Retrograde transport of horseradish peroxidase in transected axons. 3. Entry into injured axons and subsequent localization in perikaryon

Brain Research, 115 (1976) 201-213 © Elsevier Scientific Pubhshing Company, Amsterdam - Printed m The Netherlands

201

R E T R O G R A D E T R A N S P O R T OF H O R S E R A D I S H PEROXIDASE IN T R A N S E C T E D AXONS. 3. E N T R Y INTO I N J U R E D AXONS A N D SUBS E Q U E N T L O C A L I Z A T I O N IN P E R I K A R Y O N

KRISTER KRISTENSSON and YNGVE OLSSON Neuropathological Laboratory, Department of Pathology II, University of Linkbping and Neuropathological Laboratory, Institute of Pathology, University of Uppsala, Uppsala (Sweden)

(Accepted 2nd March, 1976)

SUMMARY Horseradish peroxidase (HRP) applied to crushed mouse sciatic nerves diffused through the damaged perineurium into the endoneurium. In the injured area, H R P passed into damaged myelinated and unmyelinated axons forming columns of reaction product, which extended for several millimeters proximally to the lesion. Ultrastructurally, H R P adhered to the inner surface of the axoplasm and to the surfaces of neurotubules and neurofilaments in such columns. At more proximal levels axons contained H R P in vesicular and tubular organelles and, later, nerve cell bodies of the corresponding spinal ganglia showed HRP, accumulation in cytoplasmic vesicles, cup-shaped bodies, multivesicular bodies and tubules of agranular endoplasmic reticulum. Markedly less H R P reached neurons in the spinal ganglia when applied to the nerve 30 or 60 min after the crush. After such time intervals solid H R P containing axons were also less frequently observed. Conceivably, H R P enters crushed axons momentarily after a crush as an injured cell reaction. Subsequently it is incorporated into organelles higher up in the axons, from where retrograde transport to the perikaryon will fellow. This phenomenon of a sudden non-specific influx of exogenous macromolecules into axotomized neurons and their subsequent transport to the perikaryon might be relevant for development of certain biochemical and morphological responses, e.g. lysosomal alterations, of the neuron to an axonal injury.

INTRODUCTION There are several possible mechanisms by which a nerve cell body's responses to an axonal injury are triggered4,18. One is that alterations in retrograde axonal

202 transport of macromolecules provide the ~slgnal for chromatolysls' In a series ~1 experiments we are studying changes in retrograde transport reduced by axonal mlurJes. Previously, it was found that a nerve crush interrupts axonal transport ol proteins from the p m p h e r y and that the rate of retrograde transport of a macromolecule m injured axons is within the same range as that of the ascent of '~lgnal for chromatolysls'. Further, horsera&sh peroxldase (HRP) topically apphed to a crushed or a cut nerve accumulates subsequently m the penkarya of the corresponding neurons, thus demonstrating a direct pathway by which substances from the mlured region of the axons can reach the nerve cell body 14.~5 The mm of the present study was to find out how a protein tracer following mechamcal injury can enter axons, and to study its subsequent transfer to the nerve cell body including its ultrastructural localization m the axon and the perlkaryon MATERIAL AND METHODS

Animals Female, Swiss albmo mice obtained from Anticlmex, Stockholm, 60-75 days of age, were used.

Protein tracer technique Horseradish peroxidase (HRP), type VI (Sigma), was used as a protein tracer. For histochemical demonstrations of H R P , mice were fixed by perfuslon with 2.5"o glutaraldehyde in phosphate buffer, pH 7.4. Specimens were post-fixed overnight. For hght mlcroscoplcal localization of HRP, 20 # m thick frozen sections were cut. They were incubated in 3,Y-diammobenzldine and hydrogen peroxide according to G r a h a m and Karnovsky 6. For electron microscopical examinations agar-embedded tissue slices, 30/~m thick, were cut with a Du Pont lnstrument/Sorvall TC-2 tissue sectioner. The specimens were washed in buffer overnight and incubated as above followed by treatment with ! o~ osmic acid for 2 h. The specimens were then dehydrated and embedded in Epon 812. Areas were selected for electron microscopy from 1 /zm thick unstained sections. Ultrathm sections were stained with uranyl and lead citrate and examined in an electron microscope. Unstained sections were also used.

Experimental procedure The scmtic nerve was crushed with a jeweller's forceps at a mid-thigh level about 20 m m from spinal ganglion L6. The H R P was apphed around the crushed region of the nerve, whlch was surrounded by a strip of surgical tape. In order to study the spread of H R P from the crushed region of the scmtic nerve to the spinal ganglion, 0.25 mg H R P was applied directly after the crush and groups of 3 mice were sampled 10 mm, 1, 3, 6 and 24 h thereafter. The sciatic nerve, divided into 5 m m long segments, and the lumbar spinal gangha were then examined for the presence of HRP. In order to follow the spread of H R P over several segments in the same nerve fiber, similar experiments were also performed using teased pre-

203 parations. In one group of 9 mice, 1 mg of H R P was applied to the exposed sciatic nerve and immediately thereafter the nerve was crushed. Thirty or 60 min later, the animals were killed and a specimen was then taken for examination including the crushed area and approximately 1 cm of the nerve above the lesion. It was partly teased and the resulting groups of fibers were incubated to test for H R P activity. After incubation, teasing was continued until smaller groups or single fibers were obtained. In another type of experiment, a time interval passed between the crush and the application of HRP. In this way the sciatic nerve on the left side was crushed and 0.25 mg H R P applied 10, 30 or 60 min thereafter in groups of 3-6 mice. At that time the right sciatic nerve was crushed and H R P applied directly. After an additional 30 min the mice were sacrificed and frozen sections from the sciatic nerve proximal to the crush examined for the presence of HRP. In one group of 6 mice the nerve was crushed and 30 min later 1 mg H R P was applied around the injured areas. One hour later the mice were killed and teased preparations were prepared as described above. In similar experiments groups of 6-9 mice were examined for the presence of H R P in the spinal ganglion cells after immediate and delayed H R P application around sciatic nerves. On the left side the nerve was crushed and 0.25 mg H R P applied 10, 30 or 60 min later. On the right side the same amount of H R P was applied immediately after a crush, which was performed at the same time as the delayed H R P application on the other side. Twenty-four hours later the mice were perfused and the spinal ganglia examined for the presence of HRP. Thus, the time interval for marker transport was the same in the two nerves. In each mouse the left and right ganglion were compared and evaluated for the presence of H R P in the neurons. All sections were read under coded numbers. For ultrastructural examination spinal gangha were sampled from 2 mice at 12, 24, 36, 48 and 72 h after the nerve crush and simultaneous application of HRP. Samples of the sciatic nerve at different levels were also examined 10 and 30 min, 1, 2 and 12 h after the crush. As controls served intact and crushed scmtic nerves and corresponding spinal ganglia to which no H R P had been applied. These specimens were otherwise treated in the same way as the experimental material. RESULTS Sciatic nerve

The light and electron microscopical distribution of H R P in the crushed sciatic nerves will first be described. After application of H R P directly to the crushed sciatic nerve the tracer diffused through the disrupted perineurium into the endoneurium as seen by light microscopy. Within 10 min H R P had entered many axons proximal to the lesion staining them intensely brown (Fig. 1). This diffuse staining extended in a proximal direction for about 5-6 mm, but did not appear to spread any further even in mice examined 6 and 24 h after the application. At this time H R P had reached the neurons in the spinal ganglia, where it was localized in cytoplasmic granules.

204

Figs. 1 a n d 2. Sciatic nerve proxHnal to the crush showing diffusion of H R P into endoneurJunl and a x o n s after l m m e d m t e apphcat~on ~FJg I), after apphcat~on with 30 m m dela.~, no d~ffusmn into a x o n s Js seen (Fig. 2) 1000. Figs 3 a n d 4 Spinal ganglia 24 h after H R P application to the crushed sciatic nerve Using dark field c o n d e n s e r H R P is seen as granules m m o s t n e u r o n s after immediate a p p h c a t m n (Fig, 4), but only m a few after application with 60 m l n delay at low magmficatlon. ," 170. Figs. 5 a n d 6 Teased groups o f nerve fibers proximal to the crush. Close to the c r u s h area H R P occurs as solid colums m axons (Fig. 5), at m o r e proximal levels H R P appears as g r a n u l a r materm[ (Fig 6).

205 In teased preparations extensive amounts of H R P reaction product was present in the crushed area, particularly in the mice given H R P immediately before the nerve injury. In this zone there was activity on the external surfaces of the fibers and diffusely in the axons as well (Fig. 5). More proximally (above 5 mm) the extracellular reaction product diminished but in many fibers HRP remained in axons. In areas at this level of the nerve there were columns of H R P product occupying entire segments but more often the H R P in the axons were discontinuous. In such fibers there were granules of various sizes, often occurring in groups and particularly frequent close to the nodes of Ranvier (Fig. 6). In the node we often observed dense products filling the entire axonal area. This type of granular axonal dlstribution of H R P was seen in 7 out of 9 investigated nerves. In the group of mice with H R P applied 10 min after the crush, diffusely stained axons were seen to a similar extent as after direct application. However, after a time interval of 30 or 60 min, a different picture was consistently present. In the zone of the nerve close to the crush H R P appeared in the endoneurium outlining the myelin sheaths and the infoldings at the node of Ranvier, but only a few axons contained diffuse columns of tracer (Fig. 2). This difference was also seen in the teased preparations. When H R P was applied 30 min after the crush the extracellular activity was marked, but only in 2 out of 6 mice could axonal deposits be found and, in particular, the nodal aggregations were almost absent. In a separate series of mice, electron microscopy was performed in order to obtain more detailed information on the localization of H R P in the crushed nerve. In this series H R P was applied immediately around the crushed area and 10 min-12 h thereafter, samples were obtained from this and from the proximal part of the nerve. After having passed into the endoneurium in the crushed region H R P came in contact with the basement membranes of the Schwann cells in which it was incorporated into cytoplasmic organelles. From the results of other studies we know that H R P from the basement membrane of the Schwann cells can move into the outer mesaxon of myelinated fibers and entirely surround unmyelinated fibers in the same compartment 3,9,1e. In our mice we could, m addition, find occasional myelinated fibers in which the internal mesaxon was also filled with HRP. However, any significant entry of H R P into axons by pinocytosis did not occur in this area. The most significant axonal entry of HRP occurred in the traumatized area where the earliest recognizable light microscopical alteration was presence of solid columns of brown reaction product. This pattern corresponded ultrastructurally to extensive deposits of reaction product adhering to the inner surface of the axolemma and to the surfaces of neurofilaments and neurotubuli and was present within 10 min after the crush (Fig. 8). It appeared both in myelinated and unmyehnated fibers. Interestingly, we could often find unmyelinated fibers lacking such activity lying side by side with H R P containing fibers in the same Schwann cell. In some axons the network of tubuli and filament with adhered H R P appeared condensed lying either in the center of the axon or in a peripheral zone, whereas the remaining part appeared empty. Particularly in mice taken 2 h or later some of the sohd columns contained numerous rounded vesicles or tubular formations lacking peroxidase activity, which

206

Figs 7 11 Ultrastructural Iocahzatlon o f H R P at different levels of the axon (_'lose to the injury H R P occurs diffusely in the a x o p l a s m attaching to the fibrils (Fig 8, 40,000), often e m p t y vesicular or tubular structures are seen (Fig 7, 25,000) At more proximal levels whorls of filamentous structures with attached H R P (Fig 9, 19,800) and large H R P containing organelles are seen (Fig. 10, 32,500) A small vesicle with H R P at a higher level o f the axon (Fig 11, 60,000).

207 in transverse sections gave a honeycomb-like picture (Fig. 7). Possibly, these formations represent widened portions of endoplasmic reticulum known to react in this way proximal to an axonal lesionS, 22. To get information about the routes in the axons which H R P may utilize in order to reach the nerve cell bodies, samples were also obtained from the nerve several millimeters above the crushed area and close to the spinal ganglia. Here, light microscopy had revealed granular deposits of H R P in some of the axons. Ultrastructurally, such axons contained membrane-bound bodies heavily loaded with H R P (Fig. 10) and occasionally tubular formations, presumably agranular endoplasmic reticulum, partly filled with HRP. Vesicles attached to neurotubules as described by LaVail and LaVaiP v were seen (Fig. 1 1). In addition, HRP was present m connection with whorls of a filament-like materml with an irregular outline in some axons (Fig. 9).

Spinalganglion In order to find out if there were any differences, with regard to H R P accumulation in dorsal root ganglion cells, between experiments with immediate and delayed H R P application to the crushed nerve, a special group of mice was studied. In these mice H R P was applied immediately around the crushed right sciatic nerve and after delay around the crushed left nerve. On the right side, an intense accumulation of HRP occurred m the cytoplasm of the majority, often about 80°/o~, of the neurons (Fig. 3). After a delay of 10 min between the crush and H R P apphcatmn no difference was recorded, but after a delay of 30 or 60 min, consistently less than 25 °/o, and mostly only a few, of the neurons on the left side contained H R P as seen at a low-power magnification (Fig. 4). The intensity of the staining of the neurons also appeared to be reduced compared with the right side. The detailed localization of HRP in dorsal root ganglion cells was studied in mice taken at various intervals after application of tracer around crushed sciatic nerves. Ultrastructurally, black reaction product of H R P was seen in vesicles of varying sizes in the cytoplasm of neurons examined 12 and 24 h after the crush (Figs. 12, 17). Such H R P containing vesicles could be seen close to the Golgi apparatus, but the cisterns of the Golgi remained free of tracer (Fig. 16). In some of the larger vesicles multicentric whorls or stacks of parallel lines sometimes imitating myelin structures were seen. Multivesicular bodies also contained HRP. In most of them it was diffusely distributed in the matrix contrasting to the empty small vesicles, but occasionally the reverse distribution was seen with H R P in the vesicles but not in the matrix (Fig. 19). Sometimes small H R P containing vesicles were attached to multivesicular bodies or large vesicles (Figs. 18, 20). H R P was also seen in many tubules of agranular endoplasmic reticulum. Such H R P containing tubules came very close to the outer nuclear membrane, but we were unable to trace any H R P directly to the perinuclear space (Figs. 14 and 15). Occasionally, H R P containing tubules formed cup-shaped bodies. Tubules with H R P were often attached as tails to larger vesicles and multivesicular bodies. In a previous light microscopic study, chromatolysis with fragmentation and

208

Fig. 12 HRP m vesicles of the perlkaryon 12 h after apphcatlon, prior to changes m other organelles < 6500. Fig 13. Nerve cell 48 h after the crush showing signs of chromatolysJs with dispersion of Nlssl bodies and still many HRP containing organelles. . 6000

209

5

~5"

e

•" fa

Figs 14-21. H R P m different organelles of the perikaryon including agranular endoplasmzc reticulum (Fig. 14, x 14,000), cisterns close to the nuclear membrane (F~g 15, ;: 24,000), vesicles of various sizes near Golg~ apparatus (Fig. 16, "< 31,000), complex dense bo&es (Fig. 17, "~ 40,000; Fig. 21, ~< 80,000), cup-shaped bo&es (Fig 18, ~ 41,000), m or outside multweslcular bodies (Fig. 19, ,61,000; Fig. 20, × 65,000). Figs. 19-21 are from unstained sections.

210 peripheral rimming of Nlssl bodies was evident in most neurons 36 h after the crush 1~. In neurons examined ultrastructurally after 36 h the Nissl bodies tended to fragment and be displaced to the periphery. The nucleus also became eccentrically located and often displayed an Indented membrane. This was more evident after 48 and 72 h. In these neurons H R P occurred in similar organelles as in those examined after the earlier time intervals (Fig. 13). The organelles tended to concentrate towards the center of the perikarya. Vesicles containing laminar structures and whorls were prominent (Fig. 2 I). DISCUSSION After a nerve crush the perineurial diffusion barrier to macromolecules such as H R P is broken down permitting substances to penetrate into the endoneurlum. Such damage to the perineurial diffusion barrier persists for several months after a nerve crush 26. In the present study H R P was seen to enter axons proximally to the lesion as a solid diffuse column as seen by light microscopy. Such a pattern is consistent with a so-called injured cell reaction during which an exogenous protein leaks through a damaged plasma membrane. This is in contrast to pinocytosls whereby protein is incorporated into cytoplasmic vesicles over intact plasma membranes z,s. 10,2~. Similar solid columns of H R P have previously been observed light microscopically in axons of central or peripheral origin14,21, ~4. The ultrastructural correlate appears to be a marked tendency for H R P to adhere to the inner surface of the axolemma and to surfaces of neurotubuli and neurofilament with a redistribution and condensation of such structures in the injured axons. The diffuse staining of axons extended proximally only a few millimeters. Higher up, H R P appeared in organelles in the axoplasm and the cytoplasm of the nerve cell bodies. When H R P was applied 30 or 60 min after the crush, only a few axons were stained diffusely and markedly less of the tracer reached the ganglion than after immediate apphcation. This suggests that much of the HRP, which immediately passes over the injured axoplasmic membrane, is eventually incorporated into organelles and in these transported to the nerve cell bodies. However, the mode by whlch such an intraaxonal incorporation would occur is still not known. Possible explanations for the relatively small axonal uptake of H R P at short time intervals after the crush may be a plugging of the axon due to alterations of the axoplasm at the site of the lesion or, for myelinated fibers, an ensheathment by the myelin, which would seal off the blind ends of the axons 19,~9. In a recent study, Halperin and LaVall 7 found a time delay of 0.5 h in transfer of H R P from the retina to the lsthmo-optlc nucleus in injured axons as compared to uninjured ones in the chick optic system directly after a lesion. At later stages, 6.7518 h after the injury, the neurons of the injured axons had accumulated much more H R P that the others. After 24 h the cells contained distinctly less tracer. It is difficult to make a detaded comparison between our study and that of Halperln and LaVall, since very different experimental models were used. In our model we have not compared injured with uninjured axons and we don't know if there is a similar initial time

211 delay belbre H R P is incorporated into axonal organelles after which transport will follow. The injury to the axons induced in their study by scratching the retina may be greater as compared to the crushing of the axons m the present one. Further, the neurons in their model are comparatively young and will rapidly degenerate after the axonal injury. Such factors may be involved in the increased tracer accumulation in the neurons, which in their study persisted for many hours after the injury. From our study it is not known if H R P is again taken up to a large extent when axons begin to regenerate. However, during the period ofreinnervation of a muscle a marked tracer uptake from the periphery occurs 16. It is, therefore, apparent that very complex alterations in the uptake and retrograde transport of macromolecules may follow an axonal injury. In the axons and particularly m the nerve cell bodies HRP was present in vesicles and in tubules of the agranular endoplasmic reticulum. Some of the vesicle-like structures may represent cross-sections or focal dilatations of agranular endoplasmic reticulum. The localization of HRP to these structures is in accordance with previous observations on intact neurons18,17, 24. The localization to agranular endoplasmic reticulum is of particular interest since this structure is presently implied as important in orthograde axoplasmic transport phenomena 5. Furthermore, agranular endoplasmic reticulum is apparently in direct continuity with the perinuclear space providing a compartment for transfer of materials to this space. In fact, we found that HRP taken up from the area of the nerve crush came very close to the nucleus in agranular endoplasmic reticulum. This could then be an important pathway for informative molecules to reach the nucleus and signalling the neuronal response with formation of DNA-dependent RNA-synthesis 2s. In the neuronal cytoplasm we found already 12 h after the crush many dense bodies of complex forms, e.g. mhomogeneity of matrix density, inclusion of whorls and linear densities as previously described by Matthews and Raisman22, 2a. As seen in unstained sections some of these probably contained HRP, suggesting that they may partly reflect accumulation of material coming from the injured area of the axons. Tentatively, injured axoplasm at the crush is incorporated into organelles which are transported back to the perikaryon where breakdown in lysosomal organelles may occur. This may be one factor in explaining the increase in dense bodies, autophagic vacuoles and acid phosphatases in chromatolytic neurons1,11, ~. It should, however, be pointed out that there exist alternative fates for the injured axoplasm, since Spencer and Thomas 27 showed that axoplasm may be removed by the surrounding Schwann cell. In conclusion, exogenous macromolecules can readdy diffuse into injured axons proximally to a nerve crush. This is apparently a transitory phenomenon and is probably responsible for the following marked accumulation of exogenous proteins in the nerve cell bodies. This demonstrated phenomenon, that large molecules from the injured axoplasm and its surroundings reach the perlkaryon is of great interest since it may be related to changes in the lysosomal organelles in the nerve cell body and the process by which chromatolysis, e.g. dispersion of the Nissl bodies, is signalled. It has been suggested that the onset of chromatolysis may be triggered by e~ther the

212 i n t e r r u p t i o n o f s u p p l y o f s o m e s u b s t a n c e s n o r m a l l y c a r r i e d f r o m the p e r i p h e r y by a x o p l a s m l c t r a n s p o r t o r by the a s c e n t o f a ' w o u n d s u b s t a n c e '~'° f r o m the injured a x o n f o r w h i c h a p a t h w a y as d e m o n s t r a t e d m the p r e s e n t study. A C K N O W L E D G E MENTS T h i s s t u d y was s u p p o r t e d by G r a n t s f r o m t h e S w e d i s h M e d i c a l R e s e a r c h C o u n cil, p r o j e c t nos. B75-12X-4480-01 a n d B75-12X-03020-07B.

REFERENCES 1 Barron, K. D., Chiang, T. Y., Daniels, A. C. and Doolm, P. R., Subcellular accompaniments of axon reaction in cervical motoneurons of the cat. In H. M. Zimmerman (Ed.), Progress in Neuropathology, vol I, Grune and Stratton, New York, 1971, pp. 255-280. 2 Brlghtman, M. W., The distribution within the brain of ferntin injected into the cerebrospinal fired compartments. I. Ependymal distribution, J Cell Biol., 26 (1965) 99-123 3 Bock, P. und Hanak, H., Die Vertellung exogener Peroxydase Im Endoneuralraum, Histochemte, 25 (197t) 361-371. 4 Cragg, B. G , What is the signal for chromatolysis? Brain Research, 23 (1970) 1-21 5 Droz, B., Rambourg, A. and Koenig, H L., The smooth endoplasmic reticulum, structure and role m the renewal of axonal membrane and synaptlc vesicles by fast axonal transport, Brain Research, 93 (1975) 1-14 6 Graham, R. C. and Karnovsky, M J., The early stages of absorption of injected horseradish peroxidase in the proximal tubules of mouse kidney. Ultrastructural correlates by a new techtuque, J Histochem. Cytochem., 14 (1966) 291-302. 7 Halperln, J. J and LaVafl, J H., A study of the dynamics of retrograde transport and accumulation of horseradish peroxtdase in injured neurons, Brain Research, 100 (1975) 253-269. 8 Heuser, J. E and Reese, T. S., Evidence for recycling of synapttc vesicle membrane during transmitter release at the frog neuromuscular junction, J. Cell BioL, 57 (1973) 315-344. 9 H~rano, A., Becker, N. H. and Zxmmerman, H. H., Isolation of the periaxonal space of the central myehnated nerve fiber with regard to the diffusion of proxldase, J Histochem. Cytochem., 171 (1969) 512-516 10 Holtzer, H and Holtzer, S , The in vitro uptake of fluorescem labelled plasma proteins. 1 Mature cells, C R. Lab. Carlsberg, 31 (1960) 373-408. 11 Holtzman, E., Novikoff, A. B. and Vlllaverde, H , Lysosomes and G E R L in normal and chromatolytlc neurons of the rat ganglion nodosum, J Cell Biol, 33 (1967) 419-435. 12 Krishnan, N. and Stager, M , Penetration of peroxldase into peripheral nerve fibers, ,4mer J Anat., 136 (1973) 1-14. 13 Knstensson, K. and Olsson, Y., Uptake and retrograde transport of peroxldase m hypoglossal neurons Electron microscopic locahzation in the neuronal perikaryon, Acta neuropath. (Berl), 19 (1971) 1-9 14 Kmstensson, K. and Olsson, Y., Retrograde transport of horseradish peroxidase m transected axons 1 Time relationship between transport and induction of chromatolysis, Brain Research, 79 (1974) 101-109. 15 Krlstensson, K. and Olsson, Y., Retrograde transport of horseradish peroxldase in transected axons. 2. Relations between rate of transfer from the injury to the perikaryon and onset of chromatolysis, J. Neurocytol., 4 (1975) 653-661. 16 Kristensson, K. and Sjostrand, J., Retrograde transport of protein tracer m the rabb~t hypoglossal nerve during regeneration, Brain Research, 45 (1972) 175-181. 17 l_aVail, J. H. and LaVail, M. M., The retrograde intraaxonal transport of horseradmh peroxldase m the chick visual system A hght and electron microscopic study, J. comp. Neurol., 157 (t974) 303-358. 18 Lleberman, A. R., Some factors affecting retrograde responses to axonal les~ons. In R. Bellalrs and G. G Gray (Eds.), Essay on the Nervous System, Clarendon Press, Oxford, 1974, pp. 71-105.

213 19 Lubifiska, L , Outflow from cut ends of nerve fibres, Exp. Cell Res., 10 (1956) 40--47. 20 Lubifiska, L., On axoplasmlc flow, Int. Rev. Neurobiol., 17 (1975) 241-296. 21 Lulten, P. G. M., The horseradish peroxldase technique applied to the teleostean nervous system, Brain Research, 89 (1975) 181-186. 22 Matthews, M. R., An ultrastructural study of axonal changes following constriction of postganglionic branches of the superior cervical ganglion in the rat, Phil. Trans. B, 264 (1973) 479-508. 23 Matthews, M. R. and Raisman, G , A light and electron microscoplc study of the cellular response to axonat injury in the superior cervical ganghon of the rat, Proc. roy. Soc. B, 181 (1972) 43-79. 24 Nauta, H. J., Prltz, M. B. and Lasek, R. J , Afferents to the rat caudoputamen studied with horsera&sh peroxidase. An evaluation of a retrograde neuroanatomical research method, Brain Research, 67 (1974) 219-238. 25 Olsson, Y. and Hossman, K.-A., Fine structural localization of exudated protein tracers in the brain, Acta neuropath. (BerL), 16 (1970) 103-116. 26 Olsson, Y. and Kristensson, K., The perineurmm as a diffusion barrier to protein tracers following trauma to nerves, Acta neuropath. (BerL), 23 (1973) 105-111. 27 Spencer, P. S and Thomas, P. K., Ultrastructurai studies of the dying-back process. II. The sequestration and removal by Schwann ceils and oligodendrocytes of organelles from normal and &seased axons, J. Neurocytol., 3 (1974) 763-783. 28 Torvik, A. and Heding, A., Effect of actinomycin D on retrograde nerve cell reaction. Further observations, Acta neuropath. (BerL), 14 (1969) 62-71. 29 Zelen~i, J., Lubifiska, L. and Gutmann, E., Accumulation of organelles at the ends of interrupted axons, Z. Zellforsch., 91 (1968) 200-219.