The Beneficial Clinical Effects of Teriflunomide in Experimental Autoimmune Myasthenia Gravis and the Investigation of the Possible Immunological Mechanisms

Myasthenia gravis (MG) is an autoantibody-mediated autoimmune disease characterized by skeletal muscle weakness exacerbated with exercise. There is a need for novel drugs effective in refractory MG. We aimed to test the potential of teriflunomide, an immunomodulatory drug currently used in rheumatoid arthritis and multiple sclerosis treatment, in a murine experimental autoimmune myasthenia gravis (EAMG) model. EAMG was induced by immunizations with recombinant acetylcholine receptor (AChR). Teriflunomide treatment (10 mg/kg/day, intraperitoneal) was initiated to one group of mice (n = 21) following the third immunization and continued for 5 weeks. The disease control group (n = 19) did not receive medication. Naïve mice (n = 10) received only mock immunization. In addition to the clinical scorings, the numbers of B cells and T cells, and cytokine profiles of T cells were examined by flow cytometry. Anti-AChR-specific antibodies in the peripheral blood serum were quantified by ELISA. Teriflunomide significantly reduced clinical disease scores and the absolute numbers of CD4+ T cells and some of their cytokine-producing subgroups (IFN-γ, IL 2, IL22, IL-17A, GM-CSF) in the spleen and the lymph nodes. The thymic CD4+ T cells were also significantly reduced. Teriflunomide mostly spared CD8+ T cells’ numbers and cytokine production, while reducing CD138+CD19+lambda+ plasma B cells’ absolute numbers and CD138 mean fluorescent intensities, probably decreasing the number of IgG secreting more mature plasma cells. It also led to some selective changes in the measurements of anti-AChR-specific antibodies in the serum. Our results showed that teriflunomide may be beneficial in the treatment of MG in humans.


Introduction
Myasthenia gravis (MG) is an antibody-mediated autoimmune disease characterized by skeletal muscle weakness worsened by exercise. Although a rare disease, MG impairs the daily activities of patients, and if untreated, in up to 20% percent of the cases, leads to myasthenic crisis which is characterized by exacerbated muscle weakness, respiratory failure eventually requiring intubation and mechanical ventilation (Fichtner et al. 2020;Howard 2018;Wang and Yan 2017).
MG results from antibody-mediated destruction of the postsynaptic membrane at the neuromuscular junctions (NMJ) and has several subtypes depending on the antigen specificity of the autoantibodies (Fichtner et al. 2020). In about 85% of the patients, nicotinic acetylcholine receptor (n-AChR)-specific autoantibodies are found (Vincent 2002). Muscle-specific kinase (MuSK) (6%) or lipoprotein receptor-related protein (LPR4) (4%)-specific autoantibodies have also been shown in the patients (Higuchi et al. 2011;Hoch et al. 2001;Zisimopoulou et al. 2014). About 5% of the patients are classified as seronegative MG, and no known autoantibody or antigen has been defined in those patients to this date (Carr et al. 2010;Higuchi et al. 2011;Hoch et al. 2001;Zisimopoulou et al. 2014). The majority of the AChR autoantibodies are produced by long-lived (mature) plasma cells and are of IgG1 isotype, thus can mediate complement-mediated damage of NMJs. MuSK autoantibodies, however, are mainly of IgG4 isotype and produced by short-lived plasmablasts (McConville et al. 2004;Stathopoulos et al. 2017).
The immunopathogenesis of MG involves innate and adaptive immune cells (Wang and Yan 2017). Initial antigen presentation to T cells may occur peripherally, or in the thymus due to various stimuli including infections. Elevated or activated CD4+ T helper cells subsets, including Th1, Th17 and Tfh, have been reported in MG patients suggesting their involvement at various stages of disease pathogenesis (Wang and Yan 2017). The thymus appears to be particularly involved in MG pathogenesis not just as an organ for T cell generation but acting as a site for the expression of AChR alpha subunit and being hyperplastic in 70% of AChR MG patients and harboring autoreactive B cells and T cells (Filosso et al. 2013).
Teriflunomide is an immunomodulatory drug approved by FDA and EC for use in relapsing-remitting multiple sclerosis (RRMS) in 2012 and 2013, respectively. Leflunomide, which is actively used in rheumatoid arthritis treatment since 2008, is structurally similar and is converted (70%) to teriflunomide in vivo. Thus, in vivo activity of leflunomide is thought to be largely that of teriflunomide. Teriflunomide inhibits pyrimidine de novo synthesis by reversibly inhibiting the mitochondrial enzyme dehydroorotate dehydrogenase (DHO-DH) at concentrations below 100 µM. Therefore, teriflunomide blocks DNA synthesis, and in turn, the proliferation of rapidly dividing cells such as B and T cells by reducing cellular pyrimidine pool. At higher concentrations, teriflunomide has been shown to inhibit purine synthesis as well, thereby inhibiting ATP-dependent cellular processes. Studies in mice have shown that leflunomide was effective in reducing T-dependent and independent antibody responses. A more recent study showed that teriflunomide is beneficial in the MuSK-induced EAMG model (Yilmaz et al. 2021).
Currently, long-term treatment of MG is based on a combined therapy which includes symptomatic treatment with acetylcholine esterase inhibitors and immunosuppressants like corticosteroids, azathioprine mycophenolate mofetil, as well as rituximab which has been widely used more recently (Silvestri and Wolfe 2014;Sonkar et al. 2017). Additionally, in trials, C5 inhibitor eculizumab improved generalized MG (gMG) with AChR autoantibodies, or refractory gMG, and it is currently an emerging therapy in MG. However, the duration of its use as a treatment, its long-term tolerability, its efficiency and cost of effectiveness are not clear yet. (Dhillon 2018). Despite various treatment modalities, some patients experience recurrent myasthenic crises with worsened muscle weakness. Additionally, long-term use of currently available immunosuppressant drugs has important side effects, like hypertension, diabetes mellitus, cataract and increased risk of certain type of cancers. Thus, there is a need for novel treatment options, with fewer side effects, suitable for long-term use and effective in refractory cases. In this regard, teriflunomide may be an appropriate drug with its well-characterized safety profile, with similar incidences of adverse events (AEs), serious AEs (SAEs) and AEs leading to drug discontinuation as with placebo, reported by placebo-controlled phase 3 clinical trials. The most frequent AEs with teriflunomide (incidence ≥ 10% and ≥ 2% greater than placebo) were alanine aminotransferase increases, alopecia (hair thinning), diarrhea, influenza, nausea and paresthesia (Confavreux et al. 2014;O'Connor et al. 2011). In addition to its high benefit/risk profile and high tolerability, it was shown not to cause clinical immune suppression in multiple sclerosis patients (Bar-Or et al. 2014). For these reasons, teriflunomide may be a better option in the treatment of MG with respect to the current immunosuppressants in the clinic.
In the current study, we aimed to test the therapeutic potential of teriflunomide in a murine model of AChR-autoimmune MG (EAMG), representing the autoimmune type in 85% of human patients. Teriflunomide was given to a group of EAMG mice after the disease induction by several AChR immunizations, and afterward, the disease clinical scores, the B and T cell immune responses, and the autoantibody 1 3 productions were monitored over time. Our results revealed that teriflunomide may alleviate EAMG disease severity by reducing CD4+ T cells and by changing subsequent B cell responses in a preclinical MG model.

Mice
Specific pathogen-free C57BL6 female mice were purchased from Kobay Ltd. (Ankara, Turkey) and housed at Erciyes University, Experimental Research and Application Center (DEKAM) vivarium, 5 mice per cage. The mice were fed regular chow diet. The experiments were performed on 8-week-old mice after 14 days of quarantine. The study was approved by Erciyes University animal studies institutional review board (#19/139). All the experiments were conducted according to relevant guidelines and regulations. The mice were randomly allocated into three groups: (i) Experimental autoimmune myasthenia group which would take teriflunomide (EAMG+TF), (ii) Experimental autoimmune myasthenia group (EAMG) and the (iii) Healthy control group.
Based on our clinical experience and experimental hypothesis, we considered to reach to 60% incidence at week 7 for EAMG and EAMG+TF experimental groups. Considering this hypothesis, we applied a power analysis by using type-I error of 5%, power of 80% and effect size of 0.480 (Chi-square analysis with 2 degrees of freedom). This analysis resulted to include at least 42 experimental animals to our research. Considering our research facilities, we planned to include 50 mice (10 for control group, 19 for EAMG group and 21 for EAMG+TF group) to the study. These analyses were conducted using PASS 11.0 (Power Analysis Statistical System, NCSS Inc., US) statistical software.
To test the therapeutic potential of teriflunomide on EAMG, the disease was induced by three consecutive recombinant Torpedo AChR protein immunizations on the days 0, 21 and 48. Following the third AChR immunization, teriflunomide 10 mg/kg was given intraperitoneally daily for 33 days until the mice were killed. The EAMG group received PBS.  F5881) and peptide diluted in phosphate buffered saline (PBS) (Biological Industries, cat#02-023-1A) (0.3 mg/mL) were emulsified using a stopcock and 5 cc syringes. Mice were immunized at 4 different spots with (50 µL/injection, and a total of 200 µL per immunization) on the lower flanks, both left and right, for the 1st immunization; at higher flanks for the 2nd immunization; and at the inguinal and lower flanks for the 3rd immunization. The mice were immunized on days 0, 21 and 48 and killed on day 84 (12th week). Twenty-one of AChR-immunized mice (EAMG+TF group) received teriflunomide (TF) (AUBAGIO®) starting with the day 51, when 60% of the modeled mice (n:24) developed clinical symptoms of the disease. Teriflunomide was given intraperitoneally following the third AChR immunization until the mice were killed, at 10 mg/mL, daily dose for 33 days. The EAMG group (AChR-immunized) received PBS instead of teriflunomide. The mice in the healthy control group (n = 11) were only immunized with CFA-PBS emulsion (without AChR). The experimental groups and the study design are shown in Fig. 1.

The Clinical Tests and the Scoring
The clinical evaluations of the mice were carried out in a double-blind fashion regularly with the start of the experiment in DEKAM. The weight and general condition of the mice were evaluated daily. Strength and fatiguability evaluations based on observation and examination (Ulusoy et al. 2017) were accomplished every 3 days until the second dose of AChR immunization and daily from the second dose immunization until the end of the experiment.
In addition to these assessments, the post-exercise inverted grid test (Yilmaz et al. 2021) was performed to monitor fatiguability. The clinical grade scorings of the mice were noted on a weekly base: Grade 0, normal muscle strength and no muscle weakness, even after exercise; grade 1, normal at rest but weak after exercise, with chin on the floor and inability to raise head, hunched back and reduced mobility; grade 2, grade 1 weakness at rest; grade 3, severe weakness, dehydrated and paralyzed (quadriplegic), loss of significant weight; and grade 4, found dead in the cage (Tuzun et al. 2015). These scores were treated as the status of the mice at the end of the week, with the help of the close evaluations during the week.
The clinical assessments of the animal groups were parameterized as disease incidence, disease clinical score and the ratio of actual body weights of animals to their weights at baseline, and comparisons of the groups were based on these parameters. In disease incidence assessments, mice were required to have at least a grade 1 clinical score to be considered as myasthenic.

The Enzyme Linked Immunosorbent Assay (ELISA)
Fifty µl of blood was collected from the facial veins of mice on day -1, week 7 and 12. The serum was taken after centrifugation at 6000×g for 10 min and frozen at − 80. The samples were thawed and diluted 1000 times and used for ELISA. The protocol by Yilmaz et al. was followed (Yilmaz et al. 2021). Briefly, AChR protein was diluted in coating buffer (BioLegend, cat#421701) and the 96-well ELISA plate was incubated overnight at 4 °C. The excess protein was washed with Wash Buffer [PBS 0.05% Tween (Sigma, cat#9005-64-5)]. The plates were blocked with 5% FBS (HyClone, cat#SH30071.03IR25-40) in PBS for 2 h and afterward washed 3 times with Wash Buffer. The serum was added and incubated for 2 h at room temperature. After 3 washes with Wash buffer anti-IgG1 (Southern Biotech, cat#1071-05), IgG2b (Southern Biotech, cat#1091-05), IgG3 (Southern Biotech, cat#1101-05) and IgM (Southern Biotech, cat#1021-05) were added with 5000-fold dilution. The ELISA experiments were performed two times.
The flow cytometry antibodies were purchased as 25 µg in size. Due to the limitation of our materials, we had to carry out the examinations in the thymus with different numbers of biological replicates (7, 7 and 10 mice in the control, EAMG and EAMG+TF group, respectively). For the same reason, we excluded two mice (randomly selected) of EAMG+TF group from the flow cytometry experiments. Thus, equal number of mice (n = 19) was enrolled into the EAMG and EAMG+TF groups. As a control group, all samples of 10 mice were investigated.

Statistical Analyses
Analyses were performed using IBM SPSS 25.0 statistical software. GraphPad Prism 6 program was used for the generation of the graphs. P values less than 5% was considered as statistically significant and was indicated with a single asterisk (*), while p values less than 0.01, less than 0.005 and less than 0.001 were marked with two (**), three (***) and four asterisks (****), respectively. In the analyses, no data were excluded from the study.
Shapiro-Wilk's test was applied to assess the data normality, while Levene test was used to test variance homogeneity. Mauchly's test was used to assess the sphericity assumption. A logarithmic transformation (base 10) was applied for highly skewed data.
To compare the differences between experimental groups, a two-sided independent samples t test and one-way analysis of variance (ANOVA) were used for continuous data which follows normal distribution. Alternatively, Welch t test and Welch ANOVA were performed for these data, when variance homogeneity assumption was not satisfied. Tukey and Tamhane T 2 tests were used for post-hoc comparisons. Mann-Whitney U test and Kruskal-Wallis H tests were performed for continuous data which does not follow normal distribution. Siegel-Castellan test was used for pairwise comparisons. Pearson χ 2 and Fisher-Freeman-Halton tests were applied for categorical data. Bonferroniadjusted z test was used for multiple comparisons. The 7th and 12th weeks' serum immunoglobulin levels of each individual mouse in the study groups with disease were compared with each other by using paired t test. This test was repeated for each myasthenic experimental group (EAMG+TF and EAMG).

Teriflunomide Ameliorates the Disease Scores and the Incidence in EAMG Model Induced by Recombinant Torpedo AChR Protein Immunization
Naïve unimmunized (control), immunized (EAMG) and teriflunomide-receiving immunized (EAMG+TF) groups were compared in detail with respect to weight gain, incidence and EAMG clinical scores (Figs. 2, 3, Supplementary Tables 1 to 4).
The body weight percentage was decreased in the EAMG and EAMG mice groups compared with healthy controls at 8th and 9th weeks after the first immunization (Fig. 2a,  Supplementary Tables 1, 2). At the 11th week, teriflunomide received group had a reduced median body weight percentage compared to other groups. At 12th week, there was no statistical difference among the groups with respect to the weights. Teriflunomide treatment reduced disease incidence at 12th week significantly (Fig. 2b, Supplementary Table 3). It also decreased the ratio of grade 2 mice starting from 10th week (statistically significant at 10th week and very close to statistically significant level at 12th week). Likewise, first Fig. 2 a The percentages of body weight measurements to initial body weights in all study groups and the comparison of the groups. b The incidence of the disease in all study groups and the comparison of the groups. (a) ***At the 8th week, the control group is different from the EAMG and EAMG+TF groups with regard to the percentage of body weight measurement (p < 0.005). • At the 9th week, the control group is different from the EAMG and EAMG+TF groups with regard to the percentage of body weight measurement (p < 0.05). ♦ At the 11th week, EAMG+TF group is different from the control and EAMG groups with regard to the percentage of body weight measurement (p < 0.05). (b) The disease incidence is zero between 0 and 3 weeks for all study groups. For the control group, the disease incidence is zero throughout the study period. *At the 12th week, the disease incidences are all different from each other. The lowest incidence is in the control group, the second lowest one is in the EAMG+TF group and the highest one is in the EAMG group. The disease incidence in EAMG group is different from that of EAMG+TF group at the significance level of p < 0.05 1 3 appearances of grade 2 mice were at 8th and 11th weeks in EAMG and EAMG+TF groups, respectively (Fig. 3, Supplementary Table 4). Thus, teriflunomide treatment appears to improve clinical disease states in AChR-specific EAMG.

Teriflunomide Inhibits CD4+ T Cell Responses in the Lymph Node, Spleen and the Thymus
Since teriflunomide blocks DNA synthesis in rapidly proliferating T cells (Bar-Or et al. 2014;Loffler et al. 2004;Ruckemann et al. 1998), we explored its impact on T cell numbers and cytokine production in the spleen, draining lymph nodes (inguinal) and the thymus of the three mice groups. The spleen, thymus and the lymph nodes were collected at the end of the 12th week of EAMG induction. The absolute number of splenic CD3+CD4+ T cells were comparable between the naïve unimmunized (control), immunized (EAMG) or immunized and teriflunomidereceiving (EAMG+TF) groups (Fig. 4, Supplementary Tables 5, 6). However, splenic IL-2+, IL-17A+, IFN-γ+, IL-22+, IL-10+CD3+CD4+ T cells were significantly reduced in the EAMG+TF group compared with EAMG group (Fig. 4, Supplementary Tables 5, 6). The percentages of the cytokine-producing cells among splenic CD3+CD4+ T cells were significant only for IFN-γ+ cells but not others (Fig. 5, Supplementary Tables 7, 8). However, the mean fluorescence intensity (MFI) of IFN-γ was comparable between EAMG+TF and EAMG groups (Fig. 5,Supplementary Tables 7,8). These results suggested that the reduction in the cytokine-producing splenic CD4+ T cells was due to reduced numbers of cells rather than a decrease in gene expression or protein synthesis.
In contrast to the spleen, the draining lymph node had significantly fewer CD3+ CD4+ helper T cells in the EAMG+TF mice compared with the EAMG group (Fig. 6, Supplementary Tables 9, 10). Additionally, the absolute number of IFN-γ+, GM-CSF+, IL-22+, IL-10+CD3+CD4+ T cells were significantly reduced in the lymph nodes of the EAMG+TF group compared with the EAMG mice group (Fig. 6, Supplementary Tables 9, 10). The percentage of IFN-γ+CD4+ T cells was also reduced significantly (Fig. 7,Supplementary Tables 11,12). However, similar to spleen, MFI values of IFN-γ were comparable between EAMG+TF and EAMG groups (Fig. 7,Supplementary Tables 11,12), suggesting that the reduction in the cytokine-producing splenic CD4+ T cells were due to reduced numbers of cells rather than a decrease in gene expression or protein synthesis. The rates of grade 2 mice in the groups. (a) *, **, ****The rate of grade 1 mice in the control group is statistically different from those in the EAMG and EAMG+TF groups, starting from the 6th week. (*p < 0.05; **p < 0.01; ****p < 0.001). (b) *The rate of grade 2 mice in the EAMG group is different from that in the EAMG+TF group at the 10th week in the comparison of these two groups. Thymic hyperplasia is a critical component of MG pathogenesis in 70% of MG patients with AChR autoantibodies. The hyperplastic thymus has autoreactive B and T cells. Although thymic involvement in EAMG mice was considered to be minimal (Mantegazza et al. 2016), we also assessed the impact of teriflunomide on thymic T cell output in the AChR-immunized mice. Our data revealed a significant increase in the absolute number of thymic CD3+CD4+ T cells in EAMG mice compared with the naïve control group (Fig. 8, Supplementary  Tables 13, 14). Teriflunomide treatment after disease onset significantly reduced the absolute number of CD3+CD4+ T cells. Although the reduction in CD8+ T cells was not statistically significant, CD40L expressing CD8+ T cell (activated CD8+ T cell) numbers were significantly reduced after teriflunomide treatment (Fig. 8, Supplementary Tables 13, 14). These results collectively suggest that in the EAMG context induced by AChR immunizations, teriflunomide greatly reduces both, CD4+ T cell expansion and CD8+ T cell activation.

Teriflunomide Reduces the CD138 Expression and the Lambda Positive Plasma Cells
To further assess the impact of teriflunomide on B cell or plasma cell numbers, spleen and lymph nodes of naïve unimmunized (healthy control), immunized (EAMG) or immunized and teriflunomide-receiving (EAMG+TF) groups were stained for CD19, CD138, IgG kappa or IgG lambda. The gating strategies of these B lymphocytes in the spleen and lymph nodes are shown in Supplementary  Fig. 15. Total B cells were gated as CD19+ cells (Supplementary Fig. 15) and were unaffected with teriflunomide treatment as shown by comparable CD19+ B cell numbers both in the spleen and the lymph nodes (Fig. 9,   Supplementary Tables 15, 16). Similarly, plasma B cell (marked as CD19 +/low CD138+) absolute numbers were comparable across all the groups (Fig. 9, Supplementary Tables 15, 16). However, the absolute number of IgG lambda+ plasma cells in the spleen was significantly reduced in the EAMG+TF group as compared to the EAMG group (Fig. 9, Supplementary Tables 15,  16). Furthermore, CD138 expression by CD19 +/low cells reflected by MFI CD138 was significantly reduced in the EAMG+TF group in comparison to EAMG group (Fig. 9, Supplementary Tables 15, 16). These results suggest that teriflunomide treatment after the onset of EAMG may selectively reduce IgG lambda+ plasma cells and may interfere with the functions of plasma cells by

Teriflunomide Changes the Amount of Serum Anti-AChR Antibodies in a Selective Way
Because both human MG and EAMG are antibody-mediated diseases, we directly measured the impact of teriflunomide on levels of anti-AChR antibodies (Fig. 10, Supplementary  Tables 17, 18). When basal, after the disease induction preand post-treatment (at the 7th and 12th week: on the days of 49 and 83) levels of anti-AChR immunoglobulins (IgG1, IgG2b, IgG3, IgM) were evaluated, all four AChR-specific immunoglobulin levels were elevated at 7th week in the disease groups and were further augmented following the last immunization by 12th week. The teriflunomide injections were initiated on the day 51st post-immunization and continued until mice were killed. At the end of the treatments, at the 12th week, there was no significant difference between the EAMG and EAMG+TF groups regarding AChR-specific IgG1, IgG2 or IgG3 (Fig. 10, Supplementary Tables 17, 18). However, the AChR-specific IgM levels were significantly higher by the 12th week in the EAMG+TF mice compared with the EAMG group (p = 0.001) (Fig. 10, Supplementary  Tables 17, 18). When immunoglobulin levels of each mouse at 7th (in the pre-treatment period) and 12th week (in the post-treatment period) were together considered within each disease model groups by using paired t tests, all immunoglobulins seemed to increase at the significance level of p < 0.001 in both EAMG+TF and EAMG groups after treatment. Only IgG1 increase in the EAMG+TF mice group showed a lower significance level (p < 0.05) compared to other significance levels (p < 0.001) probably reflecting a subtle reduction in the IgG1 levels (Fig. 11, Supplementary Tables 19, 20). These data collectively support that the EAMG model in the study has been successfully implemented and that teriflunomide administration after disease onset may affect IgG antibody isotypes in a selective way.

Discussion
In the current study, the therapeutic effects of the cytostatic and immunomodulatory drug teriflunomide were evaluated in a mouse model of MG induced by AChR immunizations. Our results revealed that teriflunomide treatment after disease onset decreased the incidence and clinical disease scores of murine EAMG. Teriflunomide improved EAMG symptoms by reducing the absolute CD4+ T cell have recently tested therapeutic administration of teriflunomide in MuSK-mediated EAMG and showed clinical benefit similar to the results presented herein (Yilmaz et al. 2021). AChR-specific autoantibody-mediated pathology account for 85% of MG cases in humans, while MuSK autoantibody-mediated MG accounts for 6% of MG patient population (Fichtner et al. 2020;Wang and Yan 2017). Therefore, the results of the current report have key implications for MG patients supporting the use of teriflunomide as a therapeutic treatment. In our hands, the clinical EAMG disease was first observed in some mice about the fourth week (10 days after the second immunization). Disease incidence and mean disease score reached a statistically significant level compared to the healthy control group at the sixth week after the first immunization (three weeks after the second AChR injection), which is consistent with the literature (Mantegazza et al. 2016;Shigemoto et al. 2015;Yilmaz et al. 2021). In some studies, the standards of mimicking MG treatment after disease onset are defined as the time when ≥ 60% of mice develop clinical symptoms of the disease (Tuzun et al. 2015). In our experiment, this criterion was fulfilled 3 days after the third immunization and the treatment was started at this time. Teriflunomide treatment after disease onset reduced the anti-AChR antibody-mediated EAMG incidence at 12th week, and the rate of grade 2 mice starting at 10th week till the end of the experiment compared with untreated EAMG mice. The data indicate that teriflunomide's clinical effects are detectable after three weeks of use. In Yilmaz et al.'s MuSK-mediated EAMG model, the beneficial effect of teriflunomide was also detectable starting by the 11th week of disease induction; thus, these independent studies corroborate each other's findings. Although EAMG dependent weight loss was ameliorated by teriflunomide treatment in the Yilmaz et al. study (Yilmaz et al. 2021). Our EAMG+TF group was not statistically different from EAMG group with regard to the body weight with an exception only at 11th week. The difference between the two studies can be explained by the difference in the doses of the drug, which is higher in the current study. Our findings are in accordance with the potential of teriflunomide to cause weight loss in higher doses (Committee for Medicinal Products for Human Use-CHMP 2013).
The primary mechanism of action of teriflunomide is through inhibition of mitochondrial enzyme DHO-DH. This enzyme is highly expressed in activated lymphocytes, both T and B cells which rely on de novo pyrimidine synthesis to meet their increased nucleotide demands. Because it is not a nucleotide analog, and the resting lymphocytes can use salvage pathways to meet their pyrimidine needs, teriflunomide appears to affect mostly activated lymphocytes (Bar-Or et al. 2014). In vitro proliferation assays revealed that  (Li et al. 2013). Others have shown that, in vivo, teriflunomide was particularly effective in inhibiting the proliferation of T cells with high-affinity T cell receptors for the antigens (Posevitz et al. 2012).
In our experiments regarding the T cell compartment, the most notable changes were observed in CD4+ helper T cells with teriflunomide treatment. In the Yilmaz et al. study, flow cytometric analyses of T cells (CD4+ or CD8+), B lymphocytes as well as NK cells revealed no significant difference between teriflunomide treated and untreated MuSK EAMG groups with respect to percentages of cells. However, the changes in the absolute numbers of those cells have not been documented in that study (Yilmaz et al. 2021). Our experiments also did not reveal any reduction in the percentages/frequency of CD4+ T cells or their IL-2+, IL-17+, IL-22+, GM-CSF+ subsets between teriflunomide treated and untreated AChR-EAMG groups. However, the absolute number of IFN-γ+, IL-17+, IL-2+, IL-22+, GM-CSF+ CD4+ T subsets as well as total CD4+ T cells were reduced significantly in the secondary lymphoid organs after teriflunomide treatment. The antiproliferative effect of teriflunomide on CD4+ T cells was also evident in the thymus. Thus, helper T cells responses, particularly Th1, Th17 which were shown to play critical roles in the pathogenesis of EAMG, appeared to be inhibited by teriflunomide. This finding has been supported by the existing literature (Uzawa et al. 2021;Wang and Yan 2017). When IFN-γ MFI values were examined, IFN-γ expression per cell was unaffected. These findings argue that teriflunomide, rather than inhibiting transcription/or translation events of these cytokine genes, acts by inhibiting proliferation of those T helper cell subsets. These data are in line with the human studies showing leukopenia/lymphopenia in MS patients who were treated with teriflunomide and the well-established negative impact of teriflunomide on highly proliferating T and B cells (Bar-Or et al. 2014;Confavreux et al. 2014;O'Connor et al. 2011). It is important to note that T and B cells' frequency/percentage measurements do not always reflect the changes in the absolute number of cells, and discrepancies between some reports most likely result from the lack of absolute number calculations. Fig. 11 The comparison of IgG1, IgG2b, IgG3 and IgM type anti-AChR antikor levels within the diseased animal groups (EAMG and EAMG+TF) In our experiments, teriflunomide did not reduce total CD8+ T cell numbers in vivo in EAMG mice, be it in lymph nodes, spleen or thymus. However, we observed a reduction in CD40L+ CD8+ T cell numbers (activated CD8+ T cells) in the thymus. Both CD4+ and CD8+ T cells are shown to be involved in the pathogenesis of EAMG in rats and mice, unlike human MG, which occurs predominantly through CD4+ T cell and B cell-mediated pathology (Wang and Yan 2017;Zhang et al. 1996). Accordingly, genetic or antibody-mediated depletion of CD8+ or CD4+ T cells suppressed the disease (Zhang et al. 1996). It is unclear why teriflunomide has a more robust impact on preferentially CD4+ T cell compartment in the murine EAMG model. This could be related to differential reliance of CD4+ and CD8+ T cells on DHO-DH and requires further study. Although a recent study suggested that 12-month use of teriflunomide in relapsing-remitting MS patients selectively reduced CD8+ memory T cells, absolute numbers of cells have not been examined in that study (Tilly et al. 2021). Another report performed on seven multiple sclerosis (MS) patients by Gandoglia et al. showed a trend toward reduction in helper T cells after teriflunomide use (Gandoglia et al. 2017).
Both leflunomide and teriflunomide have been shown to inhibit B cell proliferation through inhibition of DHO-DH and other targets such as cyclin D3 and cyclin A expression (Ringshausen et al. 2008). In the peripheral blood of teriflunomide-receiving MS patients, absolute numbers of mature, regulatory or CD19+ total B cells were significantly reduced (Gandoglia et al. 2017). The only report investigating B cells in the EAMG mice model was Yilmaz et al.'s anti-MuSK-mediated EAMG study. In that study, B cell percentages were found not to be significantly altered, while the absolute B cell counts were not examined (Yilmaz et al. 2021). Our data revealed that total CD19+ B cells as well as plasma cells (CD19 +/low CD138) absolute numbers were not significantly altered by teriflunomide treatment. On the other hand, Lambda+ plasma cell absolute numbers were significantly reduced in the spleen of EAMG+TF mice compared with the untreated EAMG group. Additionally, CD138 expression (mean fluorescence intensity showing mean antigen expression), which is high in long-lived mature plasma cells responsible for the production of IgG, was significantly reduced in the lymph nodes and the spleen suggesting that teriflunomide may result in significant changes in the plasma cell functions affecting antibody production in a selective way (Bortnick and Allman 2013; Nutt et al. 2015).
In our experiments, teriflunomide treatment appears to increase IgM isotype, while it also causes a subtle decrease in IgG1 level. Although IgG1 levels were still comparable between EAMG and EAMG+TF mice groups in the 12-week post-EAMG induction, the increase in IgG1 level from the basal time to the 12th week of disease induction was more significant in the EAMG group compared to the EAMG+TF group. This subtle drop in IgG1 levels in the EAMG+TF mice group is possibly continuing. Indeed, in the anti-MuSK-mediated EAMG model of Yilmaz et al., serum IgG1 levels were significantly reduced after a longer treatment with teriflunomide at the 14th week of disease induction (Yilmaz et al. 2021). Additionally reduced IgG deposition at the neuromuscular junctions was reported in that study, supporting our data. Yılmaz et al. did not measure serum IgM levels in their study (Yilmaz et al. 2021).
When both studies of teriflunomide in EAMG are evaluated together, it can be reported that this drug may cause some selective changes in the antibody responses. Upon considering the basic mechanism of action of teriflunomide that is inhibiting the rapidly progressing T and B cells, to explain these selective changes in the antibody responses seems to be difficult. Through its basic mechanism of inhibitory action on DHO-DH, teriflunomide has been proposed to have direct suppressive effect on B cells, since it reduced proliferation and lipopolysaccharide stimulated Ig secretion of B cells (Siemasko et al. 1996). Siemasko et al. in their in vivo study found that leflunomide decreases both IgG and IgM secretion, in a manner that is mostly decreasing IgG levels (Siemasko et al. 1998). They also discovered that leflunomide has some DHO-DH independent functions including tyrosine kinase inhibitory activity causing an indirect effect on B cells, in such a way that it decreases IL-4 driven class switch recombination into IgG1, causing a reduction in IgG1 production (Claussen and Korn 2012;Siemasko et al. 1998). Our results, with the increase in IgM and subtle decrease in IgG1 levels, are in accordance with the inhibition of this class switch. Similarly, in a clinical study comparing the serum immunoglobulin levels during teriflunomide and ofatumumab (human anti-CD-20 monoclonal antibody) treatment, the proportion of patients with IgG levels below the lower limit of the normal and the proportion of patients with IgM levels below the lower limit of the normal were found to be 22.9% and 6.6% in teriflunomide using patients, and 14.2% and 17.7% in ofatumumab using patients, respectively. It is interesting to see that the proportion of IgG decrease with teriflunomide use was higher than the proportion of IgM decrease below the normal lower limits. The proportion of decrease in Ig G level with teriflunomide is also higher than that of ofatumumab, which is a drug totally active on B lymphocytes (Wiendl et al. 2020).
In our study, an absolute clinical improvement has been observed with teriflunomide use and several possible mechanisms underlying this improvement have been elucidated, as all data is available and open to the readers. In addition to the subtle decrease in IgG1 antibody, reduction in T helper cell subgroups producing inflammatory cytokines may have taken a significant role in this clinical improvement. By reducing Th1, Th17 T helper cell subsets, teriflunomide seems to have lowered inflammatory cytokine levels, causing less inflammation, leading to a decrease in antigen presentation particularly by reduction in IFN-γ and to less recruitment of monocytes, macrophages and neutrophils responsible for the destruction of AChR via antibodydependent cellular cytotoxicity (Amaldi et al. 1989;Wang and Yan 2017;Uzawa et al. 2021). Meanwhile, the decrease in the number of plasma cells, the reduction of the mean CD-138+ expression (CD138 mean fluorescent intensity) meaning reduced activity of long-lived mature plasma cells, as well as the lower numbers of helper T cells by decreasing the helping effect on B cell affinity maturation, and the inhibition of the class switch to IgG1 antibody possibly through the inhibitory action on tyrosine kinase may have all contributed to the subtle decrease in IgG1 antibody in serum and possibly at neuromuscular junctions. (Bortnick and Allman 2013; Khodadadi et al. 2019;Nutt et al. 2015). IgG1 is the major immunoglobulin in the pathogenesis of AChR autoantibody-mediated MG (Lefvert et al. 1981;Rodgaard et al. 1987) and may exert its effects via complement activation, by blocking AChR signaling, or by inducing internalization of the receptor from the cell membrane (Ey et al. 1979). IgM to IgG1 isotype switch is induced by IL-4 and IL-21 cytokines (Moens and Tangye 2014), and whether their levels are altered after teriflunomide treatment in vivo requires further study.
In this study, it is exceptionally interesting to see that how a drug shows selectivity in its actions on cell types that is to say affecting helper T cells (even especially IFN-γ secreting helper T cells) more than cytotoxic T cells, affecting mostly more mature plasma cells out of all B cells and affecting some types of immunoglobulins more. These findings are most probably due to some regulations and interactions of the immune mechanisms in relation to the effects and the dose of the drug. Clausen et al. in their review explained that most of the kinase inhibitory actions of teriflunomide were observed in vitro in higher concentrations (at least one order of magnitude higher) than those used to block DHO-DH; they also claimed that it is difficult to understand whether these effects would be reliable in vivo (Claussen and Korn 2012). Indeed, our study may also show the higher concentration effects of the teriflunomide due to its use in a high dose. Likewise the differences of our findings from those of Yilmaz et al. may be explained by the differences in the doses of the drug, which was about 200-250 μg/day for each mouse in the current study and 30 μg/day in the Yilmaz et al.'s study (Yilmaz et al. 2021).
This study has a wide range of investigations with T cells, B cells and cytokines in the spleen, lymph node and the thymus in a detailed way and with serum anti-AChR immunoglobulins in the murine AChR-induced EAMG model. Nevertheless, due to some practical limitations, studies on follicular or memory T and B cells, and more elaborate examinations in the thymus could not be performed. Additionally, the prolonging of the treatment period would have yielded more pronounced results regarding the immunoglobulins. As another limitation, the specificities of the primary antibodies could not be determined by knockout validation studies. However, these antibodies have been used under similar conditions before (Supplementary Table 21). Finally, the stimulation of T cells was preferred to be performed with PMA/Ionomycin rather than more specific AChR peptide stimulation, due to its capability of giving stronger signals as a valuable method.
In conclusion, our study shows that teriflunomide has clinical benefits and prevented the progression of the disease in a murine model of MG through different possible mechanisms including suppression of immune responses by reducing the number of cytokine-producing T cells, by changing the functions of plasma cells and by leading selective changes in anti-AChR antibody quantities and types. These mechanisms need to be verified with some other studies. Additionally, extending the follow-up period in future studies and performing the experiments with different doses of teriflunomide could definitely be more informative. The data presented herein suggest that teriflunomide may be a suitable candidate for use in MG patients and even in chronic inflammatory neuromuscular diseases owing to its widespread effect on the immune system and its low side effects. Further studies, including human trials, investigating its benefit/risk ratio in MG patients are warranted.