B cells in the pathophysiology of myasthenia gravis

Abstract

Myasthenia gravis (MG) is an archetypal autoimmune disease. The pathology is characterized by autoantibodies to the acetylcholine receptor (AChR) in most patients, or to muscle-specific tyrosine kinase (MuSK) in others, and to a growing number of other postsynaptic proteins in smaller subsets. A decrease in the number of functional acetylcholine receptors (AChR), or functional interruption of the AChR, within the muscle end plate of the neuromuscular junction is caused by pathogenic autoantibodies. Although the molecular immunology underpinning the pathology is well understood, much remains to be learned about the cellular immunology contributing to the production of autoantibodies. This review will document research concerning the immunopathology of MG, bringing together evidence principally from human studies with an emphasis on the role of adaptive immunity and B cells in particular. Proposed mechanisms for autoimmunity, which take into account that different types of MG may incorporate divergent immunopathology, are offered.

Introduction

Patients with myasthenia gravis (MG) experience skeletal muscle weakness, worsened by activity ^1,2. MG is a multifactorial disease that includes immune dysregulation, predisposing genetics, and environmental factors. The disease is rare; the estimated annual incidence is 1–2 per 100,000, and the prevalence ranges from approximately 7–20 per 100,000 based on regional studies performed since 1990 ^3–5. Recent epidemiological investigations indicate that, like other autoimmune diseases, the incidence of MG is rising considerably ⁶. Such increases can be partly attributed to improved diagnostic precision and rising longevity of the populace, but a genuine rise in incidence may point toward the important role of environmental contributions.

MG is an archetype for B cell-mediated autoimmune disorders. The molecular immunopathology (Figure 1) is attributed to the presence of autoantibodies specifically targeting components of the acetylcholine receptor (AChR). The specific disease mechanism is defined by these autoantibodies and their recognition of a number of molecular elements of the AChR, which impairs neuromuscular transmission in the postsynaptic membrane. The specific end-plate abnormalities mediated by the autoantibodies include disruption of receptor signaling and complement-directed tissue damage. Unlike many autoimmune diseases, MG autoantibodies are demonstrably pathogenic ^7–12. This has been substantiated through numerous in-vitro approaches and perhaps demonstrated most convincingly through passive transfer of patient-derived serum or immunoglobulin, which reproduces features of the disease in experimental animals ¹³. Further evidence is provided by documented examples of maternal-fetal autoantibody transmission ^14,15 and neonatal transfer ^16,17, both of which can generate disease symptoms.

Figure 1.

Schematic diagram outlining the mechanistic hypothesis for the production of AChR or MuSK MG autoantibodies.

The proposed mechanistic path to autoantibody production in MG begins with naïve B cells (Steps 1 and 2), which likely encounter antigen(s) and receive T cell help in the lymph node (3). They then differentiate into memory B cells (4), antibody-secreting plasmablasts (5), and antibody-secreting long-lived plasma cells, which reside in the bone marrow (6A) and may also be present in the thymus (6B) of some patients with AChR MG. Plasmablasts and plasma cells may contribute to MG autoantibody production. B cell depletion therapy eliminates CD20+ memory and naïve B cells but does not directly eliminate plasmablasts or plasma cells, which are CD20-negative. After CD20-targeted depletion, MG serum autoantibody titers markedly diminish (especially in MuSK MG), suggesting that plasma cells are unlikely candidates for autoantibody production. Rather, short-lived plasmablasts are more viable candidates. As only a small fraction of these cells express CD20, the effectiveness of B cell depletion therapy may depend upon depletion of a pool of plasmablast-progenitor CD20+ memory B cells. Conversely, autoantibody titers that remain elevated following CD20-targeted depletion may be the product of long-lived plasma cells.

Genetic factors partly contribute to MG susceptibility ¹⁸. Although families in whom more than one member has MG are rare, limited MG twin-pair studies suggest rough approximations on MG concordance to be near 35% in monozygotic twins, and near 5% in dizygotic twins ¹⁹. These values, which are similar to a number of other autoimmune diseases, re-emphasize that varying degrees of both genetic and environmental factors contribute to the development of the disease ²⁰. Nearly all of the MG-associated genes identified to date are involved in the immune response; a pattern common to nearly all autoimmune diseases ²¹. The human leukocyte antigen (HLA) locus remains the most strongly associated risk factor for the disease, especially HLA-DQA1 ²². Examples of other genes encoding molecules that are involved in immune modulation include CTLA4, PTPN22, TNFRSF11A (RANK), ²² and TNFAIP3 interacting protein 1 (TNIP1) ²³, all of which participate in cell-signaling pathways.

The incidence of MG with AChR autoantibodies has been observed to distribute in a bimodal pattern. Cases of early-onset MG, defined as patients in whom symptoms occur before approximately age 40, are predominantly women. Conversely, the incidence of late-onset disease is higher in men than in women. MG with muscle-specific tyrosine kinase (MuSK) autoantibodies is predominantly found in women and has a peak incidence of less than 40 years of age ²⁴. Clinical classifications of MG include a number of subgroups ^25,26. Ocular MG, which is restricted to isolated ptosis, diplopia, or both, with no signs or symptoms of weakness elsewhere, is often the first manifestation of the disease. In 40–50% of ocular MG cases, autoantibodies are not detected ²⁷. This, however, does not exclude the possibility that they are present and causal. During this early stage of the disease they may be below the level of detection of commonly used assays and/or may be enriched at the neuromuscular junction (NMJ), the site of disease pathology, and thus not measurable in the serum.

Generalized MG usually includes symptoms associated with ocular disease, as well as weakness in extremity, bulbar, and/or respiratory muscles. Autoantibody status is used to classify the disease and has treatment implications in some cases. AChR, MuSK, and low-density lipoprotein (LDL) receptor-related protein 4 (LRP4) autoantibody positive and autoantibody seronegative represent additional major subsets. Within the AChR positive population, further subdivision categorizes early and late onset, early onset MG (EOMG) and late onset MG (LOMG) respectively. EOMG is often characterized by thymic hyperplasia, while cases of LOMG have thymic abnormalities less frequently. Concomitant autoimmune diseases in AChR MG patients are not uncommon, being reported in 13–22% ²⁸. The most frequently associated autoimmune disease is thyroid disease, followed by systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA).

The production of autoantibodies clearly implicates a principal role for B cells in the disease pathogenesis. Dysregulation in immune cell types extending beyond B lineages have been documented, indicating that a combination of factors contributes to disease manifestation; this theme is largely consistent throughout human autoimmunity. The contributions of B cells to the production of MG autoantibodies have been the focus of investigative interests, especially with the successful introduction of biological therapeutics that target these cells. In this review, we focus on the contributions that B cells make to the immunopathology of AChR, MuSK, and other subtypes of MG, although studies outside of AChR MG remain limited in number. We highlight T cell subsets that affect B cell responses, tissue compartments harboring B cells, B cell and antibody secreting subsets, and therapeutic intervention. Thymoma-associated MG and animal models are not included.

B cell-mediated MG immunopathology influenced by T cells

The autoantibodies in AChR MG are class switched, somatically mutated, and primarily of the IgG1 subclass. The initiation of the process that produces this affinity-driven maturation is dependent on T cell help. Accordingly, CD4 T cells are the main drivers in the immunopathogenesis of MG disease ^29–33. They play a multi-faceted role in immunity, from promoting inflammation to inducing immune tolerance and supporting B cell function. In MG, each of these roles has been demonstrated to be dysregulated, resulting in the production of autoantibodies and secretion of pro-inflammatory cytokines. Specifically, AChR-specific CD4 T cells produce IFN-γ and IL-17 in response to AChR peptide stimulation, supporting the role of Th1 and Th17 cells in the pathogenesis of MG ^34,35. Further evidence demonstrating Th1 and Th17 cells act as mediators of MG pathogenesis comes from experimental autoimmune MG (EAMG) studies used to evaluate the susceptibility of MG in the absence of Th1 and Th17 responses. Mice deficient for the IL-17 or T-bet gene, a Th1 transcription factor, were less susceptible to EAMG as demonstrated by decreases in clinical score, cytokine production, and AChR antibody production ^36,37. The only two studies profiling T cells in MuSK MG mirror the observations in AChR MG, in which CD4 T cells from MuSK MG patients exhibited enhanced frequencies of Th1 and Th17 cytokines ^38,39. Highlighting the infancy of study in this field is the fact that the identification of antigen-specific T cells in MuSK remains to be accomplished.

Regulatory T cells

The detection of AChR-specific Th1 and Th17 cells also coincides with the generation of defective regulatory T cell (Treg) responses in MG patients. Qualitatively, thymic, and peripheral Tregs from MG patients are impaired in their ability to suppress T cell responses ^40,41. This impairment of Tregs is further complicated by data demonstrating that the T cells in MG become resistant to suppression when cultured with functional Treg ⁴². Quantitatively, reports on the frequency of Tregs in MG patients have been mixed, with some groups showing a decrease in the frequency of Tregs in MG ^43–46, while others show no such alteration ^40,47,48,41. The differences could be attributed to the markers that were used to identify Tregs. Before FOXP3 was identified as the transcription factor for Tregs, Tregs were identified only by the expression of CD25. Collectively, the functional assays for Tregs provide the best support for their role in the disease, because they show that Tregs in MG patients are defective in regulating immune responses compared to those in healthy subjects.

T-follicular helper (Tfh) & T-follicular regulatory (Tfr) cells

AChR MG is a CD4 dependent B cell mediated disease; therefore, the interaction of Tfh, Tfr, and B cells are critical in dictating the development of MG disease. In 1988, before Tfh cells were recognized as a distinct CD4 T cell subset, it was demonstrated that CD4 T cells were critical in the production of anti-AChR autoantibodies by B cells ^30–33,49. Now it is well understood that CD4 T cells differentiate into Tfh cells upon expression of the master transcription factor Bcl6 and surface marker CXCR5, which allows migration into the germinal center where they support differentiation of B cells into memory B cells and antibody-secreting subsets (plasmablasts and plasma cells) ^50,51. In the thymus of MG patients, the frequency of Tfh cells and B cells increased along with the expression of the Tfh associated markers Bcl6, IL-21, PD-1, and ICOS on thymocytes ^52,53. CXCR5+ CD4 T cells are also found in the periphery, and it is suspected that they are memory Tfh cells that have migrated out of the germinal center ⁵⁴. Circulating Tfh cells, IL-21, and CXCL13 were found at a higher frequency in MG patients and correlated with disease severity ^55,56. To demonstrate the pathogenic role of Tfh cells in EAMG, knockdown of the Tfh transcription factor Bcl6 decreased the frequency of Tfh cells and IL-21 production, and importantly, diminished disease severity ⁵⁷.

A counterpart to Tfh cells, which suppress Tfh and B cell interactions in the germinal center, are follicular regulatory (Tfr) T cells ^58–60. Similar to Tfh cells, Tfr cells express CXCR5, PD-1, ICOS, and Bcl6. But unlike Tfh cells, Tfr cells are derived from natural Treg precursors, and express FOXP3 and BLIMP-1 ⁶¹. Tfr can suppress the production of IL-21 and IL-4 by Tfh cells and inhibit class switching and antibody production by B cells. In MG, the increase in Tfh cells is likely due to a dysregulation in Tfr cells as MG patients exhibit lower ratios of Tfr/Tfh cells and frequency of Tfr cells ^62,63. Moving forward, it is of interest to better understand the dynamics between Tfh, Tfr, and B cells in immune tolerance and disease pathogenesis.

B cell products: Immunoglobulins that drive MG pathology

Studies conducted many decades ago aimed to identify the circulating agent(s) that caused blocking of neuromuscular transmission in MG ^64,65. Through continued efforts that built upon these, and many other early investigations, it is now well understood that the molecular immunopathology of MG is due to the presence of circulating autoantibodies specifically targeting the AChR, MuSK, or LRP4. Patient-derived AChR, MuSK, and LRP4 autoantibodies in MG are demonstrably pathogenic ^7–12,66. Patients most often harbor only one of these autoantibody specificities. In approximately 75–80% of MG patients, AChR autoantibodies are detectable, MuSK autoantibodies are found in another 5–10% and LRP4 autoantibodies account for up to 20–30% of those without AChR or MuSK autoantibodies ^67,68. This leaves about 5–10% of MG patients without detectable serum autoantibodies to either AChR, MuSK, or LRP4; those individuals are termed seronegative ^67,68. However, the presumption that autoantibodies are absent in such cases is likely incorrect and requires alternative (autoantibody-independent) models to describe the pathology. The introduction of cell-based assays that express MuSK or clustered AChR was instrumental in demonstrating that a considerable portion of MG patients previously characterized as seronegative, did indeed have autoantibodies present in their serum ^69,70. Their autoantibodies were simply not detectable by the commercially available and commonly used radioimmunoassay ⁷¹. Accordingly, it is anticipated that seronegative MG will continue to yield to the influence of higher sensitivity assays and to the discovery of new autoantigen targets ^72–74.

Acetylcholine receptor autoantibodies

In the most common form of MG, autoantibodies directed against AChR are predominantly IgG1 and IgG3. These two subclasses effectively activate complement, leading to the formation of membrane attack complex ^75–77. Autoantibody and complement-mediated damage to the NMJ impairs neuromuscular transmission through damage to the postsynaptic neuromuscular junction in the form of simplification of postsynaptic junctional folds, removal of AChR from the membrane, and widening of the synapse ⁷⁸. Thus, AChR function is impaired by a number of autoantibody-mediated functions: complement-mediated tissue injury and removal of AChR from the muscle membrane (probably a major factor), cross-linkage of divalent AChR autoantibodies that results in internalization of AChR (by modulating autoantibodies), and rarely, a direct blockade of AChR by the autoantibodies ^79–84.

Muscle specific kinase autoantibodies

MuSK is thought to play an important role in clustering AChR at the NMJ to promote efficient neuromuscular transmission. Passive transfer and active immunization studies in animals have shown that MuSK autoantibodies are pathogenic ^85,86. In contrast to AChR MG, MuSK autoantibodies are predominantly IgG4 and do not effectively activate complement ⁸⁷. Thus, the pathogenesis of MuSK MG is likely not complement mediated, and disruption of the interaction between MuSK and the postsynaptic protein LRP4 and collagen Q may be the predominant pathologic mechanism of MuSK autoantibodies ^88–90. An interesting feature of secreted IgG4 antibodies is their ability to engage in “Fab-arm exchange” where a monospecific IgG4 protein may swap a heavy and light chain pair with another IgG4 to become bispecific ⁹¹. It has recently been shown that Fab-arm exchanged antibodies are present in MuSK MG and are pathogenic ⁹².

Other autoantibodies in MG

Recent studies of patients with AChR or MuSK autoantibodies undetectable through using standard assays have revealed additional autoantibodies specific to other NMJ proteins, including LRP4; agrin; collagen Q; cortactin; and the voltage-gated potassium channel, Kv1.4 ^93–100. LRP4 autoantibodies are predominantly IgG1, and studies suggest they are pathogenic ⁶⁶. Pathogenicity of the other autoantibodies remains unclear at the moment. Titin autoantibodies, which are known to be present in AChR autoantibody positive patients with thymoma and in LOMG, have recently been observed in patients without identifiable autoantibodies to AChR, MuSK, or LRP4 using a high-sensitivity radioimmunoprecipitation (RIA) assay ¹⁰¹. These findings suggest that titin autoantibodies may have a role as a diagnostic marker, but independent replication using a large collection of MG patients including those with and without thymoma is needed.

The contributions of B cell subsets to MG immunopathology

B cell tissue compartments in MG

The compartmentalized enrichment of disease-associated B cells can occur in tissue specific autoimmunity ¹⁰². For example, at the synovium, the site of tissue injury in patients with rheumatoid arthritis, B cells producing autoantibodies that are directed toward citrullinated protein/peptide antigens can be observed ¹⁰³. Although the site of tissue injury in AChR MG is the muscle end plate, the residence of B cells that express AChR autoantibodies is widely diverse (Figure 1). Over four decades ago, AChR-specific IgG was found in the MG thymus ¹⁰⁴. Shortly afterward, the first identification of AChR autoantibody-producing B cells that occupied the thymus was reported ¹⁰⁵. Thymic abnormality, defined by an AChR specific immune infiltrate, is now recognized as a fundamental characteristic of many (but not all) AChR MG patients.

The thymus in a subset of MG patients includes ¹⁰⁶ AChR expression by thymic epithelial cells and myoid cells, the presence of proinflammatory cytokines, and defective regulatory T cells ⁴². B cell supporting CD4+ helper T cells are also present ¹⁰⁷. B cells populating the hyperplastic thymus express markers of activation and display functional signs of activation ¹⁰⁸. B cells often organize in the hyperplastic thymus within tertiary lymphoid organs, frequently exemplifying many characteristics of germinal centers. The presence and frequency of these structures positively associate with circulating AChR autoantibodies, reflecting a contributory role in their production. While these characteristics of AChR MG thymus tissue are seminal, they are not applicable to the entire MG population given that the thymus of approximately 30% of AChR MG patients is not hyperplastic and therefore may contain only few, if any, disease-associated B cells ¹⁰⁹.

The B cell subset in the thymus tissue that is responsible for the detectable AChR autoantibody has not been precisely defined, but spontaneous production of AChR autoantibodies was demonstrated as most likely due to resident plasma cells ^110,111 with possible contributions from plasmablasts. Additional studies demonstrated that AChR autoantibody production by thymic lymphocytes can occur spontaneously or with mitogen stimulation, suggesting that heterogeneous B cell populations make such contributions ^112,113. AChR-specific CD27+ memory B cells are also likely to be present in the hyperplastic MG thymus, although specific identification of such has not been formally demonstrated. B cell repertoire sequencing and B cell immortalization studies have shown that the B cells resident in the MG thymus are broadly, clonally heterogeneous as they lack a dominant clone(s) among the infiltrate ^114,115. They harbor the characteristics of antigen experience, including somatic hypermutation ^{116–119,114} and biased usage of variable region gene segments, which include over-representation of the VH3 family at the expense of VH4 genes ¹¹⁴. Of course, among these sequenced B cells are those producing autoantibodies directed toward AChR. However, they represent a minor fraction of the total B cell infiltrate as they are not highly enriched ^120,111.

Given the major role of the thymus in MG AChR autoantibody production, thymectomy has been a long-standing treatment strategy. While thymus resection does not extinguish the disease in most cases, a recent, large, placebo-controlled clinical trial has confirmed the long-held belief that the procedure is beneficial ¹²¹. The AChR antibody titer in the majority of MG patients who have had a thymectomy does fall, but invariably does not reach undetectable levels ^122,123. Provided that all of the thymus tissue harboring immune infiltrate is surgically removed, the persistence of both disease and autoantibody is a strong indication that additional locations host autoantibody-producing B cells.

AChR autoantibody-producing B cells can also be found in the circulation ^124,30 and lymph nodes ¹²⁵. Measurement of AChR-specific B cells present in the circulation are often positively associated with thymic hyperplasia and high serum autoantibody titers ¹²⁴. AChR autoantibody-producing B cells have also been identified in the bone marrow ¹²⁶. The bone marrow is a well-recognized niche for long-lived plasma cells that directly contribute to the majority of serum immunoglobulin and are accordingly responsible for serological memory. Plasma cells in the bone marrow that produce AChR autoantibodies may contribute to persistently elevated titers after thymectomy ¹²⁷ and other treatments ¹²⁸ (discussed below).

Do lymphocytes, including autoantibody-producing B cells, gather at the site of tissue injury, the neuromuscular junction? Lymphocytes that localize to the end-plate region have been observed in the muscle tissue of AChR MG patients ^129–131. Localization in regions that include evidence of tissue damage has also been reported ¹³² although the specific lineage of such cells was not always elucidated. A subsequent investigation reported macrophages and T cells in MG muscle tissue but overall these infiltrates include few cells, are infrequently found at end-plates and are present in only a subset of patients ⁷⁶. Overall, the presence of lymphocytic infiltrates does not appear to associate with the loss of AChR from the end-plates ⁷⁶. MG autoantibody-producing B cells, at the end-plate, have not been unambiguously identified to date. Fittingly, it would also be of interest to associate such an infiltrate directly with autoimmune mechanisms to confirm that it is not non-specific inflammation that originates subsequent to the autoimmune-mediated tissue damage. Such investigations would require MG muscle biopsies, which are rarely obtained and represent a recognized constraint in the field.

While clonal B cell enrichment in the hyperplastic thymus tissue is conspicuous, perturbations in periphery are harder to observe. The circulating B cell repertoire was recently characterized through analysis of over 500,000 unique sequences, however only minor, albeit important, deviations from normal controls were evident ¹³³. This indicates that pathogenic B cells make small changes to the global repertoire that are not readily perceived without very large data sets. Finally, the thymus of patients with MuSK MG presents a picture that is quite different than that of AChR MG patients. Immune cell infiltrates and ectopic germinal centers that are frequently found in the AChR MG thymus tissue are rarely present in MuSK MG ¹³⁴. This infers that MuSK autoantibodies may develop and reside in a compartment other than the thymus.

Circulating Memory B cells and Antibody-secreting B cells

Studies directed toward understanding peripheral B cell functional abnormalities in MG are few. The few thorough immunophenotyping studies that have been conducted uncovered some abnormalities in the circulating B cell lineages. Similar to the findings of the repertoire sequencing studies, these perturbations are small and are observed only in B cell subsets. The difference in the frequency of B cell populations between AChR MG and healthy controls is unremarkable and there is no evidence of a general defect in B cell differentiation in MG patients. Double negative (CD27⁻IgD⁻) B cells, the frequency of which is elevated in some autoimmune diseases ¹³⁵ and associated with disease-specific autoantibody titer ¹³⁶, do not appear to be altered in AChR MG ¹³⁷. A subpopulation of B cells express CD5+. These cells may be associated with autoantibody production and regulation, although their role in immunobiology is not unambiguously defined. Several autoimmune diseases, such as SLE and Sjogren’s syndrome, are associated with enhanced frequencies of CD5(+) B cells. Similarly, there appears to be an increase in CD5+ B cells in a subset of AChR MG patients ¹³⁸. Plasmablast frequency, including recently activated subsets (HLA-DR^hi) in ocular MG, are elevated in a some AChR MG patients ¹³⁷. While such phenotyping studies can identify subset abnormalities that are likely related to autoimmunity, their limitation lies in that identification of autoantibody producers is not possible.

Regulatory B cells

Regulatory B cells (Bregs) are a rare B cell subset with a profound effect in promoting immune tolerance. They inhibit T cell expansion through IL-10, TGF-β, and IL-35 and support the differentiation of T cells into regulatory T cells (Tregs) ^139–141. Unlike Tregs, a specific transcription factor or a surface phenotype that identifies Bregs remains undefined, although several B cell populations including CD24^hiCD27^{+ 142}, CD24^hiCD38^{hi 143,144}, and CD25^hiCD71^{hi 145} B cells have been described to contain an enriched population of Bregs. Recent studies suggest that Bregs are antigen-specific and expand in the presence of an inflammatory event ^140,146. However, the mechanisms driving their generation and maintenance remain undefined.

One of the better-described subsets of Bregs are B10 cells, which are B cells identified by the production of IL-10, an immunosuppressive cytokine. In humans, B10 cells represent 0.3–0.8% of total B cells and decrease in number with age ¹⁴². Due to their low frequency in circulation, and in attempt to better profile B10 cells, B10 cells are expanded into B10 progenitor cells by a multi-day TLR stimulation with recombinant CD40L, followed up by a re-stimulation with phorbol 12-myristate 13-acetate (PMA) and ionomycin to detect IL-10 by flow cytometry ¹⁴². In 2014, the first MG study examining B10 cells revealed a decrease in the frequency of B10 cells and a subset of CD24⁺CD38⁺ B cells in AChR and MuSK MG compared to healthy controls ¹⁴⁷. This decrease in the frequency of B10 cells correlated with disease severity based on MGFA clinical classification assignments. Disease severity also correlated with the functional ability of B10 cells to suppress CD4 T cell proliferation ¹⁴⁸. Additionally, examination of patients undergoing rituximab treatment revealed that the repopulation of B10 cells was associated with the responsiveness to rituximab treatment. Although several EAMG and patient studies have supported the decrease in the frequency of B10 cells in MG, determining the mechanisms that limit B10 differentiation and their function remain an unmet need ^149–151. Moving forward, further research is needed to validate B10 cell frequencies as a biomarker of MG disease severity, and methods to increase B10 cell frequencies may hold promise as a treatment for MG.

Naïve B cell repertoire formation and immunological tolerance

Gene segment recombination that occurs in developing B cells is among the defining features of the adaptive immune system. The process involves stochastically recombined gene segments to generate functional antibodies (B cell receptors) that are expressed on the cell surface. The arbitrary joining is the basis for the vast diversity of the B cell repertoire needed for complete immunity, but this comes at a price; the developing B cell repertoire invariably includes autoreactive antibodies. However, control mechanisms are in place. The majority of autoreactive B cells are eliminated at two separate steps^152,153 during B cell development. First, a central checkpoint in the bone marrow between early immature and immature B cell stages removes the vast majority of developing B cells that express autoreactive antibodies. A second B cell tolerance checkpoint operates in the periphery, and selects against these autoreactive new emigrant B cells before they enter the mature naïve B cell compartment. Patients with autoimmune diseases often exhibit defective B cell tolerance checkpoints; this has been clearly demonstrated in RA, SLE, and type-1 diabetes (T1D) ^152,154,155.

B cell tolerance also functions improperly in both AChR and MuSK MG patients ¹⁵⁶. Accordingly, the naïve B cell repertoire in MG is shaped in the context of abnormal counter-selection during B cell development (that is, self-reactive cells are not eliminated). It follows that the manifestation of this abnormality is a naïve repertoire that is shaped differently from those in which B cell tolerance functions normally (Figure 1). Next generation deep sequencing allows for the comprehensive evaluation of the B cell receptor (BCR) repertoire properties in health and disease and provides the depth necessary to adequately depict the circulating peripheral repertoire, which includes up to 10¹¹ B cells in humans ¹⁵⁷. Application of this approach to the naïve B cell compartments in AChR and MuSK MG patients revealed repertoire features that were not observed where B cell tolerance functioned properly ¹³³. This emphasizes the impact of tolerance defects on peripheral autoimmune repertoires. It remains to be determined whether the autoreactive naïve B cell pool is the reservoir from which disease-associated autoantibodies are derived.

Therapeutics

Rituximab and new biologicals that target B cells

Rituximab is a clinically approved B cell directed biologic. It is a chimeric monoclonal antibody that targets the CD20 antigen found on subsets of the B cell lineage. CD20 is a 33-kDa protein expressed by all mature B cells, but not on pre-B or differentiated plasmablasts and plasma cells. Rituximab has been used as part of the standard therapy for non-Hodgkin lymphoma (NHL) and has emerged as a highly effective tool in the management of certain autoimmune diseases ^158,159. Interest in its use for MG began after a patient with both lymphoma and MG responded favorably to rituximab¹⁶⁰. Several recent studies ^{128,161–164} have demonstrated the benefits of rituximab treatment in MG. In addition to significant clinical improvement, rituximab also allowed for tapering and subsequent discontinuation of other immunotherapies in both AChR and MuSK MG patients ¹⁶³. The beneficial effect, although durable, is not permanent as post-rituximab relapses have been observed ¹⁶⁵. Interestingly, the B cell repertoire that reemerges following CD20-mediated B cell depletion therapy includes evidence of tolerance dysfunction in autoimmune conditions shown to have the checkpoint abnormalities prior to treatment ¹⁶⁶. This indicates that CD20-mediated depletion does not restore checkpoint function and suggests that disease manifestation may reemerge if the abnormal repertoire contributes to pathology.

Given that MG is a B cell mediated disease, the mechanism by which rituximab takes action may seem apparent. That is, B cells are depleted, which results in diminished pathogenic autoantibody production. However, a careful look uncovers that it may be more complicated (Figure 1). Long-lived plasma cells secrete the majority of circulating immunoglobulin. Given the absence of CD20 on their surface, they are not directly affected by rituximab. This is supported by unchanged immunization-generated and plasma cell-dependent antibody titers, such as those for tetanus, following anti-CD20 B cell depletion ¹²⁸. As described above, it is apparent that plasma cells and/or plasmablasts produce AChR autoantibodies. In AChR MG the autoantibody titer can remain positive after treatment with rituximab, even during clinical improvement ¹²⁸. Thus, it is not exactly clear how the treatment affects improvement. Similarly, NMO autoantibodies directed toward the AQP4 water channel do not fall correspondingly even though remarkable clinical improvement follows rituximab treatment ¹⁶⁷. While these collective data may reflect compartmentalization (in the muscle or CNS) of the respective autoantibodies and/or discordance of serum titer with disease activity, they may also suggest that additional mechanisms, besides those associated with autoantibody, may promote disease activity, at least in the case of NMO. In AChR MG, given the clear dependence of immunopathology on autoantibodies, it is likely that rituximab eliminates autoantibody-producing cells, but the compartment and specific cells remain to be identified. The persistence of a positive autoantibody titer in responding patients may reflect the inability of the assays to discriminate between detection of these autoantibodies and their pathogenic properties, and/or that the serum does not reflect autoantibody status in the central nervous system (CNS) tissue (in NMO) or neuromuscular junction (in MG).

The situation in MuSK MG may be quite different than that in AChR MG (Figure 1). The markedly diminished MuSK autoantibody titer, as early as three months after rituximab-mediated CD20+ B cell depletion ¹²⁸, suggests that long-lived plasma cells are not likely to be major contributors to MuSK autoantibody production. Rather, short-lived antibody-secreting cells such as plasmablasts are viable candidates as the primary driver of MuSK autoantibody production. As only a small fraction of these cells expresses CD20 ¹⁶⁸, the effectiveness of rituximab in MuSK MG may depend upon depletion of a pool of plasmablast-progenitor CD20+ memory B cells ^168,169. Our group has experimentally tested this model, the results of which suggest that plasmablasts, do indeed, contribute to the production of MuSK-specific autoantibodies ¹⁷⁰.

Provided that plasmablasts or plasma cells, the majority of which lack CD20 ¹⁶⁸, are a major source of pathogenic AChR autoantibodies, there are existing biologics that can effectively target these cells through their expression of CD19. Inebilizumab (formerly MEDI-551) is an anti-CD19 antibody with enhanced antibody-dependent cell-mediated cytotoxicity against B cell lineages. The drug is currently being evaluated in NMO given the known role of plasmablasts in AQP4 autoantibody production ^171,172. A similar evaluation in AChR MG has not been initiated, but given the similarities in these diseases, testing in rituximab-resistant patients could be considered.

MG induction; a side effect of cancer immunotherapy

The immune system is equipped with checkpoints that enforce immune homeostasis. Our immune system utilizes these checkpoints as a balance to prevent an immune response against self-antigens and thus the onset of autoimmune diseases like MG. Moreover, these checkpoints fine tune the immune response against pathogens to efficiently activate immune cells, and subsequently shut down their effector functions once the pathogen is removed to prevent immunopathogenesis. With an improved understanding of the mechanisms and the key proteins surrounding immune checkpoints, recent developments in cancer immunotherapy have focused on therapies that intervene in the biology of immune checkpoints to promote T cell activation ¹⁷³.

In the tumor microenvironment, one mechanism of immune evasion by tumor cells is the binding of inhibitory ligands with its corresponding receptor. Notably, T cells in the tumor microenvironment highly express inhibitory receptors and this interaction shuts down the effector function of T cells. Thus, the goal of checkpoint blockade therapy is to tilt this balance towards immune activation by blocking the interaction of inhibitory receptors with its ligand, promoting the generation of efficient tumor-eliminating T cells. Two of the well-studied immune checkpoint targets that have changed the landscape of cancer therapy are CTLA-4 and PD-1 ^174,175. CTLA-4 is upregulated after T cell activation and competes with CD28 for binding to B7-1 and B7-2. This competition favors CTLA-4 because CTLA-4 has a higher affinity to B7-1 and B7-2 than CD28 and subsequently results in immune suppression. PD-1 is an inhibitory receptor that is upregulated following T cell activation and inhibits T cell function.

Based on the mechanisms behind immunotherapy, it is no surprise that the main side effect of unleashing the control of the immune response has been immune-related disease manifesting as dermatologic, endocrine, gastrointestinal, and hepatic events. Although cases of MG after such immunotherapy are rare, the exacerbation of MG is a vital health concern that must be carefully monitored ¹⁷⁶. The first record of ipilimumab-induced (anti-CTLA4) MG was described in two patients with melanoma ¹⁷⁷. The functional burden caused by the development of MG contributed to the death of one patient, while the other patient improved after plasmapheresis. In lung cancer, MG symptoms were reported after combination therapy with either PD-1 or PD-L1 and CTLA-4 inhibitors ^178,179. In these particular cases, the patients developed complications associated with MG exacerbation and did not survive. Interestingly, there are also cases of checkpoint blockade therapy in patients with known history of MG ^180,181. In each of these cases, the patients were given anti-PD-1 therapy to treat melanoma, but developed an exacerbation of their MG disease. Fortunately, their MG symptoms resolved, and although their anti-PD-1 therapy was discontinued, the patients had stable disease or saw a shrinkage of their tumors.

The mechanism by which anti-CTLA-4, anti-PD-1, or anti-PD-L1 therapy induces MG disease is undefined. Available case reports describe a rapid progression to myasthenic crisis; therefore, any neurologic event must be recognized early and immunotherapy must be discontinued immediately, and then likely followed by initiation of high-dose steroids along with IVIg or plasmapheresis. Immunologically, these studies suggest that the MG-specific B and T cells are present in people with no prior clinical evidence of MG pathology. While these autoreactive cells are normally suppressed through immune-tolerance mechanisms, the introduction of checkpoint blockade therapies decreases the ability of T cells to discriminate self and non-self. It is not clear whether the checkpoint-inhibitor induced disease emulates the typical form in terms of immunobiology. Autoantibody specificity and participating cell types first need to be identified so similarities can be defined.

Collectively, with immunotherapy becoming front-line treatment for many cancer types, the number of reports describing exacerbated or emerging MG following the use of checkpoint-inhibitor immunotherapy continues to grow rapidly ^182,183,184. Therefore, prior to treatment, patients must be carefully scrutinized for any autoimmune diseases, family history of such, and current physical condition. Knowing a patient’s likelihood of an adverse event following immunotherapy assists clinicians in maximizing the power of immunotherapy while limiting side effects.

Future directions and conclusions

Our understanding of MG immunopathology remains incomplete. It appears that the mechanism used by B cells for autoantibody production in AChR and MuSK MG differ, but details of both are needed to understand the immunopathology that will guide the development of more effective therapies. Clinicians need superior biomarkers that accurately associate with disease activity and severity to help guide management. Directly measuring autoantibody-producing B cells may fulfill this requirement. The B cell response in MG most certainly requires T cell help. While it is understood that pro-inflammatory antigen-specific T cells are at work in AChR MG, very little is known about the role of T cells in MuSK or LRP4 MG.

Translational B cell studies that will have the most impact on the care and treatment of patients are likely to include further investigation of B cell depletion in MG disease subsets, including MuSK. Evaluation of second-generation B cell directed biologics such as ocrelizumab, which appear to have improved efficacy and are better tolerated over the first generation, should be considered ¹⁸⁵. However, highly specific treatment is desired over broad approaches that target an entire immune-cell lineage. Autoantigen-based chimeric immunoreceptors that can direct T cells to kill autoreactive B cells through the specificity of the B cell receptor (BCR) have been developed for the B cell-mediated autoimmune disease, pemphigus ¹⁸⁶. Such an approach may provide an effective strategy for specific targeting of autoreactive B cells in MG, and not only result in clinical improvement but also in an improved safety profile over other less specifically-targeted agents.

List of acronyms and/or abbreviations

MG

Myasthenia gravis

AChR

Acetylcholine receptor

MuSK

Muscle specific kinase

NMO

Neuromyelitis optica

NMJ

Neuromuscular junction

NINDS

Neurological Disorders and Stroke

HLA

Human leukocyte antigen

RANK

TNFRSF11A

TNIP1

TNFAIP3 interacting protein 1

LDL

Low-density lipoprotein

LRP4

Low-density lipoprotein receptor-related protein 4

EAMG

Experimental autoimmune MG

Treg

Regulatory T cell

Tfr

Follicular regulatory T cell

RIA

Radio-immunoprecipitation assay

BCR

B cell receptor

NHL

Non-Hodgkin’s lymphoma

T1D

Type 1 diabetes

BAFF

B cell activating factor

PMA

Phorbol 12 myristate 13 acetate

Tregs

Regulatory T cells

CNS

Central nervous system