- Email: [email protected]
Proteins of slow axonal transport in sciatic motoneurones of rats with streptozotocin-induced diabetes or galactosaemia
Diabetes Research and Clinical Practice, 9 (1990)
15-21
15
Elsevier DIABET 00366
Proteins of slow axonal transport in sciatic motoneurones of rats with streptozotocin-induced diabetes or galactosaemia D.R. Tomlinson*,
G. Filliatreau ‘, B. Figliomeni3, R. Hassig ‘, L. Di Giamberardino G.B. Willars4
‘Service Hospitalier FrkdPric Joliot, DPpartement de Biologic, C.E.A., U.K., ‘FIDIA
Medical College of St. Bartholomew’s Hospital, London, 4Department of Pharmacology.
’
and
91406 Orsay, France, ‘Department of Pharmacology. Research Laboratories, 35031 Abano Terme, Italy and
London School of Pharmacy, London,
U.K.
(Received 2 August 1989) (Revision received 17 October 1989) (Accepted 24 October 1989)
Summary This study examined the distribution of axonally transported tubulin and a 68 kDa polypeptide in the sciatic nerve 34 days after injection of labelled methionine into the ventral horn of the spinal cord of control rats, rats with streptozotocin-induced diabetes mellitus and rats fed a diet containing 40% galactose. The proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) of pellets produced by treatment of nerve extracts with Triton X- 100 followed by differential ultra-centrifugation. The most marked effect of both diabetes and galactosaemia was to reduce the amount of activity present in tubulin transported at a rate of 1.4 to 2.1 mm/day. The distribution of activity in the 68 kDa polypeptide band was not markedly affected by either of the experimental conditions. These findings, taken together with those of other studies, indicate that the polyol pathway may contribute to the development of some defects of nerve function in diabetic rats, but is uninvolved in others. Key words: Axonal transport;
Galactosaemia;
Neurofilament;
Introduction The axon and terminals of neurones are dependent upon the supply of macromolecules synthesised in the perikaryon for their continued Correspondence to: Prof. D.R. Tom&on, Department of Pharmacology, Medical College of St. Bartholomew’s Hospital, London EClM 6BQ, U.K. 0168-8227/90/$03.50
0 1990 Elsevier Science Publishers
Tubulin
maintenance and function. Delivery of materials is achieved by processes collectively termed ‘anterograde axonal transport’. Defects in anterograde axonal transport, which have been reported in a number of neuropathies, including experimental diabetic neuropathy, precede structural damage to the neurone [l-3]. There has been progressive recognition that distal dwindling [4,5] and degeneration of the axon in diabetes may be related to
B.V. (Biomedical
Division)
16
some perturbations in the anterograde transport of the proteins comprising the endoskeletal matrix of the axon. Thus, impaired slow transport of incorporated amino acids has been reported in both sensory and motor neurones of rats with streptozotocin (STZ)-induced diabetes [ 6-81. More detailed investigation, using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), has shown that this defect in slow anterograde transport comprises aberrations in the distribution, along the nerve, of axonally transported tubulin and neurofilament proteins [ 9-111. Furthermore, similar defects have been seen in sciatic nerves of genetically diabetic mice [ 121. Two of these studies also demonstrated defects in the visual system and showed that the defects also involved slow component b (SCb) of anterograde transport; they also correlated the transport disorders with the development of distal axon al shrinkage [ 9,111. There is sound evidence that some of the axonal transport defects of experimental diabetes are related to biochemical consequences of disordered glucose metabolism, because they are prevented or attenuated by drugs which inhibit aldose reductase and reduce flux through the polyol pathway [ 13-151. Thus, a chain of events may be established linking hyperglycaemia”to structural degeneration via impaired axonal’transport. Exaggerated flux through the polyol pathway can also be achieved in the peripheral nerves by maintaining non-diabetic rats on a diet with a substantial galactose content [ 16,171. Galactose is metabolised by aldose reductase to the polyol dulcitol, which is not readily cleared from peripheral nerves and which, therefore, has more marked effects on neurochemistry than does the more modest accumulation of sorbitol in diabetes [ 16,181. The present study was therefore designed to compare the effects of STZ-induced diabetes with those of feeding a high galactose diet on the anterograde transport of radiolabelled polypeptides of slow component a (SCa) of anterograde transport.
Materials and methods Experimental organ&at&m The study used male Sprague Dawley rats of initial weight range 240-250 g. The first intervention was to administer [ 35S]methionine to label axonally transported proteins in sciatic motoneurones (see below for details). Three days later, the rats were divided into three groups. One served as untreated controls. The rats of a second group were made diabetic by a single i.p. injection of STZ (Sigma) at 60 mg/kg administered after an overnight fast. Both of these groups were given standard laboratory rat diet and water ad libitum. The third group were fed the same diet mixed with D-galactose (Sigma) in the proportions 60 : 40 by weight. All rats were killed 34 days after injection of methionine and the nerves removed for analysis of proteins in slow anterograde axonal transport by SDS-PAGE as described below. Contralateral nerves from control and galactose-fed rats were analysed for sugars and polyols by gas chromatography. Pulse Iabelling L-[35S]Methionine was obtained either from CEN de Saclay, Service des Molecules Marquees or from The Radiochemical Centre, Amersham. Batches were freeze-dried and re-dissolved in an isotonic Pod-buffered saline at pH 7.4 to give 100 ,nCi/pl. Under halothane anaesthesia two injections, each of 1 ~1 methionine, were made into the ventral horn of the spinal cord via a partial laminectomy at L,. The wound was closed with metal clips and the animal allowed to recover. Analysis of axonally transported proteins Rats were killed 34 days later by a blow to the head and bled from the throat (blood was collected for glucose assay - see below). Both sciatic nerves were dissected out rapidly from the dorsal root ganglia at L, and L, to the division at the knee into tibial and common peroneal branches. The distance, separating the injection site (below the L, vertebra) and the L, dorsal root ganglion, was measured on the cadaver to permit subse-
17
quent calculation of axonal transport velocities. The nerve was frozen under moderate tension on a metal plate cooled with liquid N, and cut into 6 mm segments; segment 1 comprising the L, ganglion and part of the L, branch distal to its ganglion. Each segment was homogenised in 150 ~1 of a buffer containing MgCl, 5 mM, 1,4piperazinediethanesulphonic acid (Pipes) 100 mM, EGTA 5 mM, Triton X-100 0.5%, glycerol 20%, phenylmethylsulfonyl fluoride 1 mM and dimethylsulphoxide 5 %. Homogenates were centrifuged (100,000 x g, 20 mm) in a compressed air-driven micro-ultracentrifuge (Beckman). The pellet so formed contained most of the proteins of SCa [ 191 and was dissolved in 40 ~1 of a Tris-buffered SDS solution containing fl-mercaptoethanol (see [ 191 for details); dissolution was assisted by boiling in a water bath for 3 min. The dissolved pellets were analysed by SDSPAGE using gradient gels (5.5 % to 15 % acrylamide) and run exactly as described elsewhere [ 191. The bands containing tubulin and the 68 kDa polypeptide (see Fig. 1) - assumed to the lightest of the neurofilament triplet - were cut from Coomassie Blue-stained dried gels, dissolved in 30% (w : v) hydrogen peroxide and the radioactivity measured by scintillation counting. The activity in each band in each segment of each nerve was calculated as a percentage of the total activity in that band in the whole nerve.
68kDa--T---c
Proximal
Distal
Fig. 1. Autoradiograph of an SDS-PAGE gel made from pellets of 6 mm sciatic nerve segments from a control rat. The positions of tubulin (T) and the 68 kDa polypeptide were determined from molecular weight standards.
Analysis of sugars and polyols
The content of sugars and polyols of the sciatic nerve contralateral to that used for axonal transport was estimated for control and galactosaemic rats by capillary gas chromatography of trimethylsilyl derivatives exactly as described elsewhere [20]. Blood glucose was estimated by a spectrophotometric glucose oxidase assay kit (GodPerid, Boehringer).
Results The control rats showed a mean body weight gain of about 130 g over the 5 week protocol (final body weight ? SD was 383 + 12 g). The galactosaemic rats gained rather less (final weight was 320 k 20 g). Whilst the diabetic rats showed no mean weight gain (final weight was 254 + 41 g). The diabetic rats were markedly hyperglycaemic with a terminal blood glucose of 28.5 f 3.5 (SD) mmol/l, compared to controls with a terminal blood glucose of 6.0 + 0.5 mmol/l (blood glucose was not measured in the galactosaemic rats). Nerve sugar and polyol contents are shown in Table 1. The presence of galactose in the nerves of the galactosaemic rats was associated with a marked accumulation of nerve dulcitol and a profound depletion of myo-inositol. Fig. 1 shows an autoradiograph made from a typical SDS-PAGE gel of the segment pellets from a control rat. The bands corresponding to tubulin and the 68 kDa polypeptide are clearly visible in the lanes from proximal segments. The greater distal migration of the activity attributed to tubulin is also clearly visible. The profiles showing distribution along the nerve of radioactivity in the tubulin and 68 kDa polypeptide bands of the nerves are shown in Fig. 2. The distribution of labelled tubulin along the nerve was substantially different in the control rats by comparison with either the diabetic or galactosaemic rats. This difference has been highlighted by stipple shading the relevant portions of the histograms in Fig. 2. There was much more activity in the tubulin travelling between 1.4 and
18 TABLE 1 Sciatic nerve content (nmol/mg dry nerve) of monosaccharides, galactose-fed groups.
Control rats (9)
polyol pathway metabolites
and myo-inositol in the control and
Glucose
Galactose
Fructose
Sorbitol
Dulcitol
5.19 + 1.35
-
0.55 + 0.04
0.21 + 0.08
-
myo-Inositol 9.01 + 0.63 P < 0.001
Galactose-fed
rats (8)
6.03 t 1.08
3.82 + 1.24
0.88 + 0.12
0.11 f 0.06
21.83 f 2.6
4.48 + 0.31
Data are means f SEM. Numbers of rats are in parentheses. A (-) indicates a component below the level of assay detection. Comparison of analogous means was made by unpaired t-test; pairs without P values were not significantly different.
TUBULIN
30
68
KD
2.1 mm/day in the nerves of the control rats than there was in the nerves of either of the other two groups. These differences were highly significant when tested non-parametrically (P < 0.001 by Mann-Whitney U test). There were no observable differences between diabetic and galactosaemic rats in the distribution of tubulin along the sciatic nerve. In contrast, there were no marked differences between any of the groups in the distribution of the 68 kDa polypeptide, though there appeared to be some retardation of the slowestmoving activity in the nerves of the diabetic rats when compared with control nerves. This retardation was not present in the nerves of the galactosaemic rats.
PROTEIN
-
2 : e f
20
b
kl+ = E P .u ‘0 ;;; ‘ij G m ::
CONTROL 10
RATS
(9)
O-
CALACTOSE FED
(8)
Discussion
5 : ._ 1 .o 5 z
.L 7 0
30
-
m (T ‘h r .> ;
20 DIABETIC 10
(7)
^
m OL I-
O Distance
24
(mm)
48
from
0
24
Lg
dorsal
48
root
ganglion
Fig. 2. The distribution ofradioactivity attributable to tubulin and the 68 kDa polypeptide along the nerves ofthe rats of the three groups. Data are means + SEM with the numbers of rats in parentheses. See Results for the significance of the shaded portions of the tubulin activity profiles.
This study has revealed differences between diabetic and control rats in the distribution of axonally transported radiolabelled polypeptides in sciatic motoneurones. Labelled tubulin showed the more marked difference, with a much reduced activity in the fraction moving ahead of the neurofilaments in the nerves of the diabetic rats. This finding is similar to that made by others [ 91. However, one difference in technique was that we used a cytoskeleton-stabilising buffer together with ultracentrifugation to obtain a clearer image of SCa [ 191. It therefore appears that at least a substantial fraction of the tubulin, which is normally associated with SCa, was absent from the
19
radioactivity profiles in the pellets from nerves of diabetic rats. In another previous study, which examined the distribution of incorporated tritiated leucine in slow transport, there was evidence of differences between diabetic and control rats in total incorporated radioactivity (i.e., not separated by SDSPAGE) in the same range of velocities [ 71. However, it was not possible to discern any effect of treatment of diabetic rats with either an aldose reductase inhibitor or myo-inositol. It was therefore concluded that the deficit in activity carried with the front of SCa in the nerves of diabetic rats had a pathogenesis unrelated to flux through the polyol pathway [ 71. In the present study we have deliberately induced exaggerated polyol pathway flux in non-diabetic rats by feeding a high galactose diet and produced an effect on the SCa transport of tubulin which was virtually indistinguishable from that of diabetes mellitus. This brings a paradox, in that it is super&ally difficult to rationalise an apparent lack of effect of an aldose reductase inhibitor in diabetic rats [ 71 with the generation of an axonal transport defect in the nerves of galactosaemic rats which was similar to that seen in diabetic rats. It is, however, dangerous to compare the picture seen with crude measurements of total incorporated tritiated leucine with the higher resolution offered by SDS-PAGE and an examination of the effects of an aldose reductase inhibitor in diabetic rats, using the techniques described in this study, would be valuable. Secondly, there are critical differences between the nerve of a diabetic and of a galactosaemic animal which render the state of alfairs in the latter more complex than was hitherto supposed [ 20,2 11. The accumulation of dulcitol in the nerves of galactosaemic rats is accompanied by a marked increase in nerve water content which does not occur in diabetes (see [20] for discussion of the reasons for this). This water is primarily sequestered in the endoneurial space, where it increases fluid pressure and causes marked vascular compression leading to endoneurial ischaemia [ 221. In the nerves of STZ-diabetic rats there is also a reduced endoneurial blood flow
with ischaemic hypoxia [23], but in this case the cause is unlikely to be vascular compression resulting from endoneurial oedema, since the necessary increase in nerve water does not occur in the diabetic rat [ 201 and endoneurial fluid pressure is not raised in diabetic rats whilst it is in galactosefed animals [24]. There must be other causes for the ischaemic hypoxia of the endoneurium in the diabetic nerve. The list of possibilities includes decreased calibre of endoneurial resistance vessels [ 231, thickening of basement membrane [ 251, increased blood viscosity [ 261 and decreased red cell deformability [26]. It is therefore probable that both the galactose-fed and the diabetic rats used in our study had endoneurial ischaemia, albeit from different causes. Thus, the mimicry of the diabetes-induced defect by galactosaemia could stem from the common factor of axonal hypoxia rather than from more direct neurochemical consequences of polyol pathway flux. The suggestion that endoneurial hypoxia might contribute to the altered appearance of axonally transported tubulin in both models studied here is not supported by a recent study on the direct effects of hypoxia. Slow anterograde transport of total methionine-labelled proteins was studied in motoneurones of rats with central hypoxia (maintained in a chamber containing S-10% 0,) for up to 10 weeks. No abnormalities were seen in the hypoxic rats [27]. However, this study is subject to the same considerations as those cited above - findings from a low resolution technique do not preclude the possibility that abnormalities might be revealed by the use of selective sedimentation and SDS-PAGE. Thus far we have not considered the nature of the defect in the distribution profile of axonally transported labelled tubulin. It is not entirely reasonable to assume that the disappearance of activity in the front of SCa reflects a simple retardation of axonal transport. The other possibilities are the selectively reduced amount or specific activity of the material leaving the cell bodies or an alteration in the way that the transported material is handled in the axon. The latter could involve loss of label or transfer from insoluble to
20
soluble compartments, perhaps due to non-enzymatic glycosylation or galactosylation [ 251. Clearly a more detailed study is warranted. When we began this study we expected to see a well-defined retardation of neurofilament proteins [6,7,9-l 11. However, there was no clear alteration of the profile representing the 68 kDa polypeptide in the nerves of the diabetic rats and its distribution in nerves of the galactosaemic rats was similar to that for controls. Careful scrutiny of Fig. 2 indicates a slight retardation of transport of the 68 kDa polypeptide, but one must admit that such careful scrutiny was prompted by the fact that the more thorough examination performed by others did find a retardation of the 68 kDa polypeptide in diabetic rats [9]. The retardation was not marked in studies from either laboratory. We may have missed changes in transport of the slowest-moving filament proteins because ventral roots proximal to L, were not collected for study. However, we have included all material travelling at velocities greater than 0.6 mm/day. It is, of course, possible that the neurofilament defect is more marked in sciatic sensory neurones than it is in motor fibres [ 61. Again a detailed comparison would be useful because the preponderance of sensory neuropathy over motor neuropathy in the clinical situation might arise from such a fundamental basis. It does appear that any defect in axonal transport of the 68 kDa polypeptide, that might have been present in the diabetic rat, was not mimicked by galactosaemia. This is in general agreement with a previous study, in which material travelling at the slowest velocity of SCa was retarded in diabetic rats and in which the slowing was unaffected by aldose reductase inhibition [ 71. In contrast the alteration to the faster tubulin band, present in both diabetic and galactosaemic rats, may indicate polyol-related flaws affecting other proteins - again as seen elsewhere [ 131. Thus, we have an indication of mixed aetiology in the pathogenesis of peripheral nerve disorders in experimental diabetes mellitus.
Acknowledgements This study was supported by grants from INSERM and The Wellcome Trust. We are grateful to Michelle Tartas for technical assistance.
References 1 Sidenius, P. (1982) The axonopathy of diabetic neuropathy. Diabetes 31, 356-363. 2 Tomlinson, D.R. and Mayer, J.H. (1984) Defects of axonal transport in diabetes mellitus - a possible contribution to the aetiology of diabetic neuropathy. J. Auton. Pharmacol. 4, 59-72. 3 Jakobsen, J., Sidenius, P. and Brrendgaard, H. (1986) A proposal for a classification of neuropathies according to their axonal transport abnormalities. J. Neurol. Neurosurg. Psychiatry 49, 986-990. 4 Jakobsen, J. (1976) Axonal dwindling in early experimental diabetes. I. A study of cross sectioned fibres. Diabetologia 12, 539-546. 5 Jakobsen, J. (1976) Axonal dwindling in early experimental diabetes. II. A study of isolated nerve fibres. Diabetologia 12, 547-553. 6 Sidenius, P. and Jakobsen, J. (1982) Reversibility and preventability of the decrease in slow axonal transport velocity in experimental diabetes. Diabetes 31,689-693. Mayer, J.H., Tomlinson, D.R. and McLean, W.G. (1984) Slow orthograde axonal transport of radiolabelled protein in sciatic motoneurones of rats with short-term experimental diabetes: effects of treatment with an aldose reductase inhibitor or myo-inositol. J. Neurochem. 43, 1265-1270. Tomlinson, D.R., Sidenius, P. and Larsen, J.R. (1986) Slow component-a of axonal transport, nerve myo-inositol, and aldose reductase inhibition in streptozotocindiabetic rats. Diabetes 35, 398-402. Medori, R., Autilio-Gambetti, L., Monaco, S. and Gambetti, P. (1985) Experimental diabetic neuropathy: impairment of slow transport with changes in axon crosssectional area. Proc. Natl. Acad. Sci. USA 82, 7716-1120. Medori, R., Autilio-Gambetti, L., Jenich, H. and Gambetti, P. (1988) Changes in axon size and slow axonal transport related in experimental diabetic are neuropathy. Neurology 38, 597-601. Medori, R., Jenich, H., Autilio-Gambetti, L. and Gambetti, P. (1988) Experimental diabetic neuropathy: similar changes of slow axonal transport and axonal size in different animal models. J. Neurosci. 8, 1814-1821. Vitadello, M., Filliatreau, G., DuPont, J.L., Hassig, R., Gorio, A. and Di-Giamberardino, L. (1985) Altered
21
13
14
15
16
17
18
19
axonal transport of cytoskeletal proteins in the mutant diabetic mouse. J. Neurochem. 45, 860-868. Mayer, J.H. and Tomlinson, D.R. (1983) Prevention of defects of axonal transport and nerve-conduction velocity by oral administration of myo-inositol or an aldose reducstreptozotocin-diabetic rats. tase inhibitor in Diabetologia 25, 433-438. Robinson, J.P., Willars, G.B., Tomlinson, D.R. and Keen, P. (1987) Axonal transport and tissue contents of substance P in rats with long-term streptozotocin-diabetes. Effects ofthe aldose reductase inhibitor ‘statil’. Brain Res. 426, 339-348. Tomlinson, D.R., Robinson, J.P., Willars, G.B. and Keen, P. (1988) Deficient axonal transport of substance P in streptozotocin-induced diabetic rats. Effects of sorbinil and insulin. Diabetes 37, 488-493. Stewart, M.A., Sherman, W.R., Kurien, M.M., Moonsammy, G.I. and Wisgerhof, M. (1967) Polyol accumulations in nervous tissue of rats with experimental diabetes and galactosaemia. J. Neurochem. 14, 1057-1066. Stewart, M.A., Sherman, W.R. and Harris, J.T. (1969) Effects of galactose on levels of free myo-inositol in rat tissues. Ann. N. Y. Acad. Sci. 165,609-614. Gabbay, K.H. and O’Sullivan, J.B. (1968) The sorbitol pathway in diabetes and galactosemia, enzyme and substrate localization and changes in kidney. Diabetes 17, 239-243. Filliatreau, G., Denoulet, P., De Nechaud, B. and Di Giamberardino, L. (1988) Stable and metastable cytoskeletal polymers carried by slow axonal transport. J. Neurosci. 8, 2227-2233.
20 Willars, G.B., Lambourne, J.E. and Tomlinson, D.R. (1987) Does galactose feeding provide a valid model of the consequences of exaggerated polyol-pathway flux in peripheral nerve in experimental diabetes? Diabetes 36, 1425-1431. 21 Lambourne, J.E., Tomlinson, D.R., Brown, A.M. and Willars, G.B. (1987) Opposite effects of diabetes and galactosaemia on ATPase activity in rat nervous tissue. Diabetologia 30, 360-362. 22 Myers, R.R. and Powell, H.C. (1984) Galactose neuropathy: impact of chronic endoneurial edema on nerve blood flow. Ann. Neurol. 16, 587-594. 23 Tuck, R.R., Schmelzer, J.D. and Low, P.A. (1984) Endoneurial blood flow and oxygen tension in the sciatic nerves of rats with experimental diabetic neuropathy. Brain 107,935-950. 24 Powell, H.C., Costello, M.L. and Myers, R.R. (1981) Endoneurial fluid pressure in experimental models of diabetic neuropathy. J. Neuropathol. Exp. Neurol. 40, 613-624. 25 Brownlee, M., Vlassara, H. and Cerami, A. (1984) Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann. Intern. Med. 101, 527-537. 26 Brownlee, M. and Cerami, A. (198 1) The biochemistry of the complications of diabetes mellitus. Annu. Rev. Biochem. 50, 385-432. 27 Nagata, H., Brimijoin, S., Low, P. and Schmelzer, J.D. (1987) Slow axonal transport in experimental hypoxia and in neuropathy induced by p-bromophenylacetylurea. Brain Res. 422, 319-326.
15-21
15
Elsevier DIABET 00366
Proteins of slow axonal transport in sciatic motoneurones of rats with streptozotocin-induced diabetes or galactosaemia D.R. Tomlinson*,
G. Filliatreau ‘, B. Figliomeni3, R. Hassig ‘, L. Di Giamberardino G.B. Willars4
‘Service Hospitalier FrkdPric Joliot, DPpartement de Biologic, C.E.A., U.K., ‘FIDIA
Medical College of St. Bartholomew’s Hospital, London, 4Department of Pharmacology.
’
and
91406 Orsay, France, ‘Department of Pharmacology. Research Laboratories, 35031 Abano Terme, Italy and
London School of Pharmacy, London,
U.K.
(Received 2 August 1989) (Revision received 17 October 1989) (Accepted 24 October 1989)
Summary This study examined the distribution of axonally transported tubulin and a 68 kDa polypeptide in the sciatic nerve 34 days after injection of labelled methionine into the ventral horn of the spinal cord of control rats, rats with streptozotocin-induced diabetes mellitus and rats fed a diet containing 40% galactose. The proteins were separated by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE) of pellets produced by treatment of nerve extracts with Triton X- 100 followed by differential ultra-centrifugation. The most marked effect of both diabetes and galactosaemia was to reduce the amount of activity present in tubulin transported at a rate of 1.4 to 2.1 mm/day. The distribution of activity in the 68 kDa polypeptide band was not markedly affected by either of the experimental conditions. These findings, taken together with those of other studies, indicate that the polyol pathway may contribute to the development of some defects of nerve function in diabetic rats, but is uninvolved in others. Key words: Axonal transport;
Galactosaemia;
Neurofilament;
Introduction The axon and terminals of neurones are dependent upon the supply of macromolecules synthesised in the perikaryon for their continued Correspondence to: Prof. D.R. Tom&on, Department of Pharmacology, Medical College of St. Bartholomew’s Hospital, London EClM 6BQ, U.K. 0168-8227/90/$03.50
0 1990 Elsevier Science Publishers
Tubulin
maintenance and function. Delivery of materials is achieved by processes collectively termed ‘anterograde axonal transport’. Defects in anterograde axonal transport, which have been reported in a number of neuropathies, including experimental diabetic neuropathy, precede structural damage to the neurone [l-3]. There has been progressive recognition that distal dwindling [4,5] and degeneration of the axon in diabetes may be related to
B.V. (Biomedical
Division)
16
some perturbations in the anterograde transport of the proteins comprising the endoskeletal matrix of the axon. Thus, impaired slow transport of incorporated amino acids has been reported in both sensory and motor neurones of rats with streptozotocin (STZ)-induced diabetes [ 6-81. More detailed investigation, using sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), has shown that this defect in slow anterograde transport comprises aberrations in the distribution, along the nerve, of axonally transported tubulin and neurofilament proteins [ 9-111. Furthermore, similar defects have been seen in sciatic nerves of genetically diabetic mice [ 121. Two of these studies also demonstrated defects in the visual system and showed that the defects also involved slow component b (SCb) of anterograde transport; they also correlated the transport disorders with the development of distal axon al shrinkage [ 9,111. There is sound evidence that some of the axonal transport defects of experimental diabetes are related to biochemical consequences of disordered glucose metabolism, because they are prevented or attenuated by drugs which inhibit aldose reductase and reduce flux through the polyol pathway [ 13-151. Thus, a chain of events may be established linking hyperglycaemia”to structural degeneration via impaired axonal’transport. Exaggerated flux through the polyol pathway can also be achieved in the peripheral nerves by maintaining non-diabetic rats on a diet with a substantial galactose content [ 16,171. Galactose is metabolised by aldose reductase to the polyol dulcitol, which is not readily cleared from peripheral nerves and which, therefore, has more marked effects on neurochemistry than does the more modest accumulation of sorbitol in diabetes [ 16,181. The present study was therefore designed to compare the effects of STZ-induced diabetes with those of feeding a high galactose diet on the anterograde transport of radiolabelled polypeptides of slow component a (SCa) of anterograde transport.
Materials and methods Experimental organ&at&m The study used male Sprague Dawley rats of initial weight range 240-250 g. The first intervention was to administer [ 35S]methionine to label axonally transported proteins in sciatic motoneurones (see below for details). Three days later, the rats were divided into three groups. One served as untreated controls. The rats of a second group were made diabetic by a single i.p. injection of STZ (Sigma) at 60 mg/kg administered after an overnight fast. Both of these groups were given standard laboratory rat diet and water ad libitum. The third group were fed the same diet mixed with D-galactose (Sigma) in the proportions 60 : 40 by weight. All rats were killed 34 days after injection of methionine and the nerves removed for analysis of proteins in slow anterograde axonal transport by SDS-PAGE as described below. Contralateral nerves from control and galactose-fed rats were analysed for sugars and polyols by gas chromatography. Pulse Iabelling L-[35S]Methionine was obtained either from CEN de Saclay, Service des Molecules Marquees or from The Radiochemical Centre, Amersham. Batches were freeze-dried and re-dissolved in an isotonic Pod-buffered saline at pH 7.4 to give 100 ,nCi/pl. Under halothane anaesthesia two injections, each of 1 ~1 methionine, were made into the ventral horn of the spinal cord via a partial laminectomy at L,. The wound was closed with metal clips and the animal allowed to recover. Analysis of axonally transported proteins Rats were killed 34 days later by a blow to the head and bled from the throat (blood was collected for glucose assay - see below). Both sciatic nerves were dissected out rapidly from the dorsal root ganglia at L, and L, to the division at the knee into tibial and common peroneal branches. The distance, separating the injection site (below the L, vertebra) and the L, dorsal root ganglion, was measured on the cadaver to permit subse-
17
quent calculation of axonal transport velocities. The nerve was frozen under moderate tension on a metal plate cooled with liquid N, and cut into 6 mm segments; segment 1 comprising the L, ganglion and part of the L, branch distal to its ganglion. Each segment was homogenised in 150 ~1 of a buffer containing MgCl, 5 mM, 1,4piperazinediethanesulphonic acid (Pipes) 100 mM, EGTA 5 mM, Triton X-100 0.5%, glycerol 20%, phenylmethylsulfonyl fluoride 1 mM and dimethylsulphoxide 5 %. Homogenates were centrifuged (100,000 x g, 20 mm) in a compressed air-driven micro-ultracentrifuge (Beckman). The pellet so formed contained most of the proteins of SCa [ 191 and was dissolved in 40 ~1 of a Tris-buffered SDS solution containing fl-mercaptoethanol (see [ 191 for details); dissolution was assisted by boiling in a water bath for 3 min. The dissolved pellets were analysed by SDSPAGE using gradient gels (5.5 % to 15 % acrylamide) and run exactly as described elsewhere [ 191. The bands containing tubulin and the 68 kDa polypeptide (see Fig. 1) - assumed to the lightest of the neurofilament triplet - were cut from Coomassie Blue-stained dried gels, dissolved in 30% (w : v) hydrogen peroxide and the radioactivity measured by scintillation counting. The activity in each band in each segment of each nerve was calculated as a percentage of the total activity in that band in the whole nerve.
68kDa--T---c
Proximal
Distal
Fig. 1. Autoradiograph of an SDS-PAGE gel made from pellets of 6 mm sciatic nerve segments from a control rat. The positions of tubulin (T) and the 68 kDa polypeptide were determined from molecular weight standards.
Analysis of sugars and polyols
The content of sugars and polyols of the sciatic nerve contralateral to that used for axonal transport was estimated for control and galactosaemic rats by capillary gas chromatography of trimethylsilyl derivatives exactly as described elsewhere [20]. Blood glucose was estimated by a spectrophotometric glucose oxidase assay kit (GodPerid, Boehringer).
Results The control rats showed a mean body weight gain of about 130 g over the 5 week protocol (final body weight ? SD was 383 + 12 g). The galactosaemic rats gained rather less (final weight was 320 k 20 g). Whilst the diabetic rats showed no mean weight gain (final weight was 254 + 41 g). The diabetic rats were markedly hyperglycaemic with a terminal blood glucose of 28.5 f 3.5 (SD) mmol/l, compared to controls with a terminal blood glucose of 6.0 + 0.5 mmol/l (blood glucose was not measured in the galactosaemic rats). Nerve sugar and polyol contents are shown in Table 1. The presence of galactose in the nerves of the galactosaemic rats was associated with a marked accumulation of nerve dulcitol and a profound depletion of myo-inositol. Fig. 1 shows an autoradiograph made from a typical SDS-PAGE gel of the segment pellets from a control rat. The bands corresponding to tubulin and the 68 kDa polypeptide are clearly visible in the lanes from proximal segments. The greater distal migration of the activity attributed to tubulin is also clearly visible. The profiles showing distribution along the nerve of radioactivity in the tubulin and 68 kDa polypeptide bands of the nerves are shown in Fig. 2. The distribution of labelled tubulin along the nerve was substantially different in the control rats by comparison with either the diabetic or galactosaemic rats. This difference has been highlighted by stipple shading the relevant portions of the histograms in Fig. 2. There was much more activity in the tubulin travelling between 1.4 and
18 TABLE 1 Sciatic nerve content (nmol/mg dry nerve) of monosaccharides, galactose-fed groups.
Control rats (9)
polyol pathway metabolites
and myo-inositol in the control and
Glucose
Galactose
Fructose
Sorbitol
Dulcitol
5.19 + 1.35
-
0.55 + 0.04
0.21 + 0.08
-
myo-Inositol 9.01 + 0.63 P < 0.001
Galactose-fed
rats (8)
6.03 t 1.08
3.82 + 1.24
0.88 + 0.12
0.11 f 0.06
21.83 f 2.6
4.48 + 0.31
Data are means f SEM. Numbers of rats are in parentheses. A (-) indicates a component below the level of assay detection. Comparison of analogous means was made by unpaired t-test; pairs without P values were not significantly different.
TUBULIN
30
68
KD
2.1 mm/day in the nerves of the control rats than there was in the nerves of either of the other two groups. These differences were highly significant when tested non-parametrically (P < 0.001 by Mann-Whitney U test). There were no observable differences between diabetic and galactosaemic rats in the distribution of tubulin along the sciatic nerve. In contrast, there were no marked differences between any of the groups in the distribution of the 68 kDa polypeptide, though there appeared to be some retardation of the slowestmoving activity in the nerves of the diabetic rats when compared with control nerves. This retardation was not present in the nerves of the galactosaemic rats.
PROTEIN
-
2 : e f
20
b
kl+ = E P .u ‘0 ;;; ‘ij G m ::
CONTROL 10
RATS
(9)
O-
CALACTOSE FED
(8)
Discussion
5 : ._ 1 .o 5 z
.L 7 0
30
-
m (T ‘h r .> ;
20 DIABETIC 10
(7)
^
m OL I-
O Distance
24
(mm)
48
from
0
24
Lg
dorsal
48
root
ganglion
Fig. 2. The distribution ofradioactivity attributable to tubulin and the 68 kDa polypeptide along the nerves ofthe rats of the three groups. Data are means + SEM with the numbers of rats in parentheses. See Results for the significance of the shaded portions of the tubulin activity profiles.
This study has revealed differences between diabetic and control rats in the distribution of axonally transported radiolabelled polypeptides in sciatic motoneurones. Labelled tubulin showed the more marked difference, with a much reduced activity in the fraction moving ahead of the neurofilaments in the nerves of the diabetic rats. This finding is similar to that made by others [ 91. However, one difference in technique was that we used a cytoskeleton-stabilising buffer together with ultracentrifugation to obtain a clearer image of SCa [ 191. It therefore appears that at least a substantial fraction of the tubulin, which is normally associated with SCa, was absent from the
19
radioactivity profiles in the pellets from nerves of diabetic rats. In another previous study, which examined the distribution of incorporated tritiated leucine in slow transport, there was evidence of differences between diabetic and control rats in total incorporated radioactivity (i.e., not separated by SDSPAGE) in the same range of velocities [ 71. However, it was not possible to discern any effect of treatment of diabetic rats with either an aldose reductase inhibitor or myo-inositol. It was therefore concluded that the deficit in activity carried with the front of SCa in the nerves of diabetic rats had a pathogenesis unrelated to flux through the polyol pathway [ 71. In the present study we have deliberately induced exaggerated polyol pathway flux in non-diabetic rats by feeding a high galactose diet and produced an effect on the SCa transport of tubulin which was virtually indistinguishable from that of diabetes mellitus. This brings a paradox, in that it is super&ally difficult to rationalise an apparent lack of effect of an aldose reductase inhibitor in diabetic rats [ 71 with the generation of an axonal transport defect in the nerves of galactosaemic rats which was similar to that seen in diabetic rats. It is, however, dangerous to compare the picture seen with crude measurements of total incorporated tritiated leucine with the higher resolution offered by SDS-PAGE and an examination of the effects of an aldose reductase inhibitor in diabetic rats, using the techniques described in this study, would be valuable. Secondly, there are critical differences between the nerve of a diabetic and of a galactosaemic animal which render the state of alfairs in the latter more complex than was hitherto supposed [ 20,2 11. The accumulation of dulcitol in the nerves of galactosaemic rats is accompanied by a marked increase in nerve water content which does not occur in diabetes (see [20] for discussion of the reasons for this). This water is primarily sequestered in the endoneurial space, where it increases fluid pressure and causes marked vascular compression leading to endoneurial ischaemia [ 221. In the nerves of STZ-diabetic rats there is also a reduced endoneurial blood flow
with ischaemic hypoxia [23], but in this case the cause is unlikely to be vascular compression resulting from endoneurial oedema, since the necessary increase in nerve water does not occur in the diabetic rat [ 201 and endoneurial fluid pressure is not raised in diabetic rats whilst it is in galactosefed animals [24]. There must be other causes for the ischaemic hypoxia of the endoneurium in the diabetic nerve. The list of possibilities includes decreased calibre of endoneurial resistance vessels [ 231, thickening of basement membrane [ 251, increased blood viscosity [ 261 and decreased red cell deformability [26]. It is therefore probable that both the galactose-fed and the diabetic rats used in our study had endoneurial ischaemia, albeit from different causes. Thus, the mimicry of the diabetes-induced defect by galactosaemia could stem from the common factor of axonal hypoxia rather than from more direct neurochemical consequences of polyol pathway flux. The suggestion that endoneurial hypoxia might contribute to the altered appearance of axonally transported tubulin in both models studied here is not supported by a recent study on the direct effects of hypoxia. Slow anterograde transport of total methionine-labelled proteins was studied in motoneurones of rats with central hypoxia (maintained in a chamber containing S-10% 0,) for up to 10 weeks. No abnormalities were seen in the hypoxic rats [27]. However, this study is subject to the same considerations as those cited above - findings from a low resolution technique do not preclude the possibility that abnormalities might be revealed by the use of selective sedimentation and SDS-PAGE. Thus far we have not considered the nature of the defect in the distribution profile of axonally transported labelled tubulin. It is not entirely reasonable to assume that the disappearance of activity in the front of SCa reflects a simple retardation of axonal transport. The other possibilities are the selectively reduced amount or specific activity of the material leaving the cell bodies or an alteration in the way that the transported material is handled in the axon. The latter could involve loss of label or transfer from insoluble to
20
soluble compartments, perhaps due to non-enzymatic glycosylation or galactosylation [ 251. Clearly a more detailed study is warranted. When we began this study we expected to see a well-defined retardation of neurofilament proteins [6,7,9-l 11. However, there was no clear alteration of the profile representing the 68 kDa polypeptide in the nerves of the diabetic rats and its distribution in nerves of the galactosaemic rats was similar to that for controls. Careful scrutiny of Fig. 2 indicates a slight retardation of transport of the 68 kDa polypeptide, but one must admit that such careful scrutiny was prompted by the fact that the more thorough examination performed by others did find a retardation of the 68 kDa polypeptide in diabetic rats [9]. The retardation was not marked in studies from either laboratory. We may have missed changes in transport of the slowest-moving filament proteins because ventral roots proximal to L, were not collected for study. However, we have included all material travelling at velocities greater than 0.6 mm/day. It is, of course, possible that the neurofilament defect is more marked in sciatic sensory neurones than it is in motor fibres [ 61. Again a detailed comparison would be useful because the preponderance of sensory neuropathy over motor neuropathy in the clinical situation might arise from such a fundamental basis. It does appear that any defect in axonal transport of the 68 kDa polypeptide, that might have been present in the diabetic rat, was not mimicked by galactosaemia. This is in general agreement with a previous study, in which material travelling at the slowest velocity of SCa was retarded in diabetic rats and in which the slowing was unaffected by aldose reductase inhibition [ 71. In contrast the alteration to the faster tubulin band, present in both diabetic and galactosaemic rats, may indicate polyol-related flaws affecting other proteins - again as seen elsewhere [ 131. Thus, we have an indication of mixed aetiology in the pathogenesis of peripheral nerve disorders in experimental diabetes mellitus.
Acknowledgements This study was supported by grants from INSERM and The Wellcome Trust. We are grateful to Michelle Tartas for technical assistance.
References 1 Sidenius, P. (1982) The axonopathy of diabetic neuropathy. Diabetes 31, 356-363. 2 Tomlinson, D.R. and Mayer, J.H. (1984) Defects of axonal transport in diabetes mellitus - a possible contribution to the aetiology of diabetic neuropathy. J. Auton. Pharmacol. 4, 59-72. 3 Jakobsen, J., Sidenius, P. and Brrendgaard, H. (1986) A proposal for a classification of neuropathies according to their axonal transport abnormalities. J. Neurol. Neurosurg. Psychiatry 49, 986-990. 4 Jakobsen, J. (1976) Axonal dwindling in early experimental diabetes. I. A study of cross sectioned fibres. Diabetologia 12, 539-546. 5 Jakobsen, J. (1976) Axonal dwindling in early experimental diabetes. II. A study of isolated nerve fibres. Diabetologia 12, 547-553. 6 Sidenius, P. and Jakobsen, J. (1982) Reversibility and preventability of the decrease in slow axonal transport velocity in experimental diabetes. Diabetes 31,689-693. Mayer, J.H., Tomlinson, D.R. and McLean, W.G. (1984) Slow orthograde axonal transport of radiolabelled protein in sciatic motoneurones of rats with short-term experimental diabetes: effects of treatment with an aldose reductase inhibitor or myo-inositol. J. Neurochem. 43, 1265-1270. Tomlinson, D.R., Sidenius, P. and Larsen, J.R. (1986) Slow component-a of axonal transport, nerve myo-inositol, and aldose reductase inhibition in streptozotocindiabetic rats. Diabetes 35, 398-402. Medori, R., Autilio-Gambetti, L., Monaco, S. and Gambetti, P. (1985) Experimental diabetic neuropathy: impairment of slow transport with changes in axon crosssectional area. Proc. Natl. Acad. Sci. USA 82, 7716-1120. Medori, R., Autilio-Gambetti, L., Jenich, H. and Gambetti, P. (1988) Changes in axon size and slow axonal transport related in experimental diabetic are neuropathy. Neurology 38, 597-601. Medori, R., Jenich, H., Autilio-Gambetti, L. and Gambetti, P. (1988) Experimental diabetic neuropathy: similar changes of slow axonal transport and axonal size in different animal models. J. Neurosci. 8, 1814-1821. Vitadello, M., Filliatreau, G., DuPont, J.L., Hassig, R., Gorio, A. and Di-Giamberardino, L. (1985) Altered
21
13
14
15
16
17
18
19
axonal transport of cytoskeletal proteins in the mutant diabetic mouse. J. Neurochem. 45, 860-868. Mayer, J.H. and Tomlinson, D.R. (1983) Prevention of defects of axonal transport and nerve-conduction velocity by oral administration of myo-inositol or an aldose reducstreptozotocin-diabetic rats. tase inhibitor in Diabetologia 25, 433-438. Robinson, J.P., Willars, G.B., Tomlinson, D.R. and Keen, P. (1987) Axonal transport and tissue contents of substance P in rats with long-term streptozotocin-diabetes. Effects ofthe aldose reductase inhibitor ‘statil’. Brain Res. 426, 339-348. Tomlinson, D.R., Robinson, J.P., Willars, G.B. and Keen, P. (1988) Deficient axonal transport of substance P in streptozotocin-induced diabetic rats. Effects of sorbinil and insulin. Diabetes 37, 488-493. Stewart, M.A., Sherman, W.R., Kurien, M.M., Moonsammy, G.I. and Wisgerhof, M. (1967) Polyol accumulations in nervous tissue of rats with experimental diabetes and galactosaemia. J. Neurochem. 14, 1057-1066. Stewart, M.A., Sherman, W.R. and Harris, J.T. (1969) Effects of galactose on levels of free myo-inositol in rat tissues. Ann. N. Y. Acad. Sci. 165,609-614. Gabbay, K.H. and O’Sullivan, J.B. (1968) The sorbitol pathway in diabetes and galactosemia, enzyme and substrate localization and changes in kidney. Diabetes 17, 239-243. Filliatreau, G., Denoulet, P., De Nechaud, B. and Di Giamberardino, L. (1988) Stable and metastable cytoskeletal polymers carried by slow axonal transport. J. Neurosci. 8, 2227-2233.
20 Willars, G.B., Lambourne, J.E. and Tomlinson, D.R. (1987) Does galactose feeding provide a valid model of the consequences of exaggerated polyol-pathway flux in peripheral nerve in experimental diabetes? Diabetes 36, 1425-1431. 21 Lambourne, J.E., Tomlinson, D.R., Brown, A.M. and Willars, G.B. (1987) Opposite effects of diabetes and galactosaemia on ATPase activity in rat nervous tissue. Diabetologia 30, 360-362. 22 Myers, R.R. and Powell, H.C. (1984) Galactose neuropathy: impact of chronic endoneurial edema on nerve blood flow. Ann. Neurol. 16, 587-594. 23 Tuck, R.R., Schmelzer, J.D. and Low, P.A. (1984) Endoneurial blood flow and oxygen tension in the sciatic nerves of rats with experimental diabetic neuropathy. Brain 107,935-950. 24 Powell, H.C., Costello, M.L. and Myers, R.R. (1981) Endoneurial fluid pressure in experimental models of diabetic neuropathy. J. Neuropathol. Exp. Neurol. 40, 613-624. 25 Brownlee, M., Vlassara, H. and Cerami, A. (1984) Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann. Intern. Med. 101, 527-537. 26 Brownlee, M. and Cerami, A. (198 1) The biochemistry of the complications of diabetes mellitus. Annu. Rev. Biochem. 50, 385-432. 27 Nagata, H., Brimijoin, S., Low, P. and Schmelzer, J.D. (1987) Slow axonal transport in experimental hypoxia and in neuropathy induced by p-bromophenylacetylurea. Brain Res. 422, 319-326.