Original article
Ultrastructural changes in lumbar spinal cord in transgenic SOD1^G93A rats

Introduction
ALS, the most common adult motor neuron disease, is a progressive neurodegenerative disorder leading to paralysis and death. The disease is characterized by the selective degeneration and death of lower motor neurons in the brainstem and spinal cord leading to muscle weakness and atrophy. In 5-15% of cases of human ALS, mutations of the SOD1 gene were identified. Transgenic mice and rats expressing some of these mutations, for example G93A mutation, develop progressive motor neuron disease similar to human familial ALS [9,10,19].
Although numerous reviews have appeared during the last decade on neuronal death in mice bearing human mutated SOD1 transgene, the molecular mechanism by which mutations in the SOD1 gene lead to selective death of motoneurons in familial ALS remain incompletely understood. The precise mode of motoneuron death in SOD1 mutant mice is still controversial. Immunohistochemical evidence has been mounting that neuronal death is a result of apoptotic machinery [7,18,21,23], but the actual ultrastructural phenotype of motoneuron death remains undetermined.
In contrast to several different types of neurons found in the brain, spinal alpha motoneurons have been classified as fast (F), slow (S) and intermediate [25,26].
F motoneurons innervate the glycolytic, white muscle fibres and are considerably larger than S neurons innervating the oxidative, red muscle fibres [25]. S motoneurons exhibit greater NADH activity and are more abundant in mitochondria, while phosphorylase activity is greater in F motoneurons [26].
The purpose of this study was to define ultrastructural changes in motoneurons at the onset and during progression of ALS in rats expressing human mutated SOD1G93A transgene, in relation to motoneuron type.

Materials and methods
Sprague-Dawley rats expressing human mutated SOD1^G93A transgene and age-matched normal Sprague-Dawley controls were obtained from animal stocks bred in the Animal House of the Medical Research Institute of the Polish Academy of Sciences, as described in the preceding paper [10]. The animals were anaesthetized as described [10] and perfused with a solution of 4% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer pH 7.6. Tissues were kept in the same solution for further fixation. The L4 – L5 lumbar spinal cord was dissected out and postfixed with 2% osmium tetroxide in cacodylate buffer pH 7.6. After dehydration in graded alcohol, the tissue blocks were embedded in Spurr resin. Thick sections stained with methylene blue were examined with light microscopy. Thin sections stained uranyl acetate and lead citrate were examined with electron microscope JEM II.
Animals at different stages of the disease, i.e. presymptomatic (PS), age 60 and 93 days, and symptomatic displaying hind limb paralysis (S), age 120 days, were analyzed.

Results
In the control (normal) lumbar part of the spinal cord the main differences observed in the structure of the S motoneurons as compared with F motoneurons were the greater number of mitochondria, abundant rough endoplasmic reticulum and heterochromatic nuclei (Fig. 1 A, B). Motoneurons of F type had a lower number of mitochondria, more delicate and dispersed endoplasmic reticulum and larger, oval or round, euchromatic nuclei (Fig. 2 A, B).

In transgenic animals at the early PM stage (age 60 days) the ultrastructure of the lumbar spinal cord appeared normal; however, careful examination revealed that approximately 10-15% of the axons were filled with mitochondria that were abnormal in number, size and morphology (Fig. 3). Besides hypertrophic mitochondria we could distinguish giant mitochondria as well as mitochondria undergoing splitting (Fig. 4). At this stage in a few axons mitochondrial swelling was observed (Fig. 5). Such mitochondrial abnormalities were found in the proximal part of myelinated, as well as unmyelinated axons. It is worth noting that mitochondria were normal in both F and S motoneurons as well as in glial cells. At 93 days of age grossly swollen mitochondria with disrupted cristae were a prominent feature in all large axons (Fig. 6).
Dilated mitochondria appeared as vacuoles which were no longer carrying the remnants of cristae. Such vacuoles were so large that they occupied almost the entire axonal caliber (Fig. 7). Among vacuolated axons there were a few axons with dark appearance. They possessed a very dense, dark, compact, homogeneous interior (Fig. 8) with preserved dense mitochondria. At this late asymptomatic stage (age 93 days) swelling, disruption of the cristae and dilated mitochondria were observed in the S motoneurons (Fig. 9), while the F type cells had small, well preserved mitochondria.
At symptomatic stage there was still some degree of vacuolization of large axons, as well as an increase in number of axons with very dense, dark, homogeneous and compact material. The alpha motoneurons showed moderate neuronal loss, mainly of type S cells. The most interesting finding at this stage was the appearance of motoneurons with morphological signs resembling apoptotic death. Dying cells assumed fusiform shape, the nucleus showed uniformly dense, dark nucleoplasm, and cytoplasmic organelles became abnormally closely packed and condensed (Fig. 10, 11). Early (Fig. 12) and late apoptotic bodies, of various size, were observed in the neuropil as well as inside glial cells (Fig. 13). Apoptotic bodies were also observed inside the myelin tube (Fig. 14). F motoneurons in the symptomatic stage appeared normal, but careful examination revealed fragmentation of the Golgi complex and disaggregation of membrane-bound polyribosomes into free ribosomes in the cytoplasm, making the cytoplasmic matrix more granular (Fig. 15). In the smaller motoneurons apoptotic degeneration was not observed, although they showed numerous structural abnormalities. Their eccentrically located nuclei exhibited deep invaginations. The cytoplasm contained fragmented Golgi complex, autophagic vacuoles and filamentous inclusions (Fig. 16). Filamentous inclusions were also observed in the nuclei of some astrocytes (Fig. 17).

Discussion
Our ultrastructural study indicates that mitochondria in the lumbar spinal cord of transgenic rats bearing human mutated SOD1^G93A gene develop over the course of the disease major structural alterations, including variable degrees of swelling, vacuolization, cristae distortion and degeneration. Mitochondrial abnormalities appear as early as at the age of 60 days, preceding onset of the disease by two months. These findings are in keeping with previously presented data that in transgenic mice bearing human mutated SOD1 gene mitochondrial alterations in proximal axons are the first changes seen in the presymptomatic stage [11,13,14,24]. We also demonstrated for the first time that at the presymptomatic stage significant mitochondrial abnormalities appear in motoneurons of slow types while the fast motoneurons remain unaffected. These findings raise important questions: are the mitochondrial structural abnormalities a primary or secondary effect of SOD1 mutant protein, and why do these changes involve only a subset of axonal as well as slow motoneuronal mitochondria? The functional and morphological heterogeneity of mitochondria [3] may suggest that subpopulations of these organelles can carry out diverse processes within different motoneurons. The abundance of mitochondria as well as their larger size in slow motoneurons raise the possibility that they play a different role than those in fast motoneurons.

Mitochondrial dysfunction can lead to energy deficiency, ionic imbalance and oxidative damage and trigger the cell death program by releasing pro-apoptotic proteins residing in the mitochondrial interior [4,8,13,15,17]. Among the factors released from mitochondria are cytochrom C, the apoptotic inductor factor (AIF) and caspases [13,17]. The release of cytochrom C from the mitochondria to the cytosol plays a pivotal role in the apoptotic pathway that regulates the death of cells [15]. Accumulating evidence suggests that mitochondrial dysfunction causes motoneuron death in transgenic SOD1 mice by initiating the apoptotic pathway [8,20]. Consistent with massive mitochondrial degeneration seen in axons and slow motoneurons, the present study also shows that some axons at presymptomatic stage exhibit changes which mimic dark axonal degeneration. Such axons show very dark, dense axoplasm and dark mitochondria with preserved architecture which are remarkably similar to those seen during apoptosis. Recently, Alvarez et al. [1] have suggested that dark axonal degeneration may be viewed as a form of cytoplasmic apoptosis. They proposed that neurons have at least two self-destruction programs. Like other cell types, they have an intracellular death program for undergoing apoptosis when they are not needed or injured. In addition they apparently have a second molecular distinct self-destruction program in their axons. This program is activated when the axon is severed and leads to rapid degeneration of the isolated part of the cut axon [22]. This idea correlates with our findings at symptomatic stage in which moderate loss of motoneurons occurs in parallel with the appearance of apoptotic motoneuron death and self-destructed axons.
Dying motoneurons exhibit features reminiscent of apoptosis manifested by cytoplasmic and nuclear condensation and compaction with tightly packed organelles within a dark cytoplasmic matrix. The appearance of early and late apoptotic bodies in affected lumbar spinal cord provides additional evidence of motoneuron apoptosis.

Recently, accumulating evidence suggests that neuronal apoptosis plays a role in ALS [7,18,21,23]. However, although high expression of apoptosis-related proteins has been demonstrated in human mutated SOD1 transgenic mice [3], a surprisingly low fraction of motoneurons exhibit apoptotic morphology. This phenomenon may be related to the swiftness of cell death in relation to the slowness of disease. Our study indicates that in the so-called apoptotic stage the dying motoneuron is approximately one quarter of its normal diameter, the cytoplasm and nucleus are extremely condensed and the cell body adopts a fusiform shape. In addition early and late apoptotic bodies are more frequently observed than cells in the first stage of apoptosis. The nuclear condensation in affected rats differs from classic apoptosis because chromatin is not organized into uniformly dense clumps as in developmental neuronal apoptosis [2,5,6]. However, recently Leist and Jaattela [16] have classified cell death into four types: apoptosis, apoptosis-like programmed cell death, necrosis-like programmed cell death and accidental necrosis (cell lysis). Apoptosis-like programmed cell death is less compact than classical apoptosis. These authors have concluded that structural differences reflect different cell death execution machinery of mitochondrial dysfunction. Three different possible independent signals emanate from mitochondria: cytochrom C, AIF and reactive oxygen species [13]. Activated caspases triggered by cytochrom C induce classical apoptosis. When caspases cannot work, caspase-independent AIF stimulate an apoptosis-like program of cell death. Taken together, our findings suggest that dying motoneurons in transgenic rats bearing human mutated SOD1G93A gene exhibit reminiscent apoptotic morphology which is preceded by significant mitochondrial abnormalities, mainly in proximal axons and slow motoneurons. The different pattern of degeneration of slow and fast motoneurons requires further analysis.

References
1. Alvarez J, Giuditta A, Koenig E. Protein synthesis in axons and terminals: significance for maintenance, plasticity and regulation of phenotype. With a critique of slow transport theory. Prog Neurobiol 2000; 62: 1-62.
2. Clarke PG. Neuronal death in the development of the vertebrate central nervous system. Sem Neurosci 1994; 6: 291-297.
3. Collins TJ, Berridge MJ, Lipp P, Bootman MD. Mitochondria are morphologically and functionally heterogeneous within cells. EMBO J 2002; 21: 1616-1627.
4. Cregan SP, Fortin A, MacLaurin JG, Callaghan SM, Cecconi F, Yu SW, Dawson TM, Dawson VL, Park DS, Kroemer G, Slack RS. Apoptosis-inducing factor is involved in the regulation of caspase-independent neuronal cell death. J Cell Biol 2002; 158: 507-517.
5. Fidziańska A. Suicide muscle cell programme – apoptosis. Ultrastructural study. Folia Neuropathol 2002; 104: 363-368.
6. Fidziańska A, Rafałowska J. Motoneuron death in normal and spinal muscular atrophy – affected human fetuses. Acta Neuropathol 2002; 104: 363-368.
7. Guegan C, Przedborski S. Programmed cell death in amyotrophic lateral sclerosis. J Clin Invest 2003; 111: 153-161.
8. Guegan C, Vila M, Rosoklija G, Hays AP, Przedborski S. Recruitment of the mitochondrial-dependent apoptotic pathway in amyotrophic lateral sclerosis. J Neurosci 2001; 21: 6569-6576.
9. Gurney ME. Transgenic mouse of amyotrophic lateral sclerosis. N Engl J Med 1994; 331: 17.
10. Herbik M, Chrapusta SJ, Welniak-Kaminska M, Kowalczyk A, Grieb P. Maintenance of a rat transgenic model of familial amyotrophic lateral sclerosis expressing human SOD1G93A mutation. Folia Neuropathol 2006; 44: 149-153.
11. Higgins CM, Jung C, Xu Z. ALS – associated mutant SOD1G93A causes mitochondrial vacuolation by expansion of the intermembrane space and by involvement of SOD1 aggregation and proxisomes. BMC Neurosci 2003; 15: 4-16.
12. Jaarsma D, Haasdijk ED, Grashorn JA, Hawkins R, van Duijn W, Verspaget HW, London J, Holstege JC. Human Cu/Zn superoxide dismutase (SOD1) overexpression in mice causes mitochondrial vacuolization, axonal degeneration, and premature motoneuron death and accelerates motoneuron disease in mice expressing a familial amyotrophic lateral sclerosis mutant SOD1. Neurobiol Dis 2000; 7: 623-643.
13. Jordan J, Cena V, Prehn JH. Mitochondrial control of neuron death and its role in neurodegenerative disorders. J Physiol Biochem 2003; 59: 129-141.
14. Kong J, Xu Z. Massive mitochondrial degeneration in motor neurons tiggers the onset of amyotrophic lateral sclerosis in mice expressing a mutant SOD1. J Neurosci 1998; 18: 3241-3250.
15. Kroemer G, Reed JC. Mitochondrial control of cell death. Nat Med 2000; 6: 513-519.
16. Leist M, Jaattela M. Four deaths and a funeral: from caspases to alternative mechanisms. Nat Rev Mol Cell Biol 2001; 2: 589-598.
17. Manfredi G, Xu Z. Mitochondrial dysfunction and its role in motor neuron degeneration in ALS. Mitochondrion 2005; 5: 77-87.
18. Martin LJ. Neuronal death in amyotrophic lateral sclerosis is apoptosis: possible contribution of a programmed cell death mechanism. J Neuropathol Exp Neurol 1999; 58: 459-471.
19. Nagai M, Aoki M, Miyoshi I, Kato M, Pasinelli P, Kasai N, Brown RH Jr, Itoyama Y. Rats expressing human cytosolic copper-zinc superoxide dismutase transgenes with amyotrophic lateral sclerosis: associated mutations develop motor neuron disease. J Neurosci 2001; 21: 9246-9254.
20. Polster BM, Fiskum G. Mitochondrial mechanisms of neural cell apoptosis. J Neurochem 2004; 90: 1281-1289.
21. Przedborski S. Programmed cell death in amyotrophic lateral sclerosis: a mechanism of pathogenic and therapeutic importance. Neurologist 2004; 10: 1-7.
22. Raff MC, Whitmore AV, Finn JT. Axonal self-destruction and neurodegeneration. Science 2002; 296: 868-871.
23. Sathasivam S, Ince PG, Shaw PJ. Apoptosis in amyotrophic lateral sclerosis a review of the evidence. J Neuropathol Appl Neurobiol 2001; 27: 257-274.
24. Sasaki S, Warita H, Murakami T, Abe K, Iwata M. Ultrastructural study of mitochondria in the spinal cord of transgenic mice with a G93A mutant SOD1 gene. Acta Neuropathol (Berl) 2004; 107: 461-474.
25. Sato M, Mizuno N, Konishi A. Postnatal differentiation of cell body volumes of spinal cord motoneurons innervating slow – twitch and fast – twitch muscles. J Comp Neurol 1977; 175: 27-36.
26. van Asselt E, de Graaf F, van Raamsdonk W. Ultrastructural characteristics of zebrafish spinal motoneurons innervating glycolytic white, and oxidative red and intermediate muscle fibers. Acta Histochem 1993; 95: 31-44.