Acta Neuropathologica

, Volume 112, Issue 2, pp 195–204

Chemokine expression by astrocytes plays a role in microglia/macrophage activation and subsequent neurodegeneration in secondary progressive multiple sclerosis

Authors

  • Naoyuki Tanuma
    • Department of Molecular NeuropathologyTokyo Metropolitan Institute for Neuroscience
  • Hiroshi Sakuma
    • Department of Molecular NeuropathologyTokyo Metropolitan Institute for Neuroscience
  • Atsushi Sasaki
    • Department of Human PathologyGunma University Graduate School of Medicine
    • Department of Molecular NeuropathologyTokyo Metropolitan Institute for Neuroscience
Original Paper

DOI: 10.1007/s00401-006-0083-7

Cite this article as:
Tanuma, N., Sakuma, H., Sasaki, A. et al. Acta Neuropathol (2006) 112: 195. doi:10.1007/s00401-006-0083-7

Abstract

The pathological hallmarks of secondary progressive (SP) multiple sclerosis (MS) include slowly expanding demyelination and axonal damage with less inflammation. To elucidate the pathomechanisms of secondary progressive (SP) multiple sclerosis (MS), we have investigated the expression of chemokines, chemokine receptors, matrix metalloproteinase-9 (MMP-9) and immunoglobulins in the demyelinating plaques. Immunohistochemical analysis revealed that numerous hypertrophic astrocytes were observed at the rim, but not in the center, of the chronic active lesions. Microglia/macrophages phagocytosing myelin debris were also found at the lesion border. In contrast, T cell infiltration was minimal in these plaques. Characteristically, at the rim of the lesions, there were abundant immunoreactivities for monocyte chemoattractant protein-1 (MCP-1)/CCL2 and interferon-γ inducible protein-10 (IP-10)/CXCL10 and their receptors, CCR2 and CXCR3, while these immunoreactivities were weak in the center, thus forming a chemokine gradient. Double immunofluorescense staining demonstrated that cellular sources of MCP-1/CCL2 and IP-10/CXCL10 were hypertrophic astrocytes and that both astrocytes and microglia/macrophages expressed CCR2 and CXCR3. MMP-9 was also present at the rim of the lesions. These results suggest that MCP-1/CCL2 and IP-10/CXCL10 produced by astrocytes may activate astrocytes in an autocrine or paracrine manner and direct reactive gliosis followed by migration and activation of microglia/macrophages as effector cells in demyelinating lesions. Targeting chemokines in SPMS may therefore be a powerful therapeutic approach to inhibit lesional expansion.

Keywords

Multiple sclerosisProgressionImmunohistochemistryChemokine

Introduction

Multiple sclerosis (MS) is an inflammatory disease of the central nervous system (CNS) that causes demyelinating plaques with glial scar formation. Recent studies have demonstrated that MS is a disease with heterogeneous pathogenetic mechanisms in terms of clinical course, neuropathological and neuroradiological appearance of the lesions and response to therapy [20, 28]. In most patients with MS, the disease begins at about 30 years of age with episodes of acute worsening of neurologic function, followed by a variable degree of recovery between relapses, the relapsing–remitting (RR) phase of the disease. Moreover, in about half of the patients, the clinical course changes from a RR to a secondary progressive (SP) MS after 10 year and almost 90% by 25 years [28]. This shift is a serious problem because once secondary progression begins, patients appear to progress at a uniform rate regardless of how long the disease is present. SPMS responds poorly to medications which are effective in RRMS [7]. Based on pathological examination, it is generally accepted that the pathogenesis of MS consists of inflammatory and neurodegenerative phases [39]. While an inflammatory process plays a central role in reversible demyelination in RRMS, non-remitting clinical disability and disease progression in SPMS is due to degeneration of both the myelin sheath and the axon [4, 5, 40, 46]. The extent of axonal damage correlated with the numbers of macrophages/microglia [17], suggesting that these cells or their toxic products such as tumor necrosis factor-α (TNF-α), nitric oxide (NO) and matrix metalloproteinases (MMPs) are main effectors in this process [3, 9, 16]. Axons can also be damaged by antibody-mediated destruction via Fc or complement receptors [13, 32, 33]. However, the pathogenesis of neurodegeneration in MS is still a matter of debate.

Chemokines, chemotactic cytokines that selectively recruit specific subsets of leukocytes into tissues, are essential for inflammatory responses [21, 47]. So far, the role of these chemokines in MS has been mainly discussed in terms of inflammatory recruitment of leukocytes into the CNS [15, 37, 41]. For example, interferon-γ inducible protein (IP)-10/CXCL10 is produced by astrocytes and their receptor CXCR3-positive leukocytes migrate into the lesion of MS [2, 36, 38]. Monocyte chemoattractant protein (MCP)-1/CCL2 is also expressed by astrocytes and plays a role in the recruitment and activation of myelin degrading macrophages [26, 35, 42]. However, the effects of these chemokines in the neurodegenerative phase on brain parenchymal cells, in particular astrocytes and microglia, remains to be elucidated.

The purpose of the present study was to analyze in more detail the immunopathological features of SPMS in order to identify factors that are involved in neurodegeneration because the results obtained provide useful information to develop specific immunotherapy against this form of MS. Here, we have investigated the expression of chemokines, MCP-1/CCL2 and IP-10/CXCL10, and their receptor, CCR2 and CXCR3, in the brain from patients with SPMS. Consequently, we found at the rim of the plaques with ongoing demyelination that MCP-1/CCL2 and IP-10/CXCL10 produced by astrocytes and that CCR2 and CXCR3 were abundantly expressed by astrocytes and microglia/macrophages. In contrast, there were weak immunoreactivities for these chemokines and their receptors in the center of the lesions. Thus, along this chemokine gradient, microglia/macrophages expressing CCR2 and CXCR3 were activated. In addition, MMP-9 that is reported to be involved in the lesion formation in MS [6, 8] was predominantly expressed at the rim of the plaque. These findings suggest that chemokine expression by astrocytes plays an important role in activation of microglia/macrophages and expansion of demyelinating lesions in SPMS.

Materials and methods

Patients and tissue samples

The present study was approved by Ethics Committee of Tokyo Metropolitan Institute for Neuroscience and performed on postmortem brain tissues from the UK Multiple Sclerosis Tissue Bank and Department of Human Pathology, Gunma University Graduate School of Medicine. Clinical details of each patient are given in Table 1. Twenty-six tissue blocks from MS patients were studied. Control sections came from two blocks of normal appearing white matter (NAWM) of MS patients (MS58 and MS63 in Table 1) and two biopsied brain tissues, which showed no pathological findings.
Table 1

Summary of cases utilized for immunohistochemistry and histopathological findings of brain samples

Patients

Diagnosis

Age (years)/sex

Disease duration (years)

Postmortem interval (h)

Number of blocks examined

Number of MS lesions

Lesional activitya

Active

Chronic active

Chronic inactive

MS38

SPMS

42/F

18

21

6

6

1

2

3

MS53

SPMS

66/M

34

26

1

3

  

3

MS58

SPMS

51/F

21

15

3

4

  

4

MS60

SPMS

55/M

43

16

3

3

 

1

2

MS63

SPMS

66/F

30

13

3

2

  

2

MS74

SPMS

64/F

36

7

2

3

 

2

1

MS79

SPMS

49/F

21

7

2

3

1

2

 

MS80

SPMS

71/F

34

24

1

1

 

1

 

MS88

SPMS

54/F

20

22

3

4

  

4

MS114

SPMS

52/F

15

12

1

1

  

1

MS122

SPMS

44/M

12

16

1

1

 

1

 

05-02

Normalb

28/F

NA

Biopsy

1

NA

NA

NA

NA

05-30

Normalb

61/M

NA

Biopsy

1

NA

NA

NA

NA

Total

    

28

31

2

9

20

F female; M male; SPMS secondary progressive multiple sclerosis; NA not applicable

aLesional activity was determined by histological examination and divided into three categories (active, chronic active and chronic inactive) as described in the Materials and methods

bAlthough this specimen was biopsied on suspicion of brain tumor, the result was negative

Histopathology and classification of multiple sclerosis lesions

Frozen sections, 10 μm thick, were stained with hematoxylin and eosin (HE) and with Luxol Fast Blue (LFB) (Kluever–Barrera’s method) for determination of the cell morphology and myelin distribution, respectively. We also used Oil Red O staining for neutral lipid to identify myelin breakdown products. The multiple sclerosis lesions were divided into three categories on the base of the staging system described by De Groot et al. [10] with a few modifications [19]. The detailed classification of lesions was as follows: (1) active, demyelinating lesion with abundant phagocytic macrophages containing myelin components or neutral lipids; perivascular lymphocytes are present; hypertrophic astrocytes are distributed throughout the demyelinated regions; (2) chronic active, hypocellular center contains a few macrophages with some residual lipids; lymphocytes are present in the perivascular cuffs; hypercellular rim contains perivascular and parenchymal (foamy) macrophages and hypertrophic astrocytes; (3) chronic inactive, hypocellular lesion, usually containing isomorphic gliosis filling up the demyelinated region with widened extracellular spaces; perivascular and parenchymal phagocytic macrophages are not detectable.

Immunohistochemistry

A single immunoperoxidase staining was performed as previously described [23, 30]. In brief, frozen sections were mounted on slides, air dried, and fixed in ether for 10 min at room temperature. After washing with 0.01 mol/l phosphate-buffered saline (PBS, pH7.4), all slides were incubated overnight with primary antibody (Ab) at 4°C and incubated with horseradish peroxidase (HRP)-labeled secondary Ab (horse anti-goat IgG, Vector, Burlingame, CA) at room temperature for 45 min. The following primary antibodies were used: goat anti-MCP-1/CCL2, goat anti-IP-10/CXCL10, goat anti-CCR2, goat anti-CXCR3, goat anti-MMP-9 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), rabbit anti-glial fibrillary acidic protein (GFAP), mouse anti-human HLA-DP, DQ, DR (CR3/43, DAKO Japan, Kyoto), mouse anti-CD3 (OKT3, purified from the hybridoma supernatant) Abs. After washing with PBS, sections were incubated with biotinyl tyramide (1:1,000) [1] for 10 min to amplify the immunoreactivity for chemokines (MCP-1/CCL2 and IP-10/CXCL10) or their receptors (CCR2 and CXCR3). We used biotinylated anti-human IgG for detection of IgG deposition on the brain tissue. For other immunostainings, sections were incubated with biotinylated anti-mouse or anti-rabbit IgG (Vector, Burlingame, CA) as secondary Ab. Then, slides were incubated with HRP-labeled avidin–biotin complex (ABC), using VECTSTAIN Elite ABC Kit (Vector). HRP binding sites were detected in 0.005% 3,3-diaminobenzidine and 0.01% hydrogen peroxide. To confirm the specificity of the staining, the primary antibodies were omitted or replaced with normal goat serum or normal mouse IgG. In addition, antibodies against these chemokines/chemokine receptors were absorbed with blocking peptides (Santa Cruz Biotechnology, Inc.,). These controls did not show any specific staining.

Confocal microscopy

To identify the types of cells expressing MCP-1/CCL2, IP-10/CXCL10, CCR2 and CXCR3, double-label immunohistochemistry with GFAP or major histocompatibility complex (MHC) class II molecules was performed. Briefly, frozen sections were air dried and fixed as described above. After washing, sections were incubated overnight at 4°C with the first primary antibody (MCP-1/CCL2, IP-10/CXCL10, CCR2 or CXCR3) and incubated with HRP-labeled horse anti-goat IgG, (Vector) at room temperature for 45 min. After a 10-min application of biotinyl tyramide, the HRP signal was visualized with Alexa Fluor® 488-conjugated streptoavidin (Molecular Probes, Eugene, OR). After washing with PBS, slides were incubated with the second primary antibody (anti-GFAP or anti-HLA DP, DQ, DR) overnight at 4°C and the signal was detected by a 1-h incubation with rhodamine-conjugated anti-rabbit IgG or Cy3-conjugated anti-mouse IgG (Amersham Life Science, Tokyo, Japan). The observation was made using a confocal microscope TCS-SP (Leica, Heidelberg, Germany).

Quantitative analysis of chemokine- and chemokine receptor-positive cells in MS lesions

Quantitative analysis of chemokine- and chemokine receptor-positive cells in MS lesions was performed by two independent observers (N. Tanuma and H. Sakuma). Appropriate areas of the sections were selected according to the demyelinating activity within the lesions. The number of MCP-1/CCL2-, IP-10/CXCL10-, CCR2- and CXCR3-positive cells was determined at least in three standardized microscopic fields of 62,500 μm2 (defined by a morphometric grid) from each of the distinct lesional areas. In the text and figures, the mean number of cells per mm2 is given.

Statistics

Data were analyzed with the Student’s t-test or the non-parametric Mann–Whitney U test for two-group comparisons. Reported P values were two-tailed and considered significant at P value less than 0.05.

Results

We examined 28 brain tissue blocks from 11 MS patients and 2 normal controls (Table 1). Two (MS58 and MS63) of these blocks from MS patients had the NAWM defined as an area, far away from the lesion, which showed no signs of demyelination by histology and the others contained at least one demyelinating lesion. Among 31 lesions of MS, two lesions were classified as active with abundant perivascular lymphocytes, phagocytic macrophages and hypertrophic astrocytes throughout the lesions. Nine lesions were identified as chronic active with hypercellular rims and hypocellular central lesions. The other 20 lesions were classified as chronic inactive, as they were hypocellular without perivascular and parenchymal phagocytic macrophages.

Astrocytes and microglia are activated at the rim of the plaque in chronic active MS

We first characterized the nature of the chronic active lesions in their center, rim and periplaque white matter in patients with SPMS. Representative results are shown in Fig. 1. Consistent with HE and LFB staining, lipid droplets were few in number in the center (Fig. 1a). In sharp contrast, there were numerous lipid-laden foamy macrophages at the rim of the lesions (Fig. 1b), suggesting that active demyelination takes place at this site. In the periplaque area, virtually no lipid deposition was observed (Fig. 1c). In GFAP staining, numerous hypertrophic astrocytes were detected at the rim of the chronic active lesion (Fig. 1e). These astrocytes were also observed both in the center of the lesion and in the periplaque white matter, but were smaller in number and size (Fig. 1d, f). At the rim of the plaque, lipid-laden foamy macrophages were positive for MHC class II antigens (Fig. 1h). MHC class II-positive microgila/macrophages were also detected in the center of the plaque although their shapes were not foamy but more concentrated (Fig. 1g). MHC class II-positive “ramified” microglia were observed in the periplaque white matter outside the plaque (Fig. 1i). Thus, MHC class II-positive microglia/macrophages closer to the lesion site were rounder and larger. These results clearly showed that microglial activation was observed mainly at the lesion border. In these plaques, T cell infiltration was minimal and restricted to the perivascular space of a few blood vessels (Fig. 1j–l). Normal control brains and NAWM of SPMS contained no activated microglia, hypertrophic astrocytes and T cell infiltration (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-006-0083-7/MediaObjects/401_2006_83_Fig1_HTML.gif
Fig. 1

Astrocytes and microglia are activated at the rim of the plaque in chronic active MS. Lipid-laden macrophages were abundantly observed at the rim of the lesion b, but occasionally at the center a. No lipid-containing macrophages were detected in periplaque white matters around the lesion c. Numerous GFAP-positive hypertrophic astrocytes were detected at the rim of the lesions e, while immunoreactivity for GFAP was relatively weak at the center of the lesion d and periplaque white matter f. In accordance with reactive astrocytes, MHC class II-positive “foamy” microglia/macrophages were observed at the edge of the chronic active lesion h, whereas its shape was more concentrated at the center of the plaque g. In the periplaque white matter, only MHC class II-positive-“ramified” microglia were observed i. T cell infiltration was minimal and restricted to the perivascular space of a few blood vessels jl. ac Oil Red O staining, df GFAP staining, gi CR3/43 staining, jl OKT3 staining. Scale bar = 50 μm

The expression of chemokines and their receptors in different areas of MS lesions

We next analyzed the distribution of MCP-1/CCL2-, IP-10/CXCL10-, CCR2- and CXCR3-immunoreactive cells in the center and rim of the lesions, periplaque white matter and NAWM. The results from chronic active lesions are shown in Fig. 2. MCP-1/CCL2- and IP-10/CXCL10-positive cells were predominantly observed at the rim of the plaques in chronic active lesions rather than in the center (Fig. 2a, c). These immunoreactive cells possessed morphological features of hypertrophic astrocytes. Similarly, the corresponding chemokine receptors, CCR2- and CXCR3-positive cells were seen mainly at the rim of the plaques (Fig. 2b, d). Although these plaques contained some perivascular cuffs, all these lymphocytes were negative for CXCR3 (data not shown). In the periplaque white matter outside of the plaques, immunoreactivity for CCR2 (Fig. 2b) was still upregulated whereas there were no immunoreactivities except for MCP-1-positive (Fig. 2a) and CXCR3-positive (Fig. 2d) endothelial cells in these lesions. In NAWM, immunoreactivities for MCP-1/CCL2 and CXCR3 were positive in endothelial cells (Fig. 2a,d). We also examined the expression of chemokines and their receptors in normal brains. Normal brains did not express MCP-1/CCL2, IP-10/CXCL10, CCR2 and CXCR3 (Fig. 2e), while immunoreactivities for MCP-1/CCL2 and CXCR3 were positive in endothelial cells in NAWM (Fig. 2a, d). Quantitative analysis of chemokine- and chemokine receptor-positive cells clearly demonstrated that the numbers of all the positive cells at the rim were greater than that in the center of chronic active lesions (Fig. 3). The number of MCP-1/CCL2- (Fig. 3a), CCR2- (Fig. 3b), IP-10/CXCL10- (Fig. 3c) and CXCR3- (Fig. 3d) positive cells at the rim of the lesions were significantly larger than that in the center of the lesions. These findings strongly suggest that astrocytes produce MCP-1/CCL2 and IP-10/CXCL10 mainly at the lesion borders, forming a “chemokine gradient” from the center to rim within the plaque in SPMS.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-006-0083-7/MediaObjects/401_2006_83_Fig2_HTML.gif
Fig. 2

Immunohistochemistry for MCP-1/CCL2 a, CCR2 b, IP-10/CXCL10 c and CXCR3 d in the center, rim, periplaque of chronic active MS lesion, normal appearing white matter (NAWM) and normal control brain e. There were abundant immunoreactivities for these chemokines and their receptors at the rim of the lesions with ongoing demyelination, while immunoreactivities for all chemokines and their receptors examined were relatively low at the center of the lesions. In periplaque white matter around the lesion, immunoreactivity for CCR2 b was still upregulated. These immunoreactive cells possessed morphological features of astrocytes. In addition, immunohistochemistry for MCP-1/CCL2 a, CCR2 b and CXCR3 d showed staining of the blood vessel endothelium. MCP-1/CCL2 and CXCR3 were constitutively expressed on the blood vessel endothelium in NAWM. Normal control brain showed no expression of MCP-1/CCL2, IP-10/CXCL10, CCR2 and CXCR3 e. Scale bar = 50 μm

https://static-content.springer.com/image/art%3A10.1007%2Fs00401-006-0083-7/MediaObjects/401_2006_83_Fig3_HTML.gif
Fig. 3

Quantification of MCP-1/CCL2- a, CCR2- b, IP-10/CXCL10- c and CXCR3- d positive cells in chronic active MS lesion. The number of positive cells was determined in at least three standardized fields of 62,500 μm2 (defined by a morphometric grid) from each of the distinct lesional areas. Data are expressed as mean ± SD (cells/mm2). The number of all chemokine- and chemokine receptor-positive cells at the rim of the lesion was significantly higher than that at the center of the lesion

Table 2 summarizes the results of chemokine and chemokine receptor staining of the active, chronic active and chronic inactive lesions. In the chronic inactive lesion, there was a hypocellular center where none of cells expressed MCP-1/CCL2, CCR2, IP-10/CXCL10 and CXCR3. At the rim of these lesions, there were a few MCP-1/CCL2- and CXCR3-positive cells. In contrast to chronic active lesions, there were no IP-10/CXCL10- and CCR2-positive cells in chronic inactive lesions (Table 2).
Table 2

Expression of chemokines and chemokine receptors in MS lesions

Chemokines and chemokine receptors

Activea–chronic activeb

Chronic inactivec

Center

Rim

Periplaque

Center

Rim

Periplaque

MCP-1

++

+++

+

+

+

IP-10

++

+++

CCR2

++

+++

+

CXCR3

++

+++

+

+

+

− = no immunoreactive cells, + = sporadic or a few positive cells, ++ = moderate positive cells in the same area, +++ = abundant immunoreactivity

aActive is defined as the demyelinating lesion with abundant phagocytic macrophages containing myelin components or neutral lipids; perivascular lymphocytes are present; hypertrophic astrocytes are distributed throughout the demyelinated regions

bChronic active is defined as the lesion with hypocellular center containing a few macrophages with some residual lipids; lymphocytes are present in the perivascular cuffs; hypercellular rim contains perivascular and parenchymal (foamy) macrophages and hypertrophic astrocytes

cChronic inactive is defined as the hypocellular lesion, usually containing isomorphic gliosis filling up the demyelinated region with widened extracellular spaces; perivascular and parenchymal phagocytic macrophages are not detectable

Astrocytes express both chemokines and chemokine receptors but microglia express only chemokine receptors in the chronic active MS lesions

To determine the type of cells that produce MCP-1/CCL2 and IP-10/CXCL10 and cells expressing their receptor, CCR2 and CXCR3, we performed double immunofluorescense staining with astrocyte and microglia/macrophage markers. The results are shown in Fig. 4. As clearly demonstrated in Fig. 4a–l, GFAP-positive astrocytes expressed both MCP-1/CCL2 (Fig. 4c, arrows) and IP-10/CXCL10 (Fig. 4i, arrow heads), whereas CR3/43-positive microglia/macrophages were completely negative for these chemokines (Fig. 4f, l). With regard to chemokine receptors (Fig. 4m–x), both CCR2 and CXCR3 were expressed on astrocytes (Fig. 4o, u) and microglia/macrophages (Fig. 4r, x). These findings strongly suggest that hypertrophic astrocytes produce MCP-1/CCL2 and IP-10/CXCL10 and also express their receptors, while microglia/macrophages do not produce these chemokines and only express the receptors.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-006-0083-7/MediaObjects/401_2006_83_Fig4_HTML.jpg
Fig. 4

Identification of the type of cells expressing chemokines (al) and chemokine receptors (mx). Double immunofluorescense staining for MCP-1/CCL2 (green) (a, d), IP-10/CXCL10 (green) (g, j), CCR2 (M, P) and CXCR3 (green) (s, v) with GFAP (red) (b, h, n and t) and MHC class II, CR3/43 (red) (e, k, q and w) in lesions from patients with SPMS. c, f, i and l are merged images of a, b; d, e; g, h and j, k, respectively. Both MCP-1/CCL2 and IP-10/CXCL10 were expressed by astrocytes (carrows and iarrow heads), but was not expressed by microglia (f and l). o, r, u and x are merged images of m, n; p, q;s, t, and v, w, respectively. CCR2 and CXCR3 were expressed by both astrocytes (o, u) and microglia (r, x). Scale bar = 50 μm

MMP-9 are predominantly expressed at the rim of the lesion

MMPs are considered to be one of the toxic substances for demyelination and axonal injury [8, 9, 12, 22, 43]. To determine the involvement of MMPs in microglia/macrophages-mediated demyelination and axonal damage, we examined the localization of MMP-9 in the MS lesion, because MMP-9 plays a role in blood–brain barrier (BBB) breakdown, myelin degradation and axonal damage [6, 27, 34]. We found that MMP-9 was predominantly expressed at the rim of the plaques in chronic active lesions (Fig. 5b) rather than in the center (Fig. 5a). In the periplaque white matter outside of the plaques, immunoreactivity for MMP-9 was weakly detected (Fig. 5c). We also determined the IgG deposition in the lesion. Unlike MMP-9, IgG immunoreactivities were present in both the center and the rim of the plaque (Fig. 5d, e). In the periplaque white matter, faint to moderate staining was observed around blood vessels (Fig. 5f). Absorption of anti-MMP-9 and anti-human IgG with MMP-9 peptide and human sera, respectively, resulted in negative staining, indicating that these immunoreactivities were specific (data not shown). These results suggest that activated microglia might produce MMP-9 at the rim of the plaque and contribute to slowly progressive expanding lesions.
https://static-content.springer.com/image/art%3A10.1007%2Fs00401-006-0083-7/MediaObjects/401_2006_83_Fig5_HTML.gif
Fig. 5

Immunohistochemistry for MMP-9 (ac) and IgG (df) in the center (a, d), rim (b, e) and periplaque (c, f) of chronic active MS lesion. The expression of MMP-9 was predominantly observed at the rim b rather than in the center a. In the periplaque white matter outside of the plaques, immunoreactivity for MMP-9 was weakly detected c. In contrast, both the center and the rim of the plaque showed moderately intense immunoreactivity for IgG (d, e). In periplaque white matter, faint to moderately intense staining was observed around blood vessels f. Scale bar = 50 μm

Discussion

SPMS is characterized by irreversible neurological impairment without remission. Pathologically, SPMS has a slowly expanding demyelinating lesion with only modest inflammatory cell cuffing. This ongoing demyelinating lesion, namely, “progressive plaque” includes linear group of microglia engaging short segments of disrupted myelin that are associated with deposits of C3d [32]. Axonal damage is also considered to be a major cause of secondary progression with irreversible neurological impairment [4, 11, 40]. Previous studies suggested that not only T cells [17] but also microglia/macrophages and their toxic products such as tumor necrosis factor-α (TNF-α), nitric oxide (NO) and MMPs are main effectors in this process [3, 8, 16]. However, the mechanisms of microglia/macrophage activation and subsequent toxic substance production in “progressive” or chronic active plaques remain unknown.

In the present study, we have investigated the expression of chemokines, MCP-1/CCL2 and IP-10/CXCL10, and their receptor, CCR2 and CXCR3, in the brain of patients with SPMS to elucidate the pathomechanisms of secondary progression of the MS lesion where lymphocyte infiltration is minimal or absent. As clearly shown here, there were abundant immunoreactivities for MCP-1/CCL2, IP-10/CXCL10, CCR2 and CXCR3 at the rim of the lesions with ongoing demyelination, whereas these immunoreactivities were relatively weak in the center of the lesions, forming a chemokine gradient. Double immunofluorescense staining revealed that both chemokines and their receptors were expressed by hypertrophic astrocytes, while only chemokine receptors were expressed by MHC class II-positive microglia/macrophages. It should be emphasized that there was no evidence of chemokine production by microglia/macrophages. Although it was stated in a previous report [35] that MCP-1/CCL2 was expressed by astrocytes and macrophages within acute MS lesions, we demonstrated using double immunofluorescense staining that microglia/macrophages were always negative in both active and chronic active lesions. In addition, the previous studies demonstrated that CXCR3 is expressed by perivascular lymphocytes mainly in acute MS and/or relapsing–remitting MS [36, 37]. However, our study revealed that CXCR3 was not expressed by perivascular lymphocytes in chronic MS lesions, especially with minimal lymphocyte infiltration. Our result suggests that T lymphocytes are not crucial any longer in this chronic stage. The results obtained in the present study also suggest that chemokines produced by astrocytes play a role in astrocytic migration and proliferation in an autocrine or paracrine manner, resulting in reactive astrogliosis observed at the rim of the plaque in SPMS. It is likely that CCR2- or CXCR3-positeve microglia can also migrate into the rim in response to MCP-1/CCL2 and IP-10/CXCL10 produced by astrocytes. Finally, as shown here, activated microglia may secrete toxic substances such as MMP-9, one of final effectors for demyelination [8] and axonal injury [12, 27]. Recently, Omari et al. [31] reported that chemokines produced by astrocytes induces the recruitment of oligodendrocytes which expressed different CXC chemokine receptors and subsequent remyelination in MS. Thus, chemokines play an important role, not only in the leukocytes recruitment into the inflammatory site, but also in the recruitment of brain parenchymal cells, in particular astrocytes, microglia and oligodendrocytes.

We were also interested in the distribution of IgG in and around the plaque because IgGs, especially autoantibodies to myelin oligodendrocyte glycoprotein (MOG), play an important role in demyelination under a certain circumstance [13, 20, 33]. However, the recent study by Lalive et al. [18] demonstrated that native MOG-specific IgGs, which induce demyelination were most frequently found in serum of clinically isolated syndromes and RRMS, while only marginally in SPMS, suggesting that native MOG-specific IgGs may be implicated in the early pathogenesis of MS. In this study, we found that moderately intense immunoreactivities for IgG were evenly distributed both in the center and at the rim of the plaque. O’Connor et al. [29] demonstrated that the majority of CNS samples from patients with MS yielded substantially greater quantities of IgG than those from control cases by eluting IgG from plaque tissue. In addition, they also demonstrated that IgG derived from MS plaque tissue contains autoantibodies that recognize the folded (native) human MOG proteins. Thus, antibodies from inflamed CNS tissue recognize MOG. Although we did not examine the distribution of MOG-specific IgG and further investigation will be required, our results suggest that anti-myelin antibodies may not be involved in microglial activation and MMP-9 secretion at the rim of the plaque at the later disease stage. Prineas et al. [32] showed that the ongoing demyelinating lesions in SPMS include C3d deposition but this is an antibody-independent process. Taken together, it is likely that Ab-mediated activation of microglia may be less implicated in the mechanisms for lesion expansion in SPMS.

Recently, evidence is accumulating that chemokines also play an important role, not only in autoimmune diseases, but also in neurodegenerative disorders. In amyotrophic lateral sclerosis (ALS), the MCP-1/CCL2 level in cerebrospinal fluid was significantly increased compared to the control subjects [14, 44]. In ALS spinal cord tissue, MCP-1/CCL2 protein was expressed in glia, probably astrocytes [14]. Thus, MCP-1/CCL2 may recruit both astrocytes and microglia into the lesion site and promote neurodegeneration. These results together with our findings suggest that the neurodegenerative phase of MS and of neurodegenerative disorders such as ALS have common disease processes in the lesion formation. These findings also suggest that chemokines and chemokine receptors could be targets of therapies. We recently demonstrated that gene therapy with decoy chemokine receptor DNAs encoding the binding sites of the CCR2 and CXCR3 molecules prevent the disease progression of experimental autoimmune myocarditis and subsequent dilated cardiomyopathy [25]. In addition, chronic relapsing experimental autoimmune encephalomyelitis (EAE), which was recently produced by us, was effectively inhibited by anti-macrophage migration therapy using decoy chemokine and chemokine receptor DNA therapy [24]. It is also shown that deletion of macrophage-inflammatory protein (MIP)-1α/CCL3 retards neurodegeneration in mice with Sandhoff disease, a lysosomal storage disorder caused by a deficiency of β-hexosaminidases A and B [45]. The result indicates that the pathogenesis of Sandhoff disease involves an increase in MIP-1α/CCL3 that induces monocytes to infiltrate the CNS, expand the activated microglia/macrophages, and trigger apoptosis of neurons. Together, chemokines could be possible targets for therapies to slow the neurodegenerative phase in autoimmune and degenerative disorders.

In summary, we have investigated the expression of chemokines and chemokine receptors in the brain from patients with SPMS and found that MCP-1/CCL2, IP-10/CXCL10 produced by astrocytes and their receptors, CCR2, CXCR3, were abundantly expressed by astrocytes and microglia/macrophages at the rim of the lesions with ongoing demyelination. Our findings suggest that these chemokine–chemokine receptor pairs may contribute to a neurodegenerative phase of MS through recruitment of astrocytes and microglia into the site of expanding lesions. Targeting chemokines in a neurodegenerative phase of MS may therefore be a possible therapeutic approach to inhibit lesional expansion.

Acknowledgments

The authors thank Yoko Kawazoe for technical assistance. MS brain tissue samples were supplied by the UK Multiple Sclerosis Tissue Bank. This work was supported in part by grants from the Ministry of Education, Culture, Sports, Science and Technology, Japan.

Copyright information

© Springer-Verlag 2006