- Email: [email protected]
Effects of zinc on the cytoskeletal proteins in the central nervous system of the rat
Brain Research, 220 (1981) 139-149 Elsevier/North-Holland Biomedical Press
139
E F F E C T S OF Z I N C O N T H E C Y T O S K E L E T A L P R O T E I N S I N T H E C E N T R A L N E R V O U S SYSTEM OF T H E R A T
YVONNE KRESS, FELICIA GASKIN*, CELIA F. BROSNAN and SEYMOUR LEVINE Department of Pathology, Albert Einstein College of Medicine, The Bronx, N.Y. 10461 and (S.L.) Department of Pathology, New York Medical College, Valhalla, N. Y. 10595 (U.S.A.)
(Accepted January 22nd, 1981) Key words: zinc - - microtubules - - intermediate filaments - - central nervous system - - rat
SUMMARY To test for in vivo zinc neurotoxicity on the cytoskeleton of neurotubules and intermediate filaments, Zn wires were implanted into the brains of adult Lewis rats for periods of 1-35 weeks. After 16 weeks of implant, some neurons showed bundles of intermediate filaments which were often localized in the perinuclear area. At the same time, occasional 200 nm tubular-like structures were seen in swollen dendrites. These structures were morphologically similar to Zn ion-induced aggregates of pure tubulin and structures found in ZnSO4-treated dorsal root ganglion organotypic cultures. The 200 n m structures in dendrites and the intermediate filaments in neurons increased in frequency with time. After 35 weeks of Zn wire implant, few microtubules could be found in the lesion. All the animals showed an astrocyte and glial filament proliferation with axoglial membrane specialization. Other wires studied: Pt, Ni, Co, Mg, demonstrate that Zn wires have a specific effect on cytoskeletal proteins in the CNS of the rat and many of the effects can be explained by an interaction between Zn and tubulin.
INTRODUCTION Zinc ions induce microtubule protein (tubulin and microtubule-associated proteins) and tubulin to form sheets with more than 13 protofilaments and 200 nm tubular structures respectivelyS,9,1L Addition of zinc ions to microtubules in vitro results in the loss of microtubules and the formation of wide sheets only if there is sufficient tubulin in the mixture 7. This latter finding suggests that zinc does not interact directly with microtubules, but interferes with tubulin assembly. Neurotubules in organotypic cultures of dorsal root ganglia are disrupted by zinc ions and the zinc* To whom correspondence should be addressed.
g
141 treated cultures contain 200 nm structures morphologically similar to the 200 nm tubulin aggregates induced by zinc in a cell-free system. Neurofilament proliferation, especially in the perikaryon of neurons, was a striking feature in the zinc-treated cultures 1°. Other in vitro studies on the effect of zinc ions on neurofibrils show: (1) small amounts of zinc (5 × 10-6 M) stimulated rapid axonal transport of proteins and 1 mM zinc inhibited transport in frog ganglia and nerveS; (2) 10 mM ZnC12 stabilized neurotubules and induced C-shaped neurotubules in rat peripheral nerve13; and (3) colchicine binding activity (a measure of tubuliri) in soluble brain fractions was lowered by zinc ions 5. All of these results suggest that elevated zinc levels in nerve tissue would interfere with microtubule structure and function and disrupt the cytoskeleton. Levine and SowinskP 7 used zinc wire and zinc powder implants in the central nervous system of rats to show that zinc induces changes similar to a cellmediated immune reaction. We chose to use zinc wire implants for in vivo studies since the cytoskeletal components along the wire tract would be exposed to a slow continuous release of zinc ions. We used electron microscopy to follow the effects of zinc wire implants (1-35 weeks) on neurotubules, intermediate filaments and glial filaments and the possible induction of 200 nm tubules. MATERIALS AND METHODS Metal wires were implanted into the brains of adult Lewis rats approximately 200-300 g in weight, as previously described 17. Zinc, magnesium, cobalt, nickel, and platinum wires from Thiokol-Ventron, Danvers, Mass., were 0.25 mm in diameter and 7-8 mm in length. At least two rat brains with zinc wire implants were examined at 1, 2, 3, 6, 9, 16, 20, 25, 30 and 35 weeks. Brains with platinum wire implants were the control, and they were examined at 1, 3, 6, 16 and 25 weeks. Nickel wire effects were studied at 6, 9, 16, 25 and 30 weeks. Cobalt and magnesium wire implants were examined at 6 and 9 weeks. Animals were anesthetized with Nembutal and fixed by whole-body perfusion with a 4 ~ paraformaldehyde flush, followed by 5 ~o glutaraldehyde in phosphate buffer, pH 7.4, at room temperature. The brain was excised and the metal wire carefully removed from the brain. Thin slices of tissue were cut around the wire lesion, post-fixed in 2 ~ Dalton's osmium dichromate for one hour. The tissue was then dehydrated in graded ethanol solutions and flat embedded in Epon. One micron Fig. 1. a: thin section of pelleted zinc ion-induced aggregates of tubular-like structures 200 nm in diameter, cut longitudinally (ST) and in cross-section (arrows). b" dorsal root ganglion organotypic culture treated with 1 mM ZnSO4 for 24 h, showing a swollen axon (AX) with no neurotubules. Instead one sees structures (ST) resembling the zinc ion-induced 200 nm tubular structures of tubulin. c: dendrites in the lesion after a 16 week zinc wire implant in rat brain. Dendrites (D) appear swollen and contain few neurotubules (NT). Occasionally one sees structures measuring 200 nm in diameter (ST) as well as sheet-like aggregates (S). Some cross-sections of tubular-like structures are also seen (arrows). These in vivo structures appear morphologicallysimilar to the 200 nm structures of tubulin found in vitro and zinc-induced structures in tissue cultures, d-f: dendrites in 20 week lesions around zinc wire implants in rat brains. More swollen dendrites (D) containing 200 nm structures (ST) can be seen at this time point. Besides the structures the dendrites also show a few broken neurotubules (NT) and some sheet-like aggregates (S).
142 sections were cut and stained with buffered toluidine blue for light microscopy. Thin sections were stained with uranyl acetate for 20 min, followed by lead citrate for 5 min. They were examined with a Siemens electron microscope S 101.
RESULTS The lesions caused by the metal wire implants in the brains of adult Lewis rats extended 1-2 m m into the surrounding parenchyma. A further characterization of the zinc-induced mononuclear infiltrates will be described elsewhere 1. Electron microscopy was used to analyze the ultrastructural effects of the metal wire implants.
Loss o f neurotubules and induction o f 200 nm tubular structures Zinc wire implants of 1-10 weeks had little effect on neurotubules. However, a few dendrites adjacent to the wire tract were swollen and occasionally contained only neurotubule fragments. At 16 weeks of wire implant, most dendrites in the lesion were swollen and distorted and contained mainly remnants of neurotubules, amorphous material and long sheet-like aggregates. Microtubules continued to decrease with the time of zinc wire implant and by 30 weeks they were rarely found throughout the lesion. The tissue was carefully examined for structures which morphologically resemble the fixed and pelleted zinc-induced tubulin aggregates (see ref. 9, and Fig. la) and the abnormal structures found in some axons of zinc-treated mouse dorsal root ganglia organotypic cultures (see ref. 10, and Fig. lb). Serial sections of the zincinduced tubulin aggregates suggested a long tubular or wrapped sheet structure 200 nm in diameter and so these aggregates will be referred to as 200 nm tubular structures. After 16 weeks of zinc wire implant the dendrites in the lesion occasionally contained structures similar to the 200 nm tubular structures (Fig. lc). Increasing numbers of the 200 nm tubular structures were found in the dendrites in up to 25week-old lesions (Fig. ld, e, f). Platinum (1-25 weeks) and nickel wire implants (6-30 weeks) had no apparent effect on microtubules and did not induce 200 nm tubular structures. Due to their toxicity, cobalt and magnesium were only examined at 6 and 9 weeks. Neither induced the 200 nm tubular structures at these early times.
Intermediate filaments in neurons Zinc had no obvious effect on neurofilaments at early times. However, the appearance of large bundles of intermediate filaments and loss of neurotubules in some neurons became common at 16 weeks after zinc wire implantation. These filaments were often localized in the perinuclear area (Fig. 2) and measured 10 nm in diameter, were cross-bridged, and often found in close association with the nucleus. At later times (30-35 weeks) many axons and dendrites were filled with precipitated amorphous material. However, some axons still contained neurofilaments. None of the other metal wire implants (platinum, nickel, cobalt, and magnesium) produced neurofibrillary changes in neurons.
143
Fig. 2. Bundles of intermediate filaments (arrows) wrapped around the nucleus (N) in a neuron in the lesion after a 16 week zinc wire implant in rat brain, x 54,000. This neuron is associated with many synapses (not shown).
144
Astrocyte proliferation with axo-glial membrane specialization At all times a scarring astrocyte proliferation was seen near the zinc wire implants and was found in close proximity to the lymphocytic cuffs, surrounding capillaries and in close contact with myelinated and naked axons. Oligodendroglial cells were mixed with fibrous astrocytes. The processes of the fibrous astrocytes were filled with bundles of filaments measuring approximately 6-9 nm in diameter, which were not cross-bridged. The processes reached in between inflammatory cells and surrounded axons and dendrites (Fig. 3). Some axons appeared to be demyelinated or had only a thin layer of myelin left. Often one could observe a punctuated thickening of the membrane along astroglial cells, which were in close contact with unmyelinated axons and neurites. This desmosome-like junction was separated by a cleft measuring approximately 30 nm, and was darkly stained by floccular osmiophilic material (Fig. 4). Of the other metal wires we examined (platinum, cobalt, nickel and magnesium) only nickel produced a scarring astroglial proliferation and some membrane specialization. Platinum, in contrast, seemed to induce a glial fallout right along the wire tract. Nickel also produced a very severe and striking glial reaction with Rosenthal fiber formationaL
Enlarged and swollen mitochondria Many neurons and other cells in the lesion of zinc-treated animals contained mitochondria that were enlarged and swollen. Frequently only a membrane could be seen, with a few remnants of cristae visible. In general the cristae looked as if they were disintegrating. At later stages of zinc wire implant, many mitochondria also appeared very thin, elongated and darkly stained. These abnormal mitochondria were not due to fixation artefacts since mitochondria seen outside the lesion were normal. None of the other metal wire implants had an apparent effect on mitochondria.
Neurotoxic effects of zinc wire implants (30-35 weeks) Thick bundles of collagen were identified around the lesion and surrounding some blood vessels 30-35 weeks after zinc wire implantation. Most axons and dendrites were swollen and filled with amorphous material. Some axons contained intermediate filaments. Neurotubules were rarely seen throughout the lesion. Synaptic endings were swollen, distorted and filled with synaptic vesicles and the synapses had a fuzzy appearance. The neurons seemed to be shrunken and they contained a large dilated Golgi apparatus. Many cells had lysosomes filled with black granular material. Large membrane bound inclusion bodies of lipofuscin with paracrystalline arrays of tubular profiles intermixed with black osmiophilic material were seen. Lesions similar to neuritic plaques were found throughout the area. All the above described changes were found in a radius of approximately 1-2 m m surrounding the wire implant tract. Tissue examined in other areas of the brain appeared normal.
Other metal wire implants Platinum was relatively innocuous and served as a control wire implant. However, it did produce a reaction right along the wire tract approximately 0.5 m m
145
Fig. 3. Glial cell proliferation near the inflammatory response after a 16 week zinc wire implant. Oligodendroglial cells (O) and fibrous astrocytes (A) surround inflammatory cells (arrows) and myelinated and naked axons, x 19,400.
Fig. 4. Punctuated desmosome-like thickenings (arrows) along the axo-glia membranes in the lesion after a 16 week zinc wire implant (A, astrocyte; AX, axon). These desmosome-like junctions are darkly stained and show a cleft filled with fuzzy osmiophilic material. "/ 36,000.
147 into the parenchyma. Some dendrites in the lesion were swollen and empty and some macrophages were filled with dense material. It also caused some edema and a glial cell fallout. It is uncertain whether these changes are due to platinum or trauma. By 25 weeks of platinum wire implant most of the tissue appeared normal. The nickel wire implant caused a very strong reactive gliosis and by 9 weeks of implant Rosenthal fibers were found in the astrocytes 15. Some membrane specialization was also seen. Cobalt and magnesium wire implants produced large concentric zones of necrosis, and most of the tissue in the lesion was destroyed. Cobalt wire implants also caused some hemorrhage and lymphocytic cuffs. DISCUSSION Zinc ions play an important role in cellular metabolism. They are an integral part of tissues and biological fluids and one of the many homeostatic mechanisms regulating the reactivity of tissues and cells (see refs. 2 and 20 for reviews). Although many enzymes and cell processes require zinc, high levels of zinc may be toxic. Recent work by Constantinidis et al. 4 suggests an abnormality of zinc metabolism in Pick's disease. They found an enhanced concentration of zinc in brain and red cells of patients with Pick's disease, and an increase in urinary excretion of this metal. The pathology caused by zinc wire implants in rat brain cortex in these studies does not resemble that found in Pick's disease. It would be of interest to look at zinc wire implants in rat hippocampus. However, excess zinc in human hippocampus might result in a different pathology than in the rat. The studies described in this paper were done to gain information about the long-term effects of excess zinc on the central nervous system of the rat. Both the loss of microtubules and the induction of 200 nm tubular structures with increasing duration of the zinc wire implant strongly suggest that there is a slow release of zinc ions from the wire and that zinc ions affect tubulin assembly in vivo in the same way as in vitro. The in vivo studies imply that with excess zinc, dendrites first swell, then lose microtubules and gain 200 nm tubular structures, and finally contain precipitated or amorphous structures. These results are consistent with in vitro studies: (1) in a cell free system, low concentrations of zinc ions had little effect9 or induced normal microtubule assembly12; intermediate levels of zinc ions induced sheets or 200 nm tubular structures 9; and high concentrations of zinc ions resulted in precipitation 9; (2) in dorsal root ganglion cultures, swelling of axons, loss of microtubules and induction of 200 nm tubular structures in axons depended on both the zinc ion concentration and the time of incubation (there are no dendrites in these cultures) 1°; (3) preliminary results with mouse hippocampus cultures suggest that dendrites are more susceptible to zinc ions than axons; and again the appearance of 200 nm tubular structures in dendrites depended on time and zinc ion concentration la. The results reported here show that zinc has a dramatic effect on intermediate filaments in the neuronal perikaryon in the central nervous system of the rat similar to the changes we have reported in mouse dorsal root ganglion cultures 1°,
148 and should be added to the list of toxins such as aluminum, vinblastine, colchicine, hexacarbons, acrylamide carbon disulfide and B-B'-iminodipropionitrile (see refs. 11 and 19 for reviews) which cause the accumulation of large masses of neurofilaments. It is of interest that the appearance of the intermediate filament masses coincides with the disappearance ofmicrotubules. Disruption ofmicrotubules may lead to the accumulation of the intermediate filaments around the nucleus of the neuron. The axo-glial membrane specialization occurring in the central nervous system of the zinc-treated adult rats has some similarities to the abnormal membrane relationships between naked axons and scarring astrocytes found in chronic experimental allergic encephalomyelitis lesions and multiple sclerosis plaques21,2L It has been suggested that the axo-glial membrane specializations of the desmosome-type noted in demyelinated axons probably play some supportive role for the naked axons and that these junctions help reorient and reconstruct the topography of affected fiber tracts subsequent to their recovery from the initial impact of myelin loss2L As Levine and Sowinski noted 17, the rapid development of perivascular infiltration after zinc implantation did not suggest an immunological origin, and they found that zinc had no adjuvant activity. However, both the axo-glial membrane specialization and the similarities of the infiltrates to the typical lesions of experimental allergic encephalomyelitis after zinc wire implantation are intriguing. Large mitochondria are found in a variety of tissues undel different physiologic and pathological conditions is. Zinc wire implants in the central nervous system induce changes in mitochondrial structure which are consistent with the effects of zinc ions on mitochondria in other tissues which are described in a review by Chvapil et al. 3. Zinc inhibits electron transport in the mitochondrial respiratory chain and the inhibitory effect of zinc ions on respiration has been localized between cytochromes b and cl. High levels of zinc ions also inhibit succinate dehydrogenase and cytochrome oxidase. The thick bundles of collagen around the zinc wire implant after 30-35 weeks may be due to trauma or to high levels of zinc in the central nervous system. In support of the latter possibility, zinc ions have been shown to affect collagen synthesis in a specific way compared to protein synthesis in general, and to have an effect on collagen cross-linking (see Ref. 6 for a review). ACKNOWLEDGEMENTS We thank Drs. Robert D. Terry and Dikran S. Horoupian for helpful discussions. The technical help of Ms. Clemenia Cayetano is gratefully acknowledged. This work was supported in part by Grants NS-12418 and AG-00002 from the National Institutes of Health.
REFERENCES 1 Brosnan, C., Kress, Y., Gaskin, F. and Levine, S., Further studies on the inflammatory response induced by zinc wire in the central nervous system of rats, manuscript in preparation. 2 Chvapil, M.. Effect of zinc on cells and membranes, Med. Clin. N. Amer., 60 (1976) 799-812.
149 3 Chvapil, M., Elias, S. L., Ryan, J. N. and Zukoski, C. F., Pathophysiology of zinc. In Int. Rev. Neurobiol., Suppl. l, Academic Press, Inc., New York, 1972, pp. 105-124. 4 Constantinidis, J., Richard, J. and Tissot, R., Maladie de Pick et Metabolisme du Zinc., Rev. Neurol., 133 (1977) 685-696. 5 Edstr6m, A. and Mattsson, H., Small amounts of zinc stimulate rapid axonal transport in vitro, Brain Research, 86 (1975) 162-167. 6 Fernandez-Madrid, F., Prasad, A. S. and Oberleas, D., Zinc in collagen metabolism. In A. S. Prasad (Ed.), Trace Elements in Human Health and Disease, Academic Press, New York, 1976, pp. 257-268. 7 Gaskin, F., In vitro microtubule assembly regulation by divalent cations and nucleotides, Biochem., 20C (1981) 1318-1322. 8 Gaskin, F., Gethner, J. S. and Kress, Y., The characterization of tubulin preparations using laser light scattering and tubulin assembly in the presence of zinc ions. In S. V. Perry, A. Margreth, and R. S. Adelstein (Eds.), Contractile Systems in Non-Muscle Tissues, North-Holland Biomedical Press, 1976, pp. 179-190. 9 Gaskin, F. and Kress, Y., Zinc ion-induced assembly of tubulin, J. biol. Chem., 19 (1977) 6918-6924. l0 Gaskin, F., Kress, Y., Brosnan, C. and Bornstein, M., Abnormal tubulin aggregates induced by zinc sulfate in organotypic cultures of nerve tissue, Neuroscience, 3 (1978) 1117-1128. I 1 Gaskin, F. and Shelanski, M. L., Microtubules and intermediate filaments, Essays in Biochemistry, 12 (1976) 115-146. 12 Haskins, K. M., Zombola, R. R., Boling, J. M., Lee, Y. C. and Himes, R. H., Tubulin assembly induced by cobalt and zinc, Biochem. Biophys. Res. Commun., 95 (1980) 1703-1709. 13 Krammer, E. B. and Zenkar, W., Effekt von Zinkionen auf Struktur und Verteilung der Neurotubuli, Acta Neuropath. (Berl.), 31 0975) 59-69. 14 Kress, Y., Gaskin, F., Brosnan, C. and Levine, S., Long term effects of zinc in the CNS of rats, J. Neuropath. exp. NeuroL, 39 (1980) 368. 15 Kress, Y., Gaskin, F., Horoupian, D. S. and Brosnan, C., Nickel induction of Rosenthal fibers in rat brain, Brain Research, 210 (198 l) 419-425. 16 Larsson, H., Wallin, M. and Edstr6m, A., Induction of a sheet polymer of tubulin by Zn 2+, Exp. Cell Res., 100 (1976) 104-110. 17 Levine, S. and Sowinski, R., Lymphocytic inflammation produced by intracerebral implantation of zinc and other metals, J. Neuropath. exp. Neurol., 37 (1978) 471-478. 18 Munn, E. A., The Structure of Mitochondria, Academic Press, New York, 1973, pp. 1--447. 19 Norton, W. T. and Goldman, J. E., Neurofilaments. In R. A. Bradshaw and D. M. Schneider (Eds.), Proteins of the Nervous System, Raven Press, New York, 1980, pp. 301-329. 20 Prasad, A. S. (Ed.), Trace Elements in Human Health and Disease. VoL 1: Zinc and Copper, Academic Press, New York, pp. 1-470. 21 Raine, C. S., Membrane specializations between demyelinated axons and astroglia in chronic EAE lesions and multiple sclerosis plaques, Nature, (Lond.), 275 (1978) 326-327. 22 Softer, D. and Raine, C. S., Morphological analysis of axoglial membrane specializations in the demyelinated central nervous system, Brain Research, 186 (1980) 301-313.
139
E F F E C T S OF Z I N C O N T H E C Y T O S K E L E T A L P R O T E I N S I N T H E C E N T R A L N E R V O U S SYSTEM OF T H E R A T
YVONNE KRESS, FELICIA GASKIN*, CELIA F. BROSNAN and SEYMOUR LEVINE Department of Pathology, Albert Einstein College of Medicine, The Bronx, N.Y. 10461 and (S.L.) Department of Pathology, New York Medical College, Valhalla, N. Y. 10595 (U.S.A.)
(Accepted January 22nd, 1981) Key words: zinc - - microtubules - - intermediate filaments - - central nervous system - - rat
SUMMARY To test for in vivo zinc neurotoxicity on the cytoskeleton of neurotubules and intermediate filaments, Zn wires were implanted into the brains of adult Lewis rats for periods of 1-35 weeks. After 16 weeks of implant, some neurons showed bundles of intermediate filaments which were often localized in the perinuclear area. At the same time, occasional 200 nm tubular-like structures were seen in swollen dendrites. These structures were morphologically similar to Zn ion-induced aggregates of pure tubulin and structures found in ZnSO4-treated dorsal root ganglion organotypic cultures. The 200 n m structures in dendrites and the intermediate filaments in neurons increased in frequency with time. After 35 weeks of Zn wire implant, few microtubules could be found in the lesion. All the animals showed an astrocyte and glial filament proliferation with axoglial membrane specialization. Other wires studied: Pt, Ni, Co, Mg, demonstrate that Zn wires have a specific effect on cytoskeletal proteins in the CNS of the rat and many of the effects can be explained by an interaction between Zn and tubulin.
INTRODUCTION Zinc ions induce microtubule protein (tubulin and microtubule-associated proteins) and tubulin to form sheets with more than 13 protofilaments and 200 nm tubular structures respectivelyS,9,1L Addition of zinc ions to microtubules in vitro results in the loss of microtubules and the formation of wide sheets only if there is sufficient tubulin in the mixture 7. This latter finding suggests that zinc does not interact directly with microtubules, but interferes with tubulin assembly. Neurotubules in organotypic cultures of dorsal root ganglia are disrupted by zinc ions and the zinc* To whom correspondence should be addressed.
g
141 treated cultures contain 200 nm structures morphologically similar to the 200 nm tubulin aggregates induced by zinc in a cell-free system. Neurofilament proliferation, especially in the perikaryon of neurons, was a striking feature in the zinc-treated cultures 1°. Other in vitro studies on the effect of zinc ions on neurofibrils show: (1) small amounts of zinc (5 × 10-6 M) stimulated rapid axonal transport of proteins and 1 mM zinc inhibited transport in frog ganglia and nerveS; (2) 10 mM ZnC12 stabilized neurotubules and induced C-shaped neurotubules in rat peripheral nerve13; and (3) colchicine binding activity (a measure of tubuliri) in soluble brain fractions was lowered by zinc ions 5. All of these results suggest that elevated zinc levels in nerve tissue would interfere with microtubule structure and function and disrupt the cytoskeleton. Levine and SowinskP 7 used zinc wire and zinc powder implants in the central nervous system of rats to show that zinc induces changes similar to a cellmediated immune reaction. We chose to use zinc wire implants for in vivo studies since the cytoskeletal components along the wire tract would be exposed to a slow continuous release of zinc ions. We used electron microscopy to follow the effects of zinc wire implants (1-35 weeks) on neurotubules, intermediate filaments and glial filaments and the possible induction of 200 nm tubules. MATERIALS AND METHODS Metal wires were implanted into the brains of adult Lewis rats approximately 200-300 g in weight, as previously described 17. Zinc, magnesium, cobalt, nickel, and platinum wires from Thiokol-Ventron, Danvers, Mass., were 0.25 mm in diameter and 7-8 mm in length. At least two rat brains with zinc wire implants were examined at 1, 2, 3, 6, 9, 16, 20, 25, 30 and 35 weeks. Brains with platinum wire implants were the control, and they were examined at 1, 3, 6, 16 and 25 weeks. Nickel wire effects were studied at 6, 9, 16, 25 and 30 weeks. Cobalt and magnesium wire implants were examined at 6 and 9 weeks. Animals were anesthetized with Nembutal and fixed by whole-body perfusion with a 4 ~ paraformaldehyde flush, followed by 5 ~o glutaraldehyde in phosphate buffer, pH 7.4, at room temperature. The brain was excised and the metal wire carefully removed from the brain. Thin slices of tissue were cut around the wire lesion, post-fixed in 2 ~ Dalton's osmium dichromate for one hour. The tissue was then dehydrated in graded ethanol solutions and flat embedded in Epon. One micron Fig. 1. a: thin section of pelleted zinc ion-induced aggregates of tubular-like structures 200 nm in diameter, cut longitudinally (ST) and in cross-section (arrows). b" dorsal root ganglion organotypic culture treated with 1 mM ZnSO4 for 24 h, showing a swollen axon (AX) with no neurotubules. Instead one sees structures (ST) resembling the zinc ion-induced 200 nm tubular structures of tubulin. c: dendrites in the lesion after a 16 week zinc wire implant in rat brain. Dendrites (D) appear swollen and contain few neurotubules (NT). Occasionally one sees structures measuring 200 nm in diameter (ST) as well as sheet-like aggregates (S). Some cross-sections of tubular-like structures are also seen (arrows). These in vivo structures appear morphologicallysimilar to the 200 nm structures of tubulin found in vitro and zinc-induced structures in tissue cultures, d-f: dendrites in 20 week lesions around zinc wire implants in rat brains. More swollen dendrites (D) containing 200 nm structures (ST) can be seen at this time point. Besides the structures the dendrites also show a few broken neurotubules (NT) and some sheet-like aggregates (S).
142 sections were cut and stained with buffered toluidine blue for light microscopy. Thin sections were stained with uranyl acetate for 20 min, followed by lead citrate for 5 min. They were examined with a Siemens electron microscope S 101.
RESULTS The lesions caused by the metal wire implants in the brains of adult Lewis rats extended 1-2 m m into the surrounding parenchyma. A further characterization of the zinc-induced mononuclear infiltrates will be described elsewhere 1. Electron microscopy was used to analyze the ultrastructural effects of the metal wire implants.
Loss o f neurotubules and induction o f 200 nm tubular structures Zinc wire implants of 1-10 weeks had little effect on neurotubules. However, a few dendrites adjacent to the wire tract were swollen and occasionally contained only neurotubule fragments. At 16 weeks of wire implant, most dendrites in the lesion were swollen and distorted and contained mainly remnants of neurotubules, amorphous material and long sheet-like aggregates. Microtubules continued to decrease with the time of zinc wire implant and by 30 weeks they were rarely found throughout the lesion. The tissue was carefully examined for structures which morphologically resemble the fixed and pelleted zinc-induced tubulin aggregates (see ref. 9, and Fig. la) and the abnormal structures found in some axons of zinc-treated mouse dorsal root ganglia organotypic cultures (see ref. 10, and Fig. lb). Serial sections of the zincinduced tubulin aggregates suggested a long tubular or wrapped sheet structure 200 nm in diameter and so these aggregates will be referred to as 200 nm tubular structures. After 16 weeks of zinc wire implant the dendrites in the lesion occasionally contained structures similar to the 200 nm tubular structures (Fig. lc). Increasing numbers of the 200 nm tubular structures were found in the dendrites in up to 25week-old lesions (Fig. ld, e, f). Platinum (1-25 weeks) and nickel wire implants (6-30 weeks) had no apparent effect on microtubules and did not induce 200 nm tubular structures. Due to their toxicity, cobalt and magnesium were only examined at 6 and 9 weeks. Neither induced the 200 nm tubular structures at these early times.
Intermediate filaments in neurons Zinc had no obvious effect on neurofilaments at early times. However, the appearance of large bundles of intermediate filaments and loss of neurotubules in some neurons became common at 16 weeks after zinc wire implantation. These filaments were often localized in the perinuclear area (Fig. 2) and measured 10 nm in diameter, were cross-bridged, and often found in close association with the nucleus. At later times (30-35 weeks) many axons and dendrites were filled with precipitated amorphous material. However, some axons still contained neurofilaments. None of the other metal wire implants (platinum, nickel, cobalt, and magnesium) produced neurofibrillary changes in neurons.
143
Fig. 2. Bundles of intermediate filaments (arrows) wrapped around the nucleus (N) in a neuron in the lesion after a 16 week zinc wire implant in rat brain, x 54,000. This neuron is associated with many synapses (not shown).
144
Astrocyte proliferation with axo-glial membrane specialization At all times a scarring astrocyte proliferation was seen near the zinc wire implants and was found in close proximity to the lymphocytic cuffs, surrounding capillaries and in close contact with myelinated and naked axons. Oligodendroglial cells were mixed with fibrous astrocytes. The processes of the fibrous astrocytes were filled with bundles of filaments measuring approximately 6-9 nm in diameter, which were not cross-bridged. The processes reached in between inflammatory cells and surrounded axons and dendrites (Fig. 3). Some axons appeared to be demyelinated or had only a thin layer of myelin left. Often one could observe a punctuated thickening of the membrane along astroglial cells, which were in close contact with unmyelinated axons and neurites. This desmosome-like junction was separated by a cleft measuring approximately 30 nm, and was darkly stained by floccular osmiophilic material (Fig. 4). Of the other metal wires we examined (platinum, cobalt, nickel and magnesium) only nickel produced a scarring astroglial proliferation and some membrane specialization. Platinum, in contrast, seemed to induce a glial fallout right along the wire tract. Nickel also produced a very severe and striking glial reaction with Rosenthal fiber formationaL
Enlarged and swollen mitochondria Many neurons and other cells in the lesion of zinc-treated animals contained mitochondria that were enlarged and swollen. Frequently only a membrane could be seen, with a few remnants of cristae visible. In general the cristae looked as if they were disintegrating. At later stages of zinc wire implant, many mitochondria also appeared very thin, elongated and darkly stained. These abnormal mitochondria were not due to fixation artefacts since mitochondria seen outside the lesion were normal. None of the other metal wire implants had an apparent effect on mitochondria.
Neurotoxic effects of zinc wire implants (30-35 weeks) Thick bundles of collagen were identified around the lesion and surrounding some blood vessels 30-35 weeks after zinc wire implantation. Most axons and dendrites were swollen and filled with amorphous material. Some axons contained intermediate filaments. Neurotubules were rarely seen throughout the lesion. Synaptic endings were swollen, distorted and filled with synaptic vesicles and the synapses had a fuzzy appearance. The neurons seemed to be shrunken and they contained a large dilated Golgi apparatus. Many cells had lysosomes filled with black granular material. Large membrane bound inclusion bodies of lipofuscin with paracrystalline arrays of tubular profiles intermixed with black osmiophilic material were seen. Lesions similar to neuritic plaques were found throughout the area. All the above described changes were found in a radius of approximately 1-2 m m surrounding the wire implant tract. Tissue examined in other areas of the brain appeared normal.
Other metal wire implants Platinum was relatively innocuous and served as a control wire implant. However, it did produce a reaction right along the wire tract approximately 0.5 m m
145
Fig. 3. Glial cell proliferation near the inflammatory response after a 16 week zinc wire implant. Oligodendroglial cells (O) and fibrous astrocytes (A) surround inflammatory cells (arrows) and myelinated and naked axons, x 19,400.
Fig. 4. Punctuated desmosome-like thickenings (arrows) along the axo-glia membranes in the lesion after a 16 week zinc wire implant (A, astrocyte; AX, axon). These desmosome-like junctions are darkly stained and show a cleft filled with fuzzy osmiophilic material. "/ 36,000.
147 into the parenchyma. Some dendrites in the lesion were swollen and empty and some macrophages were filled with dense material. It also caused some edema and a glial cell fallout. It is uncertain whether these changes are due to platinum or trauma. By 25 weeks of platinum wire implant most of the tissue appeared normal. The nickel wire implant caused a very strong reactive gliosis and by 9 weeks of implant Rosenthal fibers were found in the astrocytes 15. Some membrane specialization was also seen. Cobalt and magnesium wire implants produced large concentric zones of necrosis, and most of the tissue in the lesion was destroyed. Cobalt wire implants also caused some hemorrhage and lymphocytic cuffs. DISCUSSION Zinc ions play an important role in cellular metabolism. They are an integral part of tissues and biological fluids and one of the many homeostatic mechanisms regulating the reactivity of tissues and cells (see refs. 2 and 20 for reviews). Although many enzymes and cell processes require zinc, high levels of zinc may be toxic. Recent work by Constantinidis et al. 4 suggests an abnormality of zinc metabolism in Pick's disease. They found an enhanced concentration of zinc in brain and red cells of patients with Pick's disease, and an increase in urinary excretion of this metal. The pathology caused by zinc wire implants in rat brain cortex in these studies does not resemble that found in Pick's disease. It would be of interest to look at zinc wire implants in rat hippocampus. However, excess zinc in human hippocampus might result in a different pathology than in the rat. The studies described in this paper were done to gain information about the long-term effects of excess zinc on the central nervous system of the rat. Both the loss of microtubules and the induction of 200 nm tubular structures with increasing duration of the zinc wire implant strongly suggest that there is a slow release of zinc ions from the wire and that zinc ions affect tubulin assembly in vivo in the same way as in vitro. The in vivo studies imply that with excess zinc, dendrites first swell, then lose microtubules and gain 200 nm tubular structures, and finally contain precipitated or amorphous structures. These results are consistent with in vitro studies: (1) in a cell free system, low concentrations of zinc ions had little effect9 or induced normal microtubule assembly12; intermediate levels of zinc ions induced sheets or 200 nm tubular structures 9; and high concentrations of zinc ions resulted in precipitation 9; (2) in dorsal root ganglion cultures, swelling of axons, loss of microtubules and induction of 200 nm tubular structures in axons depended on both the zinc ion concentration and the time of incubation (there are no dendrites in these cultures) 1°; (3) preliminary results with mouse hippocampus cultures suggest that dendrites are more susceptible to zinc ions than axons; and again the appearance of 200 nm tubular structures in dendrites depended on time and zinc ion concentration la. The results reported here show that zinc has a dramatic effect on intermediate filaments in the neuronal perikaryon in the central nervous system of the rat similar to the changes we have reported in mouse dorsal root ganglion cultures 1°,
148 and should be added to the list of toxins such as aluminum, vinblastine, colchicine, hexacarbons, acrylamide carbon disulfide and B-B'-iminodipropionitrile (see refs. 11 and 19 for reviews) which cause the accumulation of large masses of neurofilaments. It is of interest that the appearance of the intermediate filament masses coincides with the disappearance ofmicrotubules. Disruption ofmicrotubules may lead to the accumulation of the intermediate filaments around the nucleus of the neuron. The axo-glial membrane specialization occurring in the central nervous system of the zinc-treated adult rats has some similarities to the abnormal membrane relationships between naked axons and scarring astrocytes found in chronic experimental allergic encephalomyelitis lesions and multiple sclerosis plaques21,2L It has been suggested that the axo-glial membrane specializations of the desmosome-type noted in demyelinated axons probably play some supportive role for the naked axons and that these junctions help reorient and reconstruct the topography of affected fiber tracts subsequent to their recovery from the initial impact of myelin loss2L As Levine and Sowinski noted 17, the rapid development of perivascular infiltration after zinc implantation did not suggest an immunological origin, and they found that zinc had no adjuvant activity. However, both the axo-glial membrane specialization and the similarities of the infiltrates to the typical lesions of experimental allergic encephalomyelitis after zinc wire implantation are intriguing. Large mitochondria are found in a variety of tissues undel different physiologic and pathological conditions is. Zinc wire implants in the central nervous system induce changes in mitochondrial structure which are consistent with the effects of zinc ions on mitochondria in other tissues which are described in a review by Chvapil et al. 3. Zinc inhibits electron transport in the mitochondrial respiratory chain and the inhibitory effect of zinc ions on respiration has been localized between cytochromes b and cl. High levels of zinc ions also inhibit succinate dehydrogenase and cytochrome oxidase. The thick bundles of collagen around the zinc wire implant after 30-35 weeks may be due to trauma or to high levels of zinc in the central nervous system. In support of the latter possibility, zinc ions have been shown to affect collagen synthesis in a specific way compared to protein synthesis in general, and to have an effect on collagen cross-linking (see Ref. 6 for a review). ACKNOWLEDGEMENTS We thank Drs. Robert D. Terry and Dikran S. Horoupian for helpful discussions. The technical help of Ms. Clemenia Cayetano is gratefully acknowledged. This work was supported in part by Grants NS-12418 and AG-00002 from the National Institutes of Health.
REFERENCES 1 Brosnan, C., Kress, Y., Gaskin, F. and Levine, S., Further studies on the inflammatory response induced by zinc wire in the central nervous system of rats, manuscript in preparation. 2 Chvapil, M.. Effect of zinc on cells and membranes, Med. Clin. N. Amer., 60 (1976) 799-812.
149 3 Chvapil, M., Elias, S. L., Ryan, J. N. and Zukoski, C. F., Pathophysiology of zinc. In Int. Rev. Neurobiol., Suppl. l, Academic Press, Inc., New York, 1972, pp. 105-124. 4 Constantinidis, J., Richard, J. and Tissot, R., Maladie de Pick et Metabolisme du Zinc., Rev. Neurol., 133 (1977) 685-696. 5 Edstr6m, A. and Mattsson, H., Small amounts of zinc stimulate rapid axonal transport in vitro, Brain Research, 86 (1975) 162-167. 6 Fernandez-Madrid, F., Prasad, A. S. and Oberleas, D., Zinc in collagen metabolism. In A. S. Prasad (Ed.), Trace Elements in Human Health and Disease, Academic Press, New York, 1976, pp. 257-268. 7 Gaskin, F., In vitro microtubule assembly regulation by divalent cations and nucleotides, Biochem., 20C (1981) 1318-1322. 8 Gaskin, F., Gethner, J. S. and Kress, Y., The characterization of tubulin preparations using laser light scattering and tubulin assembly in the presence of zinc ions. In S. V. Perry, A. Margreth, and R. S. Adelstein (Eds.), Contractile Systems in Non-Muscle Tissues, North-Holland Biomedical Press, 1976, pp. 179-190. 9 Gaskin, F. and Kress, Y., Zinc ion-induced assembly of tubulin, J. biol. Chem., 19 (1977) 6918-6924. l0 Gaskin, F., Kress, Y., Brosnan, C. and Bornstein, M., Abnormal tubulin aggregates induced by zinc sulfate in organotypic cultures of nerve tissue, Neuroscience, 3 (1978) 1117-1128. I 1 Gaskin, F. and Shelanski, M. L., Microtubules and intermediate filaments, Essays in Biochemistry, 12 (1976) 115-146. 12 Haskins, K. M., Zombola, R. R., Boling, J. M., Lee, Y. C. and Himes, R. H., Tubulin assembly induced by cobalt and zinc, Biochem. Biophys. Res. Commun., 95 (1980) 1703-1709. 13 Krammer, E. B. and Zenkar, W., Effekt von Zinkionen auf Struktur und Verteilung der Neurotubuli, Acta Neuropath. (Berl.), 31 0975) 59-69. 14 Kress, Y., Gaskin, F., Brosnan, C. and Levine, S., Long term effects of zinc in the CNS of rats, J. Neuropath. exp. NeuroL, 39 (1980) 368. 15 Kress, Y., Gaskin, F., Horoupian, D. S. and Brosnan, C., Nickel induction of Rosenthal fibers in rat brain, Brain Research, 210 (198 l) 419-425. 16 Larsson, H., Wallin, M. and Edstr6m, A., Induction of a sheet polymer of tubulin by Zn 2+, Exp. Cell Res., 100 (1976) 104-110. 17 Levine, S. and Sowinski, R., Lymphocytic inflammation produced by intracerebral implantation of zinc and other metals, J. Neuropath. exp. Neurol., 37 (1978) 471-478. 18 Munn, E. A., The Structure of Mitochondria, Academic Press, New York, 1973, pp. 1--447. 19 Norton, W. T. and Goldman, J. E., Neurofilaments. In R. A. Bradshaw and D. M. Schneider (Eds.), Proteins of the Nervous System, Raven Press, New York, 1980, pp. 301-329. 20 Prasad, A. S. (Ed.), Trace Elements in Human Health and Disease. VoL 1: Zinc and Copper, Academic Press, New York, pp. 1-470. 21 Raine, C. S., Membrane specializations between demyelinated axons and astroglia in chronic EAE lesions and multiple sclerosis plaques, Nature, (Lond.), 275 (1978) 326-327. 22 Softer, D. and Raine, C. S., Morphological analysis of axoglial membrane specializations in the demyelinated central nervous system, Brain Research, 186 (1980) 301-313.