The Efferent Connections of the Feline Nucleus Cuneatus PETER J . HAND A N D THOMAS VAN WINKLE Department of Animal Biology, School of Veterinary Medicine and Institute of Neurological Sciences, University of Pennsylvania, Philadelphia, Pennsylvania 191 74
ABSTRACT Nucleus cuneatus projections to nucleus ventralis posterolateralis pars medialis (VPLm) and other thalamic as well as midbrain and medullary nuclei were studied in cats using the Fink-Heimer I silver technique. Single electrolytic lesions of very smiill size were made stereotaxically in different zones of nucleus cuneatus under electrophysiological control. All zones studied projected to contralateral VPLm in a pattern of discrete terminal arborizations or clusters, which were organized in onionskin-like dorso-ventral laminae. The clusters of degeneration varied in size and density according to their dorsoventral location within VPLm. Those in dorsal areas were smaller in diameter (50-125 cL) and contained less dense amounts of degeneration than clusters (150-300 p) in more ventral regions. The clustered terminal arborizations mirrored the organization of the VPLm neuronal clusters, themselves. Terminations within VPLnn were topographically organized, but were completely inverted, i.e. dorsal nucleus cuneatus projected to ventral VPLm and ventral to dorsal, lateral to medial, and medial to lateral VPLm. A ventral zone of nucleus cuneatus, which contained "deep" units, projected to a separate dorsal zone of VPLm. In addition to its classical connection with VPLm, nucleus cuneatus projected to the following contralateral brainstem or thalamic nuclei: medial and dorsal accessory olives, external nucleus of the inferior colliculus, ventrolateral part of the superior colliculus, nucleus ruber, medial geniculate nucleus pars magnocellularis, suprageniculatus, medial and lateral divisions of the posterior thalamic nuclear group, zona incerta, and Fields of Forel. Very sparse amounts of degenerat:ion were also present within nuclei ventralis posteromedialis (caudal pole) and ventralis posterolateralis pars lateralis. The brainstem and thalamic projections of the dorsocaudal part (cell nest region) of the cuneate nucleus were more restricted than those of its rostral and ventral regions. The clusters of both the VPLm neurons and cuneate terminations within VPLm provides a n anatomicall basis for the functional characteristics of synaptic security, fine grain somatotopia and modality specificity so prominent in the dorsal column nuclei-medial lemniscal system. Several anatomical and functional studies have demonstrated the relationship of the dorsal column nuclei (DCN), including nucleus cuneatus, the forelimb projection nucleus of the dorsal column system, with the thalamic nucleus ventralis posterolateralis (VPL). Anatomical studies revealed that the cuneate and gracile (hindlimb projection nucleus) nuclei project densely to a medial and a lateral part of contralateral VPL, VPLm and VlPLl (Wnvik, '68), respectively (LeGros Clark, '36; Bowsher, '58; Hand and Liu, '66; Lund and Webster, '67a; Boivie, '71b; Groenewegen et al., '75). Functional studies have J. COMP. NEUR., 171: 83-110.
shown the neurons in this system to be both modality and place specific in that a single neuron responds to only one type of stimulus (e.g., hair movement) applied to a small peripheral receptive field (Mountcastle and Henneman, '49; Kuhn, '49; Mountcastle, '57; Poggio and Mountcastle, '60; Gordon and Paine, '60; Kruger et al., '61; Per1 et al., '62; Gordon and Jukes, '64; Anderson et al., '64). In addition, VPL (also true of other levels of the dorsal col1 This investigation was supported, in part, by U.S. Public Health Service Grants NS-08410 and NS-06716. ZPreliminary reports of this work have been published (Hand and V a n Winkle, '75a,b).
83
84
PETER J . HAND AND THOMAS VAN WINKLE
umn system - Mountcastle, ’57; For a review of DCN topographical organization see Norton and Kruger, ’73) has been shown to possess a detailed somatotopic organization; e.g., the distal forelimb is represented within its ventral region and the proximal body parts more dorsally in the nucleus (Mountcastle and Henneman, ’49; Rose and Mountcastle, ’52; Poggio and Mountcastle, ’60; Anderson et al., ’64). Because the dorsal column system possesses neurons, which have such specific functional characteristics, we have an opportunity to make more exacting anatomical and physiological correlations. However, whereas single unit studies have provided us with detailed information, anatomical investigations because of either the stain used (e.g., Marchi-Ranson and Ingram, ’32; LeGros Clark, ’36; Matzke, ’51) and/ or large lesions employed (Bowsher, ’58, ’61; Kuypers and Tuerk, ’64; Hand and Liu, ’66; Lund and Webster, ’67a; Boivie, ’71b), or the “spread of the injected radioactive amino acids (Jones and Burton, ’74; Groenewegen et al., ’75) have not provided us with comparable detailed information. Therefore, this anatomical study has been designed to determine the detailed relationship of nucleus cuneatus with neurons in VPLm. In addition the answers to two other questions were sought: (1) Does the cuneate nucleus, as was suggested from earlier studies (Ranson and Ingram, ’32; Busch, ’61; Bowsher, ’61; Kuypers and Tuerk, ’64; Hand and Liu, ’66; Lund and Webster, ’67a; Ebbesson, ’68; Boesten and Voogd, ‘75; Groenewegen et al., ’75; Worden and Berkley, ’75) possess additional brainstem and thalamic projections? and (2) if so, is there a differential projection from its anatomically and functionally different regions (Hand and Liu, ’66; Lund and Webster, ’67a; Robards, ’72)? In order to answer these questions, small lesions were made within different zones of nucleus cuneatus under electrophysiological control and the resultant anterograde degeneration stained using a silver technique. The presence of an onionskin pattern of clustered terminations within VPLm, the detailed somatotopia of cuneothalamic projections, and the presence of “extra VPLm” projections primarily from
ventral and rostra1 zones of the cuneate nucleus are described and discussed. MATERIALS AND METHODS
Fifteen adult cats were anesthetized with pentobarbital sodium (36-mg/kg, intraperitoneally), placed in a David Kopf stereotaxic instrument, and either a tungsten microelectrode (5 tip diameter with 1520 p uninsulated tip) or a concentric stainless steel macroelectrode (0.25 mm in diameter with 0.5 mm exposed tip) inserted into the nucleus cuneatus by direct visualization of the overlying fasciculus cuneatus. Only one penetration was made within each nucleus. Prior to lesioning, the somatotopic subdivision of the nucleus cuneatus was determined by evoking neural activity using natural stimulation of the body somata. The stimuli included stroking of hair with a camel’s hair brush, light or stronger taps with the tip of the brush; kneading of deep tissues and manipulations of joints. Activity evoked in a cluster of cuneate units was displayed on a Tektronix 565 oscilloscope and an audiomonitor. The location of the receptive field was determined and a unit cluster categorized as responding to either superficial or deep stimuli after Millar (‘73). This determination was used for the purpose of localizing the recordinglesioning electrode within either the dorsal, cutaneous or “superficial unit” zone of NC or within its more ventrally located “deep unit” zone (Millar and Basbaum, ’75). The lesions were made by passing current (8-10 p A for 8-10 seconds, microelectrode; 1 mA for 1 second, macroelectrode) through the recording electrode. In all animals, excepting one, bilateral lesions were made for more accurate comparisons of the patterns of degeneration. This was made possible because a unilateral lesion results in unilateral (contralateral) brainstem and thalamic degeneration only. In order to minimize damage to neural tissue as a result of the recording procedure itself, each session was kept as short as possible usually being completed within 10 to 15 minutes. Seven days postoperatively, the cats were anesthetized and sacrificed by intracardial perfusion and fixation with warmed (38°C) normal saline and 10%
NUCLEUS CUNEATUS PROJECTIONS
formalin (Koenig et al., '45; Keller and Hand, '70). The brains were sliced coronally into 5 mm blocks, stored in 2 % sucrose, ,and frozen sectioned transversely at 40 pm. Histological sections were stained according to the Fink-Heimer I modification of the Nauta silver technique, which selectively stains degenerating axons, including their synaptic endings (Fink and Heimler, '67). Companion sections were stained with cresyl fast violet in order to identify thle lesion sites as well as brainstem and thalamic sites possessing axonal degeneration. Medullary and midbrain nuclei were delineated after Taber ('61) and thalamilc nuclei after Rinvik ('68) or Jones and BurtlDn ('74). Axonal degeneration was observed in the light microscope and plotted on projection drawings of the brainstem and thalamus. Because of their ideal localizatioii within particular zones of the cuneate nucleus only 12 of 28 successful lesions will be discussed.
85
of both the smaller (12 pm) and larger (20 pm) neurons, but mainly those 20 pm in size, are separated from one another by cell sparse and fiber rich zones (septal regions) (fig. 1). The cell clusters vary in diameter according to their dorso-ventral position within VPLm (also true of VPLl) (figs. 1A,B). In the most ventral parts of VPLm, they range in diameter from 175275 pm, while more dorsally they are 125150 pm in diameter (figs. 1A,B). In the most dorsal zone of VPLm, the clusters are less obvious, but range in size from 75-125 pm. The larger clusters contain more neurons than the smaller ones. For example, a cluster 275 pm in diameter contained 501 neurons, while clusters 125 pm in diameter contained 47 neurons. Thus as one proceeds more ventrally in VPLm, the clusters become larger in diameter and contain more neurons.
General pattern of terminations in VPLm A striking pattern of axonal degeneration is observed in VPLm following small lesions RESULTS in nucleus cuneatus (also true of VPLl folCytoarchitecture - VPLm lowing small lesions of the gracile nucleus, Several investigations of the cytoarchitec- unpublished observations). Terminal deture of the thalamus, including nucleus generation is in the form of dense clusters ventralis posterolateralis, are in existence 50-300 pm in diameter (figs. 2 4 ) , which (Rioch, '29; Scheibel and Scheibel, '66; are associated with the VPLm neuronal Rinvik, '68; Ralston and Herman, '69; clusters (fig. 1). The degeneration clusters Strick, '73; Jones and Burton, "74) and are organized into onionskin-like dorsothus such a discussion will not be repeated ventral laminae (figs. 3, 6, 8, 9, 12) which in detail in this study. However, because extend caudo-rostrally for varying distances. of the clustered pattern of dorsal column With large lesions as were used initially in nuclei terminations in VPLm, a descrip- this study, the clustered and onionskin tion of the general organization of the pattern of degeneration becomes obscured. In parallel with the size and cell density VPLm neurons themselves is necessary. We have observed that the cells of VPLm differences of the neuronal clusters in the are multipolar in shape and generally of dorsoventral plane of VPLm, are the size two sizes: 12 pm and 20 pm in diameter. and density differences of the arborizations Within the dorsal portion of VPLrn imme- of the cuneothalamic projections. Clusters diately adjacent to the small-celleld transi- within dorsal VPLm are smaller in diamtional (Strick, '73) or spinal afferent (Jones eter (50-125 pm) and contain less dense and Burton, '74) zone between VPL and degeneration than those (150-300 pm) in thalamic nucleus ventralis later
ABSTRACT Nucleus cuneatus projections to nucleus ventralis posterolateralis pars medialis (VPLm) and other thalamic as well as midbrain and medullary nuclei were studied in cats using the Fink-Heimer I silver technique. Single electrolytic lesions of very smiill size were made stereotaxically in different zones of nucleus cuneatus under electrophysiological control. All zones studied projected to contralateral VPLm in a pattern of discrete terminal arborizations or clusters, which were organized in onionskin-like dorso-ventral laminae. The clusters of degeneration varied in size and density according to their dorsoventral location within VPLm. Those in dorsal areas were smaller in diameter (50-125 cL) and contained less dense amounts of degeneration than clusters (150-300 p) in more ventral regions. The clustered terminal arborizations mirrored the organization of the VPLm neuronal clusters, themselves. Terminations within VPLnn were topographically organized, but were completely inverted, i.e. dorsal nucleus cuneatus projected to ventral VPLm and ventral to dorsal, lateral to medial, and medial to lateral VPLm. A ventral zone of nucleus cuneatus, which contained "deep" units, projected to a separate dorsal zone of VPLm. In addition to its classical connection with VPLm, nucleus cuneatus projected to the following contralateral brainstem or thalamic nuclei: medial and dorsal accessory olives, external nucleus of the inferior colliculus, ventrolateral part of the superior colliculus, nucleus ruber, medial geniculate nucleus pars magnocellularis, suprageniculatus, medial and lateral divisions of the posterior thalamic nuclear group, zona incerta, and Fields of Forel. Very sparse amounts of degenerat:ion were also present within nuclei ventralis posteromedialis (caudal pole) and ventralis posterolateralis pars lateralis. The brainstem and thalamic projections of the dorsocaudal part (cell nest region) of the cuneate nucleus were more restricted than those of its rostral and ventral regions. The clusters of both the VPLm neurons and cuneate terminations within VPLm provides a n anatomicall basis for the functional characteristics of synaptic security, fine grain somatotopia and modality specificity so prominent in the dorsal column nuclei-medial lemniscal system. Several anatomical and functional studies have demonstrated the relationship of the dorsal column nuclei (DCN), including nucleus cuneatus, the forelimb projection nucleus of the dorsal column system, with the thalamic nucleus ventralis posterolateralis (VPL). Anatomical studies revealed that the cuneate and gracile (hindlimb projection nucleus) nuclei project densely to a medial and a lateral part of contralateral VPL, VPLm and VlPLl (Wnvik, '68), respectively (LeGros Clark, '36; Bowsher, '58; Hand and Liu, '66; Lund and Webster, '67a; Boivie, '71b; Groenewegen et al., '75). Functional studies have J. COMP. NEUR., 171: 83-110.
shown the neurons in this system to be both modality and place specific in that a single neuron responds to only one type of stimulus (e.g., hair movement) applied to a small peripheral receptive field (Mountcastle and Henneman, '49; Kuhn, '49; Mountcastle, '57; Poggio and Mountcastle, '60; Gordon and Paine, '60; Kruger et al., '61; Per1 et al., '62; Gordon and Jukes, '64; Anderson et al., '64). In addition, VPL (also true of other levels of the dorsal col1 This investigation was supported, in part, by U.S. Public Health Service Grants NS-08410 and NS-06716. ZPreliminary reports of this work have been published (Hand and V a n Winkle, '75a,b).
83
84
PETER J . HAND AND THOMAS VAN WINKLE
umn system - Mountcastle, ’57; For a review of DCN topographical organization see Norton and Kruger, ’73) has been shown to possess a detailed somatotopic organization; e.g., the distal forelimb is represented within its ventral region and the proximal body parts more dorsally in the nucleus (Mountcastle and Henneman, ’49; Rose and Mountcastle, ’52; Poggio and Mountcastle, ’60; Anderson et al., ’64). Because the dorsal column system possesses neurons, which have such specific functional characteristics, we have an opportunity to make more exacting anatomical and physiological correlations. However, whereas single unit studies have provided us with detailed information, anatomical investigations because of either the stain used (e.g., Marchi-Ranson and Ingram, ’32; LeGros Clark, ’36; Matzke, ’51) and/ or large lesions employed (Bowsher, ’58, ’61; Kuypers and Tuerk, ’64; Hand and Liu, ’66; Lund and Webster, ’67a; Boivie, ’71b), or the “spread of the injected radioactive amino acids (Jones and Burton, ’74; Groenewegen et al., ’75) have not provided us with comparable detailed information. Therefore, this anatomical study has been designed to determine the detailed relationship of nucleus cuneatus with neurons in VPLm. In addition the answers to two other questions were sought: (1) Does the cuneate nucleus, as was suggested from earlier studies (Ranson and Ingram, ’32; Busch, ’61; Bowsher, ’61; Kuypers and Tuerk, ’64; Hand and Liu, ’66; Lund and Webster, ’67a; Ebbesson, ’68; Boesten and Voogd, ‘75; Groenewegen et al., ’75; Worden and Berkley, ’75) possess additional brainstem and thalamic projections? and (2) if so, is there a differential projection from its anatomically and functionally different regions (Hand and Liu, ’66; Lund and Webster, ’67a; Robards, ’72)? In order to answer these questions, small lesions were made within different zones of nucleus cuneatus under electrophysiological control and the resultant anterograde degeneration stained using a silver technique. The presence of an onionskin pattern of clustered terminations within VPLm, the detailed somatotopia of cuneothalamic projections, and the presence of “extra VPLm” projections primarily from
ventral and rostra1 zones of the cuneate nucleus are described and discussed. MATERIALS AND METHODS
Fifteen adult cats were anesthetized with pentobarbital sodium (36-mg/kg, intraperitoneally), placed in a David Kopf stereotaxic instrument, and either a tungsten microelectrode (5 tip diameter with 1520 p uninsulated tip) or a concentric stainless steel macroelectrode (0.25 mm in diameter with 0.5 mm exposed tip) inserted into the nucleus cuneatus by direct visualization of the overlying fasciculus cuneatus. Only one penetration was made within each nucleus. Prior to lesioning, the somatotopic subdivision of the nucleus cuneatus was determined by evoking neural activity using natural stimulation of the body somata. The stimuli included stroking of hair with a camel’s hair brush, light or stronger taps with the tip of the brush; kneading of deep tissues and manipulations of joints. Activity evoked in a cluster of cuneate units was displayed on a Tektronix 565 oscilloscope and an audiomonitor. The location of the receptive field was determined and a unit cluster categorized as responding to either superficial or deep stimuli after Millar (‘73). This determination was used for the purpose of localizing the recordinglesioning electrode within either the dorsal, cutaneous or “superficial unit” zone of NC or within its more ventrally located “deep unit” zone (Millar and Basbaum, ’75). The lesions were made by passing current (8-10 p A for 8-10 seconds, microelectrode; 1 mA for 1 second, macroelectrode) through the recording electrode. In all animals, excepting one, bilateral lesions were made for more accurate comparisons of the patterns of degeneration. This was made possible because a unilateral lesion results in unilateral (contralateral) brainstem and thalamic degeneration only. In order to minimize damage to neural tissue as a result of the recording procedure itself, each session was kept as short as possible usually being completed within 10 to 15 minutes. Seven days postoperatively, the cats were anesthetized and sacrificed by intracardial perfusion and fixation with warmed (38°C) normal saline and 10%
NUCLEUS CUNEATUS PROJECTIONS
formalin (Koenig et al., '45; Keller and Hand, '70). The brains were sliced coronally into 5 mm blocks, stored in 2 % sucrose, ,and frozen sectioned transversely at 40 pm. Histological sections were stained according to the Fink-Heimer I modification of the Nauta silver technique, which selectively stains degenerating axons, including their synaptic endings (Fink and Heimler, '67). Companion sections were stained with cresyl fast violet in order to identify thle lesion sites as well as brainstem and thalamic sites possessing axonal degeneration. Medullary and midbrain nuclei were delineated after Taber ('61) and thalamilc nuclei after Rinvik ('68) or Jones and BurtlDn ('74). Axonal degeneration was observed in the light microscope and plotted on projection drawings of the brainstem and thalamus. Because of their ideal localizatioii within particular zones of the cuneate nucleus only 12 of 28 successful lesions will be discussed.
85
of both the smaller (12 pm) and larger (20 pm) neurons, but mainly those 20 pm in size, are separated from one another by cell sparse and fiber rich zones (septal regions) (fig. 1). The cell clusters vary in diameter according to their dorso-ventral position within VPLm (also true of VPLl) (figs. 1A,B). In the most ventral parts of VPLm, they range in diameter from 175275 pm, while more dorsally they are 125150 pm in diameter (figs. 1A,B). In the most dorsal zone of VPLm, the clusters are less obvious, but range in size from 75-125 pm. The larger clusters contain more neurons than the smaller ones. For example, a cluster 275 pm in diameter contained 501 neurons, while clusters 125 pm in diameter contained 47 neurons. Thus as one proceeds more ventrally in VPLm, the clusters become larger in diameter and contain more neurons.
General pattern of terminations in VPLm A striking pattern of axonal degeneration is observed in VPLm following small lesions RESULTS in nucleus cuneatus (also true of VPLl folCytoarchitecture - VPLm lowing small lesions of the gracile nucleus, Several investigations of the cytoarchitec- unpublished observations). Terminal deture of the thalamus, including nucleus generation is in the form of dense clusters ventralis posterolateralis, are in existence 50-300 pm in diameter (figs. 2 4 ) , which (Rioch, '29; Scheibel and Scheibel, '66; are associated with the VPLm neuronal Rinvik, '68; Ralston and Herman, '69; clusters (fig. 1). The degeneration clusters Strick, '73; Jones and Burton, "74) and are organized into onionskin-like dorsothus such a discussion will not be repeated ventral laminae (figs. 3, 6, 8, 9, 12) which in detail in this study. However, because extend caudo-rostrally for varying distances. of the clustered pattern of dorsal column With large lesions as were used initially in nuclei terminations in VPLm, a descrip- this study, the clustered and onionskin tion of the general organization of the pattern of degeneration becomes obscured. In parallel with the size and cell density VPLm neurons themselves is necessary. We have observed that the cells of VPLm differences of the neuronal clusters in the are multipolar in shape and generally of dorsoventral plane of VPLm, are the size two sizes: 12 pm and 20 pm in diameter. and density differences of the arborizations Within the dorsal portion of VPLrn imme- of the cuneothalamic projections. Clusters diately adjacent to the small-celleld transi- within dorsal VPLm are smaller in diamtional (Strick, '73) or spinal afferent (Jones eter (50-125 pm) and contain less dense and Burton, '74) zone between VPL and degeneration than those (150-300 pm) in thalamic nucleus ventralis later
