Introduction

Multiple sclerosis (MS) is the most common neurological disorder in young adults in the Western world. In addition to the well-characterized inflammatory white matter, histopathological and imaging studies have suggested that the so-called ‘normal-appearing white matter’ (NAWM) has significant abnormalities [1]. Axonal injury, blood–brain barrier changes and reactive astro- and microgliosis have been reported at the histological level [2–4]. Little is known about the molecular changes taking place in the NAWM in MS. It has recently be suggested that NAWM in MS is in a subtle balance between inflammation and neuroprotection [5]. Therefore, one might expect that the understanding of molecular changes within the NAWM could be exploited to develop new treatments.

In this study, we used the well-characterized toxic cuprizone model. Cuprizone ingestion in mice induces a highly reproducible demyelination of distinct brain regions, such as the corpus callosum (CC) [6]. Some regions, however, appear to be resistant to the toxin challenge, e.g. the fimbria of the hippocampus [7]. The aim of the study was (1) to define whether the fimbria fulfils histological criteria of the NAWM in an established disease model and (2) to get a better understanding of ‘NAWM-lesion’ development in this particular animal model.

Materials and methods

Research and animal care procedures were approved by the Review Board for the Care of Animal Subjects of the district government (North Rhine-Westphalia, Germany). Demyelination was induced by feeding 8-week-old male mice (C57BL/6; 19–21 g) a diet containing 0.2% cuprizone (bis-cyclohexanone oxaldihydrazone; Sigma-Aldrich Inc., USA) mixed into a ground standard rodent chow for the indicated periods. Tissue preparation and immunohistochemistry (IHC) were performed as published in detail previously [7, 8]. Anti-proteolipid protein (PLP) antibodies (mouse IgG, 1:500; Serotec, Germany) were used to detect myelin, anti-Iba1 antibodies (rabbit IgG, 1:250; Wako, Germany) to detect microglia/macrophages, and anti-glial fibrillary acidic protein (GFAP) antibodies (rabbit IgG, 1:1,000; EnCor, USA) were used to visualize astrocytes. Anti-olig2 antibodies (Millipore, Germany) were used to detect oligodendrocytes.

Results

In a first set of experiments, animals (n = 5 per experimental group) were fed cuprizone for 5 weeks to induce acute demyelination of the CC. Definite loss of PLP immunoreactivity could be observed in the CC (Fig. 1b/e), whereas other regions appeared unaffected, such as the fimbria of the hippocampus (Fig. 1c/f). Both regions included in the subsequent study are indicated by the arrow and star in Fig. 1d. Quantitative measurement of PLP staining intensity confirmed our observational results. There was a significant decrease in PLP staining intensity in the CC (control 69.2 ± 6.8 [mean ± SEM] vs. 5 weeks cuprizone 19.4 ± 4.4; P < 0.005). This was not observed in the fimbria (control 65.3 ± 5.1 vs. 5 weeks cuprizone 61.0.4 ± 3.2; P = 0.49). Luxol fast blue staining showed similar results (data not shown). To further investigate both regions on the cellular level, we visualized microglia by anti-Iba1 and (activated) astrocytes by anti-GFAP antibodies, respectively. The quantification of fimbrial Iba1+ and GFAP+ cell numbers after acute cuprizone-induced demyelination is presented in Fig. 1u. Microglia were uniformly distributed in the CC (Fig. 1g) and fimbria (Fig. 1i) of control animals while a sustained increase in Iba1+ cell numbers was present in the demyelinated CC (Fig. 1h) and, although to a lesser extend, in the fimbria (Fig. 1j). After 5 weeks cuprizone, higher numbers of GFAP+ astrocytes compared to controls were seen within the CC (Fig. 1k/l) but not in the fimbria (Fig. 1m/n).

Fig. 1
figure 1

Myelination and inflammatory cell composition in the medial corpus callosum (CC) and fimbrial region of the hippocampus. Anti-PLP antibodies were used to visualize intact myelin in sections of control and cuprizone-treated animals (a–f). Anti-Iba1 and anti-GFAP antibodies were used to visualize microglia or astrocytes, respectively. Arrowhead in (d) indicates the medial part of the CC, which is shown in (b/e, g/h, k/l and o) in higher magnification. The star in (d) indicates the fimbrial region of the hippocampus, which is shown in (c/f, i/j, m/n, p and q-t) in higher magnification. Pictures (o/p) show hematoxylin-and-eosin-stained sections, illustrating apoptotic oligodendrocytes (arrows) within the CC and fimbria. Quantification of Iba1+ microglia and GFAP+ astrocytes (fimbria) in control and 5 weeks’ cuprizone-treated animals is given in (u). ***P < 0.001; scale bars: 500 μm (a/d); 50 μm (b/c, e/f, g–n); 10 μm (o/p); 25 μm (q–t)

Full size image

From our observation that definite microgliosis can be observed within the fimbrial region after acute cuprizone-induced demyelination, we assumed that the fimbria is affected like the CC by the toxin per se, but does not develop to a ‘classically’ demyelinated lesion. To test for this hypothesis, animals (n = 6 per experimental group) were fed cuprizone in a second experiment up to 5 weeks and sacrificed on day 2, week 1, week 2 and week 5 after the start of the cuprizone diet. Cells with morphological characteristics typical for apoptosis such as condensed and/or fragmented nuclei were seen at week 1 in both regions included in the study (Fig. 1o/p). Control animals fed with powdered food without cuprizone were lacking apoptotic cells (data not shown). In a previous study, immunohistochemistry for detection of CNPase confirmed that these apoptotic cells are oligodendrocytes [9]. Anti-Iba-1 IHC revealed that pronounced microgliosis is evident in both regions after week 1. In control animals, Iba1+ cells displayed a typical ramified morphology indicating a resting state (Fig. 1q). At week 1 (Fig. 1s) and week 2 (Fig. 1t), processes of microglia cells within the fimbria appeared to be retracted and cell bodies were swollen, which is frequently associated with microglia activation [10]. Comparable morphological changes of microglia were evident in the CC at weeks 1 and 2. While the magnitude of microgliosis progressed in the CC until week 5 (Fig. 1h), a reduction of microglia cell numbers could be observed in the fimbria. Only a moderate activation of fimbrial astrocytes was detectable at week 2 (increased number and swollen processes), whereas at all other time points in the study, astrocytic cell parameters were the same as observed in the fimbria of control animals. Within the CC, astrocytosis started at week 2; however, in contrast to the fimbria, it progressed in the CC until week 5. At week 2, the number of olig2-positive cells was dramatically decreased in the CC whereas no difference in oligodendrocyte cell numbers (and PLP staining intensity) was evident in the fimbria compared to control animals.

Discussion

Activation of microglia is a well-known characteristic histological feature of the NAWM in MS [2, 11]. Whether activated microglia are beneficial or detrimental for disease progression is controversial [11, 12]. In this study, we have shown that the fimbria of the hippocampus, which shows characteristic features of NAWM after acute cuprizone-induced demyelination, is not protected from the toxic compound per se. Both regions, the fimbria and the CC, [13] demonstrated early oligodendrocyte apoptosis and intense microglia accumulation and activation. However, only in the CC did we observe lesions that progressed to actively demyelination lesions, whereas such a course was not seen in the fimbria. Remyelination appears not to be the underlying mechanism of preserved myelination in the fimbria, since at all time points in the study anti-PLP staining intensity remained unchanged in this region.

It might be feasible that regional differences exist in microglia function, as previously suggested [14]. We assume that some intrinsic signaling contributes to a ‘switch’ from an anti-inflammatory to a pro-inflammatory phenotype of microglia within the affected CC. The former phenotype is also called ‘M1 or classically activated’, the latter one ‘M2 or alternatively activated’ [13, 15]. Therapeutic bolstering of this apparent regulation could therefore still be effective in halting the formation of new destructive lesions in human MS. In this situation, prevention of further damage rather than repair of existing damage could still be achieved.

Summary

The applied animal model with an existing regional diversity in the course of disease appears to be suitable for elucidating pathways which promote progression of the NAWM to an active lesion. We have focused on the fimbrial region of the hippocampus, since this brain region can readily be dissected without the need of costly equipment (such as laser-capture microscopy) for subsequent analyses on the molecular level.