Jul. 04, 2002
J Neuropathol Exp Neurol 2002 Jun;61(6):539-46
Peterson JW, Bo L, Mork S, Chang A, Ransohoff RM, Trapp BD.
Department of Neurosciences, Lerner Research Institute, The Cleveland Clinic Foundation, Ohio 44195, USA
Multiple inflammatory demyelinated lesions within the central nervous system (CNS) characterize multiple sclerosis (MS). MS is hypothesized to result from a peripherally mediated autoimmune reaction to CNS antigen or antigens such as the myelin proteins. Many of the inflammatory cells populating and mediating damage in MS lesions have been identified as hematogenous in origin (1-3). Infiltrating leukocytes express inducible, activation-dependent cell surface molecules, which engage counter-receptors on endothelial cells, lymphocytes, macrophages, microglia, astrocytes, myelin, and oligodendrocytes (4). These receptor/counter-receptor interactions contribute to lesion pathogenesis by promoting leukocyte entry into the CNS, leukocyte activation and proliferation, and destruction of myelin and oligodendrocytes (5).
Potential therapeutics that target cell surface molecules involved in cell trafficking are currently being developed to treat MS (6). Two immunoglobulin superfamily molecules that have been identified in MS lesions and are thought to play key roles in CNS inflammation are vascular cell adhesion molecule-1 (VCAM-1) (7, 8) and intercellular cell adhesion molecule-1 (ICAM-1) (9). In and around MS lesions, significantly increased percentages of vessels expressing ICAM-1 were identified (10). The integrin counter-receptors for VCAM-1 and for ICAM-1 have been detected on a high percentage of the infiltrating leukocytes in MS lesions (11). These include very late activation antigen-4 (VLA-4) for VCAM-1 and leukocyte function antigen-1 (LFA-1) for ICAM-1.
Therapeutic targeting of cell adhesion molecules to ameliorate inflammation has proved promising. The severity of experimental autoimmune encephalomyelitis (EAE) in animal models of MS was ameliorated by the use of ICAM-1 antibodies (12, 13). A phase I clinical study using humanized monoclonal antibodies to CD11/ CD18 an integrin ligand of ICAM-1 reduced the occurrence of new lesions detected by MRI during the 2-month period following clinical attacks of MS (14). Antibody therapies directed against VLA-4 also produced beneficial effects in various models of autoimmune and virusinduced demyelination (15-19). However, the results of VLA-4 blockade in EAE are complex (20), indicating that detailed understanding of its role in inflammatory demyelination will be required. A recent clinical trial in MS patients using intravenous injections of humanized monoclonal antibodies against VLA-4 indicated partial efficacy for ameliorating MRI disease activity in MS patients (21). Further, the beneficial effects of interferon beta in MS correlate with reducing VLA-4-dependent signaling by downregulating its expression (22) and by promoting release of soluble VCAM-1 (23). These results suggest that VLA-4 executes a functional role in the pathogenesis of MS.
VLA-4/VCAM-1 receptor-ligand coupling can mediate leukocyte-endothelial adhesion (24-26) during leukocyte extravasation into sites of inflammation (27). T and B-- lymphocytes, monocytes, mast cells, and basophils can express VLA-4 (28). VLA-4 epitopes are readily detected on infiltrating leukocytes that accumulate within the CNS (29), including MS brains (Trapp, unpublished data). VLA-4/VCAM-1 interactions are also implicated in B lymphocyte differentiation (30), T lymphocyte activation (31, 32), and proliferation (33). Increased VLA-4 expression by myelin proteolipid protein activated T lymphocytes correlate with their ability to induce EAE in adoptive transfer studies of mice (34). These observations suggest that VLA-4 is involved in leukocyte trafficking and activation in the CNS. However, the distribution of its ligand VCAM-1 in chronic human inflammatory diseases has been controversial, with its detection on endothelial cells (35) as well as tissue macrophages and other parenchymal cells in inflamed kidneys (36, 37), thyroid glands (38), and rheumatoid synovia (39, 40). In MS brain, VCAM-1 has been reported on endothelial cells (7); in vitro, it has been detected on astrocytes (41) and microglia (42), leaving the issue unresolved. Prior studies of VCAM-1 distribution within the CNS did not utilize 2-color immunofluorescence and confocal microscopy to achieve unambiguous lineage assignment for VCAM-1-- positive cells. This study examines the distribution of VCAM-1 in lesions from MS brain tissue by confocal microscopy of tissue sections immunolabeled with antibodies to VCAM-1 and lineage-specific markers.
MATERIALS AND METHODS
This study analyzed 26 tissue blocks from 9 MS brains, which contained 43 demyelinated white matter lesions. Normal control (CTL) and other neurological disease (OND) brains consisted of 1 tissue block each from 2 normal control brains, 2 encephalitis brains (ENC), 2 Alzheimer disease (AD) brains, and 1 cerebrovascular brain accident (CVA). All tissues were obtained at autopsy and fixed in 4% paraformaldehyde in 0.08 M Sorenson's phosphate buffered saline, pH 7.6. Age at the time of death and postmortem times were similar for MS and control cases. Features for all the cases studied are summarized in Table 1. Thirty-lim-thick, free-floating sections cut from each of the tissue blocks were immunostained with antibodies to myelin basic protein (MBP), major histocompatibility complex class II (MHC Class II), and vascular cell adhesion molecule (VCAM-1) using the avidin-biotin complex (ABC) procedure and DAB as described previously (43). Fifteen active, 21 chronic active, and 7 chronic inactive lesions were identified in the MS tissue blocks using previously described criteria (43).
For determination of VCAM-1-positive cell lineage and cell-- cell interactions, 30-(mu)m-thick, free-floating sections containing 5 active and 6 chronic active MS lesions were double-labeled with antibodies against VCAM-1 and specific cell lineage markers for astrocytes (GFAP), macrophages/microglia (ferritin), T lymphocytes (CD-3), or oligodendrocytes (carbonic anhydrase II). The appropriate fluorescent secondary antibodies were used. The sections were then analyzed using a Leica Aristoplan confocal microscope.
The primary antibodies used were rabbit anti-GFAP (1: 4,000), rabbit anti-MBP (1:1,000), rabbit anti-CD-3 (1:100), and mouse anti-MHC Class II (HLA-DR, 1:250) (Dako Corp., Carpinteria, CA), mouse anti-VCAM-1 (1:100) (Genzyme, Cambridge, MA), rabbit anti-ferritin (1:1,000) (Sigma, St. Louis, MO), and sheep anti-carbonic anhydrase II (1:100) (The Binding Site Ltd., Birmingham, England). Goat or rabbit secondary antibodies conjugated to biotin, Texas Red or FITC were obtained from Vector Laboratories Inc. (Burlingame, CA) and used at a dilution of 1:1,000.
Quantification of VCAM-1-Positive Cells
Six active, 7 chronic active, and 4 chronic inactive lesions from 8 MS brains immunostained for VCAM-1 by the ABC procedure in 30-(mu)m-thick, free-floating sections were selected for quantification. The numbers of VCAM-1-positive cells were quantified using a reticular grid and 40X objective. Ten fields at the lesions' borders and 10 fields from the lesions' cores were randomly selected for quantification of the total number of VCAM-1-positive cells, as well as the number of VCAM-1-- positive cells apposing or surrounding oligodendrocytes per cubic millimeter. The data was analyzed by the Student t-test.
Distribution of VCAM-1 in MS Lesions
The distribution of MBP, MHC Class II and VCAM-1 immunoreactivity was compared in adjacent sections from each tissue block. Forty-three demyelinated white matter lesions were identified by absence of MBP immunoreactivity. Lesion activity (15 active, 21 chronic active, and 7 inactive) was characterized by the distribution and density of MHC Class IT-positive cells as described previously (43). VCAM-1-positive cells were present in all active and chronic active MS lesions and particularly concentrated at the borders of active lesions. VCAM-1-positive cells were few in number and widely scattered in inactive lesions (data not shown) and in the remainder of the neurological control specimens. The distribution of VCAM-1 and MHC Class II immunoreactivity for 2 MS lesions is illustrated in Figure 1 with each dot representing an immunoreactive cell. MHC Class II-positive cells were detected throughout the MS tissue sections with the greatest density within the lesions and just outside the borders of demyelinated lesions. VCAM-1-positive cells were concentrated at the edges of MS lesions. In 3 of 43 of the MS lesions, foci of abundant VCAM-1-positive cells were detected at the border of the demyelinated zone.
A comparison of MHC Class II-positive cells and VCAM-1-positive cells in adjacent sections from 2 chronic active MS lesions revealed cells with similar morphology. MHC Class II labeled large round lipid-filled macrophages within the core of the lesions and a mixed population of macrophages and activated microglia at the lesion borders. VCAM-1-positive cells similarly varied in size and shape and could be segregated into 3 general categories based on morphology. Some VCAM-1-positive cells were small, round or oval, and predominantly localized to the perivascular space. Many VCAM-1-positive cells were large, round, and appeared to be lipid-filled. These cells represented the major phenotype of VCAM-1-positive cells identified at the cores and borders of all active and chronic active MS lesions. However, they were only a minority of the total inflammatory cells present within the MS lesions. Some VCAM-1-positive cells were detected just outside MS lesions and extended multiple processes. These cells were often located adjacent to and frequently their processes surrounded small, round VCAM-1-negative cells. VCAM-1 immunoreactivity was not detected on endothelial cells in any of the tissue specimens with the exception of the CVA tissue (data not shown).
The distribution of VCAM-1-positive microglia/macrophages was quantitated in 17 representative lesions (6 active, 7 chronic active, and 4 chronic inactive) from 8 different MS brains. Lesions of various activities were selected from multiple cases to ensure the results were not biased by data from atypical lesions or cases. The borders of active and chronic active MS lesions contained the greatest densities of VCAM-1-positive cells with an average of 102 and 29 VCAM-1-positive cells/mm^sup 3^, respectively. The densities of VCAM-1-positive cells in the cores of active (22 cells/mm3) and chronic active (8 cells/mm^sup 3^) lesions were reduced compared to the densities detected at the borders. Of the VCAM-1-positive cells detected at the borders of active and chronic active lesions, almost one fifth (17%) were closely associated with small round cells with the "fried egg" morphology characteristic of oligodendrocytes. In addition, within the cores of the active and chronic active lesions, 4%-6% of. the VCAM-1-positive cells were closely apposed to small round nuclei with "fried egg" morphology. Chronic inactive MS lesions contained very few VCAM-1-positive cells, with the greatest density at the borders and fewer within the cores.
From bright-field microscopic analysis the VCAM-1-- positive cells appeared to be macrophages/microglia by morphology. However, previous studies identified a subpopulation of astrocytes, which surrounded and engulfed oligodendrocytes in and near MS lesions (44, 45), with a distribution similar to the VCAM-1-positive cells identified in this study. In addition, another study showed that astrocytes can express VCAM-1 in vitro (41). Therefore, to determine the lineage of VCAM-1-positive cells and the identity of the small round cells they surrounded, sections containing 11 MS lesions (5 active and 6 chronic active) were double-labeled with antibodies against VCAM-1 and markers for macrophages/microglia (ferritin), astrocytes (GFAP), T lymphocytes (CD-3), or oligodendrocytes (CA-II) and then analyzed by confocal microscopy. All of the large round and process-bearing VCAM-1-positive cells were ferritin-positive indicating that the VCAM-1-positive cells were of monocyte origin. None of the VCAM-1-positive cells were GFAP immunoreactive nor did they surround or closely appose GFAP-positive perikarya. The small round cells associated with VCAM-1-positive cells were CD-3-negative and therefore not T lymphocytes (data not shown); they were CA-II-positive and thus identified as oligodendrocytes. In normal control tissue, ferritin-positive microglia were not immunopositive for VCAM-1 and they did not associate with CA-II-positive oligodendrocytes.
The pathogenesis of MS entails destruction of CNS tissue and it is well established that cells of monocyte origin are major mediators of tissue injury (46). The present study described a subset of VCAM-1-positive macrophages/microglia that surrounded oligodendrocyte perikarya at the edges of both active and chronic active lesions from 9 MS brains with clinical disease ranging from 2 wk to 28 yr. These observations raise the possibility that the expression of VCAM-1 by monocyte lineage cells is related to the depletion of oligodendrocytes in MS lesions. This conclusion rests on the selective association between VCAM-1-positive macrophages/microglia and oligodendrocyte perikarya within lesions where oligodendrocyte loss is most commonly observed.
Association between activated microglia and CNS components in MS lesions is not restricted to oligodendrocytes. Activated microglia selectively send processes to normal appearing myelin internodes at the edges of active and chronic active MS lesions (47, 48), surround terminal axonal swellings of transected axons in white matter lesions (49), and associate with and ensheath neuronal cell bodies and dendrites in cortical lesions (50). The associations between activated microglia and neuronal cell bodies have been described in a variety of pathological conditions including those that affect distal segments of axons (51). The association between activated microglia, neuronal perikarya and dendrites may result in the separation of pre- and post-synaptic terminals (52). However, it is unclear whether this deafferentation of neurons is protective or detrimental to the neuron. The association of activated microglia with myelin has been interpreted as a destructive process leading to demyelination (47, 48).
The present study identifies for the first time a potential specific molecular interaction between activated microglia and one of the CNS cells destroyed in MS lesions. Our data do not address whether VCAM-1 is induced on macrophages/microglia during the response to oligodendrocyte injury or whether VCAM-1-positive cells selectively target oligodendrocytes in MS lesions. However, the specific association between VCAM-1-positive cells and oligodendrocyte perikarya at locations where oligodendrocyte loss occurs raises the possibility that this contact is detrimental to oligodendrocyte survival. Activated macrophages/microglia secrete cytokines that can mediate oligodendrocyte and myelin damage in vitro (53). Engagement of VCAM-1 on microglia by T-cell VLA-4 can elicit tumor necrosis factor alpha production (42). The ligand for VCAM-1 on oligodendrocytes was not addressed in our present study. Integrin expression is developmentally regulated in the glial lineage (54), so that potential VCAM-1 ligands may be expressed in MS brain by such cells. There is a precedent for functional expression of alpha4 integrins by neural cells, as such epitopes have been detected on the growth cones of regenerating peripheral nerves and mediate interactions with fibronectin (55). Although microglial reaction to neuronal injury is well documented, this data is the first to describe a specific microglial-oligodendrocyte interaction in MS.
VCAM-1 has also been implicated in leukocyte trafficking, via interactions with VLA-4 expressed by leukocytes. Our studies failed to detect VCAM-1 immunoreactivity on vessels in MS lesions using a well-characterized VCAM-1 antibody that binds to both known human VCAM-1 isoforms (25). Our data are consistent with previous reports that described VCAM-1-- positive activated microglia in vitro (56). Furthermore, engagement of microglial or macrophage VCAM-1 by either fibronectin or VLA-4 can evoke pro-inflammatory responses (42, 57, 58). High concentrations of soluble VCAM-1 (sVCAM-1) have been reported in the cerebrospinal fluid (CSF) of MS patients (59, 60), although there is controversy about whether predominant expression of VCAM-1 in these studies is intrathecal or derived from circulating sources. VCAM-1 shedding from the surface of macrophages/microglia is a plausible source of CSF VCAM-1, whereas release of VCAM-1 from the luminal surface of vessels would increase serum rather than CSF levels of sVCAM- 1. There is evidence to support the hypothesis that both mechanisms are operative in MS patients. Elevated CSF VCAM-1, of putative intrathecal origin, has been associated with active, relapsing-remitting, as compared with progressive or stable disease. Fluctuations in serum VCAM-1 levels are related to therapeutic modulation of MRI-defined disease activity in MS, supporting the latter possibility. Although not directly addressed in this report, our findings raise the possibility that VLA-4-bearing cells may utilize endothelial receptors distinct from, or in addition to, VCAM-1 to enter the CNS. The major alternative ligand for VLA-4 on vessels is the connecting segment (CS)-1 epitope on fibronectin (61). Studies with small-molecule inhibitors indicate that VLA-4/CS-1 interactions are less susceptible to inhibition than are VLA-4/VCAM-1 interactions (62).
In summary, VCAM-1-positive cells in MS lesions, as defined by dual-label immunohistochemistry, are derived from the monocyte/macrophage lineage and are highly enriched at lesion edges coincident with regions of oligodendrocyte injury. Contacts between VCAM-1-positive macrophages/microglia and oligodendrocyte perikarya suggest the possibility that the molecular determinants of this interaction may provide a useful target for therapeutic intervention that may not be addressed by the current generation of anti-integrin reagents.
The authors thank Dr. Xinghua Yin for valuable input and Victoria Pickett
for manuscript editing and preparation. Dr. BO's current address is the
Department of Neurology, Haukeland Hospital, Bergen, Norway. Tissue specimens
were obtained from the National Neurological Research Specimen Bank, Veterans
Administration Medical Center (Los Angeles, CA), which is supported by
(NINDS/NIMH), National Multiple Sclerosis Society, Hereditary Disease Foundation,
and Veteran's Health Services and Research Administration.
© 2002, Journal of Neuropathology and Experimental Neurology