Chemokines are small chemotactic cytokines that modulate leukocyte recruitment and activation during inflammation. Here, we describe the role of macrophage inflammatory protein-1α (MIP-1α) during cuprizone intoxication, a model where demyelination of the CNS features a large accumulation of microglia/macrophage without T cell involvement or blood-brain barrier disruption. RNase protection assays showed that mRNA for numerous chemokines were up-regulated during cuprizone treatment in wild-type, C57BL/6 mice. RANTES, inflammatory protein-10, and monocyte chemoattractant protein-1 showed greatest expression with initiation of insult at 1–2 wk of treatment, whereas MIP-1α and β increased later at 4–5 wk, coincident with peak demyelination and cellular accumulation. The function of MIP-1α during demyelination was tested in vivo by exposing MIP-1α knockout mice (MIP-1α−/−) to cuprizone and comparing pathology to wild-type mice. Demyelination at 3.5 wk of treatment was significantly decreased in MIP-1α−/− mice (∼36% reduction), a result confirmed by morphology at the electron microscopic level. The delay in demyelination was correlated to apparent decreases in microglia/macrophage and astrocyte accumulation and in TNF-α protein levels. It was possible that larger effects of the MIP-1α deficiency were being masked by other redundant chemokines. Indeed, RNase protection assays revealed increased expression of several chemokine transcripts in both untreated and cuprizone-treated MIP-1α−/− mice. Nonetheless, despite this possible compensation, our studies show the importance of MIP-1α in demyelination in the CNS and highlight its effect, particularly on cellular recruitment and cytokine regulation.
Chemokines are a rapidly expanding superfamily of small chemoattractant cytokines (recently reviewed in Refs. 1, 2). Over 40 chemokines have been described, and they are divided into four families (CC, CXC, CX3C, and C) founded primarily on their chromosomal location and the position of two N-terminal cysteine residues. The CXC family is further divided into the glutamic acid-leucine-arginine (ELR)3 and non-ELR subfamilies based on the presence of a 3-aa motif. In addition to their chemoattractant function, chemokines can mediate many different cellular responses including proliferation, phagocytosis, cytokine secretion, apoptosis, degranulation, histamine release, and NO synthesis (3, 4, 5).
Recent focus has turned to the role of chemokines in the CNS (recently reviewed in Refs. 3, 6). Chemokines have been implicated in a variety of normal CNS functions (i.e., neuronal progenitor migration, axon guidance and adhesion, oligodendrocyte proliferation, and intercellular communication) although much more evidence supports their role in CNS disease and injury (recently reviewed in Refs. 7, 8). Chemokines are up-regulated in Alzheimer’s disease, Behcet’s disease, HIV-1-associated dementia, human T cell leukemia virus-1-associated myelopathy, brain injury, and in many animal models, including experimental brain abscess development, spinal cord injury, bacterial meningitis, and mechanical, chemical, and freeze injury (7, 9). In particular, chemokines are induced in the demyelinating disease multiple sclerosis, and numerous animal models for demyelination including experimental autoimmune encephalomyelitis (EAE), Theiler’s and hepatitis virus-induced demyelinating disease, experimental autoimmune neuritis, and twitcher, a murine model of globoid cell leukodystrophy (7, 10, 11, 12, 13).
Not only does the up-regulation of many chemokines and their receptors correlate with disease, but disruption of the chemokine-chemokine receptor system has been successful in attenuating CNS disease and pathology. Administration of antisera to RANTES attenuated both macrophage infiltration and demyelination in mouse hepatitis virus-infected mice (14). In experimental autoimmune neuritis, administration of anti-monocyte chemoattractant protein-1 (anti-MCP-1) Abs delayed the onset of disease, whereas anti-macrophage-inflammatory protein-1α (anti-MIP-1α) Abs suppressed the clinical signs as well as inhibited inflammation and demyelination (13). In EAE, although anti-RANTES had no effect on disease, anti-MIP-1α attenuated acute EAE, whereas anti-MCP-1 attenuated relapsing EAE (15, 16, 17). Knockout mice have also proven useful in studying chemokines and their receptors in EAE. Mice deficient in CCR-1, a CCR for MIP-1α, β, and RANTES, and mice deficient in CCR-2, a CC chemokine receptor for MCP-1, are resistant to myelin oligodendrocyte glycoprotein-induced EAE and have decreased numbers of leukocytes recruited into the CNS (18, 19, 20). Interestingly, mice deficient in MIP-1α and mice deficient in CCR-5, another CCR for MIP-1α, β, and RANTES, are fully susceptible to myelin oligodendrocyte glycoprotein-induced EAE (21). This discrepancy is unexplained although it highlights the differences in disease mechanism that underline various models of CNS demyelination.
As mentioned above, many of these previous studies were on EAE in which there is a breakdown of the blood-brain barrier; however, little is known about their function in CNS disorders where the blood-brain barrier remains intact and where T cells are rarely seen, such as in scrapie, Alzheimer’s, Huntington’s, and Parkinson’s disease. The current study characterizes the importance of chemokines in the cuprizone-induced model of demyelination. Cuprizone is a copper chelator shown, upon oral administration, to induce demyelination in selective regions of the CNS, specifically the corpus callosum and superior cerebellar peduncle (22, 23, 24, 25, 26). In C57BL/6 mice, demyelination is initially observed at the light microscopic level at 3 wk of treatment, and progresses rapidly to almost complete demyelination by 4 wk (23, 27). Although the exact mechanism of disease is unknown, cuprizone is believed to metabolically perturb oligodendrocytes, which down-regulate numerous genes, including myelin genes, and are subsequently lost through apoptosis (28, 29, 30, 31, 32, 33, 34). However, unlike EAE, the blood-brain barrier is not compromised (35, 36), and T cells are rarely observed and appear to have no affect on disease.4 In fact, in addition to locally reactive astrocytes, microglia/macrophage are the major responding cell type (23, 24, 25, 26). Although their accumulation parallels demyelination, it is unclear whether microglia/macrophage exacerbate pathology, or what signals are important for their appearance in the demyelinating lesion. Here, we show that numerous chemokines are up-regulated during cuprizone-induced pathology and that the lack of MIP-1α delays demyelination and cellular accumulation.
Materials and Methods
Animals and cuprizone treatment
All animals were housed in a pathogen-free facility and were maintained in accordance with National Institutes of Health guidelines and approved protocols by the University of North Carolina Institutional Animal Care and Use Committee. The generation of mice with a targeted disruption of the gene encoding MIP-1α has been previously characterized and described (37). MIP-1α−/− mice subsequently have been backcrossed to C57BL/6 for 6 generations. C57BL/6 mice were either purchased from The Jackson Laboratory (Bar Harbor, ME) or bred and maintained in the University of North Carolina animal facility. For all experiments, C57BL/6 mice were used as controls instead of littermate siblings, a practice similar to other studies (21, 38, 39, 40, 41). This was primarily due to breeding constraints because cuprizone-induced demyelination has been reported feasible only with male mice (24), and all mice in one experiment are required to be age-matched within 2 wk (G. K. Matsushima, unpublished observations). Therefore, maintaining the MIP-1α−/− line by sibling mating provided suitable numbers for comparable analyses. A scan of the MIP-1α−/− mouse genome by the Speed Congenic Service at The Jackson Laboratory demonstrated the mice to be 98.5% identical with C57BL/6 mice on all unlinked chromosomes. Only the majority of chromosome 11, the chromosome in which the mip-1α gene is located, and a small portion of chromosome 2 (maximally 23 cM in length) was still of 129 origin.
For cuprizone treatment, male mice between 8 and 10 wk of age were fed 0.2% (w/w) cuprizone (Sigma, St. Louis, MO) mixed in ground Breeder Chow 2000 (Purina, Richmond, IN) for up to 6 wk as previously described (23). Animal weights were monitored. Control, untreated mice were maintained on a normal diet for the duration of the experiment.
RNase protection assays (RPAs)
Forebrains from cuprizone-treated and untreated mice were removed, cut mid-sagitally, frozen immediately on dry ice, and stored at −70°C. Total RNA was isolated from half the forebrain using the guanidine isothiocyanate cesium chloride procedure (42). RPAs were performed using the RiboQuant multiprobe RPA system from PharMingen (San Diego, CA). All protocols were performed according to manufacturer’s instructions. For chemokine analysis, 10 μg of total RNA was used with the mCK-5 probe template (PharMingen). For chemokine receptor analysis, 20 μg of total RNA was used with the mCR-5 probe template (PharMingen). Quantitation of transcripts was accomplished using a PhosphorImager (both from Molecular Dynamics, Sunnyvale, CA) and the ImageQuant 5.0 software program. Relative units were normalized to GAPDH for each sample and compared as percentage of maximal expression for each transcript.
For histological examination of demyelination, mice were anesthetized with Metafane (Schering-Plough, Omaha, NE) and perfused intracardially with PBS followed by 4% paraformaldehyde in PBS for 10 min. Brains were removed and postfixed in 4% paraformaldehyde overnight at 4°C. They were then processed and embedded in paraffin. Coronal sections, 5 μm thick, were then cut at the fornix region of the corpus callosum (between regions 220 and 260 in the Mouse Brain Atlas; Ref. 43) and mounted on charged slides (Fisher Scientific, Pittsburgh, PA). Sections were stained with a Luxol fast blue-periodic acid-Schiff (LFB-PAS) stain, and demyelination was scored as previously described (23). Two independent readers in a double-blind fashion evaluated sections midline in the corpus callosum, as illustrated previously (31), and assigned scores ranging from 0 (no demyelination) to 3 (complete demyelination).
For electron microscopic evaluation of demyelination, brains were prepared as previously described (44, 45). Briefly, mice were anesthetized and perfused with PBS followed by 2.5% gluteraldehyde/4% paraformaldehyde in 0.1 M sodium phosphate buffer (pH 7.4). After postfixing in situ for 2 wk, brains were removed and 1-μm coronal sections were cut and stained with toluidine blue. After confirming the appropriate location of the corpus callosum above the fornix, tissues were trimmed and re-embedded in the sagittal orientation. Ultrathin sections were cut, stained with uranyl acetate and lead citrate for contrast, and examined on a transmission electron microscope (model 910; Leo, Thornwood, NY). Five micrographs (at ×5000 magnification) were then randomly taken throughout the corpus callosum of each animal, and the number of myelinated and unmyelinated axons were counted. Each micrograph contained between 200 and 550 axons. Because axons under 0.3 μm in diameter were largely unmyelinated, these were excluded (46).
Chemokine and cytokine ELISAs
Brains from untreated and cuprizone-treated mice were harvested and frozen as described above. The frozen brains were thawed in 1 ml of PBS containing a mixture of protease inhibitors (1 mM PMSF, 2.5 μg/ml aprotinin, and 2.5 μg/ml leupeptin) and homogenized. After pelleting out the larger debris, the protein concentrations of the supernatants were determined using the Bio-Rad protein assay kit (Hercules, CA) with BSA as a standard. A sandwich ELISA was used to determine the levels of MIP-1α, TNF-α, and IL-1β in the brain homogenates as previously described.5
Immunohistochemistry and lectin staining
Paraffin sections were deparaffinized with xylene and rehydrated through a series of graded ethanols. Glial fibrillary acidic protein (GFAP) immunohistochemistry was performed to identify astrocytes as previously described (47). Briefly, after quenching endogenous peroxidase activity with 0.3% H2O2 in methanol for 30 min at room temperature, sections were blocked and permeabilized in 20% normal goat serum (NGS)/0.1% Triton X-100 in PBS for 30 min. Sections were then incubated in a 1/500 dilution of rabbit anti-GFAP Ab (DAKO, Santa Barbara, CA) for 1.5 h. After rinsing, sections were then incubated in a 1/250 dilution of biotinylated goat anti-rabbit Ig Ab (PharMingen) in a 2% NGS/0.1% Triton X-100/PBS for 1 h. The Ricinus communis agglutinin-1 (RCA-1) lectin (Sigma) stain was performed to identify microglia/macrophages as previously described (23). Briefly, endogenous peroxidase activity was first quenched with 3% H2O2 in methanol for 5 min. Sections were then unmasked with a protease digestion (0.025% protease for 2 min at 43°C), blocked with a 5% BSA/ 0.1% Triton X-100/PBS solution, and incubated in biotinylated RCA-1 (Sigma) diluted 1/500 in 5% BSA/0.1% Triton X-100/PBS for 45 min at 43°C. Both stains were then visualized using the Vectastain ABC Elite reagent and diaminobenzidine tetrahydrochloride peroxidase substrate kit (Vector Laboratories, Burlingame, CA). Nuclei were stained with Gill’s Formula Hematoxylin (Vector Laboratories). Slides were then dehydrated and mounted with Permount (Fisher Scientific). Immunohistochemistry for the Pi isoform of GST (GST-Pi) was performed to identify mature oligodendrocytes as previously described (31). Briefly, deparaffinized and rehydrated sections were first blocked with 2% NGS/0.1% Triton X-100 in PBS and then unmasked with 0.1% trypsin in a 0.5 M Tris solution for 15 min at 37°C. The tissue was then incubated in anti-GST-Pi Ab (Biotrin, Newton, MA) diluted 1/1000 in 2% NGS/0.1% Triton X-100 in PBS overnight at 4°C. The sections were then rinsed and incubated in a 1/100 dilution of fluorescein-conjugated goat anti-rabbit IgG Ab (Vector Laboratories). They were then rinsed and mounted in Vectashield with 4′,6-diamidino-2-phenylindole (Vector Laboratories) to visualize nuclei.
For cellular quantification, two fields on either side of midline in the corpus callosum were examined from each animal. Analyses were performed using digitized images and Image-Pro Plus software (Media Cybernetics, Silver Spring, MD) to calculate counts and areas. Quantification of cells included only those in which a nucleus was colocalized with immunohistochemical staining.
All statistical analyses were performed using a one- or two-way ANOVA test.
Chemokines and chemokine receptors are up-regulated in the brains of cuprizone-treated mice
The dramatic accumulation of cells during cuprizone-induced demyelination (23) suggested that chemokines might be critical to the pathology. RPA on brain RNA samples from cuprizone-treated and untreated C57BL/6 mice revealed very dynamic regulation of chemokines (Fig. 1⇓A). No chemokines were detected in the brains of normal, untreated mice. However, upon exposure to cuprizone, mRNA for RANTES, MIP-1β, MIP-1α, inflammatory protein-10 (IP-10), and MCP-1 can all be detected by 1 wk, before any demyelination is observed. IP-10, MCP-1, and RANTES RNA peaked early at 1 and 2 wk, coincident with the initiation of insult. In contrast, MIP-1α and MIP-1β RNA showed maximal expression at 4–5 wk of treatment, concurrent with peak cellular accumulation and demyelination. Lymphotactin (Ltn), eotaxin, MIP-2, and T cell activation-3 (TCA-3) were not present and therefore appear not to be involved in the demyelination process. Peak expression of MIP-1α at 4–5 wk of treatment was also confirmed with RT-PCR, using GADPH as the housekeeping gene to show equal loading (data not shown). This correlated well with ELISA for MIP-1α protein, which detected expression in the brain extracts of 3-, 4-, and 5-wk-treated mice (Fig. 1⇓D). Thus, the temporal expression of specific chemokines appeared to coincide with the previously observed accumulation of microglia/macrophage and astrocytes during demyelination (23, 31).
RPA was also used to determine chemokine receptor mRNA levels in the brains of C57BL/6 mice during cuprizone treatment. CCR-5, a CCR for MIP-1α, MIP-1β, and RANTES, was detected in normal, untreated mice (wk 0), but showed a significant increase at 4–5 wk of cuprizone exposure (Fig. 1⇑B). By 5 wk of treatment, quantitation using a phosphor imager and normalizing to GAPDH showed that there was an ∼2-fold increase in CCR-5 (data not shown). Other receptors tested, such as CCR-1, CXCR-4, and CCR-8, were constitutively found in the brains of normal untreated mice, but did not appear to fluctuate with disease progression (Fig. 1⇑, B and C). CXCR-4 is expressed on neurons, microglia, astrocytes, and endothelium (48), likely accounting for its high transcript level. The combination of the induction of MIP-1α and its receptor, CCR-5, at the height of cellular accumulation and demyelination prompted further investigation into its possible functional role.
The absence of MIP-1α delays demyelination during cuprizone treatment
To test the function of MIP-1α in vivo, MIP-1α−/− mice were treated with 0.2% cuprizone, and demyelination was compared with cuprizone-treated wild-type C57BL/6 mice. Brain sections from treated and untreated mice were stained with an LFB-PAS stain, which stains myelin blue (Fig. 2⇓, A–C), and were scored for demyelination (Fig. 2⇓D). MIP-1α−/− mice showed significantly less demyelination (36% reduction) at 3.5 wk of treatment (p < 0.001). Similarly, they still showed a trend of less demyelination at 3.75 wk; however, demyelination in MIP-1α−/− mice was eventually similar to wild-type mice by 4 wk. There was no difference in weights between the MIP-1α−/− and wild-type mice during the entire time course of cuprizone treatment (data not shown), implying comparable consumption of the drug.
Because the difference in demyelination might be due to a delay in the loss of myelin-producing oligodendrocytes in the MIP-1α−/− mice, mature oligodendrocytes were identified using an anti-GST-Pi Ab and counted (Fig. 2⇑E). Only a slight increase in the number of GST-Pi+ oligodendrocytes was observed at 3.5 wk in the corpus callosum of MIP-1α−/− mice, although this difference was not statistically significant. However, as with other oligodendrocyte genes, particularly myelin genes (32), GST-Pi has been shown to be down-regulated by cuprizone treatment (33); therefore, it is possible that mature oligodendrocytes are still present but no longer adequately stain for GST-Pi in these samples.
Further corroboration for the difference in demyelination was performed using electron microscopic analysis. The number of myelinated and unmyelinated axons in sagittal ultra-thin sections of the corpus callosum were counted from the brains of untreated mice and mice exposed to 3 wk of cuprizone. This earlier time point was taken because the sensitivity of transmission electron microscopy allows detection of demyelination earlier than at the light microscopic level (46). Whereas wild-type mice showed significant demyelination by 3 wk of cuprizone exposure when compared with untreated wild-type mice, MIP-1α−/− mice showed little or no demyelination by this time point (Fig. 3⇓, A to B and C to D, respectively). Indeed, at 3 wk of treatment, the percentage of myelinated axons was significantly greater (∼39%) in the MIP-1α−/− mice when compared with the wild-type mice (p < 0.03) (Fig. 3⇓E), confirming that the lack of MIP-1α delays demyelination.
The absence of MIP-1α affects the recruitment of microglia/macrophages and astrocytes during cuprizone treatment
The function of MIP-1α is often associated with its chemotactic properties, particularly in the recruitment of microglia/macrophage. We assessed cellular accumulation in MIP-1α−/− and wild-type mice during cuprizone treatment. Nuclei counts in the corpus callosum of cuprizone-treated MIP-1α−/− mice showed a trend toward decreased cellularity at 3.5 and 4 wk of treatment, although this difference was not statistically significant at any time point (Fig. 4⇓A). In vitro analyses of macrophage, microglia, and astrocytes have revealed that MIP-1α can induce the chemotaxis of all these cell types (49, 50, 51, 52, 53). Therefore, in this study, the number of microglia/macrophages was assessed in the corpus callosum of cuprizone-treated MIP-1α−/− and wild-type mice. At 3.5 wk of cuprizone exposure, decreases in both microglia/macrophage (RCA-1+ cells; 28% reduction; p < 0.03) and astrocyte (GFAP+ cells; 24% reduction; p < 0.02) populations were observed in MIP-1α−/− mice (Fig. 4⇓, B and C, respectively). This correlated well with the reduction of demyelination and implies that MIP-1α may be exacerbating the pathology, possibly through the recruitment of these cell types.
The absence of MIP-1α decreases TNF-α production during cuprizone treatment
Because chemokines are also implicated in modulating cellular activation, it was possible that the absence of MIP-1α might be affecting cytokine secretion. Indeed, MIP-1α peptide has been shown to cause the secretion of TNF-α, IL-1α, and IL-6 in macrophages in vitro (54). The levels of TNF-α protein in the brains of cuprizone-treated MIP-1α−/− and wild-type mice were determined by ELISA (Fig. 5⇓A). MIP-1α−/− mice showed significantly decreased levels of TNF-α when analyzed by two-way ANOVA (genotype effect; p = 0.042). Although the decrease in TNF-α was distributed across all the time points, it was particularly evident at 4 wk of cuprizone treatment where there was an ∼22% reduction in TNF-α levels in the MIP-1α−/− mice.
A second proinflammatory cytokine, IL-1β, is also up-regulated in wild-type mice treated with cuprizone.5 However, no significant differences were observed in the IL-1β protein levels between MIP-1α−/− and wild-type mice, and, in fact, levels tended to be slightly elevated in the MIP-1α−/− mice at all time points (Fig. 5⇑B). Although IL-1β has many proinflammatory properties and has been shown to be detrimental to mature oligodendrocytes in vitro, we believe that it may also have neuroprotective effects through its induction of insulin-like growth factor-1, a growth factor shown to prevent oligodendrocyte apoptosis in the cuprizone model (45, 55). However, the exact function of IL-1β during cuprizone exposure has yet to be fully elucidated.
Despite a delay in demyelination in MIP-1α−/− mice during cuprizone treatment, other redundant chemokines may be compensating for the lack of MIP-1α
Because of the redundant and overlapping function in many of the chemokines and their receptors, it was possible that other related chemokines were compensating for the lack of MIP-1α in the MIP-1α−/− mice. To test this, RPA for chemokines was performed on RNA from the brains of cuprizone-treated and untreated MIP-1α−/− and wild-type mice. Surprisingly, untreated MIP-1α−/− mice showed an aberrant constitutive production in MIP-1β (5-fold increase) and TCA-3 (20-fold increase) transcripts (Fig. 6⇓). However, these levels did not increase upon cuprizone exposure. In addition, MIP-1α−/− mice showed an increase in mRNA levels of RANTES (2 wk, p < 0.02), IP-10 (1 wk, p < 0.05; 2 wk, p < 0.002), and MCP-1 (1 wk, p < 0.004; 2 wk, p < 0.001) upon cuprizone exposure when compared with wild-type levels (Fig. 6⇓B). At 1 wk of treatment, MCP-1 and IP-10 were both increased ∼3-fold in MIP-1α−/− mice. Although RANTES was elevated at 1 wk, it was not statistically significant. However, RANTES was increased ∼2-fold at 2 wk in MIP-1α−/− mice. Thus, other chemokines may be compensating for the absence of MIP-1α, making it difficult to assess the true magnitude of MIP-1α in the recruitment and activation of cells.
Chemokines are up-regulated in numerous human CNS diseases and animal CNS disease models, and likely mediate cell recruitment and activation. We have shown that chemokines are also up-regulated in the cuprizone-induced model of demyelination, a model, unlike EAE, in which the blood-brain barrier remains intact (35, 36) and T cells are not involved.4 MIP-1α, in particular, appears to play a key role because its absence delays demyelination and cellular accumulation, as well as decreases TNF-α production. This adds to a growing body of evidence that modulation of the chemokine network may prove useful in the treatment of CNS disorders.
The induction of chemokines in the CNS after exposure to cuprizone shows a very specific and dynamic pattern of expression. Expression of RANTES, MIP-1α, MIP-1β, IP-10, and MCP-1 transcripts are all observed after only 1 wk of cuprizone exposure (Fig. 1⇑A). This time point is coincident with the appearance of a few microglia/macrophage and with the perturbation of oligodendrocytes, as indicated by the down-regulation of numerous oligodendrocyte genes (30, 32, 34); yet, it precedes oligodendrocyte apoptosis, demyelination, and massive cellular recruitment (23, 31). Therefore, chemokine production appears to be one of the earliest responses to cuprizone toxicity and implies that chemokines may be key mediators of inflammation and cellular recruitment in the demyelinating lesion. RANTES, IP-10, and MCP-1 mRNA levels then proceed to crest early at 1–2 wk of cuprizone treatment; however, this is before significant cellular recruitment. In contrast, MIP-1α and β levels peak at 4–5 wk of treatment, concurrent with maximal demyelination and cellular recruitment. Therefore, the interplay between each chemokine appears complicated, and these varied expression profiles may reflect specific, temporally restricted functions. Alternatively, it may simply reflect their expression by different cellular populations.
It has been hypothesized that the chemokine profile largely regulates which cell type is recruited during inflammation. This is consistent with our data, where all the chemokines induced (MIP-1α, MIP-1β, MCP-1, IP-10, and RANTES) have been implicated in the chemotaxis of macrophages (56, 57) and microglia (50, 51, 53), the predominant cell types recruited into the demyelinating lesion during cuprizone treatment (23). MIP-2, lymphotactin, eotaxin, and TCA-3 were not observed (Fig. 1⇑A). Interestingly, most of these chemokines are implicated primarily in the recruitment of cell types not seen or rarely seen in the brain during cuprizone treatment. MIP-2 is implicated in neutrophil recruitment, lymphotactin in lymphocyte and NK cell recruitment and eotaxin in eosinophil migration (1). Although TCA-3 is a non-ELR CC chemokine and has been shown to induce macrophage chemotaxis (58), it is thought to be primarily a T cell chemoattractant (1).
The disruption of chemokines and their receptors by genetic knockout or neutralizing Ab in other models of CNS demyelination, including mouse hepatitis virus-infected mice (14) and EAE (15, 16, 17), have resulted in attenuated CNS pathology. Our model of demyelination is distinct because of the presence of an intact blood-brain barrier (35, 36) and the fact that T cells are not involved.4 It was unclear how chemokines would function in this context. In the cuprizone model, we show in this report that deficiency of MIP-1α clearly delays demyelination, a fact confirmed at the light and electron microscopic levels. The mechanism of this delay may be partly explained by reduced microglia/macrophages in the absence of MIP-1α. This reduction in microglia/macrophage numbers at 3.5 wk of cuprizone treatment could account for the decrease in demyelination by limiting the harmful effects that the microglia/macrophages might have, such as the secretion of damaging cytokines and proinflammatory mediators like NO (59, 60). Chemokines themselves have been shown to directly induce cytokine production (54, 61, 62) and NO synthesis (5, 63). Indeed, we saw significantly diminished TNF-α levels in the absence of MIP-1α (Fig. 5⇑A), although it remains unclear whether MIP-1α is directly or indirectly inducing TNF-α in the brain during cuprizone exposure.
Although approximately one-third in number compared with microglia/macrophages (23), astrocytes also accumulate in the demyelinating lesion after cuprizone exposure. The recruitment of astrocytes in response to chemokines is controversial because this cell type has long been thought of as a relatively nonmotile parenchymal cell (64). Astrocytes have been shown to migrate in vitro in response to chemokine gradients of KC, eotaxin, or MIP-1α (49, 52, 64). At 3.5 wk of cuprizone treatment, the number of astrocytes in the MIP-1α−/− mice was reduced when compared with wild-type mice (Fig. 4⇑C). Because astrocytes may exacerbate pathology through the secretion of damaging cytokines and proinflammatory mediators (65), it is possible that the reduction in astrocyte numbers contributes in part to the delay in demyelination in the absence of MIP-1α. However, it is unclear whether MIP-1α truly serves to recruit astrocytes into the demyelinating lesion and, if so, what is the source of the astrocytes. It is known that a number of astrocytes are local and proliferate as indicated by uptake of BrdU (data not shown).
MIP-1α−/− mice displayed decreased demyelination and cellular accumulation at 3.5 wk of cuprizone exposure; however, at 4 wk of treatment, MIP-1α−/− mice exhibited a similar score in demyelination to wild-type mice. A delay rather than a total abrogation was observed. A greater difference in demyelination was not observed possibly due to the aberrantly increased expression of known microglia/macrophage chemoattractants, MIP-1β, RANTES, IP-10, and MCP-1 (Fig. 6⇑) in the MIP-1α−/− mouse, which could be compensating for the absence of MIP-1α. A further examination of the protein levels of these chemokines and in vivo experiments using knockout mice or neutralizing Abs would be required to test whether these chemokines play a functional role in recruitment. It is also possible that an intact blood-brain barrier may be proficient to prevent infiltration of cells by these chemokines. Lastly, the overexpression of MIP-1β, IP-10, RANTES, and MCP-1 in the MIP-1α−/− mouse may suggest that MIP-1α normally suppresses these chemokines in vivo; however, we found no evidence for down-regulation by MIP-1α in the activation of macrophages in vitro (data not shown).
In addition to its role in exacerbating inflammation and demyelination during cuprizone treatment, the possibility also exists that MIP-1α could serve a beneficial role by recruiting microglia/macrophages that clear debris, phagocytose dead cells, and secrete cytokines and growth factors that promote remyelination (27). However, our initial experiments indicate that MIP-1α likely does not have a role in the remyelination process because MIP-1α−/− mice remyelinate as quickly as wild-type mice following 6 wk of cuprizone treatment (data not shown). Therefore, the role of MIP-1α appears to be restricted to the demyelination process.
In summary, specific and dynamic regulation of numerous chemokines correlates with cellular and morphological profiles observed during cuprizone-induced demyelination. The in vivo importance of MIP-1α is shown by the significant reduction in demyelination in MIP-1α−/− mice at 3.5 wk of cuprizone treatment, which is coincident with reductions in microglia/macrophage and astrocyte numbers, and a diminution in TNF-α protein levels. The reduction of cellular recruitment and demyelination may be even greater if it were not for the expression of potentially compensatory chemokines in the MIP-1α−/− mice. Additional crosses to other knockout mice may be fruitful in further attenuating demyelination and cellular recruitment. The results in this report may have implications toward other CNS diseases in which there is microglia/macrophage recruitment in the presence of an intact blood-brain barrier.
We thank Robert Bagnell and Victoria J. Madden for their help with the microscopy and Charita P. Langaman for preparation of tissues for electron microscopy studies. We also thank the Neurodevelopmental Disorder Research Center Morphology Core Facility for processing tissues for paraffin embedding.
↵1 This work was supported by grants from the National Multiple Sclerosis Society (RG2754B to G.K.M.), the National Institute for Dental and Craniofacial Research (P60 DE13079 to G.K.M.), and the National Institutes of Health (NS35372 to G.K.M. and NS24453 to K.S.).
↵2 Address correspondence and reprint requests to Dr. Glenn K. Matsushima, University of North Carolina-Neuroscience Center, CB 7250, Chapel Hill, NC 27599. E-mail address:
↵3 Abbreviations used in this paper: ELR, glutamic acid-leucine-arginine; MIP-1α, macrophage-inflammatory protein-1α; IP-10, inflammatory protein-10; EAE, experimental autoimmune encephalomyelitis; LFB-PAS, Luxol fast blue-periodic acid-Schiff; GFAP, glial fibrillary acidic protein; NGS, normal goat serum; GST-Pi, the Pi isoform of glutathione-S-transferase; RCA-1, Ricinus communis agglutinin-1; MCP-1, monocyte chemoattractant protein-1; RPA, RNase protection assay; m, murine; TCA-3, T cell activation-3.
↵4 M. M. Hiremath, K. Suzuki, J. P.-Y. Ting, and G. K. Matsushima. T cells are not involved in microglia-mediated demyelination during cuprizone-intoxication. Submitted for publication.
↵5 M. M. Hiremath, K. Suzuki, J. P.-Y. Ting, and G. K. Matsushima. MHC class II exacerbates demyelination in vivo independently of T cells. Submitted for publication.
- Received January 22, 2001.
- Accepted June 18, 2001.
- Copyright © 2001 by The American Association of Immunologists