Scanning Electron Microscopy

Volume 1985 Number 1 1985 Article 24

10-5-1984

The Ependyma of the Cat Central Canal, with Particular Reference to its Mitochondria-Containing Bulbs

K. Rascher University of the Saarland

K. H. Booz University of the Saarland

A. C. Nacimiento University of the Saarland

E. Donauer University of the Saarland

Follow this and additional works at: https://digitalcommons.usu.edu/electron

Part of the Biology Commons

Recommended Citation Rascher, K.; Booz, K. H.; Nacimiento, A. C.; and Donauer, E. (1984) "The Ependyma of the Cat Central Canal, with Particular Reference to its Mitochondria-Containing Bulbs," Scanning Electron Microscopy: Vol. 1985 : No. 1 , Article 24. Available at: https://digitalcommons.usu.edu/electron/vol1985/iss1/24

This Article is brought to you for free and open access by the Western Dairy Center at DigitalCommons@USU. It has been accepted for inclusion in Scanning Electron Microscopy by an authorized administrator of DigitalCommons@USU. For more information, please contact [email protected]. SCANNING ELECTRON MICROSCOPY/ 1985 / I (Pa ge l 231-2 38) 0586- 5581/85$1 . 00+. 0S SEM I nc.., AMF O'Ha11.e (Chi c.ago), IL 6066 6 USA

THE EPENDYMA OF THE CAT CENTRAL CANAL, WITH PARTICULAR REFERENCE TO ITS MITOCHONDRIA-CONTAINING BULBS

K. Rascher*, K.H. Booz*, A.C. Nacimiento** and E. Donauer***

*Department of Anatomy, **Neurosurgery Research Laboratory, ***Department of Neurosurgery, University of the Saarland, 6650 Homburg, Federal Republic of Germany

(Paper received March 20 1984, Completed manuscript rec e ived October 5 1984)

Introduction Abstract Most scanning electron microscope The ultrastructure of the ependyma studies of the ependyma have dealt with in the central canal of adult cats was the lining of the brain ventricles (see examined in both the scanning and the extensive literature in Leonhardt 1980 transmission electron microscopes (SEM and Low 1982). The ependyma of the cen­ and TEM). tral canal, however, has only rarely The same morphological details were been subjected to intensive ultrastruc­ seen in the ependyma of the central can­ tural examination (Leonhardt 1976). One al as have so frequently been described of the reasons for this relative lackof in the ependyma of the brain ventricular information on the central canal may be system, for example bundles of cilia, that its free surfaces are considerably single cilia, microvilli and occasional more difficult to expose for observation small cytoplasmic protrusions. The in the SEM than are the surfaces of the supraependymal cells and supraependymal ventricles. Another may be that the cen­ nerve fibers found in the central canal tral canal has not very often been the also resembled those seen in the ven­ subject of experimental and/or patholo­ tricular system. gical research (Hall et al. 1975). The The most striking feature of the present study was undertaken in connec­ canal ependyma were the large, spherical tion with an ongoing investi gation of bodies containing numerous mitochondria. the alterations in the c e ntral canal They are therefore called mitochondria­ caused by experimentall y induced hydro­ containing bulbs. In sections the bulbs cephalus in adult cats. were seen to be connected by long, The ependyma of the cat centralcan­ slender stalks to neurons in subependy­ al has many features in common with that mal position. In some respects the of the ventricles of higher vertebrates mitochondria-containing bulbs resemble and is quite similar to that of the rab­ the processes of cerebrospinal fluid­ bit central canal (Leonhardt 1976) but contacting neurons. does differ in several respects fromthat of the rat canal (Nakayama and Kohno 1974) Both SEM and TEM revealed that the catpo­ ssesses spherical supraependymal struc­ tures containing numerous mitochondria. They could be identified as the processes of neurons lying in subependymalposition. These neurons are therefore cerebrospi­ nal fluid-contacting neurons. Their sim­ ilarity to the well-known cerebrospinal fluid (CSF)-contacting neurons of lower vertebrates (Vigh and Vigh-Teichmann 1971) KEY WORDS: Ependyma, Spinal Cord Central Canal, Mitochondria-containing Processes, may give a clue to their possible Cat, Scanning Electron Microscopy function. Materials and Methods

*Address for correspondence: Adult cats were deeply anesthetized K. Rascher, Department of Anatomy, with nembutal (0. 35g/kg) and artificially University of the Saarland, 6650 Homburg, respirated before being transcardially Federal Republic of Germany perfused with a buffered mixture of 1% Phone No. 06841/166102 glutaraldehyde and 1% paraformaldehyde. The animals, which weighed between 2.3

231 Rascher K, Booz KH, Nacimiento AC et al.

Figure 1. Semithin secti.onof the central canal and its immediate surroundings in the cervical intumescence. L = lumen of the canal; E = ependyma; arrow= mito­ chondria-containing bulb. Figure 2. Semi thin section of the central canal in the upper lumbar region. L = lumen of the canal; E = ependyma; arrow= mitochondria-containing bulb. Figure 3. Overview of the vent ral aspect and 3.0 kilograms, received 1 lit er of of the central canal split longitudinall~ warm perfusion fixative per kilogram Mid-thoracic region. C = heavily ciliated body weight. The spinal cord was removed ependyma; arrow= mitochondria-containing and postfixed in fresh cold fixative at bulbs; B = blood vesse l with well-devel­ least overnight before being dissected oped perivascular space . into tissue blocks for scanning andtrans­ Figure 4. Cl oseup of the cut edge of the mission electron microscopy. The samples ependyma illustratin g the synchronous for SEM were split to expose the ependyma beat of the kinocilia. Caudalto the left. and subsequently postfixed in a buffered solution of 1% osmium t et roxide. The y Semithin and thin sections were cut on a were then rinsed in 2.4 % NaCl and dehy­ Reichert microtome with g la ss knives; drated in an ascending series ofmethanol. the thin sections were viewed in a Zeiss The tissue was transferred to Freon 11 EM 9 at 60 kV. and critical point dried in Freon 13 (Co­ Results hen et al. 1968) in a Polaron apparatus. The specimens were coated with gold in a The central canal of the cat va ries Balzers sputt ering chamber and examined in shape depending upon the level of the in a Cambridge Stereosca n S4 at 20 kV. spinal cord examined. In the upper cer­ The tissue samples for TEM were rin­ vical and in the sacral regions it is a sed in phosphate buffer before being post­ narrow slit with the lateral walls almost fixed in 2% osmium tetroxide. The y were touching; in the thoracic segments it is then rinsed again in buffer. The samples more or less round whereas in both the were dehydrated in isopropanol, trans­ cervical and the lumbar intumescences it ferred to propylene oxide and embedded is again slit-like but with the dorsal in an epoxy resin (ERL 4206, Spurr 1969).

232 The ependy.ma of the cat central canal

7 • pr

Figure 5. Detail of the surface of the less densely ciliated area along the dorsal cervical midline. Note the multi­ tude of microvilli on the tall protru­ sions of the apical poles of the cells. B = mitochondria-containing bulb; S smooth profile. and ventral surfaces closely apposed (compare Figures 1 and 2). The shape of the canal may be slightly altered by large blood vessels with a well-deveJoped perivascular space which penetrate into the subependymal tissue (Fig. 3). The SEM reveals that the ependyma of the can­ al has several ultrastructural features which are also characteristic of the ven­ tricular ependyma. It is fairly densely ciliated throughout (Fig. 3). The cilia are grouped in bundles so close together that the cell surfaces of the ependymo­ Figure 6. Thin section of area as in fig­ cytes can only rarely be seen. The cilia ure 5. P apical protrusion with micro­ are bent in such a manner that their beat villi; C = basal bodies of a bundle of is apparently from rostral to caudal cilia; N = nuclei of ependymocytes; arrow= (Fig. 4). The carpet of cilia becomes contacts between ependymocytes. sparser along the midline both dorsally Figure 7. Apical parts of ependymocytes and ventrally. These areas bearing fewer in ventro-median (cervical) position in­ cilia are quite narrow in the upper cer­ dicating the well-developed contacts and vical segments, broad in both intumes­ interdigitations between the cells. cences and almost invisible in thoracic levels. The less densely ciliated areas microvilli were taller along the dorsal show a multitude of microvilli on the cell aspect than along the ventral. Another apical poles. For the most part the lat­ feature which distinguished the ventral ter protrude markedly into the lumen of from the dorsal ependyma of the midlines the canal, extending their microvilli in were the cell contacts. Those of the dor­ all directions, thus giving the entire sal aspect were generally straight and surface a mossy appearance (Fig. 5). In consisted of only a few short desmosome­ thin sections the protrusions of the api­ like contacts and an occasionalapically cal poles were seen to contain a fine located gap junction. The lateral cell granular matrix and occasional microfila­ borders of the ventrally positioned epen­ ments but no larger organelles (Fig. 6). dymocytes, however, had highly developed The scanned specimens did not reveal junctional complexes with long series of any major ultrastructural differences be­ desmosome-like contacts. These cells in­ tween the dorsal and ventral surfaces of terdigitated with one another giving the the canal, but thin sectionsconsistently membranes a tortuous course (compare showed that the protrusions bearing Figures 6 and 7).

233 Rascher K, Booz KH, Nacimiento AC et al.

- . "': ... . ~ - .• . · - F ~ :'i , ,. J:.--' . ~:·,~""'t - ~- . & f t 1t· ;f .,. '.', ~ ! ; .... :~ ~ T ,. •~- r ·~ , . ..··~_:. / ~\;

. . . . ~ ·, ...

'10 . £1:""ftt~::a! ~~"' ..· •.

Figure 8. Several mitochondria-containing bulbs in ventral aspect of cervical can­ al. Note differences in size. Figure 9. Semithin section ofmitochon­ dria-containing bulb with its neuron (N) Figure 10. Two mitochondria-containing and connecting stalk (arrow). E ep e n­ bulbs. Due to the thinness of the section dyma; L = lumen of the canal. th e ir stalks (arrows) can be traced only for a short distance between the ependy­ The most remarkable feature s of the mocytes (E) . cat central canal were the large, s phe ri­ Figure 11. Closeup of three typical cal structures which could be found mitochondria-containing bulbs. throughout its entire length (Fig. 8). In light microscopic sections they could be distance down between the ependymocytes. easily recognized as round profiles with At the level of the apical cell contacts a granular cytoplasm (Figs. 1 and 2). the stalks are constricted to a narrow Semithin sections sometimes showed the neck which has desmosome-like contacts to bulbs to be connected via long stalks to the neighboring ependymal cells (Fig. 10). subependymally located cells. These cells Neurotubules in more or less parallel had larger nuclei and nucleoli than the array pass along the stalk, through the ependymal cells and fairly dark, homogen­ neck and fan ont into the base of the eously stained cytoplasm (Fig. 9). Thin bulb. The diameter of the bulbs varies sections revealed that these bodies are between 2 and 10µm. Approximately 20 % filled with mitochondria, some small ves­ each belong to the size classes 2, 3, 4 icles and dense particles (Fig. 10). The and Sµm. The final 20 % is distributed abundance of mitochondria gives the struc­ among the larger sizes. In the SEM the tures their name: mitochondria-containing bulbs are covered with a multitude of bulbs. microvilli; only rarely can a more smooth­ In serial thin sections the stalks surfaced one be seen (Figs. 5, 8, 11, 13 of the bulbs can be traced for a short and 14). Their distribution pattern is

234 The ependyma of the cat central canal

Figure 12. Supraependymal nerve fibers (F) containing microtubules and small mito­ chondria. Next to them a cytoplasmic pro­ trusion (P) of an ependymal cell, in this section not connected to it. B = basal bodies of cilia. much more easily recognized in the SEM than in serial sections. They are some­ what more frequent in the above-mentioned less densely ciliated midlines, especially in the ventral, than in lateral areas. There is an average of 45-80 bulbs per millimeter canal; only in the lumbar in­ tumescence does the number rise to 90-160 per millimeter. The mitochondria-contain­ ing bulbs differ in size, contents and surface ultrastructure from the small, smooth profiles which could occasionally be detected between the cilia and micro­ villi of the apical surfaces (Fig. 5) ~ These profiles contained no mitochondria, but were filled with a homogeneous, light­ ly staining fine granular matrix. They had no stalks but were rather protrusions of the ependymocytes themselves (Fig. 12). Supraependymal elements were also Figure 13. Supraependymal cell in cer­ found in the central canal. Supraependymal vical region. cells could be found at all levels of the Figure 14. Supraependymal cell in thor­ canal, but were most frequent in the cer­ acic region. vical region, particularly in the intum­ escence. They occurred on both the dorsal This may be due to the limitations of and the ventral surfaces. These cells TEM sampling. were extremely polymorphic, but all had broad lamellar processes which were at Discussion times closely associated with the cilia The central canal of the spinal cord and mitochondria-containing bulbs (Figs. is the caudal continuation of the ventri­ 1 3 a nd 1 4) . cular system of the brain; its ependymal In TEM sections occasional very thin cells develop in basically the samemanner unmyelinated nerve fibers were observed as those of the ventricles and it is close to the apical surfaces of the epen­ therefore not surprising that the ultra­ dymocytes, between the cilia and micro­ structure of the two are very similar. vil li (Fig. 12). They contained synaptic The linings of the vent ricles of numerous vesicles and a few small dense-cored ves­ mammalian species have been extensively icles, but no contacts between the fibers described (see Leonhardt 1980 and Low1982 and ependymocytes or other supraependymal for literature). One of the very first structures were detected in our material. SEM studies to be done on the ependyma

235 Rascher K, Booz KH, Nacimiento AC et al. dealt with that of the cat ventricles ultrastructure agree entirely with his, (Clementi and Marini 1972). Quite in con­ indicating that the bulbs appear long be­ trast to the vast body of research on ven­ fore maturity in the cat. Their frequency tricular ependyma, that of the central and distribution in the adult cat has canal has received much less attention. been quantified in serial semithin sec­ Most work on canal ependyma was done tions (Boger 1980). Our assessment of in connection with investigations on ex­ their frequency and distribution in SEM perimental hydromyelia and syringomyelia micrographs also agrees with these (Becker et al. 1972). Our findings agree earlier findings. with the light microscope observations of The mitochondria-containing bulbs these authors as well as with the earlier (Mitochondrienkolben) seen in the central TEM results of Takeichi (1966). Takeichi's canal and in the fourth ventricle of the paper, however, does not mention the pro­ rabbit (Leonhardt 1967; Leonhardt and trusions of the cell apical poles bearing Prien 1968; Lindemann and Leonhardt 1973) microvilli. Similar specializations have are similar to, but do differ in some apparently been seen only in the choroid respects from those seen in the cat. The plexus (el Gammal 1983). The supraepen­ stalks of the bulbs contain cytoplasmic dymal structures, free cells and unmyelin­ elements which indicate that they are the ated nerve fibers, are a frequent finding processes of neurons; the contacts be­ in the ventricular system (Leonhardt and tween the stalks and the ependymal cells Lindemann 1973; Coates 1973; Scott et al. are identical in the two species. The 1974). The supraependymal cells seen in position of the neurons around the cen­ the present material are most likely mac­ tral canal of the rabbit seems to be rophages; their appearance in the SEM does close to the ependyma or even intraepen­ not differ from what has already been de­ dymal. In the cat, however, neurons in the scribed of the epiplexus cells of the immediate vicinity of the ependyma are the choroid plexus and of other free cells exception and a direct connection betwe~n seen on the ventricular walls, particu­ bulbs and perikaryon could not be unequi­ larly in the infundibular recess (Mestres vocally established in TEM sections. In 1976; Mestres andBreipohl 1976; Bleier 1977). the cat the neurons are so few and far The supraependymal nerve fibers also between that they may be part of a hyp­ resemble others described on the ventri­ endymal plexus, a given neuron contribu­ cular lining (Lorez and Richards 1982; ting several bulbs, each on a long stalk. Ribas 1977). Judged by their content of Another distinct difference between the microtubules and small mitochondria the bulbs of the two species lies in their fibers are axons. They seem to be much content of mitochondria. Leonhardt (1976) thinner and less common in the cat central divided the bulbs he observed in the rab­ canal than in that of the rabbit (Leon­ bit into two main groups, the first with hardt 1976). In our material we have up many mitochondria but few microvilli, the to now not found any myelinated fibers second with fewer mitochondria but count­ nor contacts between the fibers and other less microvilli. This distinction is structures. Nor have we been able to de­ not possible among the bulbs of the cat termine the location of the neurons from canal. Here they all appear to belong to which they originate. a different group altogether: They have The ependymal protrusions seen in the many mitochondria and microvilli. Bulbs SEM as well as in thin section can readily with fewer microvilli were the exception. be classified into two markedly different The possible function of the bulbs groups: 1) the usually small, smooth pro­ has been of particular interest because trusions which can only rarely be seen they are the CSP-contacting terminals of between the microvilli and cilia and are neuronal processes. CSP-contacting neu­ a part of the ependymal cells themselves. rons have been described in the central Neither the contents nor the surface mem­ canal in all vertebrate classes (Vigh brane of this type of protrusion is spe­ and Vigh-Teichmann 1971). These authors cialized in any way. Similar protrusions found that the number of such neurons de­ have been observed in the ventricularwall creases with ascent on the phylogenetic of fetal animals (Booz 1975; Hannah and ladder. On the basis of morphological Geber 1977). What role they may play on features they proposed that the CSF­ the ependyma of the central canal in adult contacting neuronal processes might be animals could not be determined with the dendrites belonging to receptor-like present methods. 2) the considerably lar­ neurons. This seems to hold true forthe ger mitochondria-containing bulbs. They lower vertebrates. On the other hand, have been described in the central canal the corresponding cells in the rabbit of the rabbit (Schwanitz 1969; Leonhardt have been interpreted to be secretory 1976), but were apparently seen for the neurons (Leonhardt 1967) or have been first time in the kitten (Takeichi 1966). seen to possess characteristics which Takeichi (1966) does not provide any in­ belong either to receptor cells or to formation about the number or location of neurons capable of electrolyte exchange the bulbs, but our findings as to their with the CSF (Leonhardt 1976). In the

236 The ependyma of the cat central ca nal cat the bulbs pose an even more intrigu­ neurosecretion. A hitherto unknown struc­ ing problem than in the rabbit because ture in the 4th ventricle of the rabbitJ the neurons to which they belong are so z Zellforsch 79:172-184. difficult to locate. The bulbs do not Leonhardt H (1976) Die Liquorkontaktfort­ contain any cytoplasmic elements which satze im Zentralkanal des Rlickenmarkes. would clearly indicate that they are ei­ Eine raster- und transmissionselektron­ ther secretory or receptor-like. Nor does enmikroskopische Untersuchung am Kanir.­ the surface ultrastructure provide con­ chen. (The CSF-contacting profiles in clusive evidence for one or the other the central canal of the spinal cord. A function. For the present the question SEM-TEM study in the rabbit.) as to their function must remain un­ Z mikrosk-anat Forschung 90:1-15. answered. Leonhardt H (1980) Ependymund Circumven­ Acknowledgements trikulare Organe. (Ependyma and circum­ ventricular organs.) In: Oksche A and The authors gratefully acknowledge Vollrath L(eds) Handbuch der mikroskop­ the skillful technical assistance of ischen Anatomie des Menschen. NeurogliaI Mr. N. Putz, Ms. s. Brundaler and Ms. G. Nervensystem Part 10. Springer, Berlin Klihnreich. We thank Ms. G. Haag for her Heidelberg New York. pp 177-666. help with the typing. This project was Leonhardt H, Lindemann B (1973) Uber ein financed in part by the Sonderforschungs­ supraependymales Nervenzell-, Axon- und bereich 38, Homburg/Saar. Gliazellsystem. Eine raster- und trans­ References missionselektronenmikroskopische Unter­ suchung am IV. Ventrikel (Apertura lat­ Becker D, Wilson J, Watson W (1972) The eralis des Kaninchengehirns. (Observa­ spinal cord central canal: response to tions on a system of supraependyma l experimental hydrocephalus and canal nerve cells, axons and glial cells. A occlusion. J. Neurosurg 36:416-424. SEM-TEM study of the fourth ventricle Bleier R (1977) Ultrastructure of supra­ (lateral apertures) in the rabbit brainJ ependymal cells and ependyma of hypothal­ Z Zellforsch 139:285-302. amic third ventricle of mouse. J Comp Leonhardt H, Prien H (1968) Eine weitere Neural 174:359-376. Art intraventrikularer kolbenformiger Boger G Ti980) Mitochondrienkolben imZen­ Axonendigungen aus dem IV. Ventrikel des tralkanal von Katzen. (Mitochondria­ Kaninchengehirns. (An additional type of containing bulbs in the central canal of intraventricular bulb-shaped axon endings the cat spinal cord.) Dissertation, in the 4th ventricle of the rabbitbrain) Library of Univ . of the Saar l and , FRG. Z Zellforsch 92:394-399. Booz KH (1975) Secretory phenomena at the Lindemann B, Leonhardt H (1973) Supraei:en­ ependyma of the third ventricle of the dymale Neuriten, Gliazellen und Mitochon­ embryonic rat. Anat Embry o l 147:143-159. drienkolben im caudalen Abschnitt des Clementi F, Marini D (1972) The surface Bodens der Rautengrube. (Supraependymal fine structure of cerebral ventricles neurites, glial cells and mitochondria­ and of choroid plexus in cat. Z Zell­ containing bulbs in the caudal aspect of forsch 123:82-95. the floor of the 4th ventricle.) Coates P-(-1973) Supraependymal cells in Z Zellforsch 140:401-412. the recesses of the monkey third ven­ Lorez HP, Richards JG (1982) Supraepen­ tricle. Amer J Anat 136:533-539. dymal serotoninergic nerves in mammal­ Cohen AL, Marlow DP, Garner GE (1968) A ian brain; morphological, pharmacologi­ rapid critical point method using fluo­ cal and functional studies. Brain Res rocarbons (Freons) as intermediate and Bull 9:727-741. transitional fluids. J de Microscopie Low F (1982) The central nervous system in 7:331-342. scanning electron microscopy. Scanning eT Gammal S (1983) Regional surface chan­ Electron Microscopy 1982;II:869-890. ges during the development of the tel­ Mestres P (1976) The supraependymal cells encephalic choroid plexus in the chick. of the rat hypothalamus: Changes in their Cell Tissue Res 231 :251-263. morphology and cel l number during the Hall P, Mull.er J, Campbell P (1975) Expe­ ovarian cycle. Experientia (Basel) rimental hydrosyringomyelia, ischemic 32: 1329-1331. myelopathy and syringomyelia. J Neuro­ Mestres P, Breipohl W (1976) Morphology surg 43:464-470. and distribution of supraependymal cells HannahR, Geber W (1977) Specializations in the third ventricle of the albino rat. of the ependyma in the third ventricle Cell Tissue Res 168 :303-3 14. of the developing hamster. Am J Anatomy Nakayama Y, Kohnol<(1974) Number and po­ 149:597-603. larity of the ependymalcilia in the cen­ Leonhardt H (1967) Zur Frage einer intra­ tral canal of some vertebrates. J Neuro­ ventricularen Neurosekretion. Einebisher cytol 3:449-458. unbekannte nervose Struktur im IV Ven­ Ribas J-(1977) The rat epithalamus. I. trikel des Kaninchens. (Concerning the Correlative scanning and transmission problem of intraventricular electron microscopy of supraependymal

237 Rascher K, Booz KH, Nacimiento AC et al.

nerves. Cell Tissue Res 182:1-16. G. P. Kozlowski : Why is the neuron in Schwanitz W (1969) Die topographische Ver­ figure 9 so dark? Is it cons istent with teilung supraependymaler Strukturen in other examp l es of similar neurons with den Ventrikeln und im Zentralkanal des mitochondria-containing bulbs? Kaninchengehirns. (The topographical dis­ Authors: Work on the CSF-contacting tribution of supraependymal structures in neurons is still in progress , but they the ventricles and in the central canal do seem to have in common a darker­ of the rabbit.) Z Zellforsch 100:536-551. stain in g cytoplasm. This may be due to Scott DE, Kozlowski GP, Sher idan MN (1974) their comparatively numerous mitochon­ Scanning electron microscopy in the ultra­ dria and/or other osmophilic components structural analysis of the mammalianven­ of the cytoplasm. Simi larl y darker­ tricular system. Int Rev Cytol 37: 349-388. staining cells with mitocho ndria-contain­ Spurr A (1969) A low-viscosityepoxy ing bulbs have also been seen in the resin embedding medium for electron rabbit (Le onhardt 1976 ). microscopy. J Ult Research 26:31-43. G. P . Kozlowski: Could the perivascular Takeichi M (1966) The fine structure of space in figure 3 be an artifact? ependymal cells. Part I. The fine struc­ Authors: At first we also considered ture of ependymal cells in the kitten. this possibility, but in SEM and light Arch histol jap 26:483-505. micrographs the perivasc ular space was Vigh B, Vigh-Teichmann I (1971) Structure a consistent finding surrounding the lar­ of the medullospinal liquor-contacting ger vessels in the gray matter. TEM an­ neuronal system. Acta biol Acad Sci alysis showed that there were pericytes Hung 22:227-243. and collagen fibers within the perivascu­ lar space which is isolated from the tis­ Discussion with Reviewers sue around it as well as from the vesse l wal l by a basement membrane. Some of F. N. Low: Since numerous authors have these elements may be seen in figure 3; related ciliation and/or the lack of it they are even more easi ly recognized in in epe ndyma to underl y in g structures figure 15. These perivascular spaces do (nuclei, etc.), could you comment on this not appear to be unique to the cat spin­ for the central canal in the spinal cord? a l cord. We have seen id ent ical ones in Authors: The distribution pattern of cilia both the rat and the rabbit cord. in the cat central canal might be seen in connection with the nature of the under­ lying tissue. The less densely ciliated midlines are closer to white matter and the heavily ciliated lateral areas cover gray matter. This is comparable to what Page observed in the lateral ventricles of the rabbit (Page RB, 1975, Scanning electron microsc opy of the ventr icular system in normal and hydrocephalic rab­ bits. J Neurosur 42:646-664) In the lateral ventriclesof other species, how­ ever, Scott et al. (1974) observed the reverse situation. The latter authors point out that "The degree of cilial sparsity is not consistent from animal to animal, or species to species but appears random." Indeed, we have compared the surface of the canal ependyma of the cat, the rabbit and the rat (Rasch er K, Booz KH, Nacimiento AC, 1984, Comparative ultrastructure of the ependyma in the cervical central cana l of cat, rabbit and rat. An SEM study. Beitr Elektronenmikro­ skop Direktabb Oberfl 17:239-246) and seen that there are major differences between the species, for examp l e, the rat canal ependyma does not bear any bundles of cilia at all below the level of cer­ Figure 15. Blood vessels (B) with well­ vical I. In this species there are only developed perivascular space next to the one or two cilia per ependymal cell, central canal (C). Note abundance of which confirms the findings of Nakayama collagen fibers (F) and several peri­ and Kohno (1974). cytes (arrows).

238