Lack of apoptosis in mice with ALS

The role of apoptosis in neurodegenerative disorders remains controversial¹. Most of the uncertainty derives from the fact that (i) since these diseases progress over years, the possibility of finding a single dying cell in vivo is extremely remote, (ii) tissues are from end-stage patients and are therefore already depleted of neurons, (iii) even if present, death changes may be due to premortem agonal status rather than to the disease itself. Such problems may be avoided by using appropriate animal models. Severe loss of motor neurons characterizes amyotrophic lateral sclerosis (ALS), yet there is no conclusive evidence that apoptosis is involved². The etiology of sporadic ALS is unknown; in 20% of familial cases, point mutations in the superoxide dismutase 1 (SOD1) gene are present, and mutant SOD1 transgenic mice develop an ALS-like disease mediated by a toxic gain-of-function of the mutated enzyme³. Although mutant SOD1 induces apoptosis in vitro³, it is unclear whether the same happens in vivo. Overexpression of the antiapoptotic protein Bcl-2⁴ or inhibition of proapoptotic interleukin-1β-converting enzyme (ICE)⁵ modestly slow disease progression in mutant SOD1 mice. These effects, however, do not necessarily imply apoptosis regulation, since Bcl-2 might protect cells from SOD1 toxicity through its antioxidant properties, and ICE has many activities unrelated to apoptosis. Moreover, detailed ultrastructural studies in mutant SOD1 mice did not report on the presence of morphological signs of apoptosis, although they did show a wide array of ongoing degenerative changes, ranging from early vacuolar changes due to mitochondrial damage, to late formation of cytoplasmic inclusions in remaining motor neurons⁶.

To clarify whether degenerating motor neurons in mutant SOD1 mice undergo apoptosis, we used transgenic mice derived from Jackson Laboratory (Bar Harbor, Maine), that carry about 20 copies of human SOD1 gene with a glycine-to-alanine substitution at position 93 (G93A)⁷. These mice become symptomatic at around 70 days of age and die by 142±4.3 days. Sixteen transgenic and six wild-type mice combined in three age groups (9-14-19 weeks) were studied. Following perfusion fixation with 4% paraformaldehyde and paraffin embedding, the cervical and lumbar spinal cord were serially cut. Intraneuronal inclusions were immunohistochemically revealed with antibodies to human SOD1 (Sigma) and ubiquitin (Chemicon). To detect DNA fragmentation, a hallmark of apoptosis, we performed an in situ end-labeling (ISEL) assay (Boehringer)². Briefly, sections were pretreated with proteinase K (up to 50 μg/ml) and incubated with terminal deoxynucleotidyl transferase and fluorescein-11-dUTP, followed by peroxidase-conjugated anti-fluorescein sheep antisera. To evaluate the expression of apoptosis-related proteins, immunohistochemistry was performed with antibodies to c-Jun/AP-1 (Calbiochem) and PCNA (Dako), two proteins that accumulate in the nuclei of apoptotic neurons⁸, as well as with an antiserum that selectively recognizes the active form of apoptotic protease caspase-3 (MF397, kindly provided by Dr. Donald Nicholson). For each assay, at least eight sections/mouse were studied. Sections from the cerebellum of P7 mice carrying the weaver mutation, which causes massive apoptosis of granule cell precursors⁸, were stained as control.

A progressive loss of motor neurons was found in mutant SOD1 mice, amounting to 52±5% (S.E.) in the oldest mice with respect to age-matched controls. Neuronal loss was paralleled by the appearance of degenerative changes in the majority of motor neurons. Mild-to-severe vacuolation up to cell disintegration predominated in younger mice, but was occasionally found at all ages. In older mice, massive intraneuronal and intraaxonal conglomerates immunostained for SOD1 and ubiquitin were increasingly found. In spite of the prominent degeneration, none of the motor neurons showed nuclear changes (chromatin margination, pyknosis, karyorrhexis) suggestive of apoptosis. Although ISEL reveals nuclear DNA strand breaks well before the appearance of an apoptotic morphology, all motor neurons remained unlabeled, regardless of their degree of degeneration ( Fig. 1a,b). High proteinase K concentrations only produced increased background staining of all nuclei in the section (not shown). Absence of neuronal staining was also found with antibodies to c-Jun/AP-1 (Fig. 1c), PCNA and activated caspase-3. By contrast, labeled cells were readily detected with all assays in the cerebellum of weaver mice (Fig. 1d).

Figure 1: (a-c) Apoptotic changes are absent in mutant SOD1 mice.

(a,b)The ISEL assay fails to reveal DNA strand breaks in the nucleus (arrow) of degenerating motor neurons showing either (a) vacuolar changes (cresyl violet counterstain) or (b) intracytoplasmic conglomerates (double labeling ISEL/ubiquitin). Compare with the ISEL staining shown in Fig. 1d. (c) Increased c-Jun immunoreactivity is found in the nuclei of reactive astrocytes (arrow), but not in degenerating motor neurons. (d) Strong ISEL labeling is found in clusters of apoptotic granule cell precursors in weaver mouse cerebellum. All pictures taken at 1000×.

Although we cannot definitely rule out that motor neuron death in mutant SOD1 mice occurs via apoptosis, the present data do not support this hypothesis. Motor neurons degenerate either by massive vacuolation or by formation of intracytoplasmic conglomerates. Whereas the former change appears to induce motor neuron destruction by a necrotic-like mechanism, it is unclear whether cytoplasmic inclusions, that cause dysfunction to motor neurons⁹, are also responsible for their death. However, the apoptotic machinery is not apparently recruited in either case. These results should be kept in mind when devising therapeutic strategies aimed at controlling cell death in human ALS.