sign in sign up
<<
Home , Dorsal root ganglion, Ganglion, Nervous tissue, Perineurium

J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

J. Neurol. Neturosurg. Psychliat., 1950, 13, 203.

THE PENETRATION OF PARTICULATE MATTER FROM THE CEREBROSPINAL FLUID INTO THE SPINAL GANGLIA, PERIPHERAL , AND PERIVASCULAR SPACES OF THE CENTRAL BY J. B. BRIERLEY From the Departmenit of Neuropathology, Institute ofPsychiatry, Maudsley Hospital, London The precise termination of the spinal granules was observed through the membranes in around the emerging roots has been the this region, but there was an excretion of the subject of many investigations since the early work prussian blue solution into arachnoid nests of Quincke (1872) who injected 1 c.cm. of a sus- lying in close relation to and often embedded in pension of cinnabar into the lumbar subarachnoid the dura.

space of dogs. At examination two to four days A similar failure of india ink particles to pass Protected by copyright. later it was observed that cinnabar particles had into the root or along the nerve was passed on occasion along the but reported by Iwanow and Romodanowsky (1928) in more frequently along the intercostal. their experiments on dogs, although these workers Key and Retzius (1875) injected various indicators were the first to describe the outflow of cerebro- (including gelatin coloured with Berlin blue) into spinal fluid into lymphatic channels around the the subarachnoid space under a pressure of 60 mm. nerve roots. of mercury, and noted their progress for some The confinement of thorotrast to the anatomical distance along the peripheral nerves. However, in cul-de-sac of the subarachnoid space was noted by this connexion, Weed (1914) quotes Testut as having Wustmann (1933) who visualized the indicator by pointed out that the pressure employed was res- means of x rays. There was, however, a distinct ponsible for the production of damage, escape of the material into the epidural fat layers. especially in the arachnoid at the site of the emer- The behaviour of a fine suspension of india ink gence of the nerve roots. (particle size 05 p) when introduced into the Similar findings were recorded by Goldmann (1913) subarachnoid space under physiological pressure with trypan blue, although the well-known toxic conditions was investigated by Brierley and Field

effects of this indicator may detract to some extent (1948) and the outflow of ink from the cul-de-sac http://jnnp.bmj.com/ from the results obtained. into lymphatic channels was described. In this The classical experiments of Weed (1914) in which work, only the naked eye appearances in the region an isotonic mixture of potassium ferrocyanide and of the root ganglia were recorded, the ink granules ferric ammonium citrate was introduced into the not appearing to pass beyond the proximal pole of cerebrospinal fluid at a pressure only slightly in the ganglion and a few millimetres farther along the excess ofphysiological, did not permit demonstration ventral root. of prussian blue granules for more than a short Somberg (1947) in a review of the literature distance along the anterior and posterior nerve criticizes many previous workers on the grounds on September 27, 2021 by guest. roots. that the indicators employed were either toxic Elman (1923) defined the anatomical arrangement (e.g. trypan blue) or irritant (e.g. cinnabar) and that of the membranes around the roots they were often introduced under abnormally high using as indicators india ink and the Weed prussian pressures with possible production of tissue damage blue mixture. The subarachnoid space was des- as a result. cribed as terminating in an anatomical cul-de-sac Hassin (1947) maintained that the subarachnoid at the point where the arachnoid turned inwards to space is to be followed in appropriate sections over fuse with the pia. No outward passage of ink the surface of the ganglion, the arachnoid ". .. being 203 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

204 J; B. BRIERLEY in close proximity to but not blended or fused with the proved ability of such particles to pass through it-or the dura mater. That is to say, the subarach- the arachnoid and dura in certain situations .noid space plunges into the spinal ganglion where (Brierley and Field, 1948; Field and Brierley, 1948) it becomes continuous with its tissue spaces and was a reminder that similar regions of membranous finally emerges over the peripheral (post-ganglionic) permeability might be encountered elsewhere around nerve as its perineural space ". Hassin is prepared that part of the subarachnoid space under to accept the experimental evidence of Key and investigation. Retzius (1875) and of Funaoka (1930) and his The technique adopted was also cxpected to co-workers in which intraneural injection of various provide histological studies of the perivascular dyes was used to demonstrate a communication spaces of the , from which, between the endoneural and perineural spaces of in view of the survival period of 48 to 72 hours, it the peripheral nerves and the subarachnoid space. would be evident whether or not there was any Hassin states that" ... the remarkable experiments considerable penetration of these spaces under of Funaoka cannot be attributed to faulty normal cerebrospinal fluid pressure. technique ", and again " . . . the foregoing instances of marvellous experimenting and vast penetration Material and Methods of the injected dyes can hardly be explained by .-Eight adult rabbits of weights from 2-0 forced pressure . .. ", this latter in spite of an to 3 0 kg. were used. intraneural injection of 25 c.cm. of " dekalen ". Antesthetic.-The animals were anesthetized with The histological pictures presented by Hassin are sodium nembutal (2 5% solution intravenously), not convincing evidence of the continuity of the recovery taking place two to two and a half hours after subarachnoid space over the ganglion as is claimed. the operation. Such a claim could only be established (within the Indicator.-A suspension of india ink in saline, known limitations of histological technique where prepared as described previously (Brierley and Field, tissue spaces are concerned) if truly longitudinal 1948). Ninety per cent. of the particles have a diameterProtected by copyright. sections of the ganglion and post-ganglionic nerve of 0 5,u, and the range of size is from 0 4 to 1-5V.. The were presented. The two sections illustrated in suspension was dialyzed for one hour in a cellophane Hassin's paper pass obliquely through the ganglion sac suspended in running tap water and was sterilized and provide no information as to the disposition immediately before use. of the membranes in the important regions of the middle zone and distal pole. Technique That the passage of injected material from the The indicator was introduced into the cisterna magna is which was approached by a mid-line incision extending peripheral nerve into the subarachnoid space by 2 cm. above and below the arch of the atlas, the no means certain or predictable was shown by lving prone on the table with the head flexed to 900. Brierley and Field (1949) using as inoculum a The posterior spinal muscles were separated with care- volume of 0 05 ml. of radioactive phosphorus (as ful attention to hmmostasis until the margin of the phosphoric acid). It was recorded that an injection occiput was exposed. The point of attachment of the carried out in the upper third of the rabbit's atlanto-occipital membrane to the bone in the mid-line sciatic nerve was less likely to enter the cerebrospinal was defined and the periosteum stripped over the an made at a lower but adjacent four or five millimetres of the occipital surface. fluid than injection level, http://jnnp.bmj.com/ tended to pass directly into the substance of the The bone was drilled with a fine dental drill (fissure The use of a volume No. 0) at a point in the mid-line 2 mm. above the free . larger injection posterior margin of the foramen magnum, the drill may result in more constant entry into the subarach- point being directed obliquely backwards towards the noid space as a result of some mechanical tissue cistern. Perforation of the bone and dura was indicated damage and it may well be doubted if an intraneural by the free escape of cerebrospinal fluid into the groove injection much in excess of 0 05 ml. is a valid indi- between the posterior spinal muscles. The free fluid cator of normal anatomical relationships between was drawn up into a syringe to give a rough check on the volume released. A children's type lumbar puncture endoneural and perineural spaces and the sub- on September 27, 2021 by guest. arachnoid space. needle was swung into place over the hole and slowly The purpose of the present investigation was, introduced until cerebrospinal fluid could be obtained on to the normal of the slight suction. A 2-0 ml. syringe filled with india ink first, define dispositions spinal suspension was attached to the needle. The indicator membranes as they pass towards the root ganglion ran in rapidly at first, then slowed down and came and the ventral nerve root, and secondly to demon- virtually to rest after the admission of a volume rather strate the behaviour of a fine sterile particulate less than that of the cerebrospinal fluid withdrawn. In indicator introduced into the subarachnoid space no instance was the pressure of introduction allowed to under physiological pressure conditions. However, exceed 120 mm. of the indicator. Immediately on J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

THE PENETRATION OF PARTICULATE MATTER FROM C.S.F. 205 withdrawal of the needle the hole was closed with bone outward spread of ink diminished, until, in the wax and the muscles and skin edges apposed with thoracic region the apex of the " ink-cuff " coincided separate layers of silk sutures. with the proximal pole of the root ganglion. In The introduction of the indicator was repeated at the cervical region there was conspicuous migration 24-hour intervals, an average of three operations being of ink, largely in the epidural and performed on each animal. In this way it was possible reaching a maximum along the second and third one animal to introduce up to 4 5 ml. of indicator into it was without appreciably raising the pressure of the cerebro- cervical nerves where probably responsible spinal fluid. for the greater part of the filling of the deep cervical The animals were killed 12 to 24 hours after the last lymph nodes. Removal of the ink-stained epidural operation by exsanguination through the thoracic aorta tissue revealed that the ink had spread up to and on under nembutal anTsthesia. A cannula was then tied to the only in the second and into the ascending aorta and perfuision carried out with third cervical nerves, whereas in the remaining about 1000 ml. of physiological saline followed by an cervical nerves the apex of the " ink-cuff" lay at equal quantity of 10°o formol saline. the proximal pole of the root ganglion. A complete exposure of the spinal cord was obtained by carrying out a laminectomy from sacrum to coccyx. Nervous distribution of the The sciatic nerves were next exposed from the thigh Central System.-The towards the vertebral canal, fine bone forceps being ink around the circumference of the spinal cord used in the last part of their course to reveal the spinal at any one level was seldom uniform. In the root ganglia, cervical region the greater part of the ink lay on the dorsal aspect of the cord and there was little more Macroscopical Appearances than a faint speckling on the ventral surface. This Spinal Cord and Nerve Roots.-The spinal cord unequal distribution was probably due to the and meninges presented the typical appearances dorsal position of the site of introduction. In the " described in detail elsewhere, the " ink-cuffs thoracic, lumbar, and sacral regions the discrepancy Protected by copyright. being particularly well developed in all animals of between dorsal and ventral aspects is reduced while the series. Examination under the dissecting micro- the coccygeal cord, ganglia, and roots were in- scope (Fig. 1) showed that the ink had penetrated variably surrounded by a dense accumulation of some distance over the surface of the root ganglion ink which penetrated the dura to stain the epidural and almost to the distal pole in some cases. This connective tissue. was in marked contrast to the findings previously In two animals of the series the bulk of the ink in reported and may be accounted for by the intro- the spinal subarachnoid space was concentrated in a duction of larger volumes of ink at a site (cisterna well-defined strip lying along the dorsal aspect of magna) more suitable for spread of the indicator the cord and extending from the lower end of the to the spinal cord. The distal termination of the cisterna magna to a tapered point at the mid- " ink-cuff" was noticeably ragged and appeared thoracic level. In these two animals there was a to consist of more or less parallel black lines of poor development of the " ink-cuffs " in the lumbar varying thickness. There was a dense deposit of region. ink in the " axilla " between the nerve roots and the Transverse sections of the spinal cord were spinal cord. examined under the dissecting microscope and

A proportion of the india ink on the surface of showed the presence of a ring of grey haze lying http://jnnp.bmj.com/ the ganglion lay in the epidural connective tissue, immediately under the pia and most marked in the but when this latter had been dissected off the cervical region on its dorsal aspect. It was impos- linear disposition of ink particles on the surface of sible to avoid the suggestion that this appearanice the ganglion still remained. might be due to the passage of ink from the sub- The ventral nerve roots exhibited only a light arachnoid space through the pia. speckling of ink particles which died out at a level The distribution of ink particles around the corresponding to the midpoint of the ganglion, itself was reasonably uniform from case to case and although in some instances there was a thin tail of quite similar to that described by other workers on September 27, 2021 by guest. ink extending distally along the inner edge of the (Weed, 1914; Hurst, 1932). The greatest accumu- nerve from the tip of the dense " axilla " referred lation of ink was noted around the medulla, the to above. infundibulum and hypophysis, and over most of the The extent of the spread of ink over the surface of surface of the . Dense cuffs of ink had the root ganglia was by no means uniform through- formed around the emerging , particu- out the spinal nerves but was constantly maximal larly the olfactory, optic, and trigeminal. The in the three large nerve roots contributing to the intensity of the indicator on the surface- of the formation of the sciatic nerve. At higher levels the cerebral hemispheres decreased towards the vertex J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

*.4tjg I' \Q: t;SE Xdwt#k k ri k . tuts- K; .'. .* tr

0 ,. I is xi, i .. '; ' i :e ,j .. *$ A .: +: .i ,4I £:I

8 o. ^ k:f

) :... ; I, e .8

*.} *i FIG. 1.-Sacral spinal cord of rabbit i. ( x 2) 24 hours after second intro- duction of india ink. Note spread II over surface of root ganglia and dense " axilla " along inner surface Protected by copyright. of nerves. FIG. 2.-Longitudinal section of tip of spinal cord (x 60) showing dilated termination of central canal containing ink.

t.,

.1 http://jnnp.bmj.com/

...4 on September 27, 2021 by guest.

: t

FIG. 3.-Longitudinal section through proximal FIG. 4.-Longitudinal section through proximal pole pole of normal root ganglion (x 85). A. Sub- of root ganglion after introduction of ink. (x 80). arachnoid cul-de-sac. B. Reduplication of Cul-de-sac on right contains dense accumulation the arachnoid. C. Dura. D. Epidural con- of ink and polymorphonuclear cells. (K as for nective tissue. Fig. 3.) J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

THE PENETRATION OF PARTICULATE AMATTER FROM C.S.F. 207 and was in general more intense in relation to the an emerging pair of nerves) and transverse sections. cortical vessels. The animals had been perfused with physiological Examination of thick coronal slices under the saline under anesthesia followed at once by 10% dissecting microscope showed in certain regions the formol saline. The material was embedded same grey haze beneath the pia as was seen in the in celloidin and cut at 12 to 15 V. spinal cord, but this appearance was confined to In the material thus prepared, the region of certain definite localities. It was constant and greatest importance lies at the proximal pole of the most marked in the hypothalamus but was also root ganglion upon which the three spinal mem- seen in order of decreasing intensity in the ventral branes converge and where any particulate indicator, parts of the , in the cerebral cortex intrathecally introduced, can be relied upon to in relation to the choroidal fissure, and throughout accumulate. Here the question arises, does the the ventral and lateral aspects of the brain stem. subarachnoid space end blindly as a cul-de-sac or There was a conspicuous absence of ink in the is it continued as a narrow or even potential space lateral ventricles except for a faint dusting of ink over the surface of the ganglion ? Careful scrutiny on the choroid plexuses. There was a slight collec- of the longitudinal sections has left no doubt, tion of particles in the lowest part of the third that while the spinal dura passes over the ganglion ventricle and an increasing amount on passing back to form its sheath and ultimately the nerve peri- through the aqueduct to the fourth ventricle. neurium, the arachnoid has no such distal extension However, the quantity of ink even in the latter (Fig. X). Again, the subdural space is easily recog- situation was only a fraction of that to be seen in a nizable for a considerable distance over the ganglion similar volume of the basal subarachnoid space. and appears to be continuous with the sub-peri- This paucity of ink in the ventricular system was in neural space of the peripheral nerve, whereas the contrast to the situation in the central canal of the subarachnoid space becomes narrowed at the spinal cord which appeared as a distinct black dot at proximal pole of the ganglion and terminates in a Protected by copyright. all levels of the cord and which terminated in a clearly defined anatomical cul-de-sac. This is the dilated and ink-filled sac in the lower sacral region result of the dipping in of the arachnoid membrane (Fig. 2). towards the margin of the annulus of large cells that forms the proximal pole of the ganglion (Fig. 3). Histological Appearances At this point there is a reduplication of the Normal Material.-The nerve roots, ganglia, and membrane itself and the various layers pass in to mixed spinal nerves of several normal animals were blend with the connective tissue of the ganglion. examined by means of longitudinal (in the plane of It is to be noted that the apex of the cul-de-sac lies CQ orb a. Epidural fatty connective tissue. .6 C b. Dura continuous with ganglion sheath. _ d c. Sub-dural space. e d. Capsule of ganglion. 9 http://jnnp.bmj.com/ e. Ganglion cells. f f. Group of ganglion cells lying among dorsal root fibres. k g. Arachnoid. h. Subarachnoid cul-de-sac con- ?- : ~ taining ink. I j. Ganglion cells at proximal pole.

t<>_cw <. k. Dorsal nerve root. on September 27, 2021 by guest. 5C 1. Ventral cul-de-sac. m. Proximal pole of ganglion- ventral part. n -~ _-____ n. Ventral nerve root. o. Dura continuous with peri- o~~~~~~~~~~~~~~~~~~~~~~~~S neurium. p J FIG. X.-Diagrammatic longitudinal section through proximal pole of dorsal root ganglion. E J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

208 J. B. BRIERLEY in close relation to the ganglion cells from which it Lined to enter the terminals of fine lymphatic is separated by only a few layers of flattened ,els and so to be conveyed to the general lymph connective tissue cells (Fig. 4). am. Between the dorsal and ventral nerve roots there i the region of the cul-de-sac the disposition of is a similar cul-de-sac which again terminates in ink fails to reveal any anatomical continuity of close relation to ganglion cells. subarachnoid space over or into the ganglion, Around the ventral nerve root the subarachnoid re being no narrow wedge or line of particles space is gradually reduced, the arachnoid apparently nding in either of these directions. Nevertheless fusing with the and with the superficial nerous ink particles are seen to lie among the connective tissue layers of the nerve at about the glion cells whose periphery they outline to give mid point of the ganglion. iaracteristic " wire-netting " appearance (Fig. 5) briefly reported elsewhere (Field and Brierley, India Ink.-Within the general subarachnoid 8). space the ink particles tend to accumulate on the L careful study of both transverse and longitudinal deep aspect of the arachnoid rather than on the pia. ;ions shows that there are two modes of access of This distribution is also found in the basal portions particles to the typical peri-cellular position in of the " ink-cuffs " where there is also a penetration ganglion (Fig. Y.): (a) At the apical zone of of the arachnoid by the ink which comes to lie in subarachnoid cul-de-sac, ink particles are able the sub-dural space, along which it penetrates as far )ass through the layers of connective tissue and as the distal pole of the ganglion (Fig. 6). Extensive ito drift slowly around the ganglion cells towards permeation of the dura itself is also evident, so that distal pole of the ganglion. (b) The ink that has the loose, fatty epidural connective tissue contains a etrated the arachnoid to enter the sub-dural considerable quantity of ink, part of which is ce moves slowly in a distal direction over the Protected by copyright.

'.4,4.--.- ~ ~ t ; £ -+eF /aX a. Epidural fatty connective tissue. b. Dura. c. Sub-dural space. d. Lymphatic vessel draining /A w / epidura tissues. e. Arachnoid.

q ~ f. India ink in subarachnoid space. g. Subarachnoid cul-de-sac. 'Ih h. . j. Dorsal nerve root. http://jnnp.bmj.com/ 1. Outflow into proximal pole of ganglion. I *--*. J 2. Passage back from sub-dural space into peripheral part of root ganglion. 3. Outflow through arachnoid into sub-dural space and thence

through dura into lymphatics on September 27, 2021 by guest. .IF0 ,,.. in epidural connective tissue. /_

.,

.001 .(i F1ia. Y.-Diagram to show escape routes for cerebrospinal fluid from the subarachnoid cul-de-sac. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

¾.-A * 1 b

. ebr'. *

S - t %WW# r-*.AZ .* r L t S s IWt *M. *&w a. +, * -. 4. w

s- C* tJ% Srr 44 ;r ^ t %, SS

FIG. 5.-Longitudinal section through middle of FIG. 6.-Longitudinal section through distal pole root ganglion after introduction ofink (x 275). of ganglion (x 65). Note penetration of ink Note peri-cellular distribution of ink particles. particles from sub-dural space, among the Protected by copyright. ganglion cells.

t ,t *; .I I : . I *~~ ¾ t 'a. N At C' 4 a .i.1 '1 r. 4I.- A

* ¼" '. I 9 * ,^ \ ^ w v~~~~' ' 'I http://jnnp.bmj.com/ N ' I, 4I. a a' ' .revs. t& .u{ $ **?-".1'V jta"p l. ~~~~~

*. II on September 27, 2021 by guest. FIG. 7.-Longitudinal section of mixed spinal FIG. 8.-Longitudinal nerve beyond distal pole of ganglion. (x 75). section of a vessel The ink particles lie in parallel columns in the hypothal- between the nerve fibres. amus. (x 275). Ink particles lie in the perivascular space of the vessel and its tributary.

F J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

210 J. B. BRIERLEY ganglion, and a proportion of the particles pass often to a considerable depth. It was further inwards among the ganglion cells and are again observed that the perivascular spaces were only carried towards the distal pole (Fig. 6). well-filled in situations in which the subarachnoid Thus it comes about that in an animal that has space itself contained obvious accumulations of ink. received adequate amounts of ink and has survived In general, then, the filling of the spaces reflected the at least 48 hours there is a distinct capsule of ink distribution of the indicator over the hemispheres, particles around almost every cell of the root brain stem, and spinal cord. Thus, in the hypo- ganglion, the overall distribution of the ink beiig (Fig. 8) ink-enclosed vessels could be fairly uniform. traced from the basal subarachnoid space to within When the mixed spinal nerve just distal to the a short distance of the walls of the third ventricle, ganglion is examined it is seen that the progress of and in the region of the choroidal fissure of the the ink through the ganglion continues into the lateral ventricles numerous perivascular spaces con- nerve, where the particles are found lying in parallel tained ink throughout the whole thickness of the columns between the nerve fibres (Fig. 7). The cortex (Fig. 9). maximum distal spread observed was 4 to 5 mm. Isolated vessels in the thalamus (Fig. 10) showed from the distal pole of the ganglion. It is empha- limited filling while all degrees of penetration could sized that there is no peripheral concentration of ink be seen in cortical vessels, mainly on the ventral in the nerve as would be expected ifthe subarachnoid and ventro-lateral aspects of the brain (Figs. 11, 12, space were in direct continuity with the sub-peri- 13, and 16). neural space. The latter space is clearly recognizable An exceptional quantity of ink around almost beyond the distal pole where it is the continuation of every vessel was constantly observed throughout the the sub-dural space, yet at this level it is almost hippocampal (Fig. 15) lying between devoid of ink, and can therefore contribute little if the layer and the fascia dentata. anything to the total ink visible in the mixed nerve. Similar findings were observed in the globusProtected by copyright. It must be concluded then that the ink particles pallidus the majority of whose vessels were sur- lying among the nerve fibres have come from rounded by small quantities of ink (Fig. 19). peri-cellular situations in the ganglion substance, Again, in the cerebellum there was considerable as a result ofpassive transport in a slow centrifugally filling of perivascular spaces which could be traced directed stream. through the layer into the white core In the ventral nerve root, the subarachnoid cul- of the folium (Fig. 18). de-sac, as already pointed out, extends to about the In the brain stem the depth of penetration of the mid-point of the root ganglion and its walls appear indicator into the long straight perivascular spaces to possess the same permeability to ink particles as was more restricted than in the cortical and dien- does its counterpart in the dorsal root. The amount cephalic structures and may well be due to the of ink escaping by this route is, however, small, the proportionately smaller quantity of ink in the major centrifugal stream lying in the sub-dural adjacent subarachnoid space. Good penetration space which is clearly visible in all the sections was observed at several sites in the medulla (Fig. 14) (Fig. 4 and Fig. 6). Proximal to and alongside the and repeatedly in the cervical spinal cord (Fig. 17). ganglion there is little ink to be seen in the interior The existence of a demonstrable perivascular ofthe ventral root, but this amount increases distally, space was not revealed in any region of the central http://jnnp.bmj.com/ until at the point ofjunction with the dorsal root the nervous system except the hippocampus where some concentration of ink particles in the two neural unusual shrinkage may well have been responsible components is almost equal (Fig. 7). This uniform for its production. The india ink particles appeared disposition of ink is then maintained throughout to lie on the outer surface of the vessel in direct some 4 to 5 mm. of the mixed spinal nerve beyond contact with the brain substance. The penetration which it slowly decreases to disappear after a of the ink, although considerable, did not reach the further few millimetres. perineuronal spaces, which were again not visible in the celloidin material. on September 27, 2021 by guest. Perivascular Spaces of the Central Nervous The uniform absence of any perivascular cuffing System.-Serial coronal celloidin sections of the by inflammatory cells suggested that movement brain and cervical cord were cut and every tenth towards the deeper parts of the spaces was not stained with carmalum, intermediate sections being dependent upon a process of phagocytosis. Further, stained where necessary. examination under the oil immersion lens showed It was at once apparent that a certain proportion that most ink particles lay quite free on the vessel of the perivascular spaces in any one section con- adventitia, and the only cells containing ink were tained ink particles, not only in their mouths, but sparsely scattered in the J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

be Cw *.- ,^ta _-

'.B '~

eq

@ s or i>* 4 *f Ie ; .?

P - 5 ,' *> w .* X w , 9. "~ ..s ,. - AE /; S

S~ ~ . w ,. It ,I ' 's 7' s R z;'W$gbll]-\^sub-'bbs^~~4's CrA,s Protected by copyright.

A~~~; t.,'ebp .it

.sis* 1~e* *~ *vw ffi

i ~*s;ff ~ ~ 'It~';~t ? i9 9.. http://jnnp.bmj.com/ /+ + 13 .j b on September 27, 2021 by guest.

FIG. 9.-Perivascular space containing ink in cortex beside choroidal fissure. (x 125.) FIG. 10.-Vessel of thalamus surrounded by ink particles. (x 125.) FIo. 11.-Small vessel in cortex showing perivascular ink. ( x 425.) FIG. 12.-Perivascular ink deposition around sub-cortical vessels in ventral aspect of hemisphere. (x 125.) FIG. 13.-Perivascular ink deposition in cortex beside choroidal fissure. (x 375.) J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

~ 4 -.1 Atr or

a~~~~~~~~.. s V pJt;' a 14 '_ft. .4

..3. 4.. "Ill-lo V- A * I:

4tttK;

t~~ ~ ~~~1P.sA

(k's ,. . Protected by copyright. i, 'iS 1St itw #; w S~~~~~j , < t. - _. t'i * :' ."I~~~~~~.

V

7 .. -r------7 http://jnnp.bmj.com/

+ I on September 27, 2021 by guest.

FIG. 14.-Coronal section of lower medulla (x 95), ventral aspect showing ink deposit around three vessels. FIG. 15.-Coronal section of Ammon's horn (x 110), showing subcortical vessel surrounded by ink. FIG. 16.-Longitudinal section of cortical vessel (x 500) showing perivascular deposition of fine ink particles and the constriction of the perivascular space at its mouth. FIG. 17.-Longitudinal section of cervical spinal cord (x 90) showing ink around vessels in the central . FIG. 18.-Cerebellar cortex (x 55) showing perivascular ink deposits around vessels passing through Purkinje cell layer. FIG. 19.-Vessels entering globus pallidus (x 85) surrounded by ink deposits. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

THE PENETRATION OF PARTICULATE MA1TER FROM C.S.F.2213 around the perivascular space and in certain regions direction from the point of injection. It must be just below the pia. concluded then that the centrifugal fluid stream which is responsible for the transport of india ink Discussion particles through the dorsal root ganglion to the The existence of an anatomical cul-de-sac of the spinal nerve (in some 48 hours) is probably expended subarachnoid space at the proximal pole of the root within the first few millimetres of the nerve, and is ganglion and around the ventral root has been unable to influence the dispersion of an indicator established by the present investigation, and confirms injected at a more distal site in the nerve. the findings of Weed (1914), Iwanow and Romo- It is interesting to consider the reasons for the danowsky (1928), Mulder (1938), and others. No failure of Weed (1914) to note any centrifugal evidence has been found in support of the con- migration of his intrathecally introduced suspension tention of Hassin (1947) that the subarachnoid of lamp black. The most probable explanation space is in anatomical continuity with the tissue involves the factors of survival time and the- dimen- spaces of the dorsal root ganglion and the peri- sions of the particulate indicator. The duration neural space of the mixed spinal nerve. of Weed's experiments ranged from one to six hours, Continuity has, however, been demonstrated intervals too brief to permit the slow centrifugal between the sub-dural space of the nerve roots and escape of cerebrospinal fluid along the nerve to the sub-perineural space of the spinal nerve. There manifest itself. It has been noted above that some is a free passage of particulate matter from sub- 12 hours are required for the entry of ink particles arachnoid to sub-dural space and thence towards into the root ganglion, and about 48 hours for the the peripheral nerve, a finding which at first glance staining of both ganglion and mixed spinal nerve. gives the impression that the subarachnoid space Reference has been made elsewhere (Brierley and and the sub-perineural space are in direct com- Field loc. cit.) to the question of particle size and munication. egress from the subarachnoid space. The small Protected by copyright. From the point of view of physiology and patho- size of the particles employed in this work may logy the main interest of the investigation lies in well facilitate their passage through the wall of the the peculiar permeability of the walls of the sub- cul-de-sac and the intercellular spaces of the dorsal arachnoid cul-de-sac to sufficiently fine particulate root ganglion. matter. From this it is reasonable to conclude that The chief pathological problem for which the a certain proportion of the cerebrospinal fluid results of the present investigation may have signi- must escape by this route into the ganglion and ficance is that of tabes dorsalis whose pathogenesis spinal nerve. However, such an outflow can still lacks adequate explanation. Various hypotheses hardly make any appreciable contribution to have been advanced, each suggesting a different the total excretion of the cerebrospinal fluid, anatomical site as the primary point of attack of the and is likely to be of even lesser magnitude spirochkte. than the excretion into lymphatic channels At the present time the dorsal r~oot is usually con- described previously (Brierley and Field, 1948). sidered to be the first neural site to be involved in Nevertheless, this escape of cerebrospinal fluid the disease. This structure presents two so-called into the nerve must be taken into account in any vulnerable zones: (1) the point of penetration of consideration of the extent and direction of a the spinal pia where the is deficient, http://jnnp.bmj.com/ possible endoneural fluid stream (frequently and and (2) the radicular nerve of Nageotte lying just incorrectly referred to as nerve lymph), as it would proximal to the root ganglion within the sub- obviously be difficult to envisage a centripetal fluid arachnoid cul-de-sac. flow in any intraneural situation other than within It is tempting to consider the pathogenesis oftabes the axis cylinders. That the inter-fibre spaces of in the light of the behaviour of india ink particles as the peripheral nerve provide a free and extensive described above, providing the initial assumption potential pathway for injected material has been is made that the spirochetes gain access to the demonstrated by Mulder (1938), for a wide variety cerebrospinal fluid at an early stage of the disease. on September 27, 2021 by guest. of indicators, and by Brierley and Field (1949) for It would follow then that spirochetes would accum- radioactive phosphorus and suspensions of india ulate in the cul-de-sac of the spinal nerve roots and ink. The literature does not as yet provide any particularly in the lumbosacral region. The radicular convincing evidence of a physiological fluid move- nerve of Nageotte would then be open to their ment in either a central or a peripheral direction. attacks as well as the ganglion cells in relation to Mulder (1938) has shown that small intraneural the cul-de-sac. Penetration of the arachnoid (aided injections of dyes, particulate suspensions, and oily by the motility of the ) would lead to a substances are seen to diffuse equally in either permeation of the sub-dural space and the initiation J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

214 J. B. BRIERLEY of the usual hyperplastic reaction in arachnoid, from such agents as carbon monoxide, potassium dura, and epidural tissue. cyanide, ether, morphine, respiratory failure, and dinitrobenzol poisoning (reviewed by Meyer, 1937). The Perivascular Spaces.-The somewhat un- In addition siderophilic deposits occur normally in expected discovery of particulate matter in the these two zones in man and inorganic iron is fre- depths of the perivascular spaces at once conflicted quently demonstrable in and around the walls of with the current conception of an outward move- the vessels of the globus pallidus. Alexander (1942) ment of fluid from the perineuronal spaces to the has also pointed out that the two most vulnerable subarachnoid space, as advanced by Weed (1914), structures of the , the globus pallidus Schaltenbrand and Bailey (1927), and Kubie (1928). and the cornu Ammonis, are both supplied by the In the absence of passive transport by phagocytes anterior choroidal artery (in man). there is at present no explanation for this unusual The presence of perivascular spaces readily finding. The ink suspension, being isotonic with accessible to cerebrospinal fluid may be yet another the cerebrospinal fluid, cannot be held responsible aspect of the unusual vascular relations of these for the generation of an abnormal fluid flow. parts. Further, the maintenance of the animal in a normal Reference has already been made elsewhere (Field state of hydration would not set up the reversed and Brierley, 1948) to a possible relationship between flow normally associated with the intravenous the distribution of the lesions of poliomyelitis and injection of hypertonic solutions. the penetration of the central nervous system by There is a possibility that the experimental virus conveyed along perivascular spaces. The technique for the introduction of the indicator may results described above have revealed a far more be in part responsible. The free escape of cerebro- extensive invasion of perivascular spaces than was spinal fluid from the drill hole effects a partial previously envisaged, and may well raise anew the emptying of the subarachnoid space and it may well role of the cerebrospinal fluid in the pathogenesis Protected by copyright. be that some drainage of the perivascular spaces of the disease. occurs at the same time. The subsequent inflow of the ink refills the subarachnoid space, and the Summary restoration of normal cerebrospinal fluid pressure The literature relating to the subarachnoid and volume may temporarily reverse the " normal " cul-de-sac" is reviewed. -flow within the perivascular spaces to permit some A technique for introducing an india ink sus- ingress ofparticles. It seems unlikely, however, that pension into the cisterna magna of the rabbit the latter could be at all extensive in view of the without resultant rise in intracranial pressure is fact that abnormal (low) pressure conditions exist described. for only a minute or so during each introduction. Forty-eight hours after the introduction typical If there is a resumption of an outward fluid stream- ink-cuffs " are seen at the proximal poles of the ing it is to be expected that some washing out of root ganglia. the particles would take place to leave a minimal There is naked eye evidence of spread of the ink deposit in the depths ofthe spaces. The histological over the surface of the ganglia, particularly in the appearances suggest that this is not the case, as in ganglia of the sciatic nerve. many instances the deposition of the ink is uniform The distribution of ink over the cerebral hemi- http://jnnp.bmj.com/ throughout the perivascular space. spheres, brain stem, and cerebellum is described. It may be concluded, then, that under physio- Histological examination revealed (normal logical pressure conditions fine particulate sus- material) that the subarachnoid space terminates pensions in the cerebrospinal fluid are capable of near the proximal pole of the root ganglion as a penetrating the perivascular spaces in a period of well defined cul-de-sac. 24 to 48 hours and that the extent of this penetra- India ink particles accumulate in the cul-de-sac tion is related to the quantity of- the indicator in whence they pass into the sub-dural space, into the the subarachnoid space. It is--probable that with substance of the ganglion, and finally into the on September 27, 2021 by guest. adequate and uniform filling of the whole space mixed spinal nerve. there would be a considerable degree of filling of The perivascular spaces of the central nervous most of the perivascular spaces of the central system contain amounts of india ink that vary with nervous system. the degree of filling of the adjacent subarachnoid The remarkable constancy with which ink space. particles were found in the perivascular spaces of the The perivascular spaces of the hippocampus, hippocampus and globus pallidus recalled the globus pallidus, and cerebellum were well filled in involvement of these regions in the anoxia resulting all members of the series. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.13.3.203 on 1 August 1950. Downloaded from

THE PENETRATION OF PARTICULATE MATTER FROM C.S.F. 215 The physiological and pathological significance des anatomischen Institutes der Kaiserlichen Universitat of these findings is discussed. Kyoto. Series A. 1. Goldmann, E. E. (1913). "Vitalf7arbung am Zentral- nervensystem. Beitrag zur Physio-Pathologie des The author records with pleasure his indebtedness to Plexus chorioideus und der Hirnhaute." Berlin. Professor A. Meyer for his valuable criticism, to Mrs. Hassin, G. B. (1947). J. Neuropath. exp. Neurol., 6, 172. E. Beck for the photomicrographs, to Mrs. H. Cox for Hurst, E. W. (1932). J. Path. Bact., 35, 41. the histological preparations, and to Mr. P. Glenn for Iwanow, G., and Romodanowsky, K. (1928). Z. ges. skilful assistance in the operative procedures. exp. Med., 58, 596. Key, A., and Retzius, G. (1875). " Studien in der Anatomie des Nervensystems und des Bindegewebes." REFERENCES Stockholm. Alexander, L. (1942). Res. Publ. Ass. nerv. ment. Dis., Kubie, L. S. (1928). Brain, 51, 244. 21, 77. Meyer, A. (1937). Proc. R. Soc. Med., 29, 1175. Brierley, J. B., and Field, E. J. (1948). J. Anat., Loud., (Section of psychiatry, p. 49.) 82, 153. Mulder, J. D. (1938). Acta. neerl. Morphol., 1, 289. ,~9 ~~(1949). Journal of Neurology, Neurosurgery Quincke, H. (1872). Arch. Anat. Physiol., Lpz. 153. and Psychiatry, 12, 86. Schaltenbrand, G., and Bailey, P. (1927). J. Psychol. Elman, R. (1923). Bull. Johns Hopk. Hosp., 34, 99. Neurol., Lpz., 35, 199. Field, E. J., and Brierley, J. B. (1948), J. Anat., Lond., Somberg, H. M. (1947). J. Neuropath. exp. Neurol., 82, 198. 6, 166. Funaoka, S. (1930). "Eine Injektionsmethode des Weed, L. H. (1914). J. med. Res., 31, 93. Nervensystems," Arbeiten aus der dritten Abteilung Wustmann, 0. (1933). Dtsch. Z. Chir., 238, 529. Protected by copyright. http://jnnp.bmj.com/

Lu OCT r o C% on September 27, 2021 by guest.