SJL Mice Exposed to Cuprizone Intoxication Reveal Strain and Gender Pattern Differences in Demyelination

Abstract

The role of mouse strain and the influence of gender on demyelination were explored for the first time in SJL mice using the cuprizone intoxication model. We document here that SJL mice display a unique pattern of demyelination that did not follow the profile that is well‐characterized in C57BL/6 mice. The SJL mice did not readily demyelinate at the midline within the corpus callosum but showed greater demyelination immediately lateral to midline. During continuous exposure to cuprizone, demyelination was not complete and appeared to plateau after week 7. Importantly, female mice were partially resistant to demyelination, whereas male mice were more severely demyelinated. Differences in the number of mature oligodendrocytes were consistent with the extent of demyelination; however, microglia, astrocyte and oligodendrocyte precursor cell populations did not differ between male and female mice. Thus, genetic factors and gender influence susceptibility to demyelinating disease in the cuprizone model, which may provide additional insights into the variability observed in human demyelinating diseases such as multiple sclerosis.

INTRODUCTION

Multiple sclerosis (MS) is a degenerative disease of the central nervous system (CNS) in which autoimmune damage to myelin is a primary pathological finding. Each year 10 000 individuals in the United States are diagnosed with MS, with women outnumbering men by at least twofold. In addition to gender differences in susceptibility, MS may also display subtle differences in disease severity between men and women. For instance, magnetic resonance imaging scans of 281 female and 132 male MS patients showed that men had a lower number of contrast‐enhancing lesions, which represent areas of active inflammation, but a higher number of “black hole” lesions, which mark axonal loss (47). This suggests that men with MS may develop less inflammatory but more destructive lesions. Possible explanations for these gender differences include effects of gonadal hormones, inherent differences between male and female immune systems or genetic factors that may or may not involve sex chromosomes.

In the experimental autoimmune encephalomyelitis (EAE) animal model for demyelinating disease, several strains of rodents display gender differences in disease incidence and severity 3, 4, 44, 53. However, the nature of the gender difference is highly dependent upon specific strain characteristics and the method of disease induction. A comprehensive comparison of EAE clinical disease scores in several mouse strains revealed that female SJL and ASW have more severe EAE than males when induced by active immunization (44). In addition, female NZW show a higher incidence of EAE than males, but male B10.PL and PL/J have more severe EAE than females. There was no appreciable sex difference in active EAE in C57BL/6 or non‐obese diabetic mice. This indicates that genetic variability has a strong influence on gender differences in EAE. The influence of gender has also been explored in Theiler's murine encephalomyelitis virus (TMEV)‐induced demyelinating disease in mice. In this model, male SJL/J display more severe neurological deficits than female mice (1). The fact that SJL mice show gender differences in both the EAE and TMEV models indicates that the SJL strain is a logical choice for gender studies in the cuprizone model.

Early studies characterizing CNS pathology resulting from cuprizone treatment in ICI, Swiss and Swiss Webster mice used large doses of cuprizone, in the 0.5%–0.6% range, to induce demyelination 7, 31, 50. In these early cuprizone studies male mice were used, and it was noted that female Swiss mice were not susceptible to cuprizone‐induced demyelination (31). In contrast, Elsworth and Howell cited using equal ratios of male and female albino mice for their studies, but did not note whether any gender differences were observed (15). These studies focused on characterization of demyelination, oligodendrocyte depletion and the numbers of microglia and astrocytes in the cerebellar peduncles using electron microscopy (31). Demyelination appeared complete by the fifth week of treatment. In each strain, differences in cuprizone dosage, time course of demyelination and cellular changes were noted with a high degree of accuracy. In addition, staining for myelin basic protein and myelin‐associated glycoprotein by immunohistochemistry showed changes that correlated with demyelination and remyelination in the superior cerebellar peduncle (32). However, several limitations, including the genetic background, prevented full utility of the model.

In the C57BL/6 strain, initial studies indicated that a lower dose of 0.2% of cuprizone was optimal (23) to induce full demyelination of the midline corpus callosum at week 5. Coincident with mature oligodendrocyte apoptosis, demyelination was accompanied by accumulation of microglia and astrocytes within the developing lesion. Interestingly, this environment appears to promote infiltration of oligodendrocyte precursor cells (OPCs) presumably from the subventricular zone 34, 39. When cuprizone is removed from the diet, OPCs differentiate into mature oligodendrocytes and remyelination occurs over the next 5–7 weeks. Thus, acute exposure to cuprizone results in a consistent temporal pattern of demyelination and remyelination that is marked by specific cellular, biochemical and molecular changes.

In contrast, the continuous exposure of C57BL/6 mice to cuprizone results in a pattern of two cycles of demyelination and remyelination, followed by chronic demyelination from which animals do not recover 36, 39. The first remyelination is robust but incomplete, and the second remyelination is limited (39). In this strain, the inability to remyelinate is associated with a depletion of OPCs and can be stimulated by injection of functional OPCs into the lesion (38). Thus, the C57BL/6 cuprizone model has been useful to our understanding of the limitations to repair that which occur in chronically demyelinated lesions in MS 22, 49.

Our initial studies of the role of gender in C57BL/6 mice in the cuprizone model (L.C. Taylor, unpub. obs.), and the lack of sex differences in EAE in C57BL/6 mice indicate that this strain may not be a good choice for the study of gender factors in the cuprizone model. Here, the focus is on SJL mice, with the goal of determining whether gender influences the pattern of demyelination and remyelination in cuprizone toxicity. The data indicate that SJL mice show a different temporal and spatial pattern of cuprizone‐induced demyelination than C57BL/6 mice. In addition, unlike C57BL/6 mice, female SJL mice show less severe oligodendrocyte loss and demyelination than male SJL mice.

MATERIALS AND METHODS

Animals and cuprizone treatment

Adult male and female SJL/J mice were purchased from Jackson Laboratories (Bar Harbor, ME USA) and were used for experiments at 8 weeks of age. An initial dose titration was performed with 0.1%, 0.2%, 0.3%, 0.4% and 0.5% cuprizone [oxalic bis(cyclohexylidenehydrazide)] (Sigma‐Aldrich, St Louis, MO, USA) mixed into ground or Purina mouse chow and fed ad libitum for 5 weeks. In order to determine the temporal pattern of cuprizone‐induced demyelination, both male and female mice were either untreated or treated with 0.2% cuprizone (as was determined to be the optimal dose) for 3, 4, 5, 6, 7, 8, 9 or 10 weeks. The data for this time course is presented as a combination of two separate experiments, each of which included four male and four female mice at each time point, for a total of eight animals of each gender at each time point. Untreated mice were fed ground Purina chow during the time that experimental mice were being treated with cuprizone. A preliminary study of remyelination during the second time course was performed with four male and four female mice, which were treated with 0.2% cuprizone for 7 weeks then switched to a normal diet for 3 weeks. Also, four additional mice of each gender were treated with 0.2% cuprizone for 3, 6 or 8 weeks in order to assess oligodendrocyte progenitor cells, which require frozen sections. In addition, four male C57BL/6J mice bred in a pathogen‐free facility at University of North Carolina‐Chapel Hill (UNC‐CH), were treated with 0.2% cuprizone for 5 weeks along with the SJL mice as a control for the expected pattern of demyelination as previously described 23, 39. A comparison of these mice with SJL is shown in Figure 1. All animal breeding and use was performed in compliance with the NIH Guide for Care and Use of Laboratory Animals and approved by the UNC‐CH Institutional Animal Care and Use Committee.

Figure 1.

Illustration of areas scored for myelination. A. Midline scores were derived from coronal sections of the SJL brain above the fornix (bregma −0.5 to −0.7 mm) as indicated by the solid box in the left figure. Lateral scores were immediately adjacent to the midline areas of the corpus callosum that is depicted by the solid box in the right figure. Far lateral areas indicated by the dotted box also showed demyelination but were not scored. The letter V indicates a ventricle. B. Comparison of demyelination in midline vs. lateral corpus callosum of SJL and C57BL/6 mice. Luxol fast blue–periodic acid Schiff stain for myelin (blue fibers), microglia/macrophages (pink cytoplasm) and demyelinated axons (pink fibers). Left column depicts the regions that were scored for demyelination: midline and lateral, scale bars represent 200 µm length. In the two right columns, the corpus callosum is outlined by dotted lines and shows representative images of lateral and midline regions from untreated, 5‐ and 10‐week cuprizone‐treated animals. Two right column scale bars represent 50 µm length. Note the presence of two layers of fiber tracts in SJL, especially evident in the 10‐week treated brain. The bottom row shows midline and lateral areas of 5‐week treated C57BL/6 mice that were treated along side the SJL mice and are typical of the robust demyelination and gliosis previously reported in this strain [157 × 161 mm (600 × 600 dots per inch)].

Tissue preparation

To analyze OPCs, brains were removed, submerged in freezing media and immediately frozen in a super‐cooled isopentane bath. All other mice were deeply anesthetized and intracardially perfused with 0.15 M phosphate buffer followed by 4% paraformaldehyde (PFA) solution. Brains were removed, post‐fixed overnight in PFA and embedded in paraffin. Five micrometers of coronal brain sections were cut at the fornix region of the corpus callosum (approximately bregma −0.5 to −0.7 mm) and corresponding to figure 37 in The Mouse Brain In Sterotaxic Coordinates (45).

Luxol fast blue–periodic acid Schiff (LFB–PAS) stain

To examine demyelination and remyelination, paraffin sections were stained with LFB (Sigma), which stains myelin blue, and PAS (Sigma), which stains microglia/macrophages and demyelinated axons pink. Sections were scored blinded based on the amount of blue or pink fibers in the corpus callosum, on a scale of 3 (complete myelination equal to an untreated mouse) to 0 (complete demyelination, as seen during peak cuprizone demyelination). The regions analyzed were the midline corpus callosum, and a region just lateral to midline (illustrated in Figure 1).

Immunohistochemistry

The detection of mature oligodendrocytes was performed with antibody to the pi isoform of glutathione S‐transferase (GST_pi) (Biotrin, Newton, MA, USA). Paraffin sections were rehydrated and permeabilized with 0.1% Triton X‐100/ 2% normal goat serum in phosphate‐buffered saline (PBS) for 20 minutes at room temperature. Tissue was unmasked in 0.1% calcium chloride/0.1% trypsin in 0.05 M Tris, pH 7.4 for 15 minutes at 37°C. Sections were rinsed in PBS and incubated with GSTpi antibody (1:1000) or isotype control overnight at 4°C. Following appropriate rinsing, sections were incubated for 1 h at room temperature with a goat anti‐rabbit immunoglobulin G (IgG) AlexaFluor‐conjugated secondary antibody (1:400) (Molecular Probes, Eugene, OR,USA), rinsed and coverslipped with Vectasheild plus 4′,6‐diamidino‐2‐phenylindole (DAPI) (Vector Laboratories Inc, Burlingame, CA, USA) to counterstain nuclei.

Microglia/macrophages were detected with biotinylated lectin Ricinus communis agglutin‐1 (RCA‐1) (Vector Laboratories Inc.). Paraffin sections were rehydrated and unmasked with 0.025% protease, type XIV (Sigma‐Aldrich) for 2 minutes at 43°C. Following a brief rinse in PBS, they were blocked with 0.1% Triton X‐100/ 1% bovine serum albumin in PBS for 1 h at room temperature. Sections were then incubated with RCA‐1, 1:500 in blocking solution or blocking solution alone as a control, overnight at 4°C. Following appropriate rinsing, sections were incubated for 1 h at room temp with a streptavidin AlexaFluor‐conjugated secondary reagent (1:400) (Molecular Probes), rinsed and coverslipped with Vectasheild plus DAPI (Vector Laboratories Inc.) to counterstain nuclei.

Astrocytes were detected with antibody to glial fibrillary acidic protein (GFAP) (Invitrogen, Carlsbad, CA, USA). Paraffin sections were rehydrated and unmasked with 0.025% protease, type XIV (Sigma‐Aldrich) for 2 minutes at 43°C. Following a brief rinse in PBS, they were blocked with 0.1% Triton X‐100/ 2% normal goat serum in PBS for 1 h at room temperature. Sections were then incubated with GFAP antibody (1:200) or isotype control overnight at 4°C. Following rinsing, sections were incubated for 1 h at room temp with a goat anti‐rat IgG AlexaFluor‐conjugated secondary antibody (1:400) (Molecular Probes), rinsed and cover slipped with Vectasheild plus DAPI (Vector Laboratories Inc.) to counterstain nuclei.

OPCs were detected with a rabbit antibody to NG2, a kind gift from Dr W.B. Stallcup (BIMR, La Jolla, CA USA). Five‐micrometer frozen sections were fixed in 95% ethanol before being stored at −80°C. Upon removal from the freezer, sections were post‐fixed in cold acetone, rinsed in 0.05 M potassium PBS (KPBS), and blocked with 0.1% Triton X‐100/ 5% normal goat serum in KPBS for 1 h at room temperature. Sections were then incubated with NG2 antibody (1:500 in blocking solution) or isotype control overnight at 4°C. Following rinsing, sections were incubated for 1 h at room temp with a goat anti‐rabbit IgG AlexaFluor‐conjugated secondary antibody (1:600) (Molecular Probes), rinsed and cover slipped with Vectashield plus DAPI (Vector Laboratories Inc.) to counterstain nuclei.

All comparative analyses were focused in the corpus callosum at the midline or just lateral to midline (Figure 1). Immunohistochemically positive cells were analyzed using an Olympus (Melville, NY) BX40 microscope, Olympus DP70 digital camera and ImageProPlus software (Media Cybernetics, Silver Spring, MD) and quantified per square millimeter. Positive‐stained cells were counted only if a nucleus was observed.

Statistical analysis

Statistical comparisons between time points for each gender separately were carried out using a one‐factor ANOVA and Tukey's test for multiple comparisons. Statistical analysis to determine whether there was a gender effect was made using a two‐factor ANOVA and Bonferroni correction for comparisons between time points. The data are expressed as mean ± standard error of the mean.

RESULTS

Localization of demyelination within the corpus callosum of SJL mice

We first examined brains of SJL mice for demyelination at 5 weeks following exposure to cuprizone, the time when demyelination is complete in the C57BL/6 mouse 23, 35, 39. Coronal sections were stained with LFB–PAS to determine the extent of demyelination. Robust demyelination did not occur at midline as observed in C57BL/6 mice, but instead localized immediately lateral to midline (Figure 1). In addition, there appears to be two different layers of myelinated fibers, each showing distinct patterns of demyelination, with the lower lateral layer, adjacent to the fornix, showing less demyelination than the upper lateral layer (Figure 1B). Myelination scores reflect a combination of both of these layers. This pattern is more similar to that occurring in the caudal corpus callosum in the C57BL/6 mice at the level of the hippocampus, in which three myelin layers are sometimes observed (G.K. Matsushima, unpub. obs.). Although the lateral areas appear to be more susceptible to demyelination than the midline in SJL mice, we have characterized the temporal pattern of demyelination in both regions.

Titration of cuprizone in SJL mice

SJL mice were exposed to different concentrations of cuprizone (0.1, 0.2, 0.3, 0.4 and 0.5%) in the diet to determine the amount of cuprizone required to induce demyelination without significant toxicity. Again, mice were sacrificed at week 5, corresponding to the time of full demyelination in C57BL/6 mice. Mice on a diet of 0.3% or greater exhibited overt toxic effects of cuprizone indicated by diminished weight, lethargy and lack of grooming. Over half of the animals from these groups succumbed to the higher doses, similar to C57BL/6 mice. Very little demyelination was induced by 0.1% cuprizone, not significantly different from untreated. Therefore, we determined that 0.2% provided the optimal cuprizone dosage to administer to SJL mice.

Temporal pattern of demyelination at midline and lateral areas of the corpus callosum

Male and female SJL mice were treated continuously with 0.2% cuprizone for 3 to 10 weeks and analyzed at weekly intervals to determine the time frame of maximal demyelination and possible remyelinating episodes. Sections of the corpus callosum were stained for myelin with LFB–PAS and scored in a blinded manner both at midline and lateral regions. Demyelination was also observed far lateral in the corpus callosum, on the opposite side of the ventricle (Figure 1A). However, the fiber density and variability made scoring less reliable in these regions, so they were not included in the illustrated assessments. As shown in Figure 2, demyelination in male and female mice continuously exposed to cuprizone occurred gradually, from week 3 to week 10. In the midline region (Figure 2A), cuprizone‐treated male SJL mice showed significantly greater demyelination at weeks 8–10 compared with untreated mice (P < 0.05; one‐factor ANOVA with Tukey's posttest). In female SJL mice in the midline region, demyelination reached a statistically significant difference from untreated mice only at weeks 9 and 10 (P < 0.001). Comparison of male and female mice using two‐factor ANOVA showed a significant gender effect on demyelination over the entire observed time period (P = 0.001). The midline corpus callosum showed only partial demyelination and remyelination was not observed during the course of cuprizone administration.

Figure 2.

Cuprizone‐induced demyelination of the corpus callosum. Myelination scores of the corpus callosum from male and female SJL mice continuously exposed to cuprizone (eight mice per gender per group, except for the recovery time point that is composed of four mice per gender per group). A. Midline corpus callosum: two‐factor ANOVA indicates a significant gender effect (P = 0.001). B. Lateral corpus callosum: two‐factor ANOVA indicates a significant gender effect (P = 0.0305). C. Representative images of Luxol fast blue–periodic acid Schiff stain in male (M) and female (F) lateral corpus callosum at selected time points. Scale bar equals 50 µm; 187 × 302 mm (72 × 72 dots per inch).

The same coronal sections of brains from SJL mice continuously exposed to cuprizone were scored for demyelination immediately lateral to the midline area (see Figure 1). As shown in Figure 2B, the temporal pattern of demyelination in the lateral areas is dramatically different than at midline. Mild demyelination was detected in male SJL mice beginning at week 3, reaching statistical significance compared with untreated mice at weeks 5 and 6 (P < 0.05). More robust demyelination was observed in males at week 7 (P < 0.001), reaching a plateau through week 10. The partial demyelination present at week 10 persists as long as week 13 (data not shown). Thus, unlike male C57BL/6 mice, which show three cycles of demyelination while on continuous exposure to cuprizone over 13 weeks (36), SJL mice appear to demyelinate gradually once during these time points; however, the demyelination is incomplete.

Equally important in female SJL mice, demyelination showed a slow, delayed course of demyelination (Figure 2B), which is moderate from weeks 3 through 6 and does not reach a statistically significant difference from untreated until week 7 (P < 0.001). This is in contrast to males that reached statistically significant demyelination by week 5. From weeks 5 to 10, female SJL mice exhibited less demyelination than male counterparts, and this gender effect over the entire observed time period is significant (two‐factor ANOVA P = 0.0305). Thus, female SJL mice appear to be less susceptible to demyelination than male mice. In both male and female SJL there is no apparent remyelination during this time period.

A preliminary study of remyelination was performed by removing cuprizone from the diet at the 7 week time point, and allowing the mice to recover for 3 weeks. Interestingly, when compared with the 7‐week time point, significant remyelination did not occurr in either the midline or lateral regions (Figure 2A,B). Furthermore, there was no significant gender difference in myelin score at this time point.

Quantification of mature oligodendrocytes

In C57BL/6 mice continuously exposed to cuprizone, mature oligodendrocytes are depleted by week 5 and recover temporarily at week 6 before diminishing a second time (38). Here, we quantified the mature oligodendrocyte population in the midline and lateral areas. Untreated male and female mice began with similar numbers of oligodendrocytes in the corpus callosum (Figure 3). As shown in Figure 3A, the number of mature oligodendrocytes at midline is initially diminished at week 3 but did not reach a statistically significant difference from untreated mice until week 5. Thereafter the number of mature oligodendrocytes appear to plateau through week 10. Except for week 7, female mice show a trend for greater numbers of mature oligodendrocytes than male mice from week 3–10, particularly at week 8. The overall gender effect by two‐factor ANOVA is statistically significant (P = 0.0158). Thus, similar to demyelination in Figure 2, there are significantly more mature oligodendrocytes in female mice than male mice.

Figure 3.

Mature oligodendrocytes in the corpus callosum of SJL mice. Male and female mice were treated with cuprizone continuously over the time course indicated (eight mice per gender per group, except for the recovery time point that is composed of four mice per gender per group). A. Numbers of glutathione S‐transferase pi+ (GSTpi+) oligodendrocytes were quantified from the midline of the corpus callosum. Two‐factor ANOVA indicates a significant gender effect (P = 0.0158). B. Numbers of GSTpi+ oligodendrocytes were quantified from lateral areas of the corpus callosum. Two‐factor ANOVA indicates a significant gender effect (P < 0.0001). C. Representative images of mature oligodendrocytes identified by immunoreactivity to GSTpi in male (M) and female (F) lateral corpus callosum at selected time points. Scale bar equals 50 µm; 203 × 298 mm (72 × 72 dots per inch).

In the lateral regions of the corpus callosum, a gradual depletion of mature oligodendrocytes is observed in both male and female mice, beginning at week 3, compared with untreated mice (P < 0.01; Figure 3B). Female mice have more oligodendrocytes than male counterparts at all time points, and this gender effect is highly significant (P < 0.0001). Interestingly at weeks 8 and 9, there is a trend for recovery of mature oligodendrocytes in both males and females. By week 10, the number of mature oligodendrocytes again diminishes. However, similar to demyelination, full depletion of mature oligodendrocytes is not observed during this time period.

When cuprizone administration was discontinued at week 7 and mice were allowed to recover for 3 weeks, a modest increase in oligodendrocytes was observed at both midline and lateral regions; however, statistical significance was observed only in the lateral region (Figure 3B, P < 0.05 for both male and females). However, there is no gender difference in the number of oligodendrocytes during this recovery time point (7 + 3 weeks). This increase in oligodendrocytes indicates that SJL mice are capable of recovery from cuprizone‐induced oligodendrocyte loss and correlates with the remyelination observed in Figure 2A,B.

OPCs accumulate in the demyelinated lesion

The accumulation of OPCs in demyelinated lesions has been noted since the earliest cuprizone studies in both ICI and Swiss mice 7, 31. In these studies OPCs were identified by ultrastructural characteristics and were detected at 5 weeks of cuprizone treatment when the lesions in superior cerebellar peduncles were nearly fully demyelinated. The availability of antibodies to the NG2 protein, which is commonly used as one marker for OPCs, has allowed for easier quantification of these cells. In C57BL/6 mice, NG2‐positive OPCs are present in limited numbers in untreated mice, accumulate during cuprizone‐induced demyelination and reach maximal numbers at 4 and 5 weeks of intoxication (34). During recovery, OPC numbers decline slightly, presumably because of their maturation into oligodendrocytes that are responsible for the remyelination of the lesion 2, 34. However, chronic administration of cuprizone (8–12 weeks) leads to progressive depletion of OPCs from the lesion (38). Here, we used the NG2 marker to identify the pattern of OPC accumulation in SJL male and female mice at select times during demyelination. A quantification of NG2‐positive OPCs indicates that a few of these cells are present in the untreated corpus callosum, in similar numbers between male and female SJL (Figure 4). During cuprizone intoxication, OPC numbers increase gradually, becoming statistically significant at the 8‐week time point. Unlike the response of mature oligodendrocytes, the numbers of OPCs are very similar in the midline and lateral regions of the corpus callosum. In addition, OPCs seem to differ from mature oligodendrocytes in that there is no statistically significant gender difference in their numbers (Figure 4).

Figure 4.

Oligodendrocyte precursor cells (OPCs) in the corpus callosum of SJL mice. Male and female mice were treated with cuprizone continuously over the time course indicated (four mice per gender per group). A. Numbers of NG2+ OPCs were quantified from the midline of the corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. B. Numbers of NG2+ OPCs were quantified from the lateral corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. C. Representative images of OPCs identified by immunoreactivity to NG2 in male (M) and female (F) lateral corpus callosum at selected time points. Scale bar equals 50 µm; 235 × 314 mm (72 × 72 dots per inch).

Quantification of microglia

Microglia appear during demyelination and typically disappear with remyelination 2, 38, 46. Here we assessed the microglial response at midline and lateral regions to determine if their numbers correlated with the extent of demyelination. As show in Figure 5A, microglia accumulation at midline in both male and female mice is detected at weeks 3 and 5 and becomes statistically significant compared with untreated mice during weeks 6 through 10 (P < 0.05). In female mice there appears to be a plateau from week 6 through week 10, whereas male mice exhibit an upward trend in microglial numbers until week 9. At a few time points microglia accumulation appears to be less robust in female mice; however, there is not a statistically significant gender effect.

Figure 5.

Microglia in the corpus callosum of SJL mice. Male and female mice were treated with cuprizone continuously over the time course indicated (eight mice per gender per group). A. Numbers of Ricinus communis agglutin‐1‐positive (RCA‐1+) microglia were quantified from the midline of the corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. B. Numbers of RCA‐1+ microglia/macrophages were quantified from the lateral corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. C. Representative images of microglia/macrophages identified by reactivity to the lectin RCA‐1 in male (M) and female (F) lateral corpus callosum at selected time points. Scale bar equals 50 µm; 220 × 300 mm (72 × 72 dots per inch).

In the lateral regions, female and male mice show microglia accumulating between week 3 through week 6 and then the number of microglia appear to level off (Figure 5B). There were no significant differences among male and female mice in terms of microglial responses to demyelination. Interestingly, the number of microglia in the lateral regions is greater in number compared with that of the midline, probably because of the greater demyelination observed in Figure 2B.

Quantification of astrocytes

Similar to microglia (though fewer in number), astrocytes infiltrate the demyelinated lesion, but unlike microglia, they typically persist through remyelination (46). Here we assessed whether differences observed for demyelination and mature oligodendrocyte scores could be partly explained by the astrocyte populations. At midline, an endogenous population is present as expected and there is no difference between male and female mice. After exposure to cuprizone, a higher number of astrocytes begin to populate the midline of the corpus callosum and by week 5, they appear to have reached maximal numbers in both male and female mice (Figure 6A). There is no difference in the astrocyte response between male and female mice at midline.

Figure 6.

Astrocytes in the corpus callosum of SJL mice. Male and female mice were treated with cuprizone continuously over the time course indicated (eight mice per gender per group). A. Numbers of glial fibrillary acidic protein‐positive (GFAP+) astrocytes were quantified from the midline of the corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. B. Numbers of GFAP+ astrocytes were quantified from the lateral corpus callosum. Two‐factor ANOVA indicates there is no significant gender effect. C. Representative images of astrocytes identified by immunoreactivity to GFAP in male (M) and female (F) lateral corpus callosum at selected time points. Scale bar equals 50 µm; 216 × 292 mm (72 × 72 dots per inch).

In the lateral areas, the appearance of astrocytes is very similar to the midline scores. Interestingly, unlike the greater number of microglia found in the lateral areas, astrocyte numbers (∼500 cells/mm²) in the lateral regions are similar to the numbers at midline. Although female mice appear to have a slight trend for higher numbers of astrocytes at several time points, the overall differences are not statistically significant.

DISCUSSION

The SJL strain of mice revealed several differences when compared with the C57BL/6 mice in the cuprizone model of demyelination. First, the location of appreciable demyelination within the corpus callosum is immediately lateral to midline, rather than midline, as observed in the C57BL/6 mouse. Second, demyelination is only partial in SJL mice through 10 weeks of continuous treatment whereas full demyelination is observed at week 5 in C57BL/6. Third, demyelination was less severe in female SJL mice than male mice. Fourth, mature oligodendrocytes in female SJL mice were more resistant to depletion than male counterparts. There were no apparent differences between the sexes in the number of oligodendrocyte precursors, microglia or astrocytes in the demyelinated regions. Furthermore, unlike C57BL/6 mice that remyelinate quickly upon discontinuation of cuprizone, SJL mice that were treated with cuprizone for 7 weeks and then allowed to recover for 3 weeks showed only slight remyelination. However, SJL mice are similar to C57BL/6 mice in that a 0.2% diet of cuprizone promotes demyelination without the severe toxic effects observed at 0.3% diet of cuprizone or greater.

Interestingly, several strains of mice show differences in sensitivity to cuprizone intoxication. Studies of ICI and Swiss Webster mice used 0.5% to 0.6% cuprizone in their diet to induce demyelination 8, 27, 31. Though not tolerated by all of these mice, many of them mice lived for months on the cuprizone diet. In contrast, C57BL/6 mice and SJL mice are more susceptible to cuprizone, showing demyelination at the lower 0.2% dose. Similar to C57BL/6 mice (23), SJL mice die or are lethargic at 0.3%, and higher doses of 0.4, 0.5 and 0.6% resulted in death within the first couple of weeks. It is not clear why there is strain variability in sensitivity to cuprizone. Recently, strain differences have been reported for cortical demyelination induced by cuprizone. In this study BALB/cJ mice were treated with 0.2% cuprizone and compared with C57BL/6 (48). The BALB/cJ mice were reported to have a similar pattern of demyelination in the corpus callosum, but less demyelination and more microglia activation in the cortex as compared with C57BL/6 mice.

The difference in response because of genetic background also becomes important when crosses of different mice are used to analyze specific genes. Unless there is an attempt to backcross to an established genotype such as C57BL/6, the background genes of the knockout or transgenic mice are not homogeneous and could influence outcomes. This may be also true for 129 mice that have similar patterns of demyelination in the corpus callosum as C57BL/6, but demyelinate at a lower dose of 0.1% of cuprizone (data not shown). Thus, mice on the 129 background may show more severe effects if tested at the 0.2% cuprizone dose, and mice of a mixed background may give a more complicated pattern. The data in this study emphasize the importance of considering the influence of genetic background in the study of demyelination, regardless of the model, (EAE, viral or toxin‐induced demyelination).

Genetic background clearly influences the morphologic and temporal pattern of demyelination. We found SJL mice demyelinate partially at midline (Figure 2A), whereas C57BL/6 mice show robust demyelination, particularly at week 5 2, 23, 35. SJL mice show greater demyelination in the lateral region of the corpus callosum immediately adjacent to the midline; however, two distinct layers appeared to demyelinate differently. The lower one‐third is more resistant to demyelination than the upper two‐thirds (Figure 1B), hence, myelin scores in the SJL mice are a bit more complex requiring a cumulative subjective combination. Scores in Figure 2B are reflective of the entire lateral corpus callosum region indicated in Figure 1. The utility of the LFB–PAS histological stain is illustrated here as myelin fibers that are stained blue vs. unmyelinated fibers that stain pink can be estimated quickly for a large number of samples. It would be cumbersome to assess these sections by electron micrography as all axons within the tract would have to be counted because representative areas would be difficult to choose. Previous studies have shown that LFB–PAS scores are reflective to the trends in demyelination when compared with percent myelinated fibers that are counted in electron micrographs 29, 36. However, one limitation of the LFB–PAS stain is that one cannot detect whether any remyelination is occurring during the demyelination period. In order to determine whether there is any remyelination occurring during the cuprizone treatment, electron microscopy could be utilized to monitor thinly myelinated axons that are characteristic of remyelination (41). However, in C57BL/6 mice, we do not detect any apparent remyelination in the midst of demyelination, and remyelinated axons are not detected until after complete demyelination that occurs at week 5.

The reason for regional differences in the demyelination of the corpus callosum observed in SJL mice is not known. There are many examples of regional differences in dysmyelination caused by genetic manipulation of mice between areas such as the optic nerve, spinal cord, cerebellum and corpus callosum. For example, laminin‐deficient dystrophia muscularis mice exhibit hypomyelination in the corpus callosum and optic nerve, but not the spinal cord (11), whereas dominant negative β1 integrin ΔC mice are hypomyelinated in spinal cord and optic nerve, but not the corpus callosum (28). Heterogeneity of oligodendrocytes is one possible explanation for such regional differences; however, this seems unlikely in the present case, given that we are seeing differences within the corpus callosum rather than between different white matter regions. One possible explanation for the appearance of two different layers of demyelination (Figure 1B) in the lateral region of SJL corpus callosum during demyelination is that these may represent two different groups of axonal fibers coming from different types of neurons and are providing different levels of trophic support or conversely different levels of detrimental signaling. Similar layers that show differential demyelination in C57BL/6 mice occur in a more caudal regions of the corpus callosum next to the hippocampus (data not shown) and the reason for these differences is not clear. Other alternative explanations are that the cuprizone has better access to oligodendrocytes in the upper lateral layers or these oligodendrocytes are hypersensitive to insult.

Our temporal analysis of demyelination in the lateral corpus callosum indicates 7 weeks of cuprizone treatment is optimal for detecting marked demyelination in SJL mice. From 7 weeks onwards, there is only modest continued demyelination. In addition, 7 weeks is when the greatest loss of oligodendrocytes was observed in both genders. One of the more striking differences in SJL mice is the lack of total demyelination during a 10‐week period of continuous exposure to cuprizone. This is very different compared with the C57BL/6 mice that fully demyelinate by week 5. Even up to week 13 of continuous exposure to cuprizone, demyelination scores were approximately 0.75 for males and 1.0 for females at the lateral region, similar to week 10 scores in Figure 2B. Although higher doses may induce full demyelination, we could not examine SJL mice at doses of 0.3% or greater as they are susceptible to systemic toxicity. In addition, C57BL/6 mice experience a spontaneous partial remyelination after 6 weeks and then continue to demyelinate further when exposed continuously to cuprizone 36, 39. Despite a trend for an increase in the oligodendrocyte population at weeks 8 and 9, we did not detect any remyelination during continuous cuprizone exposure in SJL mice.

An additional interesting difference of SJL mice from C57BL/6 mice is the lack of robust remyelination after 3 weeks of recovery from the cuprizone diet. C57BL/6 mice showed greater than 50% remyelination within 2 weeks of discontinuation of cuprizone 2, 42, which is in stark contrast to the reduced remyelination reported here for SJL. However, it should be emphasized that the remyelination analysis performed here is limited, and a more thorough examination to determine whether remyelination progresses further would require later time points. The quantification of mature oligodendrocytes indicates that there is some recovery of these cells after 3 weeks of removal from cuprizone, and it is possible that this would lead to greater remyelination at later time points. A less robust remyelination in SJL could be explained by the less robust infiltration of microglia and astrocytes that are known to produce factors such as insulin‐like growth factor‐1 and tumor necrosis factor‐α that are important for promoting remyelination 2, 35, 37.

Another important finding is the difference between male and female SJL mice during demyelination. At many time points, male mice were more severely demyelinated than female mice in both midline and lateral regions, and the overall gender effect was statistically significant. In addition, when we analyzed the oligodendrocyte populations at midline and lateral regions, we found a statistically significant preservation of oligodendrocytes in female mice compared with males (Figure 3). Thus, female SJL mice are more resistant to oligodendrocyte loss and demyelination, and this trend was observed at most of the time points between weeks 5 through 10. A difference in the size of corpus callosum or numbers of oligodendrocytes in untreated mice is unlikely to explain the resistance to toxicity reported here. As shown in Figure 3, untreated male and female SJL mice have similar numbers of oligodendrocytes in the corpus callosum. Although the size of the corpus callosum was not measured here, data published by Bishop and Walsten (6) indicate that unlike in humans or rats, mice do not display a gender difference in corpus callosum size. The lack of difference in numbers of OPCs (Figure 4) and remyelination (Figure 2) between male and female SJL suggest that there is no gender difference in proliferation, migration or maturation of oligodendrocytes. Additional future experiments to further characterize remyelination in SJL mice may clarify this point.

It is possible that female gender affords a protection from toxicity in general, which is supported by our observations during the cuprizone dose titration that male mice died more quickly when exposed to higher doses of cuprizone (data not shown). Analysis of liver sections indicated that at the 0.2% cuprizone dose, there is little indication of toxicity in either gender as indicated by vacuolation or necrosis, but male mice do exhibit an increase in binucleated hepatocytes compared with untreated mice, whereas females do not (data not shown). Whether this indicates that female mice may handle cuprizone differently than males, or whether mechanisms for protecting oligodendrocytes in female brains are superior to those in male mice remains to be determined. A final note, Swiss female mice have been reported to be resistant to cuprizone‐induced demyelination; however, the mechanism for such differences compared with male Swiss mice is not known (31). Studies of dietary copper deficiency in rats indicate that females exhibit less severe symptoms of weight loss, anemia and lethality 16, 17, 18. The mechanism of this gender difference is not known and the potential role of endogenous sex hormones is controversial. One study found an exacerbation of some copper‐deficient symptoms, such as weight loss, in ovariectomized females (10) but no effect on severity by male or female sex hormones was detected in another study (16). It is possible that the protective effect from dietary copper deficiency or cuprizone‐chelated copper deficiency may be similar in these female subjects; however, the inherent mechanism remains elusive.

Nevertheless, these findings of gender differences are in contrast to those in the EAE model, in which female SJL mice are both more susceptible to disease induction and exhibit more severe neurological deficits 3, 44, 53. The fact that female SJL mice show increased severity in EAE, but decreased severity in cuprizone intoxication may reflect a key difference between EAE and cuprizone models: EAE has an autoimmune etiology, whereas cuprizone delivers a toxic injury to oligodendrocytes. The increased severity in EAE may be because of T‐cell immune mechanisms that are not induced by cuprizone intoxication 2, 24. Thus, in spite of a low threshold for autoimmunity in SJL females, female SJL oligodendrocytes appear to exhibit a high threshold, or resistance, to toxic injury. The utility of both models for the study of MS is supported by heterogeneity in MS lesions, reported by Lucchinetti et al (30), in which type III and IV lesions suggest primary oligodendrocyte dystrophy, whereas type I and II lesions are consistent with T‐ or B‐cell‐associated autoimmune mechanisms. Similar to our findings in the cuprizone model, female SJL mice infected with TMEV show less severe neurological deficits and a moderate (though not statistically significant) resistance to demyelination compared with males (1).

Future studies to determine the mechanism of protection of female oligodendrocytes will be an important contribution to the growing body of work addressing the role of gender in neuropathological conditions. There is much evidence for a role of sex hormones in demyelination and oligodendrocyte function. In MS there is a clear protection from relapse during the third trimester of pregnancy when several hormones, especially sex steroids, reach a peak (12). In EAE, administration of estrogens lead to functional benefits 5, 25, 43, 52. In addition, sex hormones have been shown to have direct effects on oligodendrocyte proliferation and maturation 21, 33 as well as survival 20, 51. Cuprizone intoxication of SJL mice will be a useful model to study the role of sex hormones in primary oligodendrocyte disruption.

We also monitored the microglia and astrocyte populations in the male and female mice. In the midline corpus callosum microglia accumulation began by week 3 and reached statistically significant numbers compared with untreated mice by week 6, then remained unchanged through week 10, in both male and female mice. In the lateral regions, the microglial response was more robust, with nearly 30% more microglia/mm² than at midline, and there was a slight decrease at week 10. Overall there was not a difference between male and female mice. Lastly, when we measured the astrocytic response both at midline and in the lateral regions, we observed an increase in astrocytes until week 5, when they stabilized in number. Unlike microglia, astrocyte numbers in the lateral regions remained similar to numbers at midline and did not increase appreciably. Thus, unlike microglia and astrocytes in C57BL/6 mice that respond vigorously to demyelination 2, 23, 39, these cell types in SJL mice have a guarded response at midline and in the lateral region, which correlates with subdued demyelination. There was no difference in microglia or astrocyte numbers between male and female SJL mice. This study did not address specific functions of microglia or astrocytes, which may also be influenced by sex hormones 9, 13, 14, 19, 26, 40). Future studies to address differences in cytokine and growth factor production that may provide protection for oligodendrocytes are warranted and would shed important light on the role of these cells in demyelinating disease.

In conclusion, male and female SJL mice did not fully demyelinate during a period of 10 weeks of continuous cuprizone exposure, when C57BL/6 mice would have undergone two rounds of full demyelination, indicating a clear genetic influence on susceptibility to demyelination. Importantly, SJL female mice are more resistant than males to loss of oligodendrocytes and demyelination, though there were no gender differences in microglial or astrocytic accumulation. Further study of gender and hormonal influences in SJL mice using the cuprizone model will be an important complement to EAE studies given the heterogeneity of disease pathology in human MS.