Strain differences in cuprizone induced demyelination

Multiple sclerosis (MS) is a chronic, demyelinating disease in which the myelin sheath, the insulating cover that wraps around axons of neurons, is damaged. The normal function of myelin involves increasing the speed of action potential propagation in axons as well as providing trophic support [1]. Upon damage of myelin, patients suffer from a wide range of symptoms, such as fatigue, pain, spasm, emotional changes, motor deficits and cognitive disorders [2].

The cuprizone neurotoxicity animal model is widely used in MS research [3]. The cuprizone model involves administration of the toxin cuprizone to induce oligodendrocyte apoptosis [4, 5]. Upon treatment with cuprizone, mice typically exhibit profound demyelination in the corpus callosum, the cortex and the superior cerebellar peduncles [4, 5]. In the corpus callosum, apoptosis of oligodendrocytes is observed, together with recruitment of microglia/macrophages, astrocytes and oligodendrocyte progenitor cells (OPCs) [5, 6].

Most investigators use the standard C57BL/6 mouse strain in the cuprizone model to study the different molecular components involved in the complex pathogenic processes leading to corpus callosum demyelination [5]. However, other strains are also used sometimes [7–10]. Thus it is important to know the relative sensitivity of different mouse strains in response to cuprizone toxicity. In our study, we examined the CD1 strain. By comparing it to the C57BL/6 strain, we show that CD1 mice are less vulnerable to cuprizone-induced demyelination in the corpus callosum. Therefore, genetic differences greatly influence the susceptibility of mice to cuprizone-induced damage, and that the influence of mouse strains should be taken into consideration in designing experiments using the cuprizone model.

The Black gold staining is an easy and widely used approach for visualizing myelin morphology with good resolution of individual myelin fibers [12]. We first performed Black gold staining of the 4 week-cuprizone treated samples, focusing on sections of the corpus callosum above the fornix, which is approximately between bregma −0.58 and −0.82 mm (Fig. 1a). The C57BL/6 mice showed a dramatic loss of myelin staining at this time point (Fig. 1b), which is consistent with previous reports [5]. However, the CD1 mice didn’t show significant myelin loss. This difference in response was also observed at later stages of cuprizone treatment (Fig. 1b).

An alternative histological approach that is complementary to the Black gold staining is the Luxol fast blue–periodic acid Schiff (LFB–PAS) method, which not only stains the myelin (in blue) but also stains the demyelinated areas in pink, thus providing a counter stain to contrast with the positive myelin stain. We thus performed LFB–PAS staining to confirm the results obtained with the Black gold staining (Fig. 2a). The LFB–PAS stained sections were scored blind on a scale of 3 (3 = intact and 0 = full demyelination, Additional file 1: Figure S1). As shown in Fig. 2b, the two strains of mice showed different severity of demyelination at 4–6 weeks of treatment. A close examination of the CD1 corpus callosum at 4–6 weeks of cuprizone exposure shows that CD1 mice do exhibit mild demyelination which is not apparent in the Black gold histological analysis. By 7 weeks, CD1 mice were slightly less demyelinated compared to C57BL/6 mice, but this difference is not statistically significant, although the Black gold data shows CD1 mice still exhibit more Black gold-stained fibers at this point. This slight discrepancy could be due to the different substrates that these two methods detect [13] in that LFB stains phospholipids while Black gold stains myelin proteins. Taken together, the CD1 strain exhibits a delayed and mild demyelination in comparison to C57BL/6 mice.

Because the myelin fibers are formed by oligodendrocytes, we wanted to examine whether the less severe demyelination in CD1 mice is associated with more mature oligodendrocytes being present at the lesion site. We performed immunofluorescence staining using an antibody against GSTpi, a marker for mature oligodendrocytes. As shown in Fig. 3, although both strains fed with cuprizone showed a decrease in GSTpi+ mature oligodendrocytes compared to non-cuprizone-fed controls, the decrease for C57BL/6 mice was more severe than for CD1 mice [significant differences were observed at 4, 5, 6 and 7 weeks of treatment (P < 0.05 in ANOVA at all time points)]. These observations indicate that the lesser demyelination of CD1 mice as shown in the Black gold staining and LFP-PAS stain may be attributed to more oligodendrocytes being present. We also noticed that at 7 weeks of treatment, the number of GSTpi+ oligodendrocytes in the two strains are relatively high, compared to the absence of myelin as shown in the LFB–PAS stain at this time point (compare Figs. 2 and 3 at 7 weeks). This discrepancy could be due to OPCs differentiating into pre-myelinating oligodendrocytes [14] (also see below), which are not myelinating the axons due to continuous cuprizone lesion.

Since CD1 mice show more myelin and more oligodendrocytes, we wanted to determine whether this difference is reflective of a quantitative difference in the levels of myelin proteins. To this end, we performed western blot analysis using 4 weeks-cuprizone treated corpus callosum samples (Fig. 4). We first examined the levels of myelin basic protein (MBP). As shown in Fig. 4a, c, on average, CD1 mice showed a milder reduction in MBP protein levels after cuprizone treatment compared to C57BL/6 mice, however the difference did not reach statistical significance (P > 0.05 in t-test, N = 6 for C57BL/6 and N = 8 for CD1). A comparison of untreated corpus callosum between CD1 mice and C57BL/6 mice (inset) did not reveal a significant difference (t-test, P > 0.05, N = 6 per group) between strains.

We also analyzed the level of myelin associated glycoprotein (MAG). As shown in Fig. 4a, d, the decrease in MAG level was less in CD1 mice compared to C57BL/6 mice (student’s t-test, P < 0.05, N = 6 for C57BL/6 and N = 8 for CD1), and a comparison of unlesioned corpus callosum sample between the two strains (inset) revealed a similar base line MAG level between the two strains (student’s t-test, P > 0.05, N = 6 per group).

Because the oligodendrocyte loss induced by the cuprizone lesion is usually accompanied by NG2+ OPC accumulation and differentiation [15], we analyzed the number of NG2 and to-pro-3 double positive OPCs in the cuprizone treated mice (Fig. 5). In C57BL/6 mice, NG2+ OPC number peaked at 4 and 6 weeks, while CD1 mice exhibited a gradual increase in NG2+ cell population, coinciding with increasingly severe demyelination. Although CD1 mice showed a much smaller increase in NG2+ cells at 4 and 6 weeks (P < 0.05 in ANOVA for both 4 and 6 weeks), they do show comparable numbers of NG2+ cells as C57BL/6 strain at 7 weeks. The observation that CD1 mice show less NG2+ OPC recruitment is consistent with the data that CD1 mice show more oligodendrocytes and less demyelination. We also confirmed the morphological data by western blot analysis of NG2 protein level, using 4 weeks-cuprizone treated corpus callosum samples. As shown in Fig. 4b, g, the induction of NG2 protein is significantly lower (P < 0.05, student’s t-test) in CD1 mice, compared to C57BL/6 mice.

Because cuprizone induced demyelination in the corpus callosum is also known to be associated with microglial/macrophage accumulation and prominent astrogliosis [6], we then analyzed the degree of the innate neuroinflammatory response, focusing on recruitment of microglia to the site of demyelination in the two strains. As the CD1 mice showed less myelin loss following the cuprizone lesion, we expected that CD1 mice would show less microglia compared to C57BL/6 mice. As shown in Fig. 6, the level of Iba1 and to-pro-3 double positive microglia peaked at 4 weeks of cuprizone exposure in C57BL/6 mice, and then gradually decreased in number at 5–7 weeks. In contrast, CD1 mice showed a less prominent microglial response, with the relative increase in microglia being less than C57BL/6 mice at 4, 6 and 7 weeks of cuprizone treatment (P < 0.01 in ANOVA). The numbers of Iba1+ microglia in CD1 mice stayed roughly similar throughout the time points analyzed. A western blot analysis showed that in response to 4 weeks of cuprizone administration, the levels of Iba1 was also lower in CD1 mice compared to C57BL/6 mice (Fig. 4a, e, P < 0.01 in student’s t-test), confirming the morphological observation.

We also analyzed the changes in astrocytes in the cuprizone treated mice. Both strains exhibited a gradual increase in the GFAP and to-pro-3 double positive astrocyte population from 4 to 7 weeks of treatment, with 7 weeks showing the greatest amount of astrogliosis. Compared to C57BL/6 mice, CD1 mice showed on average, a lower degree of astrocyte recruitment at all time points and statistical comparison between strains showed significant difference at 6 and 7 weeks of cuprizone treatment (Fig. 7). Western blot analysis examining GFAP levels in the two strains following 4 weeks of cuprizone treatment showed that CD1 mice exhibited on average a smaller induction of GFAP, which is consistent with the immunohistochemical analysis, but the difference did not reach statistical significance (Fig. 4a, f). Note that the arrow in the figure points to the one band/isoform of GFAP analyzed. Other bands, however, are other GFAP isoforms [16].

Finally, we asked whether the differences in demyelination could be attributed to a difference in cuprizone intake. As shown in Fig. 8, although CD1 mice are on average bigger than C57BL/6 mice (left panel), they also ingest more cuprizone containing diet (middle panel), resulting in a similar amount of food intake relative to body weight (right panel).

We have demonstrated that CD1 mice display less severe demyelination compared to C57BL/6 mice as shown using a variety of parameters. First, histology data of Black gold and LFB-PAS staining clearly indicate that CD1 mice undergo a milder and slower demyelination process. Second, immunohistochemistry data show that CD1 mice exhibit less oligodendrocyte decrease and concomitantly less recruitment of OPCs/microglia/astrocytes following cuprizone intoxication. Third, western blot analysis indicates the reduction of MAG is less in CD1 mice. Fourth, western blot analysis also confirms that CD1 mice show less increase in protein levels of NG2 and Iba1. Lastly, CD1 mice consume a similar amount of cuprizone diet per body weight as compared to C57BL/6 mice.

The earliest studies characterizing cuprizone induced pathology were done using a much higher dosage at 0.5–0.6% [7, 8, 17]. However, the cuprizone induced demyelination in the corpus callosum is best documented in the commonly used C57BL/6 strain and at a dosage of 0.2% [6, 15]. Therefore in our study, we adopted a 0.2% dosage to address cross-strain differences. For C57BL/6 mice, it was observed that the demyelination of C57BL/6 mice is robust at 4–6 weeks of cuprizone treatment, and our findings agree with the previously reported observations. However, recently, there have been several other studies that have addressed the importance of genetic background factors in influencing the extent of demyelination in the cuprizone model. For instance, SJL mice have been shown to be less susceptible to demyelination in the corpus callosum compared to the established C57BL/6 strain [10]. The authors also reported that SJL mice exhibit a unique pattern of demyelination in that the area of the corpus callosum that is immediately lateral to midline seems to demyelinate more prominently than the midline. In our study we observed several cases in which CD1 mice also showed a similar pattern (around 10% of the cuprizone treated CD1 mice, data not shown). In addition to the corpus callosum, demyelination in the cortex induced by cuprizone treatment has also been shown by Skripuletz et al. to be affected by genetic background factors [9]. In their study the authors challenged BALB/cJ mice with 0.2% cuprizone treatment of 6 weeks, and observed that cortical demyelination of BALB/cJ mice was incomplete, as opposed to C57BL/6 strain which shows no detectable cortical myelin at this stage. Taken together, strain differences and genetic background factors greatly influence the sensitivity, the time course and the pattern of cuprizone induced demyelination.

The cuprizone model of induced demyelination has been widely adopted to study the pathological process of MS for the past 20 years. However, the exact mechanism of how cuprizone induces demyelination remains elusive [18]. Nevertheless, a consensus seems to be that cuprizone selectively affects oligodendrocytes. Evidence includes giant mitochondria formation [6, 8, 19–23], increased oxidative stress [18, 19, 24–26] and endoplasmic reticulum (ER) stress [27–29] specifically within oligodendrocytes, eventually leading to oligodendrocyte apoptosis [5, 30]. Importantly this apoptosis precedes the massive demyelination caused by cuprizone administration [18], indicating that oligodendrocyte degeneration is the main cause, as opposed to a secondary immune response. This actually makes the cuprizone model particularly suitable for studying subtypes of MS which show primary oligodendrocyte dystrophy, reminiscent of toxin- or virus-, rather than autoimmunity, induced demyelination [31, 32].

In addition to the direct effect of cuprizone on oligodendrocytes, the involvement of the brain’s resident neuroinflammatory cells such as microglia and astrocyes is also believed to be responsible for the massive demyelination occurring at 4 weeks time point of cuprizone exposure [18]. Important supporting evidence includes (1), a switch from Caspase 3-dependent to Caspase 3-independent form of apoptosis is reported for later stages of cuprizone exposure [33]; (2), cuprizone-induced apoptosis in in vitro oligodendrocyte culture was successful only after addition of pro-inflammatory cytokines, which are produced by microglia and astrocytes [34]; (3), inhibition of microglial activation by minocycline prevented cuprizone induced demyelination in vivo [34]; (4), microglial processes were observed in between myelin lamellae, possibly stripping myelin from axons [8]. In our experiments, we observed a dramatic increase of microglia in C57BL/6 mice following 4 weeks of cuprizone exposure, which is similar to previous reports [6]. However, for the CD1 strain, we did not observe prominent microgliosis (Fig. 6). Even after 7 weeks of treatment, when CD1 mice already show extensive demyelination in the LFB-PAS stain (Fig. 2), the levels of Iba1+ cells are still well below that of the C57BL/6 strain. Thus we propose that the strain differences observed in this study can be at least partially accounted for by a difference in the sensitivity of the microglial cells between the two strains. Future experiments addressing the levels of pro-inflammatory cytokines during the cuprizone treatment period and functional interference with specific cells types such as ablation/inhibition of microglia and/or astrocytes should help elucidate the role of the neuroinflammatory response in contributing to the strain differences observed.

We noted that in our experiment, whereas MAG reduction is significantly less in the CD1 strain (26% reduction in CD1 compared to 73% in C57BL/6), MBP reduction upon cuprizone treatment is not significantly different between the two strains. Given that MAG molecules are localized exclusively in the periaxonal regions of the myelin sheath [35, 36] and that MBP is distributed throughout [37], a relatively abundant MAG level might indicate a relatively intact myelin-axon interface in the CD1 mice, albeit a damaged outer myelin sheath. A similar analysis of myelin protein reduction following cuprizone lesion in other strains that are less vulnerable might be able to address whether this is also the case under other genetic backgrounds.

In summary, we addressed the important role of genetic background in affecting the demyelination response induced by the neurotoxin cuprizone by comparing CD1 strain versus the commonly used C57BL/6 strain, further extending the current literature on strain differences in the cuprizone model. CD1 mice appear to show relatively mild demyelination and a blunted immune response compared to C57BL/6 mice after the same dosage of cuprizone treatment. Future study of the effects of cuprizone on transcription of myelin proteins, the subcellular morphology of oligodendrocyte and production of growth factors and inflammatory cytokines within CD1 mice might help to elucidate the mechanistic differences underlying strain differences. In turn, this might also help in understanding the wide heterogeneity observed in human MS.