Brainstem Axonal Degeneration in Mice with Deletion of Selenoprotein P

Abstract

Selenoprotein P is an abundant extracellular protein that is expressed in liver, brain, and other tissues. Studies in mice with the selenoprotein P gene deleted (Sepp ^−/− mice) have implicated the protein in maintaining brain selenium. Sepp ^−/− mice fed a normal or low selenium diet develop severe motor impairment and die, but Sepp ^−/− mice fed a high selenium diet remain clinically unimpaired. As an initial step to evaluate the effect of selenoprotein P deletion on central nervous system architecture, the brains and cervical spinal cords of Sepp ^−/− and Sepp ^{+ / +} mice fed low or high selenium diets were examined by light and electron microscopy. Brains of Sepp ^−/− mice demonstrated no gross abnormalities. At the light microscopic level, however, Sepp ^−/− mice fed either the selenium deficient diet or the high selenium diet had enlarged dystrophic axons and degenerated axons in their brainstems and cervical spinal cords. No axonal lesions were observed in the Sepp ^{+ / +} mice fed either diet. Electron microscopy demonstrated that the enlarged axons in the Sepp ^−/− mice were packed with organelles, suggesting a deficit in fast axonal transport. The similar severity of axonal lesions observed in Sepp ^−/− mice fed the 2 diets suggests that axonal dystrophy is a common phenotype for deletion of selenoprotein P regardless of selenium intake and that additional studies will be required to determine the pathogenesis of the neurological signs and mortality observed in Sepp ^−/−mice fed a low selenium diet.

Introduction

Selenium is an essential micronutrient that exerts its functions as a constituent of selenoproteins (Kryukov et al., 2003). Most selenoproteins are redox enzymes and some of them, including the glutathione peroxidases and the thioredoxin reductases, have antioxidant functions that protect cells against injury by oxidative stress. Nutritional deficiency of selenium causes a decrease in selenium concentration and consequently decreased selenoprotein expression in most tissues. The brain, however, is resistant to depletion of its selenium. Even feeding mice a selenium deficient diet for several generations does not deplete the brain of selenium (Behne et al., 1988).

The ability of the brain to maintain a high selenium concentration relative to other tissues appears to be dependent on selenoprotein P, an extracellular selenoprotein that is expressed in virtually all tissues, including the brain (Saijoh et al., 1995; Yang et al., 2000). Mice with this protein deleted (Sepp ^−/− mice) have lower brain selenium concentrations than Sepp ^{+ / +} mice (Hill et al., 2003; Schomburg et al., 2003). Moreover, Sepp ^−/− mice develop neurological dysfunction characterized by motor abnormalities with severe spasticity unless they are fed a high selenium diet (>0.25 mg selenium per kg diet). Surprisingly, the high selenium diet that prevents the motor abnormalities does not raise whole brain selenium levels significantly (Hill et al., 2003, 2004). This indicates that the relationship of selenoprotein P with neurological dysfunction is complex.

The present study was performed to explore the potential contribution of central nervous system lesions to the clinical signs previously described for Sepp ^−/− mice. Brains of Sepp ^−/− and Sepp ^{+ / +} mice maintained on either a selenium deficient or a high selenium diet were surveyed by a phenotyping protocol assessing general physical dimensions and using several histological and histochemical stains. Based on the results obtained from these surveys, more detailed morphological assessments were performed on the brainstem using light and electron microscopy. The results support brainstem axonal injury to be a phenotypic change associated with deletion of selenoprotein P, although this injury may not be sufficient to account fully for the clinical signs and mortality associated with combined selenoprotein P deletion and dietary selenium deficiency.

Materials and Methods

Animals

Weanling Sepp ^{+ / +} and Sepp ^−/− male mice were back-crossed with C57Bl/6 mice (Hill et al., 2003, 2004). They were fed a Torula yeast-based diet that was selenium deficient (<0.01 mg selenium per kg) or that was supplemented with 0.5 or 1 mg selenium per kg as sodium selenite (Hill et al., 2004). The diet was manufactured and pelleted to our specifications by Harlan-Teklad (Madison, WI). The mice were housed in a room with a 12-hour:12-hour light:dark cycle and drank tap water ad libitum. Assessments of neurological function were performed as previously described (Hill et al., 2004). All experiments were performed in accordance with the “Principles of Laboratory Animal Care” (NIH publication No. 86-23, revised 1985) and were approved by the Vanderbilt Institutional Animal Care and Use Committee.

Brain Histological Phenotyping

Quantitative and qualitative measurements were made on 16 brains from mice that had been backcrossed once. Four brains each were harvested from Sepp ^{+ / +} and Sepp ^−/− mice fed the 1.0 ppm selenium diet for 14 days beginning at weaning (21 days of age), and from Sepp ^{+ / +} and Sepp ^−/− mice fed the selenium-deficient diet for 14 days beginning at weaning. The Sepp ^−/− mice fed the selenium-deficient diet had severe neurological dysfunction at the time the brains were harvested. All other mice appeared unaffected by observation.

Each mouse was anesthetized with isoflurane and then decapitated. The brain was removed from the skull, rinsed in saline, and frozen in isopentane at −78°C. Brains were stored at −80°C until they were processed for histology.

Ten different stains were used (see 〈http://neuropheno.typing.utmem.edu〉 or 〈http://www.tnmouse.org/protocols/histology.pdf〉). Histochemical and immunocytochemical methods were applied to evaluate cellular architecture (cresyl violet, and anti-NeuN (CHEMICON International, Inc., Temecula, CA), glia (cresyl violet and anti-glial fibrillary acidic protein (Immunon, Pittsburgh, PA), the cholinergic system (acetylcholinesterase), the catecholamin-ergic system (anti-tyrosine hydroxylase [Pel-Freez Clinical Systems, LLC, Brown Deer, WI]), myelin (osmium tetroxide and Luxol fast blue), proliferating cells (anti-bromodeoxyuridine and anti-phosphohistone H3 [Becton-Dickinson, Franklin Lakes, NJ]), mitochondrial function (cytochrome oxidase), and neuropathology (anti-ubiquitin [Dako Cytomation, Carpinteria, CA]).

Calbindin immunocytochemistry (CHEMICON International, Inc.) was done in paraffin embedded tissue to assess Purkinje cells. Three additional Sepp ^{+ / +} and 3 Sepp ^−/−mice were fed the selenium-deficient diet until the Sepp ^−/−mice manifested neurological signs. All mice were then anesthetized and perfused with saline followed by 3:1 95% EtOH:acetic acid, and postfixed in the same followed by rinsing in 70% ethanol and dehydration through xylenes. Tissue was embedded in paraffin and brain sections (8 μm) were cut on a microtome and mounted on slides and processed for anti-Calbindin D-28K immunocytochemistry (Chemicon AB1778, raised in rabbit; used at 1:250).

Sagittal sections at 400 μm intervals were taken throughout the medial to lateral extent of each brain. A 10-stain series was run at each interval (see 〈http://neurophenotyping.utmem.edu〉). The anterior-to-posterior extent of the neuraxis, from olfactory bulb to brain stem was examined comparing control to experimental material.

Sections in the sagittal plane were analyzed at the level where the hippocampal formation just becomes discernible as separate regions i.e., regio superior, regio inferior and area denata. At this level cerebellar circumference and area (measuring to the depth of each folia), cortex circumference, area, and thickness at 3 anterior to posterior points (defined by, in anterior to posterior progression, the midpoint of the underlying striatum, the fornix, and the posterior bend of the hippocampal pyramidal cells) were measured. Length and area of the corpus callosum was measured at this level. Quantitative measurements were accomplished using Soft Imaging System AnalySIS Opti 3.0 on a Zeiss scope at a 35X magnification.

Perfusion Fixation for Light and Electron Microscopic Evaluation

Sepp ^{+ / +} and Sepp ^−/− (n = 3 per group and time point) were perfused at 4, 8 and 12 days after weaning and start of the selenium-deficient diet and at 12 days after weaning and start of the 0.5 ppm selenium diet. The mice were anesthetized deeply with pentobarbital (120 mg/kg given in-traperitoneally) prior to whole body perfusion. The perfusion was established through the left ventricle and the whole body was perfused with 20 ml of wash solution containing 10 mM Na phosphate, pH 7.4, 0.14 M NaCl, 3.3 mM KCl, 6 mM NaHCO₃. The wash was followed by perfusion with 70 ml of fixative solution containing 4% glutaraldehyde in 0.1 M Na phosphate, pH 7.4. Whole brains and spinal cord sections were removed, postfixed in 4% glutaraldehyde in the 0.1 M Na phosphate buffer overnight and then stored in 0.1 M Na phosphate buffer. Embedding and staining was performed as described previously (Johnson et al., 1998). Thick (1 μm) coronal sections of the pons and transverse sections of cervical spinal cord at the C₁-C₂ level were evaluated by light microscopy using an Olympus BX41 microscope equipped with an Olympus C-3040 Camedia digital camera. Thin (70 nm) sections were prepared from the pons and evaluated using a Phillips CM-12, 120 keV electron microscope with a high-resolution camera system.

Semiquantitative measurements of degenerative axonal changes were performed blinded using a single high field (40X) region of the pons sectioned at the same level for each animal. Scores for individual mice were classified as normal (0–10 affected axons/field), mild (11–40 affected axons/field), moderate (41–80 affected axons/field) or severe (>80 affected axons/field). Enlarged dystrophic axons were quantified as the mean value derived from the total number of dystrophic axons on 1 section of pons from each animal.

Brain F₂ Isoprostane Measurement

Male Sepp ^{+ / +} and Sepp ^−/− mice of backcross 7 were fed the selenium-deficient diet from weaning. After 3, 6, 9, 12, and 15 days mice were anesthetized with isoflurane and exsanguinated by withdrawal of blood from the vena cava (n = 2–3 for days 3, 9, and 15; n = 5 and 9 for day 6; and n = 1 and 3 for day 12). Brains were removed, frozen in liquid nitrogen, and stored at −80°C until used for measurement of F₂ isoprostanes (Patton and Robbins, 1990; Morrow et al., 1992).

Statistical Analysis

Because no differences were obtained for either lesion for any of the Sepp ^{+ / +} mice groups, the Sepp ^{+ / +} values for each category were pooled and used to make comparisons to Sepp ^−/− groups. Statistical comparisons for nonparametric data obtained for axonal degeneration were performed using the Kruskal–Wallis test and Dunn’s post hoc test. Statistical differences among mean values of enlarged dystrophic axons were done using one-way ANOVA and comparisons between Sepp ^{+ / +} pooled values and the individual Sepp ^−/−groups done using Dunnett’s post hoc test. Statistical comparisons of F₂ isoprostane levels were performed using the Student’s t-test.

Results

Brain Histological Phenotyping

Weanling Sepp ^−/− and Sepp ^{+ / +} mice were fed selenium-deficient or high selenium (1 mg selenium per kg) diet for 14 days, by which time the Sepp ^−/− mice fed the selenium deficient diet had become severely impaired neurologically (Hill et al., 2004). The Sepp ^{+ / +} mice were unimpaired and the Sepp ^−/−mice fed the high selenium diet had no observable impairment. Forebrain, midbrain, and hindbrain were examined using the 10 different stains listed in Materials and Methods. No quantitative differences in cortical thickness or length or in the length or thickness of the corpus callosum were found (data not shown). No qualitative differences were found in the forebrain or midbrain in any of the stained material that was examined. However, in the hindbrain, there was decreased staining of myelin bundles in the brainstem of Sepp ^−/−, whether the myelin was visualized with osmium or Luxol fast blue stain (not shown). These findings led to a more detailed examination of the brain stem.

Brain Stem and Spinal Cord Morphology in Affected Mice

Brain stems and spinal cords of Sepp ^−/− mice exhibited axonal degeneration and enlarged dystrophic axons in light microscopic examinations, but Sepp ^{+ / +} mice had no such findings (Figure 1). Within the brain-stem, the injury appeared to be limited to the axons with no structural changes observed in neuron bodies or glia (Figure 1B). Similarly, in the spinal cord sections, axonal involvement predominated and was most pronounced in the dorsal sensory fibers and ventrolateral tracts (Figure 1, panels D and F) but was not observed in the more proximal axonal regions within the ventral horns.

Ultrastructural evaluation (Figure 2) revealed the presence of axons in early and late states of degeneration as evidenced by the presence of both electron lucent and electron dense axoplasm, respectively (Figure 2, panels B and D). Large darkened profiles were seen in the spinal cord and brain stem by light microscopy (Figure 1, panels B and F); and they were identified by electron microscopy to be enlarged dystrophic axons exhibiting thinned myelin with an accumulation of dense bodies and organelles (Figure 2E), many of which appeared to be degenerating mitochondria that were enlarged with irregularly shaped cristae (Figure 2F).

Both axonal degeneration and enlarged dystrophic axons were observed in all of the Sepp ^−/− mice fed the selenium-deficient diet and in all but 1 of the Sepp ^−/− mice fed the diet containing 0.5 mg selenium/kg diet (Table 1). Statistically significant increases in lesions were observed in Sepp ^−/−mice relative to Sepp ^{+ / +} mice on day 12 (enlarged dystrophic axons (p < 0.01) and axonal degeneration (p < 0.01)) and day 4 (axonal degeneration (p < 0.05)). Greater severity scores were obtained for axonal degeneration and increased numbers of enlarged dystrophic axons were observed for the Sepp ^−/− mice fed both diets at the 12-day time point relative to the 4 and 8 day time points examined for the Sepp ^−/− mice fed the selenium-deficient diet.

Assessment of Lipid Peroxidation in the Brain

Because selenium is an antioxidant nutrient, evidence of lipid peroxidation was sought in the brains of affected mice as a potential contributing event to the observed axonal lesions. Mean F₂ isoprostane values ranged from 4.9–7.5 ng/g brain and 4.4–6.7 ng/g brain for the Sepp ^{+ / +} and Sepp ^−/− mice, respectively (Figure 3). Individual Sepp ^−/− mice developed severe neurological dysfunction within days of starting to eat the selenium deficient diet and deaths began to occur at 6 days. Sepp ^{+ / +} mice were not significantly different at days 6, 9, and 15 by Student’s t-test. Too few mice were available at days 3 and 12 for statistical testing.

Discussion

Gross features of the brain were not affected by deletion of selenoprotein P. However, there was evidence of individual cell injury in the central nervous systems of Sepp ^−/−mice. Striking axonal degeneration was present in their brainstems and spinal cords; no such injury was found in Sepp ^{+ / +} mice (Table 1). Histopathological changes identified within the brainstem and spinal cord consisted of enlarged dystrophic axons and degenerating axons. The observed lesions are somewhat nonspecific and are similar in character to those reported for ischemic injury and certain toxic axonopathies (Nukada and Dyck, 1987). The absence of nerve cell body lesions or involvement of proximal motor axons in the ventral horns of the spinal cord suggests that the lesions may originate in the mid to distal segments of axons, although a more thorough evaluation of additional brain regions will be required to assess this possibility.

The accumulation of organelles within the enlarged dystrophic axons suggests there may be a deficit in fast axonal transport. Fast axonal transport is believed to be the target for certain neurotoxins, e.g., vinca alkaloids that interfere with microtubule assembly, produce axonal dystrophy characterized by the accumulation of components of axonal transport and axonal degeneration similar to that observed here (Chan et al., 1980; Cho et al., 1983). Interestingly, vitamin E deficiency in rats has also been reported to lead to the development of axonal degeneration and axonal swellings containing various organelles (Southam et al., 1991). Similar to seleno-protein P deletion, vitamin E deficiency related lesions were the most prominent in the rostral ascending dorsal columns with little involvement of the dorsal roots, dorsal root ganglia, or neurons of the gracile or cuneate nuclei. Functional studies conducted on fast anterograde and retrograde transport in peripheral nerve in the vitamin E-deficient rats was consistent with a deficit in both transport systems. Taken together the similarities in the types and distribution of lesions observed in the Sepp ^−/−mice and vitamin E-deficient rats suggests that axonal transport may be a common target when antioxidant systems are compromised within the nervous system. Although no evidence of brain lipid peroxidation was found in the form of increased F₂ isoprostanes, the determinations were performed on whole brains; and the results do not preclude the possibility that localized elevations of lipid peroxidation may have occurred that did not significantly raise the F₂ isoprostane concentration in the whole brain analysis performed.

Sepp ^−/− mice develop severe neurological impairment with spasticity when they are fed diets containing ≤0.1 mg selenium per kg (Hill et al., 2004). The impairment is prevented by a diet containing >0.25 mg selenium per kg, although careful neurotesting has revealed subtle abnormalities in all Sepp ^−/− mice (K. E. Hill and R. F. Burk, unpublished observations). Table 1 shows that spastic Sepp ^−/− mice fed selenium-deficient diet and most nonspastic Sepp ^−/− mice fed a high selenium diet had similar axonal degeneration scores. These results do not allow a simple correlation of axonal degeneration score with severity of clinical manifestations. It is possible that feeding a low selenium diet worsens axonal degeneration but that this worsening is not reflected in the assessments we have made and there may be more advanced changes in other regions of the central or peripheral nervous systems, e.g., more distal motor tracts or peripheral nerves. Indeed, an earlier study showed that feeding a high selenium diet to neurologically impaired Sepp ^−/− mice that had been previously fed a selenium-deficient diet prevented further deterioration but did not diminish the severity of their impairment (Hill et al., 2004). That observation appears more consistent with an irreversible or slowly reversible pathogenetic mechanism, e.g., structural changes resulting in loss of axons, as opposed to a more rapidly reversible metabolic or biochemical abnormality. But it is possible that the low selenium diet has another effect on the central nervous system, which, combined with axonal degeneration, leads to the severe clinical manifestations we have observed. Further work is required to address these questions.

Selenoprotein P has at least 2 functions important to the brain. One is to transport selenium from the liver and other tissues to the brain, as has been demonstrated in earlier studies (Burk et al., 1991; Hill et al., 2003; Schomburg et al., 2003). The other function (or functions) has not been characterized but is implied by the fact that selenoprotein P is expressed by brain cells (Yang et al., 2000). This latter function might involve transport of selenium within the brain, an enzymatic function of the protein (Burk and Hill, 2005), or both. The present study shows that axonal injury occurs in mice with deletion of selenoprotein P. However, the spasticity that occurs when these mice are fed a low selenium diet appears to involve an additional injury or neurological dysfunction. More work will be required to identify the pathogenesis of the neurological signs and mortality of the Sepp ^−/− mice fed the low selenium diet and to further characterize the functions of selenoprotein P in the central nervous system.