Secondary-progressive multiple sclerosis (SPMS) is the onset of gradual worsening of disability following a relapsing-remitting course that may or may not be accompanied by relapses. There are no clear-cut demarcations (imaging, pathologic, immunologic) between the RRMS and SPMS phases, so diagnosis is generally made retrospectively.SPMS poses a number of conceptual challenges, as shown by the varying definitions that have been used. The MS Phenotype group suggested a catch-all of progressive disease (SPMS and PPMS) with four modifiers: disease activity with/without progression or progression with/without disease activity (Lublin et al. Neurology 2014;83:278-286). The MSBase group proposed SPMS criteria that required disability worsening in the absence of relapses (Lorscheider et al. Brain 2016; 139(Pt 9):2395-2405). The McDonald committee split the difference, defining progressive disease as steadily worsening disability independent of relapses, but allowing that superimposed relapses might occur (Thompson et al. Lancet Neurol 2018;17:162-173).
These definitions imply a sequence of events in which an inflammatory phase is followed by a neurodegenerative phase, in what has been described as a two-stage disease course (Leray et al. Brain 2010: 133; 1900-1913). More recent data suggest that relapses are common in SPMS and PPMS, and both are characterized by extensive inflammation that may manifest as focal lesions or more diffuse immunohistological changes. Indeed, a recent autopsy study found that among patients with progressive MS (mean disease duration 28.6 years), there was substantial inflammatory lesion activity at the time of death – lesions were present in 78% of progressive patients – and lesion load was higher than what is seen in RRMS patients (Luchetti et al. Acta Neuropathol 2018;135:511-528).
Thus, the requirement that relapses must be absent before an SPMS diagnosis, or that the presence of focal inflammation indicates RRMS rather than progressive disease, may not adequately reflect the disease process. An absence of relapses may be of value as a clinically observable phenomenon, just as relapses have been used – perhaps misleadingly – as an easy measure of disease activity and treatment response in drug trials. But such a requirement may delay recognition of SPMS onset. For example, MSBase found that with its definition of worsening disability in the absence of relapses, the median time to SPMS diagnosis was 32.6 years (Lorscheider 2016). This is considerably longer than the period of 16-20 years that is typically reported (Tremlett et al. Mult Scler 2008;14:314-324).
Pathogenic mechanisms in progressive MS
The description of progressive MS proposed by the Phenotype Group (Lublin 2016) blurs the distinction between SPMS and PPMS, which may be appropriate. As noted in a recent review (Lassmann H. Front Immunol 2019;9:3116), no qualitative differences in disease activity, lesion morphology or immunopathology have been found between SPMS and PPMS (Kuchling et al. Mult Scler 2014;20:1866-1871). There are no biomarkers that differentiate SPMS from PPMS (Lassmann 2019), and genome-wide association studies have failed to find genetic variants that differentiate the two. The slope of progression and the age at which disability milestones are reached are similar in SPMS and PPMS (Confavreux et al. N Engl J Med 2000;343:1430-1438). Once a tipping point of disability worsening has been reached, the disease process will be ‘amnesic’, to use Confavreux’s term, as to whether an RRMS phase preceded it or not (Confavreux et al. Brain 2003;126(Pt 4):770-782).
The successive stages of inflammation and neurodegeneration describe a process and an end result that may be better conceptualized as two different types of inflammation, both of which result in neurodegeneration (Lassmann 2019). In the first, more florid inflammatory phase, disruption of the blood-brain barrier permits the passage of activated lymphocytes into the CNS, where they become reactivated and promote inflammation and focal lesion formation. The predominant lymphocytes involved in early lesion formation are CD8+ T cells, found in all four lesion patterns described by Lucchinetti and colleagues (Ann Neurol 2000;47:707-717). Less numerous are CD4+ T cells, which appear to be primarily involved in the initiation of inflammation and less involved in demyelination and neurodegeneration (Lassmann 2019).
B cells promote CNS inflammation by presenting antigen to CD4+ T cells (Harp et al. Clin Immunol 2008;128:382-391) and by releasing proinflammatory cytokines (e.g. TNF-alpha, IL-6, GM-CSF) that amplify the immune response (Duddy et al. J Immunol 2004;172:3422-3427. Li et al. Sci Transl Med 2015;7:310ra166). Antigen presentation to T cells also induces B cells to differentiate to antibody-producing plasma cells (Hausser-Kinzel & Weber. Front Immunol 2019;10:201). The presence of oligoclonal bands in CSF may indicate that these plasma cells accumulate in the CNS and play a pathogenic role; immunoglobulin deposition in pattern II demyelinating lesions further suggests that they are involved in myelin and axonal damage (Lucchinetti 2000).
Infiltrating T and B lymphocytes appear to be most important in early, active lesion formation and acute axonal degeneration, and play a comparatively lesser role in progressive MS (Machado-Santos et al. Brain 2018;141:2066-2082. Frischer et al. Brain 2009;132:1175-1189). This helps to explain the finding that therapeutic agents in RRMS that reduce circulating T and B lymphocyte counts in the periphery are less effective in SPMS and PPMS. This does not mean that inflammation has no role in progressive MS; rather, a second type of inflammation appears to predominate that roughly corresponds to the shift from an adaptive to an innate immune response proposed a decade ago (Weiner HL. J Neurol 2008;255[suppl 1]:3-11). This compartmentalized inflammation is less dependent on barrier dysfunction (blood-brain, blood-meningeal and blood-CSF, reviewed in Michel et al. Front Immunol 2015;6:636) and invasion of peripheral lymphocytes into the CNS.
Of particular interest to the progressive phase is the formation of follicle-like structures in the meninges composed of cells of both the adaptive (T cells, B cells) and innate (e.g. macrophages/microglia, mast cells) immune systems. The degree of compartmentalized meningeal inflammation has been correlated with the amount of cortical grey-matter inflammation and demyelination. The extent of inflammation and tissue damage occurs along a gradient, with greater microglial activation and more extensive neuronal, astrocyte and oligodendrocyte damage in the superficial cortical layers (Magliozzi et al. Ann Neurol 2010;68:477-493. Magliozzi et al. Ann Neurol 2018;83:739-755). This gradient suggests that meningeal infiltrates release soluble factors that are directly cytotoxic (Bornstein & Appel. Ann NY Acad Sci 1965;122:280-286), or which induce pathogenic activation of microglia/macrophages. This is the topic of a poster at the upcoming AAN annual meeting (Touil et al. AAN 2019; abstract P2.2.083).
Meningeal follicles were first reported in SPMS patients and were associated with more extensive demyelination and more rapidly progressive disability (Magliozzi et al. Brain 2007;130[Pt 4]:1089-1104. Howell et al. Brain 2011;134[Pt 9]:2755-2771). However, subsequent studies have reported meningeal inflammation in PPMS and acute RRMS (Bevan et al. Ann Neurol 2018;84:829-842), suggesting that follicle formation may be an important indicator of rapidly worsening disability independent of the clinical phenotype.
Thus, it appears that B cells and T cells trapped within the CNS compartment are able to maintain a chronic inflammatory state that is associated with slowly expanding lesions and progressive tissue loss. This differs from the hypothesis that neurodegeneration becomes independent of inflammation during progressive MS (Trapp & Nave. Annu Rev Neurosci 2008;31:247-269). Rather, it appears that axonal damage is invariably associated with inflammation; this view is supported by the finding that in late-stage MS, when inflammation has resolved, the extent of ongoing axonal damage is similar to what is seen in non-MS controls (Frischer 2009).
Emerging treatment approaches
Compartmentalized inflammation poses a challenge to therapeutics. Cell-depleting therapies that target immune cells in the periphery have been highly effective in reducing relapses and early focal lesion formation, but monoclonal antibodies have limited access to the CNS compartment and would be expected to have less effect on resident B and T cells. Moreover, anti-CD20 therapies have little or no effect on antibody production since antibody-producing plasma cells do not express CD20+ (Michel 2015). A further concern is that some B cell subsets (antigen-naïve, plasmablasts, plasma cells) are a source of anti-inflammatory cytokines (e.g. IL-10, TGF-beta) required for immune regulation and tissue repair. More selective B cell targeting may be beneficial and agents that modulate B cell signalling (e.g. Bruton’s tyrosine kinase inhibitors) are now in development (Montalban et al. ECTRIMS 2018; abstract 322).
Another area of interest is modulation of sphingosine-1-phosphate (S1P) signalling in the CNS, which has been implicated in astrocyte activation and astrogliosis (Sorensen et al. Mol Pharmacol 2003;64:1199-1209). Functional antagonism of S1P1 blocks lymphocyte egress from lymphoid tissue and is presumed to be the principal mode of action of fingolimod in MS. However, preclinical studies have reported that the efficacy of fingolimod is largely dependent on loss of S1P1 signalling by astrocytes rather than on its effects on peripheral immune cells (Choi et al. PNAS USA 2011;108:751-756). Reactive astrocytes express S1P1 in chronic-inactive as well as active MS lesions (Van Doorn et al. Glia 2010;58:1465-1476), which may explain the early effects of fingolimod on diffuse inflammation and brain atrophy (De Stefano et al. Mult Scler Relat Disord 2016;7:98-101), as well as on axonal and myelin integrity in MS patients (Gurevich et al. CNS Neurosci Ther 2018;24:412-419). Experimental data suggest that S1P1 modulation also reduces microglial activation (Jackson et al. J Neuroinflammation 2011;8:76), reduces neuronal cell death (Cipriani et al. J Neuroinflammation 2015;12:86), and enhances neural cell differentiation to oligodendrocytes (Yazdi et al. Iran J Pharm Res 2018;17:1444-1457). In additional, preclinical studies of siponimod, a dual S1PR-1/5 modulator, have shown that it reduces astrocytosis and microgliosis and prevents synaptic neurodegeneration, effects that appear to occur independently of the drug’s action on T cells (Gentile A. J Neuroinflammation 2016;13:207). These findings suggest that in the CNS, S1P receptor modulation reduces activation of microglia and may promote remyelination through its effects on oligodendrocyte progenitor cells (Novgorodov AS. Faseb J 2007;21:1503-1514).
The conceptual and practical challenges of progressive MS have implications for the design and interpretation of clinical studies. Part 2 of this article will review recent trials in progressive MS and their implications for treatment.
Complete the survey on your experience with SPMS.
All responses will appear once you complete the survey and click Submit.
This poll is currently disabled.