grey matter damage in multiple sclerosis

Multiple sclerosis is a chronic immune-mediated demy-elinating disease of the CNS. Its aetiology is thought to involve both a complex genetic component, in which multiple susceptibility genes are important, and several environmental factors14. Multiple sclerosis is character-ized pathologically by inflammation, demyelination with partial restoration of myelin, axonal and neuronal dam-age, and glial scarring5,6. The pathogenetic mechanisms leading to these abnormalities are largely unclear.

As disseminated focal demyelinating lesions in the white matter are the classical hallmark of multiple scle-rosis, the disease has long been considered a typical white matter disease. Nevertheless, early histological reports noted that grey matter is also affected and that grey matter damage and cell loss may underlie some of the permanent neurological dysfunction. In 1916, in a comprehensive overview of the histology of dissemi-nated sclerosis, James W. Dawson wrote: when an area [of demyelination] is confined to the cortex, the changes are, as a rule, not nearly so marked7, refer-ring to the striking lack of glial proliferation and cel-lular paucity (an indicator that inflammation is absent) in grey matter lesions. Dawson also wondered: Is then, the process that attacks the cortex different in its nature and origin from that which affects the rest of the central nervous system?7. This tantalizing question remains highly relevanttoday.

There was little post-mortem tissue research investi-gating grey matter pathology in multiple sclerosis until 1962, when the interest in grey matter lesions in multi-ple sclerosis was rekindled by a seminal histopathology

study8 that reported that 26% of multiple sclerosis lesions were located in or around the cortical and subcortical grey matter. However, it would become apparent that this was a gross underestimation of actual grey matter damage. It was another four decades before advances in imaging techniques9,10, supported by immunohistochem-ical observations and investigations in animal models, renewed the interest in grey matter pathology in multi-ple sclerosis, and began to reveal its high prevalence and to uncover the detailed cellular pathology of grey matter demyelination1113.

Despite the extensive pathological characterization and imaging evidence of grey matter lesions in multiple sclerosis (FIG.1), the causes of grey matter damage and its relationship to white matter lesions remain unclear. This Review aims to provide an overview of existing theories regarding grey matter damage in multiple scle-rosis, answering from a histological, immunological and neuroimaging point of view some of the most pertinent questions regarding its pathogenesis.

Are white and grey matter damage linked?Since the first demonstration that the cortex is not spared by multiple sclerosis, it has remained unresolved whether cortical tissue damage is secondary to white matter pathol-ogy that is, whether cortical damage occurs via retro-grade degeneration or is a primary pathological process. Grey matter lesions are usually characterized by a relative lack of parenchymal lymphocyte infiltration, deposition of antibody and complement proteins, and bloodbrain barrier disruption when compared with white matter

Exploring the origins of grey matter damage in multiple sclerosisMassimiliano Calabrese1, Roberta Magliozzi2,3, Olga Ciccarelli4,5, Jeroen J.G.Geurts6, Richard Reynolds2 and Roland Martin7

Abstract | Multiple sclerosis is characterized at the gross pathological level by the presence of widespread focal demyelinating lesions of the myelin-rich white matter. However, it is becoming clear that grey matter is not spared, even during the earliest phases of the disease. Furthermore, grey matter damage may have an important role both in physical and cognitive disability. Grey matter pathology involves both inflammatory and neurodegenerative mechanisms, but the relationship between the two is unclear. Histological, immunological and neuroimaging studies have provided new insight in this rapidly expanding field, and form the basis of the most recent hypotheses on the pathogenesis of grey matter damage.

1Advanced Neuroimaging Laboratory of Neurology B, Department of Neurological and Movement Sciences, University Hospital Verona, Piazzale Ludovico Antonio Scuro 10, 37134, Verona, Italy. 2Division of Brain Sciences, Faculty of Medicine, Imperial College London, Hammersmith Hospital, London W12 0NN, UK.3Department of Cell Biology and Neurosciences, Istituto Superiore di Sanit, Viale Regina Elena 299, Rome, Italy.4National Institute for Health Research, University College London/University College London Hospitals NHS Foundation Trust (NIHR UCL/UCLH) Biomedical Research Centre, 149 Tottenham Court Road, London W1T 7DN, UK. 5Queen Square Multiple Sclerosis Centre, University College London, Institute of Neurology, Queen Square, London WC1N 3BG, UK. 6Section of Clinical Neuroscience, Department of Anatomy and Neurosciences, VU University Medical Center, van der Boechorststraat 7, 1081 BT Amsterdam, The Netherlands.7Neuroimmunology and Multiple Sclerosis Research Section, Department of Neurology, University Hospital Zurich, University of Zurich, Frauenklinikstrasse 26, 8091 Zurich, Switzerland. Correspondence to M.C. email: [email protected]:10.1038/nrn3900

R E V I E W S

NATURE REVIEWS | NEUROSCIENCE VOLUME 16 | MARCH 2015 | 147

2015 Macmillan Publishers Limited. All rights reserved

mailto:calabresem%40%20hotmail.it?subject=Nature%20Reviews

Complement proteinsA set of plasma proteins that coats pathogens; the coated pathogens are then cleared by phagocytes.

MRI(Magnetic resonance imaging). A non-invasive method used to obtain images of living tissue. It uses radio-frequency pulses and magnetic field gradients; the principle of nuclear magnetic resonance is used to reconstruct images of tissue characteristics (for example, proton density or water diffusion parameters).

Wallerian degenerationThe degeneration of an axon distal to a site of injury, which begins to occur approximately 1.5days after the injury.

Clinically isolated syndromeA syndrome present in a patient experiencing their first clinical episode that is suggestive of an inflammatory demyelinating disease of the CNS.

Relapsingremitting multiple sclerosis(RRMS). The early phase of multiple sclerosis characterized by several neurological episodes followed by complete or incomplete recovery.

Radiologically isolated syndromeA syndrome present in a patient who has radiological evidence of an inflammatory demyelinating disease of the CNS but no clinical signs or symptoms of such disease.

lesions12,14,15. In post-mortem samples taken from indi-viduals in the advanced stages of multiple sclerosis, high numbers of immune cells were only detected in type1 grey matter lesions (FIG.1a); all other grey matter lesions were categorized as relatively non-inflammatory1113. Moreover, according to post-mortem studies, neuronal loss occurs within both grey matter demyelinated lesions and normal-appearing grey matter12,16,17, suggesting that neuronal loss might occur independently from grey matter demyelination.

These findings imply that grey matter damage might be a result of non-inflammatory processes or local inflam-mation that have occurred within the grey matter, or a consequence of white matter lesions that are themselves caused by inflammation.

Evidence that links grey and white matter damage. Consistent with the idea that white matter damage may drive grey matter damage, several MRI studies have reported a significant correlation between white mat-ter damage and grey matter atrophy in conditions other than multiple sclerosis18. For example, advanced post-processing analyses showed a reduction of cortical grey matter volume and thinning of the primary motor cortex in patients with spinal cord injury, suggesting that dis-tant axonal transection can lead to significant cortical atrophy via retrograde degeneration18.

It has been suggested that the characteristic pattern of white matter pathology in multiple sclerosis could lead to selective retrograde injury to frontal, temporal and motor cortical areas, which would partially explain some of the post-mortem patterns of grey matter dam-age that have been observed19. Indeed, the corticospinal tract, which contains fibre connections to the precen-tral gyrus, and the frontal periventricular white matter, which consists of efferent and afferent fibres to frontal and superior temporal lobes, are common sites of white matter lesions in multiple sclerosis20. Moreover, axonal damage in active white matter lesions (that is, those that are gadolinium-enhanced in contrast MRI) could lead indirectly to anterograde and retrograde degeneration (FIG.2) of axons running within the thalamus and basal ganglia. Axon loss might also contribute to demyeli-nation and/or Wallerian degeneration by reducing local metabolic activity21,22. A correlation betweenwhite matter lesion load andthalamicatrophy was observed in patients with clinically isolated syndrome23 and early relapsingremitting multiple sclerosis (RRMS)24.

More generally, the assumption that white mat-ter lesions and grey matter atrophy are linked is sup-ported by several cross-sectional MRI studies showing significant correlations between the total grey matter volume and the total white matter volume in T1- and T2-weighted lesions2528. Moreover, a sensitive regional analysis revealed a correlation between regional cortical grey matter loss and white matter lesion volume in the corresponding or adjacent lobes in patients with pro-gressive multiple sclerosis29, although the location and extent of grey matter lesions were not studied. Finally, several longitudinal MRI studies3032 indicate an asso-ciation between accumulation of white matter lesion

volumes and grey matter volume loss in the frontal and parietal cortex of patients with RRMS. In addition, white matter lesion volume correlated with ventricular enlargement32, focal white matter damage in the optic radiations (FIG.3a) and upstream grey matter atrophy of the lateral geniculate nucleus and visual cortex in patients withRRMS33.

Evidence for independent grey matter damage. An alternative but equally important theory suggests that white matter and grey matter demyelination are two, at least partly, independent phenomena and that neu-ronal loss is not caused by white matter abnormalities perse. Proponents of this theory suggest that retro-grade changes as a result of focal white matter lesions do not satisfactorily explain the extent of the cellular pathology and diffuse nature of grey matter changes (FIG.3b,c).

Extensive pathological and imaging evidence sup-ports this argument. For example, several recent stud-ies of grey and white matter demyelination in different regions of the CNS found that the extent of corti-cal demyelination is greater than that found in white matter34. In addition, althoughcorticaldemyelina-tionsometimes occurred together withdemyelina-tionin the adjacent white matter (leukocortical lesions), in most instances thecortexwas affected independently from white matter lesions35,36. Another study observed a gradient of neuronal loss in the precentral gyrus of multiple sclerosis cases that exhibited extensive subpial demyelination, with the greatest loss in the outer corti-cal layers17. There was no relationship between this gra-dient and the white matter lesion volume or location, which argues strongly against an influence of white mat-ter lesions on grey matter subpial pathology. However, it is likely that extensive primary neuronal loss in the cortical grey matter would lead to anterograde loss of axons in downstream white matter pathways, including the spinalcord.

In addition, extensive and complete subpial demyeli-nation of individual gyri and sulci accompanied by the relative preservation of axons characterizes many cases of secondary progressive multiple sclerosis (SPMS)12,37. This pattern of damage cannot be explained by axonal damage in white matter lesions, which would lead to a more diffuse but incomplete decrease in myelin den-sity due to the loss of individual myelinaxon units. Extensive subpial demyelination was also observed in single biopsies of very early RRMS38. Invivo imaging studies confirmed the presence of grey matter lesions during the earliest phases of the disease39 in patients with very low white matter lesion volume, and some-times even in patients with radiologically isolated syn-drome (before any clinical symptoms are present)40, or preceding the occurrence of white matter lesions alto-gether41. However, it should be noted that even when using the most advanced technologies, most grey mat-ter lesions (especially type 3 subpial lesions (FIG.1c)) escape identification by MRI, making it difficult to cor-relate grey matter and white matter damage accurately by this method42.

R E V I E W S

148 | MARCH 2015 | VOLUME 16 www.nature.com/reviews/neuro


Primary and secondary progressive diseasePhases of multiple sclerosis characterized by a slow progression of disability without a well-defined clinical relapse. These phases usually follow the relapsingremitting phase (secondary progressive phase) but they can also be in the first phase of the disease (primary progressive multiple sclerosis).

TcellA lymphocyte that mediates cell-dependent immune responses by providing help (in the form of cytokines, for example) to other immune cells or by cytotoxicity (killing of a virus-infected cell).

Other radiological data suggested that diffuse grey matter atrophy is only partially related to the white matter damage visible by MRI. When brain atrophy was studied in patients with early RRMS, for example, both grey and white matter atrophy were observed, but the correlation between the two was only moderate43. A 2-year period of follow-ups of these patients revealed that fluctuations in inflammatory white matter lesions were related tochangesin white matter volume rather than to grey matter loss44. Voxel-based morphometry and tract-based spatial statistical analyses were applied to assess regional grey matter and white matter damage in 36 RRMS and 25 sex- and age-matched controls45. As expected, this analysis revealed significant differences between patients and controls. However, although sev-eral measures of white matter damage correlated with each other, none correlated with grey matter volume45. More recent imaging studies have found that the rela-tionship between grey matter atrophy and white mat-ter abnormalities was weaker in primary and secondary progressive disease than in RRMS, suggesting that neu-rodegenerative processes in patients with progressive

disease are even less related to white matter changes46. These findings could indicate that grey matter damage is either independent of white matter lesions or that white matter is affected in a way that is not detected by currently applied imaging techniques.

Finally, grey matter lesion probability maps and regional analysis of grey matter atrophy revealed a specific topographical distribution of focal and diffuse damage47. Most grey matter lesions were found in the frontotemporal lobes, with a particular prevalence in the motor regions. A high probability of focal grey mat-ter demyelination was observed in the hippocampus, deep grey matter and insula48,49, whereas a significant increase in the distribution of diffuse subpial demyeli-nation was observed by means of 7 Tesla MRI in the anterior cingulate cortex50. Grey matter demyelination has also been observed in the limbic system and in spe-cific cortical areas such as the mesial temporal lobe51. Nevertheless, the spatiotemporal topographical distri-bution of white matter lesions did not reveal any direct anatomical overlap between areas of significant reduc-tions in grey matter volume and significant increases in focal white matter lesion volume52.

Does grey matter damage occur as result of or inde-pendently of white matter damage? As described above, evidence exists for both. There is little doubt that axonal damage can lead to retrograde loss of neurons. However, this mechanism does not explain the spectrum and extent of grey matter damage in multiple sclerosis, and several studies clearly suggest that the mechanisms of damage within white matter and grey matter may be, at least partly, independent. Furthermore, primary damage to cortical neurons could give rise to anterograde axonal loss, particularly in the spinalcord.

Inflammatory grey matter damageAlthough there is limited information on the immune mechanisms that are active in the cortical layer in patients with multiple sclerosis, several causes of inflammatory demyelination within the grey matter can be envisioned (FIG.4). These include the possible expression or rela-tive overabundance of a target autoantigen for adaptive (Tcell- and antibody-mediated) immune mechanisms53 or the involvement of a Tcell with specificity for both a myelin and neuronal antigen54.

An alternative hypothesis is that the presence of inflammatory infiltrates in the meningeal space and/or in the adjacent perivascular spaces might lead to the release of cytotoxic inflammatory mediators into the grey mat-ter17,38,55. This could be caused by an infectious organism located in the adjacent meninges or by a chronic com-partmentalized inflammatory response to a self-antigen or self-antigens. Several epidemiological studies have suggested that EpsteinBarr virus (EBV) is one of the strongest candidates for this infectious agent5658. EBV proteins and RNA have been detected in B cells in the meninges and perivascular spaces of patients with mul-tiple sclerosis with extensive meningeal infiltrates and cortical demyelination5961. It has been proposed that failure to control latent EBV infection in an immune privileged site, such as the subarachnoid space, could lead

Figure 1 | Cortical lesion subtypes in multiple sclerosis. From a neuropathological point of view, grey matter lesions in multiple sclerosis tissue are grouped into several types. a | Type1 (leukocortical) lesions extend through grey matter into the white matter and do not usually reach the surface of the brain. b | Type2 lesions are contained within the grey matter and are often located around a blood vessel. Arrows in part a and b indicate the positions of lesions. c | Type3 lesions are subpial lesions that sometimes stretch around several gyri10; they are the most common type of grey matter lesion present at autopsy11 and are the most specific features of multiple sclerosis pathology as they are not seen in any other human inflammatory or neurodegenerative disease. Arrows indicate lesion borders. d | Image of the cingulate gyrus and corpus callosum, illustrating the extensive nature of some subpial lesions. Arrows indicate gyri. In this case, the lesion extends around an entire gyrus and into the next gyrus. Some investigators also define type4 lesions, which are large, cortex-spanning lesions that do not pass the grey matterwhite matter border (not shown).

Nature Reviews | Neuroscience

a b

c d

White matter

Grey matter Grey

matter

Blood vessel

Grey matter

White matter

Cingulate gyrus

Corpus callosum

R E V I E W S



BcellsLymphocytes that express immunoglobulins as surface receptors or, when they are fully mature following antigenic stimulation, release antibodies that are directed against a virus or bacteria.

to recurrent intrathecal reactivation of EBV and tissue damage in the nearby grey matter62,63. However, several studies were unable to detect EBV in the brain or lesions of patients with multiple sclerosis64,65, and this remains a controversial and highly debated issue66. The differ-ent results found in these studies are likely to be due to the different types of tissues and multiple sclerosis cases examined, the variable preservation of meningeal tissues in the samples and the differences in the sensitivity of the techniques used to detect EBV infection67.

Other viruses, such as torque teno virus, have also been hypothesized to play a part in the pathogenesis of multiple sclerosis because they can infect the CNS68, are a target for cerebrospinal fluid (CSF)-infiltrating Tcells isolated from patients with multiple sclerosis69 and interact with EBV68,69. Furthermore, myco-like viruses that are closely related to torque teno virus have recently been isolated

from cow milk and the brains of patients with multiple sclerosis, thus providing an interesting link to epidemi-ological findings on the consumption of cow milk and the geographical prevalence rates of multiple sclerosis70. Alternatively, cortical demyelination and neuronal loss could involve an infectious agent with primary tropism for oligodendrocytes and/or cortical neurons71, although no evidence exists yet for this possibility in multiple sclerosis.

Other mechanisms that might target adaptive and/or innate immune responses to the grey matter include alterations specifically linked to the function and/or degeneration of neurons, astrocytes or oligodendrocytes. These alterations might include metabolic changes72, excitatory neurotransmitter release73, expression of post-translationally modified proteins and/or peptides74, changes in electrical activity and/or ion currents, and cytokine and/or cytokine receptor expression75.

Figure 2 | Inflammatory and non-inflammatory grey matter neurodegeneration mechanisms. a | In the undisturbed grey matter, microglia typically exist in a resting state and continually survey the microenvironment with their motile processes and protrusions (upper panel). Pro-inflammatory and cytotoxic molecules released by inflammatory cells within subarachnoid or intracortical immune infiltrates, as well as cell contact-dependent mechanisms of Tcell-mediated damage, may lead to microglia and/or macrophage activation and oligodendrocyte injury (middle panel). This process can directly or indirectly lead to death of the neuronal cell body and nuclei (lower panel). Such cell death leads to morphological alterations, such as the characteristic pycnotic nuclei, shrinkage of dendrites and axonal degeneration in the cerebral cortex. This could in turn lead to dysfunction of the downstream neuronal network via anterograde trans-synaptic (Wallerian) degeneration. b | Undisturbed neuronal connectivity, and its related functions (upper panel), could be impaired by retrograde degeneration (middle panel) propagating backwards in cortical neurons whose axons have been damaged in white matter lesions or along white matter tracts such as the corticospinal tract. The white matter tract damage can lead to microglial activation (middle panel) and retrograde neuronal cell death (lower panel). White matter damage could also lead to anterograde Wallerian degeneration in the part of the neuron below the damage. Injury at the terminal endings could also spread back towards the cell body and induce microglial activation (middle panel), and therefore dying-back of neurons (bottom panel). At the end of this process, macrophages act to clear the cellular and myelin debris. Even in non-inflammatory neurodegeneration, activation of microglia and macrophages would probably contribute to the damage and its propagation to the surrounding area.


a Inflammatory neuronal damage b Non-inflammatory neuronal damage

Microglia

Activated microglia

Macrophage

CD8+ T cell

B cellCD4+ T cell

Oligodendrocyte

Pro-inflammatory and cytotoxic molecules

Anterograde Wallerian degeneration

Pycnotic nuclei

Neuronal connectivity


Retrograde degeneration


Neuronal dying back

Damage

Retrograde degeneration

Degeneration

R E V I E W S



Although none of these possible reasons has yet been proven to be a cause of the topographical distribution of grey matter damage in multiple sclerosis, there are data in support of these ideas, and we therefore believe that they can at least serve as a working hypothesis and basis for further investigation.

Role of parenchymal and perivascular lymphocytes. A recent study demonstrated that the extent of lym-phocyte infiltration in grey matter lesions depends on the disease stage: foamy macrophages, thought to be involved in ongoing demyelination, were found in grey matter active lesions of 66% of patients with early RRMS at biopsy38 but were rarely found in grey matter lesions from patients with SPMS37. The presence of perivascu-lar CD8+ Tcells was observed in 77% of intracortical lesions at biopsy with varying frequency depending on the location examined; numbers decreased from leuko-cortical regions to intracortical lesions and were low-est in subpial lesions12,38. Perivascular T cell and B cell infiltrates have also been observed in both intracorti-cal and subpial lesions in progressive multiple sclerosis cases55, particularly in subjects in whom progressive disease with signs of activate inflammation was present at death. The reasons for the heterogeneous distribu-tion of CD8+ Tcells in the cortex, their relatively higher frequency in leukocortical and intracortical plaques in comparison to subpial plaques and their localization in perivascular cuffs38 remain unclear, and further studies are clearlyneeded.

A wealth of immunological data suggests that CD4+ Tcells are inducers and drivers of multiple sclerosis2. Key findings that support this notion are, among oth-ers, the strong genetic influence of the HLA-DR15 haplotype (see below)76,77, the increased frequency of myelin-specific CD4+ Tcells with a pro-inflammatory phenotype in multiple sclerosis78,79 and their higher antigen avidity80. However, although CD4+ Tcells are present in the inflamed meninges above subpial grey matter lesions17, little information is available concern-ing their presence within cortical grey matter lesions. This is surprising if one considers that HLA classII mol-ecules (including HLA-DR alleles) serve as recognition structures for CD4+ Tcells and that the HLA-DR15 hap-lotype is by far the most important genetic risk factor in multiple sclerosis76,77. In fact, a recent study suggested that HLA-DRB1*15 status is associated with the extent of inflammation and demyelination in the motor cortex at autopsy81, indirectly suggesting that CD4+ Tcells (or an as yet unknown role of HLA classII molecules) are involved in tissue inflammation in the greymatter.

Despite the evidence implicating a role for CD4+ Tcells in multiple sclerosis, CD8+ Tcells are found at higher frequency than CD4+ Tcells in the grey mat-ter and brains of patients with multiple sclerosis at autopsy38,82, and are more frequently clonally expanded83, suggesting that CD8+ T cells have proliferated locally in the brain. However, the local proliferation of T cells within the grey matter lesions themselves has not yet been examined, and CD4+ intracerebral Tcells have generally been studied to a lesser extent than CD8+ cells. CD8+ Tcells are also thought to be more likely to directly damage neurons and other CNS cells than CD4+ Tcells because HLA classII molecules are expressed at very low levels and on a limited number of cells in the CNS, whereas HLA classI molecules, which present antigen to CD8+ Tcells, show broader expression in multiple scle-rosis brains75. However, although HLA classII molecules

Figure 3 | Is grey matter atrophy a primary or secondary pathological process? An important issue in multiple sclerosis concerns the relationship between grey and white matter damage. In particular, it is unclear whether grey matter damage is secondary to white matter pathology or is a primary pathological process. MRI can be used to study the relationship between white matter lesions and grey matter atrophy. a | MRI images of a patient with relapsingremitting multiple sclerosis (RRMS) (T1-weighted, fluid-attenuated inversion recovery and double inversion recovery sequences) showing marked demyelination of the optic radiations (arrows) and marked atrophy of the calcarine sulcus (arrowheads) b | MRI images of a patient with RRMS. These show a very low white matter lesion load and a severe cortical atrophy (arrowheads) of the right and left central sulcus (note that the sulcus is enlarged and that the patient has no spinal cord lesions). c | MRI images of a healthy control. Collectively, these images suggest that severe atrophy of a cortical region can be the consequence of severe demyelination of the white matter fibres that originate from that region (as in part a), but it can also be present in the absence of any MRI-visible white matter or grey matter lesion (as in part b). The MRI images were acquired using a 3 Tesla Philips scanner. Images taken by M.C. at the Advanced Neuroimaging Laboratory of Neurology, Radiology Unit BT, University Hospital of Verona, Italy.


a

c

b

T1-weightedFluid-attenuated inversion recovery

Double inversion recovery

R E V I E W S



are not expressed by neurons, they are expressed by activated microglia and astrocytes. Proteins and pep-tides that are released from damaged neurons can be presented by these cells to CD4+ Tcells84, although this process has not been examined yet in the grey matter of patients with multiple sclerosis. Furthermore, HLA classI molecule expression has been described in electri-cally silent neurons (those that are functionally compro-mised or damaged), and data from experimental systems suggest that the exposure of neurons to interferon- (IFN) renders them immunological targets for CD8+

Tcells75. In summary, the existing evidence suggests an involvement of both CD4+ and CD8+ Tcells in cortical damage in multiple sclerosis, although they may make different contributions to the cortical degeneration. That is, initiation and perpetuation may be driven by CD4+ Tcells, whereas effector mechanisms such as direct neu-ronal and axonal damage are more likely to be caused by CD8+ Tcells.

Several mechanisms through which CD8+ and CD4+ Tcells could damage neurons and oligodendrocytes have been proposed. These include antigen-specific

Figure 4 | Immune-mediated mechanisms of subpial cortical demyelination in progressive multiple sclerosis. Schematic depiction of the immunological and pathological events associated with neurodegeneration, oligodendrocyte and astrocyte damage, and increased microglial activation in patients with progressive multiple sclerosis. Inflammatory cells in both meningeal infiltrates (which may be both diffuse and localized in the tertiary lymphoid-like tissue) and intracortical immune infiltrates could either directly or indirectly induce neuronal, axonal and oligodendrocyte damage. Direct damage can be caused via the release of inflammatory mediators (including interferon- (IFN), typeI IFNs, tumour necrosis factor (TNF), other cytokines, chemokines and possibly antibodies) and cytotoxic factors (including matrix metalloproteinases (MMPs), granzymes and perforin). Indirect damage can be caused via intense microglial activation. Although myelin-laden macrophages have been found only in actively demyelinating cortical lesions during early multiple sclerosis, microglial activation is a diffuse event occurring in grey matter lesions and normal-appearing grey matter, and has a central role in grey matter cell injury. Microglia are involved in both the release of inflammatory stimuli (such as TNF, nitric oxide (NO) and myeloperoxidase (MPO)) and also in altered glutamate reuptake, and consequently neuronal and synaptic damage due to excitotoxicity. Astrocyte loss and/or dysfunction may also have a key role in grey matter injury. Astrocyte loss may contribute to glia limitans destruction and therefore favour the diffusion of soluble factors from the subarachnoid space towards inner cortical layers. Furthermore, astrocytes have a fundamental role in maintaining the oligodendrocyteaxonalsynaptic apparatus and in modulating the movement of molecules through the cortex. Microglial activation, astrocyte dysfunction and the potential direct cytotoxic and/or myelinotoxic effects of inflammatory cells may all be involved in oligodendrocyte damage and myelin destruction. Finally, there is accumulating evidence highlighting the important role of mitochondrial dysfunction in axonal loss and in neurodegeneration.

Subarachnoid space Antibodies IFN Type I IFNs TNF MMPs Cytokines Chemokines Granzymes Perforin

B cell

Follicular dendritic cell

Plasma cell

Meningeal tertiary lymphoid-like tissue

Macrophage

CD4+ T cell

CD8+ T cell

Dura

Pia

Dendritic cell

Follicular helper T cell


Grey matter Grey matter lesion

Microglial activation

Astrocyte Activated microglia

Intracortical immune infiltrate

Oligodendrocyte

Excitotoxic damage

Astrocyte dysfunction,glia limitans damage

Mitochondrial dysfunction

Neuron

TNF, NO, MPO

Oligodendrocyte dysfunction

Myelin destruction

Neuro-degeneration

R E V I E W S



Natural killer cellsA white blood cell population that does not express antigen-specific recognition receptors such as those expressed by T and Bcells, but recognizes cells that express fewer or no HLA classI molecules (such as virus-infected cells). Natural killer cells are important in controlling viral infections and recognition of mutated (tumour) cells.

Experimental autoimmune encephalomyelitisAn animal model of multiple sclerosis that is initiated in animals by injecting myelin proteins or peptides to raise autoreactive Tcells or by the transfer of autoreactive Tcells into naive recipients.

and antigen-independent mechanisms as well as cell contact-dependent and -independent processes85. However, all of the major cytotoxic mechanisms that have been proposed have been described mainly in experimental models. These mechanisms include the release of perforin86 or granzymes87, the involvement of FAS (also known as CD95) and FAS ligand (also known as CD95L)88, cell lysis by TRAIL (also known as TNFSF10)89, and the induction by IFN of neuron-specific, calcium-permeable complexes between the IFN receptor and glutamate receptor 1 (REF.90). It is important to note that the ability to kill target cells by perforin- or FASFAS-ligand-mediated lysis, which is typically attributed only to CD8+ Tcells, has also been shown for human CD4+ Tcells91. Furthermore, an antigen-independent lysis mechanism that involves CD56 (also known as NCAM), which is expressed on natural killer cells and a subset of CD4+ Tcells92, has been implicated in the lysis of oligodendrocytes93.

Role of meningeal inflammatory infiltrates. Several neuropathological studies in the past decade have shed light on the association between meningeal inflammation and pathology of the adjacent cerebral cortex, both in multiple sclerosis and in experimental autoimmune encephalomyelitis rodent models17,38,9498. Aberrant tertiary lymphoid-like structures contain-ing large aggregates of CD20+ B cells17,37,94, which are hallmarks of several other chronic inflammatory diseases, were found in the inflamed meninges of a sub-stantial proportion of SPMS cases that were examined post mortem. Similar accumulations of B cells have been reported in 40% of cortical biopsies from early RRMS cases, in which they are associated with underly-ing subpial demyelination38. The SPMS cases exhibiting lymphoid-like immune cell aggregates were character-ized by a high degree of inflammatory activity17,37,94,99. The study was carried out on 123 post-mortem multiple sclerosis cases with a wide range of ages at onset, pro-gression and death, and total disease duration, to avoid selection bias37. However, an independent smaller study failed to find such lymphoid structures100. As previously discussed, this disparity may be a result of differences in the technical approaches used, such as different pro-cedures of tissue processing, cutting and staining, and cohort choice65,101. Organized meningeal infiltrates were present predominantly in multiple sclerosis cases char-acterized by an earlier age at death and with evidence of ongoing inflammatory activity; such cases are not always present in autopsy collections. In addition to the organ-ised lymphoid-like structures that have been reported, extensive diffuse meningeal infiltrates have been detected at autopsy100,102,103, which may explain the widespread nature of subpial demyelination. Subpial demyelination and cortical atrophy are more pronounced within deep invaginations of the cortex16,17,35, suggesting that regional differences in CSF flow and/or stasis may result in a shielded niche (or microenvironment) for the persistence of lymphoid-like structures within cerebral sulci. It can be speculated that this may sustain a local immune response that is particularly enriched in CD20+ B lymphocytes

and plasmablasts, but also comprising CD4+ and CD8+ Tcells and macrophages. This response may chronically generate inflammatory, cytotoxic and possibly myelino-toxic mediators that, by circulating within the CSF, may diffuse freely throughout the subarachnoid space. These mediators might cross the pial membrane towards the adjacent grey matter and specifically mediate diffuse and focal subpial grey matter injury in multiple scle-rosis. The release and circulation of specific, but as yet unknown, factors in the CSF bathing the cerebral cortex could explain the fact that subpial cortical demyelination is one of the most specific features of multiple sclerosis pathology104. Among these factors, tumour necrosis fac-tor and IFN may play a fundamental role in mediat-ing subpial pathology directly or indirectly by regulating microglial activity. Indeed, the expression of tumour necrosis factor and IFN is significantly increased both in the meninges and CSF samples from post-mortem SPMS cases with higher levels of meningeal inflammation and cortical damage99. This hypothesis is also corroborated by the finding that a subset of patients with SPMS with higher levels of inflammation and frequency of immune cell infiltrates in the meninges had a higher subpial grey matter lesion volume, a higher degree of neuronal, astro-cyte and oligodendrocyte loss, and increased microglial activation17,37. The severity of these parameters followed a gradient from the external cortical layers towards the innermost cortical layers17,37.

Interestingly, the subpopulation of SPMS cases with higher levels of meningeal inflammation and grey matter damage were characterized clinically by a more rapidly progressive disease course17,37,94,105, although it should be noted that this population of patients clearly represent the more aggressive end of the spectrum of heterogeneous presentations. The link between cortical pathology and a rapidly progressive disease course has also been described in a large proportion of multiple sclerosis cases using non-conventional imaging techniques39,105. Collectively, these results support the idea that increased grey matter pathology, alone or in addition to white matter pathol-ogy, may be associated with a more rapid and aggressive disease course from early phases of the disease and that meningeal inflammation may be partly responsible for this increased pathology.

The specific link between meningeal inflammation and severity of grey matter pathology and rapidly pro-gressive disease course is further supported by studies of the spinal cord of post-mortem cases with SPMS106, in which the levels of both meningeal Tcells and activated parenchymal microglia were associated with increased diffuse axonal loss. Furthermore, a recent study exam-ining a large number of brain biopsies from patients with recently diagnosed multiple sclerosis38 has indeed revealed that cortical lesions occur early in the disease and are highly inflammatory, and associated with both focal perivascular and diffuse meningeal inflammation. The reported association between meningeal inflamma-tion and immune infiltrates in chronic subpial lesions, either in brain biopsies from patients with early multi-ple sclerosis38 or in post-mortem SPMS cases with more aggressive grey matter pathology94, indicate that both

R E V I E W S



meningeal and intraparenchymal grey matter inflam-mation may have a role in disease pathogenesis. They may have this effect either by causing cortical damage at the beginning of the disease and/or by exacerbating grey matter pathology as disease progresses.

Role of microglia. Microglia are innate immune cells that are resident in the CNS and are considered to be the dominant effector cell population in cortical grey matter injury12. Elongated microglia that are oriented perpen-dicularly to the pial surface and are closely apposed to apical dendrites and axons, together with activated stel-late microglia that extend processes to neuronal peri-karya, dendrites and axons, are observed in active and chronic grey matter lesions107. The number of activated microglia and the degree of activation correlate with the density of transected neurites in grey matter lesions12, suggesting that dendrites and axons are vulnerable to microglial activation107. Microglial activation could also be a consequence of neuronal and axonal damage and thus a sign of prior injury108. However, in post-mortem multiple sclerosis brain tissue, the number of micro-glia in cortical lesions also correlates with the degree of meningeal immune cell infiltration37. Therefore, it is unclear whether microglia have a neuroprotective or detrimental role within the cortex in multiple sclerosis, and what the sequence of eventsis.

In the mature brain, microglia typically exist in a rest-ing state and continually survey the microenvironment with their motile processes and protrusions109. Bloodbrain barrier disruption, brain injury or immunologi-cal stimuli provoke immediate and focal activation of microglia110. Depending on the circumstances of activa-tion, microglia may differentiate into a type1 phenotype, which can be pro-inflammatory and detrimental, or a type2 phenotype, which is immunomodulatory, supports survival and provides an antioxidant defence111. However, the phenotype of activated microglia in cortical multiple sclerosis lesions has yet to be determined. For microglia to remain in a chronic state of activation in the multi-ple sclerosis cortex, there would need to be a chronic stimulus. It can be hypothesized that the presence of a chronically inflammatory milieu in the subarachnoid99 and perivascular spaces61 might represent such a stimu-lus. Other stimuli that promote microglial over-activation and dysregulation include environmental toxins and neuronal death or damage. However, the conditions that determine whether microglial activation is detrimental or beneficial to neuronal survival are currently poorly understood in multiple sclerosis.

Nevertheless, it is becoming more widely accepted that microglial activation is necessary and crucial for host defence and neuronal survival, whereas their over-activation may be deleterious to neurons and oligodendrocytes112. In post-mortem homogen-ates of demyelinated and non-demyelinated cerebral cortical regions from multiple sclerosis cases, grey matter demyeli nation has been shown to be associ-ated with increased activity of myeloperoxidase113. Myeloperoxidase is expressed by a CD68+ subset of activated microglia found in active grey matter

demyelination towards the edge of the lesions but not by microglia in adjacent non-demyelinated cortex113. Moreover, the presence of activated microglia in grey matter lesions in patients with multiple sclerosis has been correlated with focal loss of the glutamate trans-porters excitatory amino acid transporter 1 (EAAT1) and EAAT2, and synaptophysin immunostaining114, suggesting that activated microglia may also per-turb astrocyte function and glutamate metabolism. Alterations in the mechanisms of glutamate reuptake found in grey matter lesions in the presence of activated microglia could be associated with signs of neuronal and synaptic damage suggestive of excitotoxicity114 (see below). Furthermore, an increase in microglial activ-ity together with neuronal and axonal loss has been found in the subpial cortex of patients with SPMS with increased diffuse and organised meningeal inflamma-tion37,94. Increased numbers of activated CD68+ micro-glia have also been observed in active and/or chronic cortical lesions of patients with primary progressive multiple sclerosis at autopsy103. Moreover, activated microglia are found along the border of grey matter lesions in a large proportion of patients with multiple sclerosis with extensive subpial demyelination115. These patients were younger at the time of death than patients without grey matter lesions or patients without rims of activated microglia in grey matter lesions100.

Non-inflammatory neurodegeneration?A recent theory highlighted the inconsistencies in the inflammatory model described above and proposed a degenerative model as the primary cause of multiple sclerosis. In the degenerative model, a primary cytode-generation that is initially focused on oligodendrocytes and/or neurons begins years before any clinical symp-toms appear. The autoimmune inflammatory reaction that is later observed would depend on the hosts predi-lection to react to the antigens released as a consequence of the cellular degeneration. This hypothesis arose as a result of inconsistencies in the literature, including observations of early myelin protein degradation before the prototypical adaptive immune response and the inability to stop disease progression using potent anti-inflammatory drugs116118.

However, although this hypothesis seems plausible, other interpretations are more likely. For example, adaptive immune processes that initiate white matter and grey mat-ter damage may later when demyelination and indirect injury to axons as well as direct axonal transections have reached a critical level lead to ongoing neurodegenera-tion that progresses independently of new inflammation. Alternatively, the efficacy of current immunomodula-tory therapies may decrease as the immune response that drives multiple sclerosis pathology becomes increasingly compartmentalized within the CNS or CSF. According to this hypothesis, ongoing inflammation would require less infiltration of cells from the peripheral immune system and therefore might not respond to systemically delivered therapies that primarily act as immunomodulators in the peripheral immune system. It is likely, as has been sug-gested previously116, that multiple sclerosis involves both

R E V I E W S



inside-out and outside-in disease mechanisms, jointly con-tributing to its variable course. However, large genome-wide association studies support a primary role of immune system alterations because the vast majority of the more than 100 common genetic variants119 that have been identified so far are immune-related.

Conclusions and perspectivesThe evidence reviewed above suggests that damage to the grey matter in multiple sclerosis proceeds in a manner that is partly independent of white mat-ter damage, although there is clearly some degree of relationship between the two types of damage (BOX 1). This relative independence is illustrated by the fact that some cases of multiple sclerosis exhibit predomi-nantly grey matter pathology with little white matter involvement, and vice versa. Neuronal degeneration inevitably follows axonal transection in white matter lesions and leads to a variable and dispersed loss of neurons in grey matter areas related to the anatomical location of the white matter lesions. Neuronal loss in the grey matter as a result of inflammatory pro-cesses or more intrinsic mechanisms of degeneration (such as a local energy deficit due to mitochondrial dysfunction) (BOX2) also leads to axon loss and degeneration of downstream neurons. Whether grey matter damage may, at some point in the disease, pro-ceed owing to a primary degenerative process inde-pendent of inflammation and independent of white matter damage remains to be resolved. Currently, there is no evidence of a primary trigger similar to that seen

in other neurodegenerative conditions. Evidence exists for a role of inflammatory processes, both innate in the parenchyma and adaptive in the meninges and perivascular spaces, alongside a role for degenerative processes involving mitochondrial dysfunction and energy deficits. However, these different processes have yet to be conclusively linked. Studies of the very early stages of the disease are required to clarify whether inflammation and/or immune mechanisms or neurodegeneration is the primary cause of the disease. Nevertheless, it is apparent that at some point both mechanisms are simultaneously active in grey matter and white matter in multiple sclerosis.

To complicate matters further, the relationship between white and grey matter damage may differ across brain regions, disease phases and multiple sclerosis phe-notypes. Additional longitudinal studies that focus on the very early stages of the disease, when either the white matter or the grey matter are almost intact, are required to clarify the relationship between white and grey mat-ter damage. As such studies are unlikely to be possible using post-mortem or biopsy tissue samples, the most promising approaches will be combinations of ever more sophisticated imaging methods such as MRI, spectros-copy and positron-emission tomography imaging with immunological, proteomics and metabolomics methods usingCSF.

In light of accumulating evidence indicating that patients with multiple sclerosis with similar white mat-ter lesion volumes undergo a more severe clinical course in the presence of higher and more diffuse grey matter

Box 1 | Possible mechanisms underlying grey matter damage

Adaptive immunity and/or autoimmune mechanismsAutoantigens. The following processes are proposed to activate T cells (preferentially CD8+ cytotoxic Tcells rather than CD4+ T helper cells): Overexpression of specific autoantigens

Epitope spreading to antigens that are more abundantly expressed in the grey matter

Post-translational modifications of molecules in the grey matter

Molecular mimicry between myelin and neuronal antigens

Infectious agents and compartmentalized immune responses. It is proposed that an infectious organism located in the adjacent meninges or an infectious agent with primary tropism for cortical neurons and/or oligodendrocytes may result in the formation of aberrant tertiary lymphoid-like structures within the meninges or the cortical layer. These tertiary lymphoid-like structures may induce a chronic compartmentalized inflammatory response against neurons or oligodendrocytes that is characterized by the presence of CD4+ and CD8+ T cells and CD20+ B cells and plasmablasts.

CNS innate inflammatory mechanismsChronic microglial activation. The presence of a chronic inflammation in the subarachnoid and perivascular spaces, environmental factors such as toxins, or neuronal damage or death, might cause an abnormal and chronic microglial over-activation that can be directly harmful to neurons. This aberrant microglial activation may also lead to a dysregulation of astrocyte functions and glutamate metabolism, resulting in further neuronal and synaptic damage.

Non-inflammatory neurodegenerative mechanismsPrimary neurodegeneration. A primary neurodegenerative event involving neurons, astrocytes or oligodendrocytes that may be caused or worsened by mitochondrial injury. Metabolic compromise may explain the substantial neuronal injury accompanied by relatively minor parenchymal inflammation, microglial activation and associated meningeal inflammation, which are usually observed in multiple sclerosis.

Retrograde degeneration due to white matter damage. Reduction of cortical grey matter volume and thinning may be the consequence of white matter pathology via retrograde degeneration, especially in the more advanced disease phase, thus explaining the significant neuronal loss observed even in the non-demyelinated cortex.

R E V I E W S



Box 2 | The role of mitochondrial injury in grey matter damage

Mitochondrial injury is found in lesions characterized by all of the typical pathological features of multiple sclerosis, including inflammation, demyelination, oligodendrocyte apoptosis and axon degeneration119. A role for inflammation in mitochondrial dysfunction is suggested by post-mortem studies showing that reactive oxygen species produced by activated microglia and macrophages can induce mitochondrial dysfunction in both white matter120 and grey matter lesions104.

Several lines of evidence have led to the hypothesis that mitochondrial injury is a primary phenomenon in multiple sclerosis121. Neurons deficient in components of the respiratory chain have been identified in normal-appearing brain tissue and are distributed diffusely in the cortex of patients with multiple sclerosis122. Focal intra-axonal mitochondrial pathology is also one of the earliest signs of damage in a mouse model of multiple sclerosis123. Abnormal mitochondrial gene expression and impaired activity of the mitochondrial respiratory chain complexes I and III, which can be found in demyelinated cortex samples of patients with progressive multiple sclerosis124, may result in mitochondrial dysfunction and a state of impaired energy production. Finally, a characteristic hypoxia-like injury caused by mitochondrial impairment might also be involved in neurodegeneration in multiple sclerosis lesions125,126.

Irrespective of the initial event, the presence of a local energy deficit in grey matter may induce a vicious cycle that leads to an increased mitochondrial production of reactive oxygen species and further deletions of mitochondrial DNA127. Energy-deficient neurons may be more prone to inflammatory insult, which induces increased energy demand in the presence of a reduced energy supply128. Although the precise causes of neuronal respiratory deficiency in multiple sclerosis are still unclear, the neuronal energy deficit is crucial for inducing axonal swelling and subsequent neuronal death, especially when it occurs as a consequence of inflammation, as is the case in grey matter regions in patients with multiple sclerosis.

1. Noseworthy,J.H., Lucchinetti,C., Rodriguez,M. & Weinshenker,B.G. Multiple sclerosis. N.Engl. J.Med. 343, 938952 (2000).

2. Sospedra,M. & Martin,R. Immunology of multiple sclerosis. Annu. Rev. Immunol 23, 683747 (2005).

3. Lill,C.M. etal. CXCR5, SOX8, RPS6KB1 and ZBTB46 are genetic risk loci for multiple sclerosis. Brain 136, 17781782 (2013).

4. Ascherio,A. Environmental factors in multiple sclerosis. Expert Rev. Neurother. 13 (Suppl. 12), 39 (2013).

5. Dutta,R. & Trapp,B.D. Pathology and definition of multiple sclerosis. Rev. Prat. 56, 12931298 (2006).

6. Lassmann,H., Brck,W. & Lucchinetti,C.F. The immunopathology of multiple sclerosis: an overview. Brain Pathol. 17, 210218 (2007).

7. Dawson,J.D. The histology of disseminated sclerosis.Trans. R.Soc. Edin. 50, 517740 (1916).

8. Brownell,B. & Hughes,J.T. The distribution of plaques in the cerebrum in multiple sclerosis. J.Neurol. Neurosurg. Psychiatry 25, 315320 (1962).

9. Chard,D. & Miller,D. Grey matter pathology in clinically early multiple sclerosis: evidence from magnetic resonance imaging. J. Neurol. Sci. 282, 5 11 (2009).

10. Calabrese,M., Filippi,M. & Gallo,P. Cortical lesions in multiple sclerosis. Nature Rev. Neurol. 6, 438444 (2010).

11. Kidd,D. etal. Cortical lesions in multiple sclerosis. Brain 122, 1726 (1999).

12. Peterson,J.W., B,L., Mrk,S., Chang,A. & Trapp,B.D. Transected neurites, apoptotic neurons, and reduced inflammation in cortical multiple sclerosis lesions. Ann. Neurol. 50, 389400 (2001).The seminal paper on cortical lesions in multiple sclerosis.

13. B,L., Vedeler,C.A., Nyland,H.I., Trapp,B.D. & Mrk,S.J. Subpial demyelination in the cerebral cortex of multiple sclerosis patients. J.Neuropathol. Exp. Neurol. 62, 723732 (2003).

14. Brink,B.P. etal. The pathology of multiple sclerosis is location-dependent: no significant complement activation is detected in purely cortical lesions. J.Neuropathol. Exp. Neurol. 64, 147155 (2005).

15. van Horssen,J., Brink,B.P., de Vries,H.E., van der Valk,P. & B,L. The bloodbrain barrier in cortical multiple sclerosis lesions. J.Neuropathol. Exp. Neurol. 66, 321328 (2007).

16. Wegner,C., Esiri,M.M., Chance,S.A., Palace,J. & Matthews,P.M. Neocortical neuronal, synaptic, and glial loss in multiple sclerosis. Neurology 67, 960967 (2006).

17. Magliozzi,R. etal. A gradient of neuronal loss and meningeal inflammation in multiple sclerosis. Ann. Neurol. 68, 477 (2010).A very interesting study demonstrating a direct relationship between meningeal inflammation, subpial demyelination and neuronal loss in multiple sclerosis.

18. Freund,P. etal. Disability, atrophy and cortical reorganization following spinal cord injury. Brain 134, 16101622 (2011).

19. Sailer,M. etal. Focal thinning of the cerebral cortex in multiple sclerosis. Brain 126, 17341744 (2003).

20. Narayanan,S. etal. Imaging of axonal damage in multiple sclerosis: spatial distribution of magnetic resonance imaging lesions. Ann. Neurol. Mar. 41, 385391 (1997).

21. Varga,A.W. etal. White matter hemodynamic abnormalities precede sub-cortical gray matter changes in multiple sclerosis. J.Neurol. Sci. 282, 2833 (2009).

22. Dziedzic,T. etal. Wallerian degeneration: a major component of early axonal pathology in multiple sclerosis. Brain Pathol. 20, 976985 (2010).

23. Henry,R.G. etal. Connecting white matter injury and thalamic atrophy in clinically isolated syndromes. J.Neurol. Sci. 282, 6166 (2009).

24. Audoin,B. etal. Localization of grey matter atrophy in early RRMS: A longitudinal study. J.Neurol. 253, 14951501 (2006).

25. De Stefano, N. et al. Evidence of early cortical atrophy in MS: relevance to white matter changes and disability. Neurology 60, 11571162 (2003).One of the seminal MRI studies on cortical atrophy in multiple sclerosis.

26. Furby,J. etal. Different white matter lesion characteristics correlate with distinct grey matter abnormalities on magnetic resonance imaging in secondary progressive multiple sclerosis. Mult. Scler. 15, 687694 (2009).

27. Sanfilipo,M.P., Benedict,R.H., Sharma,J., Weinstock-Guttman,B. & Bakshi,R. The relationship between whole brain volume and disability in multiple sclerosis: a comparison of normalized gray versus

white matter with misclassification correction. Neuroimage 26, 10681077 (2005).

28. Roosendaal,S.D. etal. Grey matter volume in a large cohort of MS patients: relation to MRI parameters and disability. Mult. Scler. 17, 10981106 (2011).

29. Ceccarelli,A. etal. A voxel-based morphometry study of grey matter loss in MS patients with different clinical phenotypes. Neuroimage 42, 315322 (2008).

30. Battaglini,M. etal. Voxel-wise assessment of progression of regional brain atrophy in relapsing-remitting multiple sclerosis. J.Neurol. Sci. 282, 5560 (2009).

31. Bendfeldt,K. etal. Association of regional gray matter volume loss and progression of white matter lesions in multiple sclerosis a longitudinal voxel-based morphometry study. Neuroimage 45, 6067 (2009).

32. Pagani,E. etal. Regional brain atrophy evolves differently in patients with multiple sclerosis according to clinical phenotype. Am. J.Neuroradiol. 26, 341346 (2005).

33. Sepulcre,J. etal. Contribution of white matter lesions to gray matter atrophy in multiple sclerosis: evidence from voxel-based analysis of T1 lesions in the visual pathway. Arch. Neurol. 66, 173179 (2009).

34. Gilmore,C.P. etal. Regional variations in the extent and pattern of grey matter demyelination in multiple sclerosis: a comparison between the cerebral cortex, cerebellar cortex, deep grey matter nuclei and the spinal cord. J.Neurol. Neurosurg. Psychiatry 80, 182187 (2009).

35. Kutzelnigg,A. etal. Widespread demyelination in the cerebellar cortex in multiple sclerosis. Brain Pathol. 17, 3844 (2007).

36. Gilmore,C.P. etal. Spinal cord neuronal pathology in multiple sclerosis. Brain Pathol. 19, 642649 (2009).

37. Howell,O. etal. Meningeal inflammation is widespread and linked to cortical pathology in multiple sclerosis. Brain 134, 27552771 (2011).

38. Lucchinetti,C.F. etal. Inflammatory cortical demyelination in early multiple sclerosis. N.Engl. J.Med. 365, 21882197 (2011).The first neuropathological characterization of inflammatory cortical lesions in early multiple sclerosis.

39. Calabrese,M. etal. Detection of cortical inflammatory lesions by double inversion recovery magnetic resonance imaging in patients with multiple sclerosis. Arch. Neurol. 64, 14161422 (2007).

lesion volumes and atrophy already from the early disease phases, it may be suggested that cortical pathology has a key role in determining the rate of disability progression in multiple sclerosis. As with other progressive neurode-generative conditions, it is likely that the accumulation of disability is directly related to increasing loss of vital neuronal populations, and the rate at which they are lost and their locations determine the relative rate of disease progression. However, in contrast to other neurodegen-erative conditions, the loss of neurons and axons is much more diffuse, which is reflected in the heterogeneous decline in diverse motor, sensory and cognitive functions. Knowledge of the mechanisms underlying grey matter damage and identification of measures of disease progres-sion associated with cortical pathology could help to iden-tify prognostic biomarkers and individualized therapies with increased efficacy for patients with a predisposition to develop more severe cortical pathology and associ-ated clinical deterioration. Whether this predisposition is linked to genetic, environmental or other factors and is common to white matter immunopathogenesis remains unknown and highlights the need for more detailed stud-ies. Despite the unknowns, from a therapeutic point of view, a combination of immunomodulatory and neuro-protective strategies appears to be the most promising to limit both grey matter degeneration and development of long-term disability.

R E V I E W S



40. Giorgio,A. etal. Cortical lesions in radiologically isolated syndrome. Neurology 77, 18961899 (2011).

41. Calabrese,M. & Gallo,P. Magnetic resonance evidence of cortical onset of multiple sclerosis. Mult. Scler. 15, 933941 (2009).

42. Seewann,A. etal. Imaging the tip of the iceberg: visualization of cortical lesions in multiple sclerosis. Mult. Scler. 17, 12021210 (2011).

43. Chard,D.T. etal. Brain atrophy in clinically early relapsing-remitting multiple sclerosis. Brain 125, 327337 (2002).

44. Tiberio,M. etal. Gray and white matter volume changes in early RRMS: a 2-year longitudinal study. Neurology 64, 10011007 (2005).

45. Sbardella,E. etal. Assessing the correlation between grey and white matter damage with motor and cognitive impairment in multiple sclerosis patients. PLoS ONE 8, e63250 (2013).

46. Steenwijk,M.D. etal. What explains gray matter atrophy in long-standing multiple sclerosis? Radiology 272, 832842 (2014).

47. Calabrese,M. etal. Imaging distribution and frequency of CLs in patients with multiple sclerosis. Neurology 75, 12341240 (2010).

48. Vercellino,M. etal. Demyelination, inflammation, and neurodegeneration in multiple sclerosis deep gray matter. J.Neuropathol. Exp. Neurol. 68, 489502 (2009).

49. Geurts,J.J. etal. Extensive hippocampal demyelination in multiple sclerosis. J.Neuropathol. Exp. Neurol. 66, 819827 (2007).

50. Cohen-Adad,J. etal. Invivo evidence of disseminated subpial T2* signal changes in multiple sclerosis at 7 T: a surface-based analysis. Neuroimage 57, 5562 (2011).

51. Audoin,B. etal. Atrophy mainly affects the limbic system and the deep grey matter at the first stage of multiple sclerosis. J.Neurol. Neurosurg. Psychiatry 81, 690695 (2010).

52. Bendfeldt,K. etal. Spatiotemporal distribution pattern of white matter lesion volumes and their association with regional grey matter volume reductions in relapsing-remitting multiple sclerosis. Hum. Brain Mapp. 31, 15421555 (2010).

53. Bettelli,E. etal. Myelin oligodendrocyte glycoprotein-specific Tcell receptor transgenic mice develop spontaneous autoimmune optic neuritis. J.Exp. Med. 197, 10731081 (2003).

54. Krishnamoorthy,G. etal. Myelin-specific T cells also recognize neuronal autoantigen in a transgenic mouse model of multiple sclerosis. Nature Med. 15, 626632 (2009).

55. Serafini,B., Rosicarelli,B., Magliozzi,R., Stigliano,E. & Aloisi,F. Detection of ectopic B-cell follicles with germinal centers in the meninges of patients with secondary progressive multiple sclerosis. Brain Pathol. 14, 164174 (2004).

56. Ascherio,A. & Munger,K.L. EpsteinBarr virus infection and multiple sclerosis: a review. J.Neuroimmune Pharmacol. 5, 271277 (2010).

57. Ascherio,A. etal. EpsteinBarr virus antibodies and risk of multiple sclerosis: a prospective study. JAMA 286, 30833088(2001).

58. Levin,L.I. etal. Multiple sclerosis and EpsteinBarr virus. JAMA 289, 15331536(2003).

59. Serafini,B. etal. Dysregulated EpsteinBarr virus infection in the multiple sclerosis brain. J.Exp. Med. 204, 28992912 (2007).

60. Serafini,B., Muzio,L., Rosicarelli,B. & Aloisi,F. Radioactive insitu hybridization for EpsteinBarr virus-encoded small RNA supports presence of EpsteinBarr virus in the multiple sclerosis brain. Brain 136, e233 (2013).

61. Magliozzi,R. etal. B-cell enrichment and EpsteinBarr virus infection in inflammatory cortical lesions in secondary progressive multiple sclerosis. J.Neuropathol. Exp. Neurol. 72, 2941 (2013).

62. Angelini,D.F. etal. Increased CD8+ Tcell response to EpsteinBarr virus lytic antigens in the active phase of multiple sclerosis. PLoS Pathog. 9, e1003220 (2013).

63. Lossius,A. etal. High-throughput sequencing of TCR repertoires in multiple sclerosis reveals intrathecal enrichment of EBV-reactive CD8+ Tcells. Eur. J.Immunol. 44, 34393452 (2014).

64. Sargsyan,S.A. etal. Absence of EpsteinBarr virus in the brain and CSF of patients with multiple sclerosis. Neurology 74, 11271135 (2010).

65. Willis,S.N. etal. EpsteinBarr virus infection is not a characteristic feature of multiple sclerosis brain. Brain 132, 33183328 (2009).

66. Lassmann,H., Niedobitek,G., Aloisi,F., Middeldorp,J.M. & NeuroproMiSeEBV Working Group. EpsteinBarr virus in the multiple sclerosis brain: a controversial issue report on a focused workshop held in the Centre for Brain Research of the Medical University of Vienna, Austria. Brain 134, 27722786 (2011).

67. Aloisi,F., Serafini,B., Magliozzi,R., Howell,O.W. & Reynolds,R. Detection of EpsteinBarr virus and B-cell follicles in the multiple sclerosis brain: what you find depends on how and where you look. Brain 133, e157 (2010).

68. Maggi,F. etal. Low prevalence of TT virus in the cerebrospinal fluid of viremic patients with central nervous system disorders. J.Med. Virol. 65, 418422(2001).

69. Sospedra,M. etal. Recognition of conserved amino acid motifs of common viruses and its role in autoimmunity. PLoS Pathog. 1, e41 (2005).

70. Lamberto,I., Gunst,K., Mller,H., Zur Hausen,H. & de Villiers,E.M. Mycovirus-like DNA virus sequences from cattle serum and human brain and serum samples from multiple sclerosis patients. Genome Announc. 2, e0084814 (2014).

71. Borkosky,S.S., Whitley,C., Kopp-Schneider,A., zur Hausen,H. & de Villiers,E.M. EpsteinBarr virus stimulates torque teno virus replication: a possible relationship to multiple sclerosis. PLoS ONE 7, e32160 (2012).

72. Zeis,T., Graumann,U., Reynolds,R. & Schaeren-Wiemers,N. Normal-appearing white matter in multiple sclerosis is in a subtle balance between inflammation and neuroprotection. Brain 131, 288303 (2008).

73. Baranzini,S.E. etal. Genetic variation influences glutamate concentrations in brains of patients with multiple sclerosis. Brain 133, 26032611 (2010).

74. Mastronardi,F.G. & Moscarello,M.A. Molecules affecting myelin stability: a novel hypothesis regarding the pathogenesis of multiple sclerosis. J.Neurosci. Res. 80, 301308 (2005).

75. Neumann,H., Cavali,A., Jenne,D.E. & Wekerle,H. Induction of MHC classI genes in neurons. Science 269, 549552 (1995).

76. Jersild,C. etal. Histocompatibility determinants in multiple sclerosis, with special reference to clinical course. Lancet 2, 12211225 (1973).

77. The International Multiple Sclerosis Genetics Consortium (IMSGC) etal. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 476, 214219 (2011).

78. Martin,R. etal. Fine specificity and HLA restriction of myelin basic protein-specific cytotoxic Tcell lines from multiple sclerosis patients and healthy individuals. J.Immunol. 145, 540548 (1990).

79. Ota,K. etal. T-cell recognition of an immunodominant myelin basic protein epitope in multiple sclerosis. Nature 346, 183187 (1990).

80. Bielekova,B. etal. Expansion and functional relevance of high-avidity myelin-specific CD4+ Tcells in multiple sclerosis. J.Immunol. 172, 38933904 (2004).

81. Yates,R.L., Esiri,M.M., Palace,J., Mittal,A. & DeLuca,G.C. The influence of HLA-DRB1*15 on motor cortical pathology in multiple sclerosis. Neuropathol. Appl. Neurobiol. http://dx.doi.org/10.1111/nan.12165 (2014).

82. Hftberger,R. etal. Expression of major histocompatibility complex classI molecules on the different cell types in multiple sclerosis lesions. Brain Pathol. 14, 4350 (2004).

83. Babbe,H. etal. Clonal expansions of CD8+ Tcells dominate the Tcell infiltrate in active multiple sclerosis lesions as shown by micromanipulation and single cell polymerase chain reaction. J.Exp. Med. 192, 393404 (2000).

84. Ulvestad,E. etal. HLA classII molecules (HLA-DR, -DP, -DQ) on cells in the human CNS studied insitu and invitro. Immunology 82, 535541 (1994).

85. Liblau,R.S., Gonzalez-Dunia,D., Wiendl,H. & Zipp,F. Neurons as targets for Tcells in the nervous system. Trends Neurosci. 36, 315324 (2013).

86. Meuth,S.G. etal. Cytotoxic CD8+ Tcell-neuron interactions: perforin-dependent electrical silencing precedes but is not causally linked to neuronal cell death. J.Neurosci. 29, 1539715409 (2009).

87. Suidan,H.S. etal. Granzyme A released upon stimulation of cytotoxic T lymphocytes activates the thrombin receptor on neuronal cells and astrocytes. Proc. Natl Acad. Sci. USA 91, 81128116 (1994).

88. Medana,I.M. etal. MHC class I-restricted killing of neurons by virus-specific CD8+ T lymphocytes

is effected through the Fas/FasL, but not the perforin pathway. Eur. J.Immunol. 30, 36233633 (2000).

89. Aktas,O. etal. Neuronal damage in autoimmune neuroinflammation mediated by the death ligand TRAIL. Neuron 46, 421432 (2005).

90. Mizuno,T. etal. Interferon-gamma directly induces neurotoxicity through a neuron specific, calcium-permeable complex of IFN- receptor and AMPA GluR1 receptor. FASEB J. 22, 17971806 (2008).

91. Vergelli,M. etal. Human autoreactive CD4+ Tcell clones use perforin- or Fas/Fas ligand-mediated pathways for target cell lysis. J.Immunol. 158, 27562761 (1997).

92. Vergelli,M. etal. A novel population of CD4+CD56+ myelin-reactive Tcells lyses target cells expressing CD56/neural cell adhesion molecule. J.Immunol. 157, 679688 (1996).

93. Zaguia,F. etal. Cytotoxic NKG2C+ CD4 Tcells target oligodendrocytes in multiple sclerosis. J.Immunol. 190, 25102518 (2013).

94. Magliozzi,R. etal. Meningeal B-cell follicles in secondary progressive multiple sclerosis associate with early onset of disease and severe cortical pathology. Brain 130, 10891104 (2007).

95. Magliozzi,R., Columba-Cabezas,S., Serafini,B. & Aloisi,F. Intracerebral expression of CXCL13 and BAFF is accompanied by formation of lymphoid follicle-like structures in the meninges of mice with relapsing experimental autoimmune encephalomyelitis. J.Neuroimmunol. 148, 1123 (2004).

96. Peters,A. etal. Th17 cells induce ectopic lymphoid follicles in central nervous system tissue inflammation. Immunity35, 986996(2011).

97. Choi,S.R. etal. Meningeal inflammation plays a role in the pathology of primary progressive multiple sclerosis. Brain 135, 29252937 (2012).

98. Kramann,N. etal. Increased meningeal T and plasma cell infiltration is associated with early subpial cortical demyelination in common marmosets with experimental autoimmune encephalomyelitis. Brain Pathol. http://dx.doi.org/10.1111/bpa.12180 (2014).

99. Gardner,C. etal. Cortical grey matter demyelination can be induced by elevated pro-inflammatory cytokines in the subarachnoid space of MOG-immunized rats. Brain 136, 35963608 (2013).

100. Kooi,E.J., Geurts,J.J., van Horssen,J., B,L. & van der Valk,P. Meningeal inflammation is not associated with cortical demyelination in chronic multiple sclerosis. J.Neuropathol. Exp. Neurol. 68, 10211028 (2009).

101. Guseo,A. & Jellinger,K. The significance of perivascular infiltrations in multiple sclerosis. J.Neurol. 211, 5160 (1975).

102. Kutzelnigg,A. etal. Cortical demyelination and diffuse white matter injury in multiple sclerosis. Brain 128, 27052712 (2005).This study provides elegant autoptic evidence of white matter and grey matter damage in multiple sclerosis.

103. Reynolds,R. etal. The neuropathological basis of clinical progression in multiple sclerosis. Acta Neuropathol. 122, 155170 (2011).

104. Fischer,M.T. etal. Disease-specific molecular events in cortical multiple sclerosis lesions. Brain 136, 17991815 (2013).

105. Calabrese,M. etal. The changing clinical course of multiple sclerosis: a matter of gray matter. Ann. Neurol. 74, 7683 (2013).An interesting paper implicating grey matter damage in the progression of clinical disability during the course of multiple sclerosis.

106. Androdias,G. etal. Meningeal Tcells associate with diffuse axonal loss in multiple sclerosis spinal cords. Ann. Neurol. 68, 465476 (2010).

107. Dutta,R. & Trapp,B.D. Pathogenesis of axonal and neuronal damage in multiple sclerosis. Neurology 68 (Suppl. 3), 2231 (2007).

108. Kreutzberg,G.W. Microglia: a sensor for pathological events in the CNS. Trends Neurosci. 19, 312318 (1996).

109. Nimmerjahn,A., Kirchhoff,F. & Helmchen,F. Resting microglial cells are highly dynamic surveillants of brain parenchyma invivo. Science 308, 13141318 (2005).

110. Aloisi,F. Immune function of microglia. Glia 36, 1 65179 (2001).

R E V I E W S



http://dx.doi.org/10.1111/nan.12165http://dx.doi.org/10.1111/nan.12165http://dx.doi.org/10.1111/bpa.12180

111. Block,M.L. & Hong,J.S. Chronic microglial activation and progressive dopaminergic neurotoxicity. Biochem. Soc. Trans. 35, 11271132 (2007).

112. Polazzi,E. & Contestabile,A. Reciprocal interactions between microglia and neurons: from survival to neuropathology. Rev. Neurosci. 13, 221242 (2002).

113. Gray,E., Thomas,T.L., Betmouni,S., Scolding,N. & Love,S. Elevated matrix metalloproteinase-9 and degradation of perineuronal nets in cerebrocortical multiple sclerosis plaques. J.Neuropathol. Exp. Neurol. 67, 888899 (2008).

114. Vercellino,M. etal. Altered glutamate reuptake in relapsing-remitting and secondary progressive multiple sclerosis cortex: correlation with microglia infiltration, demyelination, and neuronal and synaptic damage. J.Neuropathol. Exp. Neurol. 66, 732739 (2007).

115. Kooi,E.J., Strijbis,E.M., van der Valk,P. & Geurts,J.J. Heterogeneity of cortical lesions in multiple sclerosis: clinical and pathologic implications. Neurology 79, 13691376 (2012).

116. Stys,P.K., Zamponi,G.W., van Minnen,J. & Geurts,J.J. Will the real multiple sclerosis please stand up? Nature Rev. Neurosci. 13, 507514 (2012).

117. Barnett,M.H. & Prineas,J.W. Relapsing and remitting multiple sclerosis: pathology of the newly forming lesion. Ann. Neurol. 55, 458468 (2004).

118. Henderson,A.P., Barnett,M.H., Parratt,J.D. & Prineas,J.W. Multiple sclerosis: distribution of

inflammatory cells in newly forming lesions. Ann. Neurol. 66, 739753 (2009).

119. Lassmann,H., van Horssen,J. & Mahad,D. Progressive multiple sclerosis: pathology and pathogenesis. Nature Rev. Neurol. 8, 647656 (2012).A comprehensive review on the pathology and pathogenesis of progressive multiple sclerosis.

120. Fischer,M.T. etal. NADPH oxidase expression in active multiple sclerosis lesions in relation to oxidative tissue damage and mitochondrial injury. Brain 135, 886899 (2012).

121. Witte,M.E., Geurts,J.J., de Vries,H.E., van der Valk,P. & van Horssen,J. Mitochondrial dysfunction: a potential link between neuroinflammation and neurodegeneration? Mitochondrion 10, 411418 (2010).

122. Campbell,G.R. etal. Mitochondrial DNA deletions and neurodegeneration in multiple sclerosis. Ann. Neurol. 69, 481492 (2011).

123. Nikic,I. etal. A reversible form of axon damage in experimental autoimmune encephalomyelitis and multiple sclerosis. Nature Med. 17, 495499 (2011).

124. van Horssen,J., Witte,M.E. & Ciccarelli,O. The role of mitochondria in axonal degeneration and tissue repair in MS. Mult. Scler. 18, 10581067 (2012).

125. Mahad,D., Ziabreva,I., Lassmann,H. & Turnbull,D. Mitochondrial defects in acute multiple sclerosis lesions. Brain 131, 17221735 (2008).

126. Marik,C., Felts,P.A., Bauer,J., Lassmann,H. & Smith,K.J. Lesion genesis in a subset of patients with multiple sclerosis: a role for innate immunity? Brain 130, 28002815 (2007).

127. Druzhyna,N.M., Wilson,G.L. & LeDoux,S.P. Mitochondrial DNA repair in aging and disease. Mech. Ageing Dev. 129, 383390 (2008).

128. Trapp,B.D. Virtual hypoxia and chronic necrosis of demyelinated axons in multiple sclerosis. Lancet Neurol. 8, 280291 (2009).

AcknowledgementsM.C. is supported by the Progressive MS Alliance (PA-0124). R.Magliozzi is supported by an Italian Multiple Sclerosis Foundation grant (FISM 2011/R/23) and by an Italian Ministry of Health grant (GR-2010-2313255). R.R. is supported by the UK Multiple Sclerosis Society and the UK Medical Research Council. R. Martin and the Neuroimmunology and Multiple Sclerosis Research Section are supported by the Clinical Research Priority Program MS (CRPPMS) of the University of Zurich, the Swiss National Science Foundation (SNF), a European Research Council (ERC) Advanced Grant, the EU-FP7 framework programme and the Swiss Multiple Sclerosis Society.

Competing interests statementThe authors declare competing interests: see Web version for details.

R E V I E W S



http://www.nature.com/nrn/journal/v16/n3/full/nrn3900.html#affil-auth

Abstract | Multiple sclerosis is characterized at the gross pathological level by the presence of widespread focal demyelinating lesions of the myelin-rich white matter. However, it is becoming clear that grey matter is not spared, even during the earliesAre white and grey matter damage linked?Figure 1 | Cortical lesion subtypes in multiple sclerosis.From a neuropathological point of view, grey matter lesions in multiple sclerosis tissue are grouped into several types. a | Type1 (leukocortical) lesions extend through grey matter into the whitInflammatory grey matter damageFigure 2 | Inflammatory and non-inflammatory grey matter neurodegeneration mechanisms.a | In the undisturbed grey matter, microglia typically exist in a resting state and continually survey the microenvironment with their motile processes and protrusionsFigure 3 | Is grey matter atrophy a primary or secondary pathological process?An important issue in multiple sclerosis concerns the relationship between grey and white matter damage. In particular, it is unclear whether grey matter damage is secondary toFigure 4 | Immune-mediated mechanisms of subpial cortical demyelination in progressive multiple sclerosis.Schematic depiction of the immunological and pathological events associated with neurodegeneration, oligodendrocyte and astrocyte damage, and increaNon-inflammatory neurodegeneration?Conclusions and perspectivesBox 1 | Possible mechanisms underlying grey matter damageBox 2 | The role of mitochondrial injury in grey matter damage