The T cell response to interleukin 10 alters cellular dynamics and paradoxically promotes CNS autoimmunity

Abstract

IL10 is a critical anti-inflammatory cytokine, deficiency of which leads to spontaneous autoimmunity. Yet therapeutically administered or ectopically expressed IL10 may either suppress or promote disease. Distinct lineage-specific activities may explain IL10’s contradictory effects. To dissect the T cell-specific response to IL10 during organ-specific autoimmunity, we generated mice with a selective deletion of IL10Rα in T cells and analyzed its impact in an autoimmune model, experimental allergic encephalomyelitis (EAE). Surprisingly, the T-cell response to IL10 increased EAE severity. This did not result from altered T cell functional potential; T cell cytokine profile was preserved. IL10 also diminished the proliferation of T cells in situ within the target organ, an effect that would be expected to restrain disease. However, IL10 acted cell autonomously to sustain the autoreactive T cells essential for immunopathogenesis, promoting their accumulation and distorting the regulatory and effector T cell balance. Indeed, in chimeric mice and after adoptive transfer wild type T cells showed a competitive advantage over cells deficient in IL10Rα. Therefore T cell specific actions of IL10 can support autoimmune inflammation, and this appears to result from an overall increase in the long term fitness of pathologic T cells. Lineage-restricted, disease-promoting activities of IL10 should be considered in the therapeutic manipulation of the IL10 pathway.

Introduction

Interleukin 10 is critical in setting immune response amplitude (1–3). Its genetic ablation or inhibition leads to spontaneous colitis and exacerbation of autoimmune and inflammatory diseases. Polymorphisms in the IL10 gene or its receptor are associated with autoimmune susceptibility (4,5). In infections, IL10 protects against immunopathology mediated by overexuberant immune responses. Despite IL10’s considerable therapeutic potential, immunomodulatory agents that act by manipulating the IL10 pathway are not currently available.

Although IL10’s anti-inflammatory activities are well established, its actions are not exclusively immunosuppressive. It promotes B-cell growth and differentiation, and, by inhibiting apoptosis, tumor cell and lymphocyte survival (6–10). Localized overproduction of IL10 in pancreatic islets in transgenic (Tg) mice leads to islet destruction and diabetes (11). Its overexpression by tumors may promote either tolerance or rejection (12,13). Efficacy of pharmacologically administered IL10 in experimental allergic encephalomyelitis (EAE) varied with timing and mode of administration, and could even exacerbate disease (14–17). Clinical trials of IL10 failed to show benefit in Crohn’s disease (18).

Clarifying the disparate though predominantly anti-inflammatory activity of IL10, and deciphering how it may be optimally manipulated for therapy will require a dissection of its cell type-specific actions in disease states. Although T cells express only low levels of IL10R, they are potentially an important target. Studies of IL10Rβ-deficient animals can be complicated by the non-exclusive use of this receptor chain by other cytokines, particularly the IFNλs (19), but have implicated the T cell response to IL10 in disease susceptibility (20,21). Recent reports have alternatively used mice overexpressing a dominant negative IL10Rα or conditionally deficient in the leader sequence (exon 1) of IL10Rα (22–24). These identified a protective role for IL10 in intestinal inflammation by its selective suppression of pathologic Th17 responses. Studies of CD8 T cell response during infection further indicated a direct role for IL10 in the acquisition of memory, and therefore in the capacity of T cells to maintain long-term immunity (25).

Here we assess the impact of IL10 signaling into T cells in a pathophysiologically distinct autoimmune model, EAE. EAE results from immune mediated CNS demyelination orchestrated by self-Ag-specific Th1 and Th17 cells (26). Naturally produced IL10 is protective (27). In prior analyses, we demonstrated that regulatory T lymphocytes specifically re-targeted against the autoreactive T cells responsible for EAE were highly effective as a cellular immunotherapeutic. The activity of the regulatory cells was IL10 dependent, implicating direct actions of IL10 on T cells in disease abatement (28,29).

To better explore the T cell response to IL10, we developed a new mouse model conditionally deficient in the cytokine-binding domain (exon 3) of IL10Rα, the IL10 specific component of the IL10R. Mice with selective deletion of IL10Rα on T cells (denoted IL10Rα^Tdel) showed no detectable abnormalities in the absence of disease induction. Paradoxically, IL10Rα^Tdel mice demonstrated diminished EAE, indicating that while IL10 overall reduces EAE severity, it acts on T cells to increase their autoimmune potential. Our analyses demonstrate that the absence of IL10 responsiveness leads to unaltered T subset maturation, but a cell intrinsic effect of IL10 on Teff cellular dynamics during the autoimmune response; IL10 inhibited autoreactive CD4 T cell expansion while promoting their persistence. The net result of these contrasting activities was to help sustain the autoreactive T cell response. These results therefore describe a distinct IL10-mediated program that selectively acts on T cells to promote autoimmunity, highlight the contextual complexity of IL10 signaling, and provide insight into how cell-specific actions of this globally anti-inflammatory cytokine may enhance an inflammatory disease.

Materials and Methods

IL10Rα conditionally deficient mice

A 12.6 kb segment incorporating IL10Rα exon 1–5 was isolated from a C57BL/6-derived BAC, and a FRT flanked neo^r gene introduced into intron 3 and loxp sites added flanking exon 3 by recombineering (30) (Fig. 1a). This was introduced by homologous recombination into the C57BL/6-derived ES26.2 embryonic stem (ES) cell line, with correct targeting ascertained by Southern. ES cells were injected into C57BL/6J-Tyr^c-2J blastocysts to generate recombinant mice. These were bred with B6.Cg-Tg(ACTFLPe)^9205Dym/J mice to delete the neo^r leaving the loxp-flanked IL10Rα exon 3. Exon 3 incorporates much of the IL10 interacting domain and its deletion leads to an additional frameshift in downstream sequence.

Figure 1. Conditional deletion of the mouse IL10Rα locus.

(A) Upper diagram maps construct used for homologous recombination. A Frt-flanked neo^r cassette is located in intron 2, and loxP sites inserted flanking exon 3. Lower diagram shows the IL10Rα locus after deletion of the neo^r cassette in recombinant mice, leaving a minimally altered gene containing a single Frt site and loxP sites surrounding exon 3. (B) CD4 and CD8 T cells, and B cells were flow cytometrically sorted from IL10Rα^Tdel or WT splenocytes and LN cells. DNA was isolated and the quantity of IL10Rα exon 3 measured by qPCR using either of 2 oligonucleotide primer sets. Quantities calculated by the ΔΔCt method are plotted relative to the quantity observed in WT CD4 T cells. (C) WT or IL10Rα^Tdel T cells, stimulated for 4 days with αCD3/28, were stained with IL10Rα-specific Ab. Histograms show staining intensity for IL10Rα (light gray line) or isotype control (dark filled) for each population. (D) T cells were activated as in (C) and then treated with IL10. Phosphorylated Stat3 was measured by intracellular staining in IL10 treated cells (light gray line) or control cells untreated with IL10 (dark filled) for each population. (E) Equal numbers of purified CD4 T lymphocytes from IL10Rα^Tdel or WT mice were stimulated with the 1 µg/ml αCD28 and the indicated concentration of αCD3. Cells were pulsed at 72 h with [3]H-thymidine and incorporation measured 18 h later. The modest difference in proliferation seen at the 0.2 µg/ml concentration was not seen in additional experiments. (F) Purified CD4 T cells were stimulated with αCD28 with or without 2 µg/ml αCD3. At 48 h, concentrations of IL2, IFNγ, IL17, IL4, and IL10 were measured in cell free supernatant. (G) CD4⁺GFP-Foxp3⁺ Treg were flow cytometrically purified from IL10Rα^Tdel or WT splenocytes and LN cells, and co-cultured at the indicated Treg:Target cell ratio with WT CD4⁺GFP-Foxp3⁻ target cells in the presence of αCD3/CD28. Proliferation was measured by [3]H-thymidine incorporation. Representative of two or more experiments.

Mice and reagents

MOG_35–55 peptide (MEVGWYRSPFSRVVHLYRNGK) was synthesized and HPLC purified by the St. Jude Hartwell Center for Biotechnology. C57BL/6J, B6.Cg-Tg(ACTFLPe)^9205Dym/J, B6.129S2-Tcra^tm1Mom/J (TCRα^−/−), C57BL/6J-Tyr^c-2J, B6.SJL-Ptprc^a Pepc^b/BoyJ (CD45.1), and B6.PL-Thy1^a/CyJ (Thy1.1) mice were obtained from The Jackson Labs, and Tg(Cd4-cre)^1Cwi from Dr. P. Brindle. Foxp3^tm2Ayr (GFP-Foxp3) mice were the generous gift of Dr. A. Rudensky and bred >10 generations with C57BL/6J mice. WT mice controls consisted of IL10Rα^fl/fl mice or, when needed due to needs for congenic strains, matched mice lacking the IL10Rα^fl/fl allele. Studies were approved by the St. Jude Institutional Animal Care and Use Committee.

IL10Rα^Tdel mouse phenotyping and qPCR

Mouse genotyping amplified genomic DNA with primers specific for Cre (5’-GGCCCCATGGCATCCAATTTACTGACCGTACAC-3’ and 5’-TCGCTCGAGGTGATCGCCATCTTCCAGCAG-3’), floxed and wt IL10Rα alleles (5’-CTCCATACGAGAGGTCTTGGG-3’ and 5’-GCTAACTGCCCTGGGACTCCC-3’), and GFP (5’-TGCCTCTGACAAGAACCCAATGC-3’ and 5’- GGCAGGGGGTTCAAGGAAGAAG-3’). For quantitative analysis of lineage specific deletion, CD4 T, CD8 T, and B cells were isolated by flow cytometric sorting from IL10Rα^Tdel and WT (IL10Rα^fl/fl CD4-Cre⁻) mice, and genomic DNA isolated. IL10Rα exon 3 was amplified using forward primer 5’-CCGGGCAGTGGACAACA-3’ and reverse primer 5’-CCACTGTGAAGCGAGTCTCAGT-3’ (primer set 1) or forward primer 5’-AACCTGGAATGACATCCATATC-3’and the same reverse primer (primer set 2) using real time PCR. Ct data was compared with GAPDH as an internal control, and this was normalized to Ct data from simultaneously analyzed WT CD4 T cells using the comparative Ct approach (2^−ΔΔCt).

Flow cytometry and cell sorting

Analytic cytometry was performed on a Becton-Dickinson FACSCalibur or LSRII and analyzed with FlowJo software (Tree Star, Inc.). Sorting was performed with a Mo-Flo (Dako) or Reflection (iCyt) cytometer. Antibodies targeting CD4, TCRβ, CD3, CD11c, CD11b, CD205, CD8, B220, CD45R, Gr-1, F4/80, CD5, CD44, CD25, CD19, TCRγδ, NK1.1, 45Rb, CD62L, CCR2, CCR5, CD11, CXCR3, VLA-4, CD69, CD45.1, CD45.2, IL10R, and RatIgG2b were from BD Bioscience. Anti-Thy1.1 and Foxp3 were from eBioscience.

Intracellular Staining

For cytokine ICS, isolated lymphocytes were washed and stimulated with PMA (50 ng/ml) and ionomycin (1 µg/ml) for 4 h in the presence of 10 µg/ml Brefeldin A (Sigma), labeled with CD4 and TCRβ antibodies, fixed and permeabilized (Fixation Buffer, eBioscience), then labeled with anti-IFNG, IL10, and IL17 antibodies. ICS for Foxp3 was performed using the eBioscience Foxp3 staining kit. For phospho-Stat3, d4 activated T cells were stimulated with 10 ng/ml rIL-10 or medium only for 40 minutes at 37°C, fixed with 1% formaldehyde for 10 min at 37°C, washed, permeabilized with ice-cold 90% methanol for 30 min on ice, labeled with anti-P-Stat3 (Y705, Cell Signaling), washed, and stained with F(ab’)₂ goat anti-rabbit IgG (Invitrogen).

EAE induction and clinical evaluation

EAE was induced by subcutaneous immunization with 100 µg of MOG_35–55 peptide in 100 µl complete Freund’s adjuvant containing 0.4 mg M. tuberculosis H37RA (Difco). 200 ng pertussis toxin (List Biological Laboratories) was administered i.v. on days 0 and 2. Clinical scoring was: 1, limp tail; 2, hind limb paresis or partial paralysis; 3, total hind limb paralysis; 4, hind limb paralysis and body or front limb paresis or paralysis; 5, moribund.

Histopathology Scoring

Euthanized mice were perfused with 10% buffered formalin (Thermo Scientific), brain and spinal cords left in place within the skull and vertebral columns and post-fixed in 10% buffered formalin for >24 h, decalcified in formic acid (TBD-2 Decalcifier, Thermo Scientific), and cut into transverse sections of brain, brainstem, and cervical, thoracic and lumbar spinal cord. Histopathologic examination of hematoxylin and eosin sections were carried out in a blinded fashion, with sections evaluated in staggered order. Each histological lesion (mononuclear perivascular cuffing, gliosis, vacuolization/demyelination of axons, and neutrophilic infiltration) was scored independently according to predetermined criteria: 0 = lesions absent, 1 = minimal inconspicuous lesions, 2 = mild conspicuous lesions, 3 = prominent multifocal moderate lesions, 4 = marked coalescing lesions. For each animal, the individual lesion scores in all regions were totaled to determine an overall EAE severity score.

Cell Isolation and enumeration

Mice were sacrificed, lymph nodes and spleens removed, the circulation perfused with 4°C PBS, and brain and spinal cord removed. CNS lymphocytes were isolated by density centrifugation as described (31), counted on a hemacytometer and quantitatively analyzed by flow cytometry.

T cell proliferation

Splenic, DLN, and CNS cells were isolated, and total cells or magnetic bead isolated CD4 T cells stimulated as indicated at 5×10⁴ – 1 × 10⁵ cells per well. Cells were pulsed with 1 µCi [³H]thymidine after 72 h, and then harvested 18 h later on filtermat for scintillation counting. Samples were analyzed in triplicate.

Cytokine production

Lymphocytes were stimulated as above. Culture supernatants were collected at 24 or 48 h and analyzed for IL-2, IL4, IL10, IFNG, and IL-17 by Milliplex® multiplex bead assay.

Proliferation suppression assay

Flow cytometrically isolated native T cells (CD4⁺CD45RbhiCD44lo) were seeded at 5×10⁴ cells per well. Isolated GFP-Foxp3⁺ Treg were added at the indicated ratio with 0.5 µg/ml αCD3 and irradiated TCR Cα^−/− splenic APC. 1µCi 3[H]thymidine was added after 48 h and incorporation measured at 60–72 h.

In vivo proliferation

Mice were injected i.p. with 150 µl BrdU (10 mg/ml) in sterile 1xDPBS. After 16–20 hours, lymphocytes were isolated, stained with antibodies to cell surface markers, fixed and permeabilized with Cytofix/Cytoperm Buffer (BD), treated with DNase (300 µg/ml) at 37°C for 1 h, and then stained with anti-BrdU-APC (3D4, BD), and analyzed by flow cytometry.

In vivo Treg depletion

Mice were injected i.p. with 1 mg purified PC61 mAb or isotype control (IgG₁) at day -3. On day 0, depletion of CD25⁺ T cells was confirmed by analyzing CD4⁺CD25⁺ and CD4⁺Foxp3⁺ cells in peripheral blood samples and EAE was induced.

Hematopoietic chimeras and lymphocyte co-transfers

For hematopoietic chimeras, bone marrow cells were obtained from femurs and tibias of WT and IL10Rα^Tdel mice expressing the indicated congenic markers, washed, mixed in equal quantities, and 10×10⁶ cells injected i.v. into lethally irradiated (900 rad) TCRα^−/− recipients. For lymphocyte co-transfers, single cell suspensions of erythrocyte lysed LN and splenic cells from WT and IL10Rα^Tdel mice expressing the indicated congenic markers were prepared. Cells were mixed in approximately equal quantities and injected i.v. into CD45.1 congenic mice.

Statistics

Standard deviations or errors and confidence intervals were calculated with Excel® or PRISM® software. Significance between 2 groups was calculated by 2-tailed t-test. When >2 groups were present in an assay, significance was determined by 1-way, 2-tailed ANOVA using a Bonferroni correction for multiple comparisons. A p<0.05 was considered significant. For multiple comparisons, significance is only shown relative to the control group unless otherwise indicated.

Results

Characterization of mice with T cell deficiency in IL10 response

We generated C57BL/6 mice bearing a floxed IL10Rα exon 3 (IL10Rα^fl/fl; Fig. 1a) and bred these with mice expressing CRE selectively on T cells (CD4-Cre). Sorted IL10Rα^Tdel T but not B cells showed loss of genomic IL10Rα exon 3 by qPCR (Fig. 1b). IL10R expression, measured by surface staining, was absent from IL10Rα^Tdel T cells, but expressed at WT levels on B cells, macrophage, and dendritic cells (Fig. 1c and data not shown). IL10 signals through STAT3 (1), and IL10Rα^Tdel T cell STAT3 phosphorylation was absent to IL10 (Fig. 1d).

IL10Rα^Tdel mice did not display spontaneous systemic disease, clinically or at necropsy. No differences were observed in their proportions of DN1–4, DP, or SP thymocyte subsets, CD4⁺, CD8⁺, CD4⁺Foxp3⁺ (Treg), NK-T, and memory/naive TCRβ⁺ cells, TCRγδ⁺ cells, NK cells, CD19⁺B220⁺, CD5⁺, and regulatory CD5⁺CD1d⁺ B cells, myeloid, lymphoid, and plasmacytoid DCs, macrophage, or granulocytes, either in LN or spleen (Suppl. Fig. 1). To functionally assess the IL10Rα^Tdel T lymphocytes, CD4⁺ T cells were purified and stimulated with titrations of αCD3. Proliferative response was equivalent to that of cells from matched CD4-Cre⁻ IL10R^fl/fl controls (wild type, WT; Fig. 1e). No differences in the production of representative Th0 (IL2), Th1 (IFNG), Th2 (IL4), Tr1/Th2 (IL10), or Th17 (IL17) cytokines were identified (Fig. 1f). The IL10Rα^Tdel mice were further bred with GFP-Foxp3 knock-in mice for intravital demarcation of Treg. Flow cytometrically purified IL10Rα^Tdel and WT Treg demonstrated equivalent capacity to suppress naïve T cell proliferation (Fig. 1g). Therefore unmanipulated IL10Rα^Tdel mice appear normal through clinical observation, gross and histologic analysis, phenotypic characterization, and in vitro functional assessments.

Diminished EAE severity in IL10Rα^Tdel mice

Considering the documented role of IL10 in restraining EAE and other forms of autoimmunity, and of T-specific IL10 in suppressing intestinal inflammation, we anticipated that IL10Rα^Tdel mice would develop more severe EAE than WT controls. To the contrary, after immunization with myelin oligodendrocyte glycoprotein (MOG)_35–55, they showed consistently less severe disease (Fig. 2a). Cumulative and peak disease was decreased in the IL10Rα^Tdel cohorts, though not time to initial disease symptoms (Fig. 2b and Suppl. Table SI). All mortality was in the WT group. Therefore, though globally immunoregulatory, IL10 acts focally on T cells to promote MOG-induced EAE.

Figure 2. EAE development in IL10Rα^Tdel and WT mice.

(A) Clinical disease course is shown. Disease was induced with MOG_35–55 at day 0. Mean clinical score ± SEM is plotted. N=5 per cohort (B) A summary of results from 8 independent experiments, with 5–14 mice per cohort in individual experiments, is shown. Measures of disease activity include mean integrated disease score (area under the curve, AUC), mean peak disease score (Max Dis), and mean time to initial symptoms (Init Dis) for individual experiments. Ratio of value for IL10Rα^Tdel to WT mice is plotted. Mean of values is indicated by the central line, and 95% confidence interval for means by brackets to the right of this. Additional data is provided in Supplemental Table SI.

Diminished infiltrates and increased Treg/Teff ratios in IL10Rα^Tdel mice

Pathologically distinct forms of EAE can be distinguished histologically. To determine if the altered disease severity was associated with an altered immunopathologic process, we reviewed the histopathology of 6 WT and 7 IL10Rα^Tdel mice with diverse disease scores. Forebrain, brain stem, and cervical, thoracic, and lumbar spinal cord were assessed for gliosis, parenchymal vacuolation and demyelination, neutrophil infiltrate, and mononuclear cuffing. Linear regression modeling of histologic scores showed a positive correlation with clinical disease (Fig. 3a–e). However, slope and intercept for individual parameters and cumulative score did not differ between WT and IL10Rα^Tdel CNS. Therefore, WT and IL10Rα^Tdel mice develop a qualitatively similar disease that differs in magnitude of severity.

Figure 3. Histologic analysis of CNS in IL10Rα^Tdel and WT mice with EAE.

Forebrain, brain stem, and cervical, thoracic, and lumbar spinal cord from mice d 20–22 after disease induction were blindly assessed and scored for gliosis, parenchymal vacuolation and demyelination, neutrophil infiltrate, and mononuclear cuffing. Cumulative histologic scores across location and parameter (A) or for individual parameters across location (B–E) were tallied and are plotted. Linear regression modeling is overlayed and p values for slope and intercept comparing IL10Rα^Tdel and WT mice is indicated.

To quantify differences in the T cell response, CD4⁺ Teff and Foxp3⁺ Treg were enumerated in diseased CNS and spleen 14 and 21 days after induction. At the earlier time point, a significantly increased number of Teff was present in WT compared with IL10Rα^Tdel CNS, consistent with their more severe disease (Fig. 4a). This increase, however, was not associated with a parallel increase of CNS-infiltrating Treg, resulting in a diminished Treg:Teff ratio in WT compared with IL10Rα^Tdel CNS. We and others have demonstrated that Treg are potently inhibitory in MOG-EAE (28,31,32), and the Treg:Teff ratio is a critical indicator of activity. Therefore, T cell response to IL10 augments early Teff infiltration in the CNS while decreasing the proportion of Treg.

Figure 4. CNS T cell infiltrates in IL10Rα^Tdel and WT mice.

T cell infiltrates were quantified in the CNS of IL10Rα^Tdel and WT mice with early EAE (A; d14) and more advanced disease (B; d21). Upper panels display absolute numbers of CD4⁺Foxp3⁻ (Teff) and CD4⁺Foxp3⁺ T cells isolated per CNS (combined brain and spinal cord), and proportion of Foxp3⁺ T cells among the CD4⁺TCR⁺ population. Percent Foxp3⁺ T cells among splenocytes is shown for comparison. Plots show results from individual mice (dots) and mean values (bars). Lower panels show representative flow cytometric analyses of gated CD4⁺TCR⁺ cells. *, p<0.05; **, p<0.01. Data is cumulative from 2 experiments. Scores of analyzed mice were: d14 WT, 3,3,3,3,3,3,3,3,3,3,3,2,2,1; d14 IL10Rα^Tdel, 4,3,2,2,2,2,1,1,1,1,1,1,0,0; day21 WT, 3,3,3,3,3,2,2,2,2,1,1,1,1; d21 IL10Rα^Tdel, 2,2,2,2,2,2,1,1,1,1,1,1,0,0,0.

At the d21 time point, the magnitude of the infiltrate in both WT and IL10Rα^Tdel mice was substantially increased (Fig. 4b). An increase in both Teff and Treg numbers was evident in WT mice, however the growth of the Treg population exceeded that of Teff, leading to an increased Treg:Teff ratio compared with d14. A rapid expansion of CNS Treg during EAE has been previously described (33). In IL10Rα^Tdel mice, infiltrating Teff and Treg numbers were overall decreased relative to WT mice, reflecting their milder disease. The percentage of Treg was increased relative to d14, however the magnitude of this increase was not as pronounced as in the WT mice, and the proportion of Treg did not significantly differ from WT mice as it had at the earlier time point. Therefore the kinetics of CNS T cell accumulation differs between WT and IL10Rα^Tdel mice, with IL10Rα^Tdel mice showing overall decreased total T cell accumulation early and late and increased proportions of Treg early in disease.

Similar cytokine profiles and migratory phenotype of WT and IL10Rα^Tdel T cells

A mixed Th1 and Th17 response fosters the development and progression of MOG-EAE (26,34). Blockade of Th17 differentiation ameliorates disease, and direct actions of IL10 have been shown to suppress Th17 cells (22,23). An increased Th17 response would have been expected to increase disease severity in IL10Rα^Tdel mice, whereas disease was decreased. Nevertheless, to evaluate for altered Teff subset differentiation, we measured CD4⁺ T cell cytokine profiles directly by intracytoplasmic staining. Similar proportions of CNS or splenic CD4⁺ T cells producing IFNG, IL17, and IL10 were detected in WT and IL10Rα^Tdel mice (Fig. 5a–c). MOG-induced cytokine production, including IL2, IL4, IL10, IFNG, and IL17 was also similar in WT and IL10Rα^Tdel T cells (Fig. 5d). Therefore, IL10’s actions on T cells did not detectably skew subset responses during EAE.

Figure 5. Cytokine profiles and proliferation of IL10Rα^Tdel and WT T cells.

(A–C) Intracellular cytokine staining. CNS (A) or splenic (B) T cells were isolated from mice with EAE (d21), and intracellular staining for IFNG, IL17, and IL10 performed directly ex-vivo after a brief stimulation with PMA/ionomycin. Percent of cells positive for the indicated cytokines is plotted. Significant differences were not observed between the IL10Rα^Tdel and WT cells here or in similar analyses at d14–15 (not shown). Disease scores of mice analyzed in the plot were: WT, 3,3,2,2,2,1; IL10Rα^Tdel, 2,2,1,1,1,0,0. (C) Representative plots from the staining performed in (A, B) are shown. (D) Antigen induced cytokine production. Splenic T lymphocytes were purified from mice with EAE (d 16). Equal numbers were cultured without or with 100 µg/ml MOG_35–55 and secreted IL2, IFNG, IL17, IL4, and IL10 measured at 48 h. No significant differences were seen between the IL10Rα^Tdel and WT cells. Disease scores of mice analyzed in the plot were: WT, 3,3,3; IL10Rα^Tdel, 3,3,2. For proliferation studies, splenocytes were isolated 14 (E, early disease; disease scores of analyzed mice in the plot were WT: 2,1,1,1 and IL10Rα^Tdel, 2,1,1,1) or 21 d (F, progressive disease; disease scores of analyzed mice in the plot were WT: 3,3,3 and IL10Rα^Tdel, 3,3,2) after induction and stimulated with the indicated concentration of MOG_35–55 peptide. Antigen specific proliferation was measured at 72 h by [3]H-thymidine incorporation. Means±1 s.d. are plotted for individual mice. (G) Alternatively, CD4⁺ T cells were purified from the CNS of mice with EAE (d21) and equal numbers stimulated with no or 100 µg/ml MOG_35–55 in the presence of irradiated splenic APCs from non-immune WT mice. Proliferation was measured as above. Mean proliferation of individual mice is plotted. Results were not significantly different for IL10Rα^Tdel versus WT mice. **, p<0.01; ***, p<0.001 for IL10Rα^Tdel versus WT cohorts (F) or unstimulated versus stimulated (G). Data is representative of two or more experiments.

As a supplemental explanation for the diminished number of T cells in CNS infiltrates of IL10Rα^Tdel mice with EAE, we considered the possibility that they had decreased migratory potential. To examine this, levels of CCR2, CCR5, CXCR3, CD11a, and VLA-4, markers associated with T cell CNS migration, were measured on DLN, splenic and CNS T cells in mice with EAE (Suppl. Fig. S2). No differences were seen between WT and IL10Rα^Tdel populations. Therefore, defective response to IL10 does not impact phenotypic markers associated with CNS localization.

Ex vivo IL10Rα^Tdel T cell proliferation

To further define IL10’s role in shaping response magnitude, IL10Rα^Tdel and WT splenocytes were isolated from mice with EAE, stimulated with MOG_35–55, and their autoAg-specific proliferative response measured. At d14 after disease induction, responses were similar, indicating that an equivalent proportion of T cells were capable of recognizing and being effectively stimulated by antigen (Fig. 5e). By d21, however, the T cell responses diverged, with less proliferation seen among IL10Rα^Tdel splenocytes (Fig. 5f). Therefore, the peripheral T cell response to IL10 varies with time after disease induction. IL10 helps sustain the autoAg-specific response later in disease.

MOG-specific T cells are concentrated in the CNS of mice with EAE. Though fewer T cells were present in the CNS infiltrates of IL10Rα^Tdel compared with WT mice (Fig. 4), on a per cell basis, purified CD4⁺ T cells from WT or IL10Rα^Tdel mice demonstrated equivalent responsiveness (Fig. 5g, NS). In summary, CNS infiltrates in IL10Rα^Tdel mice are of diminished magnitude and early in disease bear an increased proportion of Treg compared with WT mice. With disease progression, fewer peripheral MOG-reactive IL10Rα^Tdel T cells are detected. However, at the site of autoimmune destruction, a similar fraction of IL10Rα^Tdel and WT T cells are capable of responding to MOG, and their cytokine profiles are equivalent.

Increased in situ Treg and Teff proliferation in IL10Rα^Tdel mice

The diminished late in vitro MOG-specific T cell proliferative response of IL10Rα^Tdel compared with WT T cells together with the decreased number of infiltrating T cells in the IL10Rα^Tdel CNS may indicate a diminished expansion of autoreactive T cells in situ in the IL10Rα^Tdel mice. To further examine this, we measured in situ T cell proliferation by BrdU labeling. The decreased absolute T cell numbers in the CNS of IL10Rα^Tdel mice would be expected to be associated with a decreased proportion of cycling (BrdU⁺) cells. In the spleen, only a small proportion of T cells will be MOG-specific. At d14, WT and IL10Rα^Tdel splenic CD4⁺ T cells showed similar BrdU incorporation, both for Foxp3⁺ and Foxp3⁻ cells (Fig. 6a, b). AutoAg-specific cells will, however, concentrate in the CNS. Here, IL10Rα^Tdel Treg and Teff showed significantly increased BrdU uptake compared to WT, indicating that, in contrast with our prediction, these were proliferating more rapidly at the site of autoimmunity. Analysis of total numbers of BrdU⁺ cells in the CNS did not show any difference between IL10Rα^Tdel and WT mice, either for Teff or Treg (data not shown). Therefore, despite an early increase in proliferation rate in IL10Rα^Tdel mice in situ and absolute numbers of proliferating T cells that did not differ compared with WT mice, there is a diminished accumulation of CNS T cells in the IL10Rα^Tdel mice.

Figure 6. In situ proliferation of Treg and Teff in IL10Rα^Tdel and WT mice.

(A) A pulse of BrdU was administered to IL10Rα^Tdel and WT mice with EAE (d14). At 20 h, CNS and splenic T cells were isolated and BrdU incorporation measured by ICS. Percent BrdU positive cells among CD4⁺TCR⁺Foxp3⁻ or CD4⁺TCR⁺Foxp3⁺ populations is shown. Graphs show results from individual mice (dots) and mean values (bars). (B) Sample flow cytometry plots of BrdU staining from individual mice in (A). (C) Mice were analyzed 21 d after EAE induction as in (A). ***, p<0.001. Data is cumulative from 2 experiments. Scores of analyzed mice were: d14 WT, 3,3,3,3,3,3,3,3,2,2; d14 IL10Rα^Tdel, 4,3,2,2,2,1,1,1,1,1,0; day21 WT, 3,3,3,2,2,2,2,1,0; d14 IL10Rα^Tdel, 3,2,2,1,1,1,1,1.

By d21, as clinical disease crested, rates of BrdU incorporation were diminished and similar between the two cohorts in both the CNS and spleen (Fig. 6c). Therefore, in early but not late EAE, CNS-infiltrating IL10Rα^Tdel Teff and Treg are cycling more rapidly than WT cells. This increased proliferation is consistent with the documented anti-proliferative effects of IL10 in vitro(35) and after anti-CD3 treatment in vivo(22), however, is incongruously associated with diminished total CNS IL10Rα^Tdel Teff numbers and an increased Treg proportion that would be expected to suppress proliferation (Fig. 4). It is also incongruously associated with the diminished proliferative response to MOG among IL10RαTdel splenocytes at d21 in vitro (Fig. 5f).

Retained IL10Rα^Tdel disease resistance after Treg depletion

One possible explanation for this inconsistency is that the increased proportion of Treg in the IL10Rα^Tdel CNS did not act by limiting Teff proliferation, but instead promoted Teff elimination either directly by cytolysis or indirectly by decreasing overall inflammation. To evaluate the independent effects of IL10 on Teff cells, we depleted Treg by treatment with the CD25-specific PC61 Ab prior to EAE induction. Consistent with prior observations (36), PC61 eliminated virtually all CD25⁺ and most Foxp3⁺ cells (Suppl. Fig. 3a–c). As previously described, Treg depleted mice developed more severe EAE than mice treated with control Ab (Fig. 7a). However, this was similarly true for IL10Rα^Tdel and WT mice. Indeed, IL10Rα^Tdel mice treated with PC61 demonstrated substantially less severe disease than WT mice treated with even control Ab. Like in figure 2, both cumulative and maximal disease scores were greater in WT than IL10Rα^Tdel mice (Fig. 7b). Further, the extent of this increase was similar in PC61 and control Ab-treated cohorts. This supports the presence of a Treg-independent effect in governing disease susceptibility.

Figure 7. EAE in Treg-depleted IL10Rα^Tdel and WT mice.

WT or IL10Rα^Tdel mice were treated with PC61 αCD25 Ab or isotype control on d -3 prior to EAE induction and elimination of Treg confirmed in the blood on d 0. (A) EAE was induced in WT and IL10Rα^Tdel mice and clinical disease monitored. Clinical score tracked longitudinally in mice treated with PC61 or control Ab showed that PC61 treatment led to increased disease severity in both cohorts of mice. (B) The ratio of mean cumulative disease scores (area under the curve, AUC), maximal disease scores (Max), or time to initial disease (Init) for IL10Rα^Tdel and WT mice from 3 independent experiments is plotted. Though PC61 treatment led to overall more severe disease, the ratio of each parameter between IL10Rα^Tdel and WT mice was not significantly altered by PC61 treatment.

Competitive analysis of Treg and Teff cell dynamics

To isolate the cell intrinsic effects of IL10 on T cells, we competitively analyzed mixed populations of IL10Rα^Tdel and WT T cells. Because the cells share an identical environment, we could thereby exclude indirect effects of T cell IL10 signaling that otherwise may influence outcome. We generated chimeric mice in which CD45.2⁺ GFP-Foxp3 IL10Rα^Tdel bone marrow cells were mixed with equal numbers of CD45.2⁻ GFP-Foxp3 WT cells and transplanted into lethally irradiated T-deficient TCRα^−/− hosts. This allowed delineation of the origins of the T cell populations with the CD45.2 congenic marker. EAE development in the chimeric mice was modestly delayed compared to unmanipulated mice, with first symptoms at d16–20 after induction. Cell populations were monitored by blood sampling, and after sacrifice in the spleen and CNS.

Similar proportions of IL10Rα^Tdel and WT CD4⁺ T cells, Foxp3⁺ or Foxp3⁻, were present in the blood at different time points prior to the onset of clinical symptoms (d0, 8, 15; Fig. 8a–c). However, in chimeric mice with symptomatic disease, a significant increase in WT versus IL10Rα^Tdel Teff cells was observed in the CNS. A similar trend was apparent for Treg, though this did not achieve significance. Foxp3⁻ Teff from diseased mice were further segregated into memory/activated (CD44hi) and naïve (CD44loCD62Lhi) populations. The memory/activated populations were selectively skewed toward WT T cells, and this was true in both the spleen and CNS (Fig. 8d). In contrast, the naïve populations showed a WT/IL10Rα^Tdel ratio similar to that of cells prior to immunization (pre-immune blood, 1.2±0.4; naïve spleen, 1.2±0.4; memory spleen, 2.1±0.9; naïve CNS, 1.4±0.7; memory CNS, 2.7±1.5). Therefore, activated but not naïve WT T cells outcompete corresponding IL10Rα^Tdel cells in the chimeric mice.

Figure 8. T cell competition in bone marrow chimeric mice and after co-transfer.

CD45.2⁺ IL10Rα^Tdel and CD45.2⁻ WT bone marrow cells were mixed in equal quantities and transplanted into lethally irradiated TCRα^−/− recipients. The proportion of CD45.2⁺ versus CD45.1⁺ Foxp3⁺ and Foxp3⁻ cells was determined ~8 weeks post-transplant and EAE was induced (d0). Blood was sampled on d8 and d15 prior to clinical disease development. Mice with EAE were sacrificed on d21 and 35 and proportions of splenic and CNS CD45.2⁺ T cells measured. (A, B) Plots show results from individual mice (dots) and mean values (bars). (C) Sample flow cytometry plots demonstrating discrimination of IL10Rα^Tdel and WT T cell populations from different tissue sources. CD4⁺TCR⁺GFP-Foxp3⁻ gated cells are in the upper row and Foxp3⁺ cells in the lower row. (D) Ratio of gated memory (CD44hi) and naïve (CD44loCD62Lhi) T cells in the spleen and CNS derived either from IL10Rα^Tdel or WT populations is plotted. (E–G) IL10Rα^Tdel (CD45.1⁻CD45.2⁺Thy1.1⁻) and WT (CD45.1⁻CD45.2⁺Thy1.1⁺) splenocytes were admixed and co-transferred into WT (CD45.1⁺CD45.2⁻Thy1.1⁻) hosts. EAE was induced 2 d later. The proportion of transferred IL10Rα^Tdel and WT CD4⁺TCR⁺Foxp3⁻ cells was measured in the blood prior to EAE induction, and on d7. At d16, the mice were sacrificed and proportions of WT and IL10Rα^Tdel Foxp3⁻ T cells in the spleen and CNS measured. Because of the small number of cells detected, reliable measurements of the Foxp3⁺ populations was not possible. (E) Graph shows results from individual mice (dots) and mean values (bars). (F) Sample flow cytometry plots show discrimination of the WT (Thy1.1⁺) and IL10Rα^Tdel (Thy1.1⁻) cells from gated transferred T cell (CD45.1⁻CD45.2⁺CD4⁺TCR⁺GFP-Foxp3⁻) populations. (G) Data from splenic populations was further discriminated into memory (CD44hi) and naïve (CD44loCD62Lhi) subgroups and analyzed as in (E). Because of the small percent of transferred cells and the preponderance of memory T cells in the CNS (>90%), segregated analysis of naïve T cells was not possible at that location. *, p<0.05; **, p<0.01, ***, p<0.001 versus d0 and d7 blood or for memory versus naive; +++, p<0.001 versus d7 blood and NS versus d0 blood by Bonferroni-corrected ANOVA. Data is representative of 2 experiments.

As an alternative approach to verify these findings and competitively evaluate the cell intrinsic effect of IL10 on T cell responses, one excluding thymic generation of new T cells, mature WT and IL10Rα^Tdel cells were mixed and co-transferred into WT recipients. Host and transferred populations could be distinguished by congenic markers (host, CD45.1⁺CD45.2⁻Thy1.1⁻; WT: CD45.1⁻CD45.2⁺Thy1.1⁺; IL10Rα^Tdel, CD45.1⁻CD45.2⁺Thy1.1⁻). EAE was induced and cells monitored in the blood or, after sacrifice, in organs. Due to the small numbers of transferred cells detected, event numbers precluded effective monitoring of transferred Treg, however Foxp3⁻ T cells could be readily quantified. Prior to disease induction (d0), an approximately equal number of WT and IL10Rα^Tdel CD4⁺ T cells was seen in the blood (Fig. 8e, f). This ratio was maintained at d7. By d16, after EAE development, a significantly increased proportion of WT T cells was observed in the spleen. A more dramatic increase was evident within the CNS (WT/IL10Rα^Tdel ratio: pre-immune blood d0, 0.68±0.29; d7 blood 0.83±0.03; d16 spleen, 1.7±0.3; d16 CNS, 8.7±1.2). Skewing of the splenic response toward WT cells was specific to the activated/memory population (Fig. 8g; naive d16, 1.1±0.1; memory d16, 2.3±0.4). A corresponding analysis of differential skewing among naïve and activated/memory populations was not possible in the CNS due to the paucity of naïve T cells at this location (>90% memory) and small number of transferred cells detected. Nevertheless, these results confirm findings in the bone marrow chimeras, and indicate that IL10 signaling provides a cell intrinsic competitive advantage to WT Teff cells during EAE.

Discussion

IL10 controls inflammation, whether infectious, autoimmune, or toxic. Its production, proportionate to the level of inflammation, provides negative feedback against excessive responses (37). IL10 lacks recognized pro-inflammatory activities in T cells. Despite this, we show that IL10 promotes autoimmune inflammation in EAE. Our data support a cell intrinsic role for IL10 in the long-term sustenance of the autoreactive T cells essential to this disease as a mechanism.

IL10 can regulate genes inhibiting cell cycle progression. In accordance with this, we observe that WT T cells proliferate less rapidly than IL10Rα^Tdel cells in the CNS in vivo. Implicitly, IL10 is acting to restrain autoreactive T cell expansion. The decreased rate of cell cycling in the WT CNS contrasts with the equivalent proliferation of CNS T cells measured in vitro. This indicates that environmental conditions present in the WT CNS inhibit WT T cell cycling, and that these conditions are not present when T cells are purified and re-stimulated with Ag and feeder cells in culture. The decreased cell cycling in the CNS of WT mice is also incongruously associated with an increased accumulation of CNS-infiltrating T cells and more severe clinical disease. This cannot be solely explained by the decreased proportion of Treg seen early in disease in WT mice. IL10Rα^Tdel mice depleted of Treg still showed less severe EAE than similarly treated WT mice. More severe disease would be expected if IL10’s effects were wholly Treg dependent. Moreover, WT T cells outcompete IL10Rα^Tdel T cells in mixed chimeras, both bone marrow chimeras and after the co-transfer of mature WT and IL10Rα^Tdel T cells, indicating that the effect of IL10 on T cell dominance is cell intrinsic. Therefore, IL10 supports the Teff response in EAE despite its suppression of T cell proliferation in situ, and this is independent of altered Treg quantities or function.

Our data are also inconsistent with IL10 altering the functional profile of autoreactive T cells. A classical pattern of clinical and histologic disease was seen in the IL10Rα^Tdel mice, patterns that may be altered by perturbations in T cell cytokine production (38), and cytokine profiles were unaffected. Likewise expression of relevant chemokine receptors and addressins failed to differentiate WT and IL10Rα^Tdel populations, implying that the cells had similar capacity to localize to the inflamed CNS.

The balanced inhibition of T cell subsets here contrasts with what has been observed in studies of mucosal immunity in mice bearing T cells expressing a dominant negative IL10Rα or with a selective deficiency of IL10Rα on Treg (22–24). In those, a specific effect on Th17 accumulation was documented. Indeed, we have observed a similar skewing of the IL10Rα^Tdel Th17 response in a model of sepsis (R. Alli, T. Geiger, unpublished observations), indicating a common role for IL10 in regulating Th17 production. This suggests that the impact of IL10 on T cells is highly contextual. Indeed, although in vivo data is limited, IL10 has documented activity against multiple T cell lineages and can directly inhibit Th1 IFNG secretion and Th2 IL4 and IL5 production (35,39–41). It may be anticipated that the magnitude and quality of its effects will depend on the quantity of IL10 produced as well as quality and quantity of alternative signaling inputs received by targeted cells. For instance, we may speculate that the relatively large amounts of IL10 constitutively produced in the intestines or during sepsis, by inducing SOCS3, may promote the relatively selective inhibition of IL6 signaling and Th17 differentiation (42,43). In EAE, differential regulation or altered quantities of IL10 may promote less selective pathways.

That IL10 did lead to the increased accumulation of autoreactive T cells despite their diminished expansion rate indicates that it helped to sustain the autoreactive T cell population. This is consistent with in vitro data indicating that IL10 promotes T cell survival with the withdrawal of alternative pro-survival cytokines, and the documented regulation of anti-apoptotic genes by IL10 in other cell types (44,45). More recently a role for IL10 in promoting CD8 T cell memory has been described (25). Response persistence is fundamental to autoimmunity, and we may speculate that the cytokine environment supporting the long-term maintenance of reactive T cells during autoimmunity may be similar to that promoting the generation of long-lived memory cells. T cells proliferate up to several times a day after stimulation (46), and even subtle changes in proliferation or survival rates, compounded over time, can dramatically influence cell dominance. Implicitly, the combined impact of IL10’s contrasting effects on T cell expansion and persistence integrated over time determines outcome. Cues that differentially alter these effects may thereby alter the net impact of IL10 on T cell-mediated immune responses.

Presumably, increased attrition of cells unable to respond to IL10 results from their apoptotic death. Rapid clearance of apoptotic cells in vivo prohibits direct measurement of cell death rates in situ and was therefore not directly assessed. Nevertheless, we did observe diminished MOG-induced proliferative response among peripheral IL10Rα^Tdel splenocytes and a selective competitive advantage among memory-activated WT T cells in mixed chimeras, indicating that the autoreactive T cells are not sustained in the absence of T-specific IL10 responsiveness. Cytokine withdrawal or activation-induced apoptosis would seem the most likely causes. Because apoptotic T cells are rapidly cleared in vivo, measuring ongoing rates of apoptosis is technically difficult, and analyses of isolated populations may indicate susceptibility to apoptosis with in vitro manipulation rather than in situ kinetics. Further breeding of IL10Rα^Tdel mice onto backgrounds with altered apoptosis susceptibilities will be important to directly evaluate the hypothesis that IL10 alters cell death rates, and interrogate mechanism.

Global IL10 deficiency leads to increased EAE severity (27). That T-specific IL10 unresponsiveness attenuates disease indicates that the impact of IL10 on T cells is counterbalanced by more potent activities on other cell types. Indeed, myeloid-derived APCs and effector cells express higher levels of IL10R than T cells, and these and glial population that more directly mediate myelin damage would seem probable IL10 targets. Dissecting responsible cell types is possible using conditional systems, such as that described here, which were not previously available. Nevertheless, that T cell directed IL10 has an ultimately pathologic effect does emphasize the importance of directed targeting by therapeutically administered IL10, and may explain the conflicting results on the efficacy of pharmacologically administered IL10. Our results argue that it is the balance of opposing effects of IL10 both within individual cell types, where cell preservation may coincide with decreased proliferation, as well as across lineages that defines ultimate outcome.