- Email: [email protected]
Alterations in levels of mRNAs coding for neurofilament protein subunits during regeneration
EXPERIMENTALNEUROLOGY
107,230-235 (1990)
Alterations in Levels of mRNAs Coding for Neurofilament Protein Subunits during Regeneration NANCY
A. MUMA,**~PAULN.HOFFMAN$$HILDA
IVANLIEBERBURG,IIANDDONALD
H. SLUNT,PMICHAEL L.PRIcE*~~$T#
D. APPLEGATE,*'?
Departments of *Pathology, $.Neurology, #Neuroscience, @phtkalmology, and tNeuropathology Laboratory, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205-2181; and “Athena Neuroscience, Inc., San Francisco, California
Animal models of neuronal injury can be used to explore mechanisms that regulate the expression of genes coding for cytoskeletal proteins and transmitter-related markers. In the present study, in situ hybridization was used to measure levels of messenger ribonucleic acid (mRNA) encoding each of the neurofilament subunits and &tubulin in spinal motor neurons at intervals (4 to 56 days) following a unilateral crush of the sciatic nerve. Levels of /3-tubulin mRNA increased (approximately twofold), peaked at 28 days postaxotomy, and returned to control values by 56 days postaxotomy. In contrast, levels of mRNA encoding neurofilament subunits were reduced and returned to control values at 56 days following the lesion. There were significant differences among relative levels of mRNAs coding for each subunit. Other studies have demonstrated that the ratio of pulse-labeled neurofilament subunits in motor axons remained unaltered during regeneration. Therefore, the ratios of neurofilament subunits in axons must be regulated at one of the steps that intervenes between the control of levels of mRNA and the anterograde axonal transport of assembled neurofilaments. 0 1990 Academic Press, Inc.
subunit could influence neurofilament structure that, in turn, could affect neurofilament phosphorylation, transport, and, ultimately, function. Animal models of neuronal injury are very useful for studying mechanisms that regulate the gene expression of cytoskeletal protein in neurons. For example, in one such model, i.e., axotomy, the synthesis of tubulin and actin is increased, whereas the synthesis of neurofllament proteins is diminished (5, 6, 16, 30, 38). Because levels of neurofilaments are decreased following axonal injury, this simple model can be used to ask questions concerning coordinate regulation of the gene expression of neurofilament subunits. In this study, using in situ hybridization, levels of mRNA coding for cytoskeletal proteins were measured in motor neurons of the L4-L.5 spinal cord at intervals of 4-56 days following axotomy of the sciatic nerve in rats. Our results indicate that these neurons show significant differences in the expression of genes coding for the three neurofilament subunits.
MATERIALS
AND METHODS
Axotomy INTRODUCTION Neurofilaments are composed of three protein subunits (17) and appear as lo-nm-diameter rods with periodic side arms (15,42). The central core of the neurofilament is enriched in the low molecular weight neurofilament subunit (NF-L), whereas side arms contain the high molecular weight neurofilament subunit (NF-H) (15). Each of the three subunits is coded by separate genes that are organized in a cluster on murine chromosome 8; in humans, NF-L and NF-M (middle molecular weight neurofilament subunit) genes are located on chromosome 8, whereas NF-H is found on chromosome 22 (19,20,25). An important issue in neurofilament biology is the regulation of genes coding for these proteins; differences in the regulation of genes coding for each 0014-4886/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
230 Inc. resewed.
Male Sprague-Dawley rats (n = 19; 7 weeks of age) were anesthetized with chloral hydrate (400 mg/kg) prior to crushing the sciatic nerve unilaterally at the junction of the L4 and L5 spinal nerves. At intervals following axotomy (4, 7, 14, 28, 42, or 56 days), rats were anesthetized and perfused transcardially with 0.9% saline in 0.1 M phosphate buffer (pH 7.6) followed by 4% paraformaldehyde. L4 and L5 levels of the spinal cord were removed, postfixed overnight, and embedded in paraffin.
In Situ Hybridization Using previously described methods (28), tissue sections (10 pm) were hybridized with 35S-labeled complementary dioxyribonucleic acid (cDNA) probes and then stained with cresyl violet. The cDNA probes were generous gifts from several investigators: NF-L was provided
mRNA
CODING
FOR
NEUROFILAMENT
DURING
REGENERATION
231
by Dr. Nicholas J. Cowan (24); NF-M (HNF4) was obtained from Dr. David L. Nelson (29); and ,&tubulin (RBTl) was provided by Dr. Stephen R. Farmer (3). The NF-H cDNA probe was cloned by one of the authors (25). These neurofilament probes uniquely recognize each subunit mRNA on Northern blots. To confirm the specificity of binding to RNA, tissue sections were treated with RNase A (60 min, 37°C) prior to hybridization or were hybridized with 35S-labeled PBR322 plasmid.
Image Analysis Tissue sections were analyzed using a computer-based system (Loats Associates, Inc., Westminster, MD). Neurons in spinal cord layer IX of the ventral horn (32) were evaluated for cross-sectional area, grains per neuron, and grain density. The unilateral nature of the lesion allowed for a direct comparison between axotomized neurons and contralateral control neurons in the same tissue section.
Data Analysis For each section of spinal cord analyzed, the mean number of grains per neuron and the mean grain density were calculated separately for the two ventral horns, and a value for the mean ratio of axotomized/control ventral horns was determined for each animal. An analysis of variance was computed with the mRNA species as a repeated measure. RESULTS
Controls In situ hybridization with probes for neurofilament subunits resulted in high grain densities associated with neurons but not with glial or ependymal cells (Fig. 1). With probes complementary to @-tubulin, grains were present over all cell types. For every probe, treatment with RNase A prior to hybridization eliminated clustering of grains over cells. Hybridization with a nonspecific DNA sequence, PBR322 plasmid, was associated with low background levels of grains and no specific labeling pattern. Axotomy Four days postaxotomy, levels of mRNA coding for each neurofilament subunit in motor neurons were reduced as compared to those in contralateral controls (Fig. 2). Levels of mRNA coding for NF-M and NF-H approached control levels by 56 days postaxotomy (Fig. 2). Measurements of mRNA levels can be expressed in two ways: the number of silver grains per neuron (Fig. 2A) and neuronal grain density (Fig. 2B). Significant differences occurred among ratios of axotomy/control
FIG. 1. In situ hybridization using a cDNA probe for NF-L mRNA. L4 and L5 levels of the spinal cord were removed from a rat 14 days after a unilateral crush of the sciatic nerve. A section of the spinal cord was hybridized with a 35S-labeled cDNA probe for NF-L mRNA and stained with cresyl violet. Grain densities in motor neurons on the axotomized side are reduced (upper) as compared to those in neurons in the contralateral control side of the spinal cord (lower). Bar = 8 pm.
mRNA levels (expressed as either grains per neurons or grain density) for the three neurofilament subunits. Levels of mRNAs coding for neurofilament subunits were
232
MUMA
ET AL.
8
g
0.50
x 5 Y
0.25
> f
0.00 7
14
21
28
35
42
49
I 56
f
0.00 i
I
0
7
14
DAY!3 ?0sT-AXOTOW
21
28
35
42
49
56
DAY5 POST-AXOTOW
FIG. 2. Levels of neurofilament mRNA expressed as the number of grains per neuron (A) and neuronal grain density (B). In situ hybridization was performed using cDNA probes for mRNA coding for the three neurofilament subunits (squares, NF-L, triangles, NF-M; and circles, NF-H) on spinal cord tissue from rats sacrificed at various time intervals after a unilateral crush of the sciatic nerve. Both the number of grains per neuron and the area of neuronal perikarya were calculated using a computer-based image analysis system. Mean values expressed are a ratio of the mean number of grains per neuron (A) or mean neuronal grain density (B) on the axotomized side divided by the contralateral control side of the spinal cord (2 standard error of the mean).
also significant across times postaxotomy. When these measures of mRNA levels were examined across postaxotomy times, the slopes of the curves of the number of grains per neuron for each neurofilament subunit appeared to be parallel, with the greatest reductions occurring in NF-L mRNA. Changes in neuronal grain density were also greatest for NF-L mRNA. However, it should be noted that slopes of the curves for the densities of each subunit mRNA were very different. Neuronal grain densities for NF-M mRNA were decreased maximally at 7 days postaxotomy, NF-L at 14 days postaxotomy, and NF-H at 28 days postaxotomy. These results cannot be explained on the basis of changes in cell size because there were no significant differences in perikaryal size between a-motor neurons on the axotomized side and the contralateral side of the spinal cord (N. A. Muma, personal observation). Thus, comparisons of levels of the three neurofilament subunit transcripts after axotomy demonstrate significant differences in the regulation of mRNA levels for each subunit. In contrast to the down-regulation of levels of neurofilament mRNA, the levels of mRNA coding for fl-tubulin were increased at 4 days postaxotomy (Fig. 3) and peaked at 28 days postaxotomy. These levels returned to control values by 42 days after crush of the sciatic nerve.
rons responded with reversible alterations in levels of mRNA coding for cytoskeletal proteins. The mRNA levels coding for P-tubulin increased postaxotomy and returned to control values prior to the return of neurofilament mRNA levels to control values. Although levels of mRNA coding for each of the three neurofilament subunits decreased, alterations were not identical for each of the subunits. NF-L mRNA showed a greater decline and remained at a lower level for a longer period than did mRNAs coding for NF-M and NF-H (Fig. 2). The finding of discordant regulation of genes coding for neurofilaments in motor neurons stands in contrast to results of studies of neurons of axotomized dorsal root 3.0-
2.5..
1.0+ 0
DISCUSSION
The regulation of gene expression of neuronal cytoskeletal proteins was examined in an animal model of neuronal injury (i.e., axotomy). Using in situ hybridization, levels of mRNAs coding for cytoskeletal proteins were measured at several time intervals after unilateral crush of the sciatic nerve. Axotomized spinal motor neu-
7
14
21
28
35
42
49
4 56
DAYS POST-AXOTOMY
FIG. 3.
Levels of 8-tubulin mRNA in control and axotomized motor neurons. Rats were sacrificed at various time intervals after unilateral crush of the sciatic nerve. Levels of /3-tubulin mRNA, determined by in situ hybridization, are expressed as the mean of the ratio of the mean number of grains per motor neuron (circles) and the mean neuronal grain densities (squares) for axotomized neurons divided by these values for neurons on the contralateral control side of the spinal cord (+ standard error of the mean).
mRNA
CODING
FOR
NEUROFILAMENT
ganglia in which mRNAs coding for the three neurofilament subunits appeared to be regulated coordinately (11). The lack of agreement in the results of these studies may be due to the types of neurons being examined, sensory vs motor. Precedent exists for the discordant regulation of the expression of the three neurofilament subunits. First, in the developing whole mouse brain, NF-M and NF-L mRNAs are detectable earlier than the mRNA coding for NF-H (20). Second, PC-12 cells treated with nerve growth factor (NGF) acquire a neuronal phenotype (12, 22,23,27) and show differential effects on the regulation of each of the neurofilament subunits. The NF-H subunit is altered at the post-transcriptional or translational level, NF-L and NF-M transcription rates are elevated, and there is an increase in the stability of the mRNA coding for the NF-M subunit (26). Although hybridization techniques used in these experiments are limited to measuring levels of mRNA and cannot determine whether there are alterations in the synthesis or degradation of mRNA, our findings indicate that the regulation of transcription or the control of the stability of mRNA, or both, are independent for the three neurofilament subunits. The ratio of pulse-labeled neurofilament subunits in the sciatic nerve is not altered during regeneration after axotomy (18). Therefore, the regulation of neurofilament subunit ratios in axons must be controlled at one or more of the steps that intervene between the control of mRNA levels and the transport of filaments. There are at least three possible levels of control. First, a fixed ratio of subunits could be maintained by regulating translation. Second, the ratio of subunits in the axon could be maintained by controlling the ratio of their assembly. Third, there may be differences in gating of the entry of neurofilaments into the anterograde axonal transport system (i.e., it is possible that only those filaments composed of an appropriate subunit ratio may be loaded onto the axonal transport system). Each of these alternatives would result in a different distribution of neurofilament subunits in the neuronal cell body. If the subunit ratios of neurofilaments in axons are controlled by translation, the subunit ratios within the cell bodies of axotomized neurons would be identical to those in the cell bodies of controls (and identical ratios in axons). If these ratios are controlled by assembly, then there would be proportionately more soluble NF-M and NF-H subunits in the cell bodies of regenerating neurons than in controls. Finally, if these ratios are controlled by gating, then the ratio of subunits in neurofilaments in the cell body would be different from that in the axon. Further studies are needed to determine the possible roles of these different mechanisms of regulation. Our finding provides one illustration of the ways in which animal models of cytoskeletal pathology can be
DURING
REGENERATION
233
used to elucidate mechanisms that control cytoskeletal gene expression. The response of cells to axotomy is quite different from the response of motor neurons to toxic injury. For example, aluminum toxicity results in decreased levels of neurofilament and /3-tubulin mRNAs (28), abnormal phosphorylation of neurofilaments in the cell bodies of neurons (2,4l), and accumulation of neurofilaments in cell bodies and proximal axons (4,40). Although the rate of neurofilament transport in the axon does not differ between control and intoxicated animals (1,39), aluminum-treated animals show fewer neurofilaments in axons (beyond the proximal swellings) and a concomitant atrophy of axons (39). Intoxication with &@‘-iminodipropionitrile (IDPN) causes a different type of disruption of the neuronal cytoskeleton. In this model, neurofilaments accumulate in and enormously distend the proximal axon but usually not the cell body (8,14). Slow axonal transport is impaired; beyond proximal swellings, the number of neurofilaments in axons is decreased, and axons become atrophic (13). However, IDPN exposure does not seem to affect levels of cytoskeletal mRNA in uiuo (31). On the basis of studies of these models, accumulations of neurofilaments in cell bodies (but not in proximal axons) appear to be associated with down-regulation of neurofilament mRNAs, whereas abnormalities of axonal neurofilaments or changes in transport of neurofilaments alone are not associated with altered levels of neurofilament mRNAs. Further studies in animal models, such as transgenic mice (with over- or underexpression of neurofilament subunits), may elucidate the mechanisms involved in neurofilament gene expression. Finally, perturbations of the neuronal cytoskeleton are important components of the neuronal pathology of several neurological diseases, such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (7,9,10,21,33-37). An understanding of the mechanisms that control gene expression and metabolic regulation of cytoskeletal proteins, including neurofilaments and tau, will help to form a foundation for interpreting abnormalities in the neuronal cytoskeleton in human neurological diseases. ACKNOWLEDGMENTS The authors thank Catherine A. in computerized image analysis and assistance. This work was supported Health Service (NIH ES 05398, NS Price and Muma are the recipients Alzheimer’s Disease (LEAD) Award recipient of a Javits Neuroscience 10580).
Fleischman for expert assistance Donald L. Price, Jr., for technical by grants from the U.S. Public 20164, NS 22849, NS 00896). Drs. of a Leadership and Excellence in (NIA AG 07914). Dr. Price is the Investigator Award (NIH NS
REFERENCES 1.
BIZZI, A., R. C. CRANE, L. AUTILIO-GAMBJZITI, AND P. GAMBET-IT. 1984. Aluminum effect on slow axonal transport: A novel impairment of neurofilament transport. J. Neurosci. 4: 722-731.
234
MUMA
2. BIZZI, A., AND P. GAMBE~I. 1986. Phosphorylation of neurofilaments is altered in aluminum intoxication. Actu Neuropathl. (Berlin) 71: 154-158. 3. BOND, J. F., G. S. ROBINSON, AND S. R. FARMER. 1984. Differential expression of two neural cell-specific fi-tuhulin mRNAs during rat brain development. Mol. Cell. Biol. 4: 1313-1319. 4. BUGIANI, O., AND B. GHETTI. 1982. Progressing encephalomyelopathy with muscular atrophy, induced by aluminum powder. Neurobiol. Aging 3: 209-222. 5. BURRELL, H. R., L. A. DOKAS, AND B. W. AGRANOFF. 1978. RNA metabolism in the goldfish retina during optic nerve regeneration. J. Neurochem. 3 1: 289-298. 6. BURRJXLL, H. R., A. M. HEACOCK, R. D. WATER, AND B. W. AGRANOFF. 1979. Increased tubulin messenger RNA in the goldfish retina during optic nerve regeneration. Brain Res. 168: 628632. 7. CARPENTER, S. 1968. Proximal axonal enlargement in motor neuron disease. Neurology 18: 841-851. 8. CLARK, A. W., J. W. GRIFFIN, AND D. L. PRICE. 1980. The axonal pathology in chronic IDPN intoxication. J. Neuropathol. Exp. Neurol. 39: 42-55. 9. CORK, L. C., N. H. STERNBERGER, L. A. STERNBERGER, M. F. CASANOVA, R. G. STRUBLE, AND D. L. PRICE. 1986. Phosphorylated neurofilament antigens in neurofibrillary tangles in Alzheimer’s disease. J. Neuropathol. Exp. Neurol. 46: 56-64. 10. GOLDMAN, J. E., AND S-.H. YEN. 1986. Cytoskeletal protein abnormalities in neurodegenerative diseases. Ann. Neurol. 19: 209-223. 11. GOLDSTEIN, M. E., S. R. WEISS, R. A. LAZZARINI, P. S. SHNEIDMAN, J. F. LEES, AND W. W. SCHLAEPFER. 1988. mRNA levels of all three neurofilament proteins decline following nerve transection. Mol. Brain Res. 3: 287-292. 12. GREENE, L. A., AND A. S. TISCHLER. 1982. PC12 pheochromocytoma cultures in neurobiological research. Adu. Cell. Neurobiol. 3:373-414. 13. GRIFFIN, J. W., P. N. HOFFMAN, A. W. CLARK, P. T. CARROLL, AND D. L. PRICE. 1978. Slow axonal transport of neurofilament proteins: Impairment by fl,fl’-iminodipropionitrile administration. Science 202: 633-635. 14. GRIFFIN, J. W., AND D. L. PRICE. 1980. Proximal axonopathies induced by toxic chemicals. Pages 161-178 in P. S. Spencer and H. H. Schaumburg, Eds., Experimental and Clinical Neurotoxicology. Williams & Wilkins, Baltimore. 15. HIROKAWA, N., M. A. GLICKSMAN, AND M. B. WILLARD. 1984. Organization of mammalian neurofilament polypeptides within the neuronal cytoskeleton. J. Cell Biol. 98: 1523-1536. 16. HOFFMAN, P. N., D. W. CLEVELAND, J. W. GRIFFIN, P. W. LANDES, N. J. COWAN, AND D. L. PRICE. 1987. Neurofilament gene expression: A major determinant of axonal caliber. hoc. Natl. Acad. Sci. USA 84: 3472-3476. 17. HOFFMAN, P. N., AND R. J. LASEK. 1975. The slow component of axonal transport. Identification of major structural polypeptides of the axon and their generality among mammalian neurons. J. Cell Biol. 66: 351-366. 18. HOFFMAN, P. N., G. W. THOMPSON, J. W. GRIFFIN, AND D. L. PRICE. 1985. Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J. Cell Biol. 101: 1332-1340. 19. JULIEN, J-.P., F. GROSVELD, K. YAZDANBAKSH, D. FLAVELL, D. MELJER, AND W. MUSHYNSKI. 1987. The structure of a human neurofilament gene (NF-L): A unique exon-intron organization in the intermediate filament gene family. Biochem. Biophys. Acta 909:10-20.
ET AL. 20. JULIEN, J-.P., D. MEYER, D. FLAVELL, J. HURST, AND F. GROSVELD. 1986. Cloning and developmental expression of the murine neurofilament gene family. Mol. Brain Res. 1: 243-250. 21. KOSIK, K. S. 1989. Minireview: The molecular and cellular pathology of Alzheimer neurofibrillary lesions. J. Gerontol. Biol. Sci. 44: B55-B58. 22. LEE, V. M-.Y., AND C. PAGE. 1984. The dynamics of nerve growth factor-induced neurofilament and vimentin filament expression and organization in PC12 cells. J. Neurosci. 4: 1705-1714. 23. LEE, V., J. L. WV, AND W. W. SCHLAEPFER. 1982. Induction of neurofilament triplet proteins in PC12 cells by nerve growth factor. Brain Res. 238: 169-180. 24. LEWIS, S. A., AND N. J. COWAN. 1985. Genetics, evolution, and expression of the 68,000-mol-w-t neurofilament protein: Isolation of a cloned cDNA probe. J. Cell Biol. 100: 843-850. 25. LIEBERBURG, I., N. SPINNER, S. SNYDER, J. ANDERSON, D. GOLDGABER, M. SMULOWITZ, Z. CARROLL, B. EMANUEL, J. BREITNER, AND L. RUBIN. 1989. Cloning of a cDNA encoding the rat high molecular weight neurofilament peptide (NF H): Developmental and tissue expression in the rat, and mapping of its human homologue to chromosomes 1 and 22. hoc. Natl. Acad. Sci. USA 86: 2463-2467. 26. LINDENBAUM, M. H., S. CARBONETTO, F. GROSVELD, D. FLAVELL, AND W. E. MUSHYNSKI. 1988. Transcriptional and posttranscriptional effects of nerve growth factor on expression of the three neurofilament subunits in PC-12 cells. J. Biol. Chem. 263:5662-5667. 27. LINDENBAUM, M. H., S. CARBONE~O, AND W. E. MUSHYNSKI. 1987. Nerve growth factor enhances the synthesis, phosphorylation, and metabolic stability of neurofilament proteins in PC12 cells. J. Biol. Chem. 262: 605-610. 28. MUMA, N. A., J. C. TFZONCOSO,P. N. HOFFMAN, E. H. Koo, AND D. L. PRICE. 1988. Aluminum neurotoxicity: Altered expression of cytoskeletal genes. Mol. Brain Res. 3: 115-122. 29. MYERS, M. W., R. A. LAZZARINI, V. M-.Y. LEE, W. W. SCHLAEPFER, AND D. L. NELSON. 1987. The human mid-size neurofilament subunit: A repeated protein sequence and the relationship of its gene to the intermediate filament gene family. EMBO J. 6: 1617-1626. 30. NEUMANN, D., T. SCHERSON,I. GINZBURG, U. Z. LITTAUER, AND M. SCHWARTZ. 1983. Regulation of mRNA levels for microtubule proteins during nerve regeneration. FEBS L&t. 162: 270-276. 31. PARHAD, I. M., E. A. SWEDBERG, D. A. HOAR, C. A. KREKOSIK, AND A. W. CLARK. 1988. Neurofilament gene expression following B&3’-iminodipropionitrile (IDPN) intoxication. Mol. Brain Res. 4: 293-301. 32. PAXINOS, G., AND C. WATSON. 1982. The Rat Brain in Stereotaxic Coordinates. Academic Press, Sydney. 33. PERRY, G., N. RIZZUTO, L. AUTILIO-GAMBETTI, AND P. GAMBETTI. 1985. Paired helical filaments from Alzheimer disease patients contain cytoskeletal components. Proc. N&l. Acad. Sci. USA 82: 3916-3920. 34. PRICE, D. L., R. J. ALTSCHULER, R. G. &RUBLE, M. F. CASANOVA, L. C. CORK, AND D. B. MURPHY. 1986. Sequestration of tubulin in neurons in Alzheimer’s disease. Bruin Res. 385: 305310. 35. PRICE, D. L. 1986. New perspectives on Alzheimer’s disease. Annu. Rev. Neurosci. 9: 489-512. 36. PRICE, D. L., J. C. TRONCOSO, P. J. WHITEHOUSE, J. W. GRIFFIN, P. N. HOFFMAN, AND L. C. CORK. 1986. Approaches to neurodegenerative diseases. Pages 1203-1216 in A. K. Asbury, G. M. McKhann, and W. I. McDonald, Eds., Clinical Neurobiology. Diseases of theNervous System, Vol. 2. Saunders, Philadelphia.
mRNA CODING
FOR NEUROFILAMENT
37. SCHMIDT, M. L., R. E. GIJR, R. C. GIJR, AND J. Q. TR~JANOWSKI. 1988. Intraneuronal and extracellular neurofibrillary tangles exhibit mutually exclusive cytoskeletal elements. Ann. NeuroL 23: 184-189. 38. TETZLAFF, W., M. A. BISBY, AND G. W. KREUTZBERG. 1988. Changes in cytoskeletal proteins in the rat facial nucleus following axotomy. J. Neurosci. 8: 3181-3189. 39. TRONCOSO, J. C., P. N. HOFFMAN, J. W. GRIFFIN, K. M. HESSKOZLOW, AND D. L. PRICE. 1985. Aluminum intoxication: A disorder of neurofilament transport in motor neurons. Brain Res. 342: 172-175.
DURING
REGENERATION
235
40. TRONCOSO, J. C., D. L. PRICE, J. W. GRIFFIN, AND I. M. PARHAD. 1982. Neurofibrillary axonal pathology in aluminum intoxication. Ann. Neural. 12: 273-283. 41. TRONCOSO, J. C., N. H. STERNBERGER, L. A. STERNBERGER, P. N. HOFFMAN, AND D. L. PRICE. 1986. Immunocytochemical studies of neurofilament antigens in the neurofibrillary pathology induced by aluminum. Bruin Res. 364: 295-300. 42. WILLARD, M., AND C. SIMON. 1981. Antibody decoration of neurofilaments. J. Cell Biol. 89: 198-205.
107,230-235 (1990)
Alterations in Levels of mRNAs Coding for Neurofilament Protein Subunits during Regeneration NANCY
A. MUMA,**~PAULN.HOFFMAN$$HILDA
IVANLIEBERBURG,IIANDDONALD
H. SLUNT,PMICHAEL L.PRIcE*~~$T#
D. APPLEGATE,*'?
Departments of *Pathology, $.Neurology, #Neuroscience, @phtkalmology, and tNeuropathology Laboratory, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205-2181; and “Athena Neuroscience, Inc., San Francisco, California
Animal models of neuronal injury can be used to explore mechanisms that regulate the expression of genes coding for cytoskeletal proteins and transmitter-related markers. In the present study, in situ hybridization was used to measure levels of messenger ribonucleic acid (mRNA) encoding each of the neurofilament subunits and &tubulin in spinal motor neurons at intervals (4 to 56 days) following a unilateral crush of the sciatic nerve. Levels of /3-tubulin mRNA increased (approximately twofold), peaked at 28 days postaxotomy, and returned to control values by 56 days postaxotomy. In contrast, levels of mRNA encoding neurofilament subunits were reduced and returned to control values at 56 days following the lesion. There were significant differences among relative levels of mRNAs coding for each subunit. Other studies have demonstrated that the ratio of pulse-labeled neurofilament subunits in motor axons remained unaltered during regeneration. Therefore, the ratios of neurofilament subunits in axons must be regulated at one of the steps that intervenes between the control of levels of mRNA and the anterograde axonal transport of assembled neurofilaments. 0 1990 Academic Press, Inc.
subunit could influence neurofilament structure that, in turn, could affect neurofilament phosphorylation, transport, and, ultimately, function. Animal models of neuronal injury are very useful for studying mechanisms that regulate the gene expression of cytoskeletal protein in neurons. For example, in one such model, i.e., axotomy, the synthesis of tubulin and actin is increased, whereas the synthesis of neurofllament proteins is diminished (5, 6, 16, 30, 38). Because levels of neurofilaments are decreased following axonal injury, this simple model can be used to ask questions concerning coordinate regulation of the gene expression of neurofilament subunits. In this study, using in situ hybridization, levels of mRNA coding for cytoskeletal proteins were measured in motor neurons of the L4-L.5 spinal cord at intervals of 4-56 days following axotomy of the sciatic nerve in rats. Our results indicate that these neurons show significant differences in the expression of genes coding for the three neurofilament subunits.
MATERIALS
AND METHODS
Axotomy INTRODUCTION Neurofilaments are composed of three protein subunits (17) and appear as lo-nm-diameter rods with periodic side arms (15,42). The central core of the neurofilament is enriched in the low molecular weight neurofilament subunit (NF-L), whereas side arms contain the high molecular weight neurofilament subunit (NF-H) (15). Each of the three subunits is coded by separate genes that are organized in a cluster on murine chromosome 8; in humans, NF-L and NF-M (middle molecular weight neurofilament subunit) genes are located on chromosome 8, whereas NF-H is found on chromosome 22 (19,20,25). An important issue in neurofilament biology is the regulation of genes coding for these proteins; differences in the regulation of genes coding for each 0014-4886/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form
230 Inc. resewed.
Male Sprague-Dawley rats (n = 19; 7 weeks of age) were anesthetized with chloral hydrate (400 mg/kg) prior to crushing the sciatic nerve unilaterally at the junction of the L4 and L5 spinal nerves. At intervals following axotomy (4, 7, 14, 28, 42, or 56 days), rats were anesthetized and perfused transcardially with 0.9% saline in 0.1 M phosphate buffer (pH 7.6) followed by 4% paraformaldehyde. L4 and L5 levels of the spinal cord were removed, postfixed overnight, and embedded in paraffin.
In Situ Hybridization Using previously described methods (28), tissue sections (10 pm) were hybridized with 35S-labeled complementary dioxyribonucleic acid (cDNA) probes and then stained with cresyl violet. The cDNA probes were generous gifts from several investigators: NF-L was provided
mRNA
CODING
FOR
NEUROFILAMENT
DURING
REGENERATION
231
by Dr. Nicholas J. Cowan (24); NF-M (HNF4) was obtained from Dr. David L. Nelson (29); and ,&tubulin (RBTl) was provided by Dr. Stephen R. Farmer (3). The NF-H cDNA probe was cloned by one of the authors (25). These neurofilament probes uniquely recognize each subunit mRNA on Northern blots. To confirm the specificity of binding to RNA, tissue sections were treated with RNase A (60 min, 37°C) prior to hybridization or were hybridized with 35S-labeled PBR322 plasmid.
Image Analysis Tissue sections were analyzed using a computer-based system (Loats Associates, Inc., Westminster, MD). Neurons in spinal cord layer IX of the ventral horn (32) were evaluated for cross-sectional area, grains per neuron, and grain density. The unilateral nature of the lesion allowed for a direct comparison between axotomized neurons and contralateral control neurons in the same tissue section.
Data Analysis For each section of spinal cord analyzed, the mean number of grains per neuron and the mean grain density were calculated separately for the two ventral horns, and a value for the mean ratio of axotomized/control ventral horns was determined for each animal. An analysis of variance was computed with the mRNA species as a repeated measure. RESULTS
Controls In situ hybridization with probes for neurofilament subunits resulted in high grain densities associated with neurons but not with glial or ependymal cells (Fig. 1). With probes complementary to @-tubulin, grains were present over all cell types. For every probe, treatment with RNase A prior to hybridization eliminated clustering of grains over cells. Hybridization with a nonspecific DNA sequence, PBR322 plasmid, was associated with low background levels of grains and no specific labeling pattern. Axotomy Four days postaxotomy, levels of mRNA coding for each neurofilament subunit in motor neurons were reduced as compared to those in contralateral controls (Fig. 2). Levels of mRNA coding for NF-M and NF-H approached control levels by 56 days postaxotomy (Fig. 2). Measurements of mRNA levels can be expressed in two ways: the number of silver grains per neuron (Fig. 2A) and neuronal grain density (Fig. 2B). Significant differences occurred among ratios of axotomy/control
FIG. 1. In situ hybridization using a cDNA probe for NF-L mRNA. L4 and L5 levels of the spinal cord were removed from a rat 14 days after a unilateral crush of the sciatic nerve. A section of the spinal cord was hybridized with a 35S-labeled cDNA probe for NF-L mRNA and stained with cresyl violet. Grain densities in motor neurons on the axotomized side are reduced (upper) as compared to those in neurons in the contralateral control side of the spinal cord (lower). Bar = 8 pm.
mRNA levels (expressed as either grains per neurons or grain density) for the three neurofilament subunits. Levels of mRNAs coding for neurofilament subunits were
232
MUMA
ET AL.
8
g
0.50
x 5 Y
0.25
> f
0.00 7
14
21
28
35
42
49
I 56
f
0.00 i
I
0
7
14
DAY!3 ?0sT-AXOTOW
21
28
35
42
49
56
DAY5 POST-AXOTOW
FIG. 2. Levels of neurofilament mRNA expressed as the number of grains per neuron (A) and neuronal grain density (B). In situ hybridization was performed using cDNA probes for mRNA coding for the three neurofilament subunits (squares, NF-L, triangles, NF-M; and circles, NF-H) on spinal cord tissue from rats sacrificed at various time intervals after a unilateral crush of the sciatic nerve. Both the number of grains per neuron and the area of neuronal perikarya were calculated using a computer-based image analysis system. Mean values expressed are a ratio of the mean number of grains per neuron (A) or mean neuronal grain density (B) on the axotomized side divided by the contralateral control side of the spinal cord (2 standard error of the mean).
also significant across times postaxotomy. When these measures of mRNA levels were examined across postaxotomy times, the slopes of the curves of the number of grains per neuron for each neurofilament subunit appeared to be parallel, with the greatest reductions occurring in NF-L mRNA. Changes in neuronal grain density were also greatest for NF-L mRNA. However, it should be noted that slopes of the curves for the densities of each subunit mRNA were very different. Neuronal grain densities for NF-M mRNA were decreased maximally at 7 days postaxotomy, NF-L at 14 days postaxotomy, and NF-H at 28 days postaxotomy. These results cannot be explained on the basis of changes in cell size because there were no significant differences in perikaryal size between a-motor neurons on the axotomized side and the contralateral side of the spinal cord (N. A. Muma, personal observation). Thus, comparisons of levels of the three neurofilament subunit transcripts after axotomy demonstrate significant differences in the regulation of mRNA levels for each subunit. In contrast to the down-regulation of levels of neurofilament mRNA, the levels of mRNA coding for fl-tubulin were increased at 4 days postaxotomy (Fig. 3) and peaked at 28 days postaxotomy. These levels returned to control values by 42 days after crush of the sciatic nerve.
rons responded with reversible alterations in levels of mRNA coding for cytoskeletal proteins. The mRNA levels coding for P-tubulin increased postaxotomy and returned to control values prior to the return of neurofilament mRNA levels to control values. Although levels of mRNA coding for each of the three neurofilament subunits decreased, alterations were not identical for each of the subunits. NF-L mRNA showed a greater decline and remained at a lower level for a longer period than did mRNAs coding for NF-M and NF-H (Fig. 2). The finding of discordant regulation of genes coding for neurofilaments in motor neurons stands in contrast to results of studies of neurons of axotomized dorsal root 3.0-
2.5..
1.0+ 0
DISCUSSION
The regulation of gene expression of neuronal cytoskeletal proteins was examined in an animal model of neuronal injury (i.e., axotomy). Using in situ hybridization, levels of mRNAs coding for cytoskeletal proteins were measured at several time intervals after unilateral crush of the sciatic nerve. Axotomized spinal motor neu-
7
14
21
28
35
42
49
4 56
DAYS POST-AXOTOMY
FIG. 3.
Levels of 8-tubulin mRNA in control and axotomized motor neurons. Rats were sacrificed at various time intervals after unilateral crush of the sciatic nerve. Levels of /3-tubulin mRNA, determined by in situ hybridization, are expressed as the mean of the ratio of the mean number of grains per motor neuron (circles) and the mean neuronal grain densities (squares) for axotomized neurons divided by these values for neurons on the contralateral control side of the spinal cord (+ standard error of the mean).
mRNA
CODING
FOR
NEUROFILAMENT
ganglia in which mRNAs coding for the three neurofilament subunits appeared to be regulated coordinately (11). The lack of agreement in the results of these studies may be due to the types of neurons being examined, sensory vs motor. Precedent exists for the discordant regulation of the expression of the three neurofilament subunits. First, in the developing whole mouse brain, NF-M and NF-L mRNAs are detectable earlier than the mRNA coding for NF-H (20). Second, PC-12 cells treated with nerve growth factor (NGF) acquire a neuronal phenotype (12, 22,23,27) and show differential effects on the regulation of each of the neurofilament subunits. The NF-H subunit is altered at the post-transcriptional or translational level, NF-L and NF-M transcription rates are elevated, and there is an increase in the stability of the mRNA coding for the NF-M subunit (26). Although hybridization techniques used in these experiments are limited to measuring levels of mRNA and cannot determine whether there are alterations in the synthesis or degradation of mRNA, our findings indicate that the regulation of transcription or the control of the stability of mRNA, or both, are independent for the three neurofilament subunits. The ratio of pulse-labeled neurofilament subunits in the sciatic nerve is not altered during regeneration after axotomy (18). Therefore, the regulation of neurofilament subunit ratios in axons must be controlled at one or more of the steps that intervene between the control of mRNA levels and the transport of filaments. There are at least three possible levels of control. First, a fixed ratio of subunits could be maintained by regulating translation. Second, the ratio of subunits in the axon could be maintained by controlling the ratio of their assembly. Third, there may be differences in gating of the entry of neurofilaments into the anterograde axonal transport system (i.e., it is possible that only those filaments composed of an appropriate subunit ratio may be loaded onto the axonal transport system). Each of these alternatives would result in a different distribution of neurofilament subunits in the neuronal cell body. If the subunit ratios of neurofilaments in axons are controlled by translation, the subunit ratios within the cell bodies of axotomized neurons would be identical to those in the cell bodies of controls (and identical ratios in axons). If these ratios are controlled by assembly, then there would be proportionately more soluble NF-M and NF-H subunits in the cell bodies of regenerating neurons than in controls. Finally, if these ratios are controlled by gating, then the ratio of subunits in neurofilaments in the cell body would be different from that in the axon. Further studies are needed to determine the possible roles of these different mechanisms of regulation. Our finding provides one illustration of the ways in which animal models of cytoskeletal pathology can be
DURING
REGENERATION
233
used to elucidate mechanisms that control cytoskeletal gene expression. The response of cells to axotomy is quite different from the response of motor neurons to toxic injury. For example, aluminum toxicity results in decreased levels of neurofilament and /3-tubulin mRNAs (28), abnormal phosphorylation of neurofilaments in the cell bodies of neurons (2,4l), and accumulation of neurofilaments in cell bodies and proximal axons (4,40). Although the rate of neurofilament transport in the axon does not differ between control and intoxicated animals (1,39), aluminum-treated animals show fewer neurofilaments in axons (beyond the proximal swellings) and a concomitant atrophy of axons (39). Intoxication with &@‘-iminodipropionitrile (IDPN) causes a different type of disruption of the neuronal cytoskeleton. In this model, neurofilaments accumulate in and enormously distend the proximal axon but usually not the cell body (8,14). Slow axonal transport is impaired; beyond proximal swellings, the number of neurofilaments in axons is decreased, and axons become atrophic (13). However, IDPN exposure does not seem to affect levels of cytoskeletal mRNA in uiuo (31). On the basis of studies of these models, accumulations of neurofilaments in cell bodies (but not in proximal axons) appear to be associated with down-regulation of neurofilament mRNAs, whereas abnormalities of axonal neurofilaments or changes in transport of neurofilaments alone are not associated with altered levels of neurofilament mRNAs. Further studies in animal models, such as transgenic mice (with over- or underexpression of neurofilament subunits), may elucidate the mechanisms involved in neurofilament gene expression. Finally, perturbations of the neuronal cytoskeleton are important components of the neuronal pathology of several neurological diseases, such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis (7,9,10,21,33-37). An understanding of the mechanisms that control gene expression and metabolic regulation of cytoskeletal proteins, including neurofilaments and tau, will help to form a foundation for interpreting abnormalities in the neuronal cytoskeleton in human neurological diseases. ACKNOWLEDGMENTS The authors thank Catherine A. in computerized image analysis and assistance. This work was supported Health Service (NIH ES 05398, NS Price and Muma are the recipients Alzheimer’s Disease (LEAD) Award recipient of a Javits Neuroscience 10580).
Fleischman for expert assistance Donald L. Price, Jr., for technical by grants from the U.S. Public 20164, NS 22849, NS 00896). Drs. of a Leadership and Excellence in (NIA AG 07914). Dr. Price is the Investigator Award (NIH NS
REFERENCES 1.
BIZZI, A., R. C. CRANE, L. AUTILIO-GAMBJZITI, AND P. GAMBET-IT. 1984. Aluminum effect on slow axonal transport: A novel impairment of neurofilament transport. J. Neurosci. 4: 722-731.
234
MUMA
2. BIZZI, A., AND P. GAMBE~I. 1986. Phosphorylation of neurofilaments is altered in aluminum intoxication. Actu Neuropathl. (Berlin) 71: 154-158. 3. BOND, J. F., G. S. ROBINSON, AND S. R. FARMER. 1984. Differential expression of two neural cell-specific fi-tuhulin mRNAs during rat brain development. Mol. Cell. Biol. 4: 1313-1319. 4. BUGIANI, O., AND B. GHETTI. 1982. Progressing encephalomyelopathy with muscular atrophy, induced by aluminum powder. Neurobiol. Aging 3: 209-222. 5. BURRELL, H. R., L. A. DOKAS, AND B. W. AGRANOFF. 1978. RNA metabolism in the goldfish retina during optic nerve regeneration. J. Neurochem. 3 1: 289-298. 6. BURRJXLL, H. R., A. M. HEACOCK, R. D. WATER, AND B. W. AGRANOFF. 1979. Increased tubulin messenger RNA in the goldfish retina during optic nerve regeneration. Brain Res. 168: 628632. 7. CARPENTER, S. 1968. Proximal axonal enlargement in motor neuron disease. Neurology 18: 841-851. 8. CLARK, A. W., J. W. GRIFFIN, AND D. L. PRICE. 1980. The axonal pathology in chronic IDPN intoxication. J. Neuropathol. Exp. Neurol. 39: 42-55. 9. CORK, L. C., N. H. STERNBERGER, L. A. STERNBERGER, M. F. CASANOVA, R. G. STRUBLE, AND D. L. PRICE. 1986. Phosphorylated neurofilament antigens in neurofibrillary tangles in Alzheimer’s disease. J. Neuropathol. Exp. Neurol. 46: 56-64. 10. GOLDMAN, J. E., AND S-.H. YEN. 1986. Cytoskeletal protein abnormalities in neurodegenerative diseases. Ann. Neurol. 19: 209-223. 11. GOLDSTEIN, M. E., S. R. WEISS, R. A. LAZZARINI, P. S. SHNEIDMAN, J. F. LEES, AND W. W. SCHLAEPFER. 1988. mRNA levels of all three neurofilament proteins decline following nerve transection. Mol. Brain Res. 3: 287-292. 12. GREENE, L. A., AND A. S. TISCHLER. 1982. PC12 pheochromocytoma cultures in neurobiological research. Adu. Cell. Neurobiol. 3:373-414. 13. GRIFFIN, J. W., P. N. HOFFMAN, A. W. CLARK, P. T. CARROLL, AND D. L. PRICE. 1978. Slow axonal transport of neurofilament proteins: Impairment by fl,fl’-iminodipropionitrile administration. Science 202: 633-635. 14. GRIFFIN, J. W., AND D. L. PRICE. 1980. Proximal axonopathies induced by toxic chemicals. Pages 161-178 in P. S. Spencer and H. H. Schaumburg, Eds., Experimental and Clinical Neurotoxicology. Williams & Wilkins, Baltimore. 15. HIROKAWA, N., M. A. GLICKSMAN, AND M. B. WILLARD. 1984. Organization of mammalian neurofilament polypeptides within the neuronal cytoskeleton. J. Cell Biol. 98: 1523-1536. 16. HOFFMAN, P. N., D. W. CLEVELAND, J. W. GRIFFIN, P. W. LANDES, N. J. COWAN, AND D. L. PRICE. 1987. Neurofilament gene expression: A major determinant of axonal caliber. hoc. Natl. Acad. Sci. USA 84: 3472-3476. 17. HOFFMAN, P. N., AND R. J. LASEK. 1975. The slow component of axonal transport. Identification of major structural polypeptides of the axon and their generality among mammalian neurons. J. Cell Biol. 66: 351-366. 18. HOFFMAN, P. N., G. W. THOMPSON, J. W. GRIFFIN, AND D. L. PRICE. 1985. Changes in neurofilament transport coincide temporally with alterations in the caliber of axons in regenerating motor fibers. J. Cell Biol. 101: 1332-1340. 19. JULIEN, J-.P., F. GROSVELD, K. YAZDANBAKSH, D. FLAVELL, D. MELJER, AND W. MUSHYNSKI. 1987. The structure of a human neurofilament gene (NF-L): A unique exon-intron organization in the intermediate filament gene family. Biochem. Biophys. Acta 909:10-20.
ET AL. 20. JULIEN, J-.P., D. MEYER, D. FLAVELL, J. HURST, AND F. GROSVELD. 1986. Cloning and developmental expression of the murine neurofilament gene family. Mol. Brain Res. 1: 243-250. 21. KOSIK, K. S. 1989. Minireview: The molecular and cellular pathology of Alzheimer neurofibrillary lesions. J. Gerontol. Biol. Sci. 44: B55-B58. 22. LEE, V. M-.Y., AND C. PAGE. 1984. The dynamics of nerve growth factor-induced neurofilament and vimentin filament expression and organization in PC12 cells. J. Neurosci. 4: 1705-1714. 23. LEE, V., J. L. WV, AND W. W. SCHLAEPFER. 1982. Induction of neurofilament triplet proteins in PC12 cells by nerve growth factor. Brain Res. 238: 169-180. 24. LEWIS, S. A., AND N. J. COWAN. 1985. Genetics, evolution, and expression of the 68,000-mol-w-t neurofilament protein: Isolation of a cloned cDNA probe. J. Cell Biol. 100: 843-850. 25. LIEBERBURG, I., N. SPINNER, S. SNYDER, J. ANDERSON, D. GOLDGABER, M. SMULOWITZ, Z. CARROLL, B. EMANUEL, J. BREITNER, AND L. RUBIN. 1989. Cloning of a cDNA encoding the rat high molecular weight neurofilament peptide (NF H): Developmental and tissue expression in the rat, and mapping of its human homologue to chromosomes 1 and 22. hoc. Natl. Acad. Sci. USA 86: 2463-2467. 26. LINDENBAUM, M. H., S. CARBONETTO, F. GROSVELD, D. FLAVELL, AND W. E. MUSHYNSKI. 1988. Transcriptional and posttranscriptional effects of nerve growth factor on expression of the three neurofilament subunits in PC-12 cells. J. Biol. Chem. 263:5662-5667. 27. LINDENBAUM, M. H., S. CARBONE~O, AND W. E. MUSHYNSKI. 1987. Nerve growth factor enhances the synthesis, phosphorylation, and metabolic stability of neurofilament proteins in PC12 cells. J. Biol. Chem. 262: 605-610. 28. MUMA, N. A., J. C. TFZONCOSO,P. N. HOFFMAN, E. H. Koo, AND D. L. PRICE. 1988. Aluminum neurotoxicity: Altered expression of cytoskeletal genes. Mol. Brain Res. 3: 115-122. 29. MYERS, M. W., R. A. LAZZARINI, V. M-.Y. LEE, W. W. SCHLAEPFER, AND D. L. NELSON. 1987. The human mid-size neurofilament subunit: A repeated protein sequence and the relationship of its gene to the intermediate filament gene family. EMBO J. 6: 1617-1626. 30. NEUMANN, D., T. SCHERSON,I. GINZBURG, U. Z. LITTAUER, AND M. SCHWARTZ. 1983. Regulation of mRNA levels for microtubule proteins during nerve regeneration. FEBS L&t. 162: 270-276. 31. PARHAD, I. M., E. A. SWEDBERG, D. A. HOAR, C. A. KREKOSIK, AND A. W. CLARK. 1988. Neurofilament gene expression following B&3’-iminodipropionitrile (IDPN) intoxication. Mol. Brain Res. 4: 293-301. 32. PAXINOS, G., AND C. WATSON. 1982. The Rat Brain in Stereotaxic Coordinates. Academic Press, Sydney. 33. PERRY, G., N. RIZZUTO, L. AUTILIO-GAMBETTI, AND P. GAMBETTI. 1985. Paired helical filaments from Alzheimer disease patients contain cytoskeletal components. Proc. N&l. Acad. Sci. USA 82: 3916-3920. 34. PRICE, D. L., R. J. ALTSCHULER, R. G. &RUBLE, M. F. CASANOVA, L. C. CORK, AND D. B. MURPHY. 1986. Sequestration of tubulin in neurons in Alzheimer’s disease. Bruin Res. 385: 305310. 35. PRICE, D. L. 1986. New perspectives on Alzheimer’s disease. Annu. Rev. Neurosci. 9: 489-512. 36. PRICE, D. L., J. C. TRONCOSO, P. J. WHITEHOUSE, J. W. GRIFFIN, P. N. HOFFMAN, AND L. C. CORK. 1986. Approaches to neurodegenerative diseases. Pages 1203-1216 in A. K. Asbury, G. M. McKhann, and W. I. McDonald, Eds., Clinical Neurobiology. Diseases of theNervous System, Vol. 2. Saunders, Philadelphia.
mRNA CODING
FOR NEUROFILAMENT
37. SCHMIDT, M. L., R. E. GIJR, R. C. GIJR, AND J. Q. TR~JANOWSKI. 1988. Intraneuronal and extracellular neurofibrillary tangles exhibit mutually exclusive cytoskeletal elements. Ann. NeuroL 23: 184-189. 38. TETZLAFF, W., M. A. BISBY, AND G. W. KREUTZBERG. 1988. Changes in cytoskeletal proteins in the rat facial nucleus following axotomy. J. Neurosci. 8: 3181-3189. 39. TRONCOSO, J. C., P. N. HOFFMAN, J. W. GRIFFIN, K. M. HESSKOZLOW, AND D. L. PRICE. 1985. Aluminum intoxication: A disorder of neurofilament transport in motor neurons. Brain Res. 342: 172-175.
DURING
REGENERATION
235
40. TRONCOSO, J. C., D. L. PRICE, J. W. GRIFFIN, AND I. M. PARHAD. 1982. Neurofibrillary axonal pathology in aluminum intoxication. Ann. Neural. 12: 273-283. 41. TRONCOSO, J. C., N. H. STERNBERGER, L. A. STERNBERGER, P. N. HOFFMAN, AND D. L. PRICE. 1986. Immunocytochemical studies of neurofilament antigens in the neurofibrillary pathology induced by aluminum. Bruin Res. 364: 295-300. 42. WILLARD, M., AND C. SIMON. 1981. Antibody decoration of neurofilaments. J. Cell Biol. 89: 198-205.