Adhesion Molecules in Multiple Sclerosis: Title and subTitle BreakRelation to Subtypes of Disease and Methylprednisolone Therapy FREE

MULTIPLE sclerosis (MS) is an autoimmune neurological disease characterized by multifocal areas of inflammatory demyelination within the central nervous system (CNS). The binding of circulating autoreactive T cells and macrophages to the CNS endothelial cells and subsequent migration through the blood-brain barrier is an essential step in the initiation of brain inflammation.^{1 - 2} This step of immune cell entry into the site of inflammation is mediated by adhesion molecules.³ In the CNS, T cells are exposed to the sensitization antigen that triggers inflammation by cytokines released during immunorecognition.⁴ Adhesion molecules induced by such cytokines on brain cells and immune cells promote interaction between these cells, leading to damage of myelin and oligodendrocytes with subsequent demyelination.⁵

Two adhesion molecule pathways have been well defined: the intercellular adhesion molecule 1 (ICAM-1, CD54) on endothelial cells and leukocytes together with its ligand on leukocytes, the lymphocyte function associated antigen 1 (LFA-1, CD11a), and the vascular cell adhesion molecule 1 (VCAM-1, CD106) on endothelial cell and macrophages and its ligand on monocytes and lymphocytes, the very late activation antigen 4 (VLA-4, CDw49d).⁶ Increased expression of various cell adhesion molecules has been found in the CNS in MS^{7 - 9} and its model disease, experimental allergic encephalomyelitis.^{10 - 12} The adhesion molecules are detected especially in and around cerebral microvessels and on inflammatory cells, including macrophages, microglial cells, and lymphocytes in MS lesions.^{5 ,9 ,13 - 15} Under normal conditions, cerebrovascular endothelium exhibits only low levels of ICAM-1, and no constitutive expression of other adhesion molecules has been reported.^{13 ,15}

Activation of endothelial cells and immune cells in blood results in rapid up-regulation and possible shedding of adhesion molecules in serum and cerebrospinal fluid (CSF).^{16 - 21} Elevated CSF and serum levels of certain adhesion molecules have been associated with varying activity and clinical course of MS.^{19 - 20 ,22 - 25}

An earlier study from our laboratory showed up-regulated expressions of VLA-4 and LFA-1 on immune cells from blood and CSF in relapsing-remitting MS (RRMS) in exacerbation, which were dramatically reduced after high-dose intravenous (IV) methylprednisolone sodium succinate treatment.²⁶ In the present study, we sought to ascertain whether expressions of adhesion molecules on the surface of white blood cells are increased during the clinically stable phase of secondary progressive MS (SPMS), which would indicate a sustained potential for inflammation throughout the spectrum of MS disease, regardless of its activity. We also determined the concentrations of soluble adhesion molecules in different subtypes of MS and the effect of IV methylprednisolone treatment on the levels of these proteins.

One hundred subjects were involved in the study. Seventy-one patients (31 male and 40 female) had clinically definitive MS according to the criteria of Poser et al,²⁷ and 29 subjects (9 male and 20 female) were healthy controls (Table 1). The MS group included 26 patients (9 male and 17 female) with RRMS, 20 (9 male and 11 female) with SPMS, and 25 (13 male and 12 female) with primary progressive MS (PPMS). The severity of the disease was scored using the Expanded Disability Status Scale (EDSS).²⁸ At the time of our studies, the patients with SPMS did not experience symptoms of acute neurological worsening and had not received any immunosuppressive or other therapy because of MS for 8 weeks before the study. The clinical condition of the patients with PPMS at the time of neurological examinations was stable. Eleven controls were studied because of variable neurological symptoms, and they presented no evidence of organic CNS disease. The remaining controls were healthy volunteers who were staff members of Tampere University Hospital, Tampere, Finland.

Blood and CSF samples for analysis were obtained by routine procedures, with the informed consent of the patients, as described elsewhere.²⁶ The study was approved by the Ethical Committee of Tampere University Hospital. The concentrations of circulating adhesion molecules were measured in frozen specimens stored at −70°C.

All patients underwent studies on the same 0.5-tesla magnetic resonance imaging (MRI) unit (Vectra GE; General Electric, Milwaukee, Wis), using a standard head coil. Routinely used sagittal T1-weighted, axial T2-weighted, proton density-weighted, fluid-attenuated inversion recovery–weighted, and gadolinium-pentetic acid–enhanced T1-weighted images were obtained. In addition, the imaging protocol included axial T1-weighted fast spin echo images (repetition time [TR], 300 milliseconds; echo time [TE], 20 milliseconds; number of excitations, 5; field of view, 22 cm; matrix size, 160/256; echo train length, 5; and slice thickness, 5 mm with 2-mm gap) and axial 3-dimensional T2-weighted fast spin echo images (TR, 2000 milliseconds; TE, 150 milliseconds; number of excitations, 1; field of view, 22 cm; matrix size, 192/224; echo train length, 16; and slice thickness, 2 mm with no intervening gap) for segmentation and volumetric analysis. Segmentation and volumetric analysis were performed as described previously.²⁹

The concentrations of soluble VCAM-1 (sVCAM-1), sICAM-1, and sE-selection were measured using commercially available sandwich enzyme immunoassay kits (BBE 3, BBE lB, and BBE 2B, respectively; R & D Systems, Minneapolis, Minn). Serum or CSF samples diluted 1:50 for sVCAM-1 and 1:20 for sICAM-1 and sE-selectin were added in duplicate to microtiter wells and assayed according to routine procedures. The signals obtained from standards of known concentration were used for development of a standard curve. The concentrations in patient samples were calculated using a 4-variable logistic curve fit.

Unless otherwise indicated, data are given as mean ± SEM. For phenotyping the mononuclear cells, the 3-layer indirect immunoperoxidase technique was applied.^{26 ,30} The primary antibodies used were LFA-1 (CD11a), VLA-4 (CDw49d), ICAM-1 (CD54), or VCAM-1 (CD106) (0157, 0764, 0544, and 1244, respectively; Immunotech, Marseille, France). The dilution used for the CSF specimens was 1:100; for blood specimens, 1:200. The secondary antibody was a peroxidase-conjugated rabbit anti–mouse used in dilution of 1:10 (P0161; Dako, Glostrup, Denmark) and the third was a peroxidase-conjugated goat anti–rabbit antibody used in dilution of 1:20 (L42007; Caltag, San Francisco, Calif). The specimens were analyzed at ×100 magnification. In blood, a total of 400 cells, and in CSF, a mean of 152 ± 55 mononuclear cells were analyzed for each cell surface marker. The results were expressed as the percentage of positively stained lymphocytes in the total number of lymphocytes counted and the percentage of positively stained monocytes in the total number of monocytes counted.

Independent 2-tailed t test was used in the group mean comparisons with normal distribution. Mann-Whitney test was used in cases of skewed distribution. In correlation analyses, Spearman rank order correlation coefficients were calculated. All statistical analyses were made on a microcomputer using a commercially available program package (Statistica for Windows; Statsoft, Inc, Tulsa, Okla).

The proportions of the adhesion molecule expression on blood lymphocytes and monocytes are presented in Table 2. In patients with RRMS in exacerbation, the expressions of VLA-4 and LFA-1 in blood were 1.3- to 3.5-fold higher than in controls (VLA-4 lymphocytes, 41.3 ±4.3 vs 13.1 ± 3.2 [P<.001]; LFA-1 lymphocytes, 45.9 ± 3.7 vs 13.0 ± 2.3 [P<.001]; LFA-1 monocytes, 56.6 ± 5.6 vs 24.4 ± 3.9 [P=.002]). In SPMS, the proportions of blood lymphocytes expressing these adhesion molecules were likewise at least 2.5-fold higher than in controls (VLA-4, 34.6 ± 2.6 vs 13.1 ± 3.2 [P=.003]; LFA-1, 36.3 ± 2.7 vs 13.0 ± 2.3 [P<.001]). Comparison between RRMS and SPMS showed more up-regulated LFA-1 on lymphocytes (P=.03) and ICAM-1 on monocytes (P=.03) in RRMS. No differences were found between the expressions of VCAM-1 in patients with RRMS or SPMS or in controls.

The expressions of VLA-4 and LFA-1 on the CSF lymphocytes from patients with RRMS were approximately 4-fold higher than in control CSF (Table 2) (VLA-4, 36.8 ± 8.2 vs 9.1 ± 3.4 [P=.009]; LFA-I, 45.8 ± 10.8 vs 12.0 ± 2.3 [P=.008]). In SPMS, the amounts of these adhesion proteins on the CSF lymphocytes were also at least 3-fold higher than in controls (VLA-4, 27.0 ± 3.8 vs 9.1 ± 3.4 [P=.01]; LFA-I, 42.4 ± 5.6 vs 12.0 ± 2.3 [P=.005]), but no statistically significant differences were found between RRMS and SPMS.

The immunological variables in the CSF of the 20 patients with SPMS were determined as follows: CSF leukocyte count, 0.003 ± 0.0002 × 10⁹/L (reference, ≤0.005 × 10⁹/L), IgG index, 0.88 ± 0.09 (reference, ≤0.60); and CSF–serum albumin ratio, 5.3 ± 0.6 × 10⁻³ (reference, ≤7.0 × 10⁻³). The correlation of adhesion molecule expressions to these immunological findings in the CSF revealed a significant correlation between the expressions of LFA-1 on CSF lymphocytes and IgG indexes (r = 0.62; P = .005), but no correlations were found to CSF white cell counts or blood-brain barrier permeability. The correlations between adhesion molecule expressions in blood and routine immunological findings in the CSF were not significant.

Comparison between the levels of soluble adhesion molecules in different MS subtypes revealed higher sVCAM-1 levels in SPMS than in RRMS (Table 3) (P=.005) or PPMS (P=.04), whereas the difference between SPMS and controls was not significant. However, in RRMS and PPMS, the levels of sVCAM-1 were even lower than in controls (P=.005 and P=.04, respectively). No differences were seen between both MS subtypes. No clear differences were detected in the levels of sICAM-1 among the MS subtypes or between these and control samples. In PPMS, the level of sE-selectin tended to be higher than in SPMS, but the difference did not reach statistical significance (Table 3).

The levels of soluble adhesion molecules in the CSF of patients with PPMS are shown in Table 4.

An improvement of at least 0.5 in EDSS scores 2 to 3 weeks after IV methyprednisolone treatment completion was detected in 11 of the 16 patients with RRMS in relapse. The difference between EDSS before and after methylprednisolone therapy was significant (2.9 vs.2.2; P=.01). The levels of serum sVCAM-1, sICAM-1, and sE-selectin were measured accordingly before and after high-dose IV methylprednisolone treatment. Methylprednisolone caused a significant reduction in the levels of serum VCAM-1 (433 ± 116 vs 398 ± 97 ng/mL; P<.001) and sE-selectin (40 ± 13 vs 31 ± 11 ng/mL; P<.001), but the decline in sICAM-1 level was marginal (Table 5).

The volumes of MS plaques and brain atrophy together with cell surface adhesion molecule expressions were determined in the 20 patients with SPMS. The median lesion volumes for T2- and T1-weighted plaques were 6.1 and 1.1 cm³, respectively. In the assessment of brain atrophy, we found that the median volume for total intracranial CSF spaces (total volumes of ventricles and peripheral CSF spaces) was 132 cm³. The relation of total intracranial CSF space volumes to total brain volume was 0.19. The gadolinium-enhanced T1-weighted plaques were seen in only 3 of the 20 patients with SPMS. Each of them had 2 small (2-mm) periventricular lesions. No significant correlations were found between MRI abnormalities and the expressions of adhesion molecules shown in Table 2.

The soluble adhesion molecule levels were measured in 13 of 20 patients with SPMS; the levels were correlated with corresponding MRI measurements. A correlation was found only between the volumes of T2-weighted plaques and the levels of sICAM-1 in serum (r = 0.64; P = .02). The mean (± SD) concentration of sICAM-1 in these patients was 265 ± 37 ng/mL, and the median volume of T2-weighted plaques was 7.4 cm³.

Elevated levels of soluble adhesion molecules in MS have been reported by many investigators,^{16 - 23} but much less information is available on the cell surface expression of these proteins in CSF and blood.^{24 ,31 - 32} In our study, we found increased proportions of VLA-4– and LFA-1–positive immune cells in blood of patients with RRMS and SPMS compared with healthy controls. Although the expressions of adhesion molecules were higher in RRMS in exacerbation, up-regulated VLA-4 and LFA-1 were also detected in the clinically stable phase of SPMS, which indicates leukocyte-endothelial activation in acute and stable MS. Since the ability of activated immune cells to bind to their receptors on activated endothelium facilitates their migration through the blood-brain barrier, our results imply that the potential for brain inflammation is sustained throughout the course of the disease regardless of its activity.

Previous reports on the soluble endothelial adhesion molecules in the serum of patients with MS have been conflicting, showing elevated^{16 ,18 ,20 ,22 ,25 ,33} and normal levels.^{19 ,21 ,23} This indicates that serum levels of adhesion molecules do not consistently reflect the activity of MS disease. Elevated levels of sICAM-1 or sVCAM-1 in the CSF during MS relapse seem to be more closely related to disease activity.^{17 - 19 ,21} Our results showed the level of serum sVCAM-1 in patients with SPMS to be higher than in patients with RRMS in relapse or with PPMS. Although these results were obtained in serum samples, they are consistent with the observation of Cannella and Raine⁹ on the predominance of VCAM-1 and VLA-4 in chronic active MS. Other investigators have reported higher values for serum sICAM-1 and sE-selectin in PPMS compared with RRMS, although the activity of the disease has not always been clear.^{19 ,23 ,34} The partial discrepancies between our results and previous data may be explained by differences in methods or timing of sampling in relation to disease activity. The differences between chronically progressive types of MS and RRMS in exacerbation suggest that the cellular source of adhesion molecules and the kinetics of cleavage of endothelial cell surface molecules to their soluble form may differ in relapsing and chronically progressive disease courses.

We observed that methylprednisolone reduced the concentrations of serum sVCAM-1 and sE-selectin. A lowering of sE-selectin levels recently was reported by Droogan et al.³⁵ These observations are in harmony with the current knowledge that corticosteroids suppress many important events in the inflammatory cascade.³⁶ More specifically, our results suggest a down-regulating effect of corticosteroids on the expression of adhesion molecules on circulating immune cells and the cerebral endothelium or cells within the CNS.

The correlation between the expression of adhesion molecules and MS plaques or brain atrophy in patients with SPMS showed association only between the level of serum sICAM-1 and the volumes of T2-weighted plaques. This is not surprising, since previous studies on the association of biological markers, including adhesion molecules, and MS-related MRI or EDSS changes have been mostly disappointing.^{20 ,22 ,37 - 38} The lack of clear-cut associations in this study may be explained by the lack of clinical and MRI activity during examinations. In a previous study from our laboratory on RRMS in exacerbation, significant correlations were found between the numbers of T2-weighted lesions and expressions of VLA-4 and LFA-1 on mononuclear cells from blood.²⁶

Our results imply that therapies interfering with cell adhesion should be considered as a potential objective in MS. The importance of such therapies has been demonstrated by the blockade of the VLA-4–VCAM-1 pathway to prevent and reverse experimental allergic encephalomyelitis, a T-cell–mediated neuroinflammatory model for MS.^{10 ,12 ,39 - 41} Therapies modulating adhesion molecules with resultant restriction of inflammatory cell invasion of the CNS may provide an important tool in the effort to suppress MS.

Accepted for publication November 11, 1999.

The study was supported by the Medical Research Fund of Tampere University Hospital, Tampere, Finland, and The Finnish Foundation of Cardiovascular Research, Helsinki.

We thank laboratory technologist Eija Spåre for technical assistance and Tuula Jalonen, PhD, for critical reading of the manuscript and valuable suggestions.

Reprints: Irina Elovaara, MD, PhD, Neuroimmunology Unit, Tampere University Hospital, Teiskontie 35, 33520 Tampere, Finland (e-mail: IElovaara@tays.fi).