Comparative Analysis of T-Cell Responses to Aquaporin-4 and Myelin Oligodendrocyte Glycoprotein in Inflammatory Demyelinating Central Nervous System Diseases

Autoantibodies against aquaporin-4 (AQP4-Ab) and myelin oligodendrocyte glycoprotein (MOG-Ab) are associated with rare central nervous system inflammatory demyelinating diseases like neuromyelitis optica spectrum disorders (NMOSD). Previous studies have shown that not only antibodies, but also autoreactive T-cell responses against AQP4 are present in NMOSD. However, no study has yet analyzed the presence of MOG reactive T-cells in patients with MOG antibodies. Therefore, we compared AQP4 and MOG specific peripheral T-cell response in individuals with AQP4-Ab (n = 8), MOG-Ab (n = 10), multiple sclerosis (MS, n = 8), and healthy controls (HC, n = 14). Peripheral blood mononuclear cell cultures were stimulated with eight AQP4 and nine MOG peptides selected from previous studies and a tetanus toxoid peptide mix as a positive control. Antigen-specific T-cell responses were assessed using the carboxyfluorescein diacetate succinimidyl ester proliferation assay and the detection of granulocyte macrophage colony-stimulating factor (GM-CSF), interferon (IFN)-ɤ and interleukin (IL)-4, IL-6, and IL-17A in cell culture supernatants. Additionally, human leukocyte antigen (HLA)-DQ and HLA-DR genotyping of all participants was performed. We classified a T-cell response as positive if proliferation (measured by a cell division index ≥3) was confirmed by the secretion of at least one cytokine. Reactivity against AQP4 peptides was observed in many groups, but the T-cell response against AQP4 p156-170 was present only in patients with AQP4-Ab (4/8, 50%) and absent in patients with MOG-Ab, MS and HC (corrected p = 0.02). This AQP4 p156-170 peptide specific T-cell response was significantly increased in participants with AQP4-Ab compared to those without [Odds ratio (OR) = 59.00, 95% confidence interval-CI 2.70–1,290.86]. Moreover, T-cell responses against at least one AQP4 peptide were also more frequent in participants with AQP4-Ab (OR = 11.45, 95% CI 1.24–106.05). We did not observe any significant differences for the other AQP4 peptides or any MOG peptide. AQP4-Ab were associated with HLA DQB1*02 (OR = 5.71, 95% CI 1.09–30.07), DRB1*01 (OR = 9.33, 95% CI 1.50–58.02) and DRB1*03 (OR = 6.75, 95% CI = 1.19–38.41). Furthermore, HLA DRB1*01 was also associated with the presence of AQP4 p156-170 reactive T-cells (OR = 31.67, 95% CI 1.30–772.98). To summarize, our findings suggest a role of AQP4-specific, but not MOG-specific T-cells, in NMOSD.

Autoantibodies against aquaporin-4 (AQP4-Ab) and myelin oligodendrocyte glycoprotein (MOG-Ab) are associated with rare central nervous system inflammatory demyelinating diseases like neuromyelitis optica spectrum disorders (NMOSD). Previous studies have shown that not only antibodies, but also autoreactive T-cell responses against AQP4 are present in NMOSD. However, no study has yet analyzed the presence of MOG reactive T-cells in patients with MOG antibodies. Therefore, we compared AQP4 and MOG specific peripheral T-cell response in individuals with AQP4-Ab (n = 8), MOG-Ab (n = 10), multiple sclerosis (MS, n = 8), and healthy controls (HC, n = 14). Peripheral blood mononuclear cell cultures were stimulated with eight AQP4 and nine MOG peptides selected from previous studies and a tetanus toxoid peptide mix as a positive control. Antigen-specific T-cell responses were assessed using the carboxyfluorescein diacetate succinimidyl ester proliferation assay and the detection of granulocyte macrophage colony-stimulating factor (GM-CSF), interferon (IFN)-È and interleukin (IL)-4, IL-6, and IL-17A in cell culture supernatants. Additionally, human leukocyte antigen (HLA)-DQ and HLA-DR genotyping of all participants was performed. We classified a T-cell response as positive if proliferation (measured by a cell division index ≥3) was confirmed by the secretion of at least one cytokine. Reactivity against AQP4 peptides was observed in many groups, but the T-cell response against AQP4 p156-170 was present only in patients with AQP4-Ab (4/8, 50%) and absent in patients with MOG-Ab, MS and HC (corrected p = 0.02). This AQP4 p156-170 peptide specific T-cell response was significantly increased in participants with AQP4-Ab compared to those without [Odds ratio (OR) = 59.00, 95% confidence interval-CI 2.70-1,290.86]. Moreover, T-cell responses against at least one AQP4 peptide were also more frequent in participants with AQP4-Ab (OR = 11.45, 95% CI 1.24-106.05). We did not observe any significant differences for the other AQP4 peptides or any MOG peptide. AQP4-Ab were associated with HLA DQB1 * 02 (OR = 5.71, 95% CI 1.09-30.07), DRB1 * 01 (OR = 9.33, 95% CI

INTRODUCTION
Autoantibodies targeting the aquaporin-4 (AQP4) water channel protein and the myelin oligodendrocyte glycoprotein (MOG) are associated with a broad spectrum of human central nervous system (CNS) demyelinating diseases (1,2). While AQP4-specific antibodies target the AQP4 water channel protein expressed on astrocyte end-feet processes causing a severe astrocytopathy called neuromyelitis optica (NMO) (3,4), MOG-specific antibodies target the extracellular N-terminal immunoglobulin variable (IgV)-domain of MOG expressed on myelin-forming oligodendrocytes (2,5,6). Autoantibodies against AQP4 (AQP4-Ab) have emerged as highly sensitive and specific biomarker for the diagnosis of NMO (3). However, not all patients presenting with clinical features suggestive of an NMO-disease phenotype are positive for AQP4-Ab (7), and a significant proportion of those seronegative patients harbor antibodies to MOG. This created a diagnostic uncertainty reflected in the pathogenetically undefined category of NMO spectrum disorders (NMOSD) proposed in 2015 (1,2,8).
Several lines of evidence suggest that autoreactive CD4 + T lymphocytes are key players in the pathogenesis of Abassociated demyelinating CNS diseases. First, passive transfer models using AQP4-specific human IgG are not considered pathogenic without T-cell induced disruption of the blood brain barrier (BBB) (9)(10)(11). The exact relevance of T-cell independent pathogenicity of a high affinity rodent monoclonal AQP4-Ab (12) remains to be determined, as serum AQP4-Ab in human NMOSD patients are polyclonal, with a wide range of affinities and often much lower antibody titers (9,10,(13)(14)(15). The serum concentration of AQP4-Ab is many times higher than in the cerebrospinal fluid (CSF) (13,(16)(17)(18), and peripheral B cells have the capacity to produce AQP4-Ab in vitro (19,20). Thus, it is supposed that these antibodies are produced outside the CNS and that T effector cells might initiate CNS inflammation leading to BBB disruption and entry of antibodies (6,(9)(10)(11)(21)(22)(23). Local activation of CD4 + T-cells in the CNS is indispensable for providing an inflammatory microenvironment that also enables the initiation of CNS inflammation orchestrating BBB breakdown, lesion location and formation and thus facilitates Ab-mediated disease propagation (6,11,23). Second, AQP4-Ab and MOG-Ab are class-switched complement-fixing antibodies depending on T-cell help to be generated emphasizing the pivotal role of antigen-specific T-cell responses. Finally, there is ample evidence that activated T-cells are enriched at lesion sites (11,24,25) and that the pathogenic effectors are CD4 + T-cells of either T helper (Th)-1 lineage producing pro-inflammatory interferon (IFN)-È or of Th17 lineage producing pro-inflammatory interleukin (IL)-17A (26,27). Moreover, NMOSD patients also display a higher proportion of Th17 cells or cytokines like IL-6 (28)(29)(30)(31)(32)(33)(34).
While the high diagnostic value of AQP4-Ab as hallmark serologic marker in NMOSD has been shown and AQP4specific T-cells have been examined in NMOSD patients (31,(35)(36)(37)(38), the role of MOG-Ab or MOG-specific T-cells is less clear. Since MOG-Ab can be found in up to 50% of AQP4-Ab seronegative NMOSD patients, it is possible that MOG-specific T-cells could play a role in NMOSD development. So far there is no published information about MOG-specific T-cells in NMOSD and related conditions. Until now all studies focused on MOG-specific T-cell responses from MS patients (39)(40)(41) or in experimental autoimmune encephalomyelitis (42)(43)(44).
Here, we aimed to analyze the T-cell reactivity in response to selected eight AQP4 and nine MOG peptides and their possible restriction to a particular human leukocyte antigen (HLA)-DQ and HLA-DR genotype, and to examine the functional phenotype of autoreactive CD4 + T-cells in patients with AQP4-Ab or MOG-Ab.

Patients and Control Subjects
Eight NMOSD patients with AQP4-Ab, 10 patients with MOG-Ab, 8 patients with MS and 14 healthy controls (HC) were included in this study. NMOSD and MS was diagnosed according to recently published criteria (1,45,46). Within the MOG-Ab positive group, one patient also fulfilled the 2015 diagnostic criteria for NMOSD (1), 8 of the other 9 patients had related clinical presentations (three bilateral and one unilateral monophasic optic neuritis, one recurrent optic neuritis, one monophasic and one recurrent myelitis, one acute demyelinating encephalomyelitis with recurrent optic neuritis and one recurrent demyelinating disease) and one patient fulfilled the diagnostic criteria for MS.
Demographic and clinical data of all participants are shown in Table 1.
All  Patients with MOG-Ab (one NMOSD, three bilateral and one unilateral monophasic optic neuritis, one recurrent optic neuritis, one monophasic and one recurrent myelitis, one acute demyelinating encephalomyelitis with recurrent optic neuritis, one recurrent demyelinating disease and one MS). c Significance of group differences was analyzed using the Chi-square test, *significant difference to HC group. d P-values were adjusted for 20 comparisons using Bonferroni's correction for multiple comparisons. e Data are shown as median (range). f Immunomodulatory or immunosuppressive treatment with rituximab (2) or azathioprine (2). g Immunomodulatory or immunosuppressive treatment with rituximab (2) or plasma exchange (2). h Immunomodulatory or immunosuppressive treatment with rituximab (1), interferon-β (1), natalizumab (2), azathioprine (1) and teriflunomide (1). Due to limited sample availability not all AQP4-Ab or MOG-Ab positive patients were investigated for T-cell reactivity against AQP4 or MOG peptides. Ab, antibody; AQP4, aquaporin-4; CDI, cell division index; CFSE, carboxyfluorescein succinimidyl ester; HC, healthy controls; MOG, myelin oligodendrocyte glycoprotein; MS, multiple sclerosis; n.a., not applicable/available; ns, statistically not significant after correction for multiple comparisons.
AN3041) and University of Zürich, Switzerland (KEK ZH 2013-0001) and all patients or their caregivers and controls gave written informed consent.

AQP4-Ab and MOG-Ab Detection Assays
Serum AQP4-Ab were analyzed using live cell-based immunofluorescence assays as described previously (47). Serum MOG-Ab were analyzed using recombinant live cell-based immunofluorescence assays with HEK293A cells transfected with full-length MOG (human MOG α-1 EGFP fusion protein) as described previously (47). Sera were tested at dilutions of 1:20 and 1:40 and MOG-Ab positivity was titrated with serial dilutions with a threshold of 1:160 to define MOG-Ab positivity. Isolated IgM reactivity was excluded using IgG constant chain (Fc)-specific secondary antibodies (48,49).

T-Cell Epitope Mapping Using the CFSE Proliferation Assay
Peripheral blood mononuclear cells (PBMC) were isolated by density gradient centrifugation over Histopaque 1077 (Sigma-Aldrich, St. Louis, MO, USA) according to the manufacturer's

Statistical Analysis
The primary hypothesis of this study was that T-cell responses are associated with auto-antibody responses, i.e., AQP4-specific T-cells are increased in participants with AQP4-Ab and MOGspecific T-cells are increased in participants with MOG-Ab.

AQP4-Ab Are Associated With AQP4-Specific CD4 + T-Cell Reactivity
We adopted a cell division analysis procedure based on the quantitative dilution of the fluorescent dye CFSE to investigate the CD4 + T-cell autoreactivity of individuals with AQP4-Ab, MOG-Ab, MS and HC against selected AQP4 peptides. All participants showed a positive CD4 + T-cell proliferative response with a CDI ≥ 3 to the positive control TTX (Figures 3,  4A and Table 1). T-cell proliferation with a CDI ≥ 3 for at least one AQP4 peptide was observed in the majority of patients with AQP4-Ab (88%), 43% of patients with MOG-Ab, 50% of MS patients and 29% of HC. These proliferative Tcell responses against at least one AQP4 peptide were more frequent in participants with AQP4-Ab as compared to HC (OR = 17.50, 95% CI 1. .89) or all AQP4-Ab negative participants (OR = 11.45, 95% CI 1.24-106.05; Figure 5).
The clinical presentation of the 4 patients who had increased T-cell reactivity to AQP4 p156-170 were not substantially different from the other AQP4-Ab NMOSD patients. All four cases (all female, age 20-53 years, disease duration 0.4-13.2 years) had a relapsing NMOSD disease course (2-8 relapses), three of them were treated with rituximab and the fourth patient was under high-dose corticosteroids before the initiation of rituximab treatment. Two of the four patients had a relapse at the time of blood sampling, one of them before the initiation of rituximab treatment.
The functional phenotype of proliferating T-cells was characterized by investigating the secretion of the cytokines IL-4, IL-6, IL-17A, GM-CSF, and IFN-È into cell culture supernatants of the CFSE proliferation assay using ELISA. Cytokine concentrations of IL-4 and IL-17A after stimulation with AQP4 peptides, but not after stimulation with TTX, were below the detection limit of ELISA. Quantitative and qualitative values for GM-CSF, IFN-È, or IL-6 levels are shown in Figures 3, 4B-D. A comprehensive analysis of all proliferative and cytokine responses to any AQP4 peptide and p156-170 is shown in Figure 5. From this figure it is evident that overall these responses are increased in AQP4-Ab positive patients as compared to HC or all AQP4-Ab negative participants.

No Association of MOG-Ab With MOG-Specific CD4 + T-Cell Reactivity
In a next step, we analyzed the CD4 + T-cell autoreactivity of patients with AQP4-Ab, MOG-Ab, MS and HC against selected MOG peptides. All participants showed a positive CD4 + T-cell proliferative response with a CDI ≥ 3 to the positive control TTX. We observed no statistically significant differences after challenge with the different MOG peptides between groups (Figures 6, 7A and Table 1). T-cell proliferation with a CDI ≥ 3 for at least one MOG peptide was observed in 40% of AQP4-Ab positive patients, 40% of MOG-Ab positive patients, 63% of MS patients and 50% of HC.

DISCUSSION
In this study we analyzed peripheral blood T-cell responses to AQP4 and MOG peptides in individuals with AQP4-Ab, MOG-Ab, MS, and HC. We identified significantly increased AQP4specific CD4 + T-cell reactivity to AQP4 peptide 156-170 in 4 of 8 AQP4-Ab positive NMOSD patients, but in none of the other groups. In contrast, we could not detect any significant diseasespecific T-cell response to other AQP4 or MOG peptides. AQP4 peptide 156-170 has already been described as a T-cell epitope in NMOSD patients (31) and is also one of the most important Bcell epitopes recognized by AQP4-Ab (14,15,58,59). Three other immunodominant T-cell epitopes/peptides of the AQP4 protein have been described by Varrin-Doyer et al. (31), which were also included in our study. However, we and other authors could not confirm immunodominance for these particular determinants (38,60). The possible reasons for this discrepancy could be explained by the different methods used (CFSE, 3 H-thymidine incorporation proliferation assays, cytokine secretion) and the different genetic background, i.e., HLA associations of the study populations. Several studies have identified over-representation of HLA-DPB1 * 0501, HLA-DRB1 * 0301, or HLA-DRB3 in NMO patients (31,(61)(62)(63). However, only DRB1 * 0301 but not any of the other HLA alleles was overrepresented in our study population, indicating differences in the genetic background. In contrast, we found an overrepresentation of HLA-DQB1 * 02 and HLA-DRB1 * 01 in our study.
We found no differences in MOG-specific T-cells between the four different groups with our experimental approach. The reason for the observed results could be explained by ignorance of the immune system of the MOG protein (66). In contrast to the AQP4 protein, which is also highly expressed in the periphery (67,68) and hence underlies highly regulated mechanisms of self-tolerance (69,70), MOG is only expressed in the CNS at very low levels (71,72) and therefore not subject to intense immune surveillance. This might explain why HC showed lower response to AQP4 determinants, but in vitro stimulation with MOG peptides also caused profound T cell response in some healthy subjects.
Synthetic MOG peptides used in this study may not accurately represent naturally processed antigen and MHCpresented peptides in an in vivo setting (73). One of the major pathogenic mechanisms of MOG-Ab is considered the enhanced presentation of native MOG protein to T cells via Fc receptor mediated internalization of the antigen-Ab complex (74)(75)(76). Therefore, it is possible that MOG-reactive T cells can only be detected using intact MOG protein as the antigen. Indeed, Bronge et al. were able to identify increased frequencies of IFNγ, IL-22 and IL-17A producing MOG-specific T-cells in patients with MS using bead-bound MOG as the antigen (77).
A major limitation of our study is the small number of included participants. This number reflects the expected number for our clinical centers, since NMOSD and MOGrelated disorders are rare with a worldwide prevalence of 1-4/100,000 comparatively similar in most populations. Other limitations are that most patients received immunosuppressive therapy during sample collection related to their severe clinical presentations. Most AQP4-Ab positive patients (6/8) and 4/10 MOG-Ab positive patients received immunomodulatory or immunosuppressive treatment at the time of sample collection. Although B-cell depleting therapy is known to affect T-cell responses in patients with MS (78), 3/4 AQP4-Ab patients who had increased T-cell reactivity to AQP4 peptide 156-170 were treated with rituximab and the fourth patient was under highdose corticosteroids before the initiation of rituximab treatment.
Importantly, the detailed characterization of single peptides, i.e. known "candidate antigens" based on their encephalitogenicity in animal models and/or their immunodominance in humans is crucial for a potential use in antigen-specific tolerance induction therapies (60,79). For future studies, the implementation of new unbiased approaches may provide additional perspectives. These new strategies differ from previous studies by using combinatorial peptide libraries, which e.g., cover the entire protein or which allow the discovery of novel "unknown" antigens (80). Discrepancies to other studies might be explained by the use of different assays. We and all other investigators in this field face the issue of very low precursor frequencies of CNS antigen-specific T cells in PBMC preparations. However, the CFSE dilution assay used here is a powerful and sensitive method for directly detecting proliferation of rare autoantigen-specific human T-cells (31,81). Moreover, the late addition of IL-2 during a re-stimulation step (82)(83)(84) offers high sensitivity and specificity. This strategy effectively increases the sensitivity for rare antigen-specific Tcells by selectively facilitating the proliferation of T lymphocytes that express the IL-2 receptor alpha-chain CD25 following antigen recognition (82,85). However, even though our culture conditions promote the survival of mostly proliferating T-cells, other cells might skew the cytokine response. Indeed, it is well acknowledged that the key Th17-polarizing cytokine IL-6 is produced by myeloid cells (monocytes) rather than T-cells. Furthermore, the timing of sample collection might influence the relative abundance of the different cytokines and therefore explain differences to previous studies. Finally, additional factors such as the TCR avidity or the peptide concentration in different assays may play a role for various specificities. The application of (i) a lower peptide concentration (here used 20 µg/ml per peptide is higher compared to other studies), (ii) the use of native protein antigens instead of peptides, (iii) purified memory T-cells instead of PBMC, and/or (iv) the co-culture of T-cells with autologous APC, e.g., monocytes or EBV-transformed B cells, may be critical improvements for future experiments.
To conclude, this report investigates AQP4-and MOGspecific T-cell reactivities in human individuals presenting with AQP4-Ab and MOG-Ab positive demyelinating diseases. Our in vitro data corroborates previous findings showing the involvement of AQP4-specific T-cells in AQP4-Ab positive NMOSD and confirms the AQP4 peptide 156-170 as specific T-cell epitope. In contrast, no disease-relevant MOG peptide was identified. Future confirmatory studies using an unbiased approach for epitope discovery in larger cohorts may overcome main limitations of small sample size and the use of a limited collection of synthetic peptides in this study.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/supplementary material.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the local Ethics Committee of Medical University of Innsbruck, Austria (study number AN3041) and University of Zurich, Switzerland (KEK ZH 2013-0001) and all patients and controls gave written informed consent. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AUTHOR CONTRIBUTIONS
LH and MRe performed experiments, analyzed and interpreted data, and drafted the manuscript. MRa performed experiments and participated in the analysis and interpretation of the data. VG and MRe designed the study and MRe and AL analyzed and interpreted data and revisited the article critically for important intellectual content. AP performed experiments and analyzed the data. KR, MS, HH, TB, and AL provided patient material and revisited the article critically for important intellectual content. All authors approved the final version of the manuscript. assays organized by Euroimmun (Luebeck, Germany). HH has participated in meetings sponsored by, received speaker honoraria or travel funding from Bayer, Biogen, Merck, Novartis, Sanofi-Genzyme, Siemens, Teva, and received honoraria for acting as consultant for Biogen and Teva. KR received speaker honoraria from Merck, Novartis and served as a consultant for PARADIGM-Study, Novartis with no compensation. TB has participated in meetings sponsored by and received honoraria (lectures, advisory boards, consultations) from pharmaceutical companies marketing treatments for multiple sclerosis: Almirall, Bayer, Biogen, Biologix, Bionorica, Genzyme, MedDay, Merck, Novartis, Octapharma, Roche, Sanofi/Genzyme, TG Pharmaceuticals, TEVA-ratiopharm and UCB. His institution has received financial support in the last 12 months by unrestricted research grants (Biogen, Merck, Novartis, Roche, Sanofi/Genzyme) and for participation in clinical trials in multiple sclerosis sponsored by Biogen, Merck, Novartis, Roche, Sanofi/Genzyme, and TEVA. AL received financial compensation and/or travel support for lectures and advice from Biogen, Merck, Novartis, Teva, Genzyme, Bayer, Celgene and he is a co-founder of Cellerys and co-inventor on a patent held by the University of Zurich on the use of peptide-coupled cells for treatment of MS.
The remaining 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.
Copyright © 2020 Hofer, Ramberger, Gredler, Pescoller, Rostásy, Sospedra, Hegen, Berger, Lutterotti and Reindl. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.