Since first described by Charcot almost two centuries ago (1), the field of multiple sclerosis (MS) witnessed enormous advancements. Not only we know better about the pathophysiology of the disease but also the different risk factors, clinical subtypes, and at last but not at least how attenuate the pathophysiological processes leading to neuronal demise. The success cannot be better seen than in the field of relapsing remitting multiple sclerosis (RRMS) with more than six medications got approved by the FDA and EMA since 2010 (2). However, the primary progressive multiple sclerosis (PPMS) remains a considerable challenge. Despite having more than one study fulfilling its primary end point (3, 4) to the time of writing this article, we still do not have any approved medication.

Nevertheless, the recent focus on progressive MS forms [primary and secondary progressive multiple sclerosis (PP and SPMS)] leads to a deep insight in the different pathological aspects driving the disease. In this article, we summarize the recent developments and integrate the different aspects to provide a framework explaining the sequence of events leading to the unique clinical picture.

The predominant lesions in PPMS are slowly expanding lesions with T cells, microglial, and macrophage-associated demyelination in close similar to pattern 1 demyelination (30). While the involvement of different CD4+ subtypes (Th1, Th17, and Th9) is one of the very initial events in MS (31), the main lymphocytes to be found in the lesions are CD8+ cells and correlate with the degree of axonal damage (32). sCD27, a marker of intrathecal inflammation secreted mainly by T cells, is elevated in PPMS (33). Prominent T_FH and Th17 activation in serum of PPMS patients was reported and correlated with the progression rate (34).

Evidences for B cell involvement in PPMS are numerous: the intrathecal IgG production, the detection of B cells within MS lesions, meningeal infiltrate, perivascular space and MS parenchyma, the presence of autoreactive antibodies against myelin and its products (32), and finally the success of B cell-based therapies in PPMS (35). B cells are scattered in the meninges in a diffuse manner with tertiary lymphoid follicles formation only in aggressive disease with active progressive disease (36). B- and plasma cells in PPMS lesions correlate with the severity of axonal damage (26). B cells are pathogenic through multiple pathways including antigen presentation, cytokines release, and producing of the autoantibodies (37). Their role beyond the synthesis of autoantibodies is confirmed by the fact that highly effective monoclonal anti CD-20 antibodies do not eliminate the long-lasting antibodies producing plasma cells (37). One example for non-antigen-presenting B cells is the pro-inflammatory granulocyte macrophage colony-stimulating factor (GM-CSF) B cells; through their GM-CSF secretion they induce pro-inflammatory myeloid cell response promoting the release of Th1- and Th17-differentiating cytokines like IL-6 and -12 (38).

The discovery of the B_reg cells secreting IL-10, IL-35, and TGF-b indicates the complex role of B cells in MS. B_reg can restore Th1/Th2 balance, inhibit Th1 and Th17 cell differentiation, and inhibit macrophages (37). Moreover, the secreted antibodies may play a role in regulating the immune system and inducing remyelination (39).

Phagocytic cells like the macrophages are the most common cells found in the slowly expanding lesions in PPMS (30). They are derived from blood monocytes and migrate into CNS after stimulation in the blood (40). The pro-inflammatory M1 play central role both in the demyelination and axonal damage through reactive oxygen spices, nitric oxide, and glutamate (40). CNS-infiltrating macrophages were able to induce progressive EAE through sustained secretion of TNF (41). Levels of sCD14, a marker of macrophageal activity, were higher in patients with PPMS compared to healthy controls, but similar to RRMS patients in relapses but not in remission (42). Nevertheless, macrophages (anti-inflammatory M2) are essential for the remyelination by clearing the damaged tissues in the lesions (40).

Dendritic cells can also be found in MS lesions (32). Dendritic cells form SPMS patients secret much higher levels of IL-18 than those from RRMS patients and healthy controls (43) and induce—in vitro—solely a Th1 cell response not Th1 and Th2 like in RRMS suggesting a role of dendritic cells in the disease transition into the progressive phase (44). Based on the similarities between SP and PPMS, a role of dendritic cells in PPMS cannot be excluded.

Over the last years, microglial activation (MiA) gained more interest as one of the key mechanisms for neurodegeneration and axonal demise in MS (30). In PPMS, the microglia were diffusely active in the lesions and in normal-appearing white matter (NAWM) and normal-appearing gray matter (NAGM) (45). Activated microglia in NAWM forms microglial nodules in close proximity to stressed oligodendrocytes and degenerated axons with profound release of oxygen-free radicals (30). MiA in cortical gray matter of SPMS is caused by the diffusion of inflammatory mediators from the meninges, especially from meningeal B cell infiltration and strongly correlates with clinical disability scores (46).

Similarly, the astrocytes are considered of particular importance in MS. Besides there well-known role in “scar formation,” recently the astrocytes have been identified as a potent secretor of different pro-inflammatory cytokines making them a possible target for therapeutic interventions (30).

Mitochondrial dysfunction and energy deficits gained interest as a main mechanism of neuronal demise (53). The mitochondrial dysfunction with subsequent cellular hypoxia is especially relevant for the neurodegeneration of susceptible chronically demyelinated axons commonly found in PMS through energy failure, induction of apoptosis, and enhanced production of oxygen species (53). Corresponding to that, positive correlation between CSF lactate and disease progression was reported in RRMS patients (54). We confirmed a similar correlation in PPMS patients in a large multicentric CSF cohort including 254 PPMS patients (unpublished data). A positive correlation between CSF lactate and number of inflammatory MS plaques was reported in another study with 33 clinically isolated syndrome (CIS) patients (55). Another marker, the level N-acetylaspartate (NAA) measured using MRI spectroscopy, was reduced in MS patients and correlated to clinical severity of the disease (56).

Iron accumulation in PMS is an age-dependent process leading to free-radicals’ release, glutamate toxicity, and exacerbation of the neuronal demise (57). Iron-induced T2/hypointensities were reported in the GM, WML, and periventricularly around the veins (58) and is correlated with disease progression even better than brain atrophy (59). Iron deposition in deep gray matter was elevated in SPMS patients compared to controls (60). Furthermore, the iron-storage protein “ferritin” and soluble transferrin receptors were elevated in the CSF and serum of SPMS patients compared to controls (61). A similar role can be postulated in PPMS.

Over the last years, two main hypotheses were postulated to explain the neuronal demise in MS (62): the inflammation-induced neurodegeneration and the neurodegeneration-provoked inflammation.

The inflammation-induced demyelination leads to death loss and subsequent neurodegeneration as in EAE models (63). APP axonal spheroids indicating transected injured axons are correlated with T and B cell infiltrates in MS lesions (26). Furthermore, early focal inflammatory lesions are associated with higher density of transected axons than in the later phases of the disease (64). Moreover, the meningeal inflammation is correlated with the cortical axonal loss and seems to be the driving force for active demyelination as well as neuronal, axonal, and synaptic destruction in the cerebral cortex of MS patients (65).

The second hypothesis postulates an autonomous degeneration of oligodendrocytes and myelin followed by MiA and subsequently invasion of inflammatory cells (66).

A combination of both mechanisms where the low-degree inflammation provides constant insult to the susceptible oligodendrocytes or dysfunctional axon–glial unit like in cases of disturbed iron metabolism, glutamate homeostasis, and mitochondrial dysfunction cannot be excluded (67). Nevertheless, the success of the anti CD-20 in slowing the disease progression emphasizes the role of the inflammation leading to emergence of the first hypothesis as a valid explanation for the neuronal demise in PPMS (68).

Multiple sclerosis is an autoimmune disease involving both autoreactive B and T cells. EBV infection induces formation of autoreactive B cells (B_Auto) either through antigen mimicry or through infection of the normally present B_Auto-forming apoptosis-resistant active memory B_Auto cells (first hit). On the other hand, autoreactive CD4+ T cells (T_Auto) are induced by the intestinal microbiome (second hit). Both B_Auto and T_Auto interact in the peripheral lymphoid tissue. T_Auto cells cross the BBB and are further activated by perivascular B_Auto sells. B and T cells recognize the neuronal antigens and start an inflammatory reaction leading to migration of CD8+ cells and macrophages through BBB as well as activation of the microglial cells and astrocytes leading to demise of the neurons (first event). The apoptotic neurons and oligodendrocytes release other sequestrated antigens that subsequently will be recognized by B and T cells accentuating the inflammatory reaction (second event) with wide spread low-grade inflammatory process. Other factors like mitochondrial dysfunction, glutamate cytotoxicity, and iron accumulation lead to further demise of neurons (Figure 1). We postulate that different predisposing factors may not only affect the MS course directly but also interact with each other leading to further complexification of the pathophysiology of the disease and might contribute to the determination of the clinical phenotype of the disease by focal accentuation of the inflammatory reaction with the clinical symptoms of an exacerbation (Figure 2).

Most clinical trials in PPMS were disappointing: from methylprednisolone (69), through glatiramer acetate (70), rituximab (71), interferon-beta (72, 73), and at last fingolimod (74). Lack of efficacy, inappropriate patients’ selection, short study period, and non-optimal primary outcome are the major causes of the negative results (75). The ORATORIO study was the first phase 3 study to meet its primary endpoint in PPMS (4) (Table 1). In the following section, we try to summarize the most promising treatments in PPMS.

Despite our increasing knowledge and better understanding of the underlying mechanisms, many questions remain open. Accumulating evidences support considering PPMS as a part of the MS spectrum. However, there is no solid explanation of what exactly drives the development of the clinical phenotypes. In our opinion, the core of MS may be a slowly progressive low-degree inflammatory process driven by autoreactive apoptosis-resistant EBV-infected B cells that manifests itself clinically in genetically predisposed individuals only after a specific age threshold is exceeded. In presence of other factors (like vitamin D deficiency), a superimposed fluctuating high-grade inflammatory process appears in younger age and manifests itself in the form of recurrent exacerbations. Evidences supporting this hypothesis are (1) more than half of MS patients suffer from PMS (either as PPMS from the beginning or SPMS), (2) the striking clinical and pathological similarities between PP and SPMS, (3) the almost universal positive EBV status in MS patients, (4) the presence of EBV-infected B cells in brain and meninges of MS patients, perivascular spaces, and parenchyma, (5) the well-known change in age-dependent host response to latent EBV infection, (6) the success of B cell-depleting agents in RRMS and PPMS, (7) the “preliminary” success of T-cell-based therapy against EBV-infected B cells in SPMS, (8) the presence of vitamin D deficiency in RRMS but not PPMS patients and its well-described effect on the relapse rate but not disease progression, and finally, (9) almost all pathological aspects of the progressive phase like MiA, iron accumulation, mitochondrial dysfunction, involvement of the NAWM and NAGM, cortical and cerebral atrophy, as well as meningeal infiltration can be detected very early in the disease course even in CIS patients (55, 90–93). Further work is needed to prove the exact role of EBV in PMS forms, to characterize the B_Auto population and how do they differentiate, and at last to explain the role of different risk factors in PMS and their interactions in different populations.

The current therapy options for PPMS are promisingly increasing with upcoming possibilities of targeting different aspects of the disease (Figure 3). Combination of different treatments may be a viable approach in the future, considering the suboptimal effect of every single treatment alone so far.

Primary progressive multiple sclerosis is considered a relatively rare, but very challenging phenotype in the care of MS patients. Our current knowledge supports an underlying inflammatory-driven neurodegenerative process. Autoreactive EBV-infected B cells are essential to drive the progressive inflammation. The results of ocrelizumab and biotin announce the beginning of a new era with more therapies are eventually coming to the market in next years.

AA formulated the main concept, reviewed the published data, postulated the mentioned hypothesis, and drafted the manuscript. HT supervised and reviewed the article. MW reviewed the article.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The handling editor declared a past co-authorship with one of the authors, HT, and states that the process nevertheless met the standards of a fair and objective review.