0023-6837/01/8103-263$03.00/0
`LABORATORY INVESTIGATION
`Copyright © 2001 by The United States and Canadian Academy of Pathology, Inc.
`
`Vol. 81, No. 3, p. 263, 2001
`Printed in U.S.A.
`
`BIOLOGY OF DISEASE
`Multiple Sclerosis: Current
`Pathophysiological Concepts
`Dean M. Wingerchuk, Claudia F. Lucchinetti, and John H. Noseworthy
`Department of Neurology (DMW), Mayo Clinic, Scottsdale, Arizona, and Department of Neurology (CFL, JHN), Mayo
`Clinic, Rochester, Minnesota
`
`SUMMARY: Multiple sclerosis (MS) is an often disabling disease primarily affecting young adults that exhibits extraordinary
`clinical, radiological, and pathological heterogeneity. We review the following: (a) known environmental and genetic factors that
`contribute to MS susceptibility; (b) current knowledge regarding fundamental pathophysiological processes in MS, including
`immune cell recruitment and entry into the central nervous system (CNS), formation of the plaque, and orchestration of the
`immune response; (c) descriptive and qualitative distinct pathological patterns in MS and their implications; (d) the evidence
`supporting the causative role of direct toxins, cell-mediated and humorally mediated immune mechanisms, and the concept of
`a “primary oligodendrogliopathy” in demyelination and axonal injury; (e) the potential benefits of inflammation; (f) the prospects
`for remyelination; and (g) therapeutic implications and approaches suggested by putative pathophysiological mechanisms. (Lab
`Invest 2001, 81:263–281).
`
`M ultiple sclerosis (MS) is a common, heteroge-
`
`neous disorder of the central nervous system
`(CNS) (Noseworthy, 1999; Noseworthy et al, 2000a).
`Its causes and the factors that contribute to its heter-
`ogeneity are largely unknown, although it is likely a
`complex trait with genetic and environmental compo-
`nents. The disease affects about 0.1% of the popula-
`tion in temperate climates, some 250,000 to 350,000
`people in the United States. It is a disease of young
`people (median age of onset is approximately 28
`years) but is lifelong and is often disabling; 50% of
`patients require a cane to walk 15 years after disease
`onset (Weinshenker et al, 1989).
`Early in the course of relapsing-remitting disease
`(RRMS), which affects about 85% of patients, neuro-
`logical symptoms and signs develop over several
`days, plateau, and then usually improve over days to
`weeks (Schumacher et al, 1965). These relapses typ-
`ically consist of one or a combination of the following:
`sensory symptoms, optic neuritis, Lhermitte’s sign
`(axial or limb paresthesias with neck flexion),
`limb
`weakness, gait ataxia, brain stem symptoms (diplopia;
`ataxia), Uhthoff symptom (symptomatic worsening
`with increases in body temperature), a circadian fa-
`tigue pattern (fatigue worse in mid- to late-afternoon
`concomitant with increases in core body temperature),
`and sphincter dysfunction. Inflammatory infiltrates and
`
`Received November 27, 2000.
`Address reprint requests to: Dr. John H. Noseworthy, MD, FRCP(C),
`Mayo Clinic, 200 First Street SW, Rochester, MN 55905. E-mail:
`noseworthy.john@mayo.edu
`
`demyelination in brain and spinal cord white matter
`usually accompany these clinical exacerbations. Peri-
`ods of clinical quiescence (remissions) occur between
`exacerbations; remissions vary in length and may last
`several years but are infrequently permanent. The
`remaining 15% of patients begin the disease course
`by experiencing gradually progressive neurological
`function, typically a slowly worsening myelopathy (pri-
`mary progressive disease, PPMS). Approximately
`two-thirds of patients with RRMS eventually undergo
`a similar fate; as relapse frequency lessens over time,
`progressive neurological dysfunction emerges, signal-
`ing the development of secondary progressive dis-
`ease (SPMS) (Weinshenker et al, 1989). Some patients
`who convert to a secondary progressive course con-
`tinue to experience superimposed relapses.
`The above classification system defines the proto-
`typic or classic form of MS (Lublin and Reingold,
`1996). Classification schemes for CNS demyelinating
`diseases include several uncommon syndromes with
`controversial relationships to classic MS,
`including
`complete transverse myelitis, neuromyelitis optica
`(Devic’s syndrome), acute disseminated encephalo-
`myelitis, Balo’s concentric sclerosis, and the fulminant
`Marburg variant
`(Korte et al, 1994; Mendez and
`Pogacar, 1988; Wingerchuk et al, 1999). These syn-
`dromes retain the basic inflammatory and demyelinat-
`ing pathology of MS but differ from classic disease
`with unusually acute and severe clinical presentations,
`restricted lesion topography (eg, optic nerve and spi-
`nal cord lesions in Devic’s syndrome), or distinct
`pathological features (eg, pronounced acute axonal
`destruction and necrosis in neuromyelitis optica and
`
`Laboratory Investigation • March 2001 • Volume 81 • Number 3 263
`
`Hopewell EX1009
`
`1
`
`

`

`Wingerchuk et al
`
`the Balo’s and Marburg variants). Apart from these
`entities, it has generally been accepted that similar
`pathophysiological mechanisms are operative in all
`patients with prototypic MS. Recent studies, however,
`suggest that pathological heterogeneity may also exist
`amongst patients with otherwise classic disease (Luc-
`chinetti et al, 1996, 2000a). There may be a restricted
`number of distinct pathological patterns with a single,
`dominant pattern present in all active lesions within an
`individual patient. This finding suggests that distinct
`pathogenetic mechanisms may be involved in differ-
`ent patient subgroups and has wide-ranging implica-
`tions for disease classification and future investigation
`of the causes and pathophysiological mechanisms
`that underlie MS (Lucchinetti et al, 2000a). Further-
`more, increasing attention is being paid to the role of
`axonal injury and loss, the likely correlate of progres-
`sive and irreparable injury in MS. We will review
`progress in the understanding of the etiology, patho-
`physiology, and pathology of MS and their implica-
`tions for discovering effective treatments that arrest or
`repair damage done by this disabling disease.
`
`Environment and Genetics
`
`The cause of MS is not known. Epidemiological find-
`ings support both environmental and genetic hypoth-
`eses, and these forces likely interact
`to produce
`individual disease susceptibility and influence disease
`course.
`Several observations seemingly support environ-
`mental hypotheses. The prevalence of MS generally
`increases with distance from the equator (Kurtzke,
`1980), and apparent epidemics and clusters of MS
`have been reported. Migration (and age at migration)
`may modify the disease risk, and concordance rates in
`monozygotic twins do not exceed approximately 30%
`(Ebers et al, 1986; Mumford et al, 1994). Some con-
`sider these findings as supportive of an ecological or
`infectious hypothesis for MS susceptibility. It is un-
`clear whether putative environmental factors are op-
`erative at the individual level (eg, infectious, transmis-
`sible agents) or elevate the risk of the entire population
`(eg, ecological factors, such as climate, soil condi-
`tions, or diet) (Lauer, 1997). Ecological case-control
`studies are often limited because exposures are usu-
`ally similar amongst cases and controls. Isolation of
`infectious agents and/or serological evidence of
`greater exposure in MS cases compared to controls
`have been reported frequently over several decades.
`Recent
`reports implicate human herpes virus 6
`(HHV-6) (Challoner et al, 1995; Friedman et al, 1999)
`and Chlamydia pneumoniae (Gilden, 1999; Sriram et
`al, 1999) as causative agents, but others have failed to
`confirm these observations (Boman et al 2000; Martin
`et al, 1997; Mirandola et al, 1999). To date, no single
`infectious agent has withstood the test of time.
`Genetic predisposition to MS has been established
`from the following evidence: familial aggregation un-
`explained by environmental factors (Ebers et al, 1995);
`much higher monozygotic than dizygotic twin concor-
`dance rate (31% versus 5%) (Sadovnick et al, 1993);
`
`ethnic predisposition (eg, Northern Europeans) and
`protection (many groups, including North American
`Indians and Hutterites, despite living in regions with
`high MS prevalence); and association with human
`leukocyte antigen (HLA) DR2. The exact mode of
`inheritance is unknown but does not appear to be
`Mendelian or mitochondrial in nature. In general, the
`risk to a first-degree relative is approximately 1% to
`4% (10 – 40 times the population risk), but this value
`may be substantially higher in pedigrees with multiple
`affected members.
`The genes that contribute to MS susceptibility have
`not been identified. The HLA DR2 allele has been
`associated with MS in many populations (Ebers et al,
`1995). Four entire human genome screens by linkage
`have been reported (Ebers et al, 1996; Haines et al,
`1996; Kuokkanen et al, 1997; Sawcer et al, 1996).
`Although refinement of the original genome screens
`continues (Chataway et al, 1998), the most consistent
`evidence of a susceptibility locus appears to be the
`HLA region on chromosome 6. It seems unlikely that
`any other single genes contribute a significant risk.
`Genetic factors may also determine disease course
`and severity, but HLA polymorphisms are not signifi-
`cant contributors (Weinshenker et al, 1998). Polymor-
`phisms in the interleukin-1␤-receptor and interleukin-
`1␤-receptor antagonist genes (Schrijver et al, 1999),
`the apolipoprotein E gene (Evangelou et al, 1999), and
`immunoglobulin Fc receptor genes (Myrh et al, 1999)
`have been associated with disease course. These
`associations require confirmation.
`
`Pathophysiological Features of
`Multiple Sclerosis
`
`The pathological signature of MS is the white matter
`plaque, a circumscribed area of demyelination and
`relative axonal preservation. Plaques may occur any-
`where within the white matter but favor the periven-
`tricular regions, optic nerves, brain stem, cerebellum,
`and spinal cord. Depending on their stage of develop-
`ment, they contain varying proportions of immune
`cells and immunoreactive substances. We review cur-
`rent knowledge for several questions concerning im-
`mune cell recruitment and entry into the CNS, initiation
`and propagation of active lesions, and the mecha-
`nisms and patterns of demyelination, axonal
`injury,
`remyelination, and cell loss.
`
`WhatIstheCompositionoftheMSPlaque?
`
`Multiple sclerosis plaques may be characterized as
`active or inactive (Lassmann et al, 1998). There are
`several methods for determining plaque activity, but
`the most dependable seems to be the presence in
`macrophages of specific myelin degradation products
`(reactive for myelin basic protein [MBP], myelin oligo-
`dendrocyte glycoprotein [MOG], and proteolipid pro-
`tein [PLP]) and activation markers (including MRP 14
`and 27E10)
`(Brück et al, 1995; Lucchinetti et al,
`2000a). Macrophages are especially plentiful in active
`
`264 Laboratory Investigation • March 2001 • Volume 81 • Number 3
`
`2
`
`

`

`lesions (Lassmann et al, 1998), which are hypercellular
`and contain patchy infiltrates of autoreactive T cells
`and antigen-nonspecific monocytes and macro-
`phages within a zone of myelin loss (Fig. 1). Macro-
`phages and lymphocytes form prominent perivascular
`cuffs and invade the parenchyma, whereas plasma
`cells and B cells tend to concentrate in the perivascu-
`lar region only (Prineas and Wright 1978). Most lym-
`phocytes within plaques are T cells, including both
`CD4⫹ (helper) and CD8⫹ (cytotoxic) cells; conflicting
`data exist concerning their relative proportions (Raine,
`1994) (see “Direct Cell-Mediated Injury”). The CD4⫹
`cells can be functionally divided into Th1 (secretion of
`“proinflammatory” cytokines, such as tumor necrosis
`factor-alpha [TNF-␣] and gamma-interferon [␥-IFN]) or
`Th2 (secretion of interleukins [IL]-4,-5,-6, and others)
`phenotypes; the relative proportions of these cells and
`their activity levels may contribute to lesional activity.
`Reactive astrocytes are usually present in the periph-
`ery of the lesion.
`Actively demyelinating plaques may conform to one
`of four postulated distinct pathological patterns (Luc-
`chinetti et al, 2000a). In patterns I and II, macrophages
`and T cells predominate in well-demarcated plaques
`that surround small veins and venules; pattern II is
`distinguished by the local precipitation of immuno-
`globulin (primarily IgG) and activated complement in
`regions of active myelin damage. In both patterns, the
`expression of all myelin proteins (eg, MBP, PLP, MOG,
`and myelin-associated glycoprotein [MAG]) are re-
`duced to similar degrees, and oligodendrocytes are
`variably lost at the plaque edge, with reappearance of
`oligodendrocytes within the plaque center (Fig. 2a).
`Remyelination is extensive in lesion patterns I and II.
`
`Multiple Sclerosis
`
`lesions also contain a cellular infiltrate
`Pattern III
`mainly composed of macrophages, T cells, and acti-
`vated microglia. These ill-defined plaques are not
`vessel-centered.
`Immunoglobulin and complement
`deposition are absent; however, there is a preferential
`loss of MAG compared to the other myelin proteins.
`This pattern is associated with severe oligodendrocyte
`loss and evidence of oligodendrocyte apoptosis (Fig.
`2b). Pattern IV also demonstrates macrophage and T
`cell inflammation without immunoglobulin or comple-
`ment staining, but with nonapoptotic oligodendroglial
`death in the normal-appearing periplaque white matter
`and loss of all myelin proteins at the active edge of the
`plaque. Remyelination is minimal in pattern III and IV
`lesions, and each suggests a primary injury to the
`oligodendrocyte. These conclusions are supported by
`ultrastructural studies of stereotactic brain biopsies
`from MS patients, which revealed a group of lesions
`demonstrating primary alterations in the most distal
`oligodendrocyte processes (“distal, dying-back oligo-
`dendrogliopathy”) (Rodriguez and Scheithauer, 1994).
`In autopsy cases studied thus far, all active lesions
`from an individual patient conform to a single immu-
`nopathological pattern.
`Patients with chronic MS have few active plaques.
`Chronic plaques display well-demarcated areas of
`hypocellularity with myelin pallor or loss (Fig. 3). There
`are varying degrees of axonal
`loss, usually most
`obvious in the lesional center (Barnes et al, 1991;
`Raine, 1991). There is typically a persistent but minor
`inflammatory response, with only a few scattered
`perivascular lymphocytes present, although plasma
`cells may occasionally be prominent (Prineas and
`
`Figure 1.
`Photomicrographs of an actively demyelinating multiple sclerosis lesion (immunocytochemical staining of myelin oligodendrocyte glycoprotein [brown] with
`hematoxylin counterstaining of nuclei [blue]). Left panel, At the active edge of a multiple sclerosis lesion (indicated by the asterisk), the products of myelin degradation
`are present in numerous macrophages (arrowheads). Right panel, Macrophages containing myelin debris (arrowheads) are interdigitated with degenerating myelin
`sheaths. (Both panels, Magnification, ⫻100.) (Reprinted from Noseworthy et al N Engl J Med 2000;343:938 –952. Copyright © 2000 Massachusetts Medical Society.
`All rights reserved.)
`
`Laboratory Investigation • March 2001 • Volume 81 • Number 3 265
`
`3
`
`

`

`Wingerchuk et al
`
`Figure 2.
`Photomicrographs of oligodendrocyte preservation and loss in multiple sclerosis (MS). Panel a, Oligodendrocyte preservation. Many oligodendrocytes are seen
`adjacent to and in the center of a zone of active demyelination (in situ hybridization for proteolipid [PLP] mRNA [black] and immunocytochemistry for PLP protein
`[red]). Panel b, Oligodendrocyte loss. In a second case, oligodendrocytes are absent from a zone of active demyelination but are preserved in the adjacent periplaque
`white matter. (Reprinted by permission from NatureSupplement 399: A45 copyright 1999, Macmillan Magazines Ltd.)
`
`Figure 3.
`Photomicrographs of a chronic multiple sclerosis plaque. In left panel, a well-demarcated hypocellular region of myelin loss is evident in the periventricular white
`matter (Luxol fast blue and periodic acid-Schiff myelin stain, ⫻15 magnification). In right panel, neurofilament staining for axons in the same lesion demonstrates
`a reduction in axonal density. (Reprinted from Noseworthy et al, NEnglJMed2000, 343:938 –952. Copyright © 2000, Massachusetts Medical Society. All rights
`reserved.)
`
`Wright, 1978). There are few or no oligodendrocytes,
`but there may be sizeable numbers of oligodendrocyte
`precursor cells (Wolswijk, 1998).
`Shadow plaques are circumscribed regions where
`axons maintain uniformly thin myelin sheaths; they
`may occur within acute plaques or at the edge of
`chronic ones (Fig. 4). These plaques represent areas
`of remyelination and are macroscopic evidence that
`the CNS white matter possesses the means for self-
`repair. Shadow plaques are seen in conjunction with
`actively demyelinating lesions that retain viable oligo-
`dendrocytes in the plaque center (patterns I and II).
`The next four sections consider questions that con-
`cern the inflammatory mechanisms postulated to lead
`to plaque development in patterns I and II outlined
`above. The potential processes that are operative in
`determining type III and IV pathological patterns are
`discussed in the sections on tissue injury mecha-
`nisms, axonal loss, and non-autoimmune processes
`that result in cell death.
`
`HowDotheConstituentCellsofaPlaqueEntertheCNS
`inImmune-MediatedModelsof
`InflammatoryDemyelination?
`
`An intact blood-brain barrier allows limited passage of
`T lymphocytes that may not have antigen specificity.
`This may be initiated by the interaction of adhesion
`molecules expressed on the surface of lymphocytes
`with complementary integrins present on the endothe-
`lium, resulting in T cell rolling and adherence to the
`luminal surface (Fig. 5). Examples of such molecules
`include vascular cell adhesion molecule (VCAM) and
`intercellular adhesion molecule (ICAM), each ex-
`pressed on endothelial cells; and very late antigen 4
`(VLA-4; also called ␣4-integrin) and lymphocyte
`function-associated antigen-1 (LFA-1), each displayed
`by T lymphocytes. Various selectins are also involved.
`Rolling, adherence, and diapedesis of T lymphocytes
`are modulated by VCAM/VLA-4 and ICAM/LFA-1
`interactions.
`
`266 Laboratory Investigation • March 2001 • Volume 81 • Number 3
`
`4
`
`

`

`Multiple Sclerosis
`
`Figure 4.
`Remyelination in a lesion associated with chronic multiple sclerosis. The area stained pale blue (asterisk) represents a region of partial remyelination (a shadow
`plaque) along the periventricular edge of a lesion in a patient with chronic multiple sclerosis (Luxol fast blue and periodic acid-Schiff myelin stain, ⫻15 magnification).
`NAWM denotes normal-appearing white matter. (Reprinted from Noseworthy et al, NEnglJMed 2000, 343:938 –952. Copyright © 2000 Massachusetts Medical
`Society. All rights reserved.)
`
`Immunochemical studies and gadolinium-enhanced
`brain magnetic resonance imaging (MRI) findings (Fil-
`ippi et al, 1996) indicate that the blood-brain barrier is
`disrupted in MS and experimental allergic encephalo-
`myelitis (EAE), a putative animal model of the disease.
`This disruption is present primarily in active lesions,
`but also to a lesser degree in apparently inactive
`chronic plaques. Disruption and inflammation of the
`barrier facilitates the passage of potentially patho-
`genic cells and antibodies into the CNS (Wisniewski
`and Lossinsky, 1991; Archelos et al, 1999). The mech-
`anism by which the barrier is disrupted is not known,
`but immune interactions are likely the main contribu-
`Interferon-gamma (IFN-␥) and tumor necrosis
`tors.
`factor alpha (TNF-␣), major inflammatory cytokines
`expressed in MS lesions, can induce endothelial cells
`to express VCAM and major histocompatibility com-
`plex (MHC) class II molecules. Viral infection (which
`often precedes clinical exacerbations), the presence
`of bacterial antigens or superantigens, and environ-
`mental factors such as reactive metabolites and met-
`abolic stress may also induce such changes.
`In EAE, activated T-lymphocytes may use P-selectin
`to enter the CNS very early in the disease process
`before the barrier becomes inflamed (Carrithers et al,
`2000). Molecules such as VCAM, ICAM, VLA-4, and
`
`LFA-1 do not appear to have a role in early T cell entry
`(Baron et al, 1993; Steffen et al, 1994). However, once
`the barrier is inflamed, VCAM/VLA-4 and ICAM/LFA-1
`interactions, in conjunction with other factors such as
`CD4-MHC class II binding, allow autoreactive T cell
`diapedesis and entry into the CNS (Archelos et al,
`1993; Baron et al, 1993; Engelhardt et al, 1997;
`Romanic et al, 1997; Steffen et al, 1994). P-selectin
`does not appear to have a role in later EAE stages
`(Engelhardt et al, 1997). Inhibition of VLA-4 reverses
`clinical paralysis in acute EAE and prevents relapses in
`the chronic form of the disease (Yednock et al, 1992).
`In human MS lesions, integrins are expressed on
`inflamed endothelial cells, T cells, and neural cells
`(microglia, oligodendrocytes, and astrocytes) and play
`important roles in developing and maintaining the
`plaque (Archelos et al, 1999; Cannella and Raine,
`1995). Circulating levels of ICAM-1 and VCAM-1 are
`elevated in RRMS, and the profile may differ from
`PPMS, a finding that may allow further dissection of
`the differing pathophysiology of these forms of MS
`(Durán et al, 1999; Giovannoni et al, 1997; McDonnell
`et al, 1999). There are down-regulatory systems that
`probably control the extent of inflammation. For ex-
`ample, TNF-␣-induced VCAM-1 expression may be
`followed by release of soluble VCAM-1, which may
`
`Laboratory Investigation • March 2001 • Volume 81 • Number 3 267
`
`5
`
`

`

`Wingerchuk et al
`
`Figure 5.
`Schematic of the initiation, entry, and subsequent recruitment of immune cells
`into the central nervous system. Both activated and nonactivated T lympho-
`cytes may pass through the intact blood-brain barrier assisted by selectins, but
`this process is facilitated by barrier disruption or inflammation perhaps
`triggered by viral infection or reactive metabolites, among other causes. Barrier
`injury up-regulates adhesion molecules on the endothelial surface causing
`more immune cells to roll, adhere, and diapedese into the CNS. Upon entry,
`they produce matrix metalloproteinases (MMPs), which degrade the extracel-
`lular matrix and further disrupt the blood-brain barrier allowing easier passage
`of
`immune cells and antibodies secreted by B lymphocytes. Following
`formation of a co-stimulated trimolecular complex,
`the T lymphocyte is
`(re)activated and secretes cytokines that cause surrounding immune cells and
`glia to produce chemokines. These chemoattractant substances recruit more
`cells into the CNS, amplifying the inflammatory response.
`
`block adhesion and limit cellular infiltration (Kallmann
`et al, 2000). The control of these systems is only
`partially understood. Modulation of adhesion molecule
`interactions is a logical strategy for MS preventative
`therapies.
`
`HowDoesthePlaqueDevelopinImmune-Mediated
`ModelsofInflammatoryDemyelination?
`
`Once autoreactive T cells have gained entry into the
`CNS, they invade the extracellular matrix aided by
`their secretion of matrix metalloproteinases (MMPs),
`especially MMP-9 (gelatinase B) (Yong et al, 1998b).
`These enzymes degrade the type IV collagen matrix
`and are involved in proteolysis of myelin components
`and regulation of cytokine production,
`including
`TNF-␣ (Chandler et al, 1997). MMPs facilitate conver-
`sion of a number of important pro-molecules that
`require proteolytic processing to their active forms
`(cytokines [eg, pro-TNF-␣ or -␤], cytokine receptors
`IL-6 receptor-␣] and adhesion molecules [eg,
`[eg,
`VCAM, L-selectin]) and regulate growth factor bio-
`availability (eg, releasing fibroblast growth factors
`from cell surfaces) (Izumi et al, 1998; Whitelock et al,
`1996; Yu and Stamenkovic, 2000). MMPs also play a
`role in regulating apoptotic cell death by disrupting
`cell-matrix contacts with the subsequent loss of inte-
`grin signaling (anoikis) (Alexander et al, 1996; Chen
`and Strickland, 1997). Tissue inhibitors of MMPs
`
`268 Laboratory Investigation • March 2001 • Volume 81 • Number 3
`
`(TIMPs) regulate MMP activity by forming molecular
`complexes with them.
`Several lines of evidence suggest that MMPs have a
`central role in lesion formation in EAE and MS. MMP-9
`expression in the CSF of animals with EAE correlates
`with clinical severity (Clements et al, 1997). Inhibitors
`of MMP reduce EAE disease activity and severity
`(Gijbels et al, 1994; Liedtke et al, 1998), and immuni-
`zation with myelin antigens failed to induce EAE in
`young MMP-9 genetic knockout mice (Dubois et al,
`1999). MMPs and TIMPs are present in the serum and
`CSF of MS patients and expressed in plaques (Cuzner
`et al, 1996; Maeda and Sobel, 1996). Serum MMP-9
`and TIMP-1 and TIMP-2 levels are elevated, MMP-9
`levels may be higher during relapses than periods of
`clinical inactivity, and MMP-9 levels may correlate with
`brain MRI markers of inflammation (Lee et al, 1999).
`Other studies have found elevated CSF MMP-9 levels
`in RRMS irrespective of clinical activity and without
`TIMP up-regulation (Leppert et al, 1998). Beta inter-
`ferons, which reduce exacerbation frequency and
`severity in MS, are potent MMP 9 inhibitors (Yong et
`al, 1998a). These therapies may act in part by limiting
`T cell
`infiltration and modulating the production of
`demyelinating cytokines (Uhm et al, 1999). There are
`theoretical disadvantages to these inhibition strate-
`gies, including unknown effects on the MMP activities
`that regulate cytokines and apoptosis.
`The initial invasion of T cells into the CNS is followed
`by recruitment and attraction of secondary inflamma-
`tory cells such as macrophages. Tissue injury and
`demyelination is thought to occur as a result of T cell
`attraction and activation of these cells. Integrins and
`chemokines mediate this recruitment process.
`Chemokines are a superfamily of chemoattractant
`cytokines that are secreted by leukocytes and many
`other cell types (Luster, 1998). The major subfamilies
`are C-X-C (or ␣ family) and C-C (␤ family); more than
`40 chemokines and 10 receptors have been described
`so far. Members of the ␣family attract neutrophils (eg,
`IL-8) or activated T cells (eg,
`interferon inducible
`protein 10 [IP-10]). The ␤ family seems to mediate
`chronic inflammation and individual members are at-
`tractants for monocytes, T cells, or eosinophils. Se-
`lective expression of individual chemokines may influ-
`ence the cellular composition of inflammatory lesions,
`because some chemokine receptors are associated
`with either Th1 or Th2 responses. Th1 proinflamma-
`tory cells may be associated with CCR5 (receptors for
`the chemokines RANTES, MIP-1␣, and MIP-1␤) and
`CXCR3 (receptors for IP-10 and MIG). In contrast, Th2
`anti-inflammatory cells may shift towards display of
`CCR3 (receptors for MCP-3, MCP-4, and RANTES),
`CCR4 (receptors for TARC and MDC), and CCR8
`(Bonecchi et al, 1998; Sallusto et al, 1998). In MS,
`some chemokine receptors,
`including CCR5 and
`CXCR3, may be overexpressed in peripheral and
`lesional T lymphocytes (Zhang et al, 2000), and CSF
`may contain elevated levels of the chemokines IP-10,
`RANTES, and MIG (Sorensen et al, 1999). The specific
`chemokine profiles of importance in MS are unknown,
`but there may be preferential peripheral T cell migra-
`
`6
`
`

`

`tory pattern toward RANTES and MIP-1␣ in associa-
`tion with up-regulation of their CCR5 receptor (Zang et
`al, 2000).
`
`WhichAntigensDoAutoreactiveTCellsTargetin
`Immune-MediatedModelsof
`InflammatoryDemyelination?
`
`Immune dyscontrol, perhaps involving tolerance
`mechanisms, may contribute to the initiating or prop-
`agating of a pathological state by autoreactive T cells.
`The causative autoantigen(s) in MS are still not known,
`but autoreactive T cells exist in both healthy control
`subjects and people with MS. The leading candidates
`are myelin protein constituents. The role of MBP-
`reactive T cells is the most thoroughly studied. Data
`concerning the frequency of MBP-specific autoreac-
`tive cells in MS versus controls are conflicting, but
`their level of activation may be greater in MS (Martino
`and Hartung, 1999). On the other hand, many different
`ligands can induce highly variable responses from
`single anti-MBP T cells, indicating a high degree of
`degeneracy in antigen recognition and challenging the
`role of MBP as a primary autoantigen (Hemmer et al,
`1998; Vergelli et al, 1997). The roles of the other myelin
`components, PLP, MOG, and MAG, and their autore-
`active T lymphocyte counterparts is less well studied.
`Many other putative autoantigens have been impli-
`cated in MS, also with respect to antibody-mediated
`responses (see below).
`Molecular mimicry (similarity among antigens con-
`tained in microbes and human tissue) has been hy-
`pothesized to explain immunological injury in autoim-
`mune diseases (Brocke et al, 1998). Under
`this
`schema, antigens present in or originating from an
`exogenous pathogen activate T cells; these cells then
`induce CNS demyelination by recognizing cross-
`reactive myelin antigens. This explanation has been
`used to implicate HHV-6 in MS pathogenesis (Stein-
`man and Oldstone, 1997) although a latent viral infec-
`tion, rather than mimicry, could also potentially result
`in demyelination and oligodendroglial loss. The T cell
`receptor normally maintains an extremely high level of
`cross-reactivity, probably to balance the requirement
`to recognize non-self antigens and to reduce the
`possibility of loss of self-tolerance (Mason, 1998). The
`concept of molecular mimicry remains speculative.
`
`HowDoTLymphocytesOrchestrateProcessesThat
`CauseTissueDamageinImmune-MediatedModelsof
`InflammatoryDemyelination?
`
`Myelin basic protein-specific cells exist in the T cell
`repertoire of healthy people and those with MS. These
`and other autoreactive T cells may direct the initiation
`of
`the inflammatory response through blood-brain
`barrier disruption and accumulation of secondary in-
`flammatory cells, principally monocytes and macro-
`phages, which in turn mediate tissue injury. In one EAE
`experiment, less than 4% of perivascular inflammatory
`cells were antigen-specific (Cross et al, 1990) and
`these cells remained in the perivascular space while
`the parenchyma was intensely infiltrated with leuko-
`
`Multiple Sclerosis
`
`cytes. In addition, immune mechanisms involving B
`cells, demyelinating antibodies, and complement are
`required. However, there is also evidence supporting
`the concept that the inflammatory reaction may occur
`secondary to or independently from demyelination
`(see “Are There Factors Other Than an Autoimmune
`Response That Might Be Relevant to MS Pathophys-
`iology?”). The following subsections describe pro-
`cesses based on primary immune-mediated models.
`The Trimolecular Complex. Autoreactive T cells re-
`spond to putative MS autoantigens presented by
`antigen-presenting cells through formation of a trimo-
`lecular complex (Fig. 6). Perivascular monocytes, mi-
`croglia and macrophages, parenchymal lymphocytes,
`and possibly astrocytes express MHC antigens in MS.
`There are two principal types of MHC molecules: class
`I (includes HLA-A, -B, and -C) and class II (includes
`HLA-DR, -DP, and -DQ). These molecules bind pep-
`tide antigens as part of the “processing” they require
`for presentation to different T lymphocytes. Lympho-
`cytes of the CD4⫹ type recognize antigens in con-
`
`Figure 6.
`Schema of autoantigen-induced T cell activation and orchestration of the
`immune pathways leading to demyelination and axonal loss. Once in the CNS,
`CD4⫹ T lymphocytes interact with antigen-presenting cells (APC) via the
`trimolecular complex (MHC class II molecules, the T cell receptor [TCR] and
`the inciting autoantigen). In the presence of B7-1 and B7-2 co-stimulatory
`ligands, the complex triggers specific response that depends on the T cell
`co-stimulatory molecules (eg, CD28 ⫽ immune response; CTLA-4 ⫽ anergy)
`and the cytokine milieu (IL-12 leads to Th1 differentiation; IL-4 leads to Th2
`phenotype). Th1 cells secrete TNF␣ and IFN␥, each of which activates
`phagocytic macrophages and are toxic to myelin and oligodendrocytes. IFN␥
`may also stimulate production of MHC class I-restricted CD8⫹ T lymphocytes
`that are directly cytotoxic. The Th2-type cells produce interleukins that result
`in antibody production and complement deposition. The final pathway of
`demyelination and axonal loss may incorporate some or all of these mecha-
`nisms in conjunction with processes (persistent viral infection; toxins) that
`lead to primary oligodendrocyte loss.
`
`Laboratory Investigation • March 2001 • Volume 81 • Number 3 269
`
`7
`
`

`

`Wingerchuk et al
`
`junction with MHC class II molecules, whereas CD8⫹
`lymphocytes recognize antigens in the context of
`MHC class I molecules. The trimolecular complex is
`completed by interaction