Novel B cell-dependent multiple sclerosis model using extracellular domains of myelin proteolipid protein

Therapeutic success of B cell-targeting approaches in multiple sclerosis (MS) has intensified research into the pathogenic and regulatory roles these cells play in demyelinating disease. Dissecting the function of B cells in the MS mouse model experimental autoimmune encephalomyelitis (EAE) is largely confined to induction with either the myelin oligodendrocyte glycoprotein epitope MOG35–55 or the full-length recombinant human MOG protein, the latter representing the most-used B cell-dependent EAE model. There is a clear need to investigate B cell function in additional myelin antigen contexts. Unlike MOG35–55, where lack of B cells yields more severe disease, we show here that the immunodominant myelin proteolipid protein epitope (PLP178–191) elicited identical EAE in WT and μMT mice, suggesting an absence of B cell engagement by this peptide. We hypothesized that a longer PLP antigen may better engage B cells and designed a peptide encompassing the extracellular domains (ECD) of PLP. We demonstrate here that PLPECD-immunized B cell-deficient mice failed to exhibit EAE. In contrast, PLPECD induced EAE not only in WT mice, but in B cell-sufficient mice incapable of secreting antibodies, suggesting a predominant antigen presentation role. These results establish a novel, efficient B cell-dependent EAE model.

Recent clinical trials involving the targeted deletion of B cells [1][2][3] have reinvigorated intense interest in the role of this lymphocyte in immune-mediated demyelinating diseases such as multiple sclerosis (MS). It is essential to have multiple models, whereby various angles of B cell biology (such as their potential pathogenic or regulatory roles) can be deciphered in the context of disease. Historically, myelin oligodendrocyte glycoprotein (MOG) sequence  has been used to great success in inducing demyelinating disease in WT B6 mice 4,5 , where knockout capabilities and genetic tools like the 2D2 system (mice harboring a TCR transgene specific for the MOG  epitope) are available 6 .
B cell studies in MOG-induced demyelinating disease have led to a contradictory literature, in that there is evidence for both pathogenic and regulatory roles, as well as models that require and yet do not require the presence of B cells for EAE. These discrepancies can largely be explained by differences in model type, animal strain, and inciting antigen used. For instance, antibody titers and B cell numbers in the cerebrospinal fluid (CSF) have been shown to correlate with disease severity in mice and humans [7][8][9][10] . Likewise, full length recombinant human MOG (hMOG)-reactive serum transfers from B6 mice have been shown to cause disease in healthy animals 11 . Decreased demyelination in B cell-deficient versus WT B6 mice has been observed (yet both exhibited disease) 12 while other data demonstrate that induction of EAE failed altogether in B cell-deficient mice 13 . Interestingly, these studies utilized rodent MOG versus human MOG, respectively. It has further been demonstrated that deleting B cells from MOG 35-55 EAE mice during peak disease ameliorates clinical symptoms 14 . In addition to pathogenic antibody studies, B cells have been shown to contribute to rodent MOG-induced EAE in C3H mice as well as hMOG-induced EAE in B6 mice by reactivating CD4 T cells in the CNS through a likely antigen presentation function [15][16][17][18] . These reports of B cell pathogenicity notwithstanding, it has also been demonstrated that B cell-deficient mice cannot recover from myelin basic protein-induced EAE disease in B10.PL mice 19 and have a quicker and more robust MOG  and rodent MOG disease onset compared to their WT B6 counterparts 20,21 .
Further, deletion of B cells prior to MOG  disease onset led to exacerbated disease 14 , adding to studies attributing regulatory function to B cells during EAE.
These data clearly suggest that B cell subsets have varying functional capacities (IL-10-producing Bregs 20 versus CD4 T cell-reactivating CNS B cells 15 , for example). However, the general discrepancies seen in these reports, particularly regarding experiments involving B cell-deficient mice, can be explained by differences in the length of MOG antigen used for immunization (short peptide versus full-length recombinant protein), and the origin of MOG antigen (rodent versus human). It turns out that a single amino acid at position 42 (serine in the rodent, proline in the human) in the MOG sequence 22,23 can have striking consequences for the role of B cells in MOG-induced EAE models. Ultimately, it is understood that recombinant hMOG is necessary to induce a robust B cell-dependent mouse model of MS 13,22,23 . Yet it is still unclear whether this is due to pathogenic antibody production or requirement for antigen processing and presentation. In addition, production of the hMOG protein is cumbersome and represents a significant rate-limiting step in the study of B cell functionality in vivo during demyelinating disease. Finally, heavy reliance on a single model may bias our understanding of the role of these cells in complex human disease.
Investigating the role of B cells in myelin proteolipid protein (PLP)-induced EAE models is needed. PLP is highly conserved, where complete amino acid sequence homology is shared between mice and humans, and is very abundant in the CNS, comprising 50% of total myelin protein 24 . Indeed, understanding immune responses in this context is extremely relevant and may provide insights into the pathogenicity and regulation of MS. Whereas PLP-targeted responses are a focus in the relapsing-remitting SJL mouse model of EAE 25,26 , this protein or its peptides are underutilized in the context of B6 models, where there is a greater availability of genetic tools to dissect the function of various cells and molecules. We have recently utilized PLP 178-191 -induced EAE in B6 mice to demonstrate the robust disease regulatory role of PLP 178-191 -induced CD8 T cell responses [27][28][29][30] . We therefore decided to focus on PLP-induced EAE to develop a model wherein B cell function could also be delineated. Interestingly, our findings here suggest that B cells are ancillary during the immunodominant PLP 178-191 -incduced form of EAE in B6 mice. We hypothesized that designing a longer peptide may provide a processable antigen for B cells to engage and present to T cells, thus alleviating the need for intricate whole protein production methodology. Here, we utilized a novel designed peptide (PLP ECD , an 83-mer that incorporates the extracellular sequences of PLP while excluding positions buried within or just proximal to transmembrane regions) to test B cell-dependency in the context of PLP-induced demyelinating disease in B6 mice. We offer an efficient new mouse model suitable to investigating potential roles for B cells in EAE.

Statistical analysis.
When comparing two groups, data were analyzed using the Welch's t-test. For multiparametric data, the two-way ANOVA test with Tukey post-test was performed. All graph production and statistical analyses were done using GraphPad Prism software (La Jolla, CA).

B cells are not required for EAE induction or progression in PLP 178-191 -immunized B6 mice.
www.nature.com/scientificreports www.nature.com/scientificreports/ summation of disease scores over time per mouse) was also not different (Fig. 1B). In contrast, MOG 35-55 /CFAimmunized B-cell deficient mice exhibited faster EAE kinetics and showed more severe disease (Fig. 1C,D), and is in accordance with previous literature 20 . This is also in contrast to observations in a susceptible BALB/c model, where B cells may play a regulatory role in the context of PLP [180][181][182][183][184][185][186][187][188][189][190][191][192][193][194][195][196][197][198][199] or PLP 185-206 peptide-induced EAE 33 . These data indicate that B cell engagement is suboptimal in PLP 178-191 -immunized B6 mice, and suggests that a short peptide-length PLP antigen is insufficient to drive either protective or pathogenic B cell responses in vivo in mice of this genetic background. These results also illustrate that it is not yet clear which PLP domains drive which B cell characteristics during disease.  www.nature.com/scientificreports www.nature.com/scientificreports/ vivo, and furthermore been used to read out immune responses to CNS myelin antigens 30,37-41 . Here, we performed DTH experiments as described previously 30 . Briefly, this involved making mice immune through s.c. flank injection with myelin peptide/CFA and subsequently challenging ear pinnae of immune mice with the same or similar myelin peptide (without CFA) or PBS alone (control) injection two weeks later. Using an engineer's micrometer, measuring the ear thickness of injected ears at 48 h compared to ear thickness at a background 0 h measurement, one can identify "delta ear swelling" as a readout of the DTH reactions occurring in the challenged, immune animals. Specifically, individual mice were challenged with PLP 178-191 in the right ear pinna and PLP ECD in the left ear pinna 14 days following either PLP 178-191 -or PLP ECD -immunization. Swelling was measured at 48 h post-ear challenge. PLP ECD -immunized mice exhibited robust ear swelling when challenged with either PLP ECD or PLP [178][179][180][181][182][183][184][185][186][187][188][189][190][191] (Fig. 2C), indicating that PLP ECD is sufficiently engaged, processed, and presented in these mice. As expected, PLP 178-191 -immunized mice exhibited robust swelling in PLP 178-191 -challenged ears (Fig. 2D, black symbols). PLP ECD also induced DTH responses in PLP 178-191 -immunized mice (Fig. 2D, white symbols), albeit at lower levels than in PLP ECD -immunized mice (Fig. 2E). This may be due to availability of epitopes on a per cell basis in vivo. PLP 178-191 -driven DTH was similarly elicited in both PLP 178-191 -and PLP ECD -immune mice (Fig. 2F), indicating a robust induction of responses to this pathogenic epitope. Importantly, DTH does not develop when challenging ear pinnae of PLP 178-191 -immune mice with a control PBS alone injection (Fig. 2G) or with a non-cognate ovalbumin peptide (Fig. 2H). Collectively, these results demonstrate the immunogenicity of PLP ECD .

PLP ECD induces B cell-dependent EAE in B6 mice.
To test whether PLP ECD could successfully induce EAE, and to determine whether PLP ECD could successfully engage B cells (either in a pathogenic or regulatory role), WT and B cell-deficient (µMT) B6 mice were actively immunized with PLP ECD /CFA s.c. on day 0 along with i.p. injections of pertussis toxin on days 0 and 2. Clinical disease scores were monitored over time. Repeated experimentation composited in Fig. 3 demonstrates that EAE was indeed robustly induced in WT B6 mice following immunization with PLP ECD (Fig. 3, black symbols). Intriguingly, µMT mice exhibited minimal disease induction (Fig. 3, white symbols). The two groups of mice were statistically different not only when comparing raw paralysis scores, but also with respect to day of onset, peak score of disease, and CDI (Fig. 3, EAE parameter  table). These results ultimately suggest that PLP ECD is a potent inducer of encephalitogenic responses in B6 mice, and that PLP ECD -driven EAE disease is B cell-dependent.
It is important to acknowledge that μMT mice could have compromised priming dynamics due to altered lymphoid and splenic architecture, resulting in priming loss brought about in a somewhat lymphocyte-extrinsic manner. However, this is not the case in our models, where despite the complications μMT mice present in these respects, Fig. 1 clearly shows equivalent EAE between μMT and WT mice when immunized with PLP 178-191 . Likewise, even in the face of B cell deficiency, PLP 178-191 -immune μMT and WT mice show similar PLP 178-191 -driven DTH reactions over background (Supplementary Fig. 1). This suggests that T cell priming in the context of PLP 178-191 is not only largely unaffected, but is dominated by antigen presenting cell populations other than B cells (eg. dendritic cells). Priming is likely much different in the context of the larger antigen PLP ECD . As demonstrated in Fig. 3, lack of B cells results in loss of robust EAE induction. Further, in the context of PLP ECD immunization, lack of B cells results in a diminished PLP 178-191 -driven DTH reaction, similar to background levels ( Supplementary  Fig. 1). This result indicates that unlike PLP 178-191 immunization, which is largely a B cell-independent priming, PLP ECD requires B cells as a critical antigen-presenting cell for optimal T cell priming, and that other antigen presenting cells like dendritic cells are unable to fully liberate and present the bulk of the immunodominant epitope from the novel peptide. Future studies of PLP ECD -induced germinal centers to establish how B cells interact with T cells in this model will be illuminating. Data presented includes scores from two to three independent experiments. ns = not significant; *p < 0.05; ***p < 0.001; ****p < 0.0001. (2020) 10:5011 | https://doi.org/10.1038/s41598-020-61928-w www.nature.com/scientificreports www.nature.com/scientificreports/ B cell-dependency in PLP ECD -driven EAE disease in B6 mice occurs irrespective of antibody production. Functionally, the dependency of B cells in this model likely rests on antigen-presentation and/or pathogenic antibody production. It is known that oligoclonal switched IgG antibody bands within the CSF, along with B cells, are an indicator of MS disease 9 . Antibodies have also been shown to be pathogenic in some models of EAE. Antigen presentation function may also be important for B cell pathogenicity, as B cells have been shown to reactivate encephalitogenic CD4 T cells in the CNS of EAE mice [15][16][17][18] . Clearly, B cells play a role in MS as recent clinical trials have shown success with depleting B cells 1-3 . However, these depletion strategies target CD20+ B cells and leave behind antibody-producing plasma cells and plasma blasts. Given these data, in addition to those seen here, we wondered whether antigen presentation, rather than antibody production, explains the B cell-dependency of EAE induction in this model, perhaps reflecting their role in treated MS patients. To formally test this, we utilized AID−/−μS−/− mice 42 , which have a full complement of B cells, but lack the ability to secrete antibodies. We hypothesized that unlike μMT mice, which lack mature B cells and do not exhibit EAE disease, AID−/−μS−/− mice would be susceptible to EAE induction with PLP ECD . Indeed, upon immunization, AID−/−μS−/− exhibited a similar disease course compared to WT mice (Fig. 4A). And as expected given the data in Fig. 3, μMT mice showed little EAE disease in comparison to their B cell-sufficient counterparts (Fig. 4A). In analyzing EAE disease parameters, AID−/−μS−/− mice were statistically indistinguishable from WT mice in raw paralysis scores exhibited over time, day of disease onset, peak disease score, and cumulative disease index compared to their μMT counterparts (Fig. 4A, table). These data indicate that the absence of antibody production by B cells was insufficient to render mice resistant to EAE induction. To confirm immune responses in these mice, DTH was elicited by challenging the right ear pinnae with PLP 178-191 and the left ear pinnae with PBS alone (similar to that described in Fig. 2 above) on day 23 post-immunization. As expected, given the clinical score data in Fig. 4A and DTH data in Supplementary Fig. 1, μMT mice exhibited an inferior DTH reaction compared to their B cell-sufficient counterparts, indicating that the lack of antibodies had little effect in driving inflammatory responses and EAE (Fig. 4B), and further supports that antigen-presenting cells other than B cells (eg. dendritic cells) are insufficient for optimal T cell priming in PLP ECD -immunized B6 mice. www.nature.com/scientificreports www.nature.com/scientificreports/ conclusions We present here a novel B cell-dependent, PLP-driven murine EAE model. These results indicate that PLP ECD engages WT B6 B cells more efficiently than a short, albeit immunodominant peptide (PLP [178][179][180][181][182][183][184][185][186][187][188][189][190][191] ), leading to a likely pathogenic antigen-presenting cell role. This may ultimately mimic what is occurring in human MS patients, and warrants further study into whether depletion of PLP ECD -driven B cells yields protection from demyelination. Any potential cellular and molecular immunological knowledge gained from this model may have increased translational implications given PLP ECD utilizes sequences from the highly abundant (50% of total CNS myelin protein) and highly conserved (shares 100% amino acid sequence homology between mouse and human) myelin PLP. This work contributes an additional model to the field for investigating the pathogenesis and regulation of demyelinating disease, offering multiple ways for investigators to confirm immunological findings. PLP ECD also provides an efficient way to investigate B cell-dependent demyelinating disease, as it obviates the need for various full-length protein expression systems in order to elude high titers of recombinant protein. It is also worth noting this study may represent a proof of principle that longer peptides (an 83-mer in this case) encompassing immunodominant sequences may be able to be designed for study of B cells' role in other autoimmune diseases.