Dimethyl Fumarate Selectively Reduces Memory T Cells and Shifts the Balance between Th1/Th17 and Th2 in Multiple Sclerosis Patients

Abstract

Dimethyl fumarate (DMF; trade name Tecfidera) is an oral formulation of the fumaric acid ester that is Food and Drug Administration approved for treatment of relapsing-remitting multiple sclerosis. To better understand the therapeutic effects of Tecfidera and its rare side effect of progressive multifocal leukoencephalopathy, we conducted cross-sectional and longitudinal studies by immunophenotyping cells from peripheral blood (particularly T lymphocytes) derived from untreated and 4-6 and 18-26 mo Tecfidera-treated stable relapsing-remitting multiple sclerosis patients using multiparametric flow cytometry. The absolute numbers of CD4 and CD8 T cells were significantly decreased and the CD4/CD8 ratio was increased with DMF treatment. The proportions of both effector memory T cells and central memory T cells were reduced, whereas naive T cells increased in treated patients. T cell activation was reduced with DMF treatment, especially among effector memory T cells and effector memory RA T cells. Th subsets Th1 (CXCR3⁺), Th17 (CCR6⁺), and particularly those expressing both CXCR3 and CD161 were reduced most significantly, whereas the anti-inflammatory Th2 subset (CCR3⁺) was increased after DMF treatment. A corresponding increase in IL-4 and decrease in IFN-γ and IL-17-expressing CD4⁺ T cells were observed in DMF-treated patients. DMF in vitro treatment also led to increased T cell apoptosis and decreased activation, proliferation, reactive oxygen species, and CCR7 expression. Our results suggest that DMF acts on specific memory and effector T cell subsets by limiting their survival, proliferation, activation, and cytokine production. Monitoring these subsets could help to evaluate the efficacy and safety of DMF treatment.

Figure 1. DMF treatment reduced circulating CD4⁺ T cells, CD8⁺ T cells, NKT cells and B cells

PBMCs were isolated from peripheral blood of patients with RRMS. Cells were stained with 7AAD and antibodies against TCRαβ, CD4, CD14, CD8, CD19, CD56, washed and then analyzed on a flow cytometer. (A) Typical staining profile of an RRMS patient without DMF treatment. Labels above FACS profiles indicate the gated population being analyzed. Percentages noted on the dot plots are in relation to the gated population. (B) Total absolute cell numbers (Abs #) /mL of blood of the indicated subsets were calculated based on clinical complete blood cell counts and the percentages of the indicated subsets as determined by flow cytometry. Absolute numbers of different lineages were compared between untreated (UNT) RRMS patients (n=18), and patients after 4-6 months (4-6M, n=20) or patients after 18-26 months (>18M) of DMF treatment (n=18). Scatter plots show individual patient data as dots and the median of each group as a line. (C) Longitudinal samples (n=9) were used to track changes in lineages over time within the same patient. Samples were obtained before (Pre), after 4-6 months (4-6M) and after 18-26 M (>18M) months of DMF treatment. P values from Friedman test with Dunn's multiple comparisons for pair-wise time points are shown above the data for each cell type group. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 2. DMF Treatment reduced memory T cells

PBMCs were isolated from blood of patients and stained with antibodies specific to surface markers: CD3, CD4, CD45RO, CD45RA and CCR7 in order to analyze subsets of memory cells. (A) Representative profile depicting the frequencies of memory and non-memory phenotypes among CD4⁺CD3⁺ T cells (top panels) and CD3⁺CD4^- T cells (lower panels) from an RRMS patient prior to DMF treatment, and after 4-6 months and >18 months of continuous DMF treatment (B) Frequencies of memory (CD45RA^-/CD45RO⁺) fractions as a percentage of CD3⁺CD4⁺ (upper panels) and CD3⁺CD4^- (lower panels) T cells in cross-sectional (left-hand panel) untreated (UNT, n=18), 4-6 month DMF-treated (4-6M, n=20), and 18-26 month (>18M, n=18) DMF-treated groups; as well as longitudinal cohort (n=9, right hand panel) (C) Frequencies of naïve Tn (CD45RO^-CCR7⁺), central memory Tcm (CD45RO⁺CCR7⁺), effector memory Tem (CD45RO⁺CCR7^-), and CD45RA⁺ effector memory Temra (CD45RO^-CCR7^-) of CD3⁺ CD4⁺ (upper panels) and CD3⁺CD4^- (lower panels) T cells in three cross-sectional patient groups. Scatter plots show data for individual patients as dots and the median of each group as a line. P-values from Kruskal-Wallis ANOVA with Dunn's multiple comparison tests are shown above the data for each cell type. (D) Tn, Tcm, Tem, and Temra cells as a percentage of CD3⁺CD4⁺ (upper panels) or CD3⁺CD4^- (lower panels) T cells in the longitudinal cohort at the three time points. P values from Friedman test with Dunn's multiple comparisons for pair-wise time points are shown above the data for each cell type. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 3. DMF reduced the T cell activation status

PBMCs were isolated from blood of patients and stained with antibodies against CD3, CD4, CD45RO, CCR7 and CD69 to assess activation status in T cells. From left to right, activation was shown as frequencies of CD69⁺ cells of total and corresponding Tem, and Temra subsets at three different time points (UNT: n=18; 4-6M: n=20, >18M: n=18) for CD4⁺ (upper panel) and CD4^- (lower panel) T cells. Scatter plots show data for individual patients as dots and median of each group as a line. P-values from Kruskal-Wallis ANOVA with Dunn's multiple comparison tests are shown above the data for each cell type. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 4. DMF specifically reduced Th1 and Th17, particularly those expressing CXCR3 and CD161, but increased Th2 cells

PBMCs from patients were stained with antibodies against CD3, CD4, CXCR3, CCR6, CCR3, and CD161 to evaluate frequencies of CXCR3⁺ (Th1), CCR6⁺ (Th17) of CD4⁺ T cells (UNT: n=15; 4-6M: n=19; >18M: n=18). The CCR3⁺ (Th2) of total CD4⁺ T cells were also analyzed (UNT: n=11, 4-6M: n=20, >18M: n=18). These frequencies were plotted as dots at three time points with the median of each group as a line in three cross-sectional groups (A) or in the longitudinal cohort (B). DMF effects on the frequencies of CD4⁺ T cells expressing CD161and CXCR3 with (C) or without (D) CCR6 in cross sectional (left-hand panels) and longitudinal (middle panels) studies, and corresponding correlation (right-hand panels) with frequencies of CD4⁺ T cells producing IFNγ but not IL-17. (E) IFNγ (left-hand panel) and IL-17 (middle panel) production by CD4⁺ T cells was analyzed in our longitudinal study following stimulation of PBMCs with PMA and ionomycin for 6 hours in the presence of brefeldin A. Cells were then stained with fluorochrome-conjugated antibodies against CD3, CD4, followed by intracellular staining of antibodies against IFNγ, IL-17A and IL-17F. IL-4 (right-hand panel) production was measured from ex vivo PBMC without further in vitro stimulation since it was clearly detectable (Suppl Fig 1B). PBMC were incubated in complete RPMI for 24hrs with brefeldin A added in the last 6 hours of culture. Cells were then stained with fluorochrome-conjugated antibodies against CD3, CD4, followed by intracellular staining with antibodies against human IL-4. P-values from Kruskal-Wallis ANOVA with Dunn's multiple comparison tests are shown above the data for each cell type group. *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.

Figure 5. Direct Effect of DMF on T cells in vitro.

PBMCs derived from 8 healthy donors were isolated and cultured in the absence or presence of 10μM or 100μM DMF for 48 hours. (A) DMF-induced T cell apoptosis in vitro. Cultured PBMCs were harvested, washed and surface stained with fluorochrome-conjugated antibodies against CD3, CD8, CD4, as well as 7AAD and Annexin V. The frequency of apoptotic cells (AnnexinV⁺7AAD^-) among total CD3⁺ CD4⁺ (left) and CD3⁺ CD4^- (right) T cells in the vehicle control and10μM or 100μM DMF-treated groups are shown. (B) DMF suppressed T cell activation in vitro. PBMCs from healthy donors were cultured in the presence or absence of DMF for 48 hours on plates coated with anti-CD3 and soluble anti-CD28 antibodies to activate T cells. The percentage of activated T cell blast (CD3⁺ FSC-A^hi) was shown. (C) DMF-induced reduction of ROS level in activated T cells. Plate-bound anti-CD3/soluble anti-CD28 activated PBMC cultures from healthy donors with or without DMF were harvested and stained with antibodies against CD3, CD4, followed by staining of ROS with CM-H₂DCFDA. The level of ROS on CD4⁺ (upper) and CD4^- T cell (lower) blasts (CD3⁺ FSC-A^hi) are expressed as arbitrary units (AU) of mean fluorescence intensity. We also repeated the statistical calculations with the patient showing most reduction removed, the data still showed significant reductions (p<0.05). (D) DMF inhibited T cell proliferation in vitro. T cells from healthy donors were activated in the presence or absence of DMF for 48hours on plates with immobilized anti-CD3 and soluble anti-CD28 antibodies. After culture, cells were surfaced stained with CD3, CD4 and CD69, followed by intracellular staining of Ki67 to measure cell proliferation. The frequency of Ki67⁺ cells among CD69⁺ population of CD4⁺ (upper panel) and CD4^- (lower panel) T cells are shown. (E) DMF reduced CCR7 expression on activated T cells in vitro. T cells were activated in the presence or absence of DMF for 48 hours on plates with immobilized anti-CD3 and soluble anti-CD28 antibodies. After culture, cells were surfaced stained with fluorochrome-conjugated antibodies against CD3, CD4 and CCR7. CD3⁺ T blasts were gated and CCR7 geometric mean fluorescence intensity (MFI; expressed in arbitrary units (A. U.)) was compared among vehicle control, 10μM DMF and 100μM DMF groups, respectively. P values from Wilcoxon signed rank test for the comparisons between control vs. 10 μM and control vs. 100 μM are shown above the data for each cell type group.*P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant.