European Journal of Neurology
Volume 8 Issue 6 Page 665 - November 2001
T. L. Sørensen, F. Sellebjerg, C. V. Jensen, R. M. Strieter & R. M. Ransohoff
Studies of chemokines in cerebrospinal fluid (CSF) of patients with active multiple sclerosis (MS) have indicated that specific chemokines may have important roles in disease pathogenesis. We previously reported that CSF concentrations of CXCL10 (previously known as IP-10) were elevated in MS patients in relapse, whilst levels of CCL2 (MCP-1) were reduced. Here, we report a serial analysis of CSF CXCL10 and CCL2 concentrations in 22 patients with attacks of MS or acute optic neuritis (ON) treated with methylprednisolone, and 26 patients treated with placebo in two randomized controlled trials. Chemokine concentrations were measured by enzyme linked immunosorbent assay (ELISA) in CSF obtained at baseline and after 3 weeks, and were compared with other measures of intrathecal inflammation. At baseline CSF concentrations of CCL2 were significantly lower in the patient group than in controls. The levels of CXCL10 were higher in the patient group than in controls but two outliers in the control group also had high CSF concentrations of CXCL10. The CSF concentrations of CXCL10 did not change over time or after treatment. The CSF concentration of CXCL10 was positively correlated with the CSF leukocyte count, the CSF concentration of neopterin, matrix metalloproteinase (MMP)-9, and intrathecal IgG and IgM synthesis. The concentration of CCL2 increased between baseline for 3 weeks in both groups, more distinctly so in patients treated with methylprednisolone. CCL2 correlated negatively with MMP-9 and IgG synthesis levels. CXCL10 may be involved in the maintenance of intrathecal inflammation whereas CCL2 correlates negatively with measures of inflammation, suggesting differential involvement of CXCL10 and CCL2 in CNS inflammation.
Chemokines are members of the cytokine family that act on leukocytes during physiological trafficking to skin, gut, and lymphoid tissues as well as during inflammation (Baggiolini, 1997; Luster, 1998). The chemokine system consists of a superfamily of related molecules that are divided into four subfamilies based on specific structural motifs. Many different cell types throughout the body produce chemokines, which act through receptors that are differentially expressed on leukocyte subtypes. Some chemokine receptors are expressed on only one leukocyte type and bind only one specific chemokine, others show marked redundancy in expression and ligand binding (Rollins, 1997). Recently a new nomenclature for chemokines was introduced (Zlotnik and Yoshie, 2000).
Multiple sclerosis (MS) is an inflammatory disease of the central nervous system (CNS) (Steinman, 1996; Noseworthy et al., 2000). Prevailing models of the pathogenesis of MS suggest that systemic activation of myelin-reactive T cells promotes their migration across the blood-brain barrier (BBB) where they induce disease by orchestrating inflammatory leukocyte recruitment and activation, with ensuing demyelination and axonal loss (Sørensen and Ransohoff, 1998). The process of leukocyte extravasation into the CNS involves several steps beginning with weak adhesion and rolling on the endothelium of the BBB, followed by firm arrest on the luminal side of the endothelium and subsequent diapedesis across the BBB (von Andrian and Mackay, 2000).
Chemokines are involved in several of these steps: induction and activation of leukocyte adhesion molecules that mediate firm adhesion to the endothelium, establishment of a chemotactic concentration gradient resulting in recruitment of the cell across the endothelial monolayer, and the induction of proteolytic enzymes involved in the breakdown of extracellular matrix proteins (Ransohoff, 1999).
Studies in experimental autoimmune encephalomyelitis (EAE), a well-characterized animal model of T-cell dependent inflammatory demyelination, have shown that specific chemokines are involved in leukocyte recruitment to the CNS. Furthermore, different chemokines are differentially expressed during clinical disease (Godiska et al., 1995; Karpus and Ransohoff, 1998). The chemokines CCL5 (RANTES), CXCL10 (IP-10), CCL3 (MIP-1a), and CCL2 (MCP-1) exhibit different expression patterns through disease development (Godiska et al., 1995; Karpus and Ransohoff, 1998). These descriptive findings were extended by the observations that anti-CCL3 antibodies suppressed initial attacks of adoptive transfer EAE in SJL mice (Karpus et al., 1995). Similarly, CXCL10 antisense oligodeoxynucleotides and antibodies to CXCL10, a potent chemoattractant for activated T-cells suppressed initial attacks of EAE (W.J. Karpus, personal communication). The role in EAE for CCL2, a chemoattractant for monocytes and T-cells, was potentially more complex. Anti-CCL2 blocked relapses of chronic EAE and disease could not be induced in mice that lacked CCR2, the major CCL2 receptor on monocytes and activated T-cells (Karpus and Kennedy, 1997; Fife et al., 2000; Izikson et al., 2000). However, CCL2 was also involved in the generation T helper 2 (Th2) responses in mice and was implicated in oral tolerance, a procedure that suppresses EAE (Karpus et al., 1997).
Relatively few studies have addressed chemokine expression in MS. We recently reported that CXCL10 was present in higher concentrations in CSF from patients with active MS than in cerebrospinal fluid (CSF) from neurological control subjects with non-inflammatory conditions (Sørensen et al., 1999). More than 90% of CSF T-cells expressed the CXCL10 receptor CXCR3, significantly more than found in peripheral blood. These observations, consistent with a role for the CXCL10/CXCR3 axis in MS, were supported by immunohistochemical analysis (Balashov et al., 1999; Sørensen et al., 1999; Simpson et al., 2000).
In the same study we found that the CSF concentration of CCL2 was significantly lower in CSF from patients with active MS than in neurological controls (Sørensen et al., 1999). As noted above, CCL2 induces differentiation of T-cells towards a Th2 cytokine secretion pattern, and CCL2 expression is induced by the prototype Th2 cytokine interleukin (IL)-4 and the immunomodulatory cytokine transforming growth factor (TGF)-b (Karpus et al., 1997; Lu et al., 1998; Wie and Berman, 1998; Gu et al., 2000). Reduced levels of CCL2 in the CSF could reflect the cytokine dysregulation associated with attacks of MS, as these attacks are associated with a predominant production of Th1 cytokines. However, CCL2 is expressed in reactive astrocytes in MS plaques (McManus et al., 1998; Simpson et al., 1998).
The current study was undertaken to examine serially the CSF concentrations of these two chemokines during attacks of MS. These data could be used to address whether altered CSF chemokine levels observed during MS attacks reflected temporary changes related to disease activity. This goal was addressed by measuring the CSF concentration of these chemokines in CSF samples from patients with attacks of MS or acute optic neuritis (ON) who underwent lumbar puncture before and after participation in randomized, placebo-controlled trials of oral high-dose methylprednisolone treatment (Sellebjerg et al., 1998a, 1999). This treatment was found to improve the rate of recovery in patients with ON and in attacks of MS. The clinical effects were associated with suppression of intrathecal IgG synthesis whereas there was no persistent effect on the CSF leukocyte count and activity of matrix metalloproteinase (MMP)-9 (Sellebjerg et al., 2000). The CSF concentration of neopterin, which is induced by interferon (IFN)-g, decreased over time both in patients treated with methylprednisolone and in placebo-treated patients whereas there were no significant changes in intrathecal synthesis of IgM.
Materials and methods
Forty-eight patients (36 with clinically definite MS (CDMS), 12 with monosymptomatic ON) were included in the study (Poser et al., 1988). All patients had intrathecal synthesis of IgG oligoclonal bands. The patients had been randomized to treatment with either oral high dose methylprednisolone (n=22) or placebo (n=26), and the treatment was initiated as soon as possible after the onset of symptoms (Sellebjerg et al., 1998a). Methylprednisolone was administered at a dose of 500 mg daily for 5 days followed by a 10-day-tapering course. The patients were graded on the Kurtzke Expanded Disability Status Scale (EDSS) at baseline and after 1, 3 and 8 weeks (Kurtzke, 1983). In 36 patients (nine ON and 27 CDMS) magnetic resonance imaging (MRI) was performed at study entry, and the area of gadolinium enhancing lesions was measured. All patients underwent a lumbar puncture at study entry, i.e. during the treated attack and after 3 weeks, i.e. 1 week after completing treatment with oral high-dose methylprednisolone or placebo. Nineteen patients with non-inflammatory neurological disease (seven with lumbar disc protrusion, five with low back pain, three with tension headache, two with migraine, one with torticollis and one with psychiatric symptoms) served as controls. None of the neurological control subjects had IgG oligoclonal bands or pleocytosis in CSF. The Regional Scientific Ethics Committee approved the study, and written informed consent was obtained from all participating patients.
The CSF was collected directly on ice and centrifuged within 20 min at 250 x g for 10 min. Serum samples were obtained simultaneously with lumbar puncture. The supernatant was quickly removed and stored at -80 °C. CXCL10 and CCL2 in CSF were measured by in-house developed enzyme linked immunosorbent assay (ELISA). The assay variability of the chemokine assays is less than 10%, and the lower detection limit for all chemokines is 0.01 ng/ml (Verma et al., 1997; Sørensen et al., 1999); the CSF concentration of myelin basic protein (MBP) was measured by radioimmunoassay (Sellebjerg et al., 1998b); albumin and IgG were measured by rate nephelometry (Sellebjerg and Christiansen, 1996; Sellebjerg et al., 1996); IgA, and IgM, ELISA (Sellebjerg et al., 1998b); neopterin was measured by ELISA (Sellebjerg et al., 1998a); and MMP-9 was measured by zymography (Gibjels et al., 1992; Sellebjerg et al., 1998b). The CSF/serum albumin concentration quotient was used as a measure of blood-CSF barrier integrity, and intrathecal synthesis of IgA, IgG, and IgM was calculated as previously described in detail (Sellebjerg et al., 1998b).
As none of the data showed the characteristics of normal distribution, values are presented as median values. Non-parametric analysis was employed using the Mann-Whitney test for comparison of unpaired data between groups. The Wilcoxon test was used for comparison of paired data within a group, and Spearman rank correlation analysis was used to analyse the relationships between continuous variables. A 5% significance level was employed.
CSF CXCL10 concentrations in MS patients and controls
The clinical and immunological effects of oral high-dose methylprednisolone treatment and the baseline characteristics of the patients have been reported elsewhere (Sellebjerg et al., 1998a, 1999, 2000). The CSF concentration of CXCL10 was higher at baseline in MS patients (2.6 ng/ml, range 0.3-36.1 ng/ml) than in controls (1.82 ng/ml, range 0-14.3 ng/ml) but in the present study this difference did not reach statistical significance. This discrepancy was caused by the presence in the control group of two outliers with high CSF concentrations of CXCL10 (low back pain 14.3 ng/ml, migraine 11.2 ng/ml). There was no statistical significant difference in CSF CXCL10 concentrations between patients with monosymptomatic ON (2.1 ng/ml, range 0-23 ng/ml) and patients with CDMS (3.3 ng/ml, range 0-36.1 ng/ml) allowing us to analyse them as a single group. Amongst the MS patients, CSF levels of CXCL10 did not change significantly after 3 weeks in either placebo (3.1 ng/ml, range 0.2-13 ng/ml) or methylprednisolone treatment group (3.3 ng/ml, range 0.16-20 ng/ml). To address whether CXCL10 detected in the CSF could originate from the circulation we measured the levels of CXCL10 in 10 randomly selected serum samples. CXCL10 was detected in only three of 10 serum samples, and CSF concentrations of CXCL10 were significantly higher than serum concentrations (CSF; median 3.3 ng/ml, range 1.3-36.4 ng/ml and serum; median 0, range 0-12.3 ng/ml, P < 0.05). This suggests that CXCL10 in CSF is produced intrathecally.
Relationship between CSF CXCL10
concentrations and intrathecal inflammation
(a) CSF CXCL10 concentration (ng/ml) (x-axis) correlation with CSF leukocyte count (y-axis, cells/l) at baseline for all patients (P < 0.01, =0.506);
(b) CSF CXCL10 concentration (ng/ml) (x-axis) correlation with CSF leukocyte count (y-axis, cells/l) after 3 weeks for all patients (P < 0.05, =0.444).
The CSF concentration of CXCL10 correlated significantly with the CSF leukocyte count at baseline (P < 0.01, P=0.506) (including both the treated and un-treated group) and 3 weeks later (P < 0.05, P=0.444) (including both the treated and un-treated group) (Fig. 1). Interestingly this correlation was not observed in the control group, and the two outliers in the control group had normal CSF leukocyte counts. These findings strongly suggest a role for CXCL10 in leukocyte accumulation in the acutely inflamed CNS, but not during physiological trafficking.
At baseline, CXCL10 concentrations in CSF correlated with other measures of intrathecal inflammation: neopterin (P < 0.001, P=0.509), MMP9 (P < 0.001, P=0.55), IgG synthesis (P < 0.001, P=0.504), and IgM synthesis (P < 0.05, P=0.41) (Fig. 2). Significant correlations were also observed after 3 weeks in the placebo group: neopterin (P < 0.001, P=0.479), MMP9 (P < 0.001, P=0.619), and IgG synthesis (P < 0.001, P=0.643); and in the methylprednisolone group: neopterin (P < 0.001, =0.466), MMP9 (P < 0.001, P=0.579), and IgG synthesis (P < 0.001, P=0.577).
CSF CCL2 levels in MS patients
(a) CSF CCL2 concentration (ng/ml) before and after treatment with methylprednisolone (P < 0.01);
(b) CSF CCL2 concentration (ng/ml) before and after treatment with placebo (P=not significant).
There was no difference in CSF CCL2 concentrations between patients with monosymptomatic ON (0.017 ng/ml, range 0-0.29 ng/ml) and patients with CDMS (0.018 ng/ml, range 0-0.376 ng/ml) allowing us to analyse them as a single group. The CSF concentration of CCL2 was significantly lower at baseline in patients (0.018 ng/ml, range 0-0.376 ng/ml) than in neurological controls (0.277 ng/ml, range 0-0.633 ng/ml). After 3 weeks the CSF concentration of CCL2 had increased significantly compared with baseline levels in the methylprednisolone group (0.185 ng/ml P < 0.01, 0.2-0.572 ng/ml) and somewhat but not significantly in the placebo group (0.149 ng/ml, range 0-0.699 ng/ml) (Fig. 3). No differences in baseline value between the treated and the untreated patient groups were observed (untreated, median 0.116 ng/ml, range 0-0.289 ng/ml; treated, 0.1 ng/ml, range 0-0.376 ng/ml). The increase in CCL2 concentrations was significantly higher in patients treated with methylprednisolone than in patients treated with placebo after 3 weeks (P < 0.05). To address whether CCL2 detected in the CSF could originate from the circulation we measured the levels of CCL2 in ten randomly selected serum samples. CCL2 was detected in only one of ten serum samples, and CSF concentrations of CCL2 were significantly higher than serum concentrations (CSF; median 0.148 ng/ml, range 0-0.290 ng/ml and serum; median 0 ng/ml, range 0-0.16 ng/ml, P < 0.01). This suggests that CCL2 in CSF is produced intrathecally.
Relationship between CSF CCL2 levels and intrathecal inflammation
The CSF concentration of CCL2 did not correlate with the CSF leukocyte count at any time point. CCL2 levels at baseline correlated negatively with the CSF concentrations of IgG (P < 0.01, P=-0.409) and MMP9 (P < 0.05, P=-0.325). This correlation was also seen after 3 weeks in the placebo group (IgG: P < 0.05, P=-0.301; MMP9: P < 0.05, P=-0.394), but not in the methylprednisolone group (data not shown), suggesting differential effects of treatment on these variables.
Relationship between CSF chemokine levels, clinical and MRI activity
The median improvement as estimated by the EDSS score was 1 point [interquartile range (IQR) 0-1.5] in patients treated with methylprednisolone and 0 points (IQR 0-0.5) in the placebo group. There was no significant correlation between CSF CXCL10 or CCL2 levels and clinical changes (data not shown). At baseline four of nine ON and 22 of 27 CDMS patients had enhancing lesions on T1-weighted MRI after the administration of Gadolinium (Gd)-DTPA (0.1 mmol/kg), but there was no significant difference in the area of enhancement in ON and CDMS patients (P=0.2, data not shown). There was no correlation between the area of Gd-enhancement and CSF concentrations of CCL2 or CXCL10, regardless of whether it was studied in all 36 patients or only in the 27 patients with CDMS. It was not possible to discriminate between persistently enhancing lesions and newly enhancing lesion as only a single baseline MRI was available.
CXCL10, a chemokine that attracts activated T cells through its receptor, CXCR3, has been previously shown to be present at higher concentrations in the CSF from patients with active demyelinating attacks compared with neurological controls (Sørensen et al., 1999). CXCR3 bearing T cells are enriched in the CSF (compared with peripheral blood) of patients with MS, and CXCR3 positive T cells are readily detected in perivascular cuffs in autopsy material from MS patients (Ransohoff et al., 1993; Balashov et al., 1999; Sørensen et al., 1999). CXCL10 is produced by astrocytes not only in humans but also in mice suffering from EAE (Ransohoff et al., 1993; Balashov et al., 1999; Sørensen et al., 1999).
We found a trend towards higher levels of CXCL10 in the CSF from patients with MS than in controls in the present study but this difference did not reach statistical significance, reflecting the presence of two outliers in the control group. However, we were able to confirm our finding that the CSF level of CXCL10 correlates with the CSF leukocyte count. This correlation was also observed after 3 weeks. Importantly this correlation was not observed in control subjects, indicative of a selective role for CXCL10 in leukocyte accumulation into the CSF during attacks of MS. The level of CXCL10 in the CSF also correlated with other measures of intrathecal inflammation (neopterin, IgG and IgM synthesis, and MMP9). As this correlation was maintained at follow-up we speculate that CXCL10 production in the CNS may control the recruitment of inflammatory cells into the CNS and, hence, the maintenance of CNS inflammation. The correlation between the CSF concentration of CXCL10 and neopterin may also reflect that IFN-g is crucial in the induction of both molecules.
We found no significant change in the CSF concentration of CXCL10 for 1 week after completing a 15-days course of oral high-dose methylprednisolone. This does not exclude an effect during or immediately after treatment. Other effects of methylprednisolone are, however, likely to contribute to the improvement observed after treatment. Indeed, both clinical and MRI measures of disease activity are suppressed at this time point (Sellebjerg et al., 1998a, 1999, 2000; F Sellebjerg, CV Jensen, HN Larsson, JF Frederiksen et al., unpublished observations). These findings are consistent with a key role of CXCL10 in maintaining intrathecal inflammation and may help to explain why oral high-dose methylprednisolone treatment does not influence the risk of recurrent disease activity (Sellebjerg et al., 1998a, 1999). Indeed, treatment with IFN-b results in a decrease in intrathecal inflammation as assessed by the CSF leukocyte count (Rudick, 1999). It is not known if IFN-b affects the CSF concentration of CXCL10.
We have previously shown that CCL2 levels are reduced in the CSF of MS patients as compared with controls with patients with non-inflammatory neurological disease (Sørensen et al., 1999). This finding is confirmed in the present study. We found a significant increase in the CSF concentration of CCL2 after treatment with methylprednisolone and a non-significant increase in patients treated with placebo. As CCL2 is induced by IL-4, a cardinal Th2 cytokine, and because CCL2 is able to prime T cells towards a Th2 phenotype, this finding may be secondary to a Th1 to Th2 shift during remission. Methylprednisolone treatment also induces TGF-b, which, in turn, may induce CCL2 production (Wie and Berman, 1998; Rudick et al., 1999). Alternatively CCL2 may itself skew the immune response in a Th2 direction (Karpus et al., 1997; Lu et al., 1998). We do, indeed, find that CCL2 levels are negatively correlated to measures of intrathecal inflammation such as IgG synthesis and MMP9. After treatment this correlation was not observed. This appears to reflect a differential effect of methylprednisolone treatment on different measures of intrathecal inflammation, i.e. a decrease in intrathecal IgG synthesis, no persistent effect on MMP-9 activity, and an increase in CCL2 levels (Sellebjerg et al., 2000). It remains to be established in future studies whether any of these effects are closely related to the clinical efficacy of treatment than the others.
CCL2 was detected near inflammatory infiltrates in the CNS both in EAE and in histology material from MS patients (Gourmala et al., 1997; McManus et al., 1998; Simpson et al., 1998). Evidence for a non-redundant role for CCL2 in the acute attack of EAE has come from the recent observations that mice lacking CCR2 (the CCL2 receptor) are not susceptible to EAE (Fife et al., 2000; Izikson et al., 2000). These results were interpreted as indicating a detrimental function of CCL2 in MS, although differences between EAE and MS have to be taken into account. The pathogenetic involvement of different chemokines in early relapsing-remitting disease and later chronic progressive stages is a possibility worth considering. Indeed, in vivo neutralization of CCL2 by antibody treatment attenuates disease activity in chronic but not in acute EAE (Karpus et al., 1995; 1997). It is possible that CCR2 ligands other than CCL2 may be involved in the pathogenesis of acute EAE. CCL2 may also have harmful effects during trauma and chronic inflammation, which could be more predominant in the mouse with chronic EAE than in a mouse suffering an initial attack (Glabinski and Ransohoff, 1999; Ransohoff, 1999; Muessel et al., 2000).
In this report we have described the concentrations of two biologically important chemokines in the CSF from patients with MS during relapse and remission. We show that methylprednisolone does not alter CSF CXCL10 levels and associated leukocyte accumulation in the CNS whereas this treatment augments the time dependent increase in the CSF levels of CCL2. Our data also suggest that CCL2 in CSF may reflect resolution of the proinflammatory CNS environment that occurs during relapses. These data suggest that caution must be observed if CCL2/CCR2 antagonists are applied for treatment of MS. In contrast CXCL10/CXCR3 antagonists may attenuate intrathecal inflammation and, hence, clinical disease activity.
This work was supported by P. Carl. Petersen Foundation, Niels Ydes Foundation, and Christian and Ottilia Brorsons Grant and the National Institutes of Health (USA) with grant 1PO1 NS38667 to RMR.
1 von Andrian UH&Mackay IR (2000). T-cell function and migration 2000. New Eng J Med 343:1020-1034.
2 Baggiolini M (1997). Chemokines and leukocyte traffic. Nature 392:565-568.
3 Balashov KE, Rottman JB, Weiner HL, Hancock WW (1999). CCR5+ and CXCR3+ T cells are increased in multiple sclerosis and their ligands MIP-1alpha and IP-10 are expressed in demyelinating lesions. Proc Nat Acad Sci 96:6873-6878.
4 Fife BT, Huffnagle GB, Kuziel WA, Karpus WJ (2000). CC chemokine receptor 2 is critical for induction of experimental autoimmune encephalomyelitis. J Exp Med 92:899-906.
5 Gibjels K, Masure S, Carton H, Opdenakker G (1992). Gelatinase in the cerebrospinal fluid of patients with multiple sclerosis and other inflammatory disorders. J Neuroimmunol 41:29-34.
6 Glabinski AR&Ransohoff RM (1999). Chemokines and chemokine receptors in CNS pathology. J Neurovirol 5:3-12.
7 Godiska R, Chantry D, Dietsch GN, Gray PW (1995). Chemokine expression in murine experimental allergic encephalomyelitis. J Neuroimmunol 58:167-176.
8 Gourmala NG, Buttini M, Limonta S, Sauter A, Boddeke HW (1997). Differential and time-dependant expression of monocyte chemoattractant protein-1-mRNA by astrocytes and macrophages in rat brain: effects of lipopolysaccharide administration. J Neuroimmunol 74:35-44.
9 Gu L, Tseng S, Horner RM, Tam C, Loda M, Rollins B (2000). Control of Th2 polarization by the chemokine monocyte chemoattractant protein-1. Nature 404:407-411.
10 Izikson L, Klein RS, Charo IF, Weiner HL, Luster AD (2000). Resistance to experimental autoimmune encephalomyelitis in mice lacking the CC chemokine receptor (CCR) 2. J Exp Med 192:1075-1080.
11 Karpus WJ&Kennedy KJ (1997). MIP-1alpha and MCP-1 differentially regulate acute and relapsing autoimmune encephalomyelitis as well as Th1/Th2 lymphocyte differentiation. J Leukoc Biol 62:681-687.
12 Karpus WJ, Lukacs NW, McRae BL, Strieter RM, Kunkel SL, Miller SD (1995). 1995 An important role for macrophage inflammatory protein-1alpha in the pathogenesis of the T-cell mediated autoimmune disease, experimental autoimmune encephalomyelitis. J Neuroimmunol 155:5003-5010.
13 Karpus WJ, Lukacks N, Kennedy K, Smith W, Hurst S, Barret T (1997). Differential CC Chemokine-induced enhancement of T helper cell cytokine production. J Immunol 158:4129-4136.
14 Karpus WJ&Ransohoff RM (1998). Chemokine regulation of experimental encephalomyelitis: temporal and spatial expression patterns govern disease pathogenesis. J Immunol 61:2667-2671.
15 Kurtzke JF (1983). Rating neurological impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 33:1444-1452.
16 Lu B, Rutledge BJ, Gu Let al. (1998). Abnormalities in monocyte recruitment and cytokine expression in monocyte chemoattractant protein 1-deficient mice. J Exp Med 187:601-608.
17 Luster A (1998). Chemokines-chemotactic cytokines that mediate inflammation. New Eng J Med 338:436-445.
18 McManus C, Berman JW, Brett FM, Stauton H, Farrel M, Brosnan CF (1998). MCP-1, MCP-2 and MCP-3 expression in multiple sclerosis lesions: an immunohistochemical and in situ hybridization study. J Neuroimmunol 86:20-29.
19 Muessel MJ, Berman NE, Klein RM (2000). Early and specific expression of monocyte chemoattractant protein-1 in the thalamus induced by cortical injury. Brain Res 870:211-221.
20 Noseworthy JH, Lucchinetti C, Rodriguez M, Weinshenker B (2000). Multiple sclerosis. New Eng J Med 343:938-952.
21 Poser CM, Paty DW, Scheinberg Let al. (1988). New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 38:180-185.
22 Ransohoff RM (1999). Mechanisms of inflammation in MS tissue: adhesion molecules and chemokines. J Neuroimmunol 98:57-68.
23 Ransohoff RM, Hamilton HA, Tani Met al. (1993). Astrocyte expression of mRNA encoding cytokines IP-10 and JE/MCP-1 in experimental encephalomyelitis. FASEB J 7:592-602.
24 Rollins B (1997). Chemokines. Blood 90:909-928.
25 Rudick RA (1999). Cerebrospinal fluid abnormalities in a phase III trial of Avonex (IFNbeta-1a) for relapsing multiple sclerosis. The Multiple Sclerosis Collaborative Research Group. J Neuroimmunol 93:8-14.
26 Sellebjerg F&Christiansen M (1996a). Qualitative assessment of intrathecal IgG synthesis by isoelectric focusing and immunodetection: interlaboratory reproducibility and interobserver agreement. Scand J Clin Lab Invest 56:135-143.
27 Sellebjerg F, Christiansen M, Rasmussen LS, Jaliachvili I, Nielsen PM, Frederiksen JL (1996b). The cerebrospinal fluid in multiple sclerosis. Quantitative assessment of intrathecal immunoglobulin synthesis by empirical formulae. Eur J Neurol 3:548-559.
28 Sellebjerg F, Frederiksen JL, Nielsen PM, Olesen J (1998a). Double-blind, randomized, placebo-controlled study of oral, high-dose methylprednisolone in attacks of MS. Neurology 51:529-534.
29 Sellebjerg F, Christiansen M, Nielsen PM, Frederiksen JL (1998b). Cerebrospinal fluid measures of disease activity in multiple sclerosis. Multiple Sclerosis 4:475-479.
30 Sellebjerg F, Nielsen HS, Frederiksen JL, Olesen J (1999). A randomized, controlled trial of oral high-dose methylprednisolone in acute optic neuritis. Neurology 52:1479-1484.
31 Sellebjerg F, Christiansen M, Jensen J, Frederiksen JL (2000). Immunological effects of oral high-dose methylprednisolone in acute optic neuritis and multiple sclerosis. Eur J Neurol 7:281-289.
32 Simpson J, Newcombe J, Cuzner M, Woodroofe M (1998). Expression of monocyte chemoattractant protein-1 and other -chemokines by resident and inflammatory cells in multiple sclerosis lesions. J Neuroimmunol 84:238-249.
33 Simpson JE, Newcombe J, Cuzner ML, Woodroofe MN (2000). Expression of interferon-gamma-inducible chemokines IP-10 and Mig and their receptor, CXCR3 in multiple sclerosis lesions. Neuropathol Appl Neurobiol 26:133-142.
34 Sørensen TL&Ransohoff RM (1998). Etiology and pathogenesis of multiple sclerosis. Semin Neurol 18:287-295.
35 Sørensen TL, Tani M, Jensen Jet al. (1999). Expression of specific chemokines and chemokine receptors in the central nervous system of multiple sclerosis patients. J Clin Invest 103:807-815.
36 Steinman L (1996). Multiple sclerosis: a coordinated immunological attack against myelin in the central nervous system. Cell 85:299-302.
37 Verma MJ, Lloyd A, Rager Het al. (1997). Chemokines in acute anterior uveitis. Curr Eye Res 16:1202-1208.
38 Wie JM&Berman JW (1998). Astrocyte expression of monocyte chemoattractant protein-1 is differentially regulated by transforming growth factor-beta. J Neuroimmunol 91:190-197.
39 Zlotnik A&Yoshie O (2000). Chemokines: A new classification system and their role in immunity. Immunity 2:121-127.