group of labeled cells which are separated from the remaining gray matter of the dorsal horn by a small stroke of white matter. On the right, section of S2 dorsal to the central canal. The arrows indicate four small clusters of labeled neurons in the lateral part of the gray matter with rostrocaudal intervals from 220-300 µm. Bar represents 200 µm.

Ventrolateral intermediate PAG and adjacent tegmentum Dorsal PAG In another control case, with an injection Injections in the ventrolateral part of the intermediate limited to only the dorsal PAG (case 2316 of group B; PAG and adjoining tegmentum resulted in a very small Fig. 2), labeled profiles were almost exclusively present number of labeled profiles in the lumbosacral segments in laminae I (68%) and V (30%; Figs. 10 and 11), and (49, 5, and 1 in all one out of five consecutive sections virtually absent in laminae VIII and medial VII, and in of cases 2239, 2250, 2248). Of all neurons in these cases, lamina X. However, in sharp contrast to the majority of 3% was found in lamina I, 19% in lamina V, 80% in labeled lamina I neurons in the cases of group A, the lamina VII, and 3% in lamina X, which distribution labeled lamina I and V neurons were evenly distributed pattern was similar to that in cases 2336 and 2338 (Fig. over the lumbosacral segments (Fig. 11). The labeled 10). lamina V neurons were large sized (Table I) and The tegmentum ventrolateral and ventral to the PAG multipolar, similar to the large diameter lamina V seems to receive some afferents from the sacral cord. neurons in group A, in which they constitute only a An injection in this area, which avoided the lateral and relatively small part of the total number of lumbosacral- ventrolateral PAG except for the needle track (case 2167; PAG cells (16% and 6% in cases 2300 and 2179, Fig. 2), resulted in a limited number of labeled neurons respectively compared to 30% in case 2316). The other in the lumbosacral cord (Fig. 5). dorsal PAG injections resulted in a very small number

Figure 9 On the left, dark- field-polarized light photo- graph of the dorsal horn of L6 (case 2179) showing a large sized, multipolar labeled neuron in the lateral part of lamina V which does not belong to group I. On the right, darkfield-polarized light photograph of the medial gray matter of L7 (case 2179) showing intermediate sized, multi-polar labeled neurons (group II) in the medial part of lamina VII and in lamina VIII. Bar represents 200 µm.

29 Chapter 1

A B C DM DL DM DL DM DL

VM VL VM VL VM VL ON L7 S2 -5mm level 0 5mm Dorsolateral (DL) Dorsomedial (DM) Ventrolateral (VL) Ventromedial (VM)

50

Case 40 ipsi 49% 40 contra 30 30 2179 30 30 # neurons 20 20 20 11% 12% 20 28%

10 10 10 10

0 0 0 0 total: 1072 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15

50

Case 40 66% 40 30 2182 30 30 30 # neurons 20 20 13% 20 8% 20 13% 10 10 10 10 total: 833 0 0 0 0 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 Case 50 40 66% 40 30 30 2300 30 30 # neurons 20 20 14% 20 10% 20 10%

10 10 10 10

0 0 0 0 total: 1072 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 Case 50 40 40 2316 30 30 30 30 # neurons 74% 20 20 5% 20 13% 20 8% 10 10 10 10 total: 316 0 0 0 0 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 Case 50 40 40 2338 30 30 30 30 # neurons 20 30% 20 14% 20 12% 20 44%

10 10 10 10 total: 327 0 0 0 0 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 -10 - 5 0 5 10 15 rostrallevel(mm) cau rostrallevel(mm) cau rostrallevel(mm) cau rostrallevel(mm) cau

Figure 10 Rostrocaudal distribution of labeled profiles in different quadrants of the spinal gray matter in 5 different cases. Labeled profiles were counted in every 12 alternate consecutive sections. For a definition of the quadrants and levels, see the Material and Methods section. Note that the dorsolateral region caudal to level 0 contains most labeled profiles in case 2300. In cases 2316 and 2338, this region contains a relatively small number of labeled neurons which are evenly distributed along the different levels. In the ventromedial quadrant, labeled neurons are most numerous rostral to level 0 in cases 2179 and 2338, and virtually absent in case 2316. ON= nucleus of Onuf

2179 2182 2239 2300 2316 2338

L7 L7 L7 L7 L7 L7

S2 S2 S2 S2 S2 S2

Figure 11 Schematic drawings of the distribution of retrogradely labeled neurons in the L7 and S2 segments after injections involving different parts of the dorsal midbrain. Each drawing shows the labeled cells of 12 alternate consecutive sections.At the level of L7, labeled neurons were present in lamina VII-VIII only when the injection involved the lateral PAG and/or the adjacent tegmentum. Labeled neurons in lamina I were found after injections in the PAG, but not when the injection was placed in the adjacent tegmentum. In S2, large numbers of neurons were present after injections involving the lateral or ventrolateral part of the caudal PAG, but not after injections limited to the dorsal PAG, the intermediate ventrolateral PAG, or the laterally adjacent tegmentum.

30 lumbosacral-PAG projections

of labeled profiles in the lumbosacral cord (26, 0, 2, and 16 in all one out of five consecutive sections of cases 2227, 2254, 2259, and 2262, respectively). Forty percent of these neurons was located in lamina I, 58% in lamina V, 1% in lamina VII, and 1% in lamina X.

Rostral PAG In control case 2159 (group B; Fig. 1), with an injection in the lateral part of the rostral PAG, the laminar distribution of labeled neurons in the lumbosacral cord was similar to that in group A, but the C8 number of labeled profiles was low (Fig. 4). It indicates that only a small part of the lumbosacral-PAG projections terminates in the rostral PAG.

Trajectory of axons of PAG projecting neurons in the spinal white matter T7 In cases 2161 and 2179, with large injections in the dorsal midbrain (Fig. 1), retrogradely labeled lumbo- sacral neurons were very strongly labeled and their retrogradely labeled axons could be followed in the white matter. Most of these retrogradely labeled axons crossed the midline in the ventral gray commissure, L5 slightly rostral to the level of their neurons of origin (Fig. 12 S1-L7). Retrogradely labeled axons were almost exclusively observed in the white matter ipsilateral to the midbrain injection site. From caudal to rostral in the S1-L7 segments, the retrogradely labeled axons strongly increased in number and they gradually shifted from ventromedial to dorsolateral. At the level of L5, most of them were located dorsolaterally in the white matter L7 (arrows in Fig. 12), whereas a few labeled fibers traveled through the ventromedial funiculus. In the thoracic cord, they were mainly present in the dorsolateral funiculus. At the level of C8, the bundle of labeled axons in the dorsolateral funiculus shifted into a peripheral position (arrows in Fig. 12). S1rostral The majority of these ascending labeled fibers was only visible under high magnifications (200x) in contrast to the few larger diameter anterogradely labeled fibers, which originate from a relatively small population of neurons in the ventrolateral PAG (Mouton and Holstege, caudal 1994). These thick PAG fibers descended via the lateral, S1 ventrolateral and ventromedial funiculi to terminate in the ipsilateral medial intermediate zone. Since spinal cord-projecting neurons in the PAG project densely to cervical and thoracic levels and virtually not to the sacral cord (Mouton and Holstege, 1994), the majority of S2rostral labeled fibers in the lumbosacral of figure 12 must represent retrogradely labeled ascending fibers.

S2caudal

Figure 12 Drawing showing the trajectory of retrogradely labeled axons in the white matter of case 2161, with a large injection in the dorsal midbrain. Note that at the transition between S1 and S2 the number of labeled fibers in the white matter rapidly increases. In L7 and S1, they are present in all funiculi, except the dorsal funiculus. Rostrally, they gradually shift dorsally to become located in the dorsolateral funiculus (arrows).

31 Chapter 1

Anterograde tracing study L6 Injections in L6 (cases 2249 and 2278) resulted in Although the retrograde study showed that distinct a much larger number of anterogradely labeled fibers in populations of lumbosacral neurons project to the PAG, the PAG, especially caudally, than an injection in L5. In only rough conclusions could be drawn about which the most caudal part of the PAG most labeled fibers were parts of the PAG receive these projections. This found in the dorsolateral and ventrolateral regions (Fig. necessitated an anterograde tracing study using 8 cases 15 P 1.5; Fig. 17). Slightly more rostrally, these with WGA-HRP injections at different levels of the distribution areas converged within lateral PAG, leaving lumbosacral cord. Figure 13 shows the location of the the dorsolateral and the lateral part of the ventrolateral injections and the eventual hemisections. PAG almost devoid of labeled fibers (Fig. 15, A 0.5-P 0.5). Dense projections were present in the most lateral Ipsi- and contralateral projections to the PAG peripheral portion of the lateral PAG and the adjoining In all 8 cases, both sides of the PAG contained tegmentum, but only at the levels A 1.5-P 0.5 (Figs. 15 anterogradely labeled fibers, but the projection and 17). The dorsal, ventral and central parts of the PAG contralateral to the side of the hemisection (cases 2278, received only a few fibers from L6. Rostral to level A 2295, 2302, 2309, and 2329) or injection (cases 2249, 1.5 the number of labeled axons in the lateral PAG was 2261, and 2265) was much denser. This corresponds considerably smaller than at more caudal levels (Fig. with the findings of the retrograde study. Although the 15). ipsilateral projection was much weaker, it always formed a mirror image of the contralateral projection. No L7-S3 In the 3 cases with injections in the sacral differences were observed between the lumbosacral PAG segments (2265, 2302, and 2329; Fig. 13) labeled fibers projections of male and female cats. were most numerous in the lateral parts of the caudal PAG (Figs. 16 and 17). Especially the central regions Termination sites in the PAG adjoining the ependymal layer received many afferents, L5 After an injection in the L5 segment (case 2309; but only few fibers terminated in the ependymal layer Fig. 14), only very few labeled fibers observed. They itself. The lateral parts of the lateral PAG or the adjoining were distributed in the lateral, dorsal and ventral parts tegmentum contained only few labeled fibers most of of the caudal PAG. Rostrally in the PAG, they were which seemed to represent fibers of passage on their almost absent. way to more medial parts of the lateral PAG. However,

Case Level injection WGA-HRP Rostrocaudal extent Hemisection injection site (mm)

2309 male L53 µl 2.5% 6.2 L3

2249 female caudal L5-rostral L7 4 µl 2% 9.6 ---

2278 female caudal L5-rostral L7 3 µl 2.5% 8.8 L3

2295 female caudal L6-L7 6 µl 2.5% 8.8 L4

2261 female caudal L7-S1 3 µl 2.5% 6.9 ---

2302 male S1-S2 2 µl 2.5% 6.9 L4

2329 male S1-S3 3 µl 2.5% 14.7 L5

2265 female S1-S2 2 µl 5% 9.0 ---

Figure 13 Schematic overview of the level and size of the WGA-HRP injections and eventual hemisections in the lumbosacral cord.

32 lumbosacral-PAG projections

Case 2309

Hemisection: L3

Injection site: L5 A 1.5

A 4.5 A 0.5

A. 3.5 P 0.5

A 2.5 P 1.5

Figure 14 Distribution pattern of anterogradely labeled fibers at different rostrocaudal levels in the PAG and adjacent tegmentum after an injection in L5 (case 2309). Note the sparse, diffuse projection to the lateral and dorsal parts of the caudal PAG.

33 Chapter 1

Case 2249

Injection site: L6

A 1.5

A 4.5 A 0.5

A. 3.5 P 0.5

A 2.5 P 1.5

Figure 15 Distribution pattern of anterogradely labeled fibers at different rostrocaudal levels in the PAG and adjacent tegmentum after an injection in L6 (case 2249). Note the dense projection to a circumscribed region in the lateral part of the lateral PAG and adjacent tegmental field (A 0.5 and P 0.5).

34 lumbosacral-PAG projections

Case 2329

Hemisection: L5 Injection site:

S1 S2 S3 A 1.5

A 4.5 A 0.5

A. 3.5 P 0.5

A 2.5 P 1.5

Figure 16 Distribution pattern of anterogradely labeled fibers at different levels in the PAG and adjacent tegmentum after an injection in S1-S3 (case 2329). Note the very dense projection to a circumscribed region in the medial part of the lateral PAG.

35 Chapter 1

this was not so for the tegmentum ventrolateral and observed labeled neurons in lamina I, some in lamina ventral to the PAG (A1.5; Fig. 16), which received a V, and only a few in lamina VII. In the S3 segment, they substantial projection from the sacral segments (see found neurons mainly in laminae V and VII and to only retrograde case 2167). Similar to the L6 projections, a limited extent in lamina I. Menetrey et al. (1982) in labeled fibers were sparse in the lateral part of the the lumbar enlargement of the rat found a similar ventrolateral PAG and in the dorsal and ventral PAG. distribution pattern. Their plottings show laminae I and Rostral to P 0.5, the dorsolateral PAG was almost devoid V neurons projecting mainly to the lateral PAG and of labeled fibers. lamina VII neurons projecting mainly to the adjacent In case 2261 with an injection in the caudal L7 and S1 tegmentum. segments and in case 2295 with an injection in the caudal In previous anterograde tracing studies describing L6 and L7 segments (Fig. 13; Blok et al., 1995), the lumbosacral-PAG projections, a specific projection distribution pattern of labeled fibers in the PAG was a pattern as described in the present study could not be combination of those observed in the L6 and S1-S3 found due to the very large extent of the injections injected cases. In the PAG rostral to level A 1.5, the L7 involving the L3-S2 (Yezierski, 1988 cat, rat, monkey), projections were less dense than the sacral cord L5-L7 (Wiberg and Blomqvist, 1984 cat), or L2-Co1 projections. segments (Wiberg et al., 1987 monkey). As in the present results, both Yezierski (1988) and Wiberg and Blomqvist (1984) observed a shift of terminating fibers from the DISCUSSION dorso- and ventrolateral caudal PAG to the lateral PAG at the intercollicular level. A recent electronmicros- copical study by Blok et al. (1995) provided final proof The present study demonstrates that the great majority for neurons in the lumbosacral segments to terminate in of the relatively large number of PAG-projecting neurons the PAG. in the sacral and lower lumbar cord is located in two Axons of sacral-PAG neurons ascend in the ventral distinct cell groups. quadrant in the lower lumbar cord to become located The first and by far largest group consists of small sized more dorsolaterally at thoracic and cervical levels. A neurons located in the lateral parts of lamina I of the similar shift of ascending fibers from the ventral funiculi L7-S1 segments, gradually extending from lamina I into in the caudal segments to the dorsolateral funiculi in the lateral parts of laminae V and VII at the level S2-S3. cervical segments, has been described in a study on Their main projection target in the midbrain is a spinothalamic fibers in man (Walker, 1940). The finding circumscribed area in the medial part of the lateral caudal of spinomesencephalic fibers in the dorsolateral quadrant PAG. This group will be referred to as group I (Fig. is in line with the report in the cat of Hylden et al. 18A). (1986a), and also fits with lesion studies in the cat on The second group of PAG projecting neurons is located ascending micturition (Barrington, 1933) or pain in an area at the transition of laminae VII and VIII at the pathways (Kennard, 1954). level caudal L6-rostral S1. Its main midbrain target is a circumscribed region laterally in the lateral caudal PAG, which region gradually extends into the adjacent tegmentum rostrally. This group will be referred to as The relation between lumbosacral-PAG group II (Fig. 18B). neurons and primary afferents A small number of labeled neurons does not take part in groups I and II. They are small neurons in lamina I Group I and a few large multipolar cells in the lateral part of Pelvic nerve primary afferents The pelvic nerve lamina V. In contrast to groups I and II, they are evenly contains mainly finely myelinated (A∂) and un- distributed throughout the L4-Co3 segments. The myelinated (C) fibers that convey information from injection in the L5 segment, which contains almost pelvic viscera as bladder, colon, vagina, vaginal cervix, exclusively neurons in lamina I and V, showed that these and uterine cervix to the spinal cord (de Groat et al., neurons project rather diffusely to the dorsal and lateral 1978; Morgan et al., 1981; Kawatani et al., 1990; Fig. PAG. The large lamina V neurons project also to the 19). Pelvic primary afferents enter the spinal cord via tegmentum laterally adjoining the PAG (Fig. 18C). the S1-S3 dorsal roots and project via Lissauer’s tract The lumbosacral cord projects more strongly to the to the dorsolateral part of lamina I of L7-rostral S1. The midbrain than any other part of the spinal cord, except strongest projections are to lamina I along the lateral for the upper cervical segments (in the rat: Liu, 1983; part of the dorsal horn in S1 and to the lateral laminae V cat: Wiberg and Blomqvist, 1984; VanderHorst and and VII in S2 (Morgan et al., 1981; Fig. 19). This Holstege, 1992; and monkey: Wiberg et al., 1987). distribution pattern almost exactly matches the Wiberg and Blomqvist (1984), using large midbrain distribution pattern of labeled neurons after injections injections in the cat, studied the L6-L7 and S3 segments in the lateral PAG. A limited number of pelvic fibers of the lumbosacral cord. In the L6-L7 segments, they project to more medial parts of the dorsal horn and to

36 lumbosacral-PAG projections

Injection L6 Injection S2 A 0.5 A 0.5

P 0.5 P 0.5

Figure 17 Low power photomicrographs of the caudal PAG and the immediately adjacent tegmentum after an injection in L6 (case 2249, left) and caudal S1-S2 (case 2302, right). Note that the L6 segment projects most densely to a circumscribed area in the lateral part of the lateral PAG and adjacent tegmentum, whereas the S1-S2 segments project heavily to the medial part of the lateral PAG. Bar represents 500 µm.

the dorsal gray commissure. It is not known which of Pudendal nerve primary afferents The pudendal nerve the pelvic organs project to this part of the spinal gray. conveys afferent information from somatic structures Electrophysiologically, it has been shown that inter- (skin of penis, clitoris, and perineum, the striated pelvic neurons in the lateral laminae V and VII of S2 receive muscles, and the mucosa of urethra and anus) as well as strong mono- or polysynaptic input from vesical and/or from visceral structures (urethra, vagina, and part of the intestinal afferents via the pelvic nerve (de Groat et al., uterine cervix). In the cat, the main target of pudendal 1979, 1980, 1981; Milne et al., 1981). In S2 in cats as nerve afferents is in S1-S2. Among the pudendal well as in monkey, the pelvic nerve termination pattern afferents, those from the penis and clitoris are most shows a rostrocaudal periodicity with intervals of numerous and terminate in the medial part of laminae V approximately 200 µm from centre to centre (Morgan and VII and in lamina X (Ueyama et al., 1984; Thor et et al., 1981; Nadelhaft et al., 1983). These distances are al., 1989; Kawatani et al., 1990, 1994; Fig. 19). Afferents similar to the intervals between the clusters of group I from pelvic floor muscles, perineal skin, and visceral PAG projecting neurons at this level. structures terminate laterally in laminae I, V and VII Clustering of neurons in this area has already been (Ueyama et al., 1984; Thor et al., 1989; Fedirchuk et described by Brown and Nolan (1979) using a pyridine al., 1992; Kawatani at al, 1994), which matches with silver stain. These clustered neurons are not para- the distribution pattern of retrogradely labeled neurons sympathethic preganglionic motoneurons because these after injections in the lateral PAG (Fig. 19). Electro- motoneurons do not form clusters (Nadelhaft et al., physiologically, interneurons in the lateral laminae V 1980). In all likelihood, the clustered neurons of Brown and VII of S2 receive strong monosynaptic input from and Nolan (1979) are the same cells as the clustered mechanoceptor and noci-ceptive afferents from the sex group I PAG-projecting neurons, also because their organs and perineal area (de Groat et al., 1979, 1980, location, size, and dendritic orientation are almost 1981; Milne et al., 1981; McMahon, 1986; Fedirchuk identical. Moreover, the clusters of group I sacro-PAG et al., 1992; Berkley et al., 1993). Similar to the pelvic neurons seem to precisely overlap with the bundles of nerve afferents, pudendal afferents in the lateral lamina pelvic nerve terminals. V and the dorsal gray commissure also form small

37 Chapter 1

ABC

L5 L5 L5

L7 L7 L7

S2 S2 S2

Figure 18 Schematic representation of the main conclusions, showing that the group of neurons in lamina I of L6-S2 and laminae V and VII of S2 (group I) project to the medial part of the lateral caudal PAG (A), neurons in medial lamina VII and VIII of the lower lumbar segments (group II) project to the lateral part of the lateral caudal PAG and the adjacent tegmentum (B), and that lamina I and large lamina V in all spinal segments project diffusely to the lateral and dorsal PAG (C).

clusters with intervals of 150-300 µm (Ueyama et 1988; Gillette et al., 1993a,b). However, deep afferents al.,1984; Thor et al., 1989). from different muscles and joints of hindlimbs and tail In conclusion, the laminar distribution of group I sacral- terminate in the dorsal horn at different lumbosacral PAG neurons strongly overlaps with the area of levels (Craig et al., 1988; Mense and Craig, 1988) and, termination of pelvic and pudendal afferents as far as unlike pelvic visceral afferents, are not restricted to S1- they convey information from the perineum and pelvic S3. Electrophysiologically, it has been shown that viscera. Only a limited number of sacro-PAG neurons neurons located in the same area as the group I PAG is located among afferents from the penis or clitoris. projecting neurons receive convergent input from somatic and visceral structures (Fields et al., 1970; de Small diameter primary afferents from muscles and Groat et al., 1979; McMahon and Morrison, 1979; Milne joints of the hindlimbs, back, and tail The laminar et al., 1981; Honda, 1985; Grigg et al., 1986; McMahon, distribution of deep somatic afferent fibers (type III and 1986; Schaible et al., 1987a,b; Berkeley et al., 1993). IV) from muscles and joints resembles that of fibers Many of these neurons have been shown to project to from the viscera (Craig et al. 1988; Mense and Craig, the midbrain (Hylden et al., 1986b; Yezierski and

38

. lumbosacral-PAG projections

bladder, colon, PAG-projecting vagina, vaginocervix, neurons uterine cervix, pelvic floor, perineal skin

penis and clitoris parasympathetic preganglionic motoneurons

L7 S1 rostral S1 caudal S2 S3

Figure 19 Schematic overview of the location of the pelvic and pudendal primary afferents, the group I PAG projecting neurons, and the sacral parasympathetic preganglionic motoneurons. For references of primary afferents and motoneurons see text. The PAG projecting neurons are taken from Figure 6 (case 2300).

Schwartz, 1986). Altogether, these findings lead to the the same structures as described for the group I neurons, idea that the group I PAG-projecting neurons receive i.e. pelvic viscera, muscles, joints and skin of the input from deep somatic as well as visceral parts of the hindlimbs and back (see Willis, 1985; Loewy and Spyer, body. 1990; Ritz and Greenspan, 1985). However, the even distribution of these lamina I and V neurons throughout Small diameter primary afferents from skin of the the lumbosacral segments and in all likelyhood hindlimbs, back, and tail Small diameter afferents throughout the whole spinal cord (Mouton et al., in conveying nociceptive input from the skin of hindlimbs preparation) suggests that they are not involved in the and tail terminate pre-dominantly on lamina I and II conveyance of specific information, such as from the neurons throughout the lumbosacral cord (Besson and pelvic viscera. This idea is supported by the anterograde Chaouch, 1987; Schaible et al., 1987). Many of these tracing results of the L5 case of the present study, which lamina I neurons respond exclusively to nociceptive shows a rather diffuse projection throughout the dorsal cutaneous input and project to the midbrain (Hylden et and lateral PAG. Moreover, preliminary findings of al., 1986b). This is in contrast to the neurons receiving Mouton et al. (in preparation) demonstrate that the visceral and/or deep somatic stimuli, of which only very distribution pattern within the PAG from L5 is indeed few have been reported to respond to exclusively similar to that from C8. Recent findings of Craig (1995) nociceptive or exclusively non-nociceptive input (Craig in the cat partly support this hypothesis. He found that, and Kniffki, 1985; Besson and Chaouch, 1987; Jänig, apart from a rostrocaudal shift, a similar distribution 1987; Ness and Gebhart, 1991). pattern in the PAG from lamina I neurons in C7 as from lamina I neurons in L7. In the L4-L6 and the coccygeal Group II segments, the lamina I neurons, which do not belong to Laminae VII and VIII of L6-L7 do not receive significant group I, probably receive direct input from specific primary afferent input, except from Ia afferents thermoreceptors and nociceptors in the skin of hindlimbs (Janskowska and Lindström, 1972; Ishizuka et al., 1979). and tail (Christensen and Perl, 1970; Willis and Apparently, group II lumbosacral PAG-projecting Coggeshal, 1978; Cervero et al., 1979; Craig and Kniffi, neurons receive their information mainly indirectly 1985; Hylden et al., 1986b; Besson and Chaouch, 1987; (Kow et al., 1980) from group I and II afferents as well Ferrington et al., 1987; Yezierski et al., 1987). The large as from cutaneous, joint and interosseus nerves (Edgley sized midbrain projecting neurons in lamina V have more and Jankowska, 1987). This view is supported by the complex response properties (wide dynamic range), finding that many of the midbrain projecting neurons in similar to the group II PAG projecting neurons these laminae have large, complex (excitatory and (Menetrey et al., 1982; Yezierski and Schwartz, 1986). inhibitory, ipsi- and contralaterally) receptive fields and have highly convergent (indirect), nociceptive and non- nociceptive input from cutaneous and deep receptors Functional implications of the lumbo-sacral (Yezierski and Schwartz, 1986; Yezierski, 1991). cord-PAG pathway Remaining neurons in laminae I and V The neurons in laminae I and V, not belonging to group The present study has demonstrated that different I, might receive, because of their location, input from neuronal cell groups in the lumbosacral cord project to

39 Chapter 1

different parts of the PAG. Such an organization suggests micturition (Skultety, 1963), facilitates bladder reflexes that these groups convey functionally different (Skultety, 1959; Koyama, 1962; Gjone, 1966), enhances information. In respect to the lumbosacral input to the spontaneous discharges of the vesical branche (Fukuda PAG, the medial part of the lateral PAG seems to receive and Koga, 1991), and reduces bladder capacity (Kruse mainly pelvic visceral information and to a more limited et al., 1990). Furthermore, while stimulating bladder extent input from deep somatic structures of hindlimbs afferents in the rat, the latency of potentials observed in and back. In contrast, somatosensory input from large the most caudal dorsal PAG was much shorter (13 ± 3 areas of the skin and perhaps IA afferents from muscles ms) than from those observed in the M-region (42 ± 7 of the hindlimbs and back seems to be conveyed from ms; Noto et al.,1991). the lower lumbar cord to the lateral part of the lateral The PAG might not have an exclusively facilitory effect PAG and adjacent tegmentum. The latter region overlaps on micturition, since electrical stimulation in the cat and well with the location of the so-called intercollicular dog has been shown to inhibit the micturition reflex nucleus (Danielsson and Norrsell, 1985, 1986; (Koyama et al., 1962; Gjone, 1966; Fukuda and Koga, Blomqvist et al., 1990). Some units in the intercollicular 1991). The facilitory sites in the cat were found in the nucleus receive exclusively somatosensory input, i.e. caudal and intermediate PAG (Skultety, 1959; Gjone, light tactile cutaneous stimuli from mainly the 1966), and in the tegmentum lateral to the ventrolateral contralateral body (Danielsson and Norrsell, 1985, PAG (Koyama, 1962). Inhibitory sites resided in the 1986). This is in contrast to the overlying deep layers of rostral and intermediate PAG (Gjone, 1966), especially the superior colliculus, which also respond to visual and at its lateral border and in the adjacent tegmentum auditory stimuli. The area of termination of group II (Koyama et al., 1962). However, in a more recent study fibers supports the notion of Mehler (1969) that the in the dog, Fukuda and Koga (1991) also found lateral PAG might form a single functional unit with the inhibitory sites in the caudal and intermediate PAG. adjacent tegmentum. These electrophysiological findings suggest that The fact that the projections from the remaining neurons different parts of the midbrain are involved in various in lamina I and V are derived from all spinal levels and aspects of the micturition reflex, but these effects might terminate diffusely in the lateral and dorsal PAG, also be due to electrical stimulation of fibers of passage suggests that this pathway has a general function in on their way to the PAG or the pontine micturition center. autonomic responses (Bereiter and Benetti, 1991) such as the relay of nociceptive, thermal and metabolic Lordosis information to supraspinal levels (Willis and Coggeshal, Lordosis behavior is the estrogen dependant, receptive 1978; Hylden et al., 1986b; Besson and Chaouch, 1987; posture of female rodents and cats. As with the Yezierski et al., 1987; Duranti et al., 1991; Craig, 1995). micturition reflex, lordosis is integrated at supraspinal In contrast, the specificity of the PAG projections from levels (see Pfaff et al., 1994). Lordosis behavior can be group I and II neurons suggests that they are involved initiated and strengthened by applying tactile stimuli to in more specific behaviors such as micturition, mating, the skin of the flanks, posterior rump, tailbase, and, most and agressive and defensive behaviors. effectively, perineum, vagina and cervix (Kow et al., 1979 in the rat; Michael, 1961 in the cat). The majority Micturition of these stimuli are conveyed to the sacral cord through The location of the group I sacral cord PAG-projecting the pudendal and pelvic nerves (Morgan et al., 1981; neurons strongly overlaps with the area of termination Ueyama et al. 1984), although pelvic primary afferents of pelvic nerve afferents, which convey information from seem to be more important than pudendal afferents the bladder (Morgan et al.,1981). The sacral cord sends (Gomora et al., 1994). Lordosis can still be elicited by hardly any fibers directly to the pontine micturition somatosensory and vaginocervical stimulation in center or M-region (Blok et al., 1995), which is located ovariectomized rats after precollicular decerebration in the dorsolateral pontine tegmentum (Barrington, (Rose and Flynn, 1993), which indicates that direct 1925; Nathan and Smith, 1958; Holstege et al., 1979, pathways from the sacral cord to the hypothalamus 1986; de Groat and Steers, 1990) and coordinates the (Burstein et al., 1990 in the rat; Katter et al., 1991 in the micturition reflex. However, the present study shows cat) are not strictly necessary for the lordosis response. that numerous neurons located among bladder primary These findings strongly suggest that the group I PAG afferents project densely to the medial part of the lateral projecting neurons, which receive direct pelvic and to PAG (see also Blok et al., 1995). This same area in the lesser extent pudendal primary afferents, convey sensory PAG contains neurons which send fibers to the M-region input to the PAG relevant for lordosis. It must be (in the cat: Blok and Holstege, 1994; in the rat: Valentino emphasized that the group I sacro-PAG pathway is not et al., 1994). These findings suggest that this area in the the only ascending projection that can evoke lordosis caudal PAG plays an important role in the micturition behavior. It can also be triggered by tactile stimuli reflex. This idea is supported by the findings that applied to the flanks, which enter the spinal cord at more electrical stimulation in the PAG, especially at rostral levels. intermediate and caudal levels, evokes complete The importance of the PAG in lordosis behavior has been

40 lumbosacral-PAG projections well established. Stimulation in the PAG of estrogen primed animals facilitates the lordosis response, and lesions in the PAG have been shown to decrease lordosis behavior (in the rat: Sakuma and Pfaff, 1979b; in the hamster: Floody and O’Donahue, 1980). Interestingly, in the rat low threshold sites for evoking lordosis reside in the medial part of the lateral PAG (Sakuma and Pfaff, 1979a), which, as demonstrated in the present study in the cat, is the main target of sacral-PAG group I projecting neurons. The descending pathway of the lordosis reflex has recently (VanderHorst and Holstege, 1995) been proposed to be mediated by the nucleus retroambiguus, a group of interneurons in the caudal medullary tegmental field. This nucleus receives specific projections from neurons in the lateral part of the caudal and intermediate PAG (Holstege, 1989). The NRA in turn projects to distinct axial, hindlimb, and pelvic floor motoneuronal cellgroups in the lumbosacral cord, which are thought to produce the specific motor activity described as lordosis behavior (VanderHorst and Holstege, 1995).

Agressive and defensive behaviors Stimulation in the lateral PAG and the immediately adjacent tegmentum evokes active defence behavior, associated with hypertension, tachycardia, non-opioid analgesia, and vocalization (see Abrahams et al., 1960; Skultety, 1963; Chaouch and Besson, 1987; Holstege, 1989; Bandler et al., 1991; Bandler and Depaulis, 1991; Lovick, 1993; Zhang et al., 1994). When the lateral PAG is stimulated caudally, flight can be evoked, whereas confrontational defence can be elicited in the rostral part of the lateral PAG (Bandler et al., 1991; Bandler and Depaulis, 1991). Sakuma and Pfaff (1979a) showed that electrical stimulation in lateral parts of the PAG in the rat was more effective in evoking defensive behavior than stimulation in the area of the lateral PAG near the aqueduct. This suggests that the group II projection to the lateral part of the lateral PAG and the intercollicular nucleus might play a role in defensive behavior. Since the lamina I and V neurons not belonging to group I project diffusely to all parts of the dorsal and lateral PAG it is possible that these projections are important for eliciting aggressive and defensive behavior, especially because many of these neurons are known to respond to noxious stimuli.

41

lumbosacral motoneuronal cell groups

Chapter 2

43

lumbosacral motoneuronal cell groups The organization of lumbosacral motoneuronal cell groups innervating hindlimb, pelvic floor and axial muscles in the cat

ABSTRACT In a study on descending pathways from the nucleus retroambiguus (NRA) to hindlimb motoneurons (see accompanying paper), it appeared impossible, using data from the literature, to precisely determine which muscles were innervated by the motoneurons receiving the NRA fibers. This lack of data made it necessary to produce a detailed map of the lumbosacral motoneuronal cell groups in the cat. Therefore, 50 different muscles or muscle compartments of hindlimb, pelvic floor and lower back were injected with horseradish peroxidase (HRP) in 135 cases. The respective muscles were divided into 10 groups: I: sartorius and iliopsoas; II: quadriceps; III: adductors; IV: hamstrings; V: gluteal and other proximal muscles of the hip; VI: posterior compartment of the distal hindlimb; VII: anterior compartment of the distal hindlimb; VIII: long flexors and intrinsic muscles of the foot; IX: pelvic floor muscles; and X: extensors of the lower back and tail. The L4-S2 segments were cut and incubated, and labeled motoneurons were counted and plotted. A new method was developed which made it possible, despite variations in size and segmental organization between the different cases, to compare the results of different cases. The results show that the spatial interrelation between the hindlimb and pelvic floor lumbosacral motoneuronal cell groups remains constant. This finding enabled the authors to compose an accurate overall map of the location of lumbosacral motoneuronal cell groups. The general distribution of the motoneuronal cell groups is also discussed in respect to their dorsoventral, mediolateral, and rostrocaudal position within the lumbosacral ventral horn.

INTRODUCTION Waldeyer (1888) was the first to describe that much more accurate retrograde tracing methods have motoneurons are organized into longitudinal colums. A been developed, in mammals a complete, detailed few years later, Kaiser (1891) further subdivided these oversight of the lumbosacral ventral horn motoneuronal columns into a medial and a lateral group. The medial cell groups has never been published. Some studies on motoneurons innervate, via the dorsal rami of the spinal the organization of hindlimb motoneurons focussed on motor nerves, the axial muscles, whereas the lateral the innervation of different compartments of one or two motoneurons, via the ventral rami of the spinal motor particular muscles, e.g. sartorius and tensor fasciae latae nerves, innervate the muscles of the extremities as well (Gordon et al., 1991), flexor digitorum longus (Dum et as the abdominal wall, the intercostal, and pelvic floor al., 1982), tibialis anterior (Ilya and Dum, 1984), muscles (Sprague, 1948). A similar organization exists peroneus longus (Kernell et al., 1985), and lateral in the frog (Brissaud and Bauer, 1903; Silver, 1942), gastrocnemius (Weeks and English, 1985). Other reports pigeon (Huber, 1936), chick (Landmesser, 1978), rat compared the location of motoneuronal cell groups of (Angulo, 1927; Goering, 1928; Elliott, 1944), rabbit (de only two or three muscles, e.g. gastrocnemius and soleus Neef, 1900; Romanes 1941; Elliott, 1944), guinea pig (Burke et al., 1977; Weeks and English, 1987), the (Kiesel, 1938), dog (Neef, 1900; Parhon and Parhon, different peroneal muscles (Horcholle et al., 1988), or 1902; Parhon and Nadadje, 1906; Marinesco, 1904; the extensor digitorum longus, tibialis anterior and Elliott, 1944), cat (Kiesel, 1938; Elliott, 1944), rhesus semitendinosus (Hoover and Durkovic, 1991). monkey (Elliott, 1944), macaque (Reed, 1940), gorilla The lack of a complete thorough overview of the location (Waldeyer, 1888), and human (Onuf, 1900; Van of the lumbosacral cord motoneurons appeared to be a Gehuchten and de Neef, 1900; Elliott, 1942; Sharrard, major problem in the study on the descending pathways 1955; Elliott, 1944), and seems to be most distinct in from the nucleus retroambiguus (NRA) to the lumbo- cats and humans (Elliott, 1944; Sharrard, 1955). sacral ventral horn (VanderHorst and Holstege, 1995), Only after the development of tracing techniques, more and even more precisely in the accompanying paper. detailed results could be obtained. For example, The NRA was found to project to certain distinct Romanes (1951, 1964), using the retrograde lesion- motoneuronal cell groups, for which it was not possible degeneration technique in the cat, found that moto- to precisely determine which muscles they innervated. neurons of hindlimb muscles acting on the same joints In order to solve this problem it was necessary to study were organized in the same longitudinal columns. the location of the various motoneuronal cell groups in Despite the fact that since Romanes’ (1951, 1964) study a separate project.

45 Chapter 2

A major problem concerned the comparability of the Countings and rostrocaudal distribution different cases, because the various spinal cords differ For each case, all the 1:2 collected sections were considerably in length and in subdivision into segments microscopically examined with a Zeiss Axioskop under (Sherrington, 1892; Romanes, 1951; Rexed, 1954). a combination of polarized light and darkfield condensor Romanes (1951) described the latter phenomenon in (magnification 100x) and the total number of labeled terms of pre- or postfixation of the lumbosacral plexus. neurons was counted. A labeled structure was considered According to the present findings, the only constant a motoneuron when the cell body had one or more factor was the spatial interrelation between the branches (axons or dendrites; Iliya and Dum, 1984). lumbosacral motoneuronal cell groups. Based on this All labeled motoneurons, large (probably alpha) as well observation, the authors developed a new method which, as small (probably gamma) were included. It should be for the first time, made it possible to produce an accurate emphasized that the object of this study was not to map of the location of 50 different motoneuronal cell precisely determine the number of motoneurons groups in the lumbosacral cord of the cat. innervating a particular muscle, but to determine their location in the ventral horn. MATERIALS AND METHODS A Segmental level

20 case 2191 R 10 In 74 adult male and female cats (2.9-6.4 kg), 135 0 L5 L6 L7 injections of 10% horseradish peroxidase (HRP; Sigma) S1 were made in 50 different hindlimb, pelvic floor and 30 20 case 2239 R 10 axial muscles or muscle compartments. In most cats, 0 L5 L6 L7 two different muscles (or muscle compartments) were S1 injected on the left and right side, respectively. number of labeled neurons 30 20 case 2317 R 10 The injected muscles have been divided into 10 groups. 0 L5 L6 L7 S1 For each group, an overview is given of the injected level in segments muscles, amount and type of tracer injected, survival time as well as origo, insertion, function and innervating B Length in milimeters 20 case 2191 R nerve (tables I to X). 10 0 0.00 2.4 4.8 7.2 9.6 12.0 14.4 16.8 19.2 21.6 24.024 26.4 Surgical and histological procedures 30 General surgical and histological procedures as decribed 20 10 case 2239 R in chapter 1 were used in the present study. After incision 0 0 2.4 4.8 7.2 9.6 12 14.4 16.8 19.2 21.624 24 26.4 28.8

of the skin, the muscles were exposed and identified. number of labeled neurons 30 Extreme care was taken that the covering fasciae were 20 10 case 2317 R left intact to prevent leakage of the tracer (Burke et al., 0 0 2.4 4.8 7.2 9.6 12 14.4 16.8 19.2 21.6 2424 26.4 1977). A single penetration was made through the fascia length in milimeters with a 50 or 100 µl Hamilton syringe and plain HRP was slowly injected into the muscle. Within the muscle, C Level relative to size of the enlargement the needle was moved into different directions in order 0 10203040506070 80 90 100 to more completely fill the muscle with HRP. After 20 retraction of the syringe, the area of needle entrance was 10 case 2191 R 0 immediately and thoroughly cleaned with saline, and L5 L6 L7 S1 the skin was carefully closed. 30 After 2-3 days survival time, the animals were perfused 20 10 case 2239 R 0 and the spinal cord was removed (see Chapter 1).The L5 L6L7 S1

lumbosacral cord was divided into segments based upon number of labeled neurons 30 the grouping of the dorsal rootlets. In all cases, the L4- 20 10 case 2317 R S1 segments were cut on a freezing microtome into 40 0 L5 L6 L7 S1 µm serial transverse sections. Every second section was 0 10203040506070 80 90 100 collected and processed using the tetramethyl-benzidine level (TMB) procedure according to Mesulam (1982). In the cases with injections in the intrinsic foot muscles, pelvic Figure 1 Rostrocaudal distribution of labeled medial gastrocnemius motoneurons in the lumbosacral cord in 3 floor, or tail, the S2 and S3 segments were processed as different cases. On the left (A), the rostrocaudal distribution well. Finally, the sections were mounted on slides, is related to the subdivision into segments, in the middle (B) dehydrated and coverslipped with DePeX mounting to the subdivision into millimeters, and on the right (C) to the medium. relative size of the enlargement. Note that the rostrocaudal distribution of the 3 cases only matches in C.

46 lumbosacral motoneuronal cell groups

In order to study the rostrocaudal distribution of the motoneurons of the various muscles, the number of Spatial interrelation between the motoneuronal labeled motoneurons was determined per 6 (1:2) cell groups consecutive sections, which corresponds with a rostro- In order to determine the spatial relation between the µ caudal length of the spinal cord of 480 m (6 x 2 x 40 lumbosacral motoneuronal cell groups, two character- µ m). istics of the spinal cord had to be taken into account. First, among the individual cases there exist major differences in the organization of the dorsal and ventral rootlets (Romanes, 1951; Sherrington, 1892). As a Level 0 consequence, the length and level of the segments varied A considerably (Fig. 1A). Second, the absolute length of the lumbar enlargement also varied greatly between the cases (Fig. 1B). On average, the length was 28.8mm ± 2.4mm, but it ranged between 23.0mm and 34.1mm rostral caudal (n=120). To correct for these differences between the cases, two easily recognizable landmarks were chosen rostral and caudal in the enlargement. The rostral landmark (level 0) was defined as the very beginning of the lateral extension of the ventral horn in the midlumbar cord Level 100 (caudal L4; Fig. 2A middle section), which later B appeared to correspond with the rostral pole of the psoas major motoneuronal cell group. The caudal landmark (level 100) was located in the upper sacral cord, and rostral caudal corresponds with the level at which the border of the ventral horn, instead of a rounded curvature (Fig. 2B, left section), suddenly follows a straight line between the ventromedial and dorsolateral corners of the ventral horn (Fig. 2B, middle and right sections). With this 0- Figure 2 Schematic drawing of transverse sections marking 100 scale, the spatial interrelation between the level 0 (A) and 100 (B) of the lumbosacral enlargement. Caudal motoneuronal cell groups in the different cases could to level 0, the ventral horn starts to bulge out in the white be determined adequately (Fig. 1C). matter due to the presence of psoas major motoneurons. Level 100 is characterized by a sudden change in the shape of the Plottings ventral horn due to the disappearance of a large number of motoneurons such as those innervating gluteal muscles, In each case, labeled motoneurons were plotted under hamstrings and triceps surae, i.e. at level 100 the ventrolateral darkfield illumination at low magnification using a Zeiss border of the ventral horn suddenly forms a straight line instead microscope. One drawing contains the informa-tion of of the round curvature just rostral to this level (B, left section). 12 (1:2) sections representing a length of (12 x 2 x 0.04

Table 1 Group I Iliopsoas and sartorius number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Other Remarks (1951) Iliopsoas 2242 L 50 µl type II *222 * =137 lateral/85 medial 2253 L 50 µl type II *453 * =260 lateral/193 media 2277 L 40 µl type II *319 * =159 lateral/160 media

- psoas minor O: 5 heads from caudal thoracic Flexion of pelvis Thoracic and lumbar 2348 R 15 µl type VI *119 abdominal approach and rostral lumbar vertebrae to LS column spinal nerves * =119 lateral/0 medial I: iliopectineal line, just caudal of acetabulum - psoas major O: from the 5 tendons of psoas minor, Flexion hip Lumbar 2311 L 30 µl type VI * 5 6 abdominal approach processus transversus of L5, and Fixation spine spinal nerves * =56 lateral/0 medial corpus of caudal 4 lumbar vertebrae Flexion spine 2348 L 20 µl type VI * 4 7 abdominal approach I: apex of the lesser trochanter External rotation * =47lateral/0 medial of the femur thigh - iliacus O: ventral border ilium Fixation spine Lumbar I: as for psoas major Flexion spine spinal nerves

Sartorius O: crista iliaca Flexion hip n. femoralis 2193 R 70 µl type II 3 0 3 2 7 4 Gordon et al. (1991) I: medial proximal tibia, Flexion knee 2215 L 80 µl type II 3 3 6 2 3 6 / 407/112-140 medial epicondyle, and Extension knee 2317 L 30 µl type VI 1 6 8 patella and adjacent fascia - medius O: crista iliaca Flexion hip n. femoralis 2264 R 60 µl type II 8 4 Gordon et al. (1991) I: medial proximal tibia and Flexion knee 2338 R 15 µl type VI 2 3 4 22-52 medial epicondyle 2357 R 20 µl type VI 6 1

- anterior O: crista iliaca Flexion hip n. femoralis 2264 L 60 µl type II 2 1 5 Gordon et al. (1991) I: patella and adjacent fascia Extension knee 2338 L 15 µl type VI 2 8 3 49-173

47 Chapter 2

mm)= 0.96 mm. The drawings were made using Group I a drawing tube, digitizer and Macintosh -20 -100102030405060 70 80 90 100 computer. All plottings were ordered according Iliopsoas -lateral group- to the 0 to 100 levels. For each muscle, 20 10 case 2242 L representative cases were shown. Overview 0 L5 L6 L7 S1

drawings showing the spatial interrelation of the 30 20 motoneuronal cell groups were composed by 10 case 2253 L 0 superimposing these representative cases. L5 L6 L7 S1 20 10 case 2277 L 0 L5 L6 L7 S1 RESULTS Iliopsoas -medial group- 20 10 case 2242 0 L5 L6 L7 S1 The muscles innervated by motoneurons in the 20 10 case 2253 L lumbosacral cord have been subdivided into 10 0 L5L6 L7 S1 groups, which will be presented successively. For 20 each group, a table gives an overview of the 10 case 2277 L 0 origin, insertion, function, and nerve supply of L5 L6 L7 S1 Psoas major the various muscles, and the number of labeled 20 10 case 2311 L motoneurons in each case. 0 L5 L6 L7 S1 20 10 case 2348 L 0 L5L6 L7 S1

Psoas minor Group I 30 20 10 case 2348 R Iliopsoas 0 L1 L2 L3 L4 L4 L5L6 L7 S1 L5 L6 L7 S Co Sartorius 30 20 10 case 2193 R psoas minor 0 L5 L6 L7 S1

40 30 20 10 case 2215 L 0 L5 L6L7 S1

30 lateral aspect 20 10 case 2317 L 0 L5 L6L7 S1 -anterior- L4 L5 L6 30 L7 20 S 10 case 2264 L Co 0 L5 L6 L7 S1

30 20 10 case 2338 L 0 iliacus L5 L6 L7 S1 -medius- 30 20 10 case 2264 R 0 L5 L6 L7 S1 30 lateral aspect 20 10 case 2338 R 0 L5 L6 L7 S1 L4 20 L5 L6 L7 10 case 2357 R S 0 Co L5L6 L7 S1

0 10203040506070 80 90 100 level psoas major

Figure 4 Graphs showing the rostrocaudal distribution of labeled motoneurons after injections into the iliopsoas and sartorius muscles (group I). Note that, in contrast to the lateral aspect laterally located group of iliopsoas motoneurons, the medial group of motoneurons starts at the rostral beginning of the L5 Figure 3 Schematic drawings of the iliopsoas muscles segment. The psoas major and minor have been injected via (group I). abdominal approach.

48 lumbosacral motoneuronal cell groups

AB

CD

Figure 5 Polarized light- darkfield photomicrograph of labeled motoneurons after injections into different hindlimb muscles. A: Labeled iliopsoas motoneurons at level 5 in the ventrolateral corner of the ventral horn (caudal L4 in case 2277L). B: Iliopsoas motoneurons at level 35 (mid L5 in case 2277L). Note the two groups near the ventral and medial border of the ventral horn, in all likelihood innervating the psoas major and iliacus, respectively. C: Anterior sartorius motoneurons at level 15 (rostral L5 in case 2338L). D: Adductor longus motoneurons at level 25 (L5 in case 2354R). Bar represents 400 µm.

Group I: Iliopsoas and sartorius

Iliopsoas -20 -10 010 20 L5 psoas minor case 2348R

02010 30 40 50 L5 psoas major case 2348L

L5 L6 iliopsoas case 2253L

Sartorius 02010 30 40 50 sartorius L5 L6 case 2215L

L5 L6 sartorius anterior case 2338L

L5 L6 sartorius medialis case 2338R

Figure 6 Plottings showing the location of the motoneurons innervating the iliopsoas and sartorius muscles (group I).

49 Chapter 2

Group I Group II Sartorius Quadriceps

rectus femoris

anterior sartorius

vastus lateralis

medial sartorius lateral aspect

rectus femoris medial aspect

vastus medialis Figure 7 Schematic drawing of the m. sartorius (group I).

medial aspect (superficial) Group I: mm. iliopsoas and sartorius Iliopsoas. The iliopsoas muscle is composed of three parts, the m. iliacus, and mm. psoas major and minor (Fig. 3; Table I). All three parts produce flexion of the spine (lordosis), when the hip is fixated (Sherrington, rectus femoris 1910; Crouch, 1969). The psoas major and iliacus also act as flexors of the hip when the spine is fixated (Sherrington, 1910; Crouch, 1969). vastus After injections into the m. iliopsoas near its insertion intermedius on the proximal femur, involving both the iliacus and psoas major, two groups of labeled motoneurons, a medial aspect (deep) medial and a lateral one, were found (Figs. 4, 5, and 6). The lateral group was located in the ventrolateral corner Figure 8 Schematic drawings of the quadriceps muscles of the ventral horn (Fig. 6, case 2253L; Fig. 5A). Their (group II). dendrites extended mainly dorsally and medially along the border of the gray and white matter. The rostral pole of this group, which is easily recognizable in unlabeled abdominal approach, which resulted in only one group sections, has been selected to mark the rostral end of of motoneurons in the ventrolateral part of the ventral the enlargement (level 0; see also Fig. 2A). Only horn (cases 2348L and 2348R in Fig. 6). The bulk of scattered labeled motoneurons were found further the psoas minor motoneurons was located more rostrally rostrally (until level minus 10). Caudally, the ventro- (level -25) than those of the psoas major (level -5; Figs. lateral group extended until level 40 (Fig. 4). 4 and 6). These results suggest that the iliacus muscle, The medial group was also located near the white matter, which was not injected separately, is innervated by but in the medial part of the ventral horn (Fig. 5B; see motoneurons of the medial group. also arrows in case 2253L in Fig. 6). Its motoneurons were evenly distributed between levels 10 and 60 (Fig. Sartorius. This muscle acts as a hip flexor, similar to 4), and have dendrites extending ventrally into the gray the m. iliopsoas. It can be functionally divided into a and dorsomedially into the white matter. medial and an anterior part, which in addition to hip Since the iliopsoas has its origin at the lower thoracic flexion, flex and extend the knee, respectively (Fig. 7; and upper lumbar vertebral column, a large part is Table I; Sherrington, 1910; Crouch, 1969; Loeb et al., located dorsally in the abdominal cavity and in the pelvis. 1987; Gordon et al., 1991; Pratt and Loeb, 1991). To determine the location of the motoneurons innerva- Injections involving both the medial and anterior ting the intraabdominal parts of the muscle, the psoas sartorius resulted in labeled neurons between levels 10 major and minor were respectively injected via to 50 (Fig. 6). They were most numerous rostrally (levels

50 lumbosacral motoneuronal cell groups

Table 2 Group II Quadriceps number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes (1951) Vastus lateralis O: dorsal and lateral shaft and Extension knee n. femoralis 2193 L 70 µl type II 1 4 7 3 5 8 trochanter major of femur 2209 R 90 µl type II 1 6 7 I: lateral patella 2223 L 90 µl type II 2 2 8 2327 R 20 µl type VI 1 0 5 2346 L 20 µl type VI 7 9

Vastus medialis O: shaft femur Extension knee n. femoralis 2194 R 70 µl type II 2 2 5 3 0 7 I: medial patella and 2334 L 20 µl type VI 4 0 0 ligamentum patellae 2346 R 20 µl type VI 1 9 1

Vastus intermedius O: shaft femur between vastus Extension knee n. femoralis 2235 R 30 µl type II 8 lateralis and vastus medialis Tensor capsule 2246 R 30 µl type II 6 5 I: capsule of knee joint knee joint

Rectus femoris O: lateral ilium Extension knee n. femoralis 2220 L 50 µl type II 1 5 3 3 1 4 I: lateral patella Flexion hip 2334 R 20 µl type VI 1 8 2 2356 L 25 µl type VI 1 6 5

10-30) where they formed a compact group in the Group II ventrolateral ventral horn (Figs. 5C and 4), dorsal to 00 102030405060 1020304050607700 80 80 90 90 100 100 those of the psoas major, which does not act on the knee, but only on the hip. Caudal to level 30, sartorius Vastus lateralis 20 motoneurons gradually diverged into a large medial and 10 case 2193 L 0 L5 L6L7 S1 a smaller lateral group. 30 After injections limited to the medial part of the sartorius, 20 10 case 2209 R 0 motoneurons were most numerous in the rostral part of L5 L6 L7 S1 the sartorius motoneuronal cell group, whereas the 30 20 anterior sartorius motoneurons were more evenly 10 case 2223 L 0 L5 L6 L7 S1 distributed (Fig. 4, cases 2264L and R, and 2338L and 20 R). Caudal to level 30, medial sartorius motoneurons 10 case 2327 R 0 L5 L6 L7 S1 were found both laterally and medially (case 2338R in 20 Fig. 6), whereas anterior sartorius motoneurons were 10 case 2346 L 0 L5 L6L7 S1 only found in the medial group (case 2338L in Fig. 6). Vastus medialis For a summary of the location of the motoneuronal cell 30 20 groups innervating the m. iliopsoas and sartorius see 10 case 2194 R 0 figure 37. L5 L6 L7 S1

30 20 10 case 2334 L Group II: m. quadriceps 0 L5 L6 L7 S1 The quadriceps consists of four separate muscles: the 30 20 vastus lateralis, medialis and intermedius, and the rectus 10 case 2346 R 0 femoris (Fig. 8; Table II; Sherrington, 1910; Crouch, L5 L6L7 S1 1969; Zajac, 1985). All four are extensors of the knee, Vastus intermedius 20 whereas the rectus femoris, a biarticular muscle, can 10 case 2246 R 0 L5 L6L7 S1 also act as a flexor of the hip.

20 10 case 2235 R 0 L5 L6 L7 S1 Vastus lateralis. Motoneurons of this muscle were Rectus femoris present in the lateral part of the ventral horn between 20 10 case 2220 L rostrocaudal levels 10 and 50, being most numerous 0 L5L6 L7 S1 between levels 25 and 40 (Fig. 9; Fig. 10 case 2346L).

30 20 10 case 2334 R 0 Vastus medialis. Its motoneurons were found medial L5 L6 L7 S1 to those of the vastus lateralis and were slightly more 20 10 case 2356 L predominant at rostral levels as compared to the vastus 0 L5 L6 L7 S1 lateralis motoneurons (Fig. 9; case 2346R in Fig. 10).

0 10203040506070 80 90 100 level Vastus intermedius. Injections in this relatively small Figure 9 Graphs showing the rostrocaudal distribution of muscle resulted in a limited number of labeled labeled motoneurons after injections into the quadriceps motoneurons between levels 40 and 60 (Table II; Figs. muscles (group II). 9 and 10), caudal to the majority of vastus lateralis and medialis motoneurons.

51 Chapter 2

Group II: Quadriceps 10 20 30 40 50 60

vastus lateralis L5 L6 case 2193L

L6 vastus medialis case 2346R

L6 vastus intermedius case 2246R

L5 L6 rectus femoris case 2356L

Figure 10 Plottings showing the location of the motoneurons innervating the quadriceps muscles (group II).

proximal adductor, formed a small group between levels Rectus femoris. The bulk of the rectus femoris 10 and 35 (Figs. 12 and 13; Table III). In transverse motoneurons was found between levels 30 and 50 (Fig. sections, this group was located centrally in the ventral 9), slightly caudal to the motoneurons of the sartorius, horn (Fig. 13), dorsomedial to sartorius and quadriceps the vastus lateralis, and the vastus medialis. Many rectus motoneurons. femoris motoneurons were located dorsomedially to the vastus medialis motoneurons, although there was an Adductor longus. Its motoneuronal cell group was extensive overlap between the two groups (Fig. 10). located lateral to the pectineus cell group (Fig. 12), and Figure 37 gives an overview of the location of the extended further caudally, between levels 15 and 50 (Fig. quadriceps motoneuronal cell group. 13).

Group III: Adductors Adductor brevis and magnus. Both muscles were The adductor group consists of the pectineus, adductor innervated by large groups of motoneurons, located longus, brevis, and magnus, and gracilis (Fig. 11; Table between levels 20 and 60 (Table III, Fig. 12). The two III; Sherrington, 1910; Crouch, 1969; Pratt et al., 1991). groups overlapped extensively, although the adductor All muscles adduct the leg and the gracilis in addition brevis group was located slightly more dorsally (Fig. flexes the knee. 13; see also Fig. 5D). Both motoneuronal cell groups occupied a more ventral position than those of the Pectineus. Motoneurons of the m. pectineus, the most pectineus and adductor longus, and extended much

Table 3 Group III Adductors number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes (1951) Pectineus O: cranial border os pubis Adduction thigh n. obturatorius 2233 L 40 µl type II 6 7 I: shaft femur distal to lesser If hip in 2253 R 40 µl type II 8 6 trochanter extension: 2314 R 20 µl type VI 6 3 flexion hip

Adductor longus O: cranial border os pubis, Adduction thigh n. obturatorius 2232 L 30 µl type II 7 8 medially to pectineus If hip in 2242 R 50 µl type II 2 2 3 I: linea aspera femur extension: 2316 R 25 µl type VI 9 6 flexion hip 2354 R 20 µl type VI 2 4 3

Adductor femoris O: pelvic symphyse and Adduction thigh n. obturatorius 2212 R 60 µl type II 3 4 4 brevis ramus ischium Some extension 2234 R 60 µl type II 1 4 8 I: linea aspera caudal femur hip

Adductor femoris O: pelvic symphyse and ramus Adduction thigh n. obturatorius 2206 R 90 µl type II 1 2 2 2 7 7 magnus ischium and tendon m. gracilis Extension hip 2227 L 50 µl type II 3 5 6 I: linea aspera caudal femur 2316 L 25 µl type VI 4 6 8

Gracilis O: pelvic symphyse Adduction thigh n. obturatorius 2200 R 90 µl type II 1 8 7 2 2 6 I: medial proximal tibia Flexion knee 2248 R 50 µl type II 1 1 0 Extension hip 2310 L 30 µl type VI 2 4 6 2314 L 35 µl type VI 2 6 4

52 lumbosacral motoneuronal cell groups

Group III Group III Adductors 0 10203040506070 80 90 100 Pectineus 20 10 case 2233 L 0 L5 L6 L7 S1

20 adductors 10 case 2253 R 0 L5 L6L7 S1 20 10 case 2314 R 0 L5 L6 L7 S1 Adductor longus gracilis 20 10 case 2232 L 0 L5 L6L7 S1

30 20 10 case 2242 R 0 L5 L6 L7 S1

20 10 case 2316 R 0 L5L6 L7 S1

30 20 medial aspect (superficial) 10 case 2354 R 0 L5L6 L7 Adductor brevis 30 20 10 case 2212 R 0 L5 L6 L7 S1

pectineus 20 10 case 2234 R adductor 0 L5 L7 S1 longus L6 Adductor magnus 20 adductor 10 case 2206 R brevis 0 adductor L5 L6 L7 S1 magnus 30 20 10 case 2227 L 0 L5 L6 L7 S1

40 30 20 10 case 2316 L 0 L5L6 L7 S1

Gracilis 20 case 2200 R medial aspect (deep) 10 0 L5L6 L7 S1

20 10 case 2248 R 0 Figure 11 Schematic drawings of the adductors of the thigh L5 L6 L7 S1 20 (group III). 10 case 2310 L 0 L5 L6 L7 S1

30 20 further caudally. 10 case 2314 L 0 L5 L6 L7 S1 Gracilis. The rostrocaudal level of the gracilis 0 10203040506070 80 90 100 motoneuronal cell group was almost the same as that of level the adductor brevis and magnus (Fig. 12), but gracilis Figure 12 Graphs showing the rostrocaudal distribution of motoneurons were located more dorsally (Fig. 13). labeled motoneurons after injections into the adductor muscles Figure 37 gives an overview of the location of the (group III). motoneuronal cell groups of the muscles of group III.

Group IV: Hamstrings et al., 1987; Hutchison et al., 1989; Chanaud and This group consists of the semimembranosus, semi- Macpherson, 1991; Chanaud et al., 1991; Pratt et al., tendinosus, and biceps femoris muscles (Fig. 14). 1991). Hamstring muscles are extensors of the hip and/or strong flexors of the knee, and can be subdivided functionally Semimembranosus. The majority of the motoneurons and/or anatomically (Table IV; Crouch, 1969; Sherring- of this muscle was located between levels 50 and 70 ton, 1910; Peters and Rick, 1977; Letbetter and English, (Fig. 15). They were located centrally in the ventral horn 1981; English and Letbetter, 1981; Botterman et al., (Fig. 16 and 17A), forming the caudal continuation of 1983a,b; Zajac, 1985; English and Weeks, 1987; Loeb the gracilis motoneuronal cell group. Motoneurons

53 Chapter 2

Group III: Adductors 10 20 30 40 50 60

L5 pectineus case 2314R

L5 L6 adductor longus L5 case 2354R

L6L6 L7 adductor brevis case 2212R

L6 adductor magnus case 2316L

L6 gracilis case 2314L

Figure 13 Plottings showing the location of the motoneurons innervating the adductors of the thigh (group III).

innervating the posterior part of the semimembranosus the semitendinosus muscle is anatomically divided into were located slightly more caudally than those of the a proximal and a distal part, theoretically one could anterior part (Figs. 15 and 16). Occasionally, labeled expect differences between the motoneuronal cell groups motoneurons were found at more rostral levels (level innervating these two parts. However, injections 20-30; Fig. 15). involving only one part of the muscle did not result in a difference in distribution of motoneurons within the cell Semitendinosus. This motoneuronal cell group was group. present mainly between levels 65 and 90 (Fig. 15), and overlapped with the caudal extent of the semimem- Biceps femoris anterior and posterior. Biceps femoris branosus group (Fig. 16). In all cases, a few labeled motoneurons were present between levels 75 and 100 motoneurons were found as rostral as level 40. Since (Fig. 15). Posterior biceps femoris motoneurons were

Table 4 Group IV Hamstrings number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Boyd and Davey Other (1951) (1968;α/γ) Semimembranosus O: tuberosity and ramus ischium Extension hip n. tibialis 2200 L 90 µl type II 1 4 3 4 4 8 I: medial epicondyle and distal femur Flexion knee 2206 L 90 µl type II 2 5 8 Some inward rotation 2222 L 60 µl type II 2 5 3 of thigh 2227 R 60 µl type II 4 3 3 2291 R 25 µl type VI 4 1 6

- anterior Extension hip 2291 L 25 µl type VI 2 0 2 2301 R 25 µl type VI 2 1 6

- posterior Extension hip 2292 L 50 µl type VI 1 2 6 Flexion knee 2301 L 25 µl type VI 1 4 8 2312 L 25 µl type VI 7 1

Semitendinosus O: apex of tuberosity ischium Extension hip n. tibialis 2209 L 90 µl type II 3 0 1 2 0 8 255/255 Hoover and Durkovic I: medial shaft tibia Flexion knee 2222 R 60 µl type II 3 5 2 (1991): 240-510 2312 R 25 µl type VI 2 2 0 2313 R 25 µl type VI 2 3 7

Biceps femoris O: ventral part of tuberosity ischium cran.: n. glut. inf. 2194 L 70 µl type II 2 3 6 4 5 6 med.: n. tibialis 2207 R 90 µl type II 2 2 4 caud.: n. tibialis 2313 L 25 µl type VI 2 5 7

- anterior I: lateral proximal tibia and Extension hip 2255 R 50 µl type II 1 1 7 lateral patella Abduction thigh Some outward rotation thigh Aid in extension knee

- posterior I: lateral proximal tibia and lateral Flexion knee 2255 L 50 µl type II 1 3 7 patella, distal to anterior part of Aid in hip extension and biceps femoris femoral rotation

54 lumbosacral motoneuronal cell groups

Group IV Group IV Hamstrings 0 10203040506070 80 90 100

Semimembranosus 30 20 case 2200 L 10 0 L5 L6 L7 S1

40 30 20 biceps femoris 10 case 2206 L anterior 0 L5 L6 L7 S1

40 30 20 10 case 2222 L 0 L5 L6 L7 S1

40 biceps femoris 30 20 posterior 10 case 2291 R 0 L5 L6 L7 S1 -anterior- 30 20 10 case 2291 L 0 L5 L6 L7 S1

lateral aspect 40 30 20 10 case 2301 R 0 L5 L6 L7 S1 -posterior- 30 20 10 case 2292 L 0 L5L6 L7 S1

30 20 10 case 2301 L semimembranosus 0 anterior L5 L6 L7 S1 20 10 case 2312 L 0 L5 L6 L7 S1 semimembranosus Semitendinosus posterior 40 30 20 semitendinosus 10 case 2209 L (Prox.) 0 L5 L6 L7 S1

40 30 20 10 case 2222 R 0 L5 L6 L7 S1 medial aspect 30 20 10 case 2312 R (Dist.) 0 Figure 14 Schematic drawings of the hamstring muscles L5 L6 L7 S1

(group IV). 30 20 10 case 2313 R 0 L5L6 L7 S1 found caudal to level 85 (Fig. 15), where they formed Biceps femoris the caudal continuation of the semitendinosus group 30 20 (case 2255L in Fig. 16). In contrast, anterior biceps 10 case 2194 L 0 femoris motoneurons were located more rostrally than L5L6 L7 S1 40 the posterior biceps motoneurons (Fig. 15), and lateral 30 20 to the semitendinosus motoneurons (case 2255R in Fig. 10 case 2207 R 0 16). For a summary diagram of the muscles of group IV L5 L6 L7 S1 40 see figure 37. 30 20 10 case 2313 L 0 Group V: Gluteal and other proximal muscles L5 L6 L7 S1 -anterior- This group consists of approximately 10 muscles of 30 20 10 case 2255 R which the motoneurons of the most important ones, 0 gluteus medius and maximus, caudofemoralis, tensor L5 L6L7 S1 -posterior- 30 fasciae latae, tenuissimus and obturatorius internus have 20 10 case 2255 L 0 L5 L6L7 S1

Figure 15 Graphs showing the rostrocaudal distribution of 0 10203040506070 80 90 100 labeled motoneurons after injections into the hamstring level muscles (group IV).

55 Chapter 2

Group IV: Hamstrings 50 60 70 80 90 100

L7 Semimembranosus case 2222L

Semimembranosus L7 anterior case 2301R

L7 Semimembranosus posterior case 2301L

L7 L7 S1 Semitendinosus case 2222R

L7 Biceps Femoris S1 case 2207R

Biceps Femoris S1 anterior case 2255R

Biceps Femoris S1 posterior case 2255L

Figure 16 Plottings showing the location of the motoneurons innervating the hamstring muscles (group IV).

been traced (Table V; Fig. 18). Except for the tensor number of motoneurons (Table V). Rostrocaudally, the fasciae latae, these muscles produce extension of the gluteus medius motoneuronal cell group overlapped hip and/or abduction of the thigh. The tensor fasciae extensively with the tensor fasciae latae group. However, latae is a complex muscle which is divided into an the majority of the gluteus medius motoneurons were anterior and a posterior part (Fig. 18). The anterior part located more caudally between levels 65 to 90 (Fig. 19), flexes the hip and extends the knee similar to the anterior where they formed a compact group in the most part of the sartorius. The posterior part lies over the ventrolateral part of the ventral horn at the border with gluteus medius and acts as an abductor of the thigh the white matter (Figs. 17C and 20). (Sherrington, 1910; Crouch, 1969; Loeb et al., 1987; Gordon et al., 1991; Pratt et al., 1991). Gluteus maximus. Although the name suggests other- wise, in the cat the gluteus maximus is smaller in size Tensor fasciae latae. After injections into both parts of than the gluteus medius, and is also innervated by a this muscle, labeled motoneurons were most numerous smaller number of motoneurons (Table V). Gluteus between levels 50 and 65 (Fig. 19), forming a compact maximus motoneurons are most numerous between cell group in the ventrolateral corner of the ventral horn levels 65 and 90 (Fig. 19) and are located near the ventral near the border with the white matter (Fig. 20). Further caudally (level 65 to 85), they are less numerous and border of the ventral horn, medial to the gluteus medius slightly more dispersed. In case 2327L, in which a small motoneuronal cell group (Figs. 17D and 20). injection was made into only the posterior part, labeled motoneurons were found in the same cell group, but Caudofemoralis. Caudofemoralis motoneurons were only between levels 60 to 90. most numerous between levels 85 and 100 (Fig. 19), and were located in the same area as the gluteus maximus Gluteus medius. This muscle is innervated by a large cell group (Fig. 20). Between levels 70 and 80, some

56 lumbosacral motoneuronal cell groups

A B

C D

Figure 17 Polarized light- darkfield photomicrograph of labeled motoneurons after injections into different hindlimb muscles. A: motoneurons of the semimembranosus forming a compact group in the center of the ventral horn at level 60 (caudal L6 in case 2291R). B: motoneurons of the gastrocnemius lateralis at level 80 (L7 in case 2325L). Note that the labeled motoneurons form two groups. C: gluteus medius motoneurons at level 80 (L7 in case 2229L). Note that they are located immediately adjacent to the white matter in the ventrolateral corner of the ventral horn. D: motoneurons of the gluteus maximus at level 80 (L7 in case 2254R). Note that this group is located medial to the gluteus medius group (C). Bar represents 400 µm. were located more dorsally (Fig. 20). dorsomedial to the gluteus motoneuronal cell groups Obturatorius internus and abductor cruris caudalis. (Fig. 20). The few abductor cruris caudalis motoneurons Injections into the obturatorius internus resulted in a were located in the same group as the obturatorius limited number of labeled motoneurons (Table V) which internus motoneurons. were located mainly at levels 65 to 85 (Fig. 19), just Figure 37 gives an overview of the location of the motoneuronal cell groups of the muscles of group V.

Table 5 Group V Gluteal and other proximal muscles of the hip number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Boyd and Davey Other (1951) (1986; α /γ) Tensor fasciae latae n. gluteus sup. 2196 R 90 µl type II 1 4 5 Gordon et al. (1991) - anterior O: ilium, similar to sartorius anterior Flexion hip 2229 R 60 µl type II 2 4 5 22-128 I: similar to vastus medialis Extension knee anterior: 81 - posterior O: ventral ilium and fascia mm. glutei Abduction thigh 2327 L 15 µl type VI 8 6 posterior: 65 I: fascia lata Some flexion hip Tension fascia lata

Gluteus medius O: dorsolateral ilium, gluteal fascia Abduction thigh n. gluteus sup. 2196 L 90 µl type II 2 7 4 and processi transversi S-Co Extension hip 2210 R 90 µl type II 4 1 6 I: proximal trochanter major 2229 L 70 µl type II 5 1 2 2257 R 40 µl type II 1 7 3

Gluteus maximus O: gluteal fascia and processi Abduction thigh n. gluteus inf. 2215 R 80 µl type II 1 6 0 transversi S-Co Extension hip 2230 R 40 µl type II 1 4 2 I: trochanter major 2244 R 50 µl type II 2 5 3 2254 R 40 µl type II 2 1 9

Caudofemoralis O: processi transversi Co 2-3 Abduction thigh n. gluteus inf. 2198 R 90 µl type II 1 2 6 I: lateral patella Aids in extension knee 2230 L 50 µl type II 2 5 9 2244 L 50 µl type II 1 0 1

Abductor cruris O: processus transversus Abduction thigh n. ischiadicus 2252 R 20 µl type II 1 8 20/20 caudalis I: with biceps femoris Extension hip Flexion knee

Obturatorius internus O: dorsal ramus ischium Abduction thigh n. ischiadicus 2340 L 10 µl type VI 4 4 8 2 I: trochanter (fossa) Outward rotation hip 2354 R 15 µl type VI 1 5 5 2356 R 20 µl type VI 7 2

57 Chapter 2

Group V Group V Gluteal and other proximal muscles of the hip 0 10203040506070 80 90 100 Tensor fasciae latae 20 gluteus 10 case 2196 R 0 medius L5 L6L7 S1

30 20 10 case 2229 R tensor fasciae 0 latae posterior L5 L6 L7 S1 - posterior part- tensor fasciae 20 obturator 10 case 2327 L latae anterior 0 internus L5 L6 L7 S1 abductor fascia lata Gluteus medius cruris 40 caudalis 30 20 lateral aspect 10 case 2196 L 0 L5 L6L7 S1

40 30 20 gluteus 10 case 2210 R maximus 0 L5 L6 L7 S1

caudofemoralis 60 gluteus 50 40 medius 30 20 10 case 2229 L 0 L5 L6 L7 S1

30 20 10 case 2259 R 0 L5 L6 L7 S1 lateral aspect Gluteus maximus 20 10 case 2215 R Figure 18 Schematic drawings of the gluteal and other 0 proximal muscles of the hip (group V). L5L6 L7 S1 20 10 case 2230 R 0 L5L6 L7 S1

30 20 Group VI: Posterior muscles of the distal hindlimb 10 case 2244 R 0 This group contains the triceps surae, consisting of the L5L6 L7 S1 30 mm. gastrocnemius lateralis and medialis, and soleus, 20 10 case 2254 R 0 as well as the plantaris, and tibialis posterior (Fig. 21; L5 L6 L7 S1 Table VI). The triceps surae and mm. plantaris and Caudofemoralis tibialis posterior act as plantar flexors of the foot. The 20 10 case 2198 R 0 two gastrocnemius muscles can also act as flexors of L5 L6 L7 S1

the knee (Sherrington, 1910; Crouch, 1969; Burke et 40 30 al., 1977; English and Letbetter, 1982; Weeks and 20 10 case 2230 L 0 English, 1985). L5 L6L7 S1

20 10 case 2244 L 0 Gastrocnemius medialis and lateralis. The medial L5 L6 L7 S1 gastrocnemius motoneurons were located in a rather Abductor cruris caudalis compact cell group in the dorsolateral part of the ventral 20 10 case 2252 L horn at levels 80 to 95 (Figs. 22 and 23). Motoneurons 0 L5 L6 L7 S1 innervating the lateral head of the gastrocnemius were Obturatorius internus

located medial to the motoneurons innervating the 20 10 case 2340 L medial head of this muscle, and they were not present 0 L5 L6 L7 S1

as far caudally. 20 10 case 2354 R 0 Soleus. The majority of the labeled soleus motoneurons L5 L6 L7 S1 20 were found between levels 70 and 95 (Fig. 22). Most of 10 case 2356 R 0 these motoneurons were located in the same cell group L5 L6 L7

as the medial gastrocnemius motoneurons (Fig. 23), but 0 10203040506070 80 90 100 the soleus cell group extended slightly more rostrally level and less caudally. Figure 19 Graphs showing the rostrocaudal distribution of labeled motoneurons after injections into the gluteal and other Plantaris. Plantaris motoneurons were found between proximal muscles of the hip (group V).

58 lumbosacral motoneuronal cell groups

Group V: Gluteal and other proximal muscles of the hip 50 6070 80 90 100

Tensor fasciae L7 S1 latae case 2229R

L7 S1 Gluteus medius case 2229L

L7 S1 Gluteus maximus case 2230R

L7 S1 Caudofemoralis case 2230L

Obturatorius L6 L7 L7 internus case 2354R

Figure 20 Plottings showing the location of the motoneurons innervating the gluteal and other proximal muscles of the hip (group V). levels 60 and 90, but were most numerous between levels given in figure 37. 70 and 80 (Fig. 22). The plottings show that plantaris Group VII: Anterior muscles of the distal motoneurons occupied the same area in the ventral horn hindlimb as those innervating the lateral gastrocnemius, but the The group of anterior muscles of the distal hindlimb plantaris cell group extended further rostrally and consists of the tibialis anterior, peroneus longus, and caudally. The motoneuronal cell groups of both the the extensor digitorum longus (Fig. 24; Table VII). The lateral gastrocnemius and plantaris always consisted of tibialis anterior and peroneus longus are dorsal flexors two subgroups, a dorsal and a ventral one (Figs. 17B of the foot and the extensor digitorum longus extends and 23). the toes (Sherrington, 1910; Crouch 1969; Iliya and Dum, 1984). Tibialis posterior. The relatively small tibialis posterior motoneuronal cell group (Table VI) was located in the same area as the soleus motoneurons (Figs. 22 and 23), Tibialis anterior. After injections into the tibialis but extended more rostrally. An overview of the anterior, labeled motoneurons were present between motoneuronal cell groups of the muscles of group VI is levels 50 and 85, being most numerous between 60 and

Table 6 Group VI Posterior compartment of the distal hindlimb number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Boyd and Davey Other (1951) (1968; α /γ) Gastrocnemius n. tibialis Weeks and English - lateralis O: epicondylus lateralis Plantar flexion foot 2239 L 40 µl type II 7 3 2 9 0 (1985): 238±37 I: calcaneus Flexion knee 2325 L 20 µl type VI 1 3 0

- medialis O: epicondylus medialis Plantar flexion foot 2191 R 40 µl type II 8 2 2 4 4 280/170 Burke et al. (1977) I: calcaneus Flexion knee 2239 R 40 µl type II 1 1 0 254α/83γ and 2317 R 30 µl type VI 1 5 1 2 6 6 α /94γ Weeks and English (1987): mean: 435

Soleus O: lateral head and ventral fibula Plantar flexion foot n. tibialis 2211 L 60 µl type II 9 7 9 9 155/115 Burke et al. (1977) I: laterally onto tendon gastrocnemius 2257 L 30 µl type II 2 1 5 133α/48γ and 2346 R 15 µl type VI 7 0 1 3 2 α /54γ

Plantaris O: lateral patella and lateral sesamoid; Plantar flexion foot n. tibialis 2217 L 80 µl type II 1 8 5 2 0 4 265/175 lateral gastrocnemius 2220 R 40 µl type II 2 4 2 I: tendon flexor digitorum brevis 2346 L 15 µl type VI 3 8 and calcaneus 2354 R 15 µl type VI 1 5 4

Tibialis posterior O: medial head fibula and caudal tibia Inversion ankle n. tibialis 2259 L 20 µl type II 2 1 6 6 60/75 I: lateral os naviculare and proximal Plantar flexion foot 2332 L 10 µl type VI 6 1 ventral os cuneiforme

59 Chapter 2

Group VI Group VI Posterior compartment of the distal hindlimb 0 10203040506070 80 90 100

Gastrocnemius medialis 20 10 case 2191 R 0 L5 L6 L7 S1 30 20 10 case 2239 R 0 L5 L6L7 S1

30 20 10 case 2317 R 0 L5 L6 L7 S1 Gastrocnemius lateralis 20 gastrocnemius 10 case 2239 L plantaris 0 lateralis L5 L6L7 S1 30 20 10 case 2325 L 0 L5 L6 L7 S1 Soleus 20 soleus soleus 10 case 2211 L 0 L5 L6L7 S1

30 20 10 case 2257 L 0 L5 L6L7 S1

20 10 case 2346 R 0 L5 L6L7 S1

lateral aspect (superficial) lateral aspect (deep) Plantaris 30 20 10 case 2217 L 0 L5 L6 L7 S1

30 20 10 case 2220 R 0 L5L6 L7 S1

40 30 20 10 case 2240 L 0 L5 L6 L7 S1

20 10 case 2346 L 0 gastrocnemius L5 L6L7 S1

medialis 20 10 case 2354 L 0 L5 L6 L7

soleus Tibialis posterior 20 tibialis 10 case 2259 L plantaris 0 posterior L5L6 L7 S1

tendon of 20 10 case 2332 L tibialis posterior 0 L5 L6 L7 S1

0 10203040506070 80 90 100 level Figure 22 Graphs showing the rostrocaudal distribution of labeled motoneurons after injections into the muscles of the posterior compartment of the distal hindlimb (group VI). medial aspect (superficial) medial aspect (deep)

Figure 21 Schematic drawings of the muscles of the posterior compartment of the distal hindlimb (group VI). ventral horn (Fig. 26), extending clearly further caudally than the motoneuronal cell groups of the tibialis anterior and extensor digitorum longus. Rostrally, they partly 75 (Fig. 25). They formed a small group dorsolaterally overlapped with the tibialis anterior motoneuronal cell in the ventral horn (Fig. 26), ventral to the motoneuronal group, but caudally they were located ventral to the cell group of long flexors of the foot, and dorsal to the tibialis anterior motoneurons. soleus motoneurons. Extensor digitorum longus. The motoneurons of this Peroneus longus. Its motoneurons were found between muscle were present between levels 50 and 80 (Fig. 25), the levels 50 and 90 (Fig. 25) in the lateral part of the in the same area as the tibialis anterior motoneurons,

60 lumbosacral motoneuronal cell groups

Group VI: Posterior compartment of the distal hindlimb

6070 80 90 100

S1 Gastrocnemius medialis case 2317R

Gastrocnemius lateralis case 2325L

L7 Soleus case 2211L

L7 S1 S1 Plantaris case 2217L

L7 Tibialis posterior case 2332L

Figure 23 Plottings showing the location of the motoneurons innervating the muscles of the posterior compartment of the distal hindlimb (group VI). but with a more rostral predominance (cases 2333 and formed a compact group in the dorsolateral corner of 2336 of Fig. 26). For an overview of the motoneuronal the ventral horn (Fig. 29), dorsal to the motoneurons cell groups innervating the muscles of group VII, see innervating other muscles of the distal hindlimb (groups figure 37. VI and VII). Flexor hallucis longus motoneurons were much less numerous and were located between levels Group VIII: Long flexors and intrinsic muscles 65 and 80 in the same cell group as the flexor digitorum of the foot longus (Figs. 28 and 29). This group consists of the flexor digitorum longus, flexor hallucis longus and intrinsic muscles of the foot, such Tibial nerve. Because the very small intrinsic foot as the adductors digiti, the interossei, and the flexor muscles do not have individual fasciae, the tracer digitorum brevis (Fig. 27). The first two muscles are injections into these muscles were never limited to only plantar flexors of the foot, similar to the group VI one muscle, but involved adjacent intrinsic foot muscles muscles, but they also act as flexors of the toes. The as well. In order to get an overview of all motoneurons intrinsic foot muscles are flexors, adductors or abductors innervating the intrinsic foot muscles, in one case of the toes (Sherrington, 1910; Crouch, 1969; Dum et (2270L), the tibial nerve, which supplies all intrinsic al., 1982). foot muscles, was cut just distal to the malleolus and HRP was applied to the proximal end of the nerve. Flexor digitorum longus and flexor hallucis longus. The Extremely large numbers of labeled motoneurons (Table flexor digitorum longus motoneurons were most VIII) were found between levels 65 and 120, being most numerous between levels 60 and 75 (Fig. 28). They numerous from levels 90 to 110 (Fig. 28). They formed

Table 7 Group VII Anterior compartment of the distal hindlimb number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Boyd and Davey Other (1951) (1968; α /γ) Tibialis anterior O: proximal lateral tibia and medial Ankle dorsiflexion n. peroneus 2241 L 40 µl type II 1 3 6 1 7 5 200/140 Iliya and Dum (1984) fibula, and interosseus membrane Foot inversion 2333 L 10 µl type VI 7 4 Hoover and Durkovic, I: os metatarsale I Dorsal flexor toes 2336 R 15 µl type VI 6 3 (1991): 125-400

Peroneus longus O: lateral head and shaft fibula Foot eversion n. peroneus 2210 L 90 µl type II 2 1 4 1 2 6 95/80 Horcholle-Bossavit et al., I: metatarsalia I-V Ankle dorsiflexion 2330 R 15 µl type VI 1 6 8 (1988): 129-200 2354 L 20 µl type VI 2 3 2 Kernell et al. (1985): 71-84 Extensor digitorum O: lateral epicondyle femur Ankle dorsiflexion n. peroneus 2211 R 60 µl type II 1 2 3 1 3 5 130/90 Hoover and Durkovic, longus I: phalanges II-V Dorsal flexor toes 2333 R 10 µl type VI 5 7 (1991): 90-295 Aids in extension knee 2336 L 15 µl type VI 1 6 8

61 Chapter 2

Group VII Group VII Anterior compartment of the distal hindlimb 0 10203040506070 80 90 100

Tibialis anterior 20 10 case 2198 L 0 L5L6 L7 S1 30 20 10 case 2241 L 0 L5 L6 L7 S1 20 10 case 2333 L 0 L5L6 L7 S1 20 10 case 2336 R 0 extensor L5 L6 L7 S1 digitorum Peroneus longus longus 20 10 case 2210 L peroneus 0 L5 L6L7 S1 longus 20 tibialis anterior 10 case 2330 R 0 L5 L6L7 S1 tendon 30 peroneus 20 tendon extensor 10 case 2354 L longus 0 digitorum L5 L6 L7 S1 longus Extensor digitorum longus 20 10 case 2211 R 0 L5 L6 L7 S1 lateral aspect 20 10 case 2333 R 0 L5 L6 L7 S1 Figure 24 Schematic drawings of the muscles of the anterior 20 10 case 2336 L compartment of the distal hindlimb (group VII). 0 L5 L6 L7 S1

0 10203040506070 80 90 100 level a large compact cell group in the dorsal and dorsolateral Figure 25 Graphs showing the rostrocaudal distribution of ventral horn (Fig. 29), although a few labeled motoneu- labeled motoneurons after injections into the muscles of the anterior compartment of the distal hindlimb (group VII). rons were present more ventrally, near Onuf’s nucleus.

Flexor digitorum brevis. Motoneurons innervating these muscles were mainly present between levels 90 and 100, present as far rostral as level 70 (Fig. 28). These rostral in the dorsal part of the intrinsic foot muscle moto- motoneurons were located among the flexor digitorum neuronal cell group. and hallucis longus motoneurons (Fig. 29). The adductor digiti motoneurons were slightly more numerous in the Interossei and adductor digiti. Injections into these ventral, and the interossei motoneurons in the dorsal distal muscles, resulted in very large numbers of labeled part of the intrinsic foot motoneuronal cell group (Fig. motoneurons between levels 90-110, but a few were 29). Figure 37 gives an overview of the location of the motoneuronal cell groups of the group VIII muscles.

Group VII: Anterior compartment of the distal hindlimb

6070 80 90 L7 Tibialis Anterior case 2333L

L7 Peroneus Longus case 2330R Figure 26 Plottings showing the location of the motoneurons L7 Extensor Digitorum innervating the muscles of the Longus anterior compartment of the distal hindlimb (group VII). case 2336R

62 lumbosacral motoneuronal cell groups

Table 8 Group VIII Long flexors and intrinsic muscles of the foot number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Romanes Boyd and Davey Other (1951) (1968; α /γ) Flexor digitorum O: head fibula and proximal tibia Flexion digits n. tibialis 2217 R 40 µl type II 2 0 6 1 2 2 lateral: 255/165 Dum et al. (1982) longus I: common tendon with flexor Some plantar flexion 2224 L 50 µl type II 1 7 9 medial: 155/145 1 2 5 α /75γ hallucis longus to distal phalanges foot 2241 R 40 µl type II 9 7 1 0 4 α /59γ of digits 2-5 2262 L 25 µl type II 1 4 7

Flexor hallucis O: head and shaft fibula, tibia and Flexion digits n. tibialis 2260 L 35 µl type II 3 7 1 4 5 longus ligamentum interosseum Some plantar flexion 2266 R 30 µl type II 3 2 I: common tendon with flexor foot digitorum longus to distal phalanges of digits 2-5

Flexor digitorum O: os naviculare and cuneiforme Flexion digits n. plantaris med. 2250 L 50 µl type II 1 2 6 65/35 brevis superficialis I: bases phalanges 2254 L 40 µl type II 1 6 4

Interossei O: proximal phalanges Abduction or adduction n. tibialis 2263 R 30 µl type II 4 1 6 I: distal phalanges digits

Adductor digiti O: ligament of peroneal canal Adduction digits n. tibialis 2263 L 30 µl type II 5 3 7 I: lateral proximal phalanges n. tibialis 2270 L 20 µl type II 7 4 6 740;683

Group VIII ventral part of the ventral horn (Fig. 31, 32, and 33C). Long flexors and intrinsic muscles of the foot Abductor caudae internus motoneurons were found in the same cell group as the levator ani motoneurons.

External anal sphincter. External anal sphincter motoneurons are relatively small and take part in the so-called nucleus of Onuf. They were found at levels 95 to 110 (Fig. 31). Rostrally, external anal sphincter motoneurons were found centrally in the ventral horn, just dorsomedial to the biceps femoris motoneuronal cell group (Figs. 32, and 33A and B), caudally they became located near the border of the white matter. flexor hallucis longus External urethral sphincter and ischiocavernosus. The flexor digitorum longus motoneurons of these two muscles also take part in the nucleus of Onuf and are completely intermingled (Fig. tendon 32). The rostral extent of their combined motoneuronal tibialis posterior flexor digitorum adductor digiti cell group was located caudal and ventrolateral to the brevis rostral pole of the external anal sphincter cell group. More caudally, Onuf’s nucleus, with the disappearance interossei common tendon of of the intrinsic foot cell group, became located in the flexor digitorum longus and flexor dorsolateral part of the ventral horn at level 110, where hallucis longus it continued until level 120 (Fig. 31). medial aspect Figure 27 Schematic drawings of the long flexors and intrinsic Bladder. The reason for studying the location of the muscles of the foot (group VIII). bladder preganglionics was to compare it with the location of the pelvic floor motoneurons, with which they have a close functional relationship (Holstege et Group IX: Pelvic floor muscles and bladder al., 1986). Bladder parasympathetic preganglionic motoneurons are small sized and are located in the The pelvic floor contains a group of relatively large, dorsolateral ventral horn (Fig. 32) mainly caudal to the flat muscles such as the levator ani and abductor caudae ischiocavernosus and external urethral sphincter internus (Fig. 30; Crouch, 1969; Dubrovsky and Filipini, motoneurons (levels 110-125; Fig. 31). A few rostrally 1990), as well as the smaller external urethral and anal located bladder preganglionics were located within the sphincters, and the ischiocavernosus muscles (Fig. 30; confines of Onuf’s nucleus. In conclusion, from rostral Crouch, 1969). In order to get a more complete overview, to caudal, Onuf’s nucleus motoneurons form a the detrusor muscle of the bladder, consisting of smooth continuous column with the bladder preganglionics (Fig. muscle fibers innervated by parasympathetic moto- 37). neurons, has also been included in this group. Group X: Extensors of the lower back and Levator ani and abductor caudae internus. Levator proximal tail ani motoneurons were found at levels 110-120 in the This group contains the medial longissimus, the caudal

63 Chapter 2

Group VIII

0 10203040506070 80 90 100 110 120

Flexor digitorum longus 20 10 case 2217 R 0 L5 L6 L7 S1

30 20 10 case 2224 L 0 L5 L6 L7 S1

20 10 case 2241 R 0 L5 L6 L7 S1

20 10 case 2262 L 0 L5 L6L7 S1

Flexor hallucis longus 20 10 case 2260 L 0 L5 L6 L7 S1

20 10 case 2266 R 0 L5 L6 L7 S1

n. tibialis 90 80 70 60 50 40 30 20 10 case 2270 L 0 L5 L6 L7 S1 S2 Flexor digitorum brevis 20 10 case 2250 L 0 L5 L6 L7 S1 30 20 10 case 2254 L 0 L5 L6 L7 S1 ON Interossei 50 40 30 20 10 case 2263 R 0 L5 L6L7 S1 S2 ON Adductor digiti 60 50 40 30 20 Figure 28 Graphs showing the rostrocaudal 10 case 2263 L 0 distribution of labeled motoneurons after injections L5 L6 L7 S1 S2 ON into the long flexors, intrinsic muscles of the foot,

0 10203040506070 80 90 100 110 120 and distal tibial nerve (group VIII).

level

extent of which is the extensor caudae lateralis or part of this motoneuronal cell group (levels 85 to at least sacrocaudalis dorsalis lateralis. The other muscles of 115; Fig. 36). Labeled motoneurons were found in the this group are the lumbar multifidi, which continue ventromedial corner of the ventral horn bordering the caudally as the extensor caudae medialis or sacrocaudalis white matter (Fig. 35), and ventral to levator ani moto- dorsalis medialis (Crouch, 1969; Ritz et al., 1992; Fig. neurons (Fig. 32). 34). When bilaterally activited, these muscles extend the vertebral column and, in respect to the caudal Lumbar multifidi. Motoneurons innervating these muscles, the tail. Unilateral activation causes latero- muscles were present in the medial part of the ventral flexion of the vertebral column or tail. horn at the border of the gray and white matter (Fig. 35). Their dendrites formed bundles which extended Longissimus medialis and extensor caudae lateralis. dorso-medially into the white matter (Fig. 33D). The injections were limited to the caudal portion of this Remarkably, multifidi motoneurons did not form a muscle group, at the level of the sacrum. This led to a continuous rostrocaudal column, but were located in continuous column of labeled motoneurons in the caudal small groups at the rostral end of each of the L6-S2

64 lumbosacral motoneuronal cell groups

Group VIII: Long flexors and intrinsic muscles of the foot

6070 80 90 100 110 Flexor digitorum L7 S1 longus case 2217R

L7 Flexor hallucis longus case 2260L

n. tibialis S1 S2 case 2270L

S1 Flexor digitorum brevis case 2254L

S1 S2 Interossei case 2263R

S1 S2 Adductor digiti case 2263L

Figure 29 Plottings showing the location of the motoneurons innervating the long flexors and intrinsic muscles of the foot (group VIII). segments (Fig. 36). DISCUSSION

Technical considerations branosus, caudofemoralis, abductor caudae internus, and This is the first study in which the location of the multifidi have never been traced before. The technique motoneuronal cell groups of 50 different muscles or of intramuscular injection of HRP was used, because muscle compartments of the lower part of the body are HRP diffuses well throughout the muscle, is taken up determined and presented in an overall scheme of the and transported easily, and labels a very large portion lumbosacral ventral horn. Of these muscles, the of the entire motoneuronal cell group innervating the motoneuronal cell groups of the iliopsoas, pectineus, respective muscle (Burke et al., 1977). This is a major adductor longus, anterior and posterior semimem- advantage over tracers as CTb and fluorescent

Table 9 Group IX Pelvic floor and bladder number of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present Other Remarks

Abductor caudae O: spine ischium Unilaterally: abd. tail Dorsal and ventral 2325 R 7 µl type VI 3 6 male internus I: processus transversus Co 2-5 Bilaterally: flex. tail rami of Co nerves

Levator ani O: symphysis of pelvis Compression of Ventral rami 2252 L 20 µl type II 1 0 3 female I: ventral midline of Co 3-5 vertebrae rectum; S3 and Co1 2277 R 20 µl type II 9 female Bends tail 2311 L 10 µl type VI 3 6 female 2340 R 5 µl type VI 1 3 female

External anal Circular muscle at the distal Close anus n. pudendus 2252 R 20 µl type II 1 6 4 Beattie et al. (1990) female sphincter end of the anus 2311 L 10 µl type VI 9 5 10-111 female 2354 L 15 µl type VI 6 2 Thor et al., (1989) female 163/88

External urethral Circular muscle at the distal end of Close urethra n. pudendus 2356 L 15 µl type VI 1 3 9 Beattie et al. (1990) female sphincter the urethra 16-105 Thor et al., (1989) 137/87

Ischiocavernosus O: ramus ischium Maintain erection n. pudendus 2272 R 10 µl type II 2 1 4 female I: ventral surface urogenital sinus rectal branches or outer surface crus penis

Detrusor bladder Empty bladder n. pelvicus 2343 - - - 2345 - - -

65 Chapter 2

Group IX One of the major risks of intramuscular HRP injections is leakage into surrounding muscles. During surgery, Pelvic floor extreme care was taken to minimize damage to structures adjacent to the muscle to be injected. This abductor caudae internus levator ani anus was especially important for the fasciae of the different muscles (Haase and Hrycyshyn, 1986). The fascia of a muscle group, such as the anterior compartment of the external anal sphincter hindlimb (group VII), to which the target muscle, such as the tibialis anterior, belonged had to be carefully vagina opened for exposure and identification. The fascia of a external urethral target muscle was penetrated only once with the sphincter bladder urethra injection needle, after which the needle was moved into ischiocavernosus different directions to get most effective labeling. It was

lateral aspect not necessary to damage surrounding musculature except in one case. The skin muscle cutaneus trunci or Figure 30 Schematic drawings of the muscles of the pelvic maximus, which overlies the superficial abdominal, floor and bladder (group IX). axial and lateral hindlimb muscles had to be cut, together with the skin, before injecting the superficial muscles substances as Fluorogold and Fast Blue, which label of the back or lateral hindlimb. In such cases, cutanei only a limited portion of the respective motoneuronal trunci motoneurons at the level of C8-T1 were labeled. cell group when injected intramuscularly (Richmond et Examples are the medial longissimus and biceps femoris al., 1994; personal observations). Two types of HRP (II anterior and posterior (cases 2300R, and 2255L/R). and VI) were used in this study. Type VI appeared to be However, after injections in for example the gastro- superior to type II, since small injections of type VI cnemius medialis and lateralis (cases 2239L/R), produced the same results as large injections of type II. adductor longus and iliopsoas (cases 2242L/R), vastus

Group IX

010203040506070 80 90 100 110 120 130

Levator ani 20 10 case 2252 L 0 L5 L6L7 S1 S2

20 10 case 2277 R 0 L5L6 L7 S1 S2

20 10 case 2311 R 0 L5 L6L7 S1 S2

20 10 case 2340 R 0 L5 L6L7 S1 S2 Abductor caudae internus 20 10 case 2325 R 0 L5 L6L7 S1

External anal sphincter 30 20 10 case 2252 R 0 L5 L6L7 S1 S2

20 10 case 2311 L 0 L5 L6L7 S1 S2

20 10 case 2354 L 0 L5 L6 L7 S1

Ischiocavernosus 30 20 10 case 2272 R 0 L5 L6L7 S1 S2

External urethral sphincter 30 20 Figure 31 Graphs showing the rostrocaudal 10 case 2356 L 0 distribution of labeled motoneurons after S2 L5 L6 L7 S1 injections into the muscles of the pelvic floor

0 10203040506070 80 90 100 110 120 130 (group IX). level

66 lumbosacral motoneuronal cell groups

Group IX: Pelvic floor and bladder

90 100 110 120 130

S3 Levator anii case 2311R

S1 Abductor caudae internus S2 case 2325R

S1 S2 External anal sphincter case 2354L

S1 External urethral sphincter S2 case 2356L

Ischiocavernosus S2 case 2272R

Bladder S3 case 2345L

Figure 32 Plottings showing the location of the motoneurons innervating the muscles of the pelvic floor and bladder (group IX).

AB

C D

Figure 33 Polarized light- darkfield photomicrograph of labeled motoneurons after injections into different hindlimb, pelvic floor and axial muscles. A: Motoneurons of the biceps femoris posterior at level 95 (rostral S1 in case 2255L), dividing Onuf’s nucleus into a dorsomedial and ventrolateral part. B: External anal sphincter motoneurons at level 95 (caudal L7 in case 2354L), forming a small cluster in the dorsomedial part of Onuf’s nucleus. Note the bundle of dorsally projecting dendrites. C: Motoneurons of the levator ani at level 115 (S2 in case 2252L). D: Multifidi motoneurons at level 80 (rostral L7 in case 2339L). Note the bundle of dendrites which extends dosomedially into the white matter. Bar represents 400 µm.

67 Chapter 2

Table 10 number Group X Axial muscles of neurons Muscle Origo/Insertion Function Nerve Case HRP 10% present

Longissimus O: S and Co vertebrae Extension of the Dorsal rami of 2292 R 30 µl type VI 8 9 medialis I: L, S, and Co vertebrae vertebral spinal nerves 2300 R 35 µl type VI 1 2 9 column and tail 2301 R 35 µl type VI 2 8 2339 L 15 µl type VI 8 8

Extensor caudae O: processi articulares of sacral and Extension of the Dorsal Co nerves 2292 L 30 µl type II 2 3 lateralis (DL) transverse processi of Co vertebrae tail 2300 L 15 µl type II 1 7 I: dorsal caudal vertebrae

Multifidi O: L, S, and Co1 vertebrae Extension and Dorsal rami of 2339 R 15 µl type VI 3 8 (lumbar portion) I: processus spinosus 2 vertebrae lateroflexion of spinal nerves 2346 L 15 µl type VI 5 1 rostrally to origin the spine 2346 R 15 µl type VI 8 1 Extensor caudae O: processus spinosus Sacral and Co1 Extension of the Dorsal Co nerves medialis I: processi articulares and dorsal tail (caudal multifidus) surface of caudal vertebrae

intermedius (case 2246R), sartorius medius and anterior Despite the relatively small amounts used in the present (cases 2264L/R, and 2338L/R), semimembranosus study, the number of retrogradely labeled motoneurons anterior and posterior (cases 2291L/R), gracilis (case did not seem to differ from the numbers obtained in other 2310L), and psoas major (case 2311L) no labeled cutanei studies (see Romanes, 1951; Boyd and Davey, 1968, trunci motoneurons were found in the C8-T1 spinal and others in Tables I to X). segments (see also Haase and Hrycyshyn, 1985; In two muscles, (semitendinosus and semimem- Holstege et al., 1987; Haase, 1990). branosus), a few labeled motoneurons were repeatedly Notwithstanding the fact that intact fasciae appeared to observed in motoneuronal cell groups of non-neigh- properly prevent leakage of tracer, in a few cases with bouring muscles (see for example Fig. 15). The mechan- injections into muscles belonging to groups without well ism of this phenomenon, which has also been described developed fasciae (e.g. the group of the intrinsic muscles by Burke et al. (1977) is unknown. of the foot and the group of the adductors brevis and magnus), the distribution patterns of labeled moto- Compatability of the different cases neurons not only represented the motoneurons of the The solution to the problem of comparing the location injected muscle, but, in all likelihood, also motoneurons of motoneuronal cell groups in different cases, because of other muscles of this same group. of variations in length and segmental subdivision Leakage was also prevented by limiting as much as (Sherrington, 1892; Romanes, 1951; Rexed, 1954) is possible the amounts of injected tracer. In the present one of the main assets of the present study. Until now, study, between 5µl 10% HRP (levator ani in case 2340R) the only method to overcome this problem was the use and 90µl 10% HRP (for example vastus lateralis in case of different tracers in one and the same animal. However, 2209R) was injected. In most other studies, much larger this approach does not allow integration of more than a quantities were injected (Burke et al, 1977: 500-1000µl few motoneuronal cell groups. The method used in the 50% HRP into the soleus and medial gastrocnemius; present study is based on the observation that the spatial Iliya and Dum, 1984: 600µl 10-20% HRP into the tibialis interrelation of the hindlimb and pelvic floor moto- anterior; Kernell et al., 1985: 500-1000µl 3.5-7% HRP neuronal cell groups always remained the same. into the peroneus longus; and Horcholle et al., 1988: Therefore, a relative instead of absolute (milimeter or 10-30µl 50% HRP into the peroneal muscles). segmental) subdivision was developed, using two easily detectable landmarks as level 0 and 100. With this Group X method it appeared possible to integrate the results of Axial muscles all 135 cases into one scheme (Fig. 38). lumbar multifidi Location of the various motoneuronal cell groups extensor caudae in groups I to X L1 L2 L3 L4 L5 L6 medialis L7 S Group I. The motoneuronal cell groups innervating Co the hip flexors psoas major and sartorius are present ventrolaterally in the ventral horn in the most rostral medial longissimus part of the lumbar enlargement (Figs. 37 and 38). Romanes (1951) categorized the sartorius motoneuronal cell group in his column 1, in which he also included extensor caudae lateralis the rectus femoris motoneurons. Indeed, part of the motoneurons of the rectus femoris are located among lateral aspect those of the sartorius, but another part of the rectus Figure 34 Schematic drawings of the extensors of the lower motoneuronal cell group overlapped with the cel lgroup back and proximal tail (group X). of the knee extensor vastus medialis. This partial

68 lumbosacral motoneuronal cell groups

Group X

010203040506070 80 90 100 110 120 130 140

Medial longissimus and Extensor caudae lateralis 20 10 case 2292 R 0 L5 L6 L7 S1

20 10 case 2300 R 0 L5 L6L7 S1 S2

20 10 case 2301 R 0 L5 L6 L7 S1

20 10 case 2339 L 0 L5 L6 L7 S1 S2 S3

20 10 case 2292 L 0 L5 L6 L7 S1

20 10 case 2300 L 0 L5 L6 L7 S1 S2

Multifidi and Extensor caudae medialis 20 10 case 2339 R 0 L5 L6 L7 S1

20 10 case 2346 R 0 Figure 35 Graphs showing the rostro- L5 L6L7 S1 S2

20 caudal distribution of labeled moto- 10 case 2346 L 0 neurons after injections into the extensors L5 L6L7 S1 S2 of the lower back and proximal tail

0 10203040506070 80 90 100 110 120 130 140 (group X). level

separate muscles. sartorius/rectus overlap is in agreement with the notion that the sartorius also acts as an extensor of the knee, as Group II. Quadriceps motoneurons (rectus femoris, is true for the rectus femoris and the vastus medialis. vastus medialis, lateralis and intermedius) form a The finding that the majority of the medial sartorius compact group which rostrally shows some overlap with motoneurons are located slightly more rostral than the the sartorius cell group (Figs. 37 and 38). Romanes anterior sartorius motoneurons is in agreement with the (1951) included the quadriceps motoneurons, with the findings of Gordon et al., (1991). Possibly, a study on exception of those of the rectus femoris, in his column the location of sartorius motoneurons in the dog might 2. Remarkably, he also included the motoneuronal cell produce more precise results, because in this animal the groups of the anterior compartment of the distal hindlimb functionally different medial and anterior sartorius are (group VIII in the present study) in his column 2. Neither

Group X: Axial muscles 4050 60 70 80 90 100 110 Medial longissimus S2 case 2300R

L6 L7 Multifidus S1 case 2346L

Figure 36 Plottings showing the location of the motoneurons innervating the extensors of the lower back and proximal tail (group X).

On the next pages: Figure 37 (pp. 70 and 71) Summary overview of the location of the motoneuronal cell groups in each of the 10 muscle groups. At the bottom, an overview is given of the spatial interrelation of groups I to X. Figure 38 (pp. 72 and 73) Summary overview, showing the location of the motoneuronal cell groups of all muscles of groups I to X.

69 Chapter 2 60 60 sartorius medialis iliacus psoas major sartorius anterior

50 50 90 100 80 10 50 60 70 50 60 70 80 90 100 10 20 30 40 0203040 vastus medialis vastus lateralis rectus femoris semimembranosus semitendinosus biceps femoris posterior biceps femoris anterior plantaris gastrocnemius medialis soleus gastrocnemius lateralis tibialis posterior vastus intermedius caudofemoralis obturatorius internus tensor fasciae latae gluteus medius gluteus maximus adductor longus adductor brevis adductor magnus pectineus gracilis

Group I: Iliopsoas and sartorius Group IV: Hamstrings Group V: Gluteal muscles and other proximal of the hip Group II: Quadriceps Group VI: Posterior compartment of the distal hindlimb Group III: Adductors

70 Fig. 37 lumbosacral motoneuronal cell groups 60 110 110 Group IX: Pelvic floor and bladder Group X: Axial muscles 50 130 110 120 100 120 (±L6) Group V: Gluteal and other proximal muscles of the hip Group VI: Posterior compartment of the distal hindlimb Group VII: Long flexors and intrinsic muscles of the foot Group VIII: Anterior compartment of the distal hindlimb (±S1) (±S2) 100 110 Group I: Iliopsoas and sartorius Group II: Quadriceps Group III: Adductors Group IV: Hamstrings (±L5) (±S3) 60 70 80 90 60 70 80 90 100 10 (±L7) (±S1) 0203040 60 70 80 90 100 120 130 external anal sphincter ischiocavernosus external urethral sphincter bladder levator ani extensor digitorum longus tibialis anterior peroneus longus flexor digitorum longus adductor digiti intrinsic foot (n. tibialis) flexor hallucis longus flexor digitorum brevis Group IX: Pelvic floor and bladder Group VII: Anterior compartment of the distal hindlimb Group VIII: Long flexors and intrinsic muscles of the foot

Fig. 37 71 Chapter 2 adductor brevis adductor longus adductor magnus ext. digitorum longus flexor digitorum longus gluteus medius gracilis iliacus multifidi obturatorius internus pectineus peroneus longus psoas minor and major rectus femoris sartorius medius sartorius anterior semimembranosus semitendinosus tensor fasciae latae vastus intermedius vastus lateralis vastus medialis rectus femoris flexor digitorum longus adductor longus vastus medialis 30 sartorius anterior and medius gracilis sartorius anterior obturatorius internus semimembransosus pectineus extensor digitorum longus adductor magnus vastus lateralis tensor fasciae latae (±L6) vastus medialis psoas major sartorius medius rectus femoris semimembransosus adductor magnus gracilis sartorius anterior psoas major psoas major sartorius anterior and medius adductor brevis adductor longus adductor brevis sartorius medius gracilis obturatorius internus iliacus iliacus (±L5) vastus lateralis adductor magnus level 50 level 10 level 30 vastus intermedius tensor fasciae latae adductor longus adductor brevis psoas minor and major pectineus gracilis gracilis iliacus adductor magnus rectus femoris gracilis sartorius anterior and medius adductor magnus rectus femoris multifidi vastus medialis vastus lateralis psoas major psoas minor and major adductor longus adductor brevis vastus medialis adductor longus iliacus iliacus level 40 level 20 level 0

72 Fig. 38 lumbosacral motoneuronal cell groups adductor digiti biceps femoris anterior biceps femoris posterior bladder caudofemoralis ext. digitorum longus external anal sphincter ext. urethral sphincter flexor digitorum brevis flexor digitorum longus flexor hallucis longus gastrocnemius medialis gastrocnemius lateralis gluteus maximus gluteus medius intrinsic foot ischiocavernosus levator ani longissimus medialis multifidi obturatorius internus peroneus longus plantaris semimembranosus semitendinosus soleus tensor fasciae latae tibialis posterior adductor digiti caudo- femoralis flexor digitorum brevis soleus bladder plantaris external urethral sphincter and ischiocavernosus gluteus maximus longissimus medialis external anal sphincter biceps femoris anterior peroneus longus semitendinosus flexor digitorum longus biceps femoris posterior intrinsic foot gluteus medius longissimus medialis external anal sphincter obturatorius internus tibialis anterior bladder levator ani levator ani flexor hallucis longus gluteus medius intrinsic foot ischiocavernosus biceps femoris posterior multifidi bladder gastrocnemius medialis levator ani semimembranosus and semitendinosus flexor digitorum longus multifidi peroneus longus level 70

level 90 (±S1) level 110 (±S2) gluteus medius adductors digiti tibialis posterior external urethral sphincter and ischiocavernosus caudofemoralis gastrocnemius medialis plantaris semitendinosus 50 longissimus medialis levator ani adductor digiti multifidi extensor digitorum longus biceps femoris anterior external urethral sphincter and ischiocavernosus semimembranosus peroneus longus gastrocnemius lateralis longissimus medialis gluteus medius tibialis anterior tensor fasciae latae and gluteus medius (±L7) soleus flexor digitorum longus flexor digitorum brevis gluteus maximus obturatorius internus semimembranosus semitendinosus longissimus medialis obturatorius internus level 100 level 80 level 60

73 Fig. 38 Chapter 2

anatomically, (group VIII motoneurons are not only 1977; Weeks and English, 1985; 1987) describe the located more caudally, but also more dorsally, see Figs. functional properties of the muscles of the posterior 37 and 38), nor functionally, (group VIII consists of compartment of the distal hindlimb, no studies exist on dorsal flexors of the ankle, while the quadriceps muscles the precise location of their motoneuronal cell groups. are extensors of the knee), does there seem to be an Especially the spatial interrelationship of the different overlap between groups II and VIII. posterior hindlimb motoneuronal cell groups has never been clarified. The present results indicate that the Group III. The groups adductor motoneurons are plantaris motoneurons in the caudal part of their cell located in the rostral half of the enlargement, medial group are intermingled with the motoneurons of the and dorsal to the motoneuronal cell groups of groups I lateral gastrocnemius, but the plantaris motoneuronal (iliopsoas and sartorius) and II (quadriceps; Figs. 37 and cell group extends much further rostrally. This is in 38). In the caudal part of group III, its motoneurons contrast to Romanes (1951), who reported that the (especially those innervating the gracilis) become plantaris and lateral gastrocnemius motoneuronal cell intermingled with the rostral pole of the hamstring group groups would be located at the same level, with the (IV; semimembranosus). Group III was called column plantaris motoneuronal cell group dorsal to that of the 3 by Romanes (1951). lateral gastrocnemius. Therefore, he included the plantaris motoneurons in his column 4, which also Group IV. The hamstring motoneuronal cell group (IV) contains the flexor digitorum longus and flexor hallucis is present centrally in the ventral horn in the caudal half longus, and the lateral gastrocnemius motoneuronal cell of the enlargement and has a clear rostrocaudal group in his column 5, which is located ventral to column organization (Figs. 37 and 38). Romanes split the 4. The present results clearly indicate that the difference hamstring group into a semimembranosus and semi- between the two motoneuronal groups is not dorso- tendinosus cell group (his column 3’), and a biceps ventral, but rostrocaudal, with the caudal third of the femoris posterior and anterior group (his column 3”). plantaris motoneurons intermingled with the lateral In line with the present results, Hamm et al. (1985) have gastrocnemius motoneurons. demonstrated a slight topographic organization of The medial gastrocnemius and soleus motoneurons are semimembranosus motoneurons and their Ia afferent located slightly more caudally and laterally than the connections, with the anterior motoneurons located, on lateral gastrocnemius and plantaris motoneuronal cell average, rostral to the posterior cells. Regarding the groups. The bulk of the medial gastrocnemius semitendinosus motoneuronal cell group, the present motoneurons are located caudal to those of the soleus. findings are in parallel with those of Hoover and These findings correspond with those of Burke et al., Durkovic (1991). Also in agreement with the results of (1977). Romanes (1951) included both cell groups in the present study, Letbetter and English (1981) found his column 5. the posterior biceps motoneurons caudal to those of the anterior biceps. Similarly, Botterman et al. (1983a) Group VII. The motoneurons innervating the muscles reported that Ia afferents of the posterior biceps of the anterior compartment of the distal hindlimb terminated more caudally in group IV than the Ia (peroneus longus, extensor digitorum longus and tibialis afferents of the anterior biceps. anterior) are located in the dorsolateral part of the ventral horn, laterally to group VI (posterior compartment; Figs. Group V. With the exception of the obturator internus 37 and 38). The motoneurons of all three cell groups and the abductor cruris caudalis motoneurons, all are intermingled in so far that the bulk of the extensor motoneurons of the proximal hip muscles are located at digitorum longus motoneurons are located more rostrally the border with the white matter in the ventral or within this cell group than those of the tibialis anterior. ventrolateral part of the ventral horn throughout the In contrast, Hoover and Durkovic (1991) reported tibialis caudal half of the enlargement (Figs. 37 and 38). anterior motoneurons to be located ventral to those of Romanes (1951) included the motoneurons of the tensor the extensor digitorum longus. Careful study of the fasciae latae, gluteus medius and maximus in the caudal motoneuronal cell groups of these two muscles on the part of his column 1, and those of the obturator internus left and right side in one and the same cat (cases 2333 in his column 3. Within group V of the present study, and 2336) did not leave any doubt that the bulk of the the motoneurons innervating the anterior part of the extensor digitorum motoneurons are located more rostral tensor fasciae latae were located more rostral than the than those of the tibialis anterior, but that their motoneurons of the posterior part of this muscle. This dorsoventral location is the same. subdivision was not found by Gordon et al., (1991), who, using the nerve soak technique, reported motoneurons Group VIII. The muscles acting on the most distal parts of these two parts of the tensor fasciae latae to be of the hindlimb are innervated by a very large group of completely intermingled. motoneurons, located dorsal to all other hindlimb motoneurons (Figs. 37 and 38). Within this group, the Group VI. Although a number of reports (Burke et al., more proximal muscles, flexor digitorum longus and

74 lumbosacral motoneuronal cell groups

flexor hallucis longus, have their motoneurons located rostrally, immediately dorsal to those of the plantaris and lateral gastrocnemius. This location of the flexor Lateral view digitorum longus motoneurons is in complete agreement with the findings of Dum et al., 1982. The motoneurons of the intrinsic foot muscles form the caudal two thirds of this cell group. Romanes (1951) included the long flexor motoneurons in his column 4 and those of the intrinsic foot muscles in his column 6. Group IX. From rostral to caudal, the motoneurons innervating the external anal sphincter, external urethral sphincter, ischiocavernosus and bladder form a continuous cell column (Figs. 37 and 38). The rostral two thirds of it correspond to Onuf’s nucleus, and are divided into a dorsomedial and a ventrolateral part. The dorsomedial portion innervates the external anal sphincter and the ventrolateral part innervates the ischiocavernosus/external urethral sphincter (Sato et al., superficial deep 1978; Yamamoto et al., 1978; Pullen, 1988; Thor et al., 1989; Beattie et al., 1990). At the rostral pole of Onuf’s nucleus, the two subgroups are separated by the caudal Medial view portion of the posterior biceps motoneuronal cell group. In its most caudal part, Onuf’s nucleus contains a few parasympathetic preganglionic neurons, which has also been reported by Thor et al., (1989). In agreement with the findings of Sato et al. (1978), levator ani motoneurons form a small, separate group in the ventromedial part of the ventral horn.

Group X. The axial muscles in cats extend from the occipital bone of the skull to the tip of the tail. Their main function is postural control. Throughout the spinal cord, motoneurons innervating axial muscles are located medially in the ventral horn (see Sprague, 1948; Fetcho, 1987; Kitamura and Richmond, 1994). A large number of these muscles belong to the longissimus dorsi group. superficial deep The continuous motoneuronal cell group of the lateral level 0 level 100 longissimus begins at C8 and continues until L5 (Holstege et al., 1987). The medial longissimus/extensor caudae lateralis muscles are located more caudally and their motoneurons were located in the so-called nucleus ventromedialis of Rexed (1954) between levels 85 and 115 and further caudally (Ritz et al., 1992). At levels rostral to 85 medial longissimus motoneurons were not Figure 39 Schematic drawings showing the relation between found, which is in agreement with Rexed’s (1954) the rostrocaudal location of the motoneuronal cellgroups and finding of an absence of motoneurons in the nucleus the location of the muscle they innervate. The muscles are ventromedialis at caudal lumbar levels (see also Elliott, filled with a 0-100% grayscale, corresponding to the level (0- 1942 in the human). Ritz et al. (1992) found some of 100) of the rostral pole of their motoneuronal cell group. the medial longissimus motoneurons in a more Note that for the muscles which originate from the vertebral dorsomedial position, in the so-called nucleus column, those with a rostral origo have their motoneuronal cell groups rostral in the lumbosacral cord than those with a commissuralis (Rexed, 1954), which could not be more caudal origo. For muscles with an origo in the proximal confirmed in the present study. or distal hindlimb, those which are located anterior in the The lumbosacral multifidi/extensor caudae medialis hindlimb have their motoneuronal cell groups at more rostral motoneurons were exclusively located in the nucleus levels in the lumbosacral cord than those with a more posterior commissuralis, dorsomedial to the medial longissimus origo. Moreover, medial muscles of the thigh are innervated motoneurons, which is in agreement with the findings by motoneurons located more rostrally in the lumbosacral than of Ritz et al. (1992) and Gillette et al. (1993). those of the lateral thigh muscles. Interestingly, multifidi motoneurons do not form a

75 Chapter 2

continuous column, but were found in only the most (Botterman et al., 1983a; Hamm et al., 1985), and not rostral part of the L6, L7, S1, and S2 segments. However, for muscles which act uniformly, such as the semi- since only part of the multifidi/extensor caudae medialis tendinosus (Botterman et al., 1983b). muscles were injected, it is possible that this distribution pattern continues further rostrally and caudally. Three dimensional organization of the lumbo- sacral motoneuronal cell groups Different organization of the mid-lumbar and The results presented in this paper, based on a new lower lumbosacral ventral horn rostrocaudal subdivision (level 0-100) of the lumbosacral The enlargement of the lumbosacral ventral horn is cord, clearly indicate that the three dimensional location caused by the presence of large numbers of moto- of the motoneuronal cell groups in this part of the cord neurons. Especially at caudal levels (55-110), moto- is not random. It appeared to be organized in a specific neurons occupy the entire lateral part of the ventral horn. manner in all three dimensions, rostrocaudal, dorso- Rostral to these levels, motoneurons are only present ventral and mediolateral, which will be subsequently ventrolaterally in the ventral horn, leaving the dorsal discussed. part to another kind of large neurons (Fig. 37). Rexed (1954) mistook them for motoneurons, calling this area Rostrocaudal organization. the dorsolateral nucleus of lamina IX. Romanes (1951) Figure 39 shows that, in respect to the muscles included these neurons in his column X (ex, not 10), originating from the vertebral column (iliopsoas part of but because he never observed any clearly retrogradely group I and all muscles of groups V and X), the more degenerated cells in this area, he suggested column X rostral a muscle has its origin, the more rostral the to contain interneurons. More recent studies have clearly location of its motoneuronal cell group. The same is shown that Romanes column X contains large or true for the muscles of the hindlimb: the more anterior medium sized cells projecting to the cerebellum (Burke the muscle is located within the limb, the more rostral et al., 1971; Matsushita et al., 1979; Grant and Xu, 1988). its motoneuronal cell group is located in the cord. Other conclusions that can be drawn from figure 39 are that Functional organization of lumbosacral moto- medial muscles of the upper hindlimb have their neuronal cell groups motoneuronal cell groups more rostral in the cord than In the present study, the muscles innervated by the lateral muscles, and that the more distal a muscle is lumbosacral motoneurons have been divided into 10 located, the more caudal its motoneuronal group can be groups, based upon the main functions of the respective found in the lumbosacral cord. group of muscles. There exists only slight overlap A similar rostrocaudal organization seems to exist within between the motoneuronal cell groups of these 10 groups a motoneuronal cell group of a given muscle, but only (Fig. 37). There are a few exceptions, i.e. the sartorius when that muscle shows compartmentalization, i.e. when motoneuronal cell group belonging to group I partially different regions of a muscle are separately innervated, overlaps with the motoneuronal cell groups of the vastus for example in case of the peroneus longus and triceps medialis and rectus femoris of group II, and the adductor surae (Swett et al., 1970; English and Letbetter, 1982; motoneurons of group III overlap with the semimem- Donselaar et al., 1985; Stuart et al., 1988; Weeks and branosus motoneuronal cell group of group IV. These English, 1985; 1987; Cramer and Essen, 1995). overlapping distribution patterns can be explained by Since the flexor muscles of the hip have to be located, partly overlapping functions. For example, both the by definition, anterior to the hip extensors, the hip flexor anterior part of the sartorius (group I) and vastus motoneuronal cell groups are located more rostral in medialis/rectus femoris (group II) extend the knee. the ventral horn than the hip extensors. The same is true Similarly, the motoneuronal cell groups of functionally for the knee, but reversely, i.e. the knee flexors are different parts of a muscle, for example the anterior and located by definition in a posterior and the knee extensors posterior parts of the semimembranosus and biceps in an anterior position, resulting in a more rostral location femoris, respectively, are topographically organized. of the knee extensor motoneurons in the spinal cord These findings support the concept of the “partitioning” compared to knee flexor motoneurons. of Ia excitatory post-synaptic potentials (see Stuart et al., 1988). In this concept Ia afferents from a given Dorsoventral organization. muscle make more effective connections with the From figures 37 and 38, the impression is gained that motoneurons innervating the region of the muscle from the dorsoventral location of a motoneuronal cell group wich the Ia afferents are derived, than with motoneurons depends on the joints its muscle acts on. The muscles of innervating other parts of the muscle motoneurons. the most distal joints have their motoneuronal cell groups Partitioning has been demonstrated for muscles with most dorsal in the ventral horn. Such an organization regions that are capable of independant and different was already noted by Bikeles (1905) and Romanes (1951 actions such as the anterior and posterior parts of the and 1964). A dorsoventral organization as proposed for semimembranosus and biceps femoris, respectively the brachial cord by Sterling and Kuypers (1967), in

76 lumbosacral motoneuronal cell groups which the flexors are always located dorsal to the extensors, does not apply to the lumbosacral cord, as has already been noted by these authors themselves.

Mediolateral organization. Concerning the mediolateral organization of lumbo- sacral motoneuronal cell groups it is observed that the hip adductor and abductor motoneurons are present medial in the ventral horn, those of the flexors of the hip or knee lateral to them, and the extensors of the hip or knee even further laterally. Sharrard (1955) already noticed such a mediolateral organization in his paper on the location of lumbosacral motoneuronal cell groups in the human. Regarding the motoneuronal cell groups of the muscles acting on the ankle and digits, the dorsoflexors or flexors are located lateral to the plantar flexors or extensors. At first glance, this seems to be in contrast to the organization of the hip and knee motoneuronal cell groups, where the flexors lie medial to the extensors. However, physiological experiments of Engberg (1964) have shown that the motoneurons of plantar flexors behave like flexor motoneurons. Thus, our finding that the plantar flexor motoneurons are located medial to the motoneurons of the dorsoflexors is in agreement with the hip and knee flexors moto- neuronal cell groups to be located medial to the hip and knee extensor motoneuronal cell groups.

.

77

.

NRA projections to lumbosacral motoneurons: non-estrous females

Chapter 3

79

NRA projections to lumbosacral motoneurons: non-estrous females Caudal medullary pathways to lumbosacral motoneuronal cell groups in the cat; evidence for direct projections possibly representing the final common pathway for lordosis

ABSTRACT The nucleus retroambiguus (NRA) projects to distinct brainstem, and cervical and thoracic cord motoneuronal cell groups. The present paper describes NRA projections to distinct motoneuronal cell groups in the lumbar enlargement. Lumbosacral injections of wheat germ agglutinin-horseradish peroxidase (WGA-HRP) were made to localize and quantify the retrogradely labeled neurons in the caudal medullary lateral tegmentum. These injections were combined with spinal hemisections in order to distinguish between neurons having ipsi- or contralaterally descending axons. The NRA-lumbosacral fibers desend almost exclusively contralaterally, but neurons in areas surrounding the NRA project mainly ipsilaterally. In an anterograde tracing study, injections of WGA-HRP or tritiated leucine were made in the region of the NRA to determine the NRA targets in the lumbosacral cord. Hemisections in C2 made it possible to distinguish between NRA projections and projections from neurons in the adjoining lateral tegmentum. The results show delicate NRA projections to distinct lumbosacral motoneuronal cell groups innervating specific hindlimb muscles (iliopsoas, adductors, and hamstrings), as well as axial muscles (medial longissimus and proximal tail muscles). The projection is bilateral with a contralateral predominance. Ipsilaterally terminating fibers are derived from NRA neurons which axons cross the midline at the level of the obex, descend through the contralateral spinal white matter, and recross at the level of termination. A concept is presented in which the periaqueductal gray-NRA-lumbosacral projections form the final common pathway for lordosis in the cat.

INTRODUCTION The nucleus retroambiguus (NRA), first described in anal sphincters (Holstege and Tan, 1987). humans by Olszewski and Baxter in 1954, is a Electrophysiologically, it has been shown that the NRA rostrocaudally organized column of interneurons in the contains expiration related neurons, although its rostral ventrolateral part of the medulla oblongata. In the cat, it part contains inspiration related cells as well (Merrill, extends from 1 to 8 mm caudal to the obex and 1970; 1974). NRA interneurons are known to be corresponds with the caudal part of the ventral involved in expiration, as well as in vomiting, defecation respiratory group (for review see Feldman, 1986). and vocalization (Fukuda and Fukai, 1986a,b; Miller et NRA interneurons are known to project to specific al., 1987; Holstege, 1989; Zhang et al., 1992). The NRA receives, in addition to projections from other brainstem and spinal cord motor nuclei. In the brainstem, respiration related neurons in the brainstem (for reveiw the NRA projects to somatic motoneurons located in see Feldman, 1986), specific projections from the the nucleus ambiguus, innervating the larynx, pharynx midbrain periaqeductal gray (PAG) (Holstege, 1989; and soft palate (Holstege, 1989; Zhang et al., 1992). Davis and Zhang, 1991). The PAG is the final integrator NRA neurons project to the spinal cord as well. In the for various components of emotional behavior, such as cervical cord, NRA fibers terminate in motoneuronal vocalization, jumping, blood pressure control, mictu- cell groups innervating the diaphragm, pectoral muscles, rition and lordosis (Sakuma and Pfaff, 1979b; Holstege, and cutaneus trunci muscles (Holstege and Kuypers, 1989; Bandler et al., 1991; Blok and Holstege, 1994). 1982; Feldman et al., 1985; Holstege and Blok, 1989). The intense PAG-NRA projection raises the question In the thoracic and upper lumbar cord (T1-L3), the NRA whether the NRA is involved in behavior other than projects to motoneurons innervating intercostal and expiration or expiration related activities. Therefore an abdominal muscles (Holstege and Kuypers, 1982). In attempt was made to carefully study its descending the sacral cord (S1-S2) the NRA has been shown to pathways to the spinal cord. In this paper the NRA project to the nucleus of Onuf innervating the pelvic projections to the lumbosacral cord are presented. The floor musclulature including the external urethral and results strongly suggest a role for the NRA in lordosis.

81 Chapter 3

MATERIALS AND METHODS Photomicrographs of represen-tative sections were taken. The collected sections (1:4) were grouped into Retrograde tracing study nine levels between the obex (level 0 in our study) and 8 mm caudal to the obex (level 8). Each level represented In order to localize NRA neurons projecting to the six sections. For each level, the retrogradely labeled cells lumbosacral cord, in six adult female cats (3.2-6.7 kg) were counted and plotted into one drawing by means of injections of 2.5% wheatgerm agglutinin-horseradish a drawing tube. peroxidase (WGA-HRP) were made in the L6 (cases 2249, 2278, and 2280), L7 (case 2295), S1 (case 2261) Anterograde tracing study and S2 (case 2265) segments. See Chapter 1 for the Tritiated leucine cases general surgical and histological procedures. After In seven cats (all females except for cases 1183 and laminectomy of the lumbar vertebrae, multiple injections 1472), weight 2.0 -3.5 kg, single injections with L-(4,5- of WGA-HRP were made with a glass pipette using a 3H)-leucine (specific activity > 100 Ci/mmol) were pressure pump. The injections occupied the entire placed in the region of the NRA. In two of these cases rostrocaudal extent of the respective segments, but were (1019 and 1183) the injections were made at a more kept unilaterally, except for case 2295 in which a rostral level than in the other five cases cases (1471, bilateral injection was made. Unilateral injections were 1472, 1571, 1683, and 1684). After dorsal approach and made to distinguish between neurons projecting to the exposure of the caudal medulla, the injection in each ipsi- and contralateral spinal gray matter, respectively. In order to differentiate between ipsilateral or cat was placed stereotactically with a Hamilton contralateral descending pathways, prior to the L6 microsyringe fitted with a 22-gauge needle. In all cases µ injection an ipsilateral L3 hemisection was made in case 0.5 l tritiated leucine was injected over a period of 5 µ 2278, and a contralateral hemisection was made in case minutes except for case 1571, in which 0.25 l tritiated 2280. In case 2295 the bilateral L7 injection was leucine was injected. The needle was left in place for an preceded by a left sided L3 hemisection. All hemi- additional 30 minutes to minimize the spread along the sections were made by aspiration with a glass pipette. needle track. After a survival period of 6 weeks (Holstege After 3 days survival time, the animals were perfused et al., 1979), the animals were deeply anesthetized and as described in Chapter 1. The caudal brainstem was perfused with saline followed by 10% formalin. After separated from the rest of the brain with a stereotactical postfixation in 10% formalin for at least one week, brain µ cut just caudal to the facial nucleus. The caudal medulla and spinal cords were cut into transverse 25 m frozen was separated from the spinal cord rostral to C2. The sections. One series of every tenth section was mounted, caudal medulla and C1 segment were cut on a freezing coated with Ilford G5 emulsion by dipping, and stored ° microtome into 40 µm transverse sections, keeping the in the dark at 5 C for 3 months. Subsequently, the ° angle between brainstem and spinal cord was kept as material was developed with Kodak D19 at 16 C, fixed, small as possible. One out of four sections was processed and counterstained with cresyl violet. The sections were using the tetramethylbenzidine (TMB) procedure studied with a Wild darkfield M7S microscope, and according to Mesulam (1982). To determine whether photomicrographs were taken of representative spinal the injection sites were restricted to one side of the spinal cord sections. In each experiment, the injection area was cord, sections of the injected segments were processed defined as that area in which the silver grains over the with diaminobenzidine (DAB). The hemisected cell bodies were either as numerous as, or more segments were cut and incubated in TMB in order to numerous than those over the surrounding neuropil verify if the hemisections were complete. Finally, the (Holstege et al., 1977). sections were mounted on slides, dehydrated and coverslipped with DePeX mounting medium. The WGA-HRP cases distribution of retrogradely labeled neurons was Although the autoradiographical tracing results microscopically examined with a Zeiss Axioskop under demonstrated NRA-lumbar motoneuronal projections, combination of polarized light and darkfield condensor. the material (only 1:10 non serial sections were

Table 1. Number of counted neurons in the NRA and adjoining lateral tegmental field after spinal injections of WGA-HRP

group of labeled neurons labeled neurons in NRA ventromedial to NRA absolute number % absolute number % Case Injection site WGA-HRP Hemisection ipsi contra total ipsi contra ipsi contra total ipsi contra 2249 caudal L5-rostral L7 unilateral 4 µl 2.0% - 1 1 5 3 2 6 4 4 1 2 6 7 4 4 6 8 9 0 5 5 8 8 4 1 6 2278 caudal L5-rostral L7 unilateral 3 µl 2.5% ipsi L3 1 1 9 9 1 2 8 9 3 7 0 1 5 1 5 0 1 0 0 2280 caudal L5-rostral L7 unilateral 3 µl 2.5% contra L3 2 4 3 3 4 3 5 8 7 9 3 3 4 9 4 0 3 8 9 9 0 1 0 2295 caudal L6-L7 bilateral 6 µl 2.5% left L3 3 8 4 1 9 4 0 3 9 5 5 1 8 3 1 1 3 2 9 5 9 5 2261 caudal L7-S1 unilateral 3 µl 2.5% - 1 0 1 2 5 9 3 6 0 2 8 7 2 3 5 0 1 1 2 4 6 2 7 6 2 4 2265 caudal S1-S2 unilateral 2 µl 5.0% - 1 0 9 2 4 8 3 5 7 3 0 7 0 3 3 0 9 7 4 2 7 7 8 2 3

82 NRA projections to lumbosacral motoneurons: non-estrous females processed) was not sufficient to precisely determine the for case 2278 in which the dorsal funiculus and exact location of the NRA terminations within the contralateral lamina X were not involved. In case 2265 motoneuronal cell groups. Therefore, a series of WGA- (S2 injection) the injection site extended into larger HRP tracing experiments was done, partly in com- portions of the contralateral gray. The L3 hemisections bination with hemisections, in which 1:4 sections were (cases 2278, 2280, and 2295) were complete and did processed serially. not extend across the midline. In each of 8 female cats (cases 2237, 2251, 2256, 2258, 2267, 2271, 2286, and 2290) 2-5 injections of 20-30 nl Cases without hemisections (2249, 2261, and 2.5% WGA-HRP were rostrocaudally placed in the 2265) region of the NRA. In all cases the injections were placed Ipsilateral side At the level caudal to the obex two after dorsal approach and exposure of the caudal medulla groups of labeled neurons were found; one group within and in cases 2286 and 2290 of the first two cervical the confines of the NRA and another group in the segments. The injections not only involved the NRA, tegmentum just medial to it. In the NRA the bulk of the but also the adjacent lateral tegmentum which implies labeled neurons was found at levels 2 to 6 (2-6 mm that the labeled fibers in the spinal cord could be derived caudal to the obex; Figs. 1 and 2 left). The neurons in from NRA neurons or from neurons in the adjacent the medially adjacent tegmentum formed a compact cell tegmentum. However, the retrograde tracing experiments group between levels 3 and 8. Further rostrally this cell showed that more than 93% of the NRA fibers crossed group became more dispersed and contained more the midline and descended in the contralateral funiculi, neurons. The tegmental group was found exclusively while neurons in the tegmentum medial to the NRA sent ipsilaterally. Some other dispersed neurons were found their fibers through the ipsilateral spinal cord. Therefore, in the tegmentum between NRA and central canal. Other interruption of the ipsi- or contralateral pathways at labeled neurons were present in the ventral part of the upper cervical levels would allow to distinguish between ipsilateral gracile nucleus and a few in lamina I at the the termination sites of the contralateral NRA-spinal and transition between the caudal spinal trigeminal nucleus ipsilateral reticulospinal pathways. An ipsilateral C2 and the C1 spinal cord. hemisection was made in case 2286, and a contralateral C2 hemisection was made in case 2290. Contralateral side In all 3 cases a large number of In all anterograde tracing cases the same histological labeled neurons were present in the NRA, some in the procedures were applied as described for the retrograde lateral tegmental field between NRA and central canal, WGA-HRP experiments. After perfusion, brain and and some in the ventrolateral solitary nucleus. At levels spinal cord were removed and the injection site was just caudal to the obex, some neurons were located just determined using the DAB procedure. Segments L1 to medial to the lateral reticular nucleus. µ S3 were cut on a freezing microtome into 40 m thick, In the L6-injected case (2249), the total number of transverse sections and a 1 out of 4 series of consecutive retrogradely labeled NRA neurons was slightly larger sections was processed using the TMB procedure and than in the S1- and S2-injected cases (2261 and 2265; microscopically studied using the same technique as see table 1). In each of the 3 cases, 70-74% of the NRA described for the retrograde tracing study. Photo- neurons was found on the contralateral and 26-30% on micrographs were taken from representative sections. the ipsilateral side. In case 2286 with the ipsilateral C2 hemisection, the labeled fibers of every 6 consecutive collected sections Cases with hemisections (2278, 2280, and 2295) were plotted into one drawing. In case 2290, showing Although the 3 unilaterally injected cases revealed the the ipsilateral pathway, the distribution pattern of side of termination of the lumbosacral cord projecting anterogradely labeled fibers was similar throughout the NRA neurons, the question remains whether the NRA- length of the lumbosacral cord. For each segment, one lumbosacral pathways descend ipsi- or contralaterally drawing was made showing the labeling of 3 consecutive through the spinal cord. Figure 3 depicts all six collected sections. theoretically possible trajectories (A to F). Combining the unilateral L6-injections with ipsi- and contralateral RESULTS L3 hemisections reveals the trajectories of these pathways.

Retrograde tracing study Contralateral hemisection (case 2280) In this case, Location of injections and hemisections the distribution pattern of labeled neurons in the NRA In 6 cases WGA-HRP was injected in the lumbosacral was similar to the cases without hemisection, except that cord. Table 1 shows the level of the injections and on the ipsilateral side labeled neurons in the NRA were hemisections as well as the amounts of tracer. All sparse and amounted to only 7% of the total number of unilateral injections involved the entire left side of the labeled NRA neurons, while in non hemisected cases cord and extended into the contralateral lamina X, except the ipsilaterally located NRA neurons amounted 26-30%

83 Chapter 3

Average number of retrogradely labeled neurons in the NRA per section for levels 0 to 8

unilateral L6 injection unilateral S1 injection unilateral S2 injection (case 2249) (case 2261) (case 2265)

20 20 20 ipsi ipsi ipsi

contra contra contra 10 10 10 # NRA neurons/section # NRA neurons/section 0 # NRA neurons/section 0 0 012345678 012345678 012345678 level level level

unilateral L6 injection unilateral L6 injection bilateral L7 injection contralateral L3 hemisection ipsilateral L3 hemisection left L3 hemisection (case 2280) (case 2278) (case 2295)

20 20 20 ipsi ipsi left bilateral L7 injection contra left L3 hemisectioncontra right 10 10 10 # NRA neurons/section # NRA neurons/section 0 0 # NRA neurons/section 0 012345678 012345678 012345678 unilateral L6 injection unilateral levelL6 injection level level contralateral L3 hemisction ipsilateral L3 hemisction

Figure 1 Histograms of the number of retrogradely labeled neurons in the NRA in all six cases.

(Table 1; Figs. 1 and 2 middle). Apparently, these 7% In contrast to cases 2249 and 2280, no retrogradely descended to the lumbosacral cord via trajectory A or E labeled neurons were found ventromedial to the of Figure 3, and more than 93% of all NRA neurons ipsilateral NRA (see Figs. 1 and 2 right), indicating that project to the lumbosacral segments via the contralateral this group of neurons projects to the ipsilateral cord (trajectories C or F). It also means that from all lumbosacral cord by means of ipsilaterally descending NRA neurons projecting to the ipsilateral lumbosacral axons (however, see case 2295 in table I). segments (26-30% of the total population) the great majority send their fibers by way of the contralateral Are recrossing fibers collaterals or not? cord. These fibers cross the midline twice (trajectories The question arises whether or not the recrossing fibers D or F), first at the level of the NRA itself ventral to the are collaterals of fibers crossing at the level of the NRA central canal and for the second time, since the and terminating contralaterally in the lumbosacral cord hemisection is in L3, at the level caudal to L3, probably (trajectory D or F in Fig. 3). The same question applies at the level of termination (see also the anterograde to the ipsilaterally descending NRA-lumbosacral experiments). pathways (trajectory B or E). Theoretically, a double sided injection combined with a hemisection would Ipsilateral hemisection (case 2278) After a hemisection solve this problem, provided that 1) in all cases the same ipsilateral to the side of the injection only those NRA number of NRA lumbosacral neurons exists, and 2) neurons will be labeled, which send their fibers via the WGA-HRP does not label fibers of passage. In that case, contralateral spinal cord, but which axons terminate assuming that no collateralization exists, a bilateral ipsilaterally, i.e. cross the midline twice (trajectories D injection would result in a number of labeled NRA or F in Fig. 3). Furthermore, NRA neurons will be neurons ipsilateral to the hemisection which is 1 1/3 labeled, which send their fibers through the ipsilateral times higher than the number of ipsilateral NRA neurons spinal cord, but which terminate contralaterally, i.e. cross after a unilateral hemisection (trajectories C plus D). once at the level of termination (trajectories B or E). Suppose all recrossing NRA-lumbosacral fibers are These latter neurons will be found in the NRA collaterals, no difference would be found (trajectory F). contralateral to the side of the injection. In case 2278, In case 2295 such a bilateral injection was made. The with an ipsilateral hemisection 119 labeled NRA neurons result was that not 1 1/3, but 1 1/6 times the number of were found on the ipsilateral side and 9 on the contralaterally projecting neurons was present in the contralateral side (see Table 1). Apparently, nearly one ipsilateral NRA (see Table I). This would mean that 50% third of the total number of NRA lumbosacral neurons of the contralaterally descending NRA-lumbosacral have fibers, which cross twice, while approx-imately fibers descend via trajectory F, while the other 50% 3% of the total population has fibers which descend descend via trajectories C or D. However, it must be ipsilaterally, and cross at the level of termination. emphasized that such a conclusion should be taken with

84 NRA projections to lumbosacral motoneurons: non-estrous females extreme caution. Carefull analysis of the ipsilateral NRA itself clearly In conclusion, the retrograde tracing results indicate that: shows that not all neurons within the confines of the 1) the bulk (70-74%) of the NRA-lumbosacral neurons nucleus are labeled (Fig. 4). This finding suggests that have axons descending in the contralateral cord to the NRA contains many neurons not projecting to the terminate in the contralateral lumbosacral gray matter; lumbosacral cord, and which therefore are involved in 2) 26-30% of the NRA-lumbosacral neurons project to other activities than the labeled neurons (see also the ipsilateral spinal gray, most of which descend by Holstege 1989, p. 248). way of the contralateral cord. The fibers of these neurons cross the midline twice; and 3) the impression is gained Reticulospinal pathway that part of the recrossing fibers are collaterals from The retrograde tracing results indicate that the tegmental contralaterally terminating axons and part of them are neurons medial to the NRA project mainly ipsilaterally not. to the lumbosacral gray matter. However, according to case 2295, a few (5%) of these neurons project via the Only part of the entire NRA population projects contralateral spinal cord. In respect to the cases without to the lumbosacral cord hemisections, in which considerable numbers of labeled In case 2295 with a bilateral L7 injection, according to tegmental neurons were found at the contralateral side, Table 1, 384 retrogradely labeled neurons were counted it should be recalled that in these cases the injection site in the NRA on the side ipsilateral to the hemisection. extended into the contralateral lamina X. This might have

2249 2280 2278 ECU ECU ECU G CU G G CU CU S S S Vspin. Vspin. XII XII caud XII Vspin. caud caud LTF LTF LTF LRN LRN IO LRN IO IO Obex P P Obex P

CU CU G G G Vspin. Vspin. Vspin. caud caud caud LTF NRA LTF NRA LTF NRA

LRN LRN LRN P -2-2 mm P

Vspin. G Vspin. G Vspin. G caud caud caud NRA NRA NRA P P P -4-4 mm

Vspin. G Vspin. caud G Vspin. caud G caud NRA NRA P P NRA P -6-6 mm

-8-8 mm

L3 L3 L3 L6 L6 L6

Figure 2 Schematic representation of the WGA-HRP labeled neurons in the caudal medulla in cases 2249, 2280, and 2278. Each drawing represents six sections around the levels 0, -2, -4, -6, and -8 mm caudal to the obex.

85 Chapter 3

NRA neurons NRA neurons lumbosacral levels. In case 1019 the ipsilaterally descending fibers traveled throughout the length of the spinal cord and terminated in laminae V to VIII and in lamina X of the ipsilateral gray matter. However, in case A B C D E F 1183, with the more lateral injection, in the lumbosacral gray matter labeled fibers were sparse.

Injections involving the caudal NRA Location of injections In 4 cases (1471, 1571, 1683 and 1684; see Fig. 5) injections of 3H-leucine were made in caudal portions of the NRA and adjacent tegmentum. In a fifth case (1472) the injection site avoided the NRA, but occupied the tegmentum between NRA and solitary nucleus. The injection site extended laterally into the Figure 3 Schematic representation of the different pathways caudal spinal trigeminal nucleus and dorsally into the along which the NRA neurons may send their fibers to the dorsal column nuclei. The two most caudally located lumbosacral cord. injections were made in cases 1471 and 1571. In case 1471 the injection was relatively large and involved the lateral tegmental field, including the caudal NRA and resulted in retrograde labeling of the reticulospinal the most medial part of the caudal spinal trigeminal neurons because the anterograde tracing observations nucleus. The smallest injection was made in case 1571 (see further) indicate that the reticulospinal fibers and was present in the white matter of the lateral terminate in this lamina. funiculus, but extended medially into the lateral half of the caudal NRA. Similar injection sites were present in cases 1683 and 1684, but these injections extended Anterograde tracing study further medially and further rostrally in the tegmentum tritiated leucine experiments than the injection in case 1571. Injections involving the rostral NRA Contralateral projection The contralateral pathway Location of injections In two cases 1183 and 1019 the consisted of fibers that crossed the midline at the level injection sites involved the rostral NRA and adjoining of the injection or just rostral to it and descended in the lateral reticular formation (see Fig. 5). The injection in contralateral ventral funiculus and more caudally in the case 1019 extended much further medially in the ventrolateral funiculus as well. The fibers of the tegmentum than in case 1183. Both injections were contralateral fiber bundle descended throughout the present at the level of the obex, but in case 1183 the length of the spinal cord, except in case 1472, in which injection site extended further rostrally and in case 1019 they were not found beyond upper lumbar levels. further caudally. At upper lumbar levels, similar to the rostral NRA injected cases, many labeled fibers were distributed to Contralateral projection A specific contingent of abdominal muscle motoneurons bilaterally (see Fig. 6). labeled fibers, after crossing the midline at the level of the obex, descended in the contralateral ventral funiculus throughout the length of the spinal cord. In the lumbosacral cord from these fibers many were distributed to mainly the lateral part of the L1-L3 motoneuronal cell groups, containing abdominal muscle motoneurons (Miller, 1987; Holstege et al., 1987). These projections were bilateral, with a slight contralateral preponderance. In case 1183, but not in case 1019, at the levels L6 and L7 a very small number of labeled fibers were distributed to mainly the ventral part of the contralateral ventral horn. Virtually no labeled fibers were found in the Onuf’s nucleus in both cases. In conclusion the rostral NRA does not, or to only a very limited extent, project to the L4-S3 ventral horn. Figure 4 Brightfield photomicroghraph of a Nissl-stained Ipsilateral projection In both cases many labeled fibers section, showing the NRA ipsilateral to the hemisection in descended in the ipsilateral dorsolateral funiculus, but case 2295 (bilateral L7 injection). Note the unlabeled neurons in case 1183 they were very sparsely present at within the confines of the NRA. Bar represents 200 µm.

86 NRA projections to lumbosacral motoneurons: non-estrous females

S G ECU Xd CU ECU Vsp XII CU G S Vspin 1472 1183 caudCU 1183 CU Vspin G 1684 Vspin XII 1472 caud LRN 1472 NRA caud G 1019 1683 1683 IO NRA 1471 1471 LRN P 1571 P IO 1019 NRA P 1684

Figure 5 Tritiated leucine injection sites in the NRA and adjoining caudal medullary lateral tegmentum.

At these levels, many labeled fibers were observed to ipsilateral one in the dorsolateral funiculus, distributing cross in the ventral commissure, indicating that the fibers to the intermediate zone and lamina I, and a projection to the ipsilateral abdominal muscle contralateral one in the ventral and ventrolateral motoneuronal nuclei is at least partly derived from the funiculus, distributing fibers to motoneurons innervating contralaterally descending pathway. Yet, in cases 1471 abdominal musculature. In all cases in which the caudal and 1472, but not in case 1571, a few labeled fibers NRA was injected, strong projections to motoneuronal descended in the ipsilateral ventral funiculus and may cell groups at levels L5 to S1 and to the nucleus of Onuf have contributed some fibers to the ipsilateral abdominal were found. All motoneuronal projections were bilateral, motoneuronal cell groups. Further caudally, in all cases, but in respect of the L5-S1 motoneuronal projections except 1472, a clear distribution was observed to distinct they showed a strong contralateral preponderance. parts of the motoneuronal cell groups at the levels L5 to S1 and to the nucleus of Onuf at the level S1-S2 (see Fig. 6). These motoneuronal projections were bilateral WGA-HRP experiments but, in contrast to the projections to the abdominal wall Location of injections and hemisections motor nuclei, they had a very strong contralateral In 8 cases (cases 2237, 2251, 2256, 2258, 2267, 2271, preponderance. 2286, and 2290) multiple injections of WGA-HRP were made in the caudal medulla oblongata. All injections Ipsilateral projection In all 5 cases an ipsilateral bundle involved the NRA as well as the adjoining lateral descended in the dorsolateral funiculus. In cases 1571 tegmental field and the lateral funiculus (see Fig. 7). In and 1683, with injections not extending into the lateral two of the 8 cases hemisections were made in the second tegmental field, this bundle gradually disappeared at low cervical segment, ipsilaterally in case 2286 and cervical and upper lumbar levels, respectively. In cases contralaterally in case 2290. The hemisections were 1471, 1472 and 1684 the ipsilateral bundle continued complete and did not extend across the midline. throughout the length of the spinal cord and labeled As has been shown in the retrograde tracing study more fibers were distributed to all parts of the intermediate than 90% of the NRA neurons sent their fibers through zone including lamina X (see Fig. 6). In the upper lumbar the contralateral spinal cord, while a group of neurons cord a few labeled fibers were distributed to the located in the tegmentum medial to the NRA projected intermediolateral cell column, and in the sacral cord a almost exclusively ipsilaterally. In order to distinghuish very few labeled fibers were found in the sacral between these two projection systems, the descending parasympathetic cell group . projections will be described on the basis of the findings In cases 1471, 1472 and 1684 labeled fibers were of cases 2286 for the contralateral pathway, and 2290 distributed to lamina I throughout the lumbosacral cord for the ipsilateral pathway. It should be emphasized that, (see also Fig. 4 of Holstege, 1988). These projections depending on the site of the injection, in the other 6 were found on both sides, but showed an ipsilateral WGA-HRP cases (all without hemisections) both the predominance. Lamina I projections do not seem to ipsi- and contralateral pathways were found. originate in the NRA, because the NRA is not injected in case 1472 and lamina I projections were absent or Contralateral pathway in the segments L1 to S3 sparse in the other 3 cases with injections involving the Location of descending fibers In case 2286, with the NRA. ipsilateral C2 hemisection, descending labeled fibers Summarizing: in all 5 cases with injections in the caudal were found in the dorsolateral, ventrolateral, and ventral medulla approximately the same projection pattern was funiculi. From L1 to S3 their number gradually observed i.e. two different descending pathways, an decreased. From rostral L4 to caudal L5, the labeled

87 Chapter 3

Figure 6 Darkfield photomicrographs of the L3-S1 spinal cord in cases 1471 with an injection with 3H-leucine in the NRA and medially adjacent tegmentum.

fibers in the ventrolateral funiculus gradually shifted into found to the lateral part of the ventral horn, which a more peripheral position. Only a very limited number projection was equally dense at both sides of the cord of retrogradely labeled neurons was found in the (arrow a in Fig. 8). This distribution area contains lumbosacral cord (see Fig. 8), which indicates that the motoneurons innervating the external and internal bulk of the labeled fibers in the white matter represent oblique and to a limited extent transverse muscles of descending fibers. the abdominal wall (Holstege et al., 1987; Miller, 1987). These abdominal wall motoneurons are not present in Non-motoneuronal termination sites Throughout the the caudal part of L3. In conclusion, in the rostral L3 length of the L3-S3 spinal cord, labeled fibers were segment a dense, bilateral NRA projection to moto- distributed to laminae V to VII and to lamina X neurons innervating abdominal wall muscles was bilaterally, but with a contralateral preponderance. These present. In the same segment labeled fibers were also projections were denser at sacral than at lumbar levels. present in the medial portion of the ventral horn, but it The fibers terminating ipsilaterally had recrossed in the was difficult to establish whether this area represented ventral gray commissure. At all lumbosacral levels, a termination site due to the many crossing fibers (Fig. labeled fibers terminated in areas at the transition 8 arrow b). A similar, bilateral distribution pattern to between white matter and lamina IX. In addition to this the medial portion of the ventral horn was also found in general distribution pattern, specific projections were the L4-L6 segments, but was absent in segment L7 (Fig. found to certain motoneuronal cell groups in lamina IX. 8 arrow b; Fig. 9 L5). These projections might terminate on motoneurons innervating the multifidus muscle Motoneuronal termination sites L3 In the rostral, but (VanderHorst and Holstege, Chapter 2; Gilette et al., not in the caudal half of L3, very dense projections were 1993). In segment L3 a few labeled fibers were found

88 NRA projections to lumbosacral motoneurons: non-estrous females in the area of the intermediolateral cell column, semitendinosus and m. biceps femoris, respectively containing autonomic (sympathetic) motoneurons. (Romanes, 1951; VanderHorst and Holstege, Chapter 2). Similar to the rostral L7, no labeled fibers were Motoneuronal termination sites L4 and L5 In rostral distributed to the dorsal portion of the ventral horn, L4 no specific projection to motoneurons was found, containing moto-neurons innervating the muscles but in caudal L4-rostral L5 a specific projection to a controlling the ankle joint (Romanes, 1951; Burke et motoneuronal cell group at the ventrolateral border of al., 1977; Horcholle-Bossavit et al., 1988; VanderHorst the gray matter was observed (Fig. 8 arrow c; see Fig. and Holstege, Chapter 2). 10 for detail). This area contains motoneurons inner- vating the iliopsoas muscle (VanderHorst and Holstege, Motoneuronal termination sites S1 The L7 projection Chapter 2; see also Fig. 10). In addition, in rostral L5 to the motoneurons innervating hamstring muscles (m. labeled fibers were present in the centre of the ventral semitendinosus and m. biceps femoris) continues into horn (Fig. 8 arrow d), in which area are located the rostral part of S1 (Fig. 8 arrow g). Caudally in S1 motoneurons innervating adductor muscles of the labeled fibers were present in the nucleus of Onuf (Fig. hindlimb such as m. pectineus, m. adductor longus, m. 8 arrow h; Fig. 11 S1), which contains motoneurons adductor femoris brevis and m. gracilis (Romanes 1951; innervating pelvic floor muscles. The many labeled VanderHorst and Holstege, Chapter 2). It must be fibers terminating in the area dorsomedial to the nucleus emphasized that no labeled fibers terminated in the of Onuf might represent terminations on dendrites of motoneuronal cellgroup located between the centrally Onuf motoneurons (Fig. 8 arrow i; Holstege and Tan, located adductor group and the ventrolaterally located 1987; Beattie et al., 1990). The projection to Onuf iliopsoas group. This group of motoneurons devoid of motoneurons was remarkably bilateral in contrast to the labeled fibers innervates the m. sartorius muscle projections to somatic motoneurons in more rostral (Romanes 1951; VanderHorst and Holstege, Chapter 2). portions of the lumbosacral enlargement. The projection In L5 another specific projection was found to the indicated by arrow j in Figure 8 might represent ventromedial corner of the ventral horn (Fig. 8 arrow e; terminations on motoneurons innervating the most distal Fig. 11 L5). It has not yet been clarified which of the intrinsic hindpaw muscles (VanderHorst en motoneurons occupy this area. Holstege, Chapter 2). In caudal S1, another projection was located in the ventromedial part of lamina IX. This Motoneuronal termination sites L6 In rostral L6, except projection is difficult to see in Figure 8 (arrow k), but for the projection to an area in the medial portion of the can be distinghuished in Figure 11 S1. This area contains ventral horn (Fig. 8 arrow b), only a limited number of motoneurons innervating the caudal part of the medial labeled axons were found. At this level the ventral horn longissimus muscle which inserts on the base of the tail contains many motoneurons innervating the m. (VanderHorst and Holstege, Chapter 2). quadriceps (Romanes, 1951; VanderHorst and Holstege, Chapter 2). These motoneurons apparently do not Motoneuronal termination sites S2 and S3 In S2 and to receive a substantial projection from the NRA. In caudal a limited extent in the rostral S3 another projection to L6, a strong projection was present to the centre of the tail muscle motoneurons was found (Fig. 8 arrows lamina IX (Fig. 8 arrow f; Fig. 9 L6). This area l). Motoneurons innervating the proximal tail muscles corresponds with the rostral portion of column 3' of are located within two regions of the ventral horn: the Romanes (1951, 1954), which, at this level, contains ventromedial nucleus located in the ventral part of the motoneurons innervating the m. semimembranosus ventral horn and the nucleus commissuralis in the medial (Romanes, 1951; VanderHorst and Holstege, Chapter portion of the ventral horn (Rexed, 1954; Ritz at al., 2). 1992). The intermediolateral cellgroup, containing autonomic Motoneuronal termination sites L7 The dense accumu- (parasympathetic) motoneurons innervating bladder and lation of labeled fibers observed in caudal L6 extended sexual organs also receives a moderate bilateral into the most rostral pole of L7. In the remainder of projection (Fig. 8 arrow m). rostral L7 such a dense accumulation was not found, but a rather diffuse distribution was present in the ventral Ipsilateral pathway in the segments L3 to S3 half of the ventral horn. It was not possible to draw any Location of descending fibers In case 2290, with a conclusions about which motoneurons or motoneuronal contralateral hemisection in C2, at lumbosacral levels dendrites in this area are targeted. The dorsal half of the the bulk of the labeled fibers was present in the ventral horn received only a limited number of labeled dorsolateral funiculus and only very few were located fibers at that level. In caudal L7 another accumulation in the ventrolateral, lateral and ventral funiculi. The of labeled fibers was found in the central part of the number of labeled axons gradually decreased from L1 ventral horn (Fig. 8 arrow g). This area corresponds with to S3. In contrast to case 2286 with an ipsilateral columns 3’ and 3’’ of Romanes and contains motoneu- hemisection, in case 2290 a considerable number of rons innervating hamstring muscles, at this level the m. retrogradely labeled neurons was present, mainly

89 Chapter 3

level: 0 2 4 6 8

2237

2251

2256

2258

2267

2271

2286 C2

2290 C2

Figure 7 WGA-HRP injection sites in the NRA and adjoining caudal medullary lateral tegmentum. Note that the injections in cases 2286 and 2290 are combined with an ipsilateral and contralateral hemisection, respectively.

Figure 8 ^ Schematic drawings of the labeled fibers after injection of WGA-HRP in the NRA combined with an ipsilateral hemisection (2286). Each drawing represents the labeling of 6 consecutive collected (1:4) sections. Note the accumulation of labeled fibers in distinct portions of the ventral horn motoneuronal cell groups. For reference of the arrows see text.

90 NRA projections to lumbosacral motoneurons: non-estrous females

91 Chapter 3

L6 L5 Figure 9 On the left a combined polarized light and darkfield photomicrograph of the ventral horn in the caudal L6 in case 2286 with an injection in the NRA and adjoining tegmentum combined with an ipsilateral C2 hemisection. Note the specific projection to the centre of the ventral horn (see also Fig. 8 arrow f). On the right combined polarized light and darkfield photomicrograph of the medial part of the ventral horn in the caudal L5 in case 2286. Note the specific projection to the dorsomedial ventral horn (see also Fig. 8 arrow b). Bar represents 200 µm.

contralaterally in laminae I, and V to VIII in the L7 to ventromedial to the NRA. Only a limited number of S3 segments. In the L3-L6 segments a much smaller neurons in the NRA are involved in this projection. number of labeled neurons was observed. Only very few labeled neurons were found in the lateral portion of laminae I and V. Due to the many retrogradely labeled DISCUSSION neurons in this case, some of the labeled axons found in the white matter might be derived from retrogradely labeled neurons. The present paper is the first to precisely describe nucleus retroambiguus (NRA) projections to lamina IX Termination sites The bulk of the ipsilaterally in the lumbar enlargement. The projection does not involve all lumbar motoneuronal cell groups, but a descending axons terminated in laminae V to VIII, and distinct set innervating a certain group of hindlimb and in lamina X throughout the length of the lumbosacral back muscles. In earlier studies, the NRA has been cord. Only very few labeled fibers were distributed to described to project to other distinct motoneuronal cell laminae I and IX. The sympathetic intermediolateral cell groups. Table 2 gives an overview of all the NRA- column in segments L1 to L3 (Fig. 13) and to a limited motoneuronal projections and their possible function. extent also the sacral parasympathetic nucleus in S2 and S3 received labeled afferents. In general, all projections had a very strong ipsilateral predominance and were NRA projections to motoneurons other than denser at sacral than at lumbar levels. In the contralateral gray matter, the few anterogradely labeled fibers could in the lumbosacral cord not be distinghuished from dendrites and axons of numerous retrogradely labeled neurons. According to The dorsal group of the nucleus ambiguus, containing the findings of the retrograde study, this ipsilateral motoneurons innervating soft palate and pharynx distibution is mainly derived from neurons located muscles, is known to receive strong projections from

92 NRA projections to lumbosacral motoneurons: non-estrous females

L4 L4 Figure 10 On the left a combined polarized light and darkfield photomicrograph of the ventral horn in the caudal L4 in case 2286 with a WGA-HRP injection in the NRA and adjoining tegmentum, combined with an ipsilateral C2 hemisection. Note the anterogradely labeled fibers in the lateral part of the ventral horn. On the right a combined polarized light and darkfield photomicrograph of the ventral horn in the caudal L4 , showing retrogradely labeled motoneurons after an injection of plain HRP in the m. iliopsoas. Bar represents 200 µm. the NRA, mainly contralaterally (Holstege, 1989). which are involved in expiration as well (Duron, 1973). Physiologically, NRA-laryngeal motoneuron projections The second group consists of the transverse, internal have been demonstrated by Zhang et al. (1992). These oblique, and external oblique abdominal wall motor NRA-pharynx/larynx motoneuron projections play a nuclei. Their motoneurons are located laterally in the role in expiration (Feldman, 1986) and in vocalization T5-L3 ventral horn (Holstege et al., 1987; Miller, 1987) (Holstege, 1989; Zhang et al., 1992). and receive bilateral projections from the NRA, with a Large portions of the NRA have been shown to project contralateral predominance (Holstege and Kuypers, to the phrenic nucleus (Holstege and Kuypers, 1982; 1982; Feldman et al., 1985; Holstege, 1989). The Feldman et al., 1985), but the most caudal portion of abdominal wall muscles have a function in expiration the NRA is not involved in this projection (Holstege, and expiration related activities (Campbell and Green, 1989). The NRA has also been shown to project to 1953; McCarthy and Borison, 1974). pectoral muscle motoneurons in the C7 segment (Holstege and Kuypers, 1982). These muscles serve as adductors of the shoulder (Crouch, 1969). Another NRA NRA projections to motoneurons in the projection is to the cutaneus trunci muscle motoneurons lumbosacral cord in the C8-T1 segments (Holstege and Blok, 1989). Throughout the entire length of the thoracic cord, the NRA projects to motoneuronal cell groups in the ventral hindlimb motor nuclei horn (Holstege and Kuypers, 1982; Feldman et al., iliopsoas muscles At the level L4-L5, the NRA projects 1985). The first group of motoneurons at these levels to the motoneuronal cell group innervating the m. innervates intercostal muscles, to which belong the iliopsoas. This muscle consists of two parts, the m. intercartilaginous, triangularis sterni, internal and iliacus and the m. psoas. The iliacus muscle originates external intercostal muscles. The external and internal from the ventral border of the ilium, and the psoas intercostal muscles, have a role in postural function, with muscle from the 5 most caudal vertebrae. Both parts the exception of the caudal internal intercostal muscles, insert to the the lesser trochanter of the femur. Their

93 Chapter 3

L5 S1 Figure 11 On the left a combined polarized light and darkfield photomicrograph of the ventral part of the ventral horn in L5 in case 2286 (WGA-HRP injection in the NRA and adjoining tegmentum combined with an ipsilateral C2 hemisection). Note the specific projection to the ventromedial nucleus (see also Fig. 8 arrow e). On the right a combined polarized light and darkfield photomicrograph of the S1 ventral horn in case 2286. Note in S1 the anterogade labeling in the nucleus of Onuf (see also Fig. 8 arrows h and k). Bar represents 200 µm.

Figure 12 Schematic drawings of the labeled fibers after injection of WGA-HRP in the NRA and adjoining tegmentum combined with a contralateral hemisection (2290). Each drawing represents the labeling of 3 consecutive collected (1:4) sections.

94 NRA projections to lumbosacral motoneurons: non-estrous females function is fixation of the spine and flexion of the hip sacral cord. In a later study, the nucleus of Onuf was (Crouch, 1969). found to be the main target of these NRA fibers (Holstege and Tan, 1987). In the cat, the nucleus of Onuf adductor muscles Motoneurons innervating different innervates pelvic floor muscles including the external adductor muscles of the thigh are located in column 3 anal and urethral sphincters (Sato et al., 1978). Pelvic of Romanes (1951), which extends from rostral L5 to floor muscles play a role in various activities such as caudal L6. The NRA projection to column 3 did not abdominal straining, mating behavior, micturition and involve its entire rostrocaudal extent, but was present defecation. only in the L5 and most rostral L6 segments, which matches the rostrocaudal distribution of motoneurons ventromedial nucleus of the lumbosacral cord innervating the m. pectineus, adductor longus and brevis, The ventromedial nucleus at lower lumbar levels and gracilis (Romanes, 1951; VanderHorst and Holstege, contains motoneurons innervating the lateral (Holstege Chapter 2). The mm. pectineus and adductor longus et al., 1987) and medial components of the longissimus originate from the os pubis, insert on the femur and dorsi, which serve as extensors of the back (Brink et al., adduct the thigh. The m. adductor brevis has its origo 1979, in the rat). The most caudal portion of the medial on the symphysis, inserts on the femur and functions as longissimus deviates the tail when activated unilaterally. adductor and extensor of the hip. The m. gracilis, with Motoneurons of this part of the muscle are located in its origo on the symphysis and insertion on the proximal the ventromedial nucleus at the level of S1. At the level tibia, is not only an adductor of the thigh, but can also of S2, the ventromedial nucleus contains motoneurons act as an extensor of the hip and flexor of the knee innervating intrinsic muscles of the proximal tail (Ritz (Crouch, 1969; Zajac, 1985; Pratt et al., 1991). et al., 1992). NRA projections have been observed to the ventromedial nucleus at the levels L5-L6 and S1- hamstring muscles Motoneuronal cell groups inner- S2. vating hamstring muscles (semimembranosus, semiten- dinosus and biceps femoris) also receive NRA afferents. The m. semimembranosus motoneurons are located in What is the function of the NRA-lumbo- column 3' of Romanes (1951, 1954) at the level caudal sacral motoneuronal projection? L6-rostral L7. The semimembranosus can be devided into two functionally different parts, both originating from the ischium. The anterior part inserts on the distal In only a few electrophysiological studies, the presence femur and acts as an extensor of the hip. The posterior of NRA axons caudal to the level of L3 was determined. part inserts on the proximal tibia and functions as an Miller et al. (1985) reported that of all investigated extensor of the hip and as a flexor of the knee (Crouch, expiratory NRA neurons projecting to the L1 segment, 1969; Zajac, 1985; Pratt et al., 1991). one third could also be antidromically activated from The motoneurons of the m. semitendinosus are located the white matter in L4-L5. Most expiratory neurons in column 3' of Romanes (1951,1954) at the level caudal projecting to L4-L5 were located caudal to the expiratory L7/rostral S1 and caudal to those innervating the m. neurons which projected to L1. Another physiological semimembranosus. Similar to the posterior part of the study (Sasaki et al., 1991, 1994) showed that expiratory semimembranosus and this muscle also has its origin neurons in the region of the NRA have axonal branches on the ischium and inserts on the proximal tibia. It is a terminating in the lumbosacral gray matter, but the strong flexor of the knee and helps to extend the hip authors did not reveal the exact targets. (Crouch, 1969; Zajac, 1985; Pratt et al., 1991). The motoneurons innervating the m. biceps femoris are Are NRA motoneuronal projections mono- located at the same level, but laterally to those of the synaptic? semitendinosus (column 3'’ of Romanes, 1951). The In the present paper it is demonstrated that NRA neurons biceps femoris muscle can be devided into an anterior project to motoneuronal cell groups innervating specific and a posterior part. Both parts originate from the hindlimb muscles such as the iliopsoas, adductor, and ischium, and the anterior part inserts on the distal femur hamstring muscles, as well as axial muscles such as the and the posterior part on the proximal tibia. Both muscle caudal medial longissimus and proximal tail muscles. compartments function as extensors of the hip, while The location of many labeled NRA axons within the the posterior biceps femoris also serves as a flexor of confines of distinct motoneuronal cell groups indicates the knee (Crouch, 1969; Zajac, 1985; Pratt et al., 1991; that the NRA might form direct connections with Chanaud et al., 1991a,b). lumbosacral motoneurons. Preliminary results of an electronmicroscopic double-labeling study (Vander- nucleus of Onuf Horst and Holstege, 1994) show that NRA terminals Holstege and Kuypers (1982) were the first to form asymmetric, presumably excitatory contacts with demonstrate that the NRA projects to the level of the lumbosacral motoneurons.

95 Chapter 3

parturition, jumping or lordosis. The latter two activities form examples of specific emotional behavior which can be elicited by stimulation in the PAG. The various behavioral patterns will be discussed below.

Synchronization of respiratory and locomotor rhythms It has been shown in man and other mammals that breathing and locomotion rhythms become synchronized during exercise (e.g., Bramble and Carrier, 1983). In decerebrated, vagotomized and immobilized cats, L3 respiration and locomotion are still coupled (Kawahara 1989, 1990). These studies report that diaphragmatic Figure 13 Combined polarized light and darkfield photo- and gastrocnemius muscle activities are synchronized, micrograph of the ipsilateral intermediate zone and dorsal part depending on the strength of the locomotor pattern and of the ventral horn in the rostral L3 in case 2290 (WGA-HRP end-tidal pCO2. The central nervous system mechanisms injection in the NRA and adjoining tegmentum, combined responsible for this coupling are unknown. The NRA with a contralateral C2 hemisection). Note the labeled fibers might play a role in the synchronization of locomotion in the area of the intermediolateral cellcolumn (arrow). Bar and respiratory rhythms, because it projects to both represents 200 µm. respiratory and hindlimb muscle motoneurons. However, the motoneuronal cell group innervating the m. gastrocnemius does not seem to receive afferent NRA motoneuronal projections belong to the projections from the NRA. basic motor system Defecation and parturition In general, motoneurons receive their main input from interneurons in the spinal intermediate zone. Usually Since the NRA is known to be involved in forced these projections are called propriospinal pathways. expiration, the question arises whether NRA neurons Supraspinal structures as well as afferent input from the take part in other straining related activites, such as periphery make use of these interneurons to influence defecation and parturition. During defecation, distension of the rectal wall initiates weak peristaltic waves in the the motoneurons. The propriospinal projections take part descending colon, sigmoid and rectum as well as in the basic motor system in the concept of Holstege relaxation of the internal anal sphincter. This reflex is (Holstege, 1991, 1994 ). called the intrinsic defecation reflex and is mediated by Specific, direct projections to distinct motoneuronal cell the myenteric plexus. Stimulation of afferents from the groups, bypassing the spinal interneurons, are rare; in rectum to the sacral cord initiates a parasympathetic the cat examples are Ia afferents from muscle spindles defecation reflex, which intensifies the intrinsic reflex and a very few rubrospinal fibers (Holstege, 1987). (Guyton, 1986). However, for actual defecation the In the framework of the NRA motoneuronal projections, tonically activated external anal sphincter, innervated the NRA neurons might be considered as interneurons, by Onuf motoneurons, must be relaxed (for review see similar to the ones in the spinal intermediate zone. The Dubrovsky and Filipini, 1990). The exact mechanism interneurons in the NRA do not receive afferents from by which the external anal sphincter motoneurons are the usual sources such as the peripheral Ia afferents and inhibited is not yet known. the rubro- and corticospinal tracts. Their afferents come Defecation also has an abdominal wall straining from different sources, such as the PAG. This might be component. For example, Kufuda and Fukai (1986a) the reason that they are located at a different site. In this found simultaneous activity of the abdominal wall concept the NRA motoneuronal pathway belongs to the muscles (rectus and obliquus externus), internal basic motor system, which consists of premotor intercostal muscles and the diaphragm after distension interneurons (Holstege, 1994). The same is true for the of the vagina or rectum in decerebrate dogs. The question M- and L-region neurons in the dorsolateral pons, which is whether this straining component is an integral part project to sacral parasympathetic and Onuf moto- of the defecation reflex. In patients and animals with neurons, respectively, and coordinate micturition transection of the spinal cord rostral to the sacral cord, (Holstege et al., 1986). defecation is still possible, but relaxation of the external The fact that the NRA-motoneuronal pathway is so anal sphincter and the abdominal wall straining different from other pathways to motoneurons suggests component are lacking. Possibly, the NRA-abdominal that the former projection system is involved a specific muscle motoneuronal pathway is involved in the motor activity. Candidates are synchronization of straining component (Fukuda and Fukai, 1986a,b), respiratory and locomotor rhythms or respiration although the rectus abdominis muscle, which is active independant behavioral patterns, as defecation, during straining, is reported not to receive NRA afferents

96 NRA projections to lumbosacral motoneurons: non-estrous females

TableTable 2. NRA2. projections NRA projectionsto motoneuronal cellgroups to motoneuronal cell groups motoneuronal cell group location in CNS muscle innervated function references nucleus ambiguus lateral tegmental field pharynx, soft palate, larynx expiration and vocalization Holstege, 1989; Zhang et al., 1992 phrenic nucleus C4-C6 diaphragm inspiration, vomiting, coughing and Holstege and Kuypers, 1982 abdominal straining Feldman et al., 1985 pectoralis C7 pectoralis adduction of the shoulder Holstege and Kuypers, 1982 Sterling and Kuypers, 1967; Hörner and Kümmel, 1993 cutanei trunci motor nucleus C8-T1 cutaneus trunci skin protection (shivering) Holstege and Blok, 1989 lateromedial nucleus (Rexed 1954) T1-T7 intercartilaginous inspiration Larnicol, 1982; Duron, 1973; Holstege and Kuypers, 1982; Feldman et al., 1985 lateromedial nucleus (Rexed 1954) T1-T7 triangularis sterni expiration idem latero- and ventromedial nuclei (Rexed 1954) T1-T8/ T12* external intercostal inspiration and posture (caudal part) idem latero- and ventromedial nuclei (Rexed 1954) T1-T8/ T12* internal intercostal expiration and posture (caudal part) idem lateromedial nucleus (Rexed 1954) T3-L3 external oblique expiration, abdominal straining, Holstege and Kuypers, 1982; Feldman et al., 1985; T7-L3 internal oblique vomiting, coughing, eructation, Holstege et al., 1987; Miller, 1987/89 T5-L3 transverse abdominis defaecation and vocalization iliopsoas caudal L4-L5 iliopsoas fixation of the lower spine and present paper flexion of the hip column 3 (Romanes 1951) rostral L5-rostral L6 pectineus and adductor longus, adduction of the thigh present paper adductor femoris brevis, adduction of the thigh and extension of the hip gracilis adduction of the thigh, extension of the hip, and flexion of the knee column 3' (Romanes 1951) caudal L6-rostral L7 semimembranosus extension of the hip and flexion of the knee present paper column 3' (Romanes 1951) caudal L7-rostral S1 semitendinosus extension of the hip and flexion of the knee column 3'' (Romanes 1951) caudal L7-rostral S1 biceps femoris extension of the hip and flexion of the knee ventromedial nucleus (Rexed 1954) S1 caudal part medial longissimus extension and lateral deviation of the tail present paper nucleus of Onuf S1-S2 pelvic floor contraction of external anal and Holstege and Tan, 1987 urethral sphincters; Holstege and Kuypers, 1982 (retrograde) abdominal straining ventromedial nucleus (Rexed 1954) S2 proximal tail extension and lateral deviation of the tail present paper

* Larnicol did not investigate the location of intercostal motoneurons caudal to T8. Probably they are present in T9-T12 as well.

(Holstege, 1991). stabilize the posture. Moreover, the medial longissimus The NRA-lumbosacral motoneuronal projection might muscle, when unilaterally activated, produces the typical play a role in the specific posture that accompanies deviation of the tail, and pelvic floor muscles contract defecation (Kufuda and Fukai,1986a). An argument rhythmically. Quadriceps (extensor of the knee) and against such an involvement is that only part of the triceps surae muscles (plantar flexors of the ankle) are motoneurons innervating muscles the defecation-posture not involved in this behavior. In the rat it has been shown muscles receive NRA afferents. that mechanical stimulation of the cervix, but not the In the same way, the NRA might be involved in vagina, activates the iliopsoas muscles (Martinez-Gomez parturition, which also is accompanied by abdominal et al., 1992). In the cat, the m. semitendinosus is wall straining as well as a specific posture. rhythmically active during mechanical stimulation of the cervix, whereas triceps surae muscles showed Jumping sustained EMG activity during and after cervical Jumping is one of the components of survival behavior stimulation (Cueva-Rolon et al., 1993). which can be elicited by stimulation in the lateral part The PAG is essential for the control of lordosis (see of the subtentorial PAG (Bandler et al., 1991). During Ogawa et al., 1991 for review). Stimulation of the lateral jumping, the hamstrings, quadriceps, and plantar flexors and dorsal PAG facilitates lordosis (Sakuma and Pfaff, of the ankle are activated, but the iliopsoas muscle is 1979a), whereas lesions suppress it in estrogen primed not (Zajac, 1985; Loeb, 1993). The NRA-lumbosacral rats ( Sakuma and Pfaff, 79b). projection indeed includes projections to hamstring One of the most important features of lordosis is that it muscle motoneurons, but the motoneurons innervating is dependant on estrogens (see Pfaff and Schwartz- extensor muscles of the knee (quadriceps) and plantar Giblin,1988 for review). The PAG receives a strong flexors of the ankle (e.g. medial and lateral gastro- projection from estrogen containing cells in the cnemius and plantaris) do not receive NRA afferents. ventrolateral part of the ventromedial hypothalamus (VMH; see Pfaff and Schwartz-Giblin,1988 for review). Lordosis Lordosis can also be facilitated and suppressed by Lordosis is species specific, female receptive behavior stimulation and lesions in the VMH, respectively (for and has been studied extensively in rat, hamster, guinea the rat: Pfaff and Sakuma, 1979a,b; Mathews and pig and cat (see Pfaff and Schwartz-Giblin, 1988 for Edwards, 1977; for the hamster: Malsbury et al., 1977; review). In the cat the full receptive posture consists of and for the cat: Leedy and Hart, 1985). Lesions in the crouching (forelegs collapsed), lowering of the head, VMH do not abolish PAG-facilitated lordosis in estrogen perineal elevation, tail deviation, and treading, often in primed rats (Pfaff and Sakuma, 1979b). However, when combination with calling and vulval excretion (Michael, lesions are made in the PAG, VMH stimulation is not 1961). The NRA-lumbosacral projection involves a efficient for evoking lordosis (Pfaff and Sakuma, 1979a). combination of muscles which may be activated Apparently, the VMH is important for estrogen priming, specifically during lordosis behavior. Axial muscles but the PAG controls the final motor output of lordosis. extend the lower back, iliopsoas muscles fixate the lower The PAG also receives direct ascending projections from spine to the pelvis, hamstring muscles cause extension the spinal cord (Menetrey et al., 1982; Wiberg and of the hip and flexion of the knee, and adductor muscles Blomqvist, 1984; Yezierski, 1988; Mouton et al. in prep.)

97 Chapter 3

Limbic System Cueva-Rolon et al., 1993). The specific PAG-NRA- lumbosacral motoneuronal pathway presented in this Periaqueductal Gray study, involves both axial and specific hindlimb muscles (PAG) and therefore might form the final common pathway for lordosis behavior. Nucleus Retroambiguus (NRA) Epilogue In the present study, both retro- and anterograde WGA- Motoneurons HRP experiments were done in female cats, which innervating showed no signs of estrus behavior. The anterogradely labeled NRA fibers of six sections had to be condensed into one drawing in order to produce the pattern as presented in Figure 8. Since female cats only display Iliopsoas muscles Abdominal muscles Adductor muscles lordosis behavior in pro-estrus or estrus (Michael, 1961), Larynx muscles Hamstring muscles in a new series of experiments the NRA-lumbosacral Pharynx muscles Pelvic floor muscles projections was compared in estrus versus anestrus cats. Proximal tail muscles Preliminary results showed marked differences in the density of the NRA-lumbosacral projection in estrus and anestrus cats; in the estrus cat NRA fibers were heavily labeled and could be easily observed in a single section VOCALIZATION LORDOSIS (VanderHorst and Holstege, in preparation). These results present further evidence for the concept that the Figure 14 Schematic representation of the concept for the final common pathways for vocalization and lordosis. PAG-NRA-lumbosacral pathway is the final common pathway for lordosis. The NRA in this concept would serve as a relay within the final common pathway from and is known to respond to lordosis relevant somato- the PAG to motoneuronal cell groups for vocalization sensory stimulation (Pfaff and Schwartz-Giblin, 1988 as well as lordosis (Fig. 14). for review). Lordosis can still be elicited by somato- sensory and vaginocervical stimulation in ovariec- tomized rats after precollicular decerebration (Rose and Flynn, 1993). In freely moving animals, lordosis behavior is initiated easily by applying tactile stimuli to the skin of the flanks, posterior rump, tailbase, and perineum (e.g., for the rat: Kow et al.,1979; and for the cat: Michael, 1961). It is not known how lordosis behavior is mediated from the mesencephalic region to motoneurons that innervate the muscles involved in this behavior. Direct projections from the PAG to the spinal cord have been shown in the cat by Mouton and Holstege (1994). These projections terminate in the medial ventral horn throughout the length of the spinal cord, and not specifically on lordosis motoneurons, many of which are located in lateral portions of the ventral horn. The PAG has been proposed to have a tonic effect on the gigantocellular and the lateral vestibular nuclei, which in turn control the spinal cord (Pfaff and Schwartz-Giblin, 1988; Ogawa et al., 1991). Lesions in these areas affected lordosis (Modianos and Pfaff, 1976), whereas stimulation activated deep lumbar axial muscles involved in lordosis (Schwartz-Giblin et al., 1984; Cottingham et al., 1988). The lateral vestibular nucleus and the paragiganto- cellular nucleus project to the ventromedial parts of the intermediate zone throughout the entire length of the spinal cord, but not specifically to lumbar levels (Holstege and Kuypers, 1982). Furthermore, lordosis behavior not only involves the activation of axial, but also of hindlimb muscles (Martinez-Gomez et al., 1992;

98 NRA projections to lumbosacral motoneurons: males

Chapter 4

99

NRA projections to lumbosacral motoneurons: males Nucleus retroambiguus projections to lumbosacral moto- neuronal cell groups in the male cat, possibly representing the final common pathway for mounting behavior

ABSTRACT Recently, in the female cat nucleus retroambiguus (NRA) projections have been described to distinct motoneuronal cell groups in the lumbar enlargement, possibly involved in lordosis behavior. The present paper deals with the NRA-lumbosacral pathway in the male cat. Lumbosacral injections of wheat germ agglutinin-horseradish peroxidase (WGA-HRP) were made to localize and quantify retrogradely labeled neurons in the caudal medulla. These injections were preceded by spinal hemisections in order to distinguish between neurons with ipsi- and neurons with contralaterally descending axons. The NRA-lumbosacral fibers descended almost exclusively contralaterally, but neurons in areas surrounding the NRA projected mainly ipsilaterally. Injections of WGA-HRP were made in the region of the NRA to determine its targets in the lumbosacral cord. To distinguish between the contralateral NRA pathways and the ipsilateral projections from neurons in the adjoining lateral tegmentum, the injections were preceded by ipsilateral hemisections in C2. A new scaling method was used to compare the results of the different cases, despite variations in size and segmental organization. The results show that the distribution pattern of anterogradely labeled fibers in the lumbosacral cord matched precisely the location of certain motoneuronal cell groups. The NRA projected densely to abdominal wall and pelvic floor motoneurons in Onuf’s nucleus, moderately to adductor longus, semimembranosus, and biceps femoris anterior motoneuronal cell groups, and only sparsely to iliopsoas and semitendinosus motoneuronal cell groups. Compared with the findings in the female, the NRA in the male cat projects more heavily to the biceps anterior and adductor longus, and only sparsely to the iliopsoas and semitendinosus motoneuronal cell groups. These male-female differences are discussed.

INTRODUCTION The nucleus retroambiguus (NRA) is a group of neurons (Holstege, 1989; Zhang et al., 1992) and spinal cord in the ventrolateral part of the caudal medulla oblongata motor nuclei (Merrill, 1971; Holstege and Kuypers, which has been described in humans (Olszewski and 1982; Feldman et al., 1985; Holstege and Tan, 1987; Baxter, 1954), cats (Merrill, 1970; Holstege, 1989), rats Miller et al., 1989; VanderHorst et al., 1994; VanderHorst (Ellenberger and Feldman, 1990), and songbirds (Wild, and Holstege, 1995). Regarding its involvement in 1993). In the cat, these interneurons form a column respiration, vocalization, and lordosis, the NRA receives extending between 1 and 8 mm caudal to the obex major afferent input from respiratory related neuronal (VanderHorst and Holstege, 1995). cell groups in the lateral tegmental field of the medulla NRA neurons have been shown to be respiration related and pons (for review see Feldman, 1986; Smith et al., (Merrill, 1970,1974). The caudal part of the NRA 1989; Gerrits and Holstege, in press), from two cell contains expiration related neurons, whereas rostrally groups in the ventral medullary medial tegmental field expiration and inspiration related neurons are inter- (Gerrits and Holstege, in press), as well as from the mingled (Merrill, 1970; 1974). By workers in the field midbrain periaqueductal gray (PAG; Holstege, 1989; of respiration, the NRA is referred to as the caudal part Davis and Zhang, 1991; VanderHorst and Holstege, of the ventral respiratory group (see Feldman, 1986 for 1996a). The PAG is the final integrator for vocalization review). More recent studies have demonstrated that (Zhang et al., 1994) and lordosis (Sakuma and Pfaff, NRA neurons are not only involved in expiration, but 1979; Ogawa et al., 1991), but also for various other also in vomiting (Miller et al., 1987, 1995), defecation components of emotional (aggressive and defensive) (Fukuda and Fukai, 1986, 1988), vocalization (Jürgens behavior, such as blood pressure control, nociception and Pratt, 1979; Larson and Kistler, 1984; Holstege, control, and possibly micturition (Besson and Chaouch, 1989; Larson, 1991; Zhang et al., 1992) and possibly 1987; Bandler et al., 1991; Lovick, 1993; Blok and lordosis (VanderHorst and Holstege, 1995). NRA Holstege, 1994). interneurons mediate these activities by way of very In a preceding study in the female cat, it has been distinct, direct projections to a selection of brainstem demonstrated that the NRA projects to certain moto-

101 Chapter 4

neuronal cell groups in the lumbosacral cord, innervating which implies that the labeled fibers in the spinal cord muscles of the hindlimb (i. e. iliopsoas, small adductors, would be derived from NRA neurons and from neurons and hamstrings), pelvic floor and lower back Vander- in the adjacent tegmentum. Horst and Holstege, 1995). Based on the specificity of The retrograde tracing experiments demonstrated that this projection pattern, it was hypothesized that the NRA the non-NRA spinal cord projecting neurons send their is involved in lordosis behavior, the receptive posture axons through the ipsilateral white matter (see also of the mating female. These findings raise the question VanderHorst and Holstege, 1995), whereas the NRA whether such projections also exist in males. sends its fibers through the contralateral cord. In order In order to precisely determine the lumbosacral to ascertain that the anterogradely labeled fibers are motoneuronal cell groups receiving NRA afferents, a derived from neurons in the NRA, an ipsilateral C2 detailed overview of the location of all lumbosacral hemisection was made prior to the injection in the NRA. motoneuronal cell groups was necessary. Thus far, it Segments L3 to S3 were cut and a 1 out of 4 series of was difficult to integrate retro- and anterograde tracing consecutive sections was processed and microscopically findings in the lumbosacral cord of different animals studied. because of sometimes large differences in length and segmental subdivision (Sherrington, 1892; Romanes, Comparison of motoneuronal cell groups and 1951). A new method was developed which made such NRA fibers in different cases an integration possible (see accompanying paper In order to match the rostrocaudal distribution of labeled VanderHorst and Holstege, 1996b). The results lumbosacral motoneuronal cell groups and the demonstrate that also in the male cat a specific NRA- distribution pattern of anterogradely labeled NRA fibers motoneuronal projection system exists, which precisely in the 4 cases, two problems had to be solved (see also matches the location of a distinct set of motoneuronal accompanying paper VanderHorst and Holstege, 1996b). cell groups innervating hindlimb and pelvic floor The first problem concerned the differences in the muscles. organization of dorsal and ventral rootlets forming the lumbosacral plexus (Romanes, 1951; Sherrington, 1892) MATERIAL AND METHODS between individual cases. These differences resulted in a considerable variation in length and level of the Six adult male cats were used. See Chapter 1 for the lumbosacral segments. The second problem was that the general surgical and histological procedures. absolute total length of the lumbar enlargement also varied greatly between the cases due to differences in Retrograde study size of the animals. To overcome these problems, the In order to precisely localize the neuronal cell groups in same method had to be applied in the NRA injected the caudal medulla oblongata projecting to the cases as had been used for the motoneuronal cell groups lumbosacral cord in the male cat, in 2 cases injections in the accompanying paper (VanderHorst and Holstege, µ of 3 l 2.5% WGA-HRP were made in the L5 (case 1996b). Two landmarks were chosen in the enlargement, 2309) and S1-S3 (2329) segments, respectively, after one rostrally (level 0) and one caudally (level 100). laminectomy. Both injections were made bilaterally and Caudal to level 0, ventrolateral part of the ventral horn were preceded by a hemisection a few segments rostrally. extends latterally due to the presence of psoas major This was done to separate the neurons that send their motoneurons which form the most rostral motoneuronal axons to the lumbosacral cord via the ipsi- and cell group of the hindlimb. At level 100, the ventrolateral contralateral funiculus, respectively. The lumbosacral ventral horn suddenly decreases due to the disappearance cord, the caudal medulla and the C1 segment were cut of motoneuronal cell groups innervating gluteal muscles and a 1 out of 4 series of consecutive sections was and muscles of the distal hindlimb. These landmarks processed. The retrogradely labeled neurons in the always can be identified independent of the presence of caudal medulla and C1 were counted and plotted. labeled motoneurons. In this way, it was possible to Anterograde study precisely match the distribution of motoneuronal cell The purpose of these experiments was to study all NRA groups with the distribution pattern of labeled NRA projections to the lumbosacral cord. This meant that, in fibers. In one representative case (2287) color each of 4 male cats, 3-5 injections of 30-50 nl 2.5% photomicrographs were taken from every sixth (1:4) WGA-HRP had to be made to involve all rostrocaudal collected section, representing a length of 6 x 4 x 40µm= levels of the NRA in the injection site. In all cases the 0.96 mm, and every photomicrograph was matched with caudal medulla and the first two cervical segments were the corresponding overview drawing of the lumbosacral exposed dorsally. The injections not only involved the motoneuronal cell groups of the accompanying paper NRA, but extended into the adjacent lateral tegmentum, (VanderHorst and Holstege, 1996b).

102 NRA projections to lumbosacral motoneurons: males Case 2309 Case 2329

Obex Obex ECU ECU CU G CU G Vspin. S S caud. XII XII

LTF LTF LRN IO LRN IO

P P -2 mm -2 mm CU CU Vspin. G G caud. NRA NRA

LTF LTF LRN LRN P P

-4 mm Vspin. -4 mm CU caud. G CU G NRA NRA P P

-6 mm -6 mm G NRA NRA

P P

-8 mm -8 mm

NRA LCN NRA LCN

hemisection L3 hemisection L5

injection L5 injection S1-S3

Figure 1 Plottings of labeled neurons in the caudal medulla oblongata after injections into the L5 and S1-S3 segments (cases 2309 and 2329, respectively).

103 Chapter 4

L5 injection S1-S3 injection left L3 hemisection left L5 hemisection 10 10 left right

5 5

average # neurons/section 0 average # neurons/section 0 012345678 012345678 level level Figure 2 Rostrocaudal distribution of retrogradely labeled neurons in the NRA.

RESULTS Distribution pattern of NRA-fibers in the lumbosacral cord All 4 males showed a similar distribution pattern of NRA Retrograde study fibers in the L3-S3 lumbosacral segments. Descending After bilateral injections the L5 and S1-S3 segments labeled fibers were found in the dorsolateral, ventro- (see Fig. 1 for location of the injections and hemi- lateral, and ventral funiculi of the cervical, thoracic and sections), ipsilateral to the hemisection numerous lumbosacral segments. From L1 to S3 their number retrogradely labeled neurons were found in the NRA gradually decreased. From rostral L4 to caudal L5, the and only few in the area adjacent to the NRA (Figs. 1 labeled fibers in the ventrolateral funiculus gradually and 2; Table 1). Labeled NRA neurons were present shifted into a more peripheral position. Only a very between 1 and 8 mm caudal to the obex and their axons limited number of retrogradely labeled neurons was crossed the midline ventral to the central canal, at the found in the lumbosacral cord, which indicates that the level of the NRA itself. Contralateral to the hemisection, bulk of the labeled fibers in the white matter represent the NRA contained hardly any labeled neurons, whereas descending fibers. many were present just ventromedial to the NRA (Table 1). This ventromedial group extended further rostrally Projections to laminae V-VII and X than the NRA, where it increased in size. These results A few labeled fibers were distributed to laminae V to demonstrate that the NRA projects to the lumbosacral VII and many to lamina X (Figs. 4 and 5). All these cord via the contralateral funiculus, whereas the neurons projections were bilateral, but with a strong contralateral in the tegmentum medially adjoining the NRA project preponderance, and were denser at sacral than at lumbar mainly ipsilaterally to the lumbosacral cord. levels. The ipsilateral projections were derived from fibers descending through the contralateral funiculi and that had recrossed in the ventral gray commissure at the Table 1 Number of labeled neurons NRA medial to NRA level of termination. Case left (%) right (%) left (%) right (%) 2309 168 (97) 5 (3) 26 (10) 238 (90) Projections to motoneuronal cell groups 2329 203 (96) 9 (4) 24 (8) 289 (92) In addition to the projections to laminae V-VII and X, Total number of retrogradely labeled neurons in the specific projections were found to certain motoneuronal caudal medulla counted in 1:4 sections after bilateral cell groups in lamina IX. These projections were much injections in the L5 (case 2309) and S1-S3 (case 2339), denser than the projections to the intermediate zone and preceded by left sided hemisections. equally dense as the projections to lamina X. After applying the same relative 0-100 method for the NRA cases as had been used for the drawings of motoneuronal cell groups (VanderHorst and Holstege, 1996b), the Anterograde study distribution pattern of labeled NRA fibers was matched Location of injections and hemisections with the location of the lumbosacral motoneuronal cell In all 4 cases (2287, 2294, 2326, and 2360), the groups (Fig. 5). injections involved the entire rostrocaudal extent of the NRA but extended into the adjoining lateral tegmental Level minus 220 to minus 30 In the L3 segment (± field and the lateral funiculus (Fig. 3). The ipsilateral level minus 200), which does not contain hindlimb C2 hemisections were complete and ventrally did not moto-neuronal cell groups, a very dense, bilateral extend across the midline. projection was present in the motoneuronal cell group

104 NRA projections to lumbosacral motoneurons: males

located at the rostral end of each lumbosacral segment. Obex -2mm -4mm -6mm -8mm In case 2287, levels 35-45 correspond with the rostral part of L6. In the motoneuronal cell groups of the remaining adductors of the thigh and of the quadriceps muscles 2287 C2 only occasional labeled fibers were observed (see levels 27, 30, and 47 of Fig. 5). In all likelihood, these fibers did not terminate in these cell groups, but were fibers of passage on their way to their target motoneuronal cell 2294 C2 groups.

Level 50 to120 At levels 50 to 70, a relatively dense projection was present in the center of the ventral horn 2326 C2 which precisely overlaps the semimembranosus motoneuronal cell group (see level 60 of Fig. 5). The semitendinosus motoneuronal cell group at levels 65 to 90 did receive only very few, if any, labeled fibers (see 2360 C2 level 73 of Fig. 5). At the other hand, at levels 75 to 90 a dense projection to the biceps anterior motoneuronal Figure 3 Injection sites in the caudal medulla of all 4 cell group was present (see level 83 of Fig. 5). The anterograde cases. posterior biceps motoneuronal cell group, located medial to the anterior biceps group (levels 85 to 95), also received a dense NRA projection (see level 90 of Fig. 5). Only a few labeled fibers were found in other innervating the internal and external oblique abdominal motoneuronal cell groups, innervating the gluteal wall muscles (Holstege et al., 1987; Miller, 1987; Fig. muscles, the peroneus longus and the flexor digitorum 4). This projection continued caudal until halfway the longus (see levels 60, 75, 85, and 90 of Fig. 5). L3 segment (approximately level minus 120). Around level 100, many labeled NRA fibers projected to pelvic floor motoneurons in Onuf’s nucleus (see levels Level minus 30 to 50 Just rostral to the enlargement 100 and 107 of Fig. 5). Dendrites of these motoneurons, (level minus 30 to 0, corresponding to the caudal L4), a extending dorsally into the intermediate zone (Beattie sparse projection was found to the psoas minor et al., 1990; Sasaki, 1994), also received heavy NRA motoneuronal cell group. Slightly more caudally (levels input. Although the projection to Onuf’s nucleus 0 to 10), in the most rostral part of the enlargement, a involved the entire motoneuronal cell group, it was more limited number of labeled NRA fibers was present in predominant in its dorsomedial part, containing the group containing mm. psoas minor and major motoneurons innervating the anal sphincter. Between motoneurons (level 3 of Fig. 5). The NRA projection levels 110 and 120, a few thick labeled fibers were found did not continue further caudally, in contrast to the psoas in the ventromedial corner of the ventral horn (see level major motoneuronal cell group (see Fig. 4 of Vander- 113 of Fig. 5). This very dense but restricted projection Horst and Holstege, 1996b). This finding suggests that involved the levator ani/ abductor caudae internus only psoas minor motoneurons received NRA-afferents. motoneuronal cell group (see accompanying paper A few fibers were found more medially, possibly VanderHorst and Holstege, 1996b). Around level 110, a representing terminations onto dendrites of psoas minor few labeled fibers were present dorsolaterally in the moto-neurons. ventral horn, where the motoneurons of the intrinsic foot More caudally, (levels 10 to 20), a dense projection to muscles as well as ischiocavernosus and sacral the adductor longus motoneuronal cell group was present parasympathetic neurons are located. Although it was (see levels 27 and 30 of Fig. 5). The termination pattern not possible to determine which of these motoneuronal precisely matched the entire (levels 15 to 45) cell group. cell groups was the target of the NRA afferents, it is The densest projection to this group was found at level expected that the projection concerns the ischio- 30, where adductor longus motoneurons are most cavernosus motoneurons. This idea is based upon the numerous. The sartorius motoneuronal cell group (see observation that, unlike the ischiocavernosus cell group, levels 27 and 30 of Fig. 5; VanderHorst and Holstege, at adjoining rostral and caudal levels the motoneurons 1996b) was devoid of NRA afferents. of the intrinsic foot muscles and the parasympathetic Apart from the projection to the adductor longus, motoneurons did not receive a significant NRA between levels 35 and 40 a distinct projection was projection. observed to a restricted area in the dorsomedial ventral For a summary of the involvement of lumbosacral horn, which contains multifidi motoneurons (Vander- motoneuronal cel groups in the NRA projection, see Horst and Holstege, 1996b). These motoneurons are Table 2.

105 Chapter 4

Figure 4 Low power photomicrograph showing anterogradely labeled NRA fibers in the L3 segment. Note that NRA fibers descend throughout the contralateral funiculi and at this level terminate heavily and bilaterally in the ventrolateral part of the ventral horn. Bar represents 1 mm.

level 3

pectineus

iliacus psoas minor and major

level 27 adductor sartorius anterior longus and medius gracilis vastus lateralis

iliacus

adductor magnus psoas major adductor brevis vastus medialis

level 30 adductor longus sartorius anterior and medius gracilis rectus femoris vastus lateralis iliacus vastus medialis adductor magnus adductor brevis psoas major

Figure 5 Comparison of the distribution of labeled NRA fibers (left; case 2287) and the location of lumbosacral motoneuronal cell groups (right) at different levels in the lumbosacral cord. In each of the drawings of motoneuronal cell groups, the location of motoneuronal cell groups as described by VanderHorst and Holstege (accompanying paper) was adapted to the shape of the ventral horn of case 2287 at corresponding levels. It must be emphasized that the photographs represent only one section, whereas the drawings of motoneuronal cell groups contain the information of 12 alternate sections. Bar represents 0.5 mm.

106 NRA projections to lumbosacral motoneurons: males

level 47 vastus intermedius gracilis vastus lateralis

iliacus

adductor magnus vastus adductor brevis medialis tensor fasciae latae

level 60 flexor tibialis digitorum anterior longus

semimembranosus

extensor digitorum longus

obturatorius tensor fasciae internus latae and gluteus medius

level 73 flexor digitorum longus peroneus longus plantaris

biceps semitendinosus femoris anterior

gluteus obturatorius internus medius

level 83 intrinsic gastrocnemius foot lateralis peroneus gastrocnemius longus medialis and soleus

semitendinosus

medial longissimus gluteus medius

gluteus biceps femoris maximus caudo- anterior femoralis

107 Chapter 4

level 90 intrinsic foot

gastrocnemius medialis multifidi

gluteus medial longissimus medius

caudofemoralis biceps femoris posterior

level 100 intrinsic foot

longissimus external anal medialis sphincter

level 107 intrinsic foot

external urethral sphincter and ischiocavernosus levator ani and abductor caudae internus

level 113 external urethral sphincter bladder and ischiocavernosus

levator ani and longissimus medialis abductor caudae internus

108 NRA projections to lumbosacral motoneurons: males

DISCUSSION NRA projection is stronger in the dorsomedial part of Onuf’s nucleus, innervating the anal sphincter and bulbo- cavernosus muscles, than in Onuf’s ventrolateral part, Comparison of the lumbosacral NRA innervating the urethral sphincter and the ischio- projection in males and females cavernosus (Sato et al., 1978; Yamamoto et al., 1978; Pullen, 1988; Thor et al., 1989; Beattie et al., 1990). The NRA projection to levator ani/abductor caudae The retrograde results demonstrated that in the male cat, internus motoneurons, was denser in females than in similar to the female, the NRA neurons projecting to males. Furthermore, this projection was mainly the lumbosacral cord are located between 1 and 8 mm ipsilateral in the estrous female and mainly contralateral caudal to the obex, similar to the female. in the male. No clear male/female differences were found The anterograde results of the present study showed that in the NRA projections to the psoas minor, multifidi, the NRA in the male projects to distinct hindlimb and and posterior biceps motoneuronal cell groups. The pelvic floor motoneuronal cell groups. These projections motoneuronal cell groups of the sartorius, large most heavily involved the adductor longus, semi- adductors, quadriceps and distal muscles of the hindlimb membranosus, anterior and posterior biceps femoris did not receive significant projections from the NRA in motoneuronal cell groups, and Onuf’s nucleus. either males or females. Previously, the NRA projections to the lumbosacral motoneurons in female non-estrous (VanderHorst and Holstege, 1995) and in estrous cats (VanderHorst et al., Functional implications 1995; VanderHorst and Holstege, in preparation) have been described. Comparing the results in the female with those of the present study in the male suggests some Abdominal pressure control differences in the NRA motoneuronal projection Interneurons in the NRA have been shown to be involved between males and females (Table 3). In the following in multiple behavioral patterns, such as respiration comparison between male/female NRA-motoneuronal (Merrill, 1970, 1974; Feldman, 1986), vocalization projection patterns, the findings in the estrous females (Holstege, 1989; Zhang et al., 1992, 1994), defecation will be used. In males, the iliacus, psoas major, (Fukuda and Fukai, 1986), and vomiting (Miller et al., semitendinosus and levator ani/abductor caudae internus 1987; Miller et al., 1995). These functions have in motoneuronal cell groups received much sparser common that they are accompanied by elevation of intra- projections than in the estrous female, whereas the abdominal pressure, which is caused by activation of adductor longus and anterior biceps motoneurons were abdominal wall and pelvic floor. The dense NRA more densely innervated. The projection to Onuf’s projections to motoneurons of the abdominal wall and nucleus including its dendrites is more prominent in pelvic floor muscles probably form the anatomical males than in females. In both males and females, the framework of these behaviors.

Table 2. Lumbosacral motoneuronal cell groups receiving NRA projections in the male cat

Distinct projection to entire No distinct projection* No projection motoneuronal cell group Psoas minor Psoas major Sartorius Adductor longus Iliacus Vastus medialis Semimembranosus Pectineus Vastus lateralis Biceps femoris anterior Gracilis Vastus intermedius Biceps femoris posterior Semitendinosus Rectus femoris Tensor fasciae latae Adductor femoris brevis External anal sphincter Gluteus medius Adductor femoris magnus Bulbocavernosus Gluteus maximus Abductor cruris caudalis External urethral sphincter Caudofemoralis Obturatorius internus Ischiocavernosus Peroneus longus Gastrocnemius lateralis Levator ani Flexor digitorum longus Gastrocnemius medialis Abductor caudae internus Intrinsic foot Soleus Medial longissimus (S1) Plantaris Multifidi (L6) Tibialis posterior Tibialis anterior Extensor digitorum longus Flexor hallucis longus

* The middle column represents the cell groups in which only a limited number of labeled NRA fibers was found in only part of the respective groups. In all likelihood, these fibers represent fibers of passage on their way to their target motoneuronal cell groups (left column).

109 Chapter 4

Table 3. Differences in the NRA-lumbosacral motoneuronal projection between males and females

Motoneuronal cell group Male Female Function Obliquus externus and internus +++ +++ Expiration

Psoas minor + + Flexion of lumbosacral vertebral column to pelvis Psoas major ¥ + Flexion hip; fixation spine Iliacus ¥ + Flexion of pelvis to lumbosacral vertebral column; flexion hip Adductor longus ++ + Adduction thigh Semimembranosus + ++ Extension hip; flexion knee; some inward rotation thigh Semitendinosus ¥ + Flexion knee; extension hip Biceps femoris anterior ++ + Extension hip; some extension knee; some outward rotation thigh Biceps femoris posterior + + Flexion knee

External anal sphincter/ Bulbocavernosus +++ ++ Close anus; erection External urethral sphincter/ Ischiocavernosus ++ + Close urethra; erection and anteroflexion penis Levator ani/ Abductor caudae internus + ++ Compression rectum; abduction and flexion tail

Multifidi (L6) + + Extension lower back Medial longissimus (S1) ¥ + Extension lower back

¥ indicates that there is only a very sparse or no projection (see Table 2). (see also Chapters 4 and 6).

NRA projections to cutanei trunci (Holstege and Blok, the concept is true, during mating both males and 1989), and pectoralis major (Holstege and Kuypers, females should demonstrated strong involvement of the 1982) motoneurons in the cervical cord and to iliopsoas, psoas minor, semimembranosus, biceps femoris adductor longus, and hamstring motoneurons in the posterior, and no involvement of muscles as the lumbosacral cord are more delicate and suggest that the quadriceps, adductor magnus, sartorius and triceps surae. NRA is involved in other functions than abdominal The activity of the psoas major, semitendinosus and pressure control. biceps femoris anterior would differ between males and females. Lordosis behavior In a previous paper (VanderHorst and Holstege, 1995) Brain structures known to be involved in male sexual it has been proposed that in the female cat, the NRA- behavior Lesion and electrical stimulation studies in lumbosacral pathway plays a role in receptive behavior rodents, cats, dogs, and monkeys have revealed several or lordosis behavior (Fig. 6). This behavior consists of important areas in the forebrain which play a role in crouching, elevation of the lower back (lordosis), male reproductive behavior, such as the neocortex (for treading (movements of the hindlimbs), and lateral deviation of the tail (Michael, 1960). It is only displayed by estrous females and is so strong that it overrules almost all other behavioral patterns. The findings in the present study in the male cat raise the question whether the NRA lumbosacral pathway in the male is involved in mating behavior as well.

Mounting behavior Most studies on male mating behavior have focussed on the regulation of ejaculatory and erectile (autonomic) responses. However, the neural circuitry underlying the male mating posture has never been studied before. Males show a different type of mating behavior than females. It consists of the neck grip, grasping with the forelimbs (forelimb rubbing), hindlimb stepping, arching of the back, stamping on the mounted females cat’s rump (palpation of the flanks), and pelvic thrusting (Fig. 6; Whalen, 1963). In contrast to females, adult male cats are able to mate any time if an estrous female is available. Figure 6 Photograph showing mating cats. The female In order to verify the hypothesis that NRA-lumbosacral displays crouching, elevation of the lower back, treading pathways form the final common pathway for mating (movements of the hindlimbs), and lateral deviation of the behavior, it is necessary to study the electromyographic tail, whereas the male shows the neck grip, grasping with the (EMG) activity of hindlimb muscles during mating. If forelimbs, hindlimb stepping, and arching of the back.

110 NRA projections to lumbosacral motoneurons: males review see Meisel and Sachs, 1994), the main and accessory olfactory systems in rats but not in other species (Giantonia et al., 1970; Aronson and Cooper, 1974; Hart and Haugen, 1972; Goldfoot et al., 1978), the corticomedial nucleus of the amygdala (Giantonia et al., 1970; for review see Meisel and Sachs, 1994), the caudal medial preoptic area/anterior hypothalamus (e.g. Larsson and Heimer, 1964; Heimer and Larsson, 1966; Roberts et al., 1967; Malsbury, 1971; van Dis and Larsson, 1971; Hart et al., 1973; van de Poll and van Dis, 1979; Hansen et al., 1982; Hart and Leedy, 1983; for review see Meisel and Sachs, 1994). It is not known which structures in the mesencephalon, pons, and medulla oblongata are involved in this behavior. The present study points to the NRA- lumbosacral pathway to play an important role in this respect. Although the central nucleus of the amygdala and the medial preoptic area/anterior hypothalamus do not directly project to the NRA (Hopkins and Holstege, 1978; Holstege, 1987), they may do so indirectly via the PAG (Holstege, 1989; VanderHorst and Holstege, 1996a) or the ventral medullary medial tegmental field (Gerrits and Holstege, in press). The PAG has been shown to be involved the integration of survival behavior including female mating behavior (Sakuma and Pfaff, 1979; Ogawa et al., 1991), but its role has never been studied in relation to male sexual behavior. Stimulation in the PAG is known to elicit hindlimb movements and arching of the back which have been interpretated as being part of aversive and defensive responses (Bandler et al., 1991). Possibly, part of these movements also play a role in mounting behavior.

In summary, a concept is proposed in which the PAG- NRA-lumbosacral projection is involved in both female and male mating behavior. This hypothesis is in line with reports that mounting and lordosis are not exclusively displayed by males and females, respectively (Beach, 1948; Michael, 1960; Emerey and Sachs, 1975; Krieger and Barfield, 1976; Davidson and Block, 1969; Larsson et al., 1976; Sodersten and Larsson, 1975; van de Poll and van Dis, 1977). Possibly, part of the differences in male and female mating behavior are based on the differences observed in the NRA- motoneuronal projections.

111

NRA projections to lumbosacral motoneurons: estrogen induced sprouting

Chapter 5

113

NRA projections to lumbosacral motoneurons: estrogen induced sprouting Estrogen induces axonal outgrowth in the nucleus retro- ambiguus-lumbosacral motoneuronal pathway in the adult female cat

ABSTRACT In 1995, VanderHorst and Holstege (J. Comp. Neurol. 359: 457-475) have discovered a new pathway in the cat, which originates from the nucleus retroambiguus (NRA) and terminates in a distinct set of lumbosacral hindlimb, axial and pelvic floor motoneuronal cell groups. The NRA is a compact group of interneurons located laterally in the caudal medulla oblongata. Its projection to lumbosacral motoneurons is thought to represent the final common pathway for male mounting and for female receptive or lordosis behavior. However, females only display lordosis behavior when they are in estrus, which suggests that the NRA-lumbosacral pathway is only active during estrus. This raised the question whether estrogen affects this pathway. The effect of estrogen on the NRA-lumbosacral projection was studied lightmicroscopically, using wheat germ agglutinin- horseradish peroxidase (WGA-HRP) as a tracer. The rubrospinal pathway served as control. The density of labeled NRA fibers in their target hindlimb motoneuronal cell groups appeared abundant in estrous and very weak in non-estrous cats. Such differences were not found in the rubrospinal pathway. For electronmicroscopical study, the NRA projection to the semimembranosus motoneuronal cell group was selected. In this cell group, an almost ninefold increase of labeled profiles was found in estrous versus non-estrous cats. Moreover, the semimembranosus motoneuronal cell group contained labeled growth cones in estrous, but not in non-estrous cats. These observations demonstrate that estrogen induces axonal outgrowth of NRA fibers to distinct somatic motoneuronal cell groups. The possible mechanisms underlying this outgrowth are discussed.

INTRODUCTION The nucleus retroambiguus (NRA) is a compact group hindlimb, axial, and pelvic floor muscles. Combined of interneurons in the lateral tegmentum of the caudal action of this set of muscles in the female does not serve medulla and has been described in humans, cats, rats, motor activities as stepping, jumping, scratching, hamsters, and birds (Olszewski and Baxter, 1954; running or other daily activities, but underlies aspects Merrill, 1970; Paxinos and Watson, 1986; Holstege of the receptive posture during mating. Such behavior 1989; Ellenberger and Feldman, 1990; Wild, 1993; consists of elevation of the lower back (lordosis), Gerrits et al., in preparation). Anatomically as well as rhythmic movements of the hindlimbs (treading), and physiologically, the NRA has been shown to be involved lateral deviation of the tail (Michael, 1960). This led in respiration, defecation, vomiting and vocalization VanderHorst and Holstege (1995) to postulate that in (Merrill, 1971, 1974; Feldman, 1986; Fukuda and Fukai, female cats the NRA-lumbosacral motoneuronal 1986; Miller et al., 1987; Holstege, 1989; Zhang et al., projection forms the final common pathway for lordosis 1992; Miller et al., 1995). It receives projections from behavior. Electromyographic studies are currently respiration related neurons in the brainstem (Feldman, undertaken by this lab to provide evidence for this 1986; Smith et al., 1989; Gerrits and Holstege, 1996), hypothesis. and from the midbrain periaqueductal gray (PAG; Female reproductive behavior is not displayed in the Holstege, 1989; VanderHorst and Holstege, 1996). The absence of estrogen, whereas in estrous animals it is PAG plays a crucial role in the integration of survival prominently present (Beach, 1948; Young, 1961; Clark behavior, such as defensive and agressive reactions as and Mani, 1994). If the NRA-lumbosacral motoneuronal well as mating (Bandler et al, 1991; Sakuma and Pfaff, projection represents the final common pathway for 1979a,b; Ogawa et al., 1991). lordosis behavior, the question arises whether estrogen In the female as well as in the male cat (VanderHorst et has an effect on this pathway. Therefore, the pathway al., 1994; VanderHorst and Holstege, 1995, 1996; V. G. was studied light- and electronmicroscopically in estrous J. M. VanderHorst and G. Holstege, unpublished obser- and in non-estrous cats. The results indicate that estrogen vations) the NRA projects directly to lumbosacral induces axonal outgrowth of NRA fibers to their target motoneuronal cell groups innervating a distinct set of motoneurons in the lumbosacral cord.

115 Chapter 5

MATERIALS AND METHODS pathways, the rubrospinal tract was chosen as a control pathway. In two cases (2245 and 2310), small WGA- HRP injections were made in the red nucleus, and in Ovariohysterectomy and estrogen treatment two other cases (2362 and 2363) the red nucleus was NRA series. Anterograde tracing experiments were injected with similar volumes of tracer as used in the performed in 17 adult female cats (Table 1; for general NRA injected cases (Table 1). In case 2310, prior to the surgical and histological procedures, see chapter 1). injection, an ipsilateral C2 hemisection was made. Eight females were ovariohysterectomized 4-5 weeks prior to the tracer injection. Five of them were estrogen Lightmicroscopy treated for 7 to 10 days prior to the tracing experiment, The L3-S3 segments were cut on a freezing microtome receiving daily subcutaneous injections of oestradiol into 40 µm thick, transverse sections, except for the benzoate dissolved in oil (Mycofarm, The Netherlands; respective segments of 8 cases (2251, 2256, 2288, 2307, 20 i.e. or 0.02 mg/kg/day). After 3 days of treatment, 2308, 2324 , 2337, and 2353), which were cut on a they started to display the typical lordosis behavior when vibratome into 60µm sections. Every fourth section was tapping the lower back or presenting them to a male processed using the tetramethylbenzidine (TMB; Sigma, cat. Estrogen treatment was continued until to the day The Netherlands) procedure according to Mesulam of perfusion. The remaining 3 ovariectomized control (1982). Sections of the NRA injected cases 2296 and cats received no estrogen. Progesteron was not 2299, and 2307 and 2308, and of the red nucleus cases administered, because cats are reflex ovulators, in which 2362 and 2363 were incubated simultaneously to progesteron levels start to rise only after intromission exclude eventual differences in density of labeling due by the male or intense vaginocervical stimulation to the incubation procedure. The density of anterograde (Dawson and Friedgood, 1940). labeling in the spinal cord was microscopically Of the remaining 9 non-ovariectomized cats, two were examined with a Zeiss Axioskop under combination of in full estrus, i.e. displayed the complete pattern of polarized light and darkfield condensor. Photo- estrous behavior during 3-4 days before the tracer micrographs of representative sections were taken. injection. The other 7 cats showed no signs of estrous behavior in the two weeks before or in the 3 days after Electronmicroscopy the tracer injection. Red nucleus control series. In four females, the rubro- Cases 2251, 2256, 2288, 2307, 2308, 2324, 2337, and spinal projections in estrous and non-estrous females 2353 were examined light- as well as electron- were studied (Table 1). Three cases were ovariectomized, microscopically. The L3-S3 segments of these cases of which two did and one did not receive estrogen were cut on a vibratome. Every third section was treatment prior to the tracer injections. The remaining incubated with TMB and ammoniumheptamolybdate natural case did not display estrous behavior in the two overnight (Olucha et al., 1985). The next day they were weeks before or in the 3 days after the tracer injection. processed using the slow osmication method for postfixation of Henry et al. (1985). The tissue was WGA-HRP injections stained ‘en bloc’ in 1% uranylacetate in bidest, and the NRA series. The distribution of the NRA-lumbosacral slabs were dehydrated in graded series of alcohol and neurons in the cat have been described by VanderHorst embedded in Epon between dimethyldichlorosilane- and Holstege (1995; see also Fig. 1). Since the NRA coated glass-slides (Vinores et al., 1984). A selection extends rostrocaudally over a length of 6-7 mm, multiple was made of those sections containing anterogradely needle penetrations were necessary to inject wheat germ labeled fibers in the semimembranosus motoneuronal agglutinin-horseradish peroxidase (WGA-HRP; Sigma, cell group. This is a compact cell group which does not The Netherlands) throughout its rostrocaudal extent. overlap with other motoneuronal cell groups, except for This reduced the possibility that differences in the its most rostral and most caudal poles (Romanes, 1951; injection site would result in differences in the NRA- 1964; V. G. J. M. VanderHorst and G. Holstege, un- lumbosacral projections between the cases. WGA-HRP published observations), which made it possible to was pressure injected via a glass micropipette using a identify it in unstained sections. The selected tissue was picopump after dorsal approach and exposure of the cut into ultrathin sections (60 nm) and studied electron caudal medulla. Except for cases 2237, 2251, 2256, microscopically. To determine the density of labeled 2258, 2267, and 2271, the injections were preceded by profiles per area, in each case the labeled profiles were an ipsilateral C2 hemisection, which interrupted all counted in 32 mazes covering the semimembranosus ipsilaterally descending, non-NRA pathways to the motoneuronal cell group, each maze measuring spinal cord (VanderHorst and Holstege, 1995). The 10.000µm2. The symmetry or asymmetry of the synaptic hemisections were made by aspiration with a glass membrane specialization and the content of the labeled pipette. profiles was established. In cases 2324 and 2337, the Red nucleus control series. To rule out the possibility perimeter of labeled profiles and the length of that estrogen has a similar effect on other descending postsynaptic densities was measured.

116 NRA projections to lumbosacral motoneurons: estrogen induced sprouting

Figure 1 Schematic drawings ECU showing the location of the G CU S V caud. NRA-lumbosacral neurons in XII CU spin. G V caud. the caudal medulla oblongata. G spin. CU Note that it is most prominant LTF G NRA NRA between 2 and 6 mm caudal LRN NRA P to the obex where it easily can IO LRN P P be recognized as a protrusion of gray matter into the white Obex -2 mm -4 mm -6 mm -8 mm matter (see also VanderHorst and Holstege, 1995). CU= cuneate nucleus; ECU= external cuneate nucleus; G= gracile nucleus; IO= inferior olive; LRN= lateral reticular nucleus; LTF= lateral tegmental field; NRA= nucleus retroambiguus; P= pyramidal tract; S= solitary complex; V caud. spin.= caudal spinal trigeminus complex; XII= hypoglossal nucleus.

Non-estrous Estrous natural non-estrous natural estrous

2237 2288 C2

2251 2337 C2

2256 ovariectomized-estrogen treated

2258 2299 C2

2267 2307 C2

2271 2320 C2

2286 C2 2353 C2

ovariectomized-non estrogen treated 2361 C2

2296 C2

2290 2308 C2 Figure 2 Schematic drawings of hemisections and WGA- HRP injection sites involving the nucleus retroambiguus in estrous and non-estrous cases. The core of the injections is 2324 indicated in black. C2

117 Chapter 5 AB

Figure 3 Photomicrographs showing examples of the WGA-HRP injection sites in cases 2324 (left) and 2288 (right). Bar represents 1 mm.

RESULTS

Location and size of the injection sites in the Table 1 Overview of the NRA and red nucleus cases. nucleus retroambiguus In all 17 cases, the injections involved the NRA over a considerable rostrocaudal extent (Fig. 2; Table 1). The Density hemi- WGA- number of nl/ total hemisections were complete and did not extend across NRA section HRP injections injection (nl) LM EM the midline. Examples of the DAB injection sites are Non-estrous shown in cases 2288 and 2324 (Fig. 3). natural non-estrous 2237 - 5% 3 30 90 ± ns 2251 - 5% 2 30 60 ± 25 Non-estrous cases 2256 - 5% 5 40 200 ± 9 2258 - 5% 2 25 50 ± ns Lightmicroscopy In the lumbosacral cord of all natural 2267 - 2.5% 2 30 60 ± ns and ovariectomized non-estrous cases (Table I; Fig. 2), 2271 - 2.5% 3 30 90 ± ns the abdominal wall and pelvic floor motor nuclei 2286 + 2.5% 5 45 225 ± ns ovariectomized-non estrogen treated received a substantial projection (Fig. 4; see Holstege 2296 + 2.5% 3 50 150 ± ns and Kuypers, 1982; Feldman et al., 1985; Holstege and 2308 + 2.5% 4 50 200 ± 12 Tan, 1987; Miller et al., 1989; Holstege, 1989). In 2324 + 2.5% 3 40 120 + 38 Estrous addition, labeled fibers terminated in the hindlimb natural estrous motoneuronal cell groups (Fig. 5 left), but they were so 2288 + 2.5% 4 40 160 ++++ 281 2337 + 2.5% 4 30 120 +++ 157 sparse that in single sections it was not possible to ovariectomized-estrogen treated determine which hindlimb motor nuclei were their main 2299 + 2.5% 3 50 150 +++ ns 2307 + 2.5% 4 50 200 +++ 112 target. However, superposition of 6 consecutive 1:4 2320 + 2.5% 1 30 30 ++ ns sections revealed a distinct projection pattern to the 2353 + 2.5% 4 40 160 +++ 183 motoneuronal cell groups innervating the muscles 2361 + 2.5% 4 30 120 ++++ ns iliopsoas, adductor longus, semimembranosus, semi- Red nucleus tendinosus, biceps femoris anterior and posterior, Non-estrous external anal and urethral sphincter, levator ani, abductor natural non-estrous 2245 - 5% 1 15 15 +++++ ns caudae internus, medial longissimus and multifidi ovariectomized-non estrogen treated (VanderHorst and Holstege, 1995). 2362 - 2.5% 6 25 150 ++++++ ns Estrous ovariectomized-estrogen treated Electronmicroscopy In the semimembranosus moto- 2310 + 2.5% 2 30 60 +++++ ns neuronal cell group of cases 2251, 2256, 2308, and 2324, 2363 - 2.5% 6 25 150 ++++++ ns a total number of 25, 9, 12, and 38 labeled terminals were counted per 32 mazes (320.000 µm2), respectively (Table 1). Synapses, which were all asymmetrical, were The state (estrus or non-estrus), the concentration and volume present in 22% of these profiles. Furthermore, 66% of of tracer injected, and the density of the projections to the the labeled trminals contained exclusively spherical lumbosacral cord at the light- and electronmicroscopical level are indicated. ns= not studied; LM= lightmicroscopical level; vesicles, whereas 13% contained spherical as well as a EM= electronmicroscopical level. few dense cored vesicles (Table 2). Of the remaining 22%, the vesicle content could not be identified. Apart

118 NRA projections to lumbosacral motoneurons: estrogen induced sprouting

Non estrous Estrous

L3 L3

S1 S1 Figure 4 Darkfield-polarized light photomicrographs, showing labeled NRA fibers in the abdominal wall motoneuronal cell groups at the L3 level and in Onuf’s nucleus at the S1 level in non-estrous (left; case 2324) and estrous cats (right; case 2288). Note that the NRA projections to the abdominal wall and pelvic floor motor nuclei are slightly denser in estrus than in non- estrus. Bar represents 400 µm. from spherical and dense cored vesicles, labeled estrous cases (Fig. 4). terminals frequently contained some clustered mito- Between the cases of the estrous group, some differences chondria (Fig. 6 left). Labeled terminals with flattened in density of the NRA-lumbosacral projection were or pleiomorphic vesicles were never observed. present. In natural estrous case 2288 and in ovariec- In one case (2324), the average perimeter of labeled tomized, estrogen treated case 2361, the projections were profiles was determined, and amounted 7.41µm (n=38; extremely strong. In case 2320, with a small injection ranging from 3.29 to 14.43 µm). On average, 6.9% of it in the NRA, labeled fibers were not as numerous as in was covered by synaptic junctions. Labeled profiles the other estrous cases, but still far outnumbered the predominantly contacted dendrites (Fig. 6 left), and only NRA fibers in the non-estrous cases, including those very occasionally neuronal somata. with large NRA injections (Fig. 2; Table 1).

Estrous cases Electronmicroscopy The number of labeled terminals Lightmicroscopy In all natural and ovariectomized per 320.000 µm2 in the semimembranosus motoneuronal estrous cases, the NRA-lumbosacral projection was cell group of estrous cases 2288, 2307, 2353, and 2337 remarkably dense. In sharp contrast to the non-estrous amounted 281, 112, 183, and 157, respectively (Table cases, the projections to motoneuronal cell groups 1). The average number of labeled terminals per maze innervating the iliopsoas, adductor longus, semi- differed significantly between the group of estrous and membranosus, semitendinosus, biceps femoris anterior the group of non-estrous cases (Wilcoxon-Mann- and posterior, and levator ani/abductor caudae internus Whitney test; p<0.025; Fig. 7). In 14% of the labeled were very prominent and could easily be discerned in terminals in the 4 estrous cases, synapses were observed, single sections (Fig. 5 right). The differences between which were all asymmetrical. Of the labeled profiles, estrous and non-estrous cases in respect to the density 63% contained spherical vesicles and 11% spherical as of projections to the abdominal or pelvic floor well as dense cored vesicles. In the remaining 26%, the motoneuronal cell groups were less apparent. These vesicle content could not be identified (Table 2). No projections are equally prominent in estrous and in non- labeled profiles with flattened of pleiomorphic vesicles

119 Chapter 5

Non estrous Estrous

L4 L4

L5 L5

L6 L6

L7 L7

Figure 5 Darkfield-polarized light photomicrographs of labeled NRA fibers in the motoneuronal cell groups of the iliopsoas (L4), adductor longus (L5), semimembranosus (L6), and biceps anterior (L7) in non-estrous case 2324 (left) and estrous case 2288 (right). Note that large difference in the density of NRA fibers between the non-estrous and the estrous case. Bar represents 300 µm.

120 NRA projections to lumbosacral motoneurons: estrogen induced sprouting

Table 2 Labeled NRA profiles in the semimembranosus motoneuronal cell group of non-estrous and estrous cases.

total number labeled terminals with labeled terminals in which non identifiable of labeled asymmetrical synapses no synaps was observed labeled terminals terminals spherical spherical spherical spherical vesicles and vesicles and dense cored dense cored vesicles vesicles # (%) # (%) # (%) # (%) # (%) Non-estrous (n=4) 84 14 (17) 4 (5) 41 (49) 7 (8) 18 (21) Estrous (n=4) 732 67 (9) 33 (5) 398 (54) 45 (6) 189 (26)

In each of the cases, a total area of 320.000µm2 was examined in the same region of the spinal cord.

Non-estrous Estrous

Figure 6 Electronmicrographs showing examples of labeled NRA profiles in the semimembranosus motoneuronal cell group of non-estrous case 2324 (left) and estrous case 2337 (right). In the non-estrous case, a labeled axo-dendritic profile is shown with closely packed spherical vesicles (asterisk), dense-cored vesicles (small arrows), a few mitochondria (m), and an asymmetric synaptic junction (arrowhead). The terminal is located at the initial segment of a dendrite (d), which contains a few cisternae of endoplasmic reticulum with ribosomes (rer). In the estrous case, three large labeled axo-dendritic terminals are present with densely packed spherical vesicles (asterisk), and some dense core (small arrows) and large granulated vesicles (large arrows). Two of them exhibit asymmetrical synaptic membrane specializations (arrowheads). Bar represents 1 µm.

121 Chapter 5

were observed. The labeled terminals mainly contacted labeled dendrites and only very occasionally neuronal somata. terminals In case 2337, the average perimeter of the labeled terminals was 9.49 µm (n=73; ranging from 5.25 to 22.25 7 µm), which is 1.28 times larger than in the non-estrous 6 case 2324 (7.41µm). In these labeled terminals, synaptic 5 complexes on average formed 5.0% of the perimeter, which is less than in the non-estrous case 2324 (6.9%). 4 In the area under examination (the semimembranosus 3 motoneuronal cell group), the absolute number of labeled profiles displaying synapses was much higher 2 in the estrous cases as compared to non-estrous cases 1 (100 and 18, respectively; see Table 2). In contrast, the percentage of labeled profiles showing one or more 0 synaptic junctions was larger in non-estrous cases than Non- Estrous in estrous cases (21% and 14%, respectively). estrous In all estrous cases, but not in non-estrous cases, very large labeled structures were found which contained large quantities of tightly packed mitochondria (297, Figure 7 The average number of labeled profiles/10.000µm2 ± 159, 86, and 32 mitochondria per 35.6, 25.4, 12.9, and ( S.E.M.) in the groups of non-estrous and estrous cases. µ The density of labeled profiles differs significantly (p<0.025; 8.5 m2), and some agranular reticulum, a few coated, Wilcoxon-Mann-Whitney test). dense core and large granulated vesicles, lysosomes and electron dense particles, and microtubuli (Figs. 8 left and 9). The mitochondria had a smaller diameter and were more elongated than the mitochondria in adjacent structures (Figs. 8 and 9). Similar structures have been described as the proximal or central part of neuronal growth cones in the developing central nervous system (Tennyson, 1970; Yamada et al., 1971; Bunge et al.,

AB

Figure 8 Electronmicrographs of labeled structures in the semimembranosus motoneuronal cell group of case 2337 (natural estrus). A shows one or more axonal growth cones (asterisks in A) parts of an axonal growth cone and a terminal profile (arrowheads). The growth cones contain densely packed mitochiondria, and do not form synaptic contacts. The labeled profile in B is a magnification of the labeled terminal profile in A (one section difference). It is filled with spherical synaptic vesicles, dense core and large granulated vesicles, and forms asymmetrical synaptic junctions with two dendrites (arrowheads); d=dendrite; As=astrocyte; Ax=axon. Bars represent 1 µm.

122 NRA projections to lumbosacral motoneurons: estrogen induced sprouting AB

Figure 9 Electronmicrographs representing a labeled growth cone in the semimembranosus motoneuronal cell group of estrous case 2337, for a large part surrounded by astrocytes (As). The mitochondria in the labeled profile have a small diameter, are elongated, and are oriented in or transverse to the plane of the section. Small groups of mitochondria are sequestered within agranular membranes (arrowheads in A). Apart from the mitochondria and smooth membranes, the labeled profile contains a few electron dense bodies, microtubuli (arrow in B), some spherical, dense core, and large granulated vesicles (arrowheads), and an occasional coated vesicle. The majority of the vesicles is located in close proximity to the cytoplasmic membrane. Note that some TMB reaction product appears to be incorporated into the smooth membranes and plasmamembrane. Bars represent 0.5 µm.

1973; Bridgman and Dailey, 1989; Davis et al., 1992), been described to be particularly prominent during which leads to the conclusion that the large labeled periods of synaps growth, might synthesize proteins that profiles in the present material (4.6% of the labeled are important for axonal outgrowth or maturation of the profiles in the four estrous cases) represent the proximal synapatic junction (Steward and Falk, 1986, 1991). parts of growth cones. In conclusion, many of the labeled NRA profiles in Furthermore, labeled, large terminal-like profiles were estrous cases contain growth cones and immature observed, which contained small diameter elongated terminals, which indicate that outgrowth of NRA axons mitochondria, cisternae of agranular reticulum, spherical takes place when the animal is in estrus. synaptic vesicles, dense cored vesicles, large granulated vesicles, vacuoles, coated vesicles, and a few micro- Rubrospinal control series tubules and neurofilaments (Figs. 8 and 10). Similar In 4 cases, WGA-HRP was injected in the red nucleus assemblies of organelles have been described in growth and surrounding tegmentum (Fig. 11). In all these cases, cones and immature terminals (Tennyson, 1970; Yamada the density of anterogradely labeled fibers in the et al., 1971; Bunge et al., 1973; Vaughn and Sims, 1978; intermediate zone of the lumbar enlargement was denser Knyihar-Csillik et al., 1986; Bridgeman and Dailey, than in the motoneuronal cell groups of NRA injected 1989; Peters et al., 1991). The presence of large numbers cases, and was easily visible using brightfield of coated vesicles in the NRA terminals suggests a high illumination. The two large injections resulted in a level of membrane turnover (Rees et al., 1976) which is slightly stronger projection (non-estrous case 2362 and thought to be involved in the formation of axonal estrous case 2363) than the smaller red nucleus injections collaterals (Vaughn and Sims, 1978). At the postsynaptic (non-estrous case 2245 and estrous case 2310). However, site, mitochondria, multivesicular bodies, coated vesicles comparing the estrous with the non-estrous cases with and polyribosomes were found frequently (Fig. 10). similar injection sites did not reveal any difference in Postsynaptically located polyribosomes, which have the density of labeling at the lightmicroscopical level.

123 Chapter 5

A B

C D

Figure 10 Electronmicrographs of labeled NRA axo-dendritic terminals in the semimembranosus motoneuronal cell group in estrous cases 2353 (A), 2337 (B and C), and 2288 (D). In A, a labeled terminal is shown which exhibits 3 asymmetrical complexes (arrowheads) with subsynaptic dense bodies. The profile contains a bundle of neurofilaments (f), densely arranged spherical synaptic vesicles (asterisk), a few dense core vesicles, and numerous coated vesicles, some of which seem to originate from double-membrane particles (cv). The matrix of the mitochondria (m) in the labeled profile is not as dense as that of the mitochondria in adjacent profiles. Postsynaptically, large mitochondria (m), cisternae of agranular reticulum, and free ribosomes (r) are present. B shows a large labeled terminal containing large quantities of spherical synaptic vesicles (asterisk), coated vesicles (cv), dense core and large granulated vesicles, cisternae of smooth membranes (+), and a small cluster of mitochondria (m). The terminal forms asymmetrical complexes with a dendrite (arrowhead) and a dendritic spine (arrow). C demonstrates a large labeled terminal with many densely packed spherical vesicles (asterisk) and elongated mitochondria (m) contacting a dendrite (arrowhead). The arrows indicate extensions of the terminal. In D, a labeled profile is present establishing an asymmetrical synaptic contact (arrowhead) with a small dendrite (d). The terminal contains mitochondria (m), sheets of smooth membranes (+), dense core and large granulated vesicles (arrow), spherical vesicles, and numerous small cisternae of agranular reticulum. Bar represents 1 µm.

124 NRA projections to lumbosacral motoneurons: estrogen induced sprouting Non-estrous natural non-estrous

2245 ovariectomized-non estrogen treated

2362

Estrous ovariectomized-estrogen treated

C2 2310

2363

A0.6 A1.6 A2.5 A3.3A4.1 A5.2 A6.4

Figure 11 Overview of the WGA-HRP injection sites involving the red nucleus in estrous and non-estrous cases. A0.6 to A6.4 indicate the anterior-posterior coordinates according to the atlas of Berman (1968).

DISCUSSION terminals. The finding that the percentage of labeled profiles with one or more synaptic junctions decreases in estrous as compared to non-estrous cases is in line Using a new, detailed overview of the spatial location with this hypothesis. Furthermore, electronmic- of lumbosacral motoneuronal cell groups (VanderHorst roscopical examination of the semimembranosus cell and Holstege, unpublished obser-vations), it could be group in estrous and non-estrous cases showed that the determined with great precision which of these cell labeled profiles contained mainly spherical and dense groups received NRA afferents. cored vesicles and formed asymmetrical synapses, The present results demonstrate that the density of the pathway from the NRA to hindlimb motoneuronal cell suggesting a mainly excitatory role for the NRA- groups shows significant estrogen related differences lumbosacral pathway (see Holstege, 1989). The NRA in adult female cats. The electronmicroscopical results profiles can be classified as S- or NFs-type (Conradi, confirmed the lightmicroscopical observations, 1969; McLaughlin, 1972). demonstrating an almost ninefold increase in the number Summarizing, the number of probably excitatory NRA of NRA profiles in estrous cases. These major profiles in the semimembranosus motoneuronal cell differences cannot be fully explained by the small group increases enormously in estrous cases as differences in the size of NRA profiles in estrous and compared to non-estrous cases. The presence of labeled non-estrous cases (estrous versus non-estrous= 1.28: 1). growth cones in estrous (natural as well as ovariec- The finding of labeled axonal growth cones in the tomized-estrogen treated), and the absence of such semimembranosus motoneuronal cell group in estrous structures in non-estrous cats (natural as well as cats, which were never found in any of the non-estrous ovariectomized) suggests that the difference in density cats, demonstrate that the difference in number of labeled in the NRA-lumbosacral projection is based on estrogen- profiles is probably based on the formation of new NRA induced outgrowth or sprouting of NRA axons (Fig. 12).

125 Chapter 5

Possible mechanisms underlying sprouting of The question is whether a similar mechanism underlies NRA-axons axonal outgrowth in the NRA-lumbosacral pathway. The finding that estrogen induces outgrowth of the NRA- Neither the NRA nor its target motoneuronal cell groups lumbosacral pathway leads to the question about which contain estrogen receptors or concentrate estrogen mechanisms are involved in this process. Estrogen has (Stumpf et al., 1975; Stumpf and Sar, 1976; Morrell et been demonstrated to affect neurons by two different al, 1982; Rees et al., 1980; Meijer, VanderHorst and mechanisms: induction of protein synthesis via genomic Holstege, unpublished observations). However, in the activation and changing the membrane excitability. rat the lateral PAG contains estrogen concentrating neurons some of which project to the caudal medulla Genomic activation via intracellular estrogen receptors oblongata (Corodimas et al., 1990). In the cat, estrogen Estrogen induced protein synthesis is a relatively slow receptor containing PAG neurons target the NRA response (hours to days; Pfaff and McEwen, 1983; Clark (VanderHorst, Meijer, and Holstege, unpublished and Mani, 1994; Pfaff et al., 1994). It is mediated via observations). Other estrogen concentrating structures intracellular receptors, which, when bound to estradiol, in the forebrain, such as the ventrolateral part of the activate a specific DNA target (Halachmi et al., 1994). ventromedial hypothalamus, medial preoptic area, and The resulting newly synthesized proteins are transported amygdala, do not have direct connections with the NRA down the axon (Pfaff and McEwen, 1983; Pfaff et al., (Holstege, 1991). Possibly, the estrogen-concentrating 1984; Mobbs et al., 1988), where they are thought to be neurons in the PAG play a major role in preparing the involved in plastic changes (see Pfaff et al., 1994). NRA-lumbosacral pathway for its specific action during Estrogen-related plasticity has been described in cell lordosis behavior. populations containing intracellular estrogen receptors Another possibility is that estrogen exerts an effect on (Stumpf et al., 1975; Stumpf and Sar, 1976; Pfaff and the NRA-motoneuronal pathway via the muscles Keiner, 1973; Rees et al. 1980). These cell groups have involved (see Haydon and Zoran, 1994 for review). For a facilitating effect upon lordosis behavior in adult example, in humans it has been demonstrated that the mammals. Examples are the lateral septum (Miyakawa pelvic floor muscle levator ani, but not the rectus and Arai, 1987), ventrolateral part of the ventromedial abdominis muscle, contain estrogen receptors (Smith hypothalamus (Carrer and Aoki, 1982) and PAG (Chung et al., 1990, 1993). This would correspond with the et al., 1988, 1990). Such plastic changes occur within finding that, at least in the cat, the motoneuronal cell 24-48 hours, in parallel with the estrous cycle in the rat groups innervating pelvic floor muscles receive NRA and hamster (Meisel and Luttrell, 1990; Frankfurt et al., afferents (Holstege and Tan, 1987; VanderHorst and 1990; Frankfurt and McEwen, 1991; Olmos et al., 1989; Holstege, 1995), whereas the rectus abdominus Langub et al., 1994). motoneurons are not targeted by the NRA (Holstege,

Non-estrous Estrous NRA-axon NRA-axon

terminal growth cone

motoneuron

Figure 12 Schematic illustration of the estrogen induced axonal sprouting of NRA fibers to lumbosacral motoneurons. The number of terminals reflects the almost ninefold difference between estrous and non-estrous cases.

126 NRA projections to lumbosacral motoneurons: estrogen induced sprouting

1989). Possibly, under high levels of estrogen, estrogen receptor containing muscles send a retrograde signal to their motoneurons, which, in turn, induce outgrowth of NRA axons.

Estrogen effects on membrane excitability Estrogen has also been shown to change the membrane excitability of neurons within minutes (see Alcaraz et al., 1969; Yagi, 1970; Kelly et al., 1977; Dufy et al., 1979; Levesque and Di Paolo, 1988; Schumacher, 1990; Smith, 1994). This effect seems to be mediated via membrane receptors (Schumacher, 1990) which affect calcium channels (Mermelstein et al., 1996). Such rapid effects have been described for numerous cell groups, such as the anterior hypothalamus (Kawakami et al., 1970; Cross and Dyer, 1972; cat, Alcaraz et al., 1969), medial amygdala (Nabekura et al., 1986; Minami et al., 1990), nucleus accumbens (Thompson and Moss, 1994), hippocampus (Wong and Moss, 1991, 1992), and cerebellum (Smith et al., 1987, 1989). Whether this mechanism plays a role in the NRA axonal outgrowth is not known.

Functional implications of estrogen induced axonal sprouting in the NRA-lumbosacral pathway Estrogen has been shown to induce growth or plasticity in mid- and forebrain structures (Meisel and Luttrell, 1990; Frankfurt et al., 1990; Frankfurt and McEwen, 1991; Olmos et al., 1989; Langub et al., 1994; Miyakawa and Arai, 1987; Carrer and Aoki, 1982; Chung et al., 1988, 1990), as well as in non-neuronal structures as the uterus (Burrows, 1949; Reynolds, 1951; Clark and Mani, 1994 for review). However, the present study is the first to show that estrogen induces the outgrowth of a long brainstem-spinal motor pathway, which is thought to represent the final common pathway for lordosis. The effect of estrogen on this pathway seems to be rather specific, because no estrogen-related differences could be detected in the rubrospinal motor tract. These findings are in line with the notion that estrogen or other sex steroids are necessary for activation of the reproductive neural circuitry, which appears to be latently present in non-estrous animals.

127

General discussion

General discussion

Perineum, Vagina, Cervix

129

General discussion General discussion

In this thesis, a new concept is put forward for the basic However, posture is only one of the components of circuitry (from spinal cord to brainstem to spinal cord) reproductive behavior. Female receptive behavior for for mating behavior in the cat. Relevant stimuli are example is accompanied by immobility, changes in conveyed directly from the lumbosacral cord to the PAG, nociception, cardiovascular responses, vocalization, and where this information is integrated with input from the pupil dilatation. The present chapter discusses the basic forebrain (Fig. 1). Via interneurons in the NRA, the PAG neural circuitry for receptive and mounting behavior in activates a specific set of lumbosacral motoneurons. The the context of these other behavioral components, muscles innervated by these motoneuronal cell groups estrogen effects, and species differences. In addition, are involved in the receptive posture in females and in the role of the NRA as relay for the emotional motor the mounting posture in males. system is discussed.

Female Male

Cd

F

PU BNST STT STT Cl CA Cd AD Cd AD LD CL LD SM SM AV Medial preoptic Area VA AM Pt SON R PVA

VB VL OC

RE Pu ML GP

En PVN CA OT AL BL VL Ventromedial BM CO Hypothalamus

SC SC

IC Periaqueductal IC Periaqueductal Gray Gray BC IV BC IV

CSN CSN BP ML BP ML

PC PC PON PON

CUN CUN G G Vspin. Nucleus Vspin. Nucleus caud. caud. Retroambiguus Retroambiguus NRA NRA

P P

L4-S1 L4-S1 Iliopsoas, Iliopsoas, Adductor, Adductor, Hamstring, Hamstring, Axial, and Axial, and Pelvic floor Pelvic floor muscles muscles

S2 S2 Perineum, Vagina, Perineum Cervix

Figure 1 Schematic representation of the basic neural circuitry for female receptive (left) and male mounting posture (rigth).

131 General discussion

A B

C D

E F

G H

Figure 2 Photomicrographs of mating cats. The male shows the neck grip (A), palpation of the flanks (B), and treading (C and D). The female displays receptive behavior, which consists of a crouching posture (A to E), lateral deviation of the tail (C and D), treading (D), and elevation of the lower back (A toD). After ejaculation of the male, the female emits the “estrous cry”, becomes very aggressive (E and F), and starts rolling (G) and licking (H). This behavior is called the “after-reaction”.

132 General discussion

Female and male sexual behavior in the cat of receptive behavior and marks the transition between receptive behavior and the after reaction. The “estrous During mating, male and female cats display their own cry” can only be evoked after vigorously stimulating repertoire of motor behaviors (Michael, 1961; Whalen, the vaginocervix. Normally, this is achieved during and 1963). In order to attract a male, an (pro)estrous female after ejaculation by the penis of the male cat, which bears displays so called proceptive behavior, which consists multiple little spines. of calling (vocalization), and increased locomotor activity. Olfactory stimuli from the vaginal fluid of the Analgesia, blood pressure changes, and pupil dilatation estrous female trigger male mounting behavior, which Apart from the lordotic posture and immobility, receptive starts with the neck grip (Fig. 2A). The neck grip behavior is also characterized by less overt reactions immobilizes the female, and allows the male to mount such as analgesia (Komisaruk and Wallman, 1977; the female. Mounting is characterized by grasping with Catelli et al., 1987; Gomora et al., 1994), blood pressure the forelimbs (forelimb rubbing Fig. 2B), hindlimb changes (Catelli et al., 1987), pupil dilatation (Szecht- stepping (Fig. 2C), arching of the back, stamping on man et al., 1985), and endocrine responses. All these the female cat’s rump (palpation of the flanks), and reactions can be evoked by natural or artificial finally pelvic thrusting. These sensory stimuli induce vaginocervical stimulation (Szechtman et al., 1985; the receptive posture in the female (Fig. 2A to D). In Catelli et al., 1987; Gomora et al., 1994). Mounting alone the female cat the full receptive posture consists of has been shown to be far less potent (Komisaruk and crouching (forelegs collapsed), lowering of the head, Wallman, 1977; Szechtman et al., 1985; Gomora et al., perineal elevation (lordosis), tail deviation, and treading 1994). The responses elicited by vaginocervical (rhythmical movements of the hindlimbs). This posture, stimulation are mediated via the pelvic nerve, because also called lordosis, enables intromission by the male. bilateral transection of this nerve (and hypogastric nerve; Receptive-mounting behavior usually takes several see Cunningham et al., 1991) abolishes them (Szechtman minutes and is continued until ejaculation. At that et al., 1985; Gomora et al., 1994). moment, the female emits the “estrous cry” (a distinctive form of vocalization) and turns very aggressively Immobility Immobility of the female cat can be evoked towards the male (Fig. 2E and F). She starts rubbing by vaginocervical stimulation (Naggar and Komisaruk, and rolling over the floor, and vigorously licks the 1977 in the rat). However, before mounting, the male perineal area, the tip of the tail and the toes. This so- cat licks or even bites the female in the neck, which called “after reaction” lasts for 15-20 minutes, after also immobilizes her (Whalen, 1961). Afferents which the male is allowed to mount again. Matings can conveying this sensory input from skin and muscles of continue for many hours. the neck terminate in the upper cervical cord (Abrahams et al., 1984). The basic neural circuitry for female receptive behavior Endocrine responses Vigorous stimulation of vagina and cervix during intromission by the male also triggers endocrine responses, such as the release of prolactin and I. Different components of receptive behavior and oxytocin from the pituitary (for review see Komisaruk their relevant spinal input (Fig. 3a and e) and Steinman, 1986). Prolactin in turn stimulates the Receptive posture Receptive behavior is initiated easily release of the ovarian hormone progesteron, which in freely moving animals by applying tactile stimuli to prepares the uterus for implantation. Oxytocin provokes the skin of the flanks, posterior rump, tailbase, and uterine contractions, during parturition known as the perineum (e.g., for the rat: Kow et al., 1979; for the cat: Ferguson reflex. In reflexive ovulators, such as cat and Michael, 1961). Going from the flanks to the perineum, rabbit, coital stimulation induces ovulation by the these stimuli have an increasing effect on the strength pulsatile release of luteinizing hormone-releasing of the lordosis response. The strongest response can be hormone (LH-RH, also called gonadotrophin releasing elicited by vaginocervical stimulation. Under natural hormone or GnRH), followed by the release of conditions, all these stimuli are applied by the mounting luteinizing hormone (LH) (Concannon et al., 1980; male. The lordosis relevant sensory input is conveyed Banks and Stabenfeldt, 1982; for review see Komisaruk from the perineal skin, vagina and cervix to the and Steinman, 1986). This mechanism can be abolished lumbosacral cord mainly through the pelvic and by transection of the pelvic nerves (Wildt et al., 1980; pudendal nerves (Morgan et al., 1981; Ueyama et al. Cunningham et al., 1992). 1984 ; Szechtman et al., 1985; Thor et al., 1989; Gomora et al., 1994). II. Spino-PAG pathways (Fig. 3b) Manual vaginocervical stimulation of estrogen primed Vocalization Althoug vocalization can be considered cats with a hemisection in the C2 segment of the spinal as an independant emotionally driven motor activity, the cord results in “hemi-lordosis” (unpublished obser- “estrous cry” is very distinctive. It is a typical component vations), which is characterized by extension of the lower

133 General discussion

a. Neck grip Vaginocervical stimulation

b. Upper cervical cord Lumbosacral cord

c. Ventrolateral PAG Lateral PAG

Ventral pontine and medullary Rostral ventrolateral Nucleus d. medial tegmentum medulla retroambiguus

Sympathetic preganglionics Sensory neurons Motoneurons and Sympathetic preganglionics Motoneurons of the Motoneurons of the T1-T2 in the in the premotor interneurons in the larynx, iliopsoas, intermediolateral intermediolateral dorsal horn in the ventral horn intermediolateral pharynx, adductor longus, cell column cell column and intermediate zone cell column soft palate, and hamstring, expiratory muscles pelvic floor, and axial muscles

General level of Opioid General level of Cardiovascular changes Vocalization Receptive Pupil dilatation ? e. sympathetic activity analgesia motor activity posture

Figure 3 Scheme showing the different components of receptive behavior and the pathways which are possibly involved.

back and treading of only the hindlimb located Wiberg and Blomqvist, 1984; Yezierski, 1988; contralateral to the hemisection. These findings indicate VanderHorst and Holstege, 1992; VanderHorst et al., in that receptive behavior is not mediated at the level of press; chapter 1; Mouton, VanderHorst and Holstege, in the spinal cord, but requires supraspinal control. Thus, prep.). Spino-PAG neurons are not evenly distributed in order to evoke receptive behavior, the information throughout the spinal cord. Especially the upper cervical from the genital area and the neck has to be conveyed to and sacral segments send a major projection to the PAG supraspinal levels. (VanderHorst and Holstege, 1992; Mouton, VanderHorst Without forebrain structures (after precollicular and Holstege, in prep.). Spino-PAG neurons in the upper decerebration), receptive behavior can still be elicited cervical segments might very well convey stimuli from by somatosensory and vaginocervical stimulation in the “neck-grip” to the PAG. At sacral levels, the location ovariectomized cats and rats (Bard and Macht, 1958; of spino-PAG neurons matches precisely the location Rose and Flynn, 1993). Vaginocervical stimulation of pelvic primary afferents (VanderHorst and Holstege, selectively increases the metabolic activity in the rat PAG 1992; VanderHorst et al., in press; Chapter 1), which (Allen et al., 1981), and activates the intermediate-early implicates that these convey information relevant for gene c-fos in this area (Pfaus et al., 1993; Tetel et al., receptive behavior to the PAG. 1993). Lesions in the PAG abolish receptive behavior, which demonstrate that this midbrain structure is III. Integration in the PAG (Fig. 3c) essential for an integrated pattern of this behavior After the information important for receptive behavior (Sakuma and Pfaff, 79b). has converged in the PAG, different behavioral Physiological studies have shown that the PAG receives components have to be combined into an integrated lordosis relevant somatosensory input (Pfaff and behavioral pattern. Especially the lateral and the Schwartz-Giblin, 1988 for review). Theoretically, this ventrolateral PAG are of importance for this task. information reaches the PAG indirectly or directly. A possible indirect pathway involves the nucleus tractus The lateral PAG solitarius (NTS) which responds to stimulation of the Receptive posture The sacral-PAG neurons that convey cervix, uterus, and vagina (Hubscher and Berkley, 1994; pelvic visceral information, have axons terminating 1995). Via projections from the NTS to the PAG, this medially in the lateral part of the caudal PAG information might reach the PAG (Herbert and Saper, (VanderHorst et al., in press; chapter 1). In estrogen 1992). The NTS receives vaginocervical input via a primed rats, lesions in the same area of the PAG have spinal cord pathway, and not via the vagal nerve been shown to decrease lordosis posture (Sakuma and (Hubscher and Berkley, 1995). However, after tracer Pfaff, 79b), whereas stimulation in the lateral and dorsal injections into the sacral cord, only a very sparse PAG facilitates lordosis (Sakuma and Pfaff, 1979a). projection is present in the NTS (unpublished obser- vations), suggesting a minor or no role for this indirect Estrous cry, aggressive behavior, blood pressure pathway in receptive behavior. changes, and pupil dilatation The lateral PAG is not Anatomical tracing studies in the rat, cat, and monkey only important for receptive posture. Stimulation in the have demonstrated that the PAG receives direct lateral part of the lateral PAG evokes aggressive and projections from the spinal cord (Menetrey et al., 1982; defensive behavior, which is accompanied by hyper-

134 General discussion tension, vocalization, tachycardia, pupil dilatation, and Motor system non-opioid analgesia (Sakuma and Pfaff, 1979a; Bandler Voluntary motor Emotional motor and Depaulis, 1991). Therefore, it is very well possible system system that the lateral PAG mediates the increase in blood Lateral Medial Lateral Medial pressure, pupil dilatation, as well as the estrous cry and eye, neck, specific axial and proximal emotional the sudden hostile reaction displayed by the female after body movements behaviors mating. The vigorous stimuli from the spined penis of independent gain setting systems movements of including triggering the male, which are conveyed from the sacral cord to the extremities mechanisms of rhythmical and other the lateral PAG, are likely to initiate the sudden change spinal reflexes from immobile and receptive behavior to aggressive Basic system (premotor interneurons) reactions, including the “estrous cry”.

The ventrolateral PAG Immobility and opioid dependant analgesia In the most caudal PAG, sacral cord neurons terminate in the Motoneurons ventrolateral area (VanderHorst et al., in press; Chapter 1). Stimulation in this region has been shown to elicit Figure 4 Schenatic overview of the three subdivisions of the quiescence and hyporeactivity, accompanied by motor system (from Holstege, 1996). hypotension, bradycardia, and opiod dependant analgesia (Bandler et al., 1991; Bandler and Depaulis, 1991; Carrive and Bandler, 1991; Lovick, 1993). The on motor output via the diffuse projections of the medial ventrolateral PAG also receives input from the upper emotional motor system (see Fig. 4). Both the specific cervical cord (Keay and Bandler, 1992; Mouton et al., and diffuse pathways play a role in reproductive in prep.). Therefore, this part of the PAG might be behavior. involved in the immobility evoked by the neck grip and vaginocervical stimuli. Specific pathways belonging to the lateral emo- Since analgesia evoked by vaginocervical or utero- tional motor system cervical stimulation is opioid dependant (Watkins et al., Direct PAG-spinal projections The shortest way for 1984; Hill, 1980; Gintzler and Komisaruk, 1991), this the PAG to reach the spinal cord is by means of a direct component also might be integrated in the ventrolateral pathway. Recently, direct PAG-spinal projections have PAG. been described from the ventrolateral and lateral intermediate PAG to the medial ventral horn mainly at Endocrine responses Coital stimulation triggers the the level of the cervical and upper thoracic cord (Mouton release of prolactin, oxytocin, and luteinizing hormone and Holstege, 1994). The function of this projection is from the pituitary. A direct sacral cord-hypothalamic not known, but it might play a role in the control of pathway (Burstein et al., 1990; Katter et al., 1991) might axial muscles during threat display (arching of the back; be involved in these mechanisms. However, since this Mouton and Holstege, 1994). Since only a limited tract is only very small in the cat (Katter et al., 1991), it number of fibers continue into the lumbosacral cord, is more likely that an indirect sacral cord-hypothalamic containing the motoneurons of the muscles involved in pathway is responsible for these endocrine responses. receptive behavior, it is not likely that the reproductive Possibly the PAG plays the role of intermediate structure. posture is mediated by this direct PAG-spinal projection. Moreover, it has been shown that LH-RH-containing However, this pathway might reach lordosis moto- axons of hypothalamic neurons are present in the neurons via propriospinal interneurons in the cervical ventrolateral PAG, immediately adjacent to the aqueduct cord. (Liposits and Sétáló, 1980; Shivers et al., 1983; Veening Another direct PAG-spinal projection originates from et al., 1991). Injection of LH-RH in the PAG facilitates the lateral PAG and projects directly to preganglionic lordosis behavior (Riskind and Moss, 1979; Sakuma and sympathetic motoneurons in the intermediolateral cell Pfaff, 1980; 1983; Sirinathsinghji, 1984; Pfaff et al., column of the upper thoracic cord (Holstege, 1988; see 1994 for review). This facilitating effect of LH-RH could Fig. 3). This pathway has been hypothesized to mediate also be observed in hypophysectomized animals (Pfaff, pupil dilatation (Holstege, 1988). 1973), which indicates that it is not caused by the release of LH from the pituitary. PAG-NRA-motoneuronal projections The NRA receives a specific and direct projection from the lateral, the IV. PAG-spinal pathways involved in receptive dorsal, and the caudal ventrolateral PAG (Holstege, behavior (Fig. 3d) 1989; VanderHorst and Holstege 1996; introduction) and The PAG controls specific motor patterns by means of in turn projects directly to a selective set of axial, distinct pathways belonging to the lateral emotional hindlimb, and pelvic floor motoneuronal cell groups motor system, whereas it has a global, modulatory effect (VanderHorst and Holstege, 1995; in press; chapters 3

135 General discussion

and 4; see Fig. 1). Taking into account the function of In addition, the ventrolateral PAG affects the spinal the muscles innervated by these motoneurons, the NRA dorsal horn and intermediate zone, via the ponto- pathway cannot be involved in jumping and running, medullary raphe magnus and adjacent tegmentum. The but is suitable for receptive behavior (chapter 3). In projections to the dorsal horn are well known to be conclusion, the anatomical findings indicate that the involved in opioid dependant analgesia (Oliveras et al., NRA forms a relay between the PAG and a distinctive 1974;1975; Fields and Basbaum, 1978). Similar to the set of lumbosacral motoneuronal cellgroups. Currently, modulatory effect of the ventral medullary medial electromyographical studies are underway to test the tegmentum on motoneurons in the ventral horn, neurons hypothesis that this pathway indeed controls receptive in the raphe magnus and adjacent tegmentum modulate behavior. nociceptive transmission in the dorsal horn (for review The NRA also projects to motoneurons innervating the see Mason and Leung, 1996). Thus, ventrolateral PAG- larynx, pharynx, soft palate, and expiratory muscles pontomedullary tegmentum-dorsal horn pathway might (Holstege, 1989). Thus, the PAG-NRA pathway also mediate the opioid dependant analgesia induced by controls vocalization (Holstege, 1989; Zhang et al., vaginocervical stimulation. 1992; Davis et al., 1996), which includes “calling” and the “estrous cry”. Effect of estrogen on the basic neural The rostral ventrolateral medulla The lateral PAG is circuitry for receptive behavior known to send excitatory projections to the subretro- facial nucleus, also called rostral ventrolateral medulla (Lovick et al., 1984; Lovick, 1985; Li and Lovick, 1985), Mating behavior is dependant on sex steroids. The which in turn projects to sympathetic preganglionics in female is able to display her repertoire only when she is the intermediolateral cell collumn of the thoracic and in estrus, i.e. when high levels of estrogen are present upper lumbar cord (Loewy and McKellar, 1981; Loewy and ovulation is ready to take place (Fig. 5). Estrogen is mainly produced by the ovaries, from where it reaches et al, 1981). The sympathoexcitatory neurons in the the entire body via the bloodstream. rostral ventrolateral medulla determine the level of In the CNS, estrogen induces protein synthesis in sympathetic outflow to different vascular beds (Lovick, estrogen concentrating cell groups (see DeBold and 1987; Dampney and McAllen, 1988). Thus, the lateral Malsbury, 1983; Meisel and Pfaff, 1984; 1985; McEwen, PAG-rostral ventrolateral medulla-spinal cord pathway 1988), but it can also change the excitability of might mediate the cardiovascular changes evoked by membranes of neurons in general (for review see Smith, vaginocervical stimulation (Catelli et al., 1987). 1994). Estrogen concentrating cells are most most abundant in forebrain areas. Examples of estrogen Diffuse pathways belonging to the medial emo- concentrating neuronal cell groups in the forebrain are tional motor system the medial and central nuclei of the amygdala, the medial Ventral pontine and medullary medial tegmentum Both division of the bed nucleus of the stria terminalis, the the ventrolateral and lateral PAG reach motoneurons in medial preoptic area, anterior hypothalamus, and the spinal cord indirectly by way of projections to the ventromedial hypothalamus (VMH) (in the rat: Pfaff and ventral medullary medial tegmentum (Sakuma and Pfaff, Keiner, 1973; Stumpf et al., 1975; in the mouse: Stumpf 1980; Lovick, 1993; Lovick, 1996). In contrast to the and Sar, 1975; in the cat: Rees et al., 1980; in the specific NRA-spinal projections, the pathway from the monkey: Keefer and Stumpf, 1975). These structures ventral medullary medial tegmentum, including the are all involved in the various aspects of sexual behavior caudal raphe nuclei, terminates diffusely in the ventral (Pfaus et al., 1993; Tetel et al., 1993), and all project to horn. This pathway, therefore, may play a role during the PAG (see Saper et al., 1976; Saper et al., 1976; reproduction, not by reflexively activating specific sets Hopkins and Holstege, 1978; Holstege , 1987b; Veening of motoneurons (Sakuma and Pfaff, 1980), but by et al., 1991; Shipley et al., 1991). Thus, via these modulating their general excitability via the release of pathways estrogen might affect the basic neural circuitry glutamate (extracellularly), serotonin, norepinephrine for receptive behavior. and associated neuropeptides (White et al., 1996; Since receptive behavior can also be evoked without Holstege, 1996; Lovick, 1996). input from the forebrain (Bard and Macht, 1958; Rose This ventral medullary medial tegmentum-spinal and Flynn, 1993), this chapter focusses on the effects of pathway might also play a role in an increase in estrogen on the motor and sensory pathways of the basic rhythmical (non-goal directed) motor activities (see circuitry itself. First the ventromedial hypothalamus is Holstege, 1996), such as running, rolling, and rubbing discussed, because of its close relation receptive during proceptive behavior and the after reaction. behavior in general. However, the PAG itself does not seem to trigger these latter responses, since precollicular decerebration has The ventromedial hypothalamus Especially the been demonstrated to abolish them (Bard and Macht, ventrolateral part of the VMH is well known for its role 1958). in reproductive behavior. Stimulation and lesions in the

136 General discussion

Female Male Food availability Social stress Priming pheromones Photoperiod Temperature

Pulsatile release of LH-RH from the hypothalamus

FSH and LH release from the pituitary

Estrogen from the ovaries

Central nervous system Striated muscles Reproductive organs Estrogen-concentrating Estrogen-concentrating Estrogen-concentrating neurons muscle fibers cells Increased excitability Genomic activation Genomic activation Genomic activation of neurons

VMH, Protein synthesis Protein synthesis in Proliferation MPO, in estrogen-concentrating NRA target muscles AH, PAG-NRA neurons BNSTm, anterograde growth retrograde growth Amygdala, transport factors? transport factors? Dorsal horn NRA NRA target motoneurons

Growth of NRA-axons towards their target motoneurons

Neural cicuitry for receptive behavior Pheromones from vaginal fluid Receptive behavior Mounting behavior

Mating

Vaginocervical stimulation

Prolactin Phasic release of LH-RH Sperm from the pituitary from the hypothalamus Progesteron from the ovaries LH surge Maturation uterus Reflex ovulation

Fertilization and Implantation

Figure 5 Schematic overview of a concept for the mechanisms necessary for succesful mating.

VMH respectively facilitate and suppress lordosis (Pfaff evoking receptive behavior (Pfaff and Sakuma, 1979a). and Sakuma, 1979a,b for the rat; Mathews and Edwards, Apparently, the VMH is important for estrogen priming, 1977; Malsbury et al., 1977 for the hamster; Leedy and whereas the PAG is the final integrator for receptive Hart, 1985 for the cat). However, lesions in the VMH behavior, in the same way as it is for vocalization (Davis do not abolish PAG-facilitated lordosis in estrogen et al., 1996). primed rats (Pfaff and Sakuma, 1979b), but lesions in The question remains how the estrogen sensitive VMH the PAG render VMH stimulation unsufficient for acts on PAG neurons. VMH neurons specifically project

137 General discussion

to the medial part of the lateral caudal PAG (Canteras et The nucleus retroambiguus According to Rose and al., 1994). Estrogen affects this pathway by decreasing Sutin, estrogen also increases the responsivity of the stimulation threshold of VMH-PAG neurons “medullary neurons located in the medial lateral reticular (Sakuma and Pfaff, 1982), and by inducing protein nuclei and the region extending dorsomedially to the synthesis (see DeBold and Malsbury, 1983; Meisel and nucleus ambiguus”. These neurons responded at short Pfaff, 1984; 198; Yahr and Ulibarri, 1986; McEwen, latency to PAG stimulation (Rose and Sutin, 1973) and 1988) and transport along the axons (Meyerson, 1982; represent in all likelihood NRA neurons. Since neurons Pfaff et al., 1984). This latter mechanism is consistent in this area do not contain estrogen receptors (un- with the finding that inhibition of protein synthesis published observations by Meijer, VanderHorst and blocks the effect of estrogen on reproductive behavior Holstege), these changes might be mediated indirectly (Meyerson, 1982) by the projection from estrogen concentrating cells in the PAG to the NRA (cat; unpublished observations by Motor pathways Meijer, VanderHorst and Holstege). Periaqueductal gray Estrogen not only increases the Recently, it has been shown that estrogen induces sensory responsivity of hypothalamic neurons (Alcaraz outgrowth of the pathway from the NRA to the et al., 1969; Beyer et al., 1971), but also of mesence- lumbosacral cord, involved in the receptive posture phalic neurons (Beyer et al., 1971). For example, female (VanderHorst and Holstege, see chapter 5). Thus, apart sex steroids strongly facilitate unit responsiveness to from an effect on the responsivity of NRA neurons, lordosis relevant tactile stimulation (Rose and Bieber, estrogen positive PAG-NRA projections might also play 1984), whereas implants of estradiol in the PAG in a role in the sprouting reaction of the NRA-lumbosacral hamsters facilitates vocalization and receptive behavior terminal fibers. (Floody et al., 1986). The PAG is the most caudal structure in the brain that Motoneurons and muscles Since outgrowth of the NRA- contains large quantities of estrogen concentrating lumbosacral axons is directed only to the NRA target neurons (in the rat: Pfaff and Keiner, 1973; Stumpf et motoneuronal cell groups, it is also possible that these al., 1975; in the mouse: Stumpf and Sar, 1975; in the motoneurons attract the NRA fibers. Motoneuronal cell cat: Rees et al., 1980; in the monkey: Keefer and Stumpf, groups themselves do not concentrate estrogen (in the 1975). In the cat, they are mainly located in the lateral rat: Pfaff and Keiner, 1973; Stumpf et al., 1975; Morrell PAG and adjacent tegmentum, and to lesser extent in et al., 1982; in the mouse: Stumpf and Sar, 1975; in the the dorsal PAG (Rees et al., 1980). Thus, apart from an cat: Rees et al., 1980; in the monkey: Keefer and Stumpf, indirect facilitating effect of estrogen via the VMH, the 1975), but some muscles such as the levator ani muscle PAG itself is also estrogen sensitive. PAG-estrogen do (Smith et al., 1990; 1993 in human). In bovine calfs, concentrating neurons do not only exert a local effect, estrogen receptor concentrations in neck, shoulder and but some of them project to the region of the NTS hindlimb muscles did not differ between males and (Corodimas and Morrell, 1990 in the rat) or more females, but in males the concentration in abdominal precisely to the NRA (unpublished observations of wall muscles was lower (Sauerwein and Meyer, 1989). Meijer, VanderHorst and Holstege in the cat). The In rats, no estrogen receptors have been found in the mechanisms by which estrogen affects the PAG-NRA quadriceps (Michel and Baulieu, 1980), which is not pathway might be similar to its effect on the VMH-PAG targeted by the NRA. These findings indicate that projection, i.e. estrogen decreases the stimulation estrogen receptors in muscles might be selectively threshold of the PAG-NRA neurons, and induces protein distributed. Possibly, outgrowth of NRA fibers to distinct synthesis and transport along the PAG-NRA axons. sets of motoneurons is mediated via a retrograde signal, induced by estrogen, from estrogen concentrating The ventral medullary medial tegmentum At the level muscle fibers to their motoneurons. By means of growth of the medulla, estrogen affects the responsivity of factors, these motoneurons in turn might attract afferent neurons in the ventral medullary medial tegmentum fibers, such as the NRA fibers. This mechanism will (Kow and Pfaff, 1982). This effect might play a role in only work if the NRA target muscles, in contrast to other proceptive behavior and the afterreaction, which muscles, contain estrogen receptors. This possibility is behaviors are characterized by a high level of motor presently under investigation. activity. Since the ventral medullary medial tegmentum does not contain estrogen concentrating neurons, the Sensory pathways changes in responsivity might be the result of an effect Sensory peripheral pathways Regarding sensory stimuli of estrogen on the excitability of neuronal membranes relevant for recepetive behavior, the genital sensory field in general (Smith, 1994). A more likely explanation is in female rats has been shown to become larger by that this area is affected indirectly, because estrogen also estrogen administration (Komisaruk et al., 1972), and increases the excitability of the numerous PAG cells that the response sensitivity of pelvic afferent fibers is larger project to the ventral medullary medial tegmentum on the day of proestrus (Robbins et al., 1992; however (Sakuma and Pfaff, 1980b). see Kow and Pfaff, 1979). This effect might be mediated

138 General discussion by estrogen receptors in the dorsal root ganglion be involved. In this concept, the PAG in turn activates (Sohrabji et al., 1994). the somatic motoneurons of the muscles involved in the mounting posture and vocalization via a relay in the Spinal sensory neurons Extracellular recording has NRA (chapter 4). For urine marking, the M-region in demonstrated that estrogen does not significantly affect the pons forms the relay between PAG and/or MPO and the responses of lumbosacral units to lordosis relevant bladder motoneurons (Holstege et al., 1986; Blok and sensory input (Kow et al., 1980). However, laminae I Holstege, 1996). and V throughout the spinal cord contain numerous Adult male cats have constant levels of sex steroids and estrogen concentrating (Morrell et al., 1982), or estrogen are able to mate at any time with an estrous female. receptor containing neurons (Amandusson et al., 1995; This indicates that the neural circuitry for mating in the rat; Meijer, VanderHorst, and Holstege, behavior in the male cat is not subjected to changes. unpublished observations in the cat). These spinal Preliminary results of VanderHorst and Holstege show neurons are sensitive for estrogen and must exert an that there exist no significant differences in number of effect, either at spinal or at supraspinal levels. Possibly, labeled NRA-lumbosacral terminals between castrated that some of these estrogen receptor containing neurons and non-castrated adult males. The strength of the NRA convey lordosis relevant information to the PAG. This projection in the castrated and non-castrated males was option is currently studied. twofold higher than in non-estrous females. These results are in line with common knowledge that castration of adult male house cats is not effective in abolishing urine spraying and mounting behavior. Possibly, after The basic neural cicuitry for male mounting castration of male cats at a younger age, the strength of behavior the NRA-lumbosacral pathway in these males is less that in males that have been exposed to male sex steroids, Male reproductive behavior is evoked primarily by and is similar to that in non-estrous females. This option olfactory or vomeronasal input (Meisel and Sachs, 1994 is currently under investigation. for review). Somatosensory stimuli from the penis, which reach the lumbosacral cord via the pudendal nerve, are important for intromission and ejaculation, Species differences but play a minor role in the mounting posture (see Lucio et al., 1994). Lesion studies have revealed several Behavior important areas in the forebrain which play a role in Reproductive behavior is displayed by all species, but male reproductive behavior. All these areas are involved even between mammalian species there exist consid- in the processing of olfactory input. Bilateral lesions of erable variations in motor patterns. For example, female the olfactory bulbs, affecting both the main and hamsters and rats only display the lordotic posture accessory olfactory systems, or the corticomedial accompanied with hindlimb extension, and do not show nucleus of the amygdala eliminates copulation in rodents treading with the hindlimbs which is so typical for the (Meisel and Sachs, 1994; Giantonia et al., 1970). cat. Furthermore, receptive behavior in the hamster is However, lesions of the olfactory system do not have such effects in cats (Aronson and Cooper, 1974), dogs very tonic and continues for several minutes after ending (Hart and Haugen, 1972), or the rhesus monkey sensory stimulation. In rats, the lordotic posture is more (Goldfoot et al., 1978). In the medial preoptic area, “phasic” and lasts for only a few seconds. electrolytic or neurotoxic lesions in cats and rats (e.g. Heimer and Larsson, 1966; Hart et al., 1973; Kelley and Pathways Pfaff, 1978; van de Pol and Dis, 1979; Hart and Leedy, Differences in the organization of the CNS must underly 1983; Meisel and Sachs, 1994) have been shown to affect these species differences in behavioral patterns. Indeed, or disrupt male sexual behavior, whereas electrical subtle, but significant species differences appear to exist stimulation facilitates copulatory behavior (rat: for the organization of the NRA-spinal, as well as for Malsbury, 1971; van Dis and Larsson, 1971; opossum: the spino-PAG pathways. Roberts et al., 1967). The MPO appears not only to be The NRA has been identified in humans (Olszewski and involved mounting, but also in other aspects of male Baxter, 1954), monkey (Holstege, unpublished mating behavior such as mating vocalization (in observations), cat (Merrill, 1970; Holstege, 1989), rat Mongolian gerbils; Holman et al., 1991 ; in mice: Nyby (Paxinos and Watson, 1986; Ellenberger and Feldman, et al., 1992) and urine marking (in cats: Hart and Voith, 1990), hamster (Gerrits et al., in prep.), and bird (Wild, 1978; in mice: Nyby et al., 1992). 1993; Fig. 6). In cat, monkey, and human, the NRA is It is not known which structures are involved in relatively large and clearly protrudes into the white mounting behavior at the level of the midbrain. Since matter. In hamster and rat, the NRA is less pronounced, the MPO is known to send a major projection to the consisting of a rather diffuse group located among other PAG (Saper et al., 1978; Holstege, 1987), the PAG might neurons. Physiologically, cervical stimulation in female

139 General discussion

rats has been shown to elicit activation of the psoas major and iliacus muscles (Martinez-Gomez et al., 1992), Pigeon XII Dh NRA which are also involved in the lordosis posture in cats. LCN Ssp Tracing studies in this laboratory indicate that the NRA- P lumbosacral pathway exists in the hamster and rat (Kerstens et al., 1996; Gerrits et al., 1996). Moreover, XII Dh in rat, cat, and hamster, these pathways seem to be Finch LCN NRA organized differently, which is in line with the concept P Ssp that differences in sexual behavior of these animals are caused by different NRA-lumbosacral motoneuronal projections. Hamster Regarding the sensory pathways involved in reproduc- Vspin. caud NRA tive behavior, lumbosacral-PAG projections have been demonstrated in monkey (Wiberg et al., 1987; Yeziersky, P LRN 1988; Zhang et al., 1990), cat (Wiberg and Blomqvist, Rat 1984a; Yeziersky, 1988; VanderHorst et al., in press; G Chapter 1; see Fig. 7), rat (Yeziersky, 1988), hamster Vspin. Cun NTS caud NRA (Gerrits et al., in prep.). In the monkey and to a lesser XII extent in the cat (see chapter 1), this pathway terminates P LRN in two distinct regions within the PAG, i.e. in the lateral part of the lateral PAG and adjacent tegmentum, and in Cat the medial part of the lateral PAG. In contrast, in the rat and hamster the projections involve the entire lateral Cun Vspin. G PAG and show no subdivisions. caud NRA Plasticity P In the rat and hamster, with estrous cycles of a few days, synaptic plasticity has been demonstrated in the ventromedial nucleus, arcuate nucleus, and the preoptic area which occurred in parallel with the time interval of Monkey the estrous cycle (Meisel and Luttrell, 1990; Frankfurt et al., 1990; Frankfurt and McEwen, 1991; Olmos et G al., 1989; Langub et al., 1994). In the cat, with estrous Cun intervals of 2-3 weeks up to a few months, plastic Vspin. caud changes might be more prominent and might play a more important role than in animals with very short cycles. NRA The estrogen induced sprouting in the NRA-lumbosacral pathway (chapter 5) supports this hypothesis. Seasonal P breeders display their mating behavior only once a year, depending on the amount of light, the temperature, the availability of food and water, and social cues such as primer pheromones and the absence of stress (for review see Bronson and Heideman, 1994). These external cues Human influence neuronal cell groups in the hypothalamus,

Figure 6 Schamatic drawings of transverse sections through the lower brainstem, showing the NRA in the pigeon and the Vspin. G caud finch (Wild, 1993), hamster (Gerrits, VanderHorst and Holstege, in prep.), rat (Holstege, Kerstens, Moes, and VanderHorst, in press), cat (VanderHorst and Holstege, 1995), NRA monkey (Holstege, unpublished), and human (Olszewski and Baxter, 1954). P Dh=dorsal horn; XII=Nu. hypoglossus; P=pyramidal tract; Ssp CL=lateral cervical nucleus; Ssp=supra-spinal nucleus; LRN=lateral reticular nucleus; NTS=nucleus of the solitary tract; G=gracile bucleus; Cun=cuneate nucleus; Vspin. caud=caudal spinal trigeminal complex

140 General discussion

RAT CAT MONKEY

Figure 7 Termination pattern of lumbosacral-PAG projecting neurons in the rat, cat, and monkey (from Yezierski, 1988). Note that the projection is most diffuse in the rat, and most specific in the monkey. which in turn change the release of LHRH from the neural circuitry for sexual behavior. For example, vaginal pituitary, and thus the production of gonadal hormones stimulation in women, similar to the rat, has been shown (for review see Turek and Van Cauter, 1994). In these to elevate the pain threshold (Whipple and Komisaruk, animals, plasticity as described for the NRA-spinal 1985). The major difference between humans and other pathway in the cat might play a major role in preparing animal species does not concern the basic cicuitry, but the CNS for reproduction. the facilitating and inhibitory effects of structures in the forebrain and cortex. Humans, with their large Human reproductive behavior forebrains, are more capable than other mammals to The question remains how reproductive behavior is control sexual behavior, as well as other emotional organized in humans. In all likelihood, the human behaviors, such as anger. species, like all other species, is equipped with a basic Moreover, in contrast to other species, women are also

141 General discussion

sexually active when the ovaries do not release estrogen, distinct group of retrogradely labeled cells in the medial although gonadal hormones are well known for their part of the lateral PAG (see introduction; VanderHorst effect on mood and performance (see Schreiner-Engel and Holstege, 1996), whereas more rostral injections et al., 1981; Logue and Moos, 1988). Most other resulted in labeled neurons in the lateral PAG as well as mammals only display this potentially risky mating in the adjacent tegmentum (Holstege, 1989). Other behavior when conception is most probable. Humans support comes from physiological studies in which are able to generate a safe environment, and apparently vocalization and lordosis have been elicited by do not need this mechanism anymore. stimulation of the lateral and medial part of the lateral PAG, respectively (Sakuma and Pfaff, 1979a). These results indicate that indeed different groups of PAG- The NRA as multifunctional interneuronal NRA neurons might be involved in the PAG-NRA- cell group for the emotional motor system vocalization pathway and the PAG-NRA-lordosis pathway. Some overlap, however, cannot be excluded.

Multifunctional role of the NRA The NRA as relay for the emotional motor system NRA neurons have been shown to project directly to a In general, spinal motoneurons receive their main input distinct set of brainstem, and spinal cord motor nuclei from interneurons in the spinal intermediate zone. (for review see VanderHorst and Holstege, 1996; see Usually these projections are called propriospinal also Table 2 of chapter 3). The NRA is known to play a pathways. Supraspinal structures as well as afferent input role in expiration (Merrill, 1970;1974), vomiting (Miller from the periphery make use of these interneurons to et al., 1987; 1995), defecation (Fukuda and Fukai, 1986; influence the motoneurons. The propriospinal projec- 1988), and vocalization (Holstege, 1989; Zhang et al., tions take part in the basic motor system in the concept 1992; 1995). The anatomical results of this thesis of Holstege (Holstege, 1991, 1994; see introduction). strongly suggest that the NRA is also involved in Specific, direct projections to distinct motoneuronal cell reproductive behavior. In this context, the NRA appears groups, bypassing the spinal interneurons, are rare. In to form a relay between the PAG and specific sets of the cat examples are Ia afferents from muscle spindles, somatic motoneurons involved in emotional behavior and a very few rubrospinal fibers to C8-T1 digit in general. It is very well possible that the NRA serves motoneurons (Holstege, 1987; McCurdy et al., 1987). even more behaviors, such as the fetus expulsion reflex NRA neurons also project directly to motoneurons, during parturition. In periparturient rats, electrical because they themselves are the premotor interneurons, stimulation of the pelvic nerve has been shown to similar to the ones in the spinal intermediate zone. reflexively activate abdominal wall muscles (Cueva- Premotor interneurons of somatic (hindlimb) moto- Rolon et al., 1995), which reflex might be conveyed via neurons usually receive peripheral Ia afferents or input the sacral-PAG-NRA pathway. The extensor responses from cortico- and rubrospinal tracts, but this is not the of newborn rats during micturition are very much alike case for the premotor interneurons in the NRA. A the lordotic posture (Beach, 1966; Williams, 1987; probable explanation is that the NRA interneurons have Williams and Lorang, 1987), and might be mediated by a special function, i.e. they activate a set of motoneuronal the same pathways. cell groups in the lower brainstem and spinal cord, which The question is how the NRA is organized functionally activation has to be totally independent of voluntary or anatomically, in order to control these functions. input (from cortico- and rubrospinal pathways) or spinal Physiologically, it has been demonstrated that the NRA reflexes. In this context, the NRA receives afferents from is capable to control completely different motor other sources such as the PAG (Holstege, 1989; Davis activities. For example, different populations of NRA and Zhang, 1991; VanderHorst and Holstege, 1996), neurons projecting to Onuf’s nucleus are active during respiratory related neuronal cell groups in the pons and the retching and expulsion phase of vomiting (Miller et medulla, such as the NTS for vagal afferent input, al., 1995). Kölliker Fuse and lateral parabrachial nuclei (Bianchi Anatomically, a somatotopical, rostrocaudal organ- and Barillot, 1975; Feldman, 1986 for review; Smith et ization for NRA-spinal neurons has not been revealed. al., 1989; Gerrits and Holstege, in press), and from two In contrast, it has been shown that some NRA neurons cell groups in the ventral medullary medial tegmental project to both the C4-C6 and the L1-L3 segments field (Gerrits and Holstege, in press). (Portillo et al., 1994). These findings implicate that the NRA neurons are functionally organized, with for example the group of NRA-lordosis neurons located caudal to the group of NRA-vocalization neurons. A similar organization might not only exist for the NRA- motoneuronal projections, but also for the PAG-NRA projections. This idea is supported by the findings that discrete tracer injections in the caudal NRA result in a

142

References References

Abrahams, V. C., S. M. Hilton, and A. W. Zbrozyna (1960) Motoneurons innervating the external anal and urethral Active muscle vasodilatation produced by stimulation of sphincters of the female cat have different patterns of the brain stem: its significance in the defense reaction. J. dendritic arborization. Neurosci. Lett. 111: 69-74. Physiol. (London) 154: 491-513. Bereiter, D. A., and A. P. Benetti (1991) Microinjections of Abrahams, V. C., F. J. Richmond, and J. Keane (1984) calcitonin gene-related peptide within the trigeminal Projections from C2 and C3 nerves supplying muscles subnucleus caudalis of the cat affects adrenal and and skin of the cat neck: a study using transganglionic autonomic function. Brain Res. 558: 53-62. transport of horseradish peroxidase. J. Comp. Neurol. 230: Berkley, K. J., C. H. Hubscher, and P. D. Wall (1993) Neuronal 142-154. responses to stimulation of the cervix, uterus, colon, and Alcaraz, M., C. Guzman-Flores, M. Salas, and C. Beyer (1969) skin in the rat spinal cord. J. Neurophysiol. 69: 545-556. Effect of estrogen on the responsivity of hypothalamic Berman, A. L. (1968) The brain stem of the cat. A and mesencephalic neurons in the female cat. Brain Res. cytoarchitectonic atlas with stereotaxic coordinates. The 15: 439-446. university of Wisconsin Press. Madison, Milwaukee, Allen, T. O., N. T. Adler, J. H. Greenberg, and M. Reivich London. (1981) Vaginocervical stimulation selectively increases Besson, J. -M, and A. Chaouch (1987) Peripheral and Spinal metabolic activity in the rat brain. Science 211: 1070-1072. Mechanisms of Nociception. Physiological Reviews 67: Amandusson, A., O. Hermanson, and A. Blomqvist (1995) 67-186. Estrogen receptor-like immunoreactivity in the medullary Beyer, C., J. Almanza, L. de la Torre, and C. Guzmanflores and spinal dorsal horn of the female rat. Neurosci. Lett. (1971) Effects of genital stimulation on the brain stem 196: 25-28. multiunit activity of anestrous and estrous cats. Brain Res. Angulo y Gonzalez, A. W. (1927) The motor nuclei in the 32: 143-150. cervical cord of the albino rat at birth. J. Comp. Neurol. Bianchi, A. L., and J. C. Barillot (1975) Activity of medullary 43: 115-142. respiratory neurones during reflexes from the lungs in cats. Aronson, L. R., and M. L. Cooper (1974) Olfactory deprivation Respir. Physiol. 25: 335-352. and mating behavior in sexually experienced male cats. Bikeles (1905) Zur Lokalisation im Rückenmark. Deutsche Behav. Biol. 11: 459-480. Zeitschrift für Nervenheilkunde 29: 180-207. Bandler, R., and A. Depaulis (1991) Midbrain periaqueductal Blok, B. F. M., and G. Holstege (1994) Direct projections gray control of defensive behavior in the cat and the rat. from the periaqueductal gray to the pontine micturition In R. Bandler and A.Depaulis (eds): The midbrain center (M-region). An anterograde and retrograde tracing periaqueductal gray matter. Plenum Press, New York pp. study in the cat. Neurosci. Lett. 166: 93-96. 175-198. Blok, B. F. M., H. DeWeerd, and G. Holstege (1995) Bandler, R., P. Carrive, and S. P. Zhang (1991) Integration of Ultrastructural evidence for a paucity of projections from somatic and autonomic reactions within the midbrain the lumbosacral cord to the pontine micturition center or periaqueductal grey: viscerotopic, somatopic and M-region in the cat: A new concept for the organization functional organization. Brain Res. 87: 269-305. of the micturition reflex with the periaqueductal gray as Banks, D. H., and G. Stabenfeldt (1982) Luteinizing hormone central relay. J. Comp. Neurol. 359: 300-309. release in the cat in response to coitus on consecutive days Blok, B. F. M., and G. Holstege (1996) The neuronal control of estrus. Biology of Reproduction 26: 603-611. of micturition and its relation to the emotional motor Bard, P., and M. B. Macht (1958) The behavious of chronically system. In: The Emotional Motor System. G. Holstege, decerebrate cats. Ciba Foundation symposium on the R. Bandler, and C. Saper (eds.). Progr. Brain Res. 107: neurological basis of behavior pp. 267-273 113-126. Barrington, F. J. F. (1925) The effect of lesions of the hind- Blomqvist, A., I. Danielsson, and U. Norrsell (1990) The and mid-brain on micturition in the cat. Quart. J. Exp. somatosensory intercollicular nucleus of the cat’s Physiol. Cogn. Med. 15: 81-102. mesencephalon. J. Physiol. (Lond) 429 : 191-203. Barrington, F. J. F. (1933) The localization of the paths Botterman, B. R., T. M. Hamm, R. M. Reinking, and D. G. subserving micturition in the spinal cord of the cat. Brain Stuart (1983a) Localization of monosynaptic Ia excitatory 56: 126-148. post-synaptic potentials in the motor nucleus of the cat Beach, F. A. (1948) Hormones and behavior: a survey of biceps femoris muscle. J. Physiol. 338: 355-377. interrelationships between endocrine secretions and Botterman, B. R., T. M. Hamm, R. M. Reinking, and D. G. patterns of overt response. New York: Hoeber. Stuart (1983b) Distribution of monosynaptic Ia excitatory Beattie, M. S., Q. Li, M. G. Leedy, and J. C. Bresnahan (1990) post-synaptic potentials in the motor nucleus of the cat

145 References

semitendinosus muscle. J. Physiol. 338: 379-393. Chanaud, C. M., and J. M. Macpherson (1991a) Functionally Boyd, I. A., and M. R. Davey (1968) Composition of peripheral complex muscles of the cat hindlimb. III. Differential nerves. Livingstone, Edinburgh and London activation within biceps femoris during postural Bramble, D.M., and D.R. Carrier (1983) Running and perturbations. Exp. Brain Res. 85: 271-280. breathing in mammals. Science 219: 251-256. Chanaud, C. M., C. A. Pratt, and G. E. Loeb (1991b) Bridgman, P. C., M. E. Dailey (1989) The organization of Functionally complex muscles of the cat hindlimb. II. mysosin and actin in rapid frozen nerve growth cones. J. Mechanical and architectural heterogenity within the Cell Biol. 108: 95-109. biceps femoris. Exp. Brain Res. 85: 257-270. Brink, E. E., J. I. Morrell, and D. W. Pfaff (1979) Localization Christensen, B. N., and E. R. Perl (1970) Spinal neurons of lumbar epaxial motoneurons in the rat. Brain Res. 170: specifically excited by noxious or thermal stimuli: 23-43. Marginal zone of the dorsal horn. J. Neurophysiol. 33: 293-307. Brissaud, E., and A. Bauer (1903) Recherches sur les localisations motrices spinales. J. de Neurol. 8: 303-312. Chung, S. K., D. W. Pfaff, and R. S. Cohen (1988) Estrogen- Bronson, F. H., and P. D. Heideman (1994) Seasonal regulation induced alterations in synaptic morphology in the midbrain of reproduction in mammals. In: E. Knobil and J. D. Neill central gray. Exp. Brain Res. 69: 522-530. (eds.) The physiology of reproduction. Raven Press, New Chung, S. K., R. S. Cohen, and D. W. Pfaff (1990) York: Vol. 2, pp. 541-584. Transneuronal degeneration in the midbrain central gray Brown, H. K., and M. F. Nolan (1979) Ultrastructure and following chemical lesions in the ventromedial nucleus: quantitative synaptology of the sacral parasympathetic a qualitative and quantitative analysis. Neurosci. 38: 409- nucleus. J. Neurocytol. 8: 167-179. 426. Bunge, M. B. (1973) Fine structure of nerve fibers and growth Clark, J. H., and S. K. Mani (1994) Actions of ovarian steroid cones of isolated sympathetic neurons in culture. J. Cell hormones. In: The physiology of reproduction (E. Knobil Biol. 56: 713-735. and J.D.Neill, eds), Vol. 1. pp1011-1059. New York: Raven Press . Burke, R. A., A. Lundberg, and F. Weight (1971) Spinal border cell origin of the ventral spinocerebellar tract. Exp. Brain Concannon, P., B. Hodgson, and D. Lein (1980) Reflex LH Res. 12: 283-294. release in estrous cats following single and multiple copulation. Biology of Reproduction 23: 111-117. Burke, R. E., P. L. Strick, K. Kanda, C. C. Kim, and B. Walmsley (1977) Anatomy of medial gastrocnemius and Conradi, S. (1969) On motoneuron synaptology in adult cats. soleus motor nuclei in cat spinal cord. J. Neurophysiol. Acta Physiol. scand. Suppl. 332: 1-115. 40: 667-680. Corodimas, K. P., and J. I. Morrell (1990) Estradiol- Burrows H (1949) Biological actions of sex hormones. concentrating forebrain and midbrain neurons project Cambridge: Cambridge University Press directly to the medulla. J. Comp. Neurol. 291: 609-620. Burstein, R., K. D. Cliffer, and G. J. Giesler (1990) Cells of Cottingham, S. L., P. A. Femano, and D. W. Pfaff (1988) origin of the spinohypothalamic tract in the rat. J. Comp. Vestibulospinal and reticulospinal interactions in the Neurol. 291: 329-344. activation of back muscle EMG in the rat. Exp. Brain Res. 73: 198-208. Cabot, J. B. Some principles of the spinal organization of the sympathetic preganglionic outflow. (1996) In: The Craig, A. D. (1995) Distribution of brainstem projections from Emotional Motor System. G. Holstege, R. Bandler, and spinal lamina I neurons in the cat and the monkey. J. Comp. C. Saper (eds.). Progr. Brain Res. 107: 29-42. Neurol. 361: 225-248. Campbell, E. J. M., and J. H. Green (1953) The expiratory Craig, A. D., and K. D. Kniffi (1985) Spinothalamic function of the abdominal muscles in man. An electro- lumbosacral lamina I cells responsive to skin and muscle myographic study. J. Physiol. 120: 409-418. stimulation in the cat. J. Physiol. (London) 365: 197-221. Canteras, N. S., R. B. Simerly, and L. W. Swanson (1994) Craig, A. D., B. Heppelmann, and H. G. Schaible (1988) The Organization of projections from the ventromedial nucleus projection of the medial and posterior articular nerves of of the hypothalamus: a Phaseolus vulgaris-Leuco- the cat’s knee to the spinal cord. J. Comp. Neurol. 276: agglutinin study in the rat. J. Comp. Neurol. 348: 41-79. 279-288. Carrer, H. F., and A. Aoki (1982) Ultrastructural changes in Cramer, K. S., and D. C. Van Essen (1995) Lack of topography the hypothalamic ventromedial nucleus of ovariectomized in the spinal cord projection to the rabbit soleus muscle. rats after estrogen treatment. Brain Res. 240: 221-33. J. Comp. Neurol. 351: 404-414. Carrive, P., and R. Bandler (1991) Viscerotopic organization Cross, B. A., and R. G. Dyer (1972) Ovarian modulation of of neurons subserving hypotensive reactions within the unit activity in the anterior hypothalamus of the cyclic midbrain periaqueductal grey: a correlative functional and rat. J Physiol. (London) 222: 25P. anatomical study. Brain Res. 541: 206-215. Crouch, J. E. (1969) Text-Atlas of Cat Anatomy. Philadelphia, Catelli, J. M., A. F. Sved, and B. R. Komisaruk (1987) Lea and Febiger Vaginocervical probing elevates blood pressure and Cueva-Rolon, R., E. J. Munoz-Martinez, R. Delgado-Lezama, induces analgesia by separate mechanisms. Physiol. J. G. Raya, and G. Gonzalez-Santos (1993) Substained Behav. 41: 609-612. activation of the triceps surae muscle produced by Cervero, F., A. Iggo, and V. Molony (1979) Ascending mechanical stimulation of the genital tract of the female projections of nociceptor-driven lamina I neurones in the cat. Brain Res. 600: 33-38. cat. Exp. Brain Res. 35: 135-149. Cueva-Rolon, R., L. E. Gomez, B. R. Komisaruk, and E. J.

146 References

Munoz Martinez (1995) Inhibition of withdrawal of the urinary bladder and sex organs. In A.D. Loewy and responses by pelvic nerve electrical stimulation. Brain Res. K.M. Spyer (eds): Central regulation of autonomic 679: 267-273. functions. Oxford University Press, Oxford, pp. 310-333. Cunningham, S. T., J. L. Steinman, B. Whipple, A. D. Mayer, DeBold, J. F., and C. W. Malsbury (1983) Inhibition of sexual and B. R. Komisaruk (1991) Differential roles of receptivity after intracranial cycloheximide infusions in hypogastric and pelvic nerves in the analgesic and motoric female hamsters. Brain Res. Bull. 11: 633-6. effects of vaginocervical stimulation in rats. Brain Res. Donselaar, Y., D. Kernell, O. Eerbeek, and B. A. Verhey (1985) 559: 337-343. Somatotopic relations between spinal motoneurons and Cunningham, S. T., J. S. Rosenblatt, and B. R. Komisaruk muscle fibres of the cat’s m. peroneus longus. Brain Res. (1992) Reflexive ovulation in the rat, induced by caesarean 335: 81-88. section, is blocked by pelvic and/or hypogastric nerve Dubrovsky, B., and D. Filipini (1990) Neurobiological aspects transection. Neuroendocrinology 56: 393-396. of the pelvic floor muscles involved in defecation. Dampney, R. A. l., and R. M. McAllen (1988) Differential Neurosci. Biobehav. Rev. 14: 157-168. control of sympathetic fibers supplyin ghinlimb skin and Dufy, B., J.-D. Vincent, H. Fleury, P. DuPasquier, D. Gourdji, muscle by subretrofacial neurones in the cat. J. Physiol. and A. Tixier-Vidal (1979) Membrane effects of Lond. 395: 41-56. thyrotropin-releasing hormone and estrogen shown by Danielsson I., and U. Norrsell (1985) Somatosensory, short intracellular recording from pituitary cells. Science 204: latency activation in the intercollicular region of the cat’s 509-511. mesencephalon. Acta Physiol. Scand. 124: 479-482. Dum, R. P., R. E. Burke, M. J. O’Donovan, J. Toop, and J. A. Danielsson I., and U. Norrsell (1986) Somatosensory units in Hodgson (1982) Motor-unit organization in flexor the cat’s intercollicular region. Acta Physiol. Scand. 128: digitorum longus muscle of the cat. J. Neurophysiol. 47: 579-586. 1108-1125. Davidson, J. M., and G. J. Bloch (1969) Neuroendocrine Duranti, R., T. Pantaleo, F. Bellini, F. Bongianni, and G. Scano aspects of male reproduction. Biol. of Reprod. Suppl. 1: (1991)Respiratory responses induced by the activation of 67-92. somatic nociceptive afferents in humans. J. Appl. Physiol. Davis, L., P. Dou, M. DeWit, and S. B. Kater (1992) Protein 71: 2440-2448. synthesis within neuronal growth cones. J. Neurosci. 12: Duron, B. (1973) Postural and ventilatory functions of 4867-4877. intercostal muscles. Acta Neurobiol. Exp. 33: 355-380. Davis, P. J., and S. P. Zhang (1991) What is the role of the Edgley, S. A., and E. Jankowska (1987) An interneuronal relay midbrain periaqueductal gray in respiration and for group I and II afferents in the midlumbar segments of vocalization? In: A. Depaulis and R. Bandler (eds): The the cat spinal cord. J. Physiol. (London)389: 647-674. midbrain periaueductal Gray matter. New York: Plenum Press. 57-65. Ellenberger, H. H., and J. L. Feldman (1990a) Brainstem connections of the rostral ventral respiratory group of the Davis, P. J., S. P. Zhang, and R. Bandler (1996) Midbrain and rat. Brain Res. 513: 35-42. medullary regulation of respiration and vocalization. In: The Emotional Motor System. G. Holstege, R. Bandler, Ellenberger, H. H., and J. L. Feldman (1990b) Subnuclear and C. Saper (eds.). Progr. Brain Res. 107: 315-326. organization of the lateral tegmental field of the rat. I: Nucleus ambiguus and ventral respiratory group. J. Comp. Dawson, A. B., and H. B. Friedgood (1940) The time and Neurol. 294: 202-11. sequence of preovulatory changes in the cat ovary after mating or mechanical stimulation of the cervix uteri. Anat. Elliott, H. C. (1942) Studies on the motor cells of the spinal Rec. 76: 411-424. cord. I. Distribution in the normal human cord. Am. J. Anat. 70: 95-117. de Groat, W. C., I. Nadelhaft, C. Morgan, and T. Schauble (1978) Horseradish peroxidase tracing of visceral efferent Elliott, H. C. (1944) Studies on the motor cells of the spinal and primary afferent pathways in the cat’s sacral spinal cord. IV. Distribution in experimental animals. J. Comp. cord using benzidine processing. Neurosci. Lett. 10: 103- Neurol. 81: 97-103. 108. Emery, D. E., and B. D. Sachs (1975) Ejaculatory pattern in de Groat, W. C., A. M. Booth, J. Krier, R. J. Milne, C. Morgan, female rats without androgen treatment. Science 190: 484- and I. Nadelhaft (1979) Neural control of the urinary 486. bladder and large intestine. In C. Brooks, K. Koizumi and Engberg, I. (1964) Reflexes to foot muscles in the cat. Acta A. Sato (eds): Integrative functions of the autonomic Physiol. Scand. 62: 1-64. nervous system. Tokyo University press, Tokyo/Elsevier, English A. W., and W. D. Letbetter (1981) Intramuscular Amsterdam chapter 4: pp. 50-67. “compartmentalization” of the cat biceps femoris and de Groat, W. C., A. M. Booth, R. J. Milne, and C. Morgan semitendinosus muscles: Anatomy and EMG patterns. (1980) Properties of interneurons in the sacral autonomic Abstr. Soc. Neurosci. 7: 557. nucleus of the cat. Proc. Int. Union Physiol. Sci., 14: 373. English, A. W., and W. D. Letbetter (1982) Anatomy and de Groat, W. C., I. Nadelhaft, R. J. Milne, A. M. Booth, C. innervation patterns of cat lateral gastrocnemius and Morgan, and K. Thor (1981) Organization of the sacral plantaris muscles. Am. J. Anat. 164: 67-77. parasympathetic reflex pathways to the urinary bladder English, A. W., and O. I. Weeks (1987) An anatomical and and large intestine. J. Auton. Nerv. Syst. 3: 135-160. functional analysis of cat biceps femoris and semi- de Groat, W.C., and W.D. Steers (1990) Autonomic regulation tendinosus muscles. J. Morphol. 191: 161-175.

147 References

Fedirchuk, B., L. Song, J. W. Downie, and S. J. Shefchyk Holstege (1996) Descending projections from the caudal (1992) Spinal distribution of extracellular field potentials ventrolateral medulla to the lumbosacral cord of the male generated by electrical stimulation of pudendal and hamster: possible role in reproductive behavior. Neurosci. perineal afferents in the cat. Exp. Brain Res. 89: 517-520. Abstr. Feldman, J. L. (1986) Neurophysiology of breathing in Giantonio, G. W., N. L. Lund, and A. A. Gerall (1970) Effect mammals. In: F.E. Bloom (eds): Handbook of Physiology, of diencephalic and rhinencephalic lesions on the male Sect. 1: The Nervous System, vol. IV. Intrinsic Regulatory rat’s sexual behavior. J. Comp. Physiol. Psychol. 73: 38- Systems of the Brain. Bethesda, Md: Am. Physiol. Soc., 46. pp. 463-524. Gillette, R. G., R. C. Kramis, and W. J. Roberts (1993a) Spinal Feldman, J. L., A. D. Loewy, and D. F. Speck (1985) projections of cat primary afferent fibers innervating Projections from the ventral respiratory group to phrenic lumbar facet joints and multifidus muscle. Neurosci. Lett. and intercostal motoneurons in cat: an autoradiographic 157: 67-71. study. J. Neurosci. 5: 1993-2000. Gillette, R. G., R. C. Kramis, and W. J. Roberts (1993b) Ferrington, D. G., L. S. Sorkin, and W. D. Willis (1987) Characterization of spinal somatosensory neurons having Responses of spinothalamic tract cells in the superficial receptive fields in lumbar tissues of cats. Pain. 54: 85-98. dorsal horn of the primate lumbar spinal cord. J. Physiol. Gjone, R. (1966) Excitatory and inhibitory bladder responses (London) 388: 681-703. to stimulation of limbic, diencephalic and mesencephalic Fetcho, J. R. (1987) A review of the organization and evolution structures in the cat. Acta Physiol. Scand. 66: 91-102. of motoneurons innervating the axial musculature of Goering, J. H. (1928) Experimental analysis of motor-cell vertebrates. Brain Res. Rev. 12: 243-280. columns in the cervical enlargement of the spinal cord in Fields, H. L., and A. I. Basbaum (1978) Brainstem control of the albino rat. J. Comp. Neurol. 46: 125-151. spinal pain-transmission neurons. annu. Rev. Physiol. 40: Goldfoot, D. A., S. M. Essock-Vitale, C. Asa, J. E. Thornton, 217-248. and A. L. Leshner (1978) Anosmia in male rhesus monkeys Fields, H. L., G. A. Meyer, and L. D. Partridge (1970) does not alter copulatory activity with cycling females. Convergence of visceral and somatic input onto cat spinal Science 199: 1095-1096. neurons. Exp. Neurol. 26: 36-52. Gomora, P., C. Beyer, G. Gonzalez Mariscal, and B. R. Floody, O. R., and T. L. O’Donohue (1980) Lesions of the Komisaruk (1994) Momentary analgesia produced by Mesencephalic Central Gray Depress Ultrasound copulation in female rats. Brain Res. 656: 52-58. Production and Lordosis by Female Hamsters. Physiol. Gordon, D. C., G. E. Loeb, and F. J. Richmond (1991) Behav. 24: 79-85. Distribution of motoneurons supplying cat sartorius and Floody, O. R., R. D. Lisk, and A. J. Vomachka (1986) tensor fasciae latae, demonstrated by retrograde multiple- Facilitation of lordosis by estradiol in the mesencephalic labelling methods. J. Comp. Neurol. 304: 357-372. central gray. Physiol. Behav. 37: 587-595. Grant, G., and Q. Xu (1988) Routes of entry into the Frankfurt, M., and B. S. McEwen (1991) Estrogen increases cerebellum of spinocerebellar axons from the lower part axodendritic synapses in the VMN of rats after ovari- of the spinal cord. An experimental anatomical study in ectomy. Neuroreport 2: 380-382. the cat. Exp. Brain Res. 77: 543-561. Frankfurt, M., E. Gould, C. S. Woolley, and B. S. McEwen Grigg, P., H. G. Schaible, and R. F. Schmidt (1986) Mechanical (1990) Gonadal steroids modify dendritic spine density sensitivity of group III and IV afferents from posterior in ventromedial hypothalamic neurons: a Golgi study in articular nerve in normal and inflamed cat knee. J. the adult rat. Neuroendocrinol. 51: 530-535. Neurophysiol. 55: 635-643. Fukuda, H., and K. Fukai (1986a) Postural change and Guyton, A. C. (1986) Textbook of medical physiology. straining induced by distension of the rectum, vagina and Chapters 63, 66 and 82. Philadelphia: W.B. Saunders. urinary bladder of decerebrate dogs. Brain Res. 380: 276- Haase, P. (1990) Explanation for the labeling of cervical 286. motoneurons in young rats following the introduction of Fukuda, H., and K. Fukai (1986b) Location of the reflex centre horseradish peroxidase into the calf. J. Comp. Neurol. 297: for straining elicited by activation of pelvic afferent fibres 471-478. of decerebrate dogs. Brain Res. 380: 287-296. Haase, P., and A. W. Hrycyshyn (1985) Labeling of Fukuda, H., and K. Fukai (1988) Discharges of bulbar motoneurons supplying the cutaneous maximus muscle respiratory neurons during rhythmic straining evoked by in the rat, following injection of the triceps brachii muscle activation of pelvic afferent fibers in dogs. Brain Res. 449: with horseradish peroxidase. Neurosci. Lett. 60: 313-318. 157-166. Haase, P., and A. W. Hrycyshyn (1986) On the diffusion of Fukuda, H., and T. Koga (1991) Midbrain stimulation inhibits horseradish peroxidase into muscles and the “spurious” the micturition, defecation and rhythmic straining reflexes labeling of motoneurons. Exp. Neurol. 91: 399-402. elicited by activation of sacral vesical and rectal afferents Halachmi, S., E. Marden, G. Martin, H. MacKay, C. in the dog. Exp. Brain Res. 83: 303-316. Abbondanza, and M. Brown (1994) Estrogen receptor- Gerrits, P. O. , and G. Holstege (1996) Pontine and medullary associated proteins: possible mediators of hormone- projections to the nucleus retroambiguus: A WGA-HRP induced transcription. Science 264: 1455-1458. and autoradiographic tracing study in the cat. J. Comp. Hamm, Th., M. W. Koehler, D. G. Stuart, and S. Vanden Noven Neurol. in press. (1985) Partitioning of monosynapotic Ia excitatory post- Gerrits, P. O., L. Kerstens, V. G. J. M. VanderHorst, and G. synaptic potentials in the motor nucleus of the cat

148 References

semimembranosus muscle. J. Physiol. 369: 379-398. Holstege, G. (1994) The basic, somatic, and emotional Hansen, S., C. Kohler, M. Goldstein, and H. V. M. Steinbusch components of the motor system in mammals. In : The rat (1982) Effects of ibotenic acid-induced neuronal nervous system, Academic Press Inc. San Diego in press degeneration in the medial preoptic area and the lateral Holstege, G., and B. F. Blok (1989) Descending pathways to hypothalamic area on sexual behavior in the male rat. Brain the cutaneous trunci muscle motoneuronal cell group in Res. 239: 213-232. the cat. J. Neurophysiol. 62: 1260-1269. Hart, B. L., and C. M. Haugen (1972) Scent marking and Holstege, G., and H. G. J.M. Kuypers (1982) The anatomy of sexual behavior maintained in anosmic dogs. Commun. brain stem pathways to the spinal cord in the cat. A labeled Behav. Biol. 6: 131-135. amino acid tracing study. Progr. Brain Res. 57: 145-175. Hart, B. L., and M. G. Leedy (1983) Female sexual responses Holstege, G., and J. Tan (1987) Supraspinal control of in male cats facilitated by olfactory bulbectomy and medial motoneurons innervating the striated muscles of the pelvic preoptic/anterior hypothalamic lesions. Behavioural floor including urethral and anal sphincters in the cat. Brain Neruroscience 97: 608-614. 110: 1323-1344. Hart, B. L., and V. L. Voith (1978) Changes in urine spraying, Holstege, G., H. G. J.M. Kuypers, and J. J. Dekker (1977) feeding and sleep behavior of cats following medial The organization of the bulbar fibre connections to the preoptic-anterior hypothalamic lesions. Brain Res. 145: trigeminal, facial and hypoglossal motor nuclei II. An 406-409. autoradiographic tracing study in cat. Brain 100: 265-286. Hart, B. L., C. M. Haugen, and D. M. Petersen (1973) Effects Holstege, G., H. G. J. M. Kuypers, and R.C. Boer (1979) of medial preoptic-anterior hypothalamic lesions on Anatomical evidence for direct brain stem projections to mating behavior of male cats. Brain Res. 54: 177-191. somatic motoneuronal cell groups and autonomic Haydon, P. G., and M. J. Zoran (1994) Retrograde regulation preganglionic cell groups in cat spinal cord. Brain Res. of presynaptic development during synaptogenesis. J. 171: 329-333. Neurobiol. 25: 694-706. Holstege, G., J. Van Neerven, and F. Evertse (1987) Spinal Heimer, L. and K. Larsson (1966) Impairment of mating cord location of the motoneurons innervating the behavior in male rats following lesions in the preoptic- abdominal, cutaneous maximus, latissimus dorsi and anterior hypothalamic continuum. Brain Res. 3: 248-263. longissimus dorsi muscles in the cat. Exp. Brain Res. 67: Heimer, L. and K. Larsson (1967) Mating behavior of male 179-194. rats after olfactory bulb lesions. Physiol. Behav. 2: 207- Holstege, G., D. Griffiths, H. De Wall, and E. Dalm (1986) 209. Anatomical and physiological observations on supraspinal Henry, M. A., L. E. Westrum, and L. R. Johnson (1985) control of bladder and urethral sphincter muscles in the Enhanced ultrastructural visualization of the horseradish cat. J. Comp. Neurol. 250: 449-461. peroxidase-tetramethylbenzidine reaction product. J. Holstege , J. C. (1989) Brainstem projections to motoneurons Histochem. Cytochem. 33: 1256-1259. in the lumbar spinal cord. An ultrastructural study in the Herbert, H., and C. B. Saper (1992) Organization of medullary rat. Thesis Erasmus University Rotterdam, The Nether- adrenergic and noradrenergic projections to the peri- lands. aqueductal gray matter in the rat. J. Comp. Neurol. 315: Holstege, J. C. (1996) The vetnro-medial medullary 34-52. projections to spinal motoneurons: ultrastructure, Holman, S. D., and J. B. Hutchison (1991) Lateralized action transmitters and functional aspects. In: The Emotional of androgen on development of behavior and brain sex Motor System. G. Holstege, R. Bandler, and C. Saper differences. Brain Res. Bull. 27: 261-265. (eds.). Progr. Brain Res. 107: 159-182. Holstege, G. (1987a) Anatomical evidence for an ipsilateral Honda, C. N. (1985) Visceral and somatic afferent convergence rubrospinal pathway and for direct rubrospinal projections onto neurons near the central canal in the sacral spinal to motoneurons in the cat. Neurosci. Lett. 74: 269-274. cord of the cat. J. Neurophysiol. 53: 1059-1078. Holstege, G. (1987b) Some anatomical observations on the Hoover, J. E., and R. G. Durkovic (1991) Morphological projections from the hypothalamus to brainstem and spinal relationships among extensor digitorum longus, tibialis cord: an HRP and autoradiographic tracing study in the anterior, and semitendinosus motor nuclei of the cat: an cat. J. Comp. Neurol. 260: 98-126. investigation employing the retrograde transport of Holstege, G. (1988) Direct and indirect pathways to lamina I multiple fluorescent tracers. J. Comp. Neurol. 303: 255- in the medulla oblongata and spinal cord of the cat. Progr. 266. Brain Res. 77: 47-94. Hopkins, D. A., and G. Holstege (1978) Amygdaloid Holstege, G. (1989) Anatomical study of the final common projections to the mesencephalon, pons, and medulla pathway for vocalization in the cat. J. Comp. Neurol. 284: oblongata in the cat. Exp. Brain Res. 32: 529-547. 242-252. Horcholle-Bossavit, G., L. Jami, D. Thiesson, and D. Zytnicki Holstege, G. (1991a) Descending pathways from the (1988) Motor nuclei of peroneal muscles in the cat spinal periaqueductal gray and adjacent areas. In R. Bandler and cord. J. Comp. Neurol. 277: 430-440. A.Depaulis (eds.): The midbrain periaqueductal gray Hörner, M., and H. Kümmel (1993) Topographical repre- matter. Plenum Press, New York pp. 239-265. sentation of shoulder motor nuclei in the cat spinal cord Holstege, G. (1991a) Desending motor pathways and the spinal as revealed by retrograde fluorochrome tracers. J. Comp. motor system: limbic and non-limbic components. Progr. Neurol. 335: 309-319. Brain Res. 87: 307-412. Huber, J. F. (1936) Nerve roots and nuclear groups in the

149 References

pigeon. J. Comp. Neurol. 65: 405-414. squirrel monkey. J. Comp. Neurol. 160: 419-442. Hubscher, C. H., and K. J. Berkley. (1994) Responses of Kelly, M. J., R. L. Moss, and C. A. Dudley (1977) The effects neurons in caudal solitary nucleus of female rats to of microelectrophoretically applied estrogen, cortisol and stimulation of vagina, cervix, uterine horn and colon. Brain acetylcholine on medial preoptic-septal unit activity Res. 664: 1-8. throughout the estrous cycle of the female rat. Exp. Brain Hubscher, C. H., and K. J. Berkley. (1995) Spinal and vagal Res. 30: 53-64. influences on the responses of rat solitary nucleus neurons Kennard, M. A. (1954) The course of ascending fibers in the to stimulation of uterus, cervix and vagina. Brain Res. spinal cord of the cat essential to the recognition of painful 702: 251-254. stimuli. J. Comp. Neurol. 100: 511-524. Hutchison, D. L., R. R. Roy, S. Bodine-Fowler, J. A. Hodgson, Kernell, D., B. A. Verhey, and O. Eerbeek (1985) Neuronal and V. R. Edgerton (1989) Electromyographic (EMG) and muscle unit properties at different rostro-caudal levels amplitude patterns in the proximal and distal com- of cat’s motoneurone pool. Brain Res. 335: 71-79. partments of the cat semitendinosus during various motor Kerstens, L., V. G. J. M. VanderHorst, and G. Holstege (1996) tasks. Brain Res. 479: 56-64. The nucleus retroambiguus in the female rat projects to Hylden, J. L. K., H. Hayashi, and G. J. Bennett (1986a) Lamina several distinct motoneuronal cell groups in the I spinomesencephalic neurons in the cat ascend via the lumbosacral cord: possible role in lordosis behavior. dorsolateral funiculi. Somatosens. Res. 4: 31-42. Neurosci. Abstr. Hylden, J. L. K., H. Hayashi, R. Dubner, and G. J. Bennett Kiesel, I. (1938) Nerve-cell groups in the anterior column of (1986b) Physiology and morphology of the lamina I caudal part of spinal cord and localisation of motor spinomesencephalic projection. J. Comp. Neurol. 247: branches of sciatic nerve in cats and guinea-pigs. Fol. 505-515. Neuropath. eston. 17: 86-101. Iliya, A. R., and R. P. Dum (1984) Somatotopic relations Kitamura, S., and F. J. R. Richmond (1994) Distribution of between the motor nucleus and its innervated muscle fibers motoneurons supplying dorsal and ventral suboccipital in the cat tibialis anterior. Exp. Neurol. 86: 272-292. muscles in the feline neck. J. Comp. Neurol. 347: 25-35. Ishizuka, N., H. Mannen, T. Hongo, and S. Sasaki (1979) Knyihar-Csillik, E., P. Rakic, and B. Csillik (1986) Reactive Trajectory of group Ia afferent fibers stained with synapto-neogenesis in the upper dorsal horn of the adult horseradish peroxidase in the lumbosacral spinal cord of primate: regenrative or collateral sprouting? In: the cat: three dimensional reconstructions from serial Development and plasticity in the mammalian spinal cord sections. J. Comp. Neurol. 186: 189-213. (M. Goldberger, A. Gorio, M. Murray, eds.), pp191-210. Padova, Italy: Liviana Press. Jänig, W. (1987) Neuronal mechanisms of pain with special emphasis on visceral and deep somatic pain. Acta Komisaruk, B. R. (1972) Induction of lordosis in ovariec- Neurochir Suppl (Wien) 38 : 16-32. tomized rats by stimulation of the vaginal cervix: hormonal Jürgens, U., and R. Pratt (1979) Role of the periaqueductal and neural interrelationships. UCLA Forum Med. Sci. 15: grey in vocal expression of emotion. Brain Res. 167: 367- 127-135. 379. Komisaruk, B. R., and J. L. Steinman (1986) Genital Kaiser, O. (1891) Die Functionen der Ganglienzellen des stimulation as a trigger for neuroendocrine and behavioral Halsmarkes. Haag. control of reproduction. Ann. N. Y. Acad. Sci. 474: 64-75. Katter, J. T., R. Burstein, and G. J. Jr. Giesler (1991) The cells Komisaruk, B. R., and J. Wallman (1977) Antinociceptive of origin of the spinohypothalamic tract in cats. J. Comp. effects of vaginal stimulation in rats: neurophysiological Neurol. 303: 101-112. and behavioral studies. Brain Res. 137: 85-107. Kawahara, K., and M. Suzuki (1990) Interactions between Komisaruk, B. R., and B. Whipple (1986) Vaginal stimulation- locomotor and respiratory activities during hindlimb produced analgesia in rats and women. Ann. N. Y. Acad. stepping on a stationary surface in decerebrate cats. Sci. 467: 30-39. Neurosci. Lett. 113: 23-28. Komisaruk, B. R., N. T. Adler, and J. Hutchison (1972) Genital Kawahara, K., Y. Nakazono, Y. Yamauchi, and Y. Miyamoto sensory field: enlargement by estrogen treatment in female (1989) Coupling between respiratory and locomotor rats. Science 178: 1295-1298. rhythms during fictive locomotion in decerebrate cats. Kow, L. -M, and D. W. Pfaff (1979) Responses of single units Neurosci. Lett. 103: 326-330. in sixth lumbar dorsal root ganglion of female rats to Kawakami, M., E. Terasawa, and T. Ibuki (1970) Changes in mechanostimulation relevant for Lordosis Reflex. J. multiple unit activity of the brain during the estrous cycle. Neurophysiol.. 42: 203-214. Neuroendocrinol. 6: 30. Kow, L. -M, and D. W. Pfaff (1982) Responses of medullary Kawatani, M., C. Takeshige, and W. C. de Groat (1990) Central reticulospinal and other reticular neurons to somatosensory distribution of afferent pathways from the uterus of the and brainstem stimulation in anesthetized or freely-moving cat. J. Comp. Neurol. 302: 294-304. ovariectomized rats with or without estrogen treatment. Kawatani, M., M. Tanowitz, and W. C. de Groat (1994) Exp. Brain Res. 47: 191-202. Morphological and electrophysiological analysis of the Kow, L. -M, M. O. Montgomery, and D. W. Pfaff (1979) peripheral and central afferent pathways from the clitoris Triggering of Lordosis Reflex in Female Rats with of the cat. Brain Res. 646: 26-36. Somatosensory Stimulation: Quantitative Determination Keefer, D. A, and W. E. Stumpf (1975) Atlas of estrogen- of Stimulus Parameters. J. Neurophysiol. 42: 195-203. concentrating cells in the cetnral nervous system of the Kow, L. -M, F. P. Zemlan, and D. W. Pfaff (1980) Responses

150 References

of lumbosacral spinal units to mechanical stimuli related cat: interanimal variability of locomotor activity and to analysis of lordosis reflex in female rats. J. Neuro- cutaneous reflexes. Exp. Brain Res. 96: 125-140. physiol. 43: 27-46. Loeb, G. E., C. A. Pratt, C. M. Chanaud, and F. J. R. Richmond Koyama, Y., A. Makuya, and M. Kuru (1962) Vesico-motor (1987) Distribution and innervation of short, interdigitated areas in the cat midbrain. Jap. J. Physiol. 12: 63-80. muscle fibers in parallel-fibered muscles of the cat Krieger, M. S., and R. J. Barfield (1976) Masculine sexual hindlimb. J. Morphol. 191: 1-15. behavior: pacing and ejaculatory patterns in female rats Loewy, A. D., and S. McKellar (1981) Serotonergic projections induced by electrical shock. Physiol. Behavior 16: 671- from the ventral medulla to the intermediolateral cell 675. column in the rat. Brain Res. 21: 146-152. Kruse, M. N., H. Noto, J. R. Roppolo, and W. C. de Groat Loewy, A. D., and K. M. Spyer (1990) Central regulation of (1990) Pontine control of the urinary bladder and external autonomic functions. New York: Oxford. urethral sphincter in the rat. Brain Res 532: 182-190. Loewy, A. D., J. H. Wallach, and S. McKellar (1981) Efferent Landmesser, L. (1978) The distribution of motoneurones connections of the vetnral medulla oblongata in the rat. supplying chick hind limb muscles. J. Physiol. 284: 371- Brain Res. Rev. 3: 63-80. 389. Logue, C. M., and R. H. Moos (1988) Positive perimenstrual Langub Jr., M. C., B. E. Maley, and R. E. Watson Jr. (1994) changes: toward a new perspective on the menstrual cycle. Estrous cycle-associated axosomatic synaptic plasticity J. Psychosom. Res. 32: 31-40. upon estrogen receptive neurons in the rat preoptic area. Lovick, T. A. (1985) Projections from the diencephalon and Brain Res. 641: 303-310. mesencephalon to nucleus paragigantocellularis lateralis Larnicol, N., D. Rose, D. Marlot, and B. Duron (1982) in the cat. Neurosci. 14: 853-861. Anatomical organization of cat intercostal motor nuclei Lovick, T. A. (1987) Differential control of cardiac and as demonstrated by HRP retrograde labelling. J. Physiol. vasomotor activity by neurones in nucleus paragigan- Paris 78: 198-206. tocellularis lateralis in the cat. J. Physiol. 389: 23-35. Larson, C. R. (1991) On the relation of PAG neurons to Lovick, T. A. (1993) Integrated activity of cardiovascular and laryngeal and respiratory muscles during vocalization in pain regulatory systems: role in adaptive behavioural the monkey. Brain Res. 552: 77-86. responses. Pro. Neurobiology 40: 631-644. Larsson, K., and L. Heimer (1964) Mating behavior in male Lovick, T. A. (1996) Midbrain and medullary regulation of rats after lesions in the preoptic area. Nature 202: 413- defensive cardiovascular functions. In: The Emotional 414. Motor System. G. Holstege, R. Bandler, and C. Saper Larson, C. R., and M. K. Kistler (1984) Periaqueductal gray (eds.). Progr. Brain Res. 107: 310-314. neuronal activity associated with laryngeal EMG and Lovick, T. A., P. R. Smith, and S. M. Hilton (1984) Spinally- vocalization in the awake monkey. Neurosci. Lett. 46: 261- projecting neurones near the ventral surface of the medulla 266. in the cat. J. AuÑØn˙ Nerv. Sys. 11: 27-33. Larsson, K., G. Perez-Palacios, G. Morali, and C. Beyer (1975) Lucio, R. A., J. Manzo, M. Martinez-Gomez, B. D. Sachs, Effects of dihydrotestosterone and estradiol pretreatment and P. Pacheco (1994) Participation of pelvic nerve upon testosterone-induced sexual behavior in the castrated branches in male rat copulatory behavior. Physiol. Behav. male rat. Horm. Behav. 6: 1-8. 55: 241-246. Leedy, M. G., and B. L. Hart (1985) Female and male sexual Malsbury, C. W. (1971) Facilitation of male rat copulatory responses in female cats with ventromedial hypothalamic behavior by electrical stimulation of the medial preoptic lesions. Behav. Neurosci. 99: 936-941. area. Physiol. Behav. 7: 797-805. Letbetter, W. D., and A. W. English (1981) The relationship Malsbury, C.W., L. M. Kow, and D.W. Pfaff (1977) Effects of between peripheral intramuscular “compartments” and medial hypothalamic lesions on the lordosis response and spatial arrangement of biceps femoris and semitendinosus other behaviors in female golden hamsters. Physiol. Behav. motor nuclei in the cat lumbr spinal cord. Abstr. Soc. 19: 223-237. Neurosci. 7: 557. Marinesco, G. (1904) Recherches sur les localisations motrices Levesque, D., and T. DiPaolo (1988) Rapid conversion of high spinales. cited from: Elliott, H. C. (1944) Studies on the into low striatal D2-dopamine receptor agonist binding motor cells of the spinal cord. IV. Distribution in states after an acute physiological dose of 17ß-estradiol. experimental animals. J. Comp. Neurol. 81: 97-103. Neurosci. Lett. 88: 113-118. Martinez-Gomez, M., R. Chirino, C. Beyer, B. R. Komisaruk, Li, P., and T. A. Lovick (1985) Excitatory projections from and P. Pacheco (1992) Visceral and postural reflexes hypothalamic and midbrain defence areas to nucleus evoked by genital stimulation in urethane-anesthetized paragigantocellularis lateralis in the rat. Exp. Neurol. 89: female rats. Brain Res. 575: 279-284. 543-553. Mason, P, and C. G. Leung (1996) Physiological functions of Liposits, Zs., and G. Sétáló (1980) Descending luteinizing pontomedullary raphe and medial reticular neurons. In: hormone-releasing hormone (LH-RH) nerve fibers to the The Emotional Motor System. G. Holstege, R. Bandler, midbrain of the rat. Neurosci. Lett. 20: 1-4. and C. Saper (eds.). Progr. Brain Res. 107: 269-284. Liu, R. P. C. (1983) Laminar origins of spinal projection Mathews, D., and D. A. Edwards (1977) Involvement of the neurons to the periaqueductal gray of the rat. Brain Res. ventromedial and anterior hypothalamic nuclei in the 264: 118-123. hormonal induction of receptivity in the female rat. Loeb, G. E. (1993) The distal hindlimb musculature of the Physiol. Behav. 19: 319-326.

151 References

Matsushita, M., Y. Hosoya. and M. Ikeda (1979) Anatomical and light microscopic analysis. In M.-M. Mesulam (ed): organization of the spinocerebellar system in the cat, as Tracing neural connections with horseradish peroxidase. studied by retrograde transport of horseradish peroxidas. Chisester: John Wiley pp. 1-151. J. Comp. Neurol. 184: 81-106. Meyerson, B. J. (1982) Colchicine delays the estrogen-induced McCarthy, L. E., and H. L. Borison (1974) Respiratory copulatory response in the ovariectomized female rat. mechanics of vomiting in decebrate cats. Am. J. Physiol. Brain Res. 253: 281-286. 226: 738-743. Michael, R. P. (1961) Observations upon the sexual behaviour McEwen, B. S. (1988) Genomic regulation of sexual behavior. of the domestic cat (Felis catus L.) under laboratory J. Steroid. Biochem. 30: 179-83. conditions. Behaviour 18: 1-24. McLaughlin, B. J. (1972) The fine structure of neurons and Michael, R. P., D. Zumpe, and R. W. Bonsall (1990) Estradiol synapses in the mostor nuclei of the cat spinal cord. J. administrationand the sexual activity of castrated male Comp. Neurol. 144: 429-460. rhesus monkeys (Macaca Mulatta). Horm. Behav. 24: 71- McMahon, S.B. (1986) Sensory-motor integration in urinary 88. bladder function. In F. Cervero and J.F.B. Morrison (eds.), Michel, G., and E. E. Baulieu (1980) Androgen receptor in Progress in Brain Research, Vol. 67, Elsevier, Amsterdam, rat skeletal muscle: characterization and physiological pp. 245-253. variations. Endocrinology 107: 2088-2098. McMahon, S.B., and J.F.B. Morrison (1979) Convergence of Miller, A. D. (1987) Localization of motoneurons innervating visceral and somatic axons on neurones which ascend in individual abdominal muscles of the cat. J. Comp. Neurol. the lateral funiculus of the cat. J. Physiol. , 275: 56P. 256: 600-606. Mehler, W. (1969) Some neurological species differences - A Miller, A. D., K. Ezure, and I. Suzuki (1985) Control of posteriori. Ann. N.Y. Acad. Sci. 167: 424-468. abdominal muscles by brain stem respiratory neurons in Meisel, R. L., and V. R. Luttrell (1990) Estradiol increases the cat. J. Neurophysiol. 54: 155-167. the dendritic length of ventromedial hypothalamic neurons Miller, A. D., L. K. Tan, and I. Suzuki (1987) Control of in female Syrian hamsters. Brain Res. Bull. 25: 165-168. abdominal and expiratory intercostal muscle activity Meisel, R. L., and D. W. Pfaff (1984) RNA and protein during vomiting: role of ventral respiratory group synthesis inhibitors: effects on sexual behavior in female expiratory neurons. J. Neurophysiol. 57: 1854-1866. rats. Brain Res. Bull. 12: 187-193. Miller, A. D., L. K. Tan, and S. F. Lakos (1989) Brainstem Meisel, R. L., and D. W. Pfaff (1985) Specificity and neural projections to cats’ upper lumbar spinal cord: implications sites of action of anisomycin in the reduction or facilitation for abdominal muscle control. Brain Res 493: 348-356. of female sexual behavior in rats. Horm. Behav. 19: 237- Miller, A. D., S. Nonaka, M. S. Siniaia, and J. Jakus (1995) 251. Multifunctional ventral respiratory group: bulbospinal Meisel, R. L., and B. D. Sachs (1994) The physiology of male expiratory neurons paly a role in pudendal discharge sexual behavior. In: E. Knobil and J. D. Neill (eds.) The during vomiting. J. Auton. Nerv. Syst. 54: 253-260. physiology of reproduction. Raven Press, New York: Vol. Milne, R. J., R. D. Foreman, G. J. jr. Giesler, and W. D. Willis 2, pp. 3-106. (1981) Convergence of cutaneous and pelvic visceral Menetrey, D., A. Chaouch, D. Binder, and J. M. Besson (1982) nociceptive inputs onto primate spinothalamic neurons. The origin of the spinomesencephalic tract in the rat: an Pain 11: 163-183. anatomical study using the retrograde transport of Miyakawa, M., and Y. Arai (1987) Synaptic plasticity to horseradish peroxidase. J. Comp. Neurol. 206: 193-207. estrogen in the lateral septum of the adult male and female Mense, S., and A. D. Craig (1988) Spinal and supraspinal rats. Brain Res. 436: 184-188. terminations of primary afferent fibers from the Mobbs, C. V., R. E. Harlan, M. R. Burrous, and D. W. Pfaff gastrocnemius-soleus muscle in the cat. Neurosci. 26: (1988) An estradiol-induced protein synthesized in the 1023-1035. ventral medial hypothalamus and transported to the Mermelstein, P. G., J. B. Becker, and D. J. Surmeier (1996) midbrain central gray. J. Neurosci. 8: 113-118. Estradiol reduces calcium currents in rat neostriatal Modianos, D. T., and D. W. Pfaff (1976) Brain stem and neurons via a membrane receptor. J. Neurosci. 16: 595- cerebellar lesions in female rats. II. Lordosis reflex. Brain 604. Res. 106: 47-56. Merrill, E. G. (1970) The lateral respiratory neurones of the Morgan, C., I. Nadelhaft, and W. C. de Groat (1981) The medulla: their associations with nucleus ambiguus, distribution of visceral primary afferents from the pelvic nucleus retroambigualis, the spinal accessory nucleus and nerve to Lissauer’s Tract and the spinal gray matter and the spinal cord. Brain Res. 24: 11-28. its relationship to the sacral parasympathetic nucleus. J. Merrill, E. G. (1971) The descending pathways from the lateral Comp. Neurol. 201: 415-440. respiratory neurons in cats. J. Physiol. 218: 82-83. Morrell, J.I., T. D. Wolinsky, M. S. Krieger, and D. W. Pfaff Merrill, E. G. (1974) Finding a respiratory function for the (1982) Autoradiographic identification of estradiol- medullary respiratory neurons. In R. Bellairs and E.G. concentrating cells in the spinal cord of the female rat. Gray (eds): Essays on the Nervous System. Oxford: Exp. Brain Res. 45: 144-50. Clarendon pp. 451-486. Mouton, L. J., and G. Holstege (1994) The periaqueductal Mesulam, M. -M (1982) Principles of horseradish peroxidase gray in the cat projects to lamina VIII and the medial part neurohistochemistry and their applications for tracing of lamina VII throughout the length of the spinal cord. neural pathways-axonal transport, enzyme histochemistry Exp. Brain Res. 101: 253-264.

152 References

Nabekura, J., Y. Oomura, T. Minami, Y. Mizuno, and A. Fukuda groups of the sacral region of the spinal cord in man. Arch. (1986) Mechanism of the rapid effect of 17ß-estradiol on Neurol. Psych. 3: 3. medial amygdala neurons. Science 233: 226-228. Parhon, C., and G. Nadadje (1906) Nouvelles contributions a Nadelhaft, I., W. C. de Groat, and C. Morgan (1980) Location l’etude des localisations des noyaux medullaires des nerfs and morphology of parasympathetic preganglionic craniens et rachidiens chez l’homme et le chien. J. de neurons in the sacral spinal cord of the cat revealed by Neurol. 11: 129-140. retrograde axonal transport of horseradish peroxidase. J. Parhon, C., and Mme. C. Parhon (1902) Sur les centres spinaux Comp. Neurol. 193: 265-281. des muscles de la jambe. J. de Neurol. 7: 283-293. Nadelhaft, I., J. Roppolo, C. Morgan, and W. C. de Groat Paxinos, G., C. and Watson (1986) The rat brain stereotaxic (1983) Parasympathetic preganglionic neurons and coordinates. Academic Press, Sydney. visceral primary afferents in monkey sacral spinal cord Peters, A., S. L. Palay, and HdeF. Webster (1991) The fine revealed following application of horseradish peroxidase structure of the nervous system. Neurons and their to pelvic nerve. J. Comp. Neurol. 216: 36-53. supporting cells. third edition. New York, Oxford: Oxford Naggar, A. N., and B. R. Komisaruk (1977) Facilitation of University Press. tonic immobility by stimulation of the vaginal cervix in Peters, S. E., and Rick, C. (1977) The actions of three the rat. Physiol. Behav. 19: 441-444. hamstring muscles of the cat: a mechanical analysis. J. Nathan, P. W., and M. C. Smith (1958) The centrifugal pathway Morphol. 152: 315-328. for micturition within the spinal cord. J. Neurol. Pfaff, D. W. (1973) Luteinizing hormone releasing factor Neurosurg. Psychiat. 21: 177-189. (LRF) potentiates lordosis behavior in hypophysectomized Neef de, C. (1900) Les localisations motrices medullaires chez ovariectomized rats. Science 182: 1148-1149. le chien et le lapin. Nevraxe 2: 71-111. Pfaff, D. W., and M. Keiner (1973) Atlas of estradiol- Ness, T. J., and G. F. Gebhart (1991) Interactions between concentrating cells in the central nervous system of the visceral and cutaneous nociception in the rat. II. Noxious female rat. J. Comp. Neurol. 151: 121-158. visceral stimuli inhibit cutaneous nociceptive neurons and Pfaff, D. W., and B. S. McEwen (1983) Actions of estrogens reflexes. J. Neurophysiol. (1991) 66: 29-39. and progestins on nerve cells. Science 219: 808-14. Noto, H., J. R. Roppolo, W. D. Steers, and W. C. de Groat Pfaff, D. W., and Y. Sakuma (1979a) Facilitation fo the lordosis (1991) Electrophysiological analysis of the ascending and refles of female rats from the ventromedial nucleus of the descending components of the micturition reflex pathway hypothalamus. J. Physiol. 288: 189-203. in the rat. Brain Res 549: 95-105. Pfaff, D. W., and Y. Sakuma (1979b) Deficit in the lordosis Nyby, J., J. A. Matochik, and R. J. Barfield (1992) Intracranial reflex of female rats caused by lesions in the ventromedial androgenic and estrogenic stimulation of male-typical nucleus of the hypothalamus. J. Physiol. 288: 203-211. behaviors in house mice (Mus domesticus). Horm. Behav. 26: 24-45. Pfaff, D. W., and S. Schwartz-Giblin (1988) Cellular mechanisms of female reproductive behaviors. In E. Ogawa, S., Kow, L. M., McCarthy, M. M., Pfaff, D. W., and Knobil and J. Neil et al. (eds): The physiology of Schwartz-Giblin, S. (1991) Midbrain PAG control of reproduction. New York, Raven Press, pp. 1487-1568. female reproductive behavior: In vitro electrophysiological characterization of actions of lordosis-relevant substances. Pfaff, D. W., L. Rosello, and P. Blackburn (1984) Proteins In: A. Depaulis and R. Bandler (eds) The midbrain synthesized in medial hypothalamus and transported to midbrain in estrogen-treated female rats. Exp. Brain Res. periaqueductal gray matter, New York: Plenum Press, pp. 57: 204-207. 211-235. Pfaff, D. W., S. Schwartz-Giblin, M. M. McCarthy, and L. M. Oliveras, J. L, J. M. Besson, G. Guilbaud, and J. C. Liebeskind Kow (1994) Cellular and molecular mechanisms of female (1974) Behavioral and electrophysiological evidence of reproductive behaviors. In E. Knobil and J.D.Neill (eds): pain inhibition from midbrain stimulation in the cat. Exp. The physiology of reproduction, Vol. 2. New York: Raven Brain Res. 20: 32-44. Press, pp.107-220. Oliveras, J. L, F. Redjemi, G. Guilbaud, and J. M. Besson Pfaus, J. G., S. P. Kleopoulus, C. V. Mobbs, R. B. Gibbs, and (1975) Analgesia induced by electrical stimulation of the D. W. Pfaff (1993) Sexual stimulation activates c-fos inferior cetnralis nucleus of the raphe in the cat. Pain 1: within estrogen-concentrating regions of female rat 139-145. forebrain. Brain Res. 624: 253-267. Olmos, G., F. Naftolin, J. Perez, and P. A. Tranque (1989) Portillo, F., L. Grélot, S. Milano, and A. L. Bianchi (1994) Synaptic remodeling in the rat arcuate nucleus during the Brainstem neurons with projecting axons to both phrenic estrous cycle. Neurosci. 32: 663-667. and abdominal motor nuclei: a double fluorescent labeling Olszewski, J., and D. Baxter (1954) Cytoarchitecture of the study in the cat. Neurosci. Lett. 173: 50-54. Human Brain Stem. Switzerland, J.B. Lippincott Pratt, C. A., and G. E. Loeb (1991) Functionally complex Company. muscles of the cat hindlimb. I. Patterns of activation across Olucha, F., F. Martinez-Garcia, and C. Lopez-Garcia (1985) sartorius. Exp. Brain Res. 85: 243-256. A new stabilizing agent for the tetramethyl benzidine Pratt, C. A., C. M. Chanaud, and G. E. Loeb (1991) (TMB) reaction product in the histochemical detection of Functionally complex muscles of the cat hindlimb. IV. horseradish peroxidase (HRP). J. Neurosci. Meth. 13: 131- Intramuscular distribution of movement command signals 138. and cutaneous reflexes in broad, bifunctional thigh Onuf, B. (1900) On the arrangement and function of cell muscles. Exp. Brain Res. 85: 281-299.

153 References

Pullen, A. H. (1988) Quantitative synaptology of feline medulla on mating and responses to genital stimulation motoneurones to external anal sphincter muscle. J. Comp. in female cats. Brain Res. 56: 350-345. Neurol. 269: 414-424. Sakuma, Y., and D. W. Pfaff (1979a) Facilitation of female Reed, A. F. (1940) The nuclear masses of the cervical cord of reproductive behavior from mesencephalic central gray Macaca mulatta. J. Comp. Neurol. 72: 187-206. in the rat. Am. J. Physiol. 237: R278-R284. Rees, R.P., M. B. Bunge, and R. P. Bunge (1976) Morphol- Sakuma Y., and D. W. Pfaff (1979b) Mesencephalic ogical changes in the neuritic growth cone and target mechanisms for integration of female reproductive neuron during synaptic junction development in culture. behavior in the rat. Am. J. Physiol. 237: R285-R290. J. Cell. Biol. 68: 240-263. Sakuma, Y., and D. W. Pfaff (1980a) Cells of origin of Rees, H. D., G. M. Switz, and R. P. Michael (1980) The medullary projections in central gray of rat mesecephalon. estrogen-sensitive neural system in the brain of female J. Neurophysiol. 44: 1002-1011. cats. J. Comp. Neurol. 193: 789-804. Sakuma, Y., and D. W. Pfaff (1980b) Excitability of female Rexed, B. (1954) A cytoarchitectonic atlas of the spinal cord rat central gray cells with medullary projections: changes in the cat. J. Comp. Neurol. 100: 297-380. produced by hypothalamic stimulation and estrogen Reynolds, S. R. M. (1951) Determinants of uterine growth treatment. J. Neurophysiol. 44: 1012-1023. and activity. Physiol. Rev. 1: 244-273. Sakuma, Y., and D. W. Pfaff (1980c) LH-RH in the Richmond, F. J. R., R. Gladdy, J. L. Creasy, S. Kitamura, E. mesencephalic central gray can potentiate lordosis reflex Smits, and D. B. Thomson (1994) Efficacy of seven of female rats. Nature 283: 566-567. retrograde tracers, compared in multiple-labelling studies Sakuma, Y., and D. W. Pfaff (1982) Properties of ventromedial of feline motoneurones. J. Neurosci. Meth. 53: 35-46. hypothalamic neurons with axons to midbrain central gray. Riskind, P., and R. L. Moss (1979) Midbrain central gray. Exp. Brain Res. 46: 292-300. LHRH infusion enhance lordotic behavior in estrogen- Sakuma, Y., and D. W. Pfaff (1983) Modulation of the lordosis primed ovariectomized rats. Brain Res. Bull. 4: 203-205. reflex of female rats by LHRH, its antiserum and analogs Risold, P. Y., N. S. Canteras, and L. W. Swanson (1994) Organ- in the mesencephalic central gray. Neuroendocrinology ization of projections from the anterior hypothalamic 36: 218-224. nucleus: a Phaseolus vulgaris-Leucoagglutinin study in Saper, C. B., L. W. Swanson, and W. M. Cowan (1976) The the rat. J. Comp. Neurol. 348: 1-40. efferent connections of the ventromedial nucleus of the Ritz, L. A., and J. D. Greenspan (1985) Morphological features hypothalamus of the rat. J. Comp. Neurol. 169: 409-442. of lamina V neurons receiving nociceptive input in cat Saper, C. B., L. W. Swanson, and W. M. Cowan (1978) The sacraocaudal spinal cord. J. Comp. Neurol. 238: 440- efferent connections of the anterior hypothalamic area of 452. the rat, cat and monkey. J. Comp. Neurol. 182: 575-600. Ritz, L. A., S. M. Bailey, C. R. Murray, and M. L. Sparkes Saper, C. B., L. W. Swanson, and W. M. Cowan (1979a) An (1992) Organizational and morphological features of cat autoradiographic study of the efferent connections of the sacrocaudal motoneurons. J. Comp. Neurol. 318: 209-221. lateral hypothalamic area in the rat. J. Comp. Neurol. 183: Robbins, A., K. J. Berkley, and Y. Sato (1992) Estrous cycle 669-707. variation of afferent fibers supplying reproductive organs Saper, C. B., L. W. Swanson, and W. M. Cowan (1979b) Some in the female rat. Brain Res. 596: 353-356. efferent connections of the rostral hypothalamus in the Roberts, W. W., M. L. Steinberg, and L. W. Means (1967) squirrel monkey (Saimiri sciureus) and cat. J. Comp. Hypothalamic mechanisms for sexual, aggressive, and Neurol. 184: 205-243. other motivational behaviors in the opossum, Ddelphis Sasaki, M. (1994) Morphological analysis of external urethral virginiana. J. Comp. Physiol. Psychol. 64: 1-15. and external anal sphincter motoneurons of cat. J. Comp. Romanes, G. J. (1941) The development and significance of Neurol. 349: 269-287. the cell-columns in the ventral horn of the cervical and Sasaki, S., H. Uchino, M. Imagawa, T. Miyake, and Y. Uchino upper thoracic spinal cord of the rabbit. J. Anat. 76: 112- (1991) Lower lumbar branching of caudal medullary 130. expiratory neurons of the cat. Brain Res. 553: 159-162. Romanes, G. J. (1951) The motor cell columns of the lumbo- Sasaki, S., H. Uchino, and Y. Uchino (1994) Axon branching sacral spinal cord of the cat. J. Comp. Neurol. 94: 313- of medullary expiratory neurons in the lumbar and sacral 363. spinal cord of the cat. Brain Res. 648: 229-238. Romanes, G. J. (1964) The Motor Pools of the Spinal Cord. Sato, M., N. Mizuno, and A. Konishi (1978) Localization of Progr. Brain Res. 11: 93-119. motoneurons innervating perineal muscles: a HRP study Rose, J. D., and S. L. Bieber (1984) Joint and separate effects in cat. Brain Res. 140: 149-154. of estrogen and progesterone on responses of midbrain Sauerwein, H., and H. H. Meyer (1989) Androgen and estrogen neurons to lordosis-controlling somatic stimuli in the receptors in bovine skeletal muscle: relation to steroid- female golden Syrian hamster. J. Neurophysiol. 51: 1040- induced allometric muscle growth. J. Anim. Sci. 67: 206- 1054. 212. Rose, J. D., and F. W. Flynn (1993) Lordosis response Schaible, H. G., R. F. Schmidt, and W. D. Willis (1987a) components can be elicited in decerebrate rats by Convergent inputs from articular, cutaneous and muscle combined flank and cervix stimulation. Physiol. Behav. receptors onto ascending tract cells in the cat spinal cord. 54: 357-361. Exp. Brain Res. 66: 479-488. Rose, J. D., and J. Sutin (1973) Effects of lesions in the lateral Schaible, H. G., R. F. Schmidt, and W. D. Willis (1987b)

154 References

Enhancement of the responses of ascending tract cells in steroids modulate motor-correlated increases in cerebellar the cat spinal cord by acute inflammation of the knee joint. discharge. Brain Res. 476: 307-316. Exp. Brain Res. 66: 489-499. Södersten, P. ,and K. Larsson (1975) Lordosis behavior and Schreiner-Engel, P., R. C. Schiavi, H. Smith, and D. White mounting behavior in male rats: effects of castration and (1981) Sexual arousability and the menstrual cycle. treatment with estradiol benzoate or testosterone Psychosom. Med. 43: 199-214. propionate. Physiol. Behav. 14: 159-164. Schumacher, M. (1990) Rapid membrane effects of steroid Sohrabji, F., L. A. Greene, R. C. Miranda, and C. D. Toran hormones: an emerging concept in neuroendocrinology. Allerand (1994) Reciprocal regulation of estrogen and Trends in Neurosci. 13: 359-362. NGF receptors by their ligands in PC12 cells. J. Neurobiol. Schwartz-Giblin, S., P. A. Femano, and D. W. Pfaff (1984) 25: 974-988. Axial electromyogram and intervertebral length gauge Sprague, J. M. (1948) A study of motor cell localization in responses during lordosis behavior in rats. Exp. Neurol. the spinal cord of the rhesus monkey. Am. J. Anat. 82: 1- 85: 297-315. 26. Sharrard, W. J. W. (1955) The distribution of the permanent Sterling, P., and H. G. J.M. Kuypers (1967) Anatomical paralysis in the lower limb in poliomyelitis. J. Bone Jt. organization of the brachial spinal cord of the cat. II. The Surg. 37: 540-558. motoneuron plexus. Brain Res. 4: 16-32. Sherrington, C. S. (1892) Notes on the arrangement of some Steward, O., and P. M. Falk (1986) Protein synthetic machinery motor fibers in the lumbo-sacral plexus. J. Physiol. at postsynaptic sites during synaptogenesis: A quantitative (London) 13: 621-772. study of the association between polyribosomes and Sherrington, C. S. (1910) Flexion-reflex of the limb, crossed developing synapses. J. Neurosci. 6: 412-423. extension-reflex, and reflex stepping and standing. J. Steward, O., and P. M. Falk (1991) Selective localization of Physiol. (London) 40: 28-121. polyribosomes beneath developing synapses: aquantitative Shipley, M. T., A. Z. Murphy, T. A. Rizvi, M. Ennis, and M. analysis of the relationships between polyribosomes and M. Behbehani (1996) In: The Emotional Motor System. developing synapses in the hippocampus and dentate G. Holstege, R. Bandler, and C. Saper (eds.). Progr. Brain gyrus. J. Comp. Neurol. 314: 545-557. Res. 107: 355-378. Stuart, D. G., Th. M. Hamm, and S. Vanden Nocen (1988) Shivers, B. D., R. Harlan, J. I. Morrell, and D. W. Pfaff (1983) Partitioning of monosynaptic Ia EPSP connections with Immunocytochemical localization of luteinizing hormone- motoneurons according to neuromusclular topography: releasing hormone in male and female rat brains. generality and functional implications. Progr. Neurobiol. Neuroendocrinology 36: 1-12. 30: 437-447. Silver, M. L. (1942) The motoneurons of the spinal cord of Stumpf, W. E., and M. Sar (1975) Hormone-architecture of the frog. J. Comp. Neurol. 77: 1-39. the mouse brain with 3H-Estradiol. In: Anatomical Sirinathsinghji, D. J. S. (1984) Modulation of lordosis behavior Neuroendocrinology Int. Conf. Neurobiol. of CNS- of female rats by naloxone, ß-endorphin and its antiserum Hormone interactions, Chapel Hill 1974 pp. 82-103. in the mesencephalic cetnral gray: possible mediation via Karger, Basel. GnRH. Neuroendocrinol. 39: 222-230. Stumpf, W. E., and M. Sar (1976) Steroid hormone target sites Skultety, F. M. (1959) Relation of periaqueductal gray matter in the brain: the differential distribution of estrogen, to stomach and bladder motility. Neurology 9: 190-198. progestin, androgen and glucocorticosteroid. J. Ster. Biochem. 7: 1163-1170. Skultety, F. M. (1963) Stimulation of Periaqueductal Gray and Hypothalamus. Arch. Neurol. 8: 608-620. Stumpf, W. E., M. Sar, and D. A. Keefer (1975) Atlas of estrogen target cells in rat brain. In: Anatomical Smith, J. C., D. E. Morrison, H. H. Ellenberger, M. R. Otto, Neuroendocrinology Int. Conf. Neurobiol. of CNS- and J. L. Feldman (1989) Brainstem projections to the Hormone interactions, Chapel Hill 1974 pp. 104-119. major respiratory neuron populations in the medulla of Karger, Basel. the cat. J. Comp. Neurol. 281: 69-96. Smith, P., G. Heimer, A. Norgren, and U. Ulmsten (1990) Swett, J. E., E. Eldred, and J. S. Buchwald (1970) Somatotopic Steroid hormone receptors in pelvic muscles and ligaments cord-to-muscle relations in efferent innervation of cat in women. Gynecol. Obstet. Invest. 30: 27-30. gastrocnemius. Am. J. Physiol. 219: 762-766. Smith, P., G. Heimer, A. Norgren, and U. Ulmsten (1993) Szechtman, H., N. T. Adler, and B. R. Komisaruk (1985) Localization of steroid hormone receptors in the pelvic Mating induces pupillary dilatation in female rats: role of muscles. J. Obstet. Gynecol. and Reprod. Biol. 50: 83- pelvic nerve. Physiol. Behav. 35: 295-301. 85. Tennyson, V. M. (1970) The fine structure of the axon and Smith, S. S. (1994) Female sex steroid hormones: from growth cone of the dorsal root neuroblast of the rabbit receptors to networks to performance-actions on the embryo. J. Cell Biol. 44: 62-79. sensorimotor system. Progr. Neurobiol. 44: 55-86. Tetel, M. J., M. J. Getzinger, and J. D. Blaustein (1993) Fos Smith, S. S., B. D. Waterhouse, and D. J. Woodward (1987) expression in the rat brain following vaginal-cervical Sex steroid effects on extrahypothalamic CNS. I. Estrogen stimulation by mating and manual probing. J. Neuro- augments neuronal responsiveness to iontophoretically endocrinol. 5: 397-404. applied glutamate in the cerebellum. Brain Res. 422: 40- Thompson, T. L., and Moss R.L. (1994) Estrogen regulation 51. of dompamine release in the nucleus accumbens: genomic- Smith, S. S., D. J. Woodward, and J. K. Chapin (1989) Sex and nongenomic-mediated effects. J. Neurochem. 62:

155 References

1750-1756. Estrogen increases the strength of a medullary-lumbo- Thor, K. B., C. Morgan, I. Nadelhaft, M. Houston, and W. C. sacral motoneuronal projection, possibly involved in de Groat (1989) Organization of afferent and efferent lordosis behavior in the cat. Soc. Neurosci. Abstr. 21: pathways in the pudendal nerve of the female cat. J. Comp. 1462. Neurol. 288: 263-279. Vaughn, J. E., and T. J. Sims (1978) Axonal growth cones and Turek, F. W., and E. Van Cauter (1994) Rhythms in developing axonal collaterals form synaptic junctions in reproduction. In: E. Knobil and J. D. Neill (eds.) The embryonic mouse spinal cord. J. Neurocytol. 7: 337-363. physiology of reproduction. Raven Press, New York: Vol. Veening, J., P. Buma, G. J. Ter Horst, T. A. P. Roeling, P. G. 2, pp. 3-106. M. Luiten, and R. Nieuwenhuys (1991) Hypothalamic Ueyama, T., N. Mizuno, S. Nomura, A. Konishi, K. Itoh, and projections to the PAG in the rat: topographical, immuno- H. Arakawa (1984) Central distribution of afferent and electronmicroscopical and funcional aspects.In: A. efferent components of the pudendal nerve in cat. J. Comp. Depaulis and R. Bandler (eds. ) The midbrain peri- aqueductal gray matter. pp. 387-415. Neurol. 222: 38-46. Valentino, R. J., M. E. Page, P. H. Luppi, Y. Zhu, E. Van Vinores, S. A., M. M. Herman, L. J. Rubinstein, and P. J. Bockstaele, and G. Aston-Jones (1994 ) Evidence for Marangos (1984) Electron microscopic localization of widespread afferents to Barrington’s nucleus, a brainstem neuron-specific enolase in rat and mouse brain. J. region rich in corticotropin-releasing hormone neurons. Histochem. Cytochem. 32: 1295-1302. Neuroscience 62: 125-143. Waldeyer, W. (1888) Das Gorilla-Rückenmark. Abh. der Königlichen Akad. der Wissensch. Berlin. Phys.-Math. Van de Poll, N. E., and H. Van Dis (1977) Hormone induced Classe, Abh. III, pp.1-147. lordosis and its relation to sexual activity in male rats. Horm. Behav. 8: 1-7. Walker, A. E. (1940) The spinothalamic tract in man. Arch. Neur. and Psychiat. 43: 284-298. Van de Poll, N. E., and H. Van Dis (1979) The effect of medial preoptic-anterior hypothalamic lesions on bisexual Watkins, L. R., P. L. Faris, B. R. Komisaruk, and D. J. Mayer behavior of the male rat. Brain Res. Bull. 4: 505-511. (1984) Dorsolateral funiculus and intraspinal pathways mediate vaginal stimulation-induced suppression of Van Gehuchten, A., and C. de Neef (1900) Les noyaux moteurs nociceptive responding in rats. Brain Res. 294: 59-65. de la moelle lombo-sacree chez l’homme. Nevraxe. 1. 201. Weeks, O. I., and A. W. English (1985) Compartmentalization VanderHorst, V. G. J.M., and G. Holstege (1992) Afferent of the cat lateral gastrocnemius motor nucleus. J. Comp. projections from the spinal cord to the periaqueductal gray Neurol. 135: 255-267. in the cat. Soc. Neurosci. Abstr. 18: 311. Weeks, O. I., and A. W. English (1987) Cat triceps surae motor VanderHorst, V. G. J. M., and G. Holstege (1995) Caudal nuclei are organized topologically. Exp. Neurol. 96: 163- medullary pathways to lumbosacral motoneuronal cell 177. groups in the cat: evidence for direct projections possibly representing the final common pathway for lordosis . J. Whalen, R. E. (1963) Sexual behaviour of cats. Behaviour Comp. Neurol. 359: 457-475. 20: 321-342. VanderHorst, V. G. J. M., and G. Holstege (1996a) A concept Whipple, B., and B. R. Komisaruk (1985) Elevation of pain for the final common pathway of vocalization and lordosis threshold by vaginal stimulation in women. Pain 21: 357- behavior in the cat. Progr. Brain Res. 107: 327-342. 367. VanderHorst, V. G. J. M., and G. and Holstege (1996b) The White, S. R., S. J. Fung, D. A. Jackson, K. M. Imel (1996) organization of lumbosacral motoneuronal cell groups Serotonin, norepinephrine and associated neuropeptides: innervating hindlimb, pelvic floor and axial muscles in effects on somatic motoneurons excitability. In: The the cat. Submitted to J. Comp. Neurol. Emotional Motor System. G. Holstege, R. Bandler, and C. Saper (eds.). Progr. Brain Res. 107: 183-200. VanderHorst, V. G. J. M., L. J. Mouton, B. F. M. Blok, and G. Holstege Distinct cell groups in the lumbosacral cord of Wiberg, M, and A. Blomqvist (1984a) The spinomesen- the cat project to different areas in the periaqueductal gray. cephalic tract in the cat: its cells of origin and termination J. Comp. Neurol. in press. pattern as demonstrated by the intraaxonal transport method. Brain Res. 291: 1-18. VanderHorst, V. G. J. M., and G. Holstege Nucleus retroambiguus projections to lumbosacral motoneuronal Wiberg, M., and A. Blomqvist (1984b) The projection to the cell groups in the male cat, possibly representing the final mesencephalon from the dorsal column nuclei. An common pathway for mounting behavior. Submitted to J. anatomical study in the cat. Brain Res. 311: 225-244. Comp. Neurol. Wiberg, M., J. Westman, and A. Blomqvist (1987) Somato- VanderHorst, V. G. J. M., and G. Holstege Estrogen induces sensory projection to the mesencephalon: an anatomical axonal outgrowth in the nucleus retroambiguus- study in the monkey. J. Comp. Neurol. 264: 92-117. lumbosacral motoneuronal pathway in the adult female Wild, J. M. (1993) The avian nucleus retroambigualis: a cat. Submitted to J. Neurosci. nucleus for breathing, singing and calling. Brain Res. 606: VanderHorst, V. G. J. M., H. DeWeerd and G. Holstege (1994) 319-324. Direct projections from the nucleus retroambiguus to Wildt, D. E., S. W. Seager, and P. K. Chakraborty (1980) Effect hindlimb motoneurons; a combined electron- and of copulatory stimuli on incidence of ovulation and on lightmicroscopic study in the cat. Soc. Neurosci. Abstr. serum luteinizing hormone in the cat. Endocrinol. 107: 20: 1584. 1212-1217. VanderHorst, V. G. J. M., H. DeWeerd and G. Holstege (1995) Williams, C. L. (1987) Estradiol benzoate facilitates lordosis

156 References

and ear wiggling of 4- to 6-day-old rats. Behav. Neurosci. Zhang, S. P., P. J. Davis, and R. Bandler (1995) Brain stem 101: 718-723. integration of vocalization: Role of the nucleus retro- Williams, C. L., and D. Lorang (1987) Brain transections ambiguus. J. Neurophysiol. 74: 2500-2512. differentially alter lordosis and ear wiggling of 6-day-old rats. Behav. Neurosci. 101: 819-826. Willis, W. D. (1985) The pain system. Basel: Karger. Willis, W. D., and R. E. Coggeshall (1978) Sensory mechanisms of the spinal cord. New York: Plenum. Wong, M., and R. L. Moss (1991) Electrophysiological evidence for a rapid membrane action of the gonadal steroid, 17ß-estradiol, on CA1 pyramidal neurons in the rat hippocampus. Brain Res. 543:148-152. Wong, M., and R. L. Moss (1992) Long-term and short-term electrophysiological effects of estrogen on the synaptic properties of hippocampal CA1 neurons. J. Neurosci. 12: 3217-3225. Yagi, K. (1970) Effects of estrogen on the unit activity of the rat hypothalamus. J Physiol. Soc. Jpn. 32: 692. Yamada K. M., B. S. Spooner, and N. K. Wessells (1971) Ultrastructure and function of growth cones and axons of cultured nerve cells. J. Cell Biol. 49: 614-135. Yamamoto, T., H. Satomi, H. Ise, H. Takatama, and K. Takahashi (1978) Sacral spinal innervations of the rectal and vesical smooth muscles and the sphincteric striated muscles as demonstrated by the horseradish peroxidase method. Neurosci. Lett. 7: 41-48. Yezierski, R. P. (1988) Spinomesencephalic tract: projections from the lumbosacral spinal cord of the rat, cat, and monkey. J. Comp. Neurol. 267: 131-146. Yezierski, R. P., and C. M. Mendez (1991) Spinal distribution and collateral projections of rat spinomesencephalic tract cells. Neuroscience 44: 113-130. Yezierski, R. P., and R. H. Schwartz (1986) Response and receptive-field properties of spinomesencephalic tract cells in the cat. J. Neurophysiol. 55: 76-96. Yezierski, R. P., L. S. Sorkin, and W. D. Willis (1987) Response properties of spinal neurons projecting to midbrain or midbrain-thalamus in the monkey. Brain Res. 437: 165- 170. Young W. C. (1961) The hormones and mating behavior. In: Sex and internal secretions W. C. Young (ed.), Baltimore: Wiliams and Wilkins. Zajac, F. E. (1985) Thigh muscle activity during maximum- height jumps by cats. J. Neurophysiol. 53: 979-994. Zhang, D., S. M. Carlton, L. S. Sorkin, and W. D. Willis (1990) Collaterals of primate spinothalamic tract neurons to the periaqueductal gray. J. Comp. Neurol. 196: 277-290. Zhang, S. P., R. Bandler, and P. Carrive (1990) Flight and immobility evoked by excitatory amino acid micro- injection within distinct parts of the subtentorial midbrain periaqueductal gray of the cat. Brain Res. 520: 73-82. Zhang, S. P., P. J. Davis, P. Carrive, and R. Bandler (1992) Vocalization and marked pressor effect from the region of the nucleus retroambigualis in the caudal ventrolateral medulla of the cat. Neurosci. Lett. 140: 103-107. Zhang, S. P., P. J. Davis, R. Bandler, and P. Carrive (1994) Brain stem integration of vocalization: Role of the midbrain periaqueductal gray. J. Neurophysiol. 72: 1337- 1356.

157 References Abbreviations

AL lateral amygdaloid nucleus BC brachium conjunctivum BIC brachium of the inferior colliculus BL basolateral amygdaloid nucleus BM basomedial amygdaloid nucleus BNST bed nucleus of stria terminalis BNSTL lateral part of the bed nucleus of stria terminalis CA central amygdaloid nucleus Cd caudate nucleus CGMp medial geniculate body, principal part CI capsula interna CL claustrum CO cortical amygdaloid nucleus CS superior colliculus CU nucleus cuneatus CUN cuneiform nucleus Dh dorsal horn ECU external cuneate nucleus En entopeduncular nucleus F fornix G nucleus gracilis IC inferior colliculus IO inferior olive LCN lateral cervical nucleus LH lateral hypothalamic area LRN lateral reticular nucleus LTF lateral tegmental field ML medial lemniscus MLF medial longitudinal fasciculus NRA nucleus retroambiguus OC optic chiasm OT optic tract P pyramidal tract PAG periaqueductal gray PC pedunculus cerebri Pu putamen PVN paraventricular hypothalamis nucleus S solitary complex SON supraoptic nucleus Ssp supraspinal nucleus VMH ventromedial nucleus of the hypothalamus III oculomotor nucleus IV trochlear nucleus Vspin. caud. spinal trigeminal complex pars caudalis XII hypoglossal nucleus

158 Summary

Much of our behavior is not under voluntary control, second and third groups probably play a role in agressive but is regulated by emotions. Emotional behavior, such behavior and pain control. as fighting, flighting, and screaming is crucial for the Among many other emotional behavior components, survival of the individual, whereas reproduction is stimulation and lesion studies have demonstrated that essential for the continuation of the species. The areas the PAG is essential for vocalization (sound production), in the central nervous system controlling this un- and reproductive behavior. Specific projections from the voluntary emotional behavior belong to the so-called PAG to the nucleus retroambiguus (NRA) are important “Emotional Motor System”. The periaqueductal gray for vocalization. The NRA is an elongated, compact (PAG), a large group of neurons surrounding the group of premotor interneurons located at the transition aqueduct of Silvius in the midbrain, plays a crucial role between spinal cord and brain stem. Its interneurons in within the emotional motor system, because it acts as turn project to the motoneuronal cell groups of the the final integrator for the individual components of larynx, pharynx, soft palate, and expiratory muscles. survival behavior. For example, during mating the PAG These pathways have been shown to be involved in not only controls the mating posture, but also regulates expiration, vomiting, and vocalization. In addition to the necessary changes in blood pressure and pain its projections to the larynx, pharynx and abdominal perception. In order to integrate these basic behaviors, straining muscles, the NRA also appeared to project to the PAG receives afferent information via sensory hindlimb motoneurons in the lumbosacral cord. To find pathways and sends the results of the integration to out what the function is of this other NRA pathway, it premotor interneurons in the brain stem and spinal cord. was necessary to precisely identify the lumbosacral The aim of this thesis is to describe these pathways as motoneuronal cell groups receiving NRA input. far as they are related to sexual behavior in general and female receptive and male mounting behavior in particular. SC Sensory en motor pathways can be visualized light- and IC Periaqueductal gray electronmicroscopically by means of neuronal tracers. BC IV

CSN After injecting a tracer in the nervous system, it is taken BP ML

PC up by neuronal tissue and transported from the nerve PON fiber (axon) to the cell body (retrograde transport) and/ or from the cell body to the fiber ending (anterograde

CUN G transport). Vspin. caud. For the integration of receptive behavior, the PAG Nucleus retroambiguus requires input from the pelvic viscera and sex organs. P This information in conveyed to the lumbosacral cord via the pelvic and pudendal nerves, and reaches the PAG L4-S1 largely via direct projections. The first chapter describes Iliopsoas, Adductor, the lumbosacral-PAG pathways and demonstrates that Hamstring, they can be divided into three groups. The first and Axial, and Pelvic floor largest group was present in L7-S3 segments and muscles consisted of small, oval and fusiform neurons. It extended from the dorsolateral part of Rexed’s lamina I in L7, along the lateral part of the dorsal horn in S1 into S2 Group I lamina V of S2. In the lateral part of S2, some of its Perineum, ± µ Vagina, neurons formed clusters with intervals of 230 m. The Cervix location of the first group overlapped extensively with the termination area of pelvic and pudendal afferents. The main midbrain target of the first group was the medial part of the lateral PAG. From its location it can be concluded that this first group is involved in the Figure 1 Schematic representation of the sensory and motor micturition reflex and reproductive behavior. The smaller pathways involved in receptive posture.

159 Therefore, in Chapter 3 the location of 50 different mounting behavior. hindlimb, lower back, and pelvic floor muscles, divided In contrast to the male cat, females only display receptive into 10 groups, has been described. A new method had behavior when they are in estrus (in heat), i.e. when to be developed to compare the results of different cases, there exist high levels of the female sex steroid estrogen. despite variations in size and segmental organization Apparently, the NRA-lumbosacral pathway functions between the different cases. The method is based on the differently during estrous than during non-estrous, which finding that the spatial interrelation between the is studied in Chapter 6 using light- and electron- hindlimb and pelvic floor lumbosacral motoneuronal cell microscopical techniques. Lightmicroscopically, the groups is always the same. With this method it appeared density of labeled NRA fibers in their target hindlimb possible to compose an accurate overall map of the motoneuronal cell groups appeared abundant in estrous location of the lumbosacral motoneuronal cell groups. and very weak in non-estrous cats. Such differences were Chapters 4, 5, and 6 deal with the NRA terminations in not found in the rubrospinal pathway, which served as the lumbosacral cord. NRA interneurons projecting to control. Electronmicroscopical examination of the NRA the lumbosacral cord are located between 1 to 8 mm projection to the semimembranosus motoneuronal cell caudal to the obex. The axons of these cells reach the group revealed an almost ninefold larger number of lumbosacral cord via the contralateral white matter, labeled profiles in estrous versus non-estrous cats. whereas the neurons located ventromedial to the NRA project ipsilaterally. In the female cat, delicate NRA Moreover, the semimembranosus motoneuronal cell projections terminate in distinct lumbosacral moto- group contained labeled growth cones in estrous, but neuronal cell groups innervating specific hindlimb not in non-estrous cats. Apparently, estrogen induces muscles (iliopsoas, adductor longus, and hamstrings), axonal outgrowth of the NRA-motoneuronal pathway axial muscles (medial longissimus and proximal tail in the spinal cord of the adult cat, which retracts when muscles), and pelvic floor muscles. This combination estrogen levels are low. These variations in strenght or of muscles led to the hypothesis that the PAG-NRA- plasticity of the NRA-lumbosacral motoneuronal lumbosacral motoneuronal projection form the final pathway explain the presence of female receptive common pathway for female receptive behavior. In behavior during estrous, and its absence in non-estrous chapter 4, this pathway is described in the non-estrous females. A possible mechanism for estrogen induced female cat, and in Chapter 5 in the male cat. In the male growth of NRA axons is presented. cat, the NRA projects more heavily to the biceps anterior The results of these chapters are discussed in the and adductor longus, and only sparsely to the iliopsoas framework of an overview of the brain stem and spinal and semitendinosus motoneuronal cell groups. These pathways for reproductive and other behaviors. Finally, male-female differences suggest that the PAG-NRA- a concept is presented of the basic spinal-brain stem- lumbosacral projections not only form the final common spinal reflex pathways for receptive and mounting pathway for female receptive behavior, but also for male behavior in the cat.

160 Samenvatting

Een groot deel van ons gedrag wordt niet willekeurig plaats van deze eerste groep. De locatie van deze groep gestuurd, maar wordt bepaald door emoties. Emotioneel cellen duidt erop dat ze van belang is voor de mictiereflex gedrag, zoals vluchten, vechten, schreeuwen en voort- en het voortplantingsgedrag. De veel kleinere, tweede planten is noodzakelijk voor het overleven van het en derde groep spelen mogelijk een rol bij agressief individu en de instandhouding van de soort. De gebieden gedrag en de controle van pijn. in het centrale zenuwstelsel die betrokken zijn bij het Studies waarbij gestimuleerd of gelaedeerd is, hebben tot stand brengen van dit onwillekeurige, emotionele aangetoond dat het PAG essentieel is voor de motorische gedrag worden ondergebracht in het zogenoemde component van onder andere vocalisatie (het produceren Emotioneel Motorisch Systeem. Het periaquaductale van geluid) en paringsgedrag. Voor vocalisatie zijn grijs (PAG), een grote hoeveelheid neuronen die het specifieke projecties van het PAG naar de nucleus retro- aquaduct van Silvius in de middenhersenen omgeeft, ambiguus (NRA) van belang. De NRA is een lang- speelt een centrale rol binnen dit emotioneel motorisch gerekte, compacte groep premotor-interneuronen systeem. Het PAG is verantwoordelijk voor de uit- gelegen op de overgang tussen ruggemerg en hersen- eindelijke integratie van de afzonderlijke componenten stam, die vervolgens naar de motorneuronen van de van overlevingsgedrag. Zo regelt het PAG tijdens de stemband- en de uitademingsspieren projecteert. De voortplanting niet alleen de paringshouding, maar NRA bleek echter ook te projecteren naar motor- produceert het ook de noodzakelijke veranderingen in neuronen in het lumbosacrale ruggemerg, die been- bloeddruk en pijndrempel. spieren innerveren. Om een uitspraak te kunnen doen Dit proefschrift heeft als doel de sensorische en over de functie van deze verbindingen, was het nood- motorische baansystemen te beschrijven in de hersen- zakelijk de lumbosacrale motorneurongroepen te stam en het ruggemerg waarmee het PAG voort- plantingsgedrag regelt. Deze baansystemen zijn licht- en elektronenmicroscopisch zichtbaar gemaakt door middel van zogenaamde neuronale tracers. Nadat deze SC in het zenuwstelsel zijn geïnjecteerd, worden ze door IC Periaquaductale grijs zenuwweefsel opgenomen en getransporteerd van BC IV

CSN zenuwvezel naar cellichaam (retrograad transport) en/ BP ML

PC of van cellichaam naar het uiteinde van de zenuwvezel PON (anterograad transport). Voor de integratie van het paringsgedrag heeft het PAG CUN G Vspin. informatie over de voortplantingsorganen nodig. Deze caud. informatie komt via de n. pelvicus en n. pudendus het Nucleus retroambiguus lumbosacrale ruggemerg binnen en bereikt vervolgens, P voor een belangrijk deel via directe verbindingen, het PAG. Het eerste hoofdstuk beschrijft hoe lumbosacrale L4-S1 Spieren van de neuronen naar het PAG projecteren en laat zien dat deze - achterpoot iliopsoas kunnen worden onderverdeeld in drie groepen. De eerste adductoren hamstrings en grootste groep lumbosacrale cellen die naar het PAG - lage rug projecteert, is gelegen in de ruggemergsegmenten L7- - bekkenbodem S3 en bestaat uit kleine, ovale en fusiforme neuronen. De groep strekt zich uit van het dorsolaterale deel van S2 Groep I lamina I in L7, langs het laterale deel van de dorsale Perineum, Vagina, hoorn in S1 naar lamina V in S2. Sommige van deze Cervix neuronen in het laterale deel van S2 vormen clusters met een interval van ± 230 µm. De ligging van deze groep en van de clusters overlapt in grote mate met het eindigingsgebied van vezels afkomstig van de n. pelvicus Figuur 1 Schema van de sensorische en motorische banen en n. pudendus. In de middenhersenen vormt het mediale in het ruggemerg en de hersenstam die verantwoordelijk zijn deel van het laterale PAG de belangrijkste eindigings- voor de receptieve houding.

161 identificeren waarop de NRA vezels eindigden. In (mounting). Hoofdstuk 2 wordt daarom de locatie van de motor- In tegenstelling tot katers vertonen vrouwelijke katten neurongroepen van 50 verschillende spieren van de alleen paringsgedrag wanneer ze in oestrus (krols) zijn, achterpoot, bekkenbodem en lage rug in de kat be- dat wil zeggen wanneer er voldoende van het vrouwelijk schreven. Een nieuwe methode werd ontwikkeld om de geslachtshormoon oestrogeen aanwezig is. Blijkbaar zijn resultaten van verschillende experimenten te kunnen alleen dan de verbindingen van de NRA naar het vergelijken, ondanks de variaties in grootte en segmen- lumbosacrale ruggemerg functioneel. tale organisatie tussen de verschillende dieren. De Met behulp van zowel licht- als electronenmicros- resultaten laten zien dat de locatie van de groepen copische technieken wordt in Hoofdstuk 5 de vraag motorneuronen van de achterpoot en de bekkenbodem beantwoord of oestrogenen inderdaad een effect hebben onderling constant blijft. Door deze vinding werd het op deze verbinding. Lichtmicroscopisch blijkt de mogelijk om een gedetailleerd overzicht van de ligging dichtheid van NRA-axonen in de groepen motor- van de motorneuronen in het lumbosacrale ruggemerg neuronen van de achterpoot hoog te zijn in krolse en te maken. laag in niet- krolse katten. Dergelijke verschillen werden De hoofdstukken 3, 4 en 5 hebben betrekking op de niet gevonden in de rubrospinale baan, die een rol speelt verbindingen van de NRA naar het lumbosacrale bij de willekeurige aansturing van de spieren van de ruggemerg. In Hoofdstuk 3 worden deze beschreven in ledematen. Bij electronenmicroscopisch onderzoek van de niet-krolse vrouwelijke kat. De axonen van de de NRA- projectie naar de motorneuronen van de m. interneuronen in de NRA die naar het lumbosacrale semimembranosus bleken er bijna negen keer zoveel ruggemerg projecteren, bereiken dit via de contralaterale gelabelde NRA-eindigingen in krolse dan in niet-krolse witte stof, terwijl cellen gelegen naast de NRA katten te zijn. Bovendien waren in deze motorneuron- ipsilateraal projecteren. De NRA-vezels eindigen in groep in krolse katten groeiconussen aanwezig die specifieke groepen motoneuronen die spieren van de onderdeel uitmaakten van NRA-axonen. Dit bleek niet achterpoot (m. iliopsoas, m. adductor longus en de het geval te zijn voor niet-krolse katten. Deze bevin- hamstrings), de lage rug (mm. multifidi, m. longissimus dingen tonen aan dat oestrogenen de eindigingen van medialis en de proximale staartspieren) en bekkenbodem een afdalende baan in het ruggemerg van volwassen innerveren. Deze combinatie van spieren leidde tot het katten kunnen laten uitgroeien en stellen daarmee de concept dat de verbindingen van het PAG via de NRA centrale controle van paringsgedrag in een geheel nieuw naar lumbosacrale motorneuronen ten grondslag liggen perspectief. aan het receptief gedrag van de vrouwelijke kat. De resultaten van deze hoofdstukken zijn samengebracht In Hoofdstuk 4 worden de verbindingen van de NRA in een overzicht van de baansystemen in het ruggemerg naar het lumbosacrale ruggemerg beschreven in de en de hersenstam die een rol spelen bij paringsgedrag mannelijke kat. Ook hier blijkt dat de NRA naar (zie Fig. 1). Bij de vrouwelijke kat kan dit gedrag worden specifieke groepen lumbosacrale motorneuronen pro- opgewekt door specifieke stimuli zoals het prikkelen van jecteert. Er bestaan echter enkele verschillen. In de flanken, het perineum of de vagina. Deze informatie vergelijking tot de vrouwelijke kat zijn de projecties naar komt het centrale zenuwstelsel binnen ter hoogte van de motorneuronen van de m. biceps femoris anterior en het lumbosacrale ruggemerg en wordt direct doorge- de m. adductor longus in de mannelijke kat sterker, maar zonden naar het PAG. Het PAG stuurt vervolgens, via naar de motorneurongroepen van de m. iliopsoas en de interneuronen in de NRA, motoneuronen aan van de m. semitendinosus zwakker. Deze verschillen duiden spieren die bij het paringsgedrag betrokken zijn. Volgens erop dat de verbindingen van de PAG via de NRA naar dit concept is het receptieve gedrag van de kat gebaseerd lumbosacrale motorneuronen niet alleen verantwoor- op specifieke verbindingen binnen het centrale zenuw- delijk zijn voor receptief gedrag in de vrouwelijke kat, stelsel. De wisselende sterkte (plasticiteit) van dit baan- maar ook voor de paringshouding van de kater systeem bepaalt het uiteindelijke effect ervan.

162 Dankwoord

Dit proefschrift sluit een periode af waarin ik heb kunnen leren wat hersenonderzoek is. Vanaf het begin heb ik het gevoel gehad dat het soort onderzoek bij mij pastte en ben al snel gegrepen door het Emotioneel Motorisch Systeem. Het onderzoek vormde (en vormt nog steeds) een uitdaging die ik voorheen niet heb kunnen vinden. Mogelijk zou ik nu nog steeds op zoek zijn, ware het niet dat Prof. Dr. G. Holstege niet net zijn intrede had gedaan in Groningen. Daarom, beste Gert, wil ik je als eerste bedanken voor deze kans, voor je vertrouwen in een goede afloop en voor de ruimte die ik gekregen heb om mijn draai te kunnen vinden. Verder zal ik nog lang terugdenken aan al die zaterdagen die nodig waren om ongestoord de manuscripten door te nemen.

Velen hebben direct of indirect (al was het maar in de vorm van koekjes) bijgedragen aan dit proefschrift. In het bijzonder wil ik de volgende personen hiervoor bedanken: Drs. L. J. Mouton, beste Noor. Jij hebt me de principes van het experimenteren en het verwerken van gegevens bijgebracht, waarbij ik de voordelen van een gestructureerde manier van werken heb mogen ervaren. Je opgeruimde natuur heeft verder kunnen voorkomen dat mijn deel van “onze kamer” daadwerkelijk in een opslagruimte voor oud papier is veranderd.

Mw. Ing. E. Meijer en Dhr. K. van Linschoten, beste Ellie en Klaas. In dit proefschrift is de informatie verwerkt van duizenden coupes die één voor één gesneden, gesleept en gepoetst moesten worden. Dankzij jullie is dit werk nu al klaar en niet pas in het jaar 2000.

Dhr. H. de Weerd, beste Henk. Ik wil je bedanken voor je inspanningen en geduld om gelabelde eindigingen op te sporen en vast te leggen. De uiteindelijke resultaten zijn van grote waarde geweest voor Hoofdstuk 5.

Dhr. P. van der Sijde, beste Peter. Harder of zachter, lichter of donkerder, groter of kleiner, groener, blauwer of roder, digitaal of “lekker ouderwets”, uiteindelijk heb je van iedere foto iets moois weten te maken.

“de Boerderij” in Nijmegen. Alle betrokkenen van “de Boerderij” wil ik bedanken voor de prettige samenwerking die ertoe heeft bijgedragen dat ik op tijd de juiste experimenten hebben kunnen doen.

Dr. H. van der Want en Dr. J. Holstege, beste Han en Joan. Ik wil jullie bedanken voor jullie advies met betrekking tot het electronen microscopische gedeelte van het onderzoek. Ik heb nog wel even wat tijd nodig om álles uit te werken.

Dr. P. O. Gerrits, beste Peter. Bedankt voor al je goede adviezen, niet alleen die met betrekking tot de afwerking van dit proefschrift!

Voor allerhande probleempjes heb ik altijd kunnen rekenen op de hulp van Gerrie (de barra’s houden we erin!), Jan (je wist heel wat overstromingen te voorkomen), Lenka (sorry voor de dodelijke blikken), Fritsina (nu krijg ik tenminste weer post), Bertil (voor een onuitputtelijke hoeveelheid literatuur), Jos (ondanks het “foute” computermerk) en Roel (voor de tangconstructie).

Diane, with your defence in mind nothing can go wrong!

Lieve papa en mama. Dit proefschrift is ook een beetje van jullie. De nodige nieuwsgierigheid is er altijd al geweest, maar van jullie heb ik de instelling meegekregen om eruit te halen wat erin zit. Bovendien hebben jullie me een grote mate van vrijheid gegeven, en op de juiste momenten wat goede raadgevingen, waardoor ik altijd met een gerust hart mijn gang heb kunnen gaan.

163 Publications

VanderHorst, V. G. J. M., and G. Holstege (1995) Caudal medullary pathways to lumbosacral motoneuronal cell groups in the cat: evidence for direct projections possibly representing the final common pathway for lordosis. J. Comp. Neurol. 359: 457-475.

VanderHorst, V. G. J. M., and G. Holstege (1996) A concept for the final common pathway of vocalization and lordosis behavior in the cat. Progr. Brain Res. 107: 327-341.

VanderHorst, V. G. J. M., L. J. Mouton, B. F. M. Blok, and G. Holstege Distinct cell groups in the lumbosacral cord of the cat project to different areas in the periaqueductal gray. J. Comp. Neurol. in press.

VanderHorst, V. G. J. M., and G. Holstege The organization of lumbosacral motoneuronal cell groups innervating hindlimb, pelvic floor and axial muscles in the cat. Accepted for J. Comp. Neurol.

VanderHorst, V. G. J. M., and G. Holstege Nucleus retroambiguus projections to lumbosacral motoneuronal cell groups in the male cat, possibly representing the final common pathway for mounting behavior. Accepted for J. Comp. Neurol.

VanderHorst, V. G. J. M., and G. Holstege Estrogen induces axonal outgrowth in the nucleus retroambiguus- lumbosacral motoneuronal pathway in the adult female cat. Submitted to J. Neurosci.

VanderHorst, V. G. J. M., and G. Holstege Estrogen induced sprouting in a sexual motor pathway in adult female cats. Submitted to Science

Holstege, G., L. Kerstens, M. C. Moes, and V. G. J. M. VanderHorst Evidence for a periaqueductal gray - nucleus retroambiguus - spinal cord pathway in the rat. Submitted to Neurosci.

Abstracts

VanderHorst, V. G.J.M., and G. Holstege (1992) Afferent projections from the spinal cord to the periaqueductal gray in the cat. Neurosci. Abstr. 18(1): 311

VanderHorst, V. G.J.M., and G. Holstege (1993) Nucleus retroambiguus projections to hindlimb motoneurons in the cat; involvement of the Emotional Motor System. The Physiologist 36(4): A-23 Graduate Student Award of the American Physiological Society

VanderHorst, V. G.J.M., and G. Holstege (1994) Specific projections from the nucleus retroambiguus to axial and hindlimb motoneurons in the cat. The Neural Control of Movement, Maui, Hawaii

VanderHorst, V. G. J. M., H. DeWeerd, and G. Holstege (1994) Direct projections from the nucleus retroambiguus to hindlimb motoneurons; a combined electron- and lightmicroscopic study in the cat. Neurosci. Abstr. 2: 1584.

VanderHorst, V. G. J. M., H. DeWeerd, and G. Holstege (1995) Estrogen increases the strength of a medullary- lumbosacral motoneuronal projection, possibly involved in lordosis behavior in the cat. Neurosci. Abstr. 21: 1462.

164 Curriculum vitae

Veronique van der Horst is geboren als derde in een serie van vier, in Heeze op 21 mei 1969. In het bosrijke Leende heeft ze kennis gemaakt met alles wat fladdert en rondspringt. Ze heeft het VWO gevolgd op het Hertog Jan College te Valkenswaard (1981-1984) en het Mencia de Mendoza te Breda (1984-1987). In 1986 is ze geselecteerd om op uitnodiging van de duitse regering een maand lang in Duitsland te wonen en te reizen. Na het eindexamen is ze in Groningen gaan wonen om Geneeskunde te studeren. Hoogtepunt van deze periode was de beklimming van de Mont Blanc in 1990. Dit avontuur vormde vervolgens de aanleiding tot een eerste kennismaking met de atletiek. Na het behalen van het laatste doctoraal tentamen in 1991 (cum laude) heeft ze besloten om “de witte jas” nog even te laten hangen om eens wat anders te doen en heeft ze zich opgegeven als studentassistent Hoofd-Hals bij de vakgroep Anatomie. Binnen enkele weken verplaatsten de werkzaamheden zich echter van de snijzaal naar het laboratorium. In april 1992 is ze vervolgens als AIO in dienst getreden bij deze vakgroep. Voor een deel van haar onderzoek kreeg ze in 1993 de “Graduate Student Award” van de “American Physiological Society” toegekend.

165